close

Вход

Забыли?

вход по аккаунту

?

Soil moisture observations using L-, C-, and X -band microwave radiometers

код для вставкиСкачать
NOTE TO USERS
Page(s) not included in the original manuscript and are
unavailable from the author or university. The manuscript
was scanned as received.
134
This reproduction is the best copy available.
®
UMI
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
SOIL MOISTURE OBSERVATIONS USING L-, C-, AND X- BAND
MICROWAVE RADIOMETERS
by
John Dennis Bolten
Bachelor of Science
West Virginia University, 1998
Masters of Science
University of South Carolina, 2001
Submitted in Partial Fulfillment of the Requirements
for the Degree of Doctor of Philosophy in the
Department of Geological Sciences
University of South Carolina
2005
Major Professor
Committee Member
Chairman, Examimhg Committee
Dean of the Graduate School
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
UMI Number: 3201309
INFORMATION TO USERS
The quality of this reproduction is dependent upon the quality of the copy
submitted. Broken or indistinct print, colored or poor quality illustrations and
photographs, print bleed-through, substandard margins, and improper
alignment can adversely affect reproduction.
In the unlikely event that the author did not send a complete manuscript
and there are missing pages, these will be noted. Also, if unauthorized
copyright material had to be removed, a note will indicate the deletion.
®
UMI
UMI Microform 3201309
Copyright 2006 by ProQuest Information and Learning Company.
All rights reserved. This microform edition is protected against
unauthorized copying under Title 17, United States Code.
ProQuest Information and Learning Company
300 North Zeeb Road
P.O. Box 1346
Ann Arbor, Ml 48106-1346
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
To Mom, Dad, and brothers Bolten
ii
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Acknowledgments
I would first like to thank my advisor, Dr. Venkat Lakshmi, for his unmatched
encouragement and enthusiasm. It has been a wonderful journey working with him. He is
an excellent teacher, mentor, and friend. I am amazed by his gift for ‘synergistically’
mixing work and play. Also, he never hesitated to remind me to “relax”. His confidence
in my abilities was always apparent. He has shown me that anything is possible if you
aim high; whether it is academic, on the basketball court, or a pushup contest in the
middle o f soil physics class.
I am also thankful to Dr. Tom Jackson for helping me gain a clearer
understanding o f my research topic. He has devoted many hours to this research during
field campaigns and our numerous discussions. I thank him for his involvement and look
forward to working with him in the future.
I am grateful for the participation o f Dr. Claudia Benitez-Nelson and Dr. John
Jensen. They provided many insightful ideas and comments regarding my study and
career. It has been inspiring to discuss my research with them. Dr. Benitez-Nelson
deserves special thanks for her support during the last year when things were particularly
hectic. Her positive attitude and calmness was always welcomed.
There are many people who have stood by me through the years. I owe many
thanks for their companionship and support. They are the best friends I could ask for, and
I will never forget them. I would like to extend my deepest gratitude to Mark
iii
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Woodworth for always, always being there for me. I am amazed by his genuine kindness,
and somewhat frightened by his madness regarding organic geochemistry and mountain
bikes. Martha McConnell has been a great friend and is thanked for her support and
generosity.
I would like to acknowledge the fellow students in my lab who have shared this
experience with me. It was a joy to work with them and I look forward to our future
collaboration and long friendships.
My dog Doctor is my ultimate companion. He cannot be thanked enough for
sticking with me during all the late nights and long weekends. I only wish he had caught
just one squirrel in the courtyard before I finished my Ph.D.
My father and brothers are noted for always taking an interest in my studies, and
never failing to remind me that I am, after all, the baby o f the family.
A special dedication goes to my mother.
“Now this is not the end. It is not even the beginning o f the. end. But it is, perhaps, the end
o f the beginning”- Sir Winston Churchill
iv
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
ABSTRACT
Soil Moisture Observations Using L-, C-, and X- Band Microwave Radiometers
John Dennis Bolten
The purpose o f this thesis is to further the current understanding o f soil moisture
remote sensing under varying conditions using L-, C-, and X-band. Aircraft and satellite
instruments are used to investigate the effects o f frequency and spatial resolution on soil
moisture sensitivity. The specific objectives o f the research are to examine multi-scale
observed and modeled microwave radiobrightness, evaluate new EOS Aqua Advanced
Microwave Scanning Radiometer (AMSR-E) brightness temperature and soil moisture
retrievals, and examine future satellite-based technologies for soil moisture sensing.
The cycling o f Earth’s water, energy and carbon is vital to understanding global
climate. Over land, these processes are largely dependent on the amount o f moisture
within the top few centimeters of the soil. However, there are currently no methods
available that can accurately characterize Earth’s soil moisture layer at the spatial scales
or temporal resolutions appropriate for climate modeling.
The current work uses ground truth, satellite and aircraft remote sensing data
from three large-scale field experiments having different land surface, topographic and
climate conditions. A physically-based radiative transfer model is used to simulate the
observed aircraft and satellite measurements using spatially and temporally co-located
v
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
surface parameters. A robust analysis o f surface heterogeneity and scaling is possible due
to the combination o f multiple datasets from a range of microwave frequencies and field
conditions. Accurate characterization of spatial and temporal variability o f soil moisture
during the three field experiments is achieved through sensor calibration and algorithm
validation. Comparisons o f satellite observations and resampled aircraft observations are
made using soil moisture from a Numerical Weather Prediction (NWP) model in order to
further demonstrate a soil moisture correlation where point data was unavailable. The
influence o f vegetation, spatial scaling, and surface heterogeneity on multi-scale soil
moisture prediction is presented.
This work demonstrates that derived soil moisture using remote sensing provides
a better coverage o f soil moisture spatial variability than traditional in-situ sensors.
Effects o f spatial scale were shown to be less significant than frequency on soil moisture
sensitivity. Retrievals o f soil moisture using the current methods proved inadequate under
some conditions; however, this study demonstrates the need for concurrent spacebome
frequencies including L-, C, and X-band.
Major Professor: Dr. Venkat Lakshmi
vi
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Table of Contents
Page
Dedication................................................................................................................................... ii
Acknowledgements.................................................................................................................. iii
Abstract........................................................................................................................................ v
Table o f contents...................................................................................................................... vii
List of tables..............................................................................................................................
xi
List of figures..........................................................................................................................
xiii
Chapter 1. Introduction..................................................................................................... 1
1.1. Motivation for research........................................................................................
1
1.2. Research questions................................................................................................ 4
1.3. Theoretical background........................................................................................ 4
1.4. Overview o f field experiments........................................................................... 6
1.4.1.
Southern Great Plains 1999 Experiment (SGP99).......................
7
1.4.2.
Soil Moisture Experiments 2002 (SMEX02)................................
8
1.4.3.
Soil Moisture Experiments 2004 (SMEX04)................................
9
1.5. Outline o f dissertation chapters........................................................................ 10
1.6. References............................................................................................................
vii
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
12
Chapter 2. Airborne Passive Microwave Remote Sensing Observations of Soil
Moisture at L- and C- band.............................................................................................. 15
Abstract......................................................................................................................... 15
2.1. Introduction..........................................................................................................
16
2.2. Data and intercomparison methods..................................................................
18
2.2.1. Southern Great Plains 1999 Experiment (SGP99).......................... 18
2.2.2. Passive / Active L- and S- Band Sensor (PALS)...............................20
2.2.3. C- Band Polarimetric Scanning Radiometer (PSR/C)......................21
2.3. Instrument comparison.......................................................................................
22
2.3.1. Co-location and comparison.............................................................. 22
2.3.2. Sensitivity and re-sampling analysis................................................ 23
2.3.3. Radiative transfer m odel.................................................................... 24
2.3.4. Heterogeneity....................................................................................... 25
2.4. Instrument comparison results........................................................................... 27
2.4.1. Co-location and field site comparison............................................. 27
2.4.2. Spatial heterogeneity......................................................................... 28
2.5. Conclusions and discussion..............................................................................
31
2.6. References............................................................................................................. 33
Chapter 3. Soil Moisture Estimates From Aircraft and Satellite Microwave Sensors
During the Soil Moisture Experiments 2002 (SMEX02)............................................
60
Abstract......................................................................................................................... 60
3.1. Introduction........................................................................................................... 62
viii
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
3.2. SM EX02.............................................................................................................. 64
3.2.1. Advanced Microwave Scanning Radiometer (AMSR-E).............. 67
3.2.2. Polarimetric Scanning Radiometer Instrument (PSR)...................... 69
3.2.3. Radiative transfer................................................................................... 70
3.2.4. Co-location and comparison techniques............................................. 73
3.3. Results..................................................................................................................
75
3.3.1 AMSR-E 10.65 GHz observations and simulations.......................... 76
3.3.2. PSR C/X observations and simulations............................................ 77
3.3.3. Comparison o f AMSR-E and PSR C/X ..........................................
78
3.4. Soil moisture prediction...................................................................................... 79
3.5. Heterogeneity....................................................................................................... 80
3.6. Radio Frequency Interference...........................................................................
81
3.7. Conclusions and discussion............................................................................... 82
3.8. References............................................................................................................. 83
Chapter 4. AMSR-E Observations During the Soil Moisture Experiments 2004
(SMEX04).......................................................................................................................... 114
4.1. Introduction.......................................................................................................
114
4.2. M ethods..............................................................................................................
115
4.2.1. SMEX04............................................................................................ 117
4.2.2. Topography........................................................................................ 117
4.2.3. Operational model moisture estimates..........................................
118
4.2.4. AMSR-E products............................................................................. 119
ix
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
4.3. Preliminary results............................................................................................ 120
4.4. Discussion........................................................................................................... 123
4.5. References........................................................................................................... 125
Chapter 5. Summary....................................................................................................... 134
5.1. References......................................................................................................... 141
Cumulative bibliography.................................................................................................. 143
x
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
List of Tables
Page
Table 2.1. Field site characteristics........................................................................................ 38
Table 2.2. PALS and PSR/C instrument characteristics..................................................... 39
Table 2.3. Statistics o f PALS and PSR/C brightness temperatures at the fieldand PSR- scales for all field sites.................................................................................... 40
Table 2.4. Inputs for the emission m odel.............................................................................. 41
Table 2.5. Mean and standard deviations o f the input parameters for spatial
Heterogeneity simulations............................................................................................... 42
Table 2.6. Range and standard deviations o f brightness
temperatures from all PALS (PSR/C-scale) and PSR/C channels.............................
43
Table 2.7. Summary o f statistics for PALS 1.4 GHz horizontally polarized footprints
falling entirely within a PSR/C footprint........................................................................ 44
Table 2.8. Means and standard deviations o f the modeled 1.4 GHz and 6.9 GHz
brightness temperature polarization differences........................................................... 45
Table 2.9. Cross-channel correlation coefficients o f brightness temperatures over the
SGP99 field sites for each day........................................................................................
46
Table 3.1. AMSR-E instrument characteristics...................................................................... 89
Table 3.2. PSR/C Instrument characteristics for SMEX02 region flights.......................... 90
xi
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Table 3.3. Radiative transfer model inputs for AMSR and PSR C/X simulations
91
Table 3.4. Observed SMEX02 regional mean daily statistics forinstrument and in situ
data....................................................................................................................................... 92
Table 3.5. R2 o f observed TB and near surface soil moisture............................................. 93
Table 3.6. Correlation statistics o f estimated and observed AMSR footprint volumetric
soil moisture....................................................................................................................... 94
xii
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
List of Figures
Page
Figure 2.1. PALS brightness temperatures for 6 flight lines were linearly interpolated
giving a smooth field o f brightness temperatures o f approximately 6 x 40
km in size (from Njoku et al. 2002).............................................................................. 47
Figure 2.2. PSR/C horizontally polarized 7.325 GHz brightness temperature maps
acquired on July 8-9, 11, 14, 15 and July 19, 1999(from Jackson et a l 2002).... 48
Figure 2.3. Comparison o f PALS 1.4 GHz horizontally polarized TB
resampled to field- and PSR/C-scale with individual 6.9 GHz horizontally
polarized TBover the field sites....................................................................................... 49
Figure 2.4. Relationship o f PALS 1.4 GHz horizontally-polarized brightness
temperatures at field- and PSR/C-scaleto in situ soil moisture................................. 50
Figure 2.5. Relationship between PALS and PSR/C brightness temperatures
and 0-5 cm in situ gravimetric soil moisture..................................................................
51
Figure 2.6. PALS and PSR/C brightness temperature data and in situ moisture data
plotted daily for site LW 23................................................................................................ 52
Figure 2.7. PALS and PSR/C brightness temperature data and
in situ moisture data plotted daily for site LW 24........................................................... 53
Figure 2.8. PALS and PSR/C brightness temperature data and in situ
moisture data plotted daily for site LW 25....................................................................... 54
xiii
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Figure 2.9. Modeled A T b verses soil moisture for L-band (1.4 GHz)
and C-band (6.9 GHz) channels for three values o f vegetation water
content wc, values at 55° incidence................................................................................. 55
Figure 2.10. Co-located PALS and PSR/C horizontally polarized brightness
temperatures throughout the entire region on July 9, 1999......................................... 56
Figure 2.11. Co-located PALS and PSR/C horizontally polarized brightness
temperatures throughout the entire region on July 11, 1999....................................... 57
Figure 2.12. 1.4 and 6.9 GHz channel comparison for single horizontally
polarized values which were located closest to the field site centers......................... 58
Figure 2.13. Modeled brightness temperature plotted with co-located PALS
(1.4 GHz) and PSR/C (6.9 GHz) observations.............................................................. 59
Figure 3.1. Location o f SMEX02 region.............................................................................
95
Figure 3.2. Mean observed AMSR 10.7 GHz average T b vs . observed
0-1 cm volumetric soil moisture...................................................................................... 96
Figure 3.3. Spatial distribution o f PSR 7.3 GHz Horizontally-polarized TB................... 97
Figure 3.4. Comparison o f mean observed and mean
simulated AMSR 10.7 GHz footprint-scale brightness temperatures......................... 98
Figure 3.5. Comparison o f simulated AMSR10.7 GHz horizontal and
vertical polarized TB with in situ soil moisture and observed TB................................
99
Figure 3.6. Mean estimated vs. observed average PSR 7.3 GHz T b ................................ 100
Figure 3.7. Mean estimated vs. observed average PSR 10.7 GHz TB.............................
xiv
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
101
Figure 3.8. Comparison of simulated PSR 10.7 GHz horizontal and vertical
polarized TB with in situ soil moisture and observed TB..........................................
102
Figure 3.9. Comparison of simulated PSR 7.3 GHz horizontal and
vertical polarized TB with in situ soil moisture and observed TB.............................
103
Figure 3.10. Spatial distribution of simulated and
observed PSR 7.3 GHz horizontally polarized TB....................................................... 104
Figure 3.11. Spatial distribution o f simulated and observed
PSR 10.7 GHz horizontally polarized TB.................................................................... 105
Figure 3.12. Comparison o f mean observed domain-scale
AMSR 10.7 GHz and PSR 10.7 GHz horizontally polarized TB.............................. 106
Figure 3.13. AMSR 10.7 GHz predicted soil moisture
vs. in situ 0-1 cm soil moisture..................................................................................... 107
Figure 3.14. PSR 7.3 GHz predicted soil moisture vs.in situ 0-1cm soil m oisture.... 108
Figure 3.15. Range of mean observed footprint-scale PSR 10.7 GHz horizontallypolarized Tb v s . observed AMSR 10.7 GHz horizontally polarized TB..................
109
Figure 3.16. PSR C/X horizontally polarized TB............................................................... 110
Figure 3.17. Mean o f PSR C/X horizontally-polarized TB within each AMSR-E
footprints..........................................................................................................................
Ill
Figure 3.18. Observed vs. simulated AMSR 6.9 GHz TB................................................. 112
Figure 3.19. Observed AMSR 6.9 GHz average TB vs. observed PSR
7.3 GHz average TB........................................................................................................
113
Figure 4.1. Digital Elevation Map o f SMEX04 study areas..........................................
128
xv
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Figure 4.2. Mapped AMSR-E estimated volumetric soil moisture
for August 5th, 2004 over NAME TIER 1 region......................................................
129
Figure 4.3. AMSR-E 1:30 AM estimates o f soil moisture for W alnut Gulch
Watershed, AZ plotted with mean in-situ precipitation and soil moisture, and
ECMWF estimated volumetric soil moisture.............................................................
130
Figure 4.4. AMSR-E 1:30 PM estimates of soil moisture for Walnut Gulch
Watershed, AZ plotted with mean in-situ precipitation and soil moisture, and
ECMWF estimated volumetric soil moisture.............................................................
131
Figure 4.5. AMSR-E 6.9 and 10.7 GHz vs. in situ soil moisture for 1:30 AM
overpass for Walnut Gulch Watershed, A Z ................................................................
132
Figure 4.6. 1:30 AM overpass AMSR-E soil moisture (- 0.09 cm3/cm3) vs.in situ soil
moisture for the Walnut Gulch Watershed from Junel to Sept 30th,2004............ 133
xvi
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
CHAPTER 1
INTRODUCTION
1.1. Motivation for research
Accurate observations o f land-atmosphere processes are essential for a better
understanding o f meteorological forcings and interactions such as precipitation,
infiltration, runoff, and river drainage. Soil moisture influences many processes that drive
regional and global climate. For example, soil moisture determines surface heat fluxes
and near-surface moisture gradients by partitioning rainfall into infiltration and runoff,
and contrails atmospheric effects and climate conditions (Shukla & Mintz 1982).
Non-uniform weather patterns, soil types, topography, and vegetation lead to an
inherent heterogeneous nature o f soil moisture distribution. This surface heterogeneity
makes accurate large-scale soil moisture observations difficult. There are many methods
used to calculate soil moisture: electrical resistivity, neutron probing, gamma ray, Time
Domain Reflectivity, and satellite/ground remote sensing. O f these, remote sensing has
been shown to be the only method capable o f providing time and cost effective soil
moisture estimation over large areas (Schmugge et al. 2002). Microwave remote sensing
(via satellite or aircraft), is capable o f large-scale soil moisture estimation; providing
spatial averages o f soil physical and hydraulic properties over lightly vegetated areas.
1
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Therefore, in order to improve the predictive capability o f large-scale coupled
hydrologic-meteorological models; the verification o f remotely sensed large-scale soil
moisture dynamics is essential (Brubaker & Entekhabi 1996, Jackson et al. 1996).
However, collection o f large-scale in situ soil moisture measurements needed for
validation o f these remote sensing estimates are difficult.
The current research is motivated by the need for a reliable soil moisture
algorithm that can be applied to current and future satellite-based soil moisture missions.
The following sections introduce the field and instrument data used to produce
quantitative and qualitative estimates of multi-scale, near-surface soil moisture. Data
were collected during three robust large-scale field experiments focused on field and
model validation and calibration o f three newly developed microwave remote sensing
instruments (airborne and satellite). By combining spatially and temporally co-located
field and remote sensing data, an in-depth analysis o f the effects o f microwave frequency,
surface heterogeneity, and scaling on soil moisture prediction is presented. Substantial
work includes the investigation of the influence o f spatial resolution on soil moisture
sensitivity. Observations at the field and satellite- scales are used in this analysis. The
application o f aircraft sensors provides a critical link between these spatial scales.
An excellent opportunity to evaluate the spatial and temporal variability of
estimated soil moisture and the problems associated with these methods is provided
through modeling o f the microwave emission using the coupled ground and remote
sensing data. A detailed analysis of the scaling, heterogeneity and vegetation effects on
modelled microwave emission will benefit future missions including NASA’s Hydros
2
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
(launch date-2010) (Entekhabi et al. 2004) and the European Space Agency Soil Moisture
Ocean Salinity (SMOS) mission (launch date-2007) (Kerr et al. 2001).
Further investigation o f soil moisture estimates is done through comparisons of
satellite-derived soil moisture products to an operational model-based soil moisture data
set. By comparing soil moisture derived from a Numerical W eather Prediction (NWP)
model to that from remote sensing sensors, temporal and spatial patterns o f soil moisture
patterns are evaluated. The operational data set used in this study is provided by the
European Centre for Medium-range Weather Forecasts (ECMWF). The ECMWF soil
moisture dataset consists of four soil layers at the T511 (~40km) spectral resolution
(Douville et al. 2000, van den Flurk et al. 2000). The ECMWF products are widely used
in climatology research (Drusch et al. 2004). However, evaluations with in situ
observations or model inter-comparisons are indispensable for a better understanding of
parameterization systems and improving reanalysis (Seuffert et al. 2004). The current
methodology builds upon previous efforts by incorporating the satellite-based Advanced
Microwave Scanning Radiometer (AMSR-E) observations in our retrieval methods over
terrain with pronounced topography and vegetation variation.
3
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
1.2. Research questions
This research work builds upon previous efforts in the field o f microwave remote
sensing by investigating three large-scale field studies in an attempt to answer the
following questions:
(1) Does lower frequency outweigh higher spatial resolution in applications o f passive
microwave radiometry fo r soil moisture sensitivity?
(2) Do measurements o f microwave brightness using L-, C-, and X-band provide
reliable estimates o f soil moisture under varying land use/land cover, climatic and
topographic regions?
1.3. Theoretical background
Remote sensing o f soil moisture is possible due to the large contrast in dielectric
constant o f dry soil (~ 4) and water (~ 80) and the resulting dielectric properties of soilwater mixtures (~ 4-40) in the microwave spectrum (Schmugge 1985). These large
differences in dielectric properties allow quantitative estimates o f soil moisture to be
provided using physically based expressions. In summary, microwave technology is the
only remote sensing method that is able to measure a direct response to the absolute
amount o f water in the surface soil under both bare soil and vegetated conditions.
Passive microwave remote sensing is based on the measurement o f thermal
radiation from the land surface in the centimeter wave band, and is largely determined by
the physical temperature and the emissivity of the radiating body (air/soil/water interface,
overlying vegetation) (Njoku 1977, Ulaby 1986). Modeling o f this thermal radiation
4
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
(emission) is possible through parameterization the land/atmosphere emission paths and
solving the radiative transfer equations, such as in (Njoku & Li 1999). Variations of this
model have been used to analyze ground, aircraft and satellite data, and have been shown
to be valid for frequencies in the range 1-20 GHz (Bolten et al. 2003, Narayan et al.
2004). In the current approach, the composition o f the soil (moisture content, bulk
density, sand and clay mass fractions) is used to determine the dielectric constant and
resulting emissivity from the effective soil moisture depth using the empirical mixing
model and calculations described in (Wang & Schmugge 1980, Dobson et al. 1985,
Peplinski et al. 1995). The emitting soil depth is assumed to be roughly one-tenth of a
wavelength in the medium (i.e. ~ 0.4 cm depth for the 7.3 GHz channel used in this
study) (Ulaby 1986). The Fresnel equations are used to estimate surface reflectivities for
each polarization and incidence angle over this surface, as derived in (Ulaby 1986).
Resulting reflectivities are then modified for rough surfaces by a height parameter h and
polarization mixing parameter Q (Choudhury 1979, Njoku & Li 1999). Vegetation is
represented in the model as a single scattering layer above the emitting soil surface,
parameterized by vegetation water content and empirically-derived values. Referred to as
the vegetation opacity, it is related to the vegetation water content wc, by:
bw
T‘ =
cos a0
where 0 is the incidence angle and the parameter b is thought to be approximately
proportional to frequency and depend weakly on vegetation type at low frequencies, but
may depend more on canopy structure at higher frequencies (such as 6-10 GHz used in
this study) (Jackson & Schmugge 1991, LeVine & Karam 1996, Wigneron et al. 2004).
The brightness temperature at the top of the vegetation layer is calculated for each
5
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
polarization p, as a function o f the aforementioned soil brightness temperature and
reflectivity with the addition o f vegetation opacity t, vegetation single-scattering albedo
co, effective soil temperature Ts, and vegetation effective temperature Tce:
TBp = T s ( 1 - rp )e~* + Tce(1 - «)(1 - ^
)(1 + r,* -* )
Soil moisture is calculated using an iterative, least-squares-minimization method.
In the iterative procedure, the volumetric soil moisture is adjusted to minimize the
weighted-sum of squared differences between measured and computed average
brightness temperatures. The estimated moisture for the (field site) location is that which
satisfies the minimization calculation. This moisture value represents the mean field-scale
soil moisture content and is later up-scaled for comparison with the remote sensing
instruments and regional-scales.
1.4. Overview of field experiments
The lower frequency ( 1 - 1 0 GHz) microwave region is considered to be optimal
for soil moisture remote sensing. Much effort has been made towards designing a future
satellite mission within these wavelengths having the temporal and spatial resolutions
required for soil moisture sensing. However, microwave measurements from space have
been limited by the physical limitations of large aperture antennas required to obtain
reasonable spatial resolutions. Airborne instruments having similar frequencies to current
and future satellite missions are being designed and tested. The current work uses
observations from a recently launched spacebome passive microwave instrument and two
airborne instruments, as described in the following sections.
6
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Three large-scale soil moisture field campaigns were conducted in 1999, 2002 and
2004 in the United States and Mexico. The objective of these experiments was to
represent the soil moisture dynamics within the study areas through the collection of
ground-based samples in conjunction with aircraft flights and satellite overpasses. Colocation o f these efforts (spatially and temporally) provided validation o f the remotely
sensed data. Each field experiment covered a range o f spatial scales and various ground
and remote sensing data including; surface soil moisture and temperature, soil roughness,
soil bulk density, vegetation water content, and precipitation. The advantage o f using data
from the three studies is that each study has unique spatial domain, vegetation cover, and
extensiveness o f remote sensing and ground data collected. Analysis o f these experiments
in union provided an insightful look into not only the progress o f the experiments
themselves, but also the application o f a general radiative transfer model in different
terrain. The following sections present a brief overview o f each field experiment used in
this study. More detailed descriptions of the individual experiment goals, methods and
remote sensing instruments are found in the chapters in which they are discussed.
1.4.1. Southern Great Plains 1999 Experiment (SGP99)
The SGP99 experiment included a variety o f airborne C-, S-, and L- band
microwave instruments to provide large-scale soil moisture mapping in the Little Washita
Basin (603 km2), near Chickasha, Oklahoma. Data acquisition during the study took
place between July 9th and 20th 1999. A description o f instruments used during SGP99
and
experiment
details
can
be
found
on
the
experiment
web
site
[http://daac.gsfc.nasa.gov/CAMPAIGN_DOCS/SGP99/index.html], Forest cover within
7
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
the watershed is very sparse and typically follows streams constituting a small portion of
the watershed (Njoku et al. 2002). The basin consists mostly o f rolling hills (maximum
relief is less than 200 m), rangeland and pasture. Within the watershed, the ground truth
data collection included eleven field sites (0.8 km x 0.8 km), chosen in five types o f landcover: rangeland, wheat, com, alfalfa and fallow; ranging in vegetation water content
from 0-7.18 kg m '2. Collected ground data applicable to this dissertation include:
gravimetric soil moisture (0 - 2.5 cm and 0 - 5 cm), surface roughness, soil bulk density,
and vegetation water content. A major rain event occurred on the third day (July 10th,
1999). As a result, a subsequent dry-down period was observed throughout the basin
resulting in a change in gravimetric soil moisture content from about 24.3% to 3.4% over
a period o f 8 days.
1.4.2. Soil Moisture Experiments 2002 (SMEX02)
In contrast to SGP99 (nominal vegetation water content < 2 kg m '2), the Soil
Moisture Experiment 2002 (SMEX02) was designed to extend the existing soil moisture
retrieval algorithms to areas o f moderate to heavy vegetation water content conditions (48 kg/m2) (Bolten et al. 2004). In addition, a large amount o f effort was focused on
providing intense ground tmth efforts and passive microwave aircraft programs designed
to support AMSR algorithm development and validation (Bindlish et al. 2005). Thus,
SMEX02 involved a wider range of vegetation cover, spatial extent and remote sensing
instruments. The study was conducted from June 25th through July 12th, 2002 in the
Walnut Creek watershed (-100 km2) in central Iowa. The terrain is undulating and the
land cover type for the watershed region is primarily agricultural with com and soybeans
8
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
being the major crops. SMEX02 provided unique conditions o f high initial biomass
content and significant change in biomass over the course of the experiment.
The W alnut Creek watershed has been studied extensively and is well
instrumented for in situ sampling of hydrologic parameters. Similar to SGP99, an
extensive in situ dataset consisting of: gravimetric soil moisture, surface temperature,
bulk density, surface roughness, vegetation water content, crop type and radiometer and
radar observations were collected. This dataset was used to construct the emission models
for interpretation of the remote sensing data as discussed in Chapters 2-5.
Notable radio frequency interference (RFI) from anthropogenic sources in the
region was observed in both the AMSR-E and PSR C/X data during SMEX02. These
effects are manifested by higher than expected brightness temperatures, particularly in the
ASMR-E 6.9 GHz frequency (Li et al. 2004). The degree of RFI effects was analyzed in
the current analysis by comparing observations from the interference-prone channels to
similar frequencies that were not affected.
1.4.3. Soil Moisture Experiments 2004 (SMEX04)
SMEX04 compliments the preceding experiments by focusing on satellite
footprint heterogeneity. The study included diverse levels o f topography and vegetation
and also strengthened the soil moisture components o f the North American Monsoon
Experiment (NAME) [http://www.joss.ucar.edu/name/]. NAME is an internationally
coordinated study aimed at determining the sources and limits o f predictability o f warm
season precipitation over North America. Soil moisture plays an important role in this
process by influencing convection and intraseasonal monsoon patterns. Improved climate
9
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
prediction on seasonal-to-interannual time scales depends on the ability to quantify the
initial states and forecast the evolution o f the surface forcing variables (e.g., sea surface
temperature and soil moisture). Thus, accurate characterization o f soil moisture spatial
and temporal variability is critical to NAME.
SMEX04 was conducted in August 2004 during the rainy season in southwestern
U. S. and northern Mexico. Two regional study sites (-5 0 km by 75 km) were established
in Arizona and Sonora, Mexico. The study consisted o f corresponding efforts similar to
SGP99 and SMEX02 including a system o f in situ soil moisture networks, and aircraft
and satellite remote sensing products. The extensive network o f in situ measurements and
remote sensing products allowed focused monitoring o f the core monsoon region as well
as on a regional continental scales.
1.5. Outline of dissertation chapters
The overall goal of this dissertation is to answer the previously mentioned
research questions by applying similar analytical methods to observations from three
remote sensing instruments and in situ observations over three different areas. Chapter 2
compares observations o f two airborne instruments o f L- and C-band during SGP99. The
influence of scale and frequency effects on soil moisture sensitivity is addressed in an
analysis o f simulated and observed brightness temperatures.
Chapter 3 uses data from the SMEX02 experiment to investigate satellite and
airborne observations in C-band. The same radiative transfer model applied to the SGP99
data is used to simulate the airborne and AMSR-E observations. Effects o f spatial
heterogeneity during SMEX02 are studied using a statistical analysis o f sub-footprint and
10
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
in situ observations and radio brightness. Comparisons o f estimated soil moisture using
remote sensing instruments and in situ observations are discussed.
Chapter 4 extends the microwave soil moisture remote sensing observations and
retrieval algorithms to more challenging vegetation and topographic conditions. AMSR-E
observations during SMEX04 are analyzed over rolling terrain near Tombstone, Arizona.
Soil moisture estimates from an operational data assimilation model provided by the
European Center for Medium Range Weather Forecasting (ECMWF) are used in this
analysis. These values are combined with in situ observations o f soil moisture and
precipitation to validate AMSR-E brightness temperature and soil moisture retrievals.
11
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
References
Bindlish R, Jackson TJ, Gasiewski AJ, Klein M, Njoku EG (2005) Soil moisture mapping
and AMSR-E validation using the PSR in SMEX02. Remote Sensing of
Environment
Bolten JD, Lakshmi V, Njoku EG (2003) Soil moisture retrieval using the passive/active
L- and S-band radar/radiometer. IEEE Transactions on Geoscience and Remote
Sensing 41:2792-2801
Bolten JD, Narayan U, Guijarro L, Lakshmi V (2004) Passive-Active Microwave Remote
Sensing o f Soil Moisture at Both L- and C-band: A Comparison o f Two Field
Experiments. Italian Journal of Remote Sensing 30/31:65-86
Brubaker KL, Entekhabi D (1996) Asymmetric recovery from wet versus dry soil
moisture anomalies. Journal of Applied Meteorology 35:94-109
Choudhury BJS, T. J.; Newton, R.W.;Chang, A.T.C. (1979) Effect o f surface roughness
on microwave emission o f soils. Journal of Geophysical Research 84:5699-5706
Dobson MC, Ulaby FT, Hallikainen MT, Elrayes MA (1985) Microwave Dielectric
Behavior o f Wet Soil .2. Dielectric Mixing Models. IEEE Transactions on
Geoscience and Remote Sensing 23:35-46
Douville H, Viterbo P, M ahfouf JF, Beljaars ACM (2000) Evaluation o f the Optimum
Interpolation and Nudging Techniques for soil moisture analysis using FIFE data.
Monthly W eather Review 128:1733-1756
Drusch M, Wood EF, Gao H, Thiele A (2004) Soil moisture retrieval during the Southern
Great Plains Hydrology Experiment 1999: A comparison between experimental
remote sensing data and operational products. Water Resources Research 40
Entekhabi D, Njoku EG, Houser P, Spencer M , Doiron T, Kim YJ, Smith J, Girard R,
Belair S, Crow W, Jackson TJ, Kerr YH, Kimball JS, Koster R, M cDonald KC,
O'Neill PE, Pultz T, Running SW, Shi JC, Wood E, van Zyl J (2004) The
hydrosphere state (Hydros) satellite mission: An earth system pathfinder for
global mapping o f soil moisture and land freeze/thaw. IEEE Transactions on
Geoscience and Remote Sensing 42:2184-2195
12
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Jackson TJ, Schmugge J, Engman ET (1996) Remote sensing applications to hydrology:
Soil moisture. Hydrological Sciences Joumal-Joumal Des Sciences
Hydrologiques 41:517-530
Jackson TJ, Schmugge TJ (1991) Vegetation Effects on the Microwave Emission of
Soils. Remote Sensing of Environment 36:203-212
Kerr YH, Waldteufel P, Wigneron JP, Martinuzzi JM, Font J, Berger M (2001) Soil
moisture retrieval from space: The Soil Moisture and Ocean Salinity (SMOS)
mission. IEEE Transactions on Geoscience and Remote Sensing 39:1729-1735
LeVine DM, Karam MA (1996) Dependence o f attenuation in a vegetation canopy on
frequency and plant water content. IEEE Transactions on Geoscience and Remote
Sensing 34:1090-1096
Li L, Njoku EG, Im E, Chang PS, Germain KS (2004) A preliminary survey of radio­
frequency interference over the US in Aqua AMSR-E data. IEEE Transactions on
Geoscience and Remote Sensing 42:380-390
Narayan U, Lakshmi V, Njoku EG (2004) Retrieval o f soil moisture from passive and
active L/S band sensor (PALS) observations during the Soil Moisture Experiment
in 2002 (SMEX02). Remote Sensing o f Environment 92:483-496
Njoku EG, Li L (1999) Retrieval of land surface parameters using passive microwave
measurements at 6-18 GHz. IEEE Transactions on Geoscience and Remote
Sensing 37:79-93
Njoku EG, W ilson WJ, Yueh SH, Dinardo SJ, Li FK, Jackson TJ, Lakshmi V, Bolten J
(2002) Observations of soil moisture using a passive and active low-frequency
microwave airborne sensor during SGP99. IEEE Transactions on Geoscience and
Remote Sensing 40:2659-2673
Njoku EGK, N. (1977) Theory for passive microwave remote sensing o f near-surface soil
moisture. Journal of Geophysical Research 82
Peplinski NR, Ulaby FT, Dobson MC (1995) Dielectric-Properties o f Soils in the 0.3-1.3Ghz Range. IEEE Transactions on Geoscience and Remote Sensing 33:803-807
13
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Schmugge T (1985) Remote sensing of soil moisture. In: Burt MGAaTP (ed)
Hydrological Forcasting. John Wiley and Sons, New York, p 604
Schmugge TJ, Kustas WP, Ritchie JC, Jackson TJ, Rango A (2002) Remote sensing in
hydrology. Advances in Water Resources 25:1367-1385
Seuffert G, Wilker H, Viterbo P, Drusch M, M ahfouf JF (2004) The usage o f screen-level
parameters and microwave brightness temperature for soil moisture analysis.
Journal of Hydrometeorology 5:516-531
Shukla J, Mintz Y (1982) Influence of Land-Surface Evapo-Transpiration on the Earths
Climate. Science 215:1498-1501
Ulaby FTM, R.K., Fung, A.K. (1986) Microwave remote sensing: Active and passivevoume scattering and emission theory, advanced systems and applications.
Dedham, MA: Artech House.
van den Hurk B, Viterbo P, Belajaars A, Betts A (2000) Offline validation o f the ERA40
surface scheme. ECMWF Tech Memo 295:42
Wang JR, Schmugge TJ (1980) An empirical model for the complex dielectric
permittivity o f soil as a function o f water content. IEEE Transactions on
Geoscience and Remote Sensing GE-23:35-46
Wigneron JP, Parde M, Waldteufel P, Chanzy A, Kerr Y, Schmidl S, Skou N (2004)
Characterizing the dependence of vegetation model parameters on crop structure,
incidence angle, and polarization at L-band. IEEE Transactions on Geoscience
and Remote Sensing 42:416-425
14
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
CHAPTER 2
AIRBORNE PASSIVE MICROWAVE REMOTE SENSING
OBSERVATIONS OF SOIL MOISTURE AT L- AND C-BAND
Abstract
The 1999 Southern Great Plains Experiment (SGP99) provided concurrent largescale observations o f L- and C-band radiometric brightness temperatures for a range of
soil moisture and vegetation conditions. Observations were collected using the
Passive/Active L- and S-band (PALS) and C-band Polarimetric Scanning Radiometer
(PSR/C) airborne sensors. The data provide new information on the sensitivities o f low
frequency measurements to soil moisture at spatial scales of 2300 m, 800 m, and 400 m.
Comparisons o f calculated and observed L- and C-band (1.4 and 6.9 GHz) brightness
temperatures with in situ validation data are made. The difference in footprint size and
sampling density o f the two instruments allowed the effect of sub-footprint heterogeneity
on the received radio brightness to be observed. Resampling o f PALS data to field- and
PSR/C-scales was used in conjunction with modeled emission to investigate the issue of
heterogeneity and its role in the interpretation of spatial scale variability. The PALS and
PSR/C data were shown to be comparable at all spatial resolutions. This work provides a
unique analysis o f C- and L-band microwave remote sensing that is relevant to current
(AMSR-E, C-band) and next generation (Hydros, L-band) spacebome radiometers.
15
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
2.1. Introduction
The need for readily available, large-scale soil moisture measurements has
increased in recent years. Soil moisture dominates biogeochemical cycles (Gulledge &
Schimel 1998), heat exchange (Bindlish et al. 2001, Basara & Crawford 2002) and the
infiltration rates (Wu et al. 1996, Koster & Milly 1997) at the land/atmosphere boundary,
all of which are driving forces in local and regional climate (Shukla & Mintz 1982,
Delworth & Manabe 1989). These relationships influence many disciplines such as
agriculture, meteorology and hydrology. Hence, soil moisture is an important variable for
understanding these relationships and predicting their impact and correlation in the
development of accurate circulation models, applicable on a global scale.
The inherent heterogeneous nature of soil moisture (a result o f weather patterns,
vegetation, soil types, vegetation and topography) makes accurate large-scale soil
moisture characterization difficult. O f the many current methods used to sense soil
moisture, electrical resistivity (Michot et al. 2003), neutron attenuation, gamma ray
attenuation (Fetter 1994), time domain reflectivity, and satellite/ground remote sensing
(Owe et al. 1988, Choudhury 1991, Famiglietti et al. 1999), remote sensing has been
shown to be the only method to provide time and cost effective soil moisture estimation
over large areas (Schmugge 1985, Owe et al. 1999). Microwave remote sensing (via
satellite or aircraft) permits large-scale soil moisture estimation, which in effect provides
spatial averages of soil physical and hydraulic properties over bare and lightly vegetated,
i.e., vegetation water content < 2.5 kg/m2, areas and estimates o f soil moisture and texture
properties. Therefore, remote sensing of large-scale soil moisture dynamics (using remote
sensing) and coupled model verification is potentially beneficial for improvement of
16
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
large-scale coupled hydrologic-meteorological models (Jackson et al. 1999, Bindlish et
al. 2001).
A number of studies have focused on the use of experimentally observed passivemicrowave remote sensing data to verify theoretical brightness temperature/soil moisture
model relationships for various target (land cover, soil roughness, soil moisture and
temperature) and sensor (frequency/wavelength, sensor platform: tower, truck, aircraft,
satellite) parameters (Dobson & Ulaby 1986, Jackson & Schmugge 1989, Wang et al.
1989, Schmugge et al. 1992, Hollenbeck et al. 1996, Laymon et al. 1999, Owe et al.
1999, Laymon et al. 2001, Jackson et al. 2002, Bindlish et al. 2003, Bolten et al. 2003,
Njoku et al. 2003). These efforts have demonstrated the sensitivity o f microwave
radiometry using vertical and horizontal polarizations at low microwave frequencies, i.e.
1-10 GHz, to soil type, roughness, vegetation type and soil moisture content in the top 05 cm o f the surface (Eom 1992, Galantowicz et al. 2000, M ettemicht & Zinck 2003,
Wigneron et al. 2003).
Such studies have generally focused on analyses o f sensor data at single
frequencies, e.g., L- band or C-band but seldom jointly at multiple frequencies. Direct
comparisons o f L- and C- band with respect to their abilities to map spatial
inhomogeneities as well as the affect o f different spatial resolutions, have yet to be
thoroughly investigated.
In this paper we investigate the multi-polarized, (H,V) low-
frequency, (L- and C-band) brightness temperature sensitivity to soil moisture for
vegetation conditions ranging from bare soils to dense com canopy using airborne
radiometers. The two airborne instruments used are the Passive and Active L- and S-band
(PALS) sensor (Njoku et al. 2002) and the Polarimetric Scanning Radiometer (PSR/C)
17
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
sensor (Jackson et al. 2002). The study is related to previous studies o f soil moisture
retrieval using these sensors (Wilson et al. 2001, Jackson et al. 2002, Njoku et al. 2002)
and corresponding detailed studies of the influence of vegetation type and density on
brightness temperature (Bolten et al. 2003). Data from four overflight days of the Little
Washita watershed on July 8th, 9th, 11th and 14th during the 1999 Southern Great Plains
(SGP99) experiment are used. By applying spatial averaging to the co-registered PALS
and PSR/C samples, we compare and simulate the multi-scale L- and C-band brightness
temperatures over varying vegetation types and soil moisture conditions.
Section 2.1. presents data from the SGP99 field experiment, description of the Land C-band airborne sensors, and the intercomparison strategy. Section 2.2. discusses the
intercomparison results, and the conclusions and discussion are presented in section 2.3.
2.2. Data and Intercomparison Methods
2.2.1. 1999 Southern Great Plains Experiment (SGP99)
The 1999 Southern Great Plains Experiment focused on soil moisture retrieval
and remote sensing using several instruments to provide large-scale airborne mapping of
soil moisture. Data collection included ground, aircraft and satellite-based remote sensing
measurements over selected field sites (locations of ground data collection) for each day
of the study. The experiment used a variety o f airborne L-, S-, and C-band microwave
instruments in the Little Washita Basin near Chickasha, Oklahoma (35 27’ N, 097 57
W) in the Southern Great Plains Region of the United States between July 8, 1999 and
July 20, 1999. Basin topography consists mostly of rolling hills with maximum relief less
than 200 m for rangeland and pasture over an area o f 603 km2. Forest cover within the
18
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
watershed is very sparse, constituting only a small portion o f the watershed (Jackson &
Hsu 2001) and primarily limited to riparian zones. Thirteen field sites (0.8 km x 0.8 km)
were chosen with five types o f vegetation cover: fallow, harvested wheat (stubble), grass
(rangeland), alfalfa, and com with vegetation water contents from essentially zero for
bare soil to over 7 kg/m2 in the com fields (Table 2.1).
The watershed includes a range of soil textures from coarse to fine (Jackson &
Hsu 2001). Weather conditions were ideal during the study producing a major rain event
on the second day (July 10th, 1999) of the study, which allowed a subsequent drydown
period to be observed throughout the basin. A change in gravimetric soil moisture content
from about 24.3 % to 3.4 % was observed over a period of 8 days. The precipitation was
non-uniform throughout the basin (31.0 mm in the western part o f the watershed to
approximately 9.6 mm in the eastern regions) causing a heterogeneous soil moisture
pattern across the watershed.
The extensive ground data collected during the experiment included gravimetric
soil moisture, surface temperature, bulk soil density, vegetation water content, and
surface roughness, as described in (Jackson & Hsu 2001). Soil moisture and temperature
were collected daily whereas bulk density, vegetation water content, and surface
roughness were collected once during the experiment and assumed to remain constant for
the duration o f the experiment. Surface conditions were assumed to be homogeneous
within each sampled field.
19
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
2.2.2. Passive / Active L- and S- Band Sensor (PALS)
This study uses data from the JPL Passive/Active L- and S-band (PALS) sensor
for soil moisture observations within the watershed (Njoku et al. 2002, Bolten et al.
2003). PALS was developed to study the utilization of dual-frequency, dual-polarization,
passive and active measurements for spacebome remote sensing o f ocean salinity and soil
moisture (Wilson et al. 2001). The instrument provides geolocated horizontal and vertical
polarized L- and S-band brightness temperatures, backscatter coefficient, and surface
temperature. PALS radiometer channels are centered at 1.4 and 2.69 GHz and its radar
channels are at 1.26 and 3.15 GHz. Table 2.2 lists the key PALS characteristics. A more
thorough description o f the instrument can be found in (Wilson et al. 2001). This chapter
focuses on the PALS passive data for comparison with passive PSR/C. The L- band data
show similar soil moisture sensitivity and retrieval results to S- band (Bolten et al. 2003)
and therefore are discussed exclusively.
The PALS instrument was flown on the NCAR C-130 aircraft for a total of 6 days
during SGP99. A nominal altitude of 1000 m and incidence angle o f 39 provided a mean
footprint size o f 300 x 400 m. The C-130 flights were generally made between 8:30 am
and 1:00 pm Central Daylight Time (CDT). An attempt was made to conduct all flights at
the same speed, altitude, and direction and on a daily basis. Flight lines were selected
over the SGP99 field sites and in incremental latitude steps to provide flanked coverage
of the Little Washita region. The PALS instrument acquired data before and after the
significant rain event on July 10 thus enabling observations o f soil wetting and drying
patterns (Figure 2.1).
20
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
2.2.3. C- BandPolarim etric Scanning Radiometer (PSR/C)
The C-band Polarimetric Scanning Radiometer (PSR/C) is an airborne microwave
imaging radiometer system developed for the purpose of obtaining high-resolution multi­
band polarimetric emission imagery of a broad range o f geophysical phenomena
(Piepmeier & Gasiewski 2001). PSR/C provides simultaneous vertical and horizontal
polarized measurements within four adjacent frequency bands at 5.80-6.20, 6.30-6.70,
6.75-7.10, and 7.15-7.50 GHz, with respective channel center frequencies of 6.00, 6.50,
6.925, and 7.325 GHz. The 3-dB beamwidths are 10 degrees for all bands. In addition,
the 6.75-7.10 GHz channel (which corresponds to the AMSR-E C-band channel) also
measured the third and fourth Stokes parameters although only the V and H polarization
data are used in this study. Instrument characteristics are listed in Table 2.2.
The PSR/C was installed on the NASA P3-B aircraft and operated in a conical
scanning mode at an incidence angle of 55° from nadir on July 8, 9, 11, 14, 15, and 19,
1999. Flown at a nominal altitude o f 8.23 km, the PSR/C provided a swath width of 25.5
km with and an average footprint size of 2.3 km, yielding the first high-resolution C-band
polarimetric imagery o f soil moisture variations. The nominal daily time window for
observations was 8:30-11:30 am CDT (13:30-16:30 UTC).
The PSR/C data was subject to occurrences o f RFI due to the unprotected C-band
frequencies and leakage from strong RFI during some calibration views. However, a
spectral interference correction algorithm using the four PSR/C sub-bands was
successfully applied to the data (Gasiewski et al. 2002). Figure 2.2 shows the
horizontally-polarized 7.325 GHz TB images for each PSR/C flight during the study. The
spatial structure of the mapped TB images is consistent with data observed from the
21
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Tropical Rainfall Mapping Mission (TRMM) Microwave Imager (TMI) and moisture
patterns observed during the study (Jackson et al. 2002).
2.3. Instrument comparison
2.3.1. Co-location and comparison
The study of spatial scaling is vital to remote sensing applications. Surface
parameters derived at different resolutions (e.g., from different sensors) can be
considerably different even if derived using the same algorithms and methodology. A
number o f recent studies (Hu & Islam 1997, Drusch et al. 1999, Jackson et al. 1999,
Guha & Lakshmi 2002, Oldak et al. 2002), have investigated this scaling issue and shown
that retrievals o f regional soil moisture values can be performed with confidence using an
aggregation o f small-scale retrievals. The large difference in spatial resolution of the
PALS and PSR/C instruments allows this issue to be addressed in a unique fashion.
In an effort to best compare the PALS and PSR/C sensors, three methods were
used for co-location. First, the PALS and PSR/C footprint ellipse location, size and
orientation were calculated from in-flight navigational data. To minimize the effect of
antenna beam contributions from regions outside the PALS and PSR/C 3 db footprints,
analysis was done only on the PALS ellipses that were located entirely within a PSR/C
footprint. These co-located PALS brightness temperatures were averaged to provide a
mean value o f the PALS brightness temperature at the PSR/C-scale. In addition to the
comparisons between multiple PALS footprints with co-located PSR/C footprints, single
co-located footprints from each instrument were compared with the in situ field data. The
single PALS and PSR/C footprints that fell closest to and within 300 m of the field
22
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
centers were selected for comparison. A third resampling method was performed to
calculate the mean PALS brightness temperature from each field site. Field-scale data
were calculated as the mean value o f the brightness temperatures located entirely within
the field sites.
2.3.2. Sensitivity and resampling analysis
This analysis focuses on the comparison of two instruments that have a significant
difference in footprint size and sampling density. In order to verify the use of such
differing datasets, a sensitivity analysis was performed by comparing resampled PALS
1.4 GHz horizontally-polarized brightness temperatures at both the field- and PSR/Cscale to PSR/C 6.9 GHz horizontally-polarized brightness temperatures of the field sites.
The effect of spatial resolution on soil moisture sensitivity is evident from the decrease in
slope in Figures 2.3 and 2.4. Decreasing the resolution o f the PALS horizontal
polarization L-band TB by resampling it to PSR/C-footprint resolution results in lower
slopes compared to the original high-resolution data. Assuming a near-linear relationship
between TB and soil moisture at these frequencies, the decrease in slope is a result of
lower spatial resolution. Figure 2.3 shows that with lower spatial resolution, the
sensitivity (slope) o f L-band TB to soil moisture approaches that o f C-band. The slope of
the average PALS 1.4 GHz H-pol brightness temperatures at PSR-scale decreases to 2.58 from -3.53 at field-scale. Figure 2.4 shows horizontally-polarized L-band data at the
field-scale and PSR-scale for high (VWC > 1.0 kg m'2) and low vegetated fields (VWC <
1.0 kg m '2). It is seen that the high spatial resolution L-band data results in steeper slopes
o f Tb
vs.
soil moisture curves (higher sensitivity) for both VWC ranges. The range and
23
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
standard deviation o f the PALS data are smaller at the resampled PSR/C-scale than at the
field-scale due to less variance of surface variables at the larger spatial resolution.
However, when resampled to the spatial resolution o f the PSR/C instrument, PALS data
follow similar trends to soil moisture effects as the PSR/C data and therefore are found to
be comparable, supporting the assumption that scaling is linear. Table 2.3 summarizes
these results.
2.3.3. Radiative transfer model
Estimates of L- and C-band brightness temperatures were calculated using a
physically-based microwave radiative transfer model (Njoku & Li 1999). Observed
surface soil and vegetation parameters (Table 2.4) were assumed homogeneous, resulting
in average effective values over the radiometer footprint. These values were used within
the formalism o f (Dobson et al. 1985, Hallikainen et al. 1985) to calculate the dielectric
constant as approximated by the quadratic expressions.
The radiative transfer model for bare-soil emissivity was modified for roughness
and vegetation, assuming uniform moisture content to the penetration depth o f the sensor.
Emissivity is increased by scattering from both vegetation and surface roughness, each of
which can reduce the sensitivity of microwave emission to soil moisture variability by
decreasing the range o f measurable emission.
The soil surface is parameterized by a roughness height parameter h and a vertical
and horizontal mixing parameter Q. The parameter h is related to the surface height
standard deviation (a) by the expression:
24
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
a\2
h; = i( A ncj cos 0)
X
where X is the wavelength in cm and 6 is the viewing angle from nadir (Njoku & Li
1999). Surface roughness values of h=Q.2 and Q=0 were assigned for all fields based on
observed roughness data.
The radiative transfer model assumes that the vegetation canopy consists of a
uniform layer above the soil parameterized in the model by optical thickness xc,
vegetation water content wc and single scattering albedo co. The optical thickness
parameter is dependent on wc and has been shown to follow an approximately linear
relationship as described in (Njoku & Li 1999):
r c - bw c / cos 0
where the factor cos 0 accounts for the slant observation path through vegetation and the
b parameter is a coefficient depending on vegetation type and frequency. Vegetation
opacity coefficients have been assigned to each band according to frequency and
simulated vegetation cover in reference to previous studies (Jackson & Schmugge 1991).
2.3.4. Heterogeneity
Land surface parameters exhibit heterogeneity over a wide range o f spatial and
temporal scales including those smaller than the scale o f these aircraft measurements.
Given such differences in spatial resolution, soil moisture retrieval methods developed at
small-scales must be examined if they are to be extrapolated to satellite scales. At
satellite scales, the information received is spatially averaged over a region as large as
2500 km2.
25
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Analysis o f sub-footprint variability was accomplished by identifying PALS
footprints within the larger PSR/C footprints. Such variability is further exemplified
using simulated brightness temperatures at these frequencies to provide a controlled
environment for the computation of true heterogeneity effects. Previous studies by
(Chang & Wetzel 1991, Drusch et al. 1999, Jackson & Hsu 2001, Burke et al. 2002, Guha
& Lakshmi 2002) have demonstrated that land cover effects on remote sensing systems
vary when applying soil moisture retrieval techniques developed at high resolutions
(ground and aircraft data) to coarser satellite-scale resolutions. This study is a
continuation o f these experiments in that an aggregation scheme is attempted with a
passive model for the purpose o f evaluating the effects o f surface emission non-linearities
in L- and C-band simulations. The method employed is an introduction to the
heterogeneity problem and uses an arithmetic-averaging framework.
A simulation study was also performed using randomly generated surface
parameters to simulate brightness temperatures at a spatial scale similar to the PSR/C
instrument using the radiative transfer model. Heterogeneity was investigated by
analyzing the differences between estimates o f area-averaged brightness temperatures
calculated from area-averaged geophysical variables and mean footprint brightness
temperatures. The surface variables are distributed uniformly over the footprint; no
covariance or spatial correlations between the surface parameters have been assumed.
The antenna pattern is considered constant within the footprint, thus the observed
brightness temperature is an area average of the component brightness temperatures.
The land surface parameters; mv, Te, wc, and h were generated on a 1000 x 1000
grid. Each value was randomly generated with a given standard deviation, normalized to
26
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
a desired mean value. These randomly generated parameters were introduced into the
passive model resulting in 100,000 outputs o f polarization difference brightness
temperature A TB for both 1.4 and 6.9 GHz. The next step involved aggregating both the
input parameters and the output TB to a 100 x 100 grid (i.e., output scheme). The 100 x
100 averaged values o f surface parameters were then applied to the radiative transfer
model (i.e., the input scheme). Both input and output results were aggregated in a similar
fashion to 10 x 10 and 1 x 1 grids. Table 2.5 lists the standard deviations assigned for the
surface parameters used in the aggregation model. It is seen that as the input variables are
averaged from 1000 x 1000 field to a 1 x 1 field, the standard deviation of the fields
approach zero, whereas the mean field value remains constant. The influence o f sub­
footprint heterogeneity is found by comparison o f the input and output results.
2.4. Instrument comparison results
2.4.1. Co-location and fie ld site comparisons
Figure 2.5 shows the field site in situ 0-5 cm gravimetric soil moisture data
compared to the nearest 6.9 and 1.4 GHz brightness temperatures for all days of the
study. Both instruments responded to changes in soil moisture. Decreases in brightness
temperature of 85 K and 43 K were observed for the 1.4 GHz and 6.9 GHz horizontallypolarized channels, respectively. As expected, the L-band channel was more sensitive to
soil moisture changes.
The observed sensitivities to soil moisture and vegetation are shown in Figures
2.6, 2.7 and 2.8. These figures show brightness temperature data and in-situ moisture data
plotted daily for the bare (LW24), harvested wheat (LW23), and com (LW25) fields.
27
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Vegetation emission is mainly un-polarized, and decreases the observed polarized
emission from the underlying soil (Ulaby et al. 1983). As a result, vegetation reduces the
brightness temperature polarization difference, A Tb, at off-nadir angles. The harvested
wheat field (LW 23) underwent the greatest soil moisture increase after the July 10th
rainfall and resulted in the largest brightness temperature decrease. In the com field (LW
25), the C-band horizontally-polarized brightness temperature show a comparable, or
even lower value than the corresponding L-band channel. The near-equivalence of these
channels over heavy vegetation (7.18 kg/m2) could be instrument related and may be
exaggerated by the lower moisture values at the com fields.
Figure 2.9 illustrates calculated A T b for various biomass densities. The modeled
A T b
increases with increasing soil moisture for both the 1.4 GHz and 6.9 GHz channels,
but exhibits greater sensitivity at the lower frequency band. The values o f
A T b
also
decrease with an increasing vegetation water content and saturate at lower moisture
values for higher vegetation water content and higher frequency. Similarly, the observed
data show more evidence o f the polarization response in the L-band channels than at Cband. This is not observed in the com field (Figure 2.8), rather, the C-band ATB is
greater than L-band for each day. This could be the result o f footprint heterogeneity, as
the com field (LW25) is surrounded by lower vegetated fields.
2.4.2. Spatial heterogeneity
Figures 2.10 and 2.11 show co-located 1.4 and 6.9 GHz horizontally-polarized
brightness temperatures (PSR/C-scale) throughout the entire region on July 9th and 11th.
The decrease in brightness temperature between the two days (due to precipitation and
28
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
increased soil moisture) increases the correlation of the two channels from almost zero
(July 9th) to an R2 of 0.47 (July 11th). Table 2.6 summarizes the influence o f soil moisture
change on each channel for both instruments. Generally, the lower frequency (1.4 H)
shows a larger range as compared to the higher frequency (6.9 H) both before (July 9th)
and after (July 11th) the rainfall. In addition, the higher frequency (6.9 H) shows almost
as much (~57 K) o f an increase in the range o f brightness temperatures as the 1.4 H band
(~75K). This demonstrates the sensitivity of C-band to soil moisture at low vegetation
water contents.
Sub-pixel heterogeneity was evaluated by examining the average range and
standard deviation o f the PSR/C-scale PALS data within each PSR/C footprint. Table 2.7
shows the range (max TB- min TB) and standard deviation of PALS 1.4 GHz TB for each
o f the four co-located days. It is seen that the range(std. dev.) increases from 11.17(3.5) K
to 19.03(6.2) K after the precipitation, as the rainfall was not uniform over the PSR/C
footprints. As expected, the LH-band has the greatest range and standard deviation
(greater moisture sensitivity) for all channels before and after the rainfall 7.71(2.58) K on
July 8th, 8.05(2.71)K on July 14th.
The effects o f the heterogeneity are also shown by analyzing the difference
between the mean and standard deviations o f the inputs and outputs from the passive
model simulations (Table 2.8). This analysis is different from the observed aggregation
method. In this case the surface conditions are altered within the same footprint; the
previous method used multiple observed footprints with different surface conditions. The
L-band A TB standard deviation decreases (4.21 K from 1000 x 1000 to 10 x 10 grid)
when aggregating. Similarly, the C-band simulation results in a decrease in standard
29
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
deviation (5.10 K from 1000 x 1000 to 10 x 10 grid). It is seen that when similar soil
moisture values are encountered in regions with varying vegetation, the region with
higher vegetation will show a larger decrease in A T b in C-band (1.48 K) than L-band
(0.12 K). The differences in input and output schemes do not exceed 2 K for all simulated
conditions.
When focusing on a one-to-one comparison o f the two sensors, heterogeneity
effects (as seen in Figure 2.10) were reduced by comparing only the single PALS and
PSR/C footprints that fell closest to the field site centers. Figure 2.12 compares 1.4 and
6.9 GHz horizontally-polarized brightness temperatures for all days o f the study. As seen
in the multiple footprint comparison, L-band values have a larger range in brightness
temperature. Moderate R2 values were achieved (up to 0.69), likely due to the difference
in emission sources for the two footprints (PALS 400 m, PSR/C 2.3 km). A summary of
correlation coefficients is listed in Table 2.9. The highest correlation (between
instruments) is 0.829 for the 1.4 H and 7.3 H channels, and 0.986 for the 1.4 V and 2.7 H.
We have included all channels here for sake of completion.
Figure 2.13 shows a comparison between the estimated brightness temperature
and the PALS and PSR/C observations (over all fields) during SGP99. The curves
indicate that the radiative transfer model agrees well with the brightness temperature
observations below 20% moisture, the average absolute difference between estimated and
measured brightness temperature being less than 5 K.
30
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
2.5. Conclusions and discussion
It has been well established that the long wavelength region o f the microwave
spectrum ( - 2 1 cm) is valuable for near surface soil moisture retrieval. The recent
addition o f moderate resolution passive microwave spacebome missions has presented
the opportunity to apply retrieval methods and algorithms developed over the past
decades from ground and airborne studies to a global scale. However, due to the
influence o f surface heterogeneity, roughness and vegetation on the microwave signal, a
more complete understanding o f system design needed to minimize these effects still
requires additional field studies. In this study, we have compared brightness temperatures
measured by two distinct microwave remote sensing instruments in an attempt to study
the effects o f changes in soil moisture, vegetation, and sub-footprint heterogeneity on Land C-bands.
The incomplete filling o f the PSR/C footprint by the PALS footprints results in a
difficult one-to-one comparison. However, a resampling of the PALS data to field- and
PSR/C-scales are a more valid approach for comparison with the PSR/C data. Both of the
sensors showed a significant response to increased soil moisture on July 10th for all
vegetation conditions encountered. An increase in brightness temperature range and
standard deviation was observed with decreasing vegetation; the largest range was
observed in the L-band channels. The 1.4 GHz horizontally-polarized channel was most
sensitive to these changes and showed good correlation (R2=0.83) with the PSR/C 7.3
GHz H- polarized channel.
31
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
This work shows that there is a significant spectral heterogeneity within even a ~2
km sized footprint and that this variability increases with increased soil moisture. It was
established using a model simulation that the potential error due to spatially
heterogeneous surface types can be reduced by spatial averaging (simulating the
aggregating effect o f satellite-scale sensors). Therefore, this study suggests that spatially
heterogeneous footprints should not pose a significant problem for the retrieval of
average soil moisture within the C- to L-band frequencies.
Aggregation of TB reduced the effect of soil moisture and vegetation variability
within the larger PSR/C footprint. In fact, resampling the high resolution L-band
brightness temperature to the C-band spatial resolution does decrease the sensitivity as
seen from the lower slope o f the TB vs. soil moisture curves. This lower sensitivity
prevails in all fields. However, lower spatial resolution L-band still possesses higher
sensitivity than the same resolution C-band data.
This research forms a framework for future microwave satellite sensors. In
particular, we have the C-band Advanced Microwave Scanning Radiometer (AMSR) in
space at the moment with two L-band missions, viz., SMOS (Soil Moisture and Ocean
Salinity - ESA) and Hydros (Hydrospheric States Mission - NASA) to be launched in a
few years time. In fact, if we do achieve concurrent observations from the L- and C-band
instruments, we would have a task o f merging these observations for better retrievals of
soil moisture. Studies like the one presented in this paper should go a long way in this
regard.
32
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
References
Basara JB, Crawford KC (2002) Linear relationships between root-zone soil moisture and
atmospheric processes in the planetary boundary layer. Journal o f Geophysical
Research-Atmospheres 107:1-18
Bindlish R, Jackson TJ, Wood E, Gao HL, Starks P, Bosch D, Lakshmi V (2003) Soil
moisture estimates from TRMM Microwave Imager observations over the
Southern United States. Remote Sensing o f Environment 85:507-515
Bindlish R, Kustas WP, French AN, Diak GR, Mecikalski JR (2001) Influence of near­
surface soil moisture on regional scale heat fluxes: Model results using
microwave remote sensing data from SGP97. IEEE Transactions on Geoscience
and Remote Sensing 39:1719-1728
Bolten JD, Lakshmi V, Njoku EG (2003) Soil moisture retrieval using the passive/active
L- and S-band radar/radiometer. IEEE Transactions on Geoscience and Remote
Sensing 41:2792-2801
Burke EJ, Bastidas LA, Shuttleworth WJ (2002) Exploring the potential for multipatch
soil-moisture retrievals using multiparameter optimization techniques. IEEE
Transactions on Geoscience and Remote Sensing 40:114-120
Chang JT, Wetzel PJ (1991) Effects of Spatial Variations o f Soil-Moisture and
Vegetation on the Evolution of a Prestorm Environment - a Numerical CaseStudy. Monthly Weather Review 119:1368-1390
Choudhury BJ (1991) Passive Microwave Remote-Sensing Contribution to Hydrological
Variables. Surveys in Geophysics 12:63-84
Delworth T, Manabe S (1989) The Influence o f Soil Wetness on near-Surface
Atmospheric Variability. Journal o f Climate 2:1447-1462
Dobson MC, Ulaby FT (1986) Active Microwave Soil-Moisture Research. IEEE
Transactions on Geoscience and Remote Sensing 24:23-36
Dobson MC, Ulaby FT, Hallikainen MT, Elrayes MA (1985) Microwave Dielectric
Behavior o f Wet Soil .2. Dielectric Mixing Models. IEEE Transactions on
Geoscience and Remote Sensing 23:35-46
33
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Drusch M, Wood EF, Simmer C (1999) Up-scaling effects in passive microwave remote
sensing: ESTAR 1.4 GHz measurements during SGP '97. Geophysical Research
Letters 26:879-882
Eom HJ (1992) A Thermal Microwave Emission Model for Row-Structured Vegetation.
International Journal of Remote Sensing 13:2975-2982
Famiglietti JS, Devereaux JA, Laymon CA, Tsegaye T, Houser PR, Jackson TJ, Graham
ST, Rodell M, van Oevelen PJ (1999) Ground-based investigation of soil moisture
variability within remote sensing footprints during the Southern Great Plains 1997
(SGP97) Hydrology Experiment. Water Resources Research 35:1839-1851
Fetter CW (1994) Applied Hydrologeology, Vol. Macmillan College Publishing
Company, New York
Galantowicz JF, Entekhabi D, Njoku EG (2000) Estimation o f soil-type heterogeneity
effects in the retrieval o f soil moisture from radiobrightness. IEEE Transactions
on Geoscience and Remote Sensing 38:312-316
Gasiewski AJ, Klein M, Yevgrafov A, Leuskiy V (2002) Interference mitigation in
passive microwave radiometry International Geoscience and Remote Sensing
Symposium, Toronto, Canada
Guha A, Lakshmi V (2002) Sensitivity, spatial heterogeneity, and scaling of C-band
microwave brightness temperatures for land hydrology studies. IEEE
Transactions on Geoscience and Remote Sensing 40:2626-2635
Gulledge J, Schimel JP (1998) Moisture control over atmospheric CH4 consumption and
C 02 production in diverse Alaskan soils. Soil Biology & Biochemistry 30:11271132
Hallikainen MT, Ulaby FT, Dobson MC, Elrayes MA, Wu LK (1985) Microwave
Dielectric Behavior o f Wet Soil .1. Empirical-Models and ExperimentalObservations. IEEE Transactions on Geoscience and Remote Sensing 23:25-34
Hollenbeck KJ, Schmugge TJ, Homberger GM, Wang JR (1996) Identifying soil
hydraulic heterogeneity by detection o f relative change in passive microwave
remote sensing observations. Water Resources Research 32:139-148
34
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Hu ZL, Islam S (1997) Effects of spatial variability on the scaling of land surface
parameterizations. Boundary-Layer Meteorology 83:441-461
Jackson TJ, Gasiewski AJ, Oldak A, Klein M, Njoku EG, Yevgrafov A, Christiani S,
Bindlish R (2002) Soil moisture retrieval using the C-band polarimetric scanning
radiometer during the Southern Great Plains 1999 Experiment. IEEE Transactions
on Geoscience and Remote Sensing 40:2151-2161
Jackson TJ, Hsu AY (2001) Soil moisture and TRMM microwave imager relationships in
the Southern Great Plains 1999 (SGP99) Experiment. IEEE Transactions on
Geoscience and Remote Sensing 39:1632-1642
Jackson TJ, Le Vine DM, Hsu AY, Oldak A, Starks PJ, Swift CT, Isham JD, Haken M
(1999) Soil moisture mapping at regional scales using microwave radiometry: The
Southern Great Plains Hydrology Experiment. IEEE Transactions on Geoscience
and Remote Sensing 37:2136-2151
Jackson TJ, Schmugge TJ (1989) Passive Microwave Remote-Sensing System for SoilMoisture - Some Supporting Research. IEEE Transactions on Geoscience and
Remote Sensing 27:225-235
Jackson TJ, Schmugge TJ (1991) Vegetation Effects on the Microwave Emission of
Soils. Remote Sensing of Environment 36:203-212
Koster R, Milly PCD (1997) The interplay between transpiration and runoff formulations
in land surface schemes used with atmospheric models. Journal o f Climate
10:1578-1591
Laymon CA, Crosson WL, Jackson TJ, Manu A, Tsegaye TD (2001) Ground-based
passive microwave remote sensing observations o f soil moisture at S-band and Lband with insight into measurement accuracy. IEEE Transactions on Geoscience
and Remote Sensing 39:1844-1858
Laymon CA, Crosson WL, Soman VV, Belisle WR, Coleman TL, Fahsi A, Manu A,
Senwo ZN, Tsegaye TD, O'Neill PE, Jackson TJ (1999) Huntsville'96: an
experiment in ground-based microwave remote sensing o f soil moisture.
International Journal of Remote Sensing 20:823-828
Mettemicht GI, Zinck JA (2003) Remote sensing o f soil salinity: potentials and
constraints. Remote Sensing of Environment 85:1-20
35
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Michot D, Benderitter Y, Dorigny A, Nicoullaud B, King D, Tabbagh A (2003) Spatial
and temporal monitoring of soil water content with an irrigated com crop cover
using surface electrical resistivity tomography. Water Resources Research 39:14
Njoku EG, Jackson TJ, Lakshmi V, Chan TK, Nghiem SV (2003) Soil moisture retrieval
from AMSR-E. IEEE Transactions on Geoscience and Remote Sensing 41:215229
Njoku EG, Li L (1999) Retrieval of land surface parameters using passive microwave
measurements at 6-18 GHz. IEEE Transactions on Geoscience and Remote
Sensing 37:79-93
Njoku EG, Wilson WJ, Yueh SH, Dinardo SJ, Li FK, Jackson TJ, Lakshmi V, Bolten J
(2002) Observations o f soil moisture using a passive and active low-frequency
microwave airborne sensor during SGP99. IEEE Transactions on Geoscience and
Remote Sensing 40:2659-2673
Oldak A, Pachepsky Y, Jackson TJ, Rawls WJ (2002) Statistical properties of soil
moisture images revisited. Journal o f Hydrology 255:12-24
Owe M, Chang A, Golus RE (1988) Estimating Surface Soil-Moisture from Satellite
Microwave Measurements and a Satellite Derived Vegetation Index. Remote
Sensing o f Environment 24:331-345
Owe M, Van de Griend AA, de Jeu R, de Vries JJ, Seyhan E, Engman ET (1999)
Estimating soil moisture from satellite microwave observations: Past and ongoing
projects, and relevance to GCIP. Journal o f Geophysical Research-Atmospheres
104:19735-19742
Piepmeier JR, Gasiewski AJ (2001) High-resolution passive polarimetric microwave
mapping o f ocean surface wind vector fields. IEEE Transactions on Geoscience
and Remote Sensing 39:606-622
Schmugge T (1985) Remote sensing of soil moisture. In: Burt (ed) Hydrological
Forcasting. John Wiley and Sons, New York, p 604
Schmugge T, Jackson TJ, Kustas WP, Wang JR (1992) Passive Microwave RemoteSensing of Soil-Moisture - Results from Hapex, Fife and Monsoon-90. ISPRS
Journal of Photogrammetry and Remote Sensing 47:127-143
36
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Shukla J, Mintz Y (1982) Influence o f Land-Surface Evapo-Transpiration on the Earths
Climate. Science 215:1498-1501
Ulaby FT, Razani M, Dobson MC (1983) Effects o f Vegetation Cover on the Microwave
Radiometric Sensitivity to Soil-Moisture. IEEE Transactions on Geoscience and
Remote Sensing 21:51-61
Wang JR, Shiue JC, Schmugge TJ, Engman ET (1989) Mapping Surface Soil-Moisture
with L-Band Radiometric Measurements. Remote Sensing o f Environment
27:305-312
Wigneron JP, Calvet JC, Pellarin T, Van de Griend AA, Berger M, Ferrazzoli P (2003)
Retrieving near-surface soil moisture from microwave radiometric observations:
current status and future plans. Remote Sensing o f Environment 85:489-506
Wilson WJ, Yueh SH, Dinardo SJ, Chazanoff SL, Kitiyakara A, Li FK, Rahmat-Samii Y
(2001) Passive active L- and S-band (PALS) microwave sensor for ocean salinity
and soil moisture measurements. IEEE Transactions on Geoscience and Remote
Sensing 39:1039-1048
Wu JQ, Zhang RD, Yang JZ (1996) Analysis o f rainfall-recharge relationships. Journal of
Hydrology 177:143-160
37
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Field
Site
Vegetation
Cover
wc
(kg m '2)
2
3
4
5
12
13
21
22
23
24
25
26
27
Grass (range)
Grass (range)
Grass (range)
Grass (range)
Grass (range)
Grass (range)
Wheat (stubble)
Wheat (stubble)
Wheat (stubble)
Bare
Com
Com
Alfalfa
0.16
2.38
0.48
0.34
0.50
0.14
0.12
0.02
0.36
0.00
7.18
5.19
1.01
Table 2.1. Field site characteristics.
38
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
PSR/C and PALS Instrument Descriptions
PSR/C
PALS
Dates
Flown
(July)
Frequencies
(GHz)
Bandwidths
(MHz)
8, 9, 11, 14, 15, 19
8, 9, 11, 12, 13, 14
6.00, 6.50, 6.925, 7.325
1.41,2.69
400, 400, 350, 350
20, 60
O
00
V and H, 3rd and 4th
Stokes' parameter
VandH
Polarization
capability at
6.75-7.10 GHz
Incidence
55°
Angle (deg) (Full 360° Azimuth Scan)
Mean
2300
300 x 400
Spatial
( 8.2 km altitude and 55° (1 km altitude and 38°
Resolution
incidence angle)
incidence angle)
(m)
Table 2.2. PALS and PSR/C instrument characteristics.
39
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Field-Scale
1.4
1.4
1.4
6.9
6.9
GHz
GHz
GHz
GHz
GHz
H-pol
H-pol
H-pol
H-pol
H-pol
(all fields)
(low wc)
(high wc)
(all fields)
(low wc)
6.9 GHz H-pol (high wc)
range(K)
84.85
83.41
49.36
ct(K)
21.8
23.16
15.42
PSR/C-Scale
slope R2 range(K )
71.34
-3.53 0.91
-3.66 0.91
66.17
-3.11 0.88
37.88
41.95
41.95
21.52
CT (K)
15.2
16.46
slope
-2.58
-2.69
R2
0.75
0.72
12.69
10.66
12.8
-2.48
-1.45
-1.84
0.95
0.64
0.77
7
-1.15
0.52
Table 2.3. Statistics o f PALS and PSR/C brightness temperatures at the field- and PSRscales for all field sites.
40
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
(a) Media & Sensor Parameters
Vegetation:
Single scattering albedo co
Opacity xc (1.4 , 6.9 GHz)
Soil:
Roughness coefficient, h
Bulk density (g cm'3)
Sand and clay mass fractions (%)
Sensor:
Viewing angle PALS, PSR/C (deg)
Frequency (GHz)
Polarization
(b) Media Variables
Surface soil moisture (g cm'3)
Vegetation water content wc (kg m '2)
Surface temperature (K)
0
0.1, 0.3
0.2
1.14
30,20
3 9 , 55
1.4, 2.7, 6.0
6.5, 6.9, 7.3
H, V
in situ
in situ
in situ
Table 2.4. Inputs for the emission model.
41
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
1000X1000 100X100 10X10
a
a
a
Input Parameters
Soil Moisture
(cm3/cm3)
0.32
0.075
0.017
0.003
Surface
Temperature (C)
33.9
0.397
0.088
0.022
Vegetation Water
Content (kg/m2)
1.25
0.333
0.076
0.014
Canopy
Temperature (C)
31.90
0.397
0.088
0.022
Surface Roughness
0.10
0.013
0.003
0.001
Table 2.5. Mean and standard deviations o f the input parameters for
spatial heterogeneity simulations.
42
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Channel
1.4 V
1.4H
2.7H
2.7V
6.OH
6.0V
6.5H
6.5V
6.9V
6.9H
7.3H
7.3V
July 9th
range
CT
17.3
2.5
24.8
5.0
19.4
2.7
14.7
1.9
35.3
5.5
25.8
4.6
31.1
4.8
23.6
3.8
31.6
4.7
24.2
3.7
23.0
3.6
31.1
4.7
July 11th
range
a
64.2
8.6
101.6
13.0
79.5
11.2
54.1
8.0
79.3
10.6
7.2
53.6
79.0
10.5
52.2
7.2
49.7
81.0
82.7
51.8
7.0
10.3
10.3
7.1
Table 2.6. Range and standard deviations o f brightness
temperatures from all PALS (PSR/C-scale) and PSR/C channels.
43
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Range (K) / a (K)
Average #
footprints
1.4H
1.4 V
July 8th
12
7 .7 1 /2 .5 8
6.02 / 2.03
July 9th
18
11.17/3.55
6 .7 5 /2 .1 2
July 11th
18
19.03/6.15
12.24/3.96
July 14th
11
8.05/2.71
5 .1 6 /1 .7 6
Table 2.7. Summary o f statistics for PALS 1.4 GHz horizontally polarized footprints
falling entirely within a PSR/C footprint.
44
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Frequency
1.4 GHz
6.925 GHz
1.4 GHz
6.925 GHz
Input
Output
Averaged
Averaged
ct(K)
°(K )
1000X1000
4.4070
4.4070
5.5157
5.5157
100X100
.9820
2.3527
.9810
2.4440
10X10
1.4 GHz
6.925 GHz
.2022
0.4536
0.2020
0.4731
Table 2.8. Means and standard deviations
o f the modeled 1.4 GHz and 6.9 GHz brightness
temperature polarization differences.
45
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
6.0v
6.Oh
6.5v
6.5h
6.9v
6.9h
7.3v
7.3h
6.0v
1.00
0.76
0.98
0.80
0.97
0.82
0.94
0.81
1.4h 0.77
1.4v 0.76
2.7h :';:0 i7 5 i
2.7v 0.74
6.Oh
6.5v
1.00
0.79
0.99
0.78
0.98
0.80
0.95
1.00
0.83
0.99
0.85
0.98
0.84
0.74
0.75
0.74
0.77
0.80
0.80
0.78
0.78
6.5h
1.00
0.85
0.99
0.84
0.98
0.80
0.81
0.80
0.82
6.9v
6.9h
1.00
0.86
0.99
0.85
1.00
0.84
0.99
0.82
0.82
0.80
0.79
0.81
0.81
0.80
0.80
7.3v
1.00
0.83
0.79
0.80
0.77
0.77
7.3h
1.4h
1.00
0.83
1.00
0.82
0.81
0.80
0.98
0.98
0.93
1.4v
2.7h
1.00
0.99
1.00
0.98
0.97
2.7v
1.00
Table 2.9. Cross-channel correlation coefficients o f brightness temperatures over the
SGP99 field sites for each day.
46
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
■98.30
-98.20
-98.10
-98 00
-98 30
-98.20
•98.10
•98.00
Longitude
210 220 230 240 250 260 270 280 290
Kelvin
Figure 2.1. PALS brightness temperatures for 6 flight lines were linearly interpolated
giving a smooth field of brightness temperatures o f approximately 6 x 40 km in size
(from Njoku et al. 2002).
47
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Northing (m)
4100000
1000000
1900000
3000000
Uiyll
000000
700000
030000
000000*
700000
'w o o o
Jbly 14
600000
July 15
*700000
60COOO
700000
600000
700000
Easting (m)
190
210
230
2S0
270
290
310
Brightness Temperature (K)
Figure 2.2. PSR/C horizontally polarized 7.325 GHz brightness temperature maps
acquired on July 8-9, 11, 14, 15 and July 19, 1999 (from Jackson et al. 2002).
48
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
290
•Q*
272
254
i236
218
6.9 GHz H-pol
1.4 GHz H-pol PSR-scale
1.4 GHz H-pol Field-scale
200
0
10
20
30
S oil M oistu re (%)
Figure 2.3. Comparison of PALS 1.4 GHz horizontally polarized TB resampled to fieldand PSR/C-scale with individual 6.9 GHz horizontally polarized TB over the field sites.
49
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
290
om
272
254
m
I236
PSR-scale High VW C
218
PSR-scale Low VW C
Field-scale High VW C
Field-scale Low VW C
200
0
10
20
30
Soil Moisture (%)
Figure 2.4. Relationship o f PALS 1.4 GHz horizontally-polarized brightness temperatures
at field- and PSR/C-scale to in situ soil moisture.
50
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
290
270
/T)
O
250
*
CO
230
210
O
6.9 GHz H-pol
•
1.4 GHz H-pol
190
10
20
30
Soil M oisture (%)
Figure 2.5. Relationship between PALS and PSR/C brightness temperatures and 0-5 cm
in situ gravimetric soil moisture.
51
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
LW 23 (Harvested Wheat)
0.36 Kg/m2
310
290
270
-A — 6.9V GHz
6.9H GHz
- 0 — 1.4H GHz
£ 250
V
■A -- 1.4V GHz
♦
6,9 GHz Diff +200
-<>— 1.4 GHz Diff +200
210
190
8
11
9
14
July Day
Figure 2.6. PALS and PSR/C brightness temperature data and in situ moisture data
plotted daily for site LW23.
52
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Brightness Temperature (K)
LW 24 (Bare Field)
310
30
290
25
270
20
- -A - 6.9V GHz
— ■ — 6.9H GHz
□
15 £
250
1.4H GHz
-■A - 1.4V GHz
♦
6.9 GHz Diff +200
—0 — 1.4 GHz Diff +200
10
230 -
210
190
8
11
9
14
July Day
Figure 2.7. PALS and PSR/C brightness temperature data and in situ moisture data
plotted daily for site LW24.
53
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
LW 25 (Corn)
7.18 Kg/m2
310
Brightness Temperature (K)
2 90
270
- -A - 6.9V GHz
6.9H GHz
- B — 1.4H GHz
15 .2
2 50
- A - - 1.4V GHz
♦
6.9 GHz D iff+200
- 0 — 1.4 GHz D iff+200
2 30
210
190
July Day
Figure 2.8. PALS and PSR/C brightness temperature data and in situ moisture data
plotted daily for site LW25.
54
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
35
i r t f""i t |
'□
t i i T""r"T".|"T"T-T r.T.„rT„r „T,. |..r-| rilT
II,r„1„„r „t r„„j r „1 r r „r tM
M
,,...r„,T,,y ,,r r ,r yi
6 .9 GHz T b (V-H)
1.4 GHz T b (-V-H)
30
W o- 0 Kg m - 2
25
a, 20
H
<
:
:
15
10
/
/
I >11 M
1
illtttttttttttill
........ ...... 1 1 1 1 * f *
*
*
*
f 111
|DDODDDOOO □O POOOO □ □ D O D Q O O O O g
□ODD1
5
0
□ □ n o o D ODDnnDDaDDODDDODDDDDDDDDDDDDDD
I 1I I 1
1.1 L
....1
. I I. I
10
20
IL...I. 1 i l l I
W c- 1.5 Kg m - 2 -
Wc - 3 , 0 Kg m - 2
W c - 1 .5 Kg m - 2 .
Wc= '3’°
1 I I 1 . I .1 I I I I I t ■I . I I I I . ■1 .
30
40
50
Soil Moisture (%)
Figure 2.9. Modeled A TB verses soil moisture for L-band (1.4 GHz) and C-band (6.9
GHz) channels for three values o f vegetation water content wc, values at 55° incidence.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
290
y = 0.0763x + 249.96
R2 = 0.01
270
u
m 250
H
O
i
n
X
230
210
190
------------- 1------170 I--------------1--------------1------------- 1------------- 1------------- ■
170
190
210
230
6.9
250
270
290
GHz H-pol TB(K )
Figure 2.10. Co-located PALS and PSR/C horizontally polarized
brightness temperatures throughout the entire region on July 9, 1999.
56
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
290 -
270 -
g
09 250
H
oo.
-
* 230 -
a
o
~
210
-
190 -
170 170
200
230
260
290
6.9 GHz H-pol TB (K)
Figure 2.11. Co-located PALS and PSR/C horizontally polarized
brightness temperatures throughout the entire region on July 11, 1999.
57
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
315
y = 1 .7 8 3 8 x -225.43
R2 = 0.66
295
i t /
275
*
/♦ *
.
*
^0
255
Q.
1
X
N
X
O
235
215
195
175
175
195
215
235
6.9
255
275
295
315
GHz H-pol T b (K)
Figure 2.12. 1.4 and 6.9 GHz channel comparison for single horizontally polarized values
which were located closest to the field site centers.
58
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
"v""J'""f T"‘1 J I—
E""T1 T*"T r“1 m
f V"5"I rTTTTT’TTTT"
3 QO
e
**
280
6.9 GHz V
i 2«0
&
6.9 GHz H
- 240
- 1.4 GHz V
I■B 220
1.4 GHz H
M
□
O 1 . 4 GHz H-pol ob served
200
□
o
A 1 .4 GHz V-poI o b se rv e d
+- 6 . 9 GHz H-pol ob served
* 6 .9 GHz V -pol o b se rv e d
1 8 0 l.,i i, i. ,i,. i i > i i i
i i i
10
i i i i
i i i
i i i i >, l i i i l
20
JO
Soil Moisture <3 Orov)
i,, i,„i i i i i . i i
40
Figure 2.13. Modeled brightness temperature plotted with co-located PALS (1.4 GHz)
and PSR/C (6.9 GHz) observations.
59
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
CHAPTER 3
SOIL MOISTURE ESTIMATES FROM AIRCRAFT AND SATELLITE
MICROWAVE SENSORS DURING SMEX02
(SOIL MOISTURE EXPERIMENT 2002)
Abstract
The launch o f NASA’s Aqua satellite on 4 May, 2002 provided the first
opportunity for high temporal and large scale observation of soil moisture. The AMSR-E
(Advanced Microwave Scanning Radiometer for the Earth Observing System) instrument
onboard Aqua improves the spatial resolution and frequency range o f earlier generations
o f passive microwave instruments. This chapter deals with near simultaneous
observations o f in situ soil moisture and AMSR-E brightness temperatures (TB) at a
regional site (100 x 50) in the Walnut Creek watershed near Ames, Iowa in June-July
2002. Brightness temperatures from the airborne C/X-band Polarimetric Scanning
Radiometer (PSR C/X) are used to compare moisture and vegetation effects on the
microwave emission at both the satellite (-56 km) and aircraft (-2 km) spatial
resolutions. Simulations o f expected AMSR-E, and PSR C/X brightness temperatures are
performed in an attempt to analyze sub-pixel heterogeneity and moisture retrieval
methodology for comparison with the observed data. Trends in model errors are related to
60
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
variations in observed model parameter inputs and observed brightness temperature
biases. A description o f the model simulation and heterogeneity effects is given in
addition to moisture retrievals and in situ comparisons. These studies will help us to gain
confidence in using the AMSR-E derived soil moisture in regions that lack in situ
observations.
61
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
3.1 Introduction
The importance o f near surface soil moisture observations has been well
established. Many studies have shown the importance o f soil moisture in the realm of
agriculture, hydrology and climatology by demonstrating a unique ability to determine
surface runoff and infiltration, and similarly influence surface heat flux and local climate
(Delworth & Manabe 1989, Brubaker & Entekhabi 1996, Koster & Milly 1997, Pielke
2001, Basara & Crawford 2002, Montaldo & Albertson 2003). These boundary condition
relationships enable a better understanding o f land-surface hydrologic processes with the
use of large-scale soil moisture observations. However, soil moisture varies in space and
time. The heterogeneous nature of soil moisture makes accurate point measurements at
high spatial and temporal scales difficult. Still, observations are needed at all scales for
hydrologic modeling, weather forecasting, climate prediction, flood and drought
monitoring and other water and energy cycle applications (Jackson et al. 1996, Fennessy
& Shukla 1999, Dirmeyer et al. 2000, Kustas et al. 2003).
Previous studies (Njoku 1977, Wang et al. 1987, Schmugge & Jackson 1994,
Njoku & Entekhabi 1996, Njoku et al. 2002, Wigneron et al. 2003, Drusch et al. 2004,
Gao et al. 2004, Lee & Anagnostou 2004) have demonstrated the strong relationship
between microwave brightness temperature TB and surface soil moisture on the order of
approximately 0-6 cm depth over bare and lightly vegetated (<5 kg/m2 biomass) surfaces.
These recent approaches have led to large-scale soil moisture observations using aircraft
and satellite-based instruments. In this study, the EOS-Advanced Microwave Scanning
Radiometer (AMSR-E) is used to evaluate soil moisture changes during the Soil Moisture
Experiments 2002 (SMEX02) conducted in central Iowa. A description o f the AMSR-E
62
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
instrument and an introduction to the current soil moisture retrieval algorithm used for
standard product development is described in (Njoku et al. 2003).
A microwave emission model is used in this study to evaluate soil moisture
predictions within the AMSR-E frequencies. The model is a semi empirical, four
component dielectric-mixing model is used to model the dielectric behavior o f the soil
water mixture following that o f (Dobson et al. 1985). Variations o f the radiative transfer
model used in this study has been tested and shown to hold true for areas o f sparse
vegetation; formulation and discussion of the model parameters can be found in (Mo et
al. 1982, Kerr & Njoku 1990, Njoku & Li 1999). Soil moisture is estimated by comparing
modeled and observed brightness temperatures in conjunction with in situ observations.
The fairly linear relationship o f soil moisture and microwave brightness temperature
should allow this retrieval algorithm to be applied to both coarse and high resolutions
(Hollenbeck et al. 1996, Jackson & Hsu 2001). Aggregation o f the finer resolution PSR
C/X footprints was executed for comparison o f the two instruments.
Section 3.2 gives a brief description of SMEX02 with focus on ground truth
protocols and climate observations, followed by instrument descriptions for both AMSRE and PSR C/X. An outline o f the retrieval algorithm and the parameter/brightness
temperature aggregation scheme will be provided. Section 3.3 describes the results of this
paper, with comparisons o f the forward model and observations obtained by the AMSR-E
and PSR C/X and the retrieved soil moisture. Also analyzed are the effects of observed
local Radio Frequency Interference (RFI), particularly in the AMSR-E 6.9 GHz
brightness temperatures.
63
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Previous studies on the use of microwave remote sensing data during SMEX02
have centered around analysis of the aircraft data (Bindlish et al. 2005) and ASMR-E and
PSR C/X (McCabe et al. 2004, McCabe et al. 2005). Our work differs from these studies
by focusing on radiative transfer modeling for comparison of simulated and observed T b.
Results by (Bindlish et al. 2005) and (McCabe et al. 2004, McCabe et al. 2005) provide
significant contributions to algorithm development and validation and focus on
vegetation and scaling effects for fine-scale AMSR-E soil moisture retrievals during
SMEX02. These studies were motivated towards the application o f hydrometeorological
data (McCabe et al., 2005) and analysis of field-scale moisture validations (Bindlish et
al., 2005) within the Walnut Creek Watershed. While the current study addresses similar
issues, our interests are centered more on quantitatively comparing the modeled and
observed domain-scale TB in order to better understand the effects of frequency and scale,
footprint heterogeneity, and RFI on satellite-scale moisture prediction. In doing so, we
present a quantitative statistical analysis o f the remote sensing and biophysical
parameters during SMEX02 and hope to gain a better understanding o f these inter-related
factors.
3.2. SMEX02 field experiment
The primary objectives o f SMEX02 included the extension o f microwave remote
sensing observations and retrieval algorithms to more diverse and challenging conditions.
Intense sampling within part of the study area was designed to improve understanding of
the effects o f soil moisture and field-scale heterogeneity on land-atmosphere fluxes.
Upon completion, a comprehensive data set was acquired enabling validation of AMSR-
64
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
E brightness temperature and soil moisture retrievals (Njoku 2004). An extensive in situ
dataset consisting of gravimetric soil moisture, surface temperature, bulk density, surface
roughness, vegetation water content, and crop type observations were collected during
the SMEX02 campaign [http://hydrolab.arsusda.gov/smex02]. Regional field sampling
design provided average surface volumetric soil moisture values at a scale equivalent to
two spacebome AMSR-E footprints at -5 0 km. Comparisons were made with the
observed field observations and brightness temperatures from the airborne C/X-band
Polarimetric Scanning Radiometer (PSR C/X) also flown during the study (Bindlish et al.
2005). Efforts were made to sample the region as close in time to the AMSR-E ascending
overpass as possible. In addition, the PSR C/X was flown as close in time to the ASMR-E
overpass and in situ sampling times as possible, usually between the 1200 and 1500 local
time. The dual-polarized 6.9 and 10.65 GHz AMSR-E and PSR C/X brightness
temperature data for 7 days o f the study were analyzed with emphasis on the influence of
spatial resolution differences (56 km for AMSR-E and 2 km for PSR C/X).
Figure 3.1 shows the designated SMEX02 region in central Iowa. The
experiments were conducted from 25th June through 12th July, 2002. A grid of 47
individual field sites were sampled each day covering a domain o f approximately 50 km
by 100 km. Field data for these regional sites was collected for 16 days during the
experiment (June 25th - July 12th except June 28th).
For each day, each o f the 47 sites was sampled as close to 1200 local time as
possible for soil moisture and temperature. The primary soil moisture measurements
taken in each regional site were 3 theta probe samples and 1 gravimetric soil moisture
measurement at 0-1 cm and 1-6 cm soil depths. The gravimetric measurements were
65
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
coincident with one o f the theta probe locations and were used to calibrate the theta probe
observations and also to provide estimates of bulk density for the fields. The average of
the
calibrated
theta
probe
measurements
were
used
as
the
field
average
[http://hydrolab.arsusda. go v/SMEX02].
The SMEX02 region is primarily agricultural. Approximately 95% of the region
is used for row crop agriculture, a majority o f this being com (50%) and soybean (4045%) and the remaining 5-10% being forage and grains. During the study, the soybean
fields grew from essentially bare soils to a vegetation water content o f 1 - 1.5 kg/m2
while the com fields grew from 2-3 kg/m2 to 4-5 kg/m2. SMEX02 provided unique
conditions of high initial biomass content and significant change in biomass over the
course of the experiment. In situ sampling o f vegetation water content wc occurred up to
four times during the experiment for all watershed sites. wc was extrapolated for the
entire SMEX02 region for all days by using 30 m resolution images o f Normalized
Difference W ater Index (NDWI) obtained from five days o f Landsat (5&7) Thematic
Mapper data (Gao 1996). The NDWI index is defined as:
NDWI= (Band4 - Band5) / (Band4 + Band5)
Crop specific relationships between NDWI and wc>were established and applied to each
o f second Landsat NDWI images to provide an extensive mapping o f vegetation
parameters over the study area. A detailed description o f the data source and algorithms
in estimation o f vegetation water content is given by (Gao 1996, Jackson et al. 2004,
Jackson & Cosh 2003).
The surface temperature was sampled using handheld infrared thermometers
(IRT). Soil temperatures were obtained using temperature probes inserted to depths o f 1
66
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
cm, 5 cm, and 10 cm. Air temperature was collected from the USDA/NRCS Soil Climate
Analysis Network (SCAN) instrument located in Boone Co. Temperature sampling was
done at the four locations where gravimetric soil moisture samples were also collected.
The soil type data for the SMEX02 study area were obtained from the CONUS-SOIL
dataset (Miller & White 1998). Wetting o f the region resulted from a series of
thunderstorms that occurred over the region on July 4th and 7th and July 10th.
3.2.1. Advanced Microwave Scanning Radiometer (AMSR-E)
The Advanced Microwave Scanning Radiometer (AMSR-E) instrument on the
NASA EOS Aqua satellite provides global passive microwave measurements of
terrestrial, oceanic, and atmospheric variables for the investigation o f global water and
energy cycles (Kawanishi et al. 2003, Njoku et al. 2003, Shibata et al. 2003). The satellite
follows a sun-synchronous orbit with equatorial crossing at approximately 1330 LST.
The instrument measures brightness temperatures at six frequencies, 6.92, 10.65, 18.7,
23.8, 36.5, and 89.0 GHz, with vertical and horizontal polarizations at each frequency, for
a total o f twelve channels. With a fixed incidence angle of 54.8° and an altitude o f 705
km, AMSR-E provides a conically scanning footprint pattern with a swath width of 1445
km. The mean footprint diameter ranges from 56 km at 6.92 GHz to 5 km at 89 GHz. A
summary o f the Aqua AMSR-E instrument characteristics are listed in Table 3.1.
A frequent revisit time is important for hydrologic applications. Temporal
resolution is particularly vital to obtain adequate soil moisture sampling o f surface
wetting and drying between precipitation events. AMSR-E revisit coverage is on a ~2day
period at the equator and more frequently at higher latitudes. Within the SMEX02 region,
67
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
the ascending and descending overpasses occur at approximately 130 and 1330 local
time. All data used in this study is ascending (1330 local time overpass) data, as it
corresponds to the time o f the in situ data collection as well as aircraft-mounted remote
sensing observations. The AMSR-E ascending data that was without artifacts and within
the SMEX02 timeframe were used (June 25, July 1st-15th except July 12th).
The AMSR-E data used in this study was the Level-2A (AE L2A) data product
acquired from the National Snow and Ice Data Center (NSIDC) (Ashcroft & Wentz
2003).
The
Level-2A
data
are
generated
by
Remote
Sensing
Systems
[http://www.ssmi.com/] from the NASDA Level 1A product. Resampled Level-2A
brightness temperatures are used for all higher order AMSR-E level 2B and level 3
products. The B01 version algorithm data were used for this study. These products are the
original, pre-calibrated, at-launch version of the data.
Figure 3.2 illustrates the relation between observed mean footprint-scale 0-1 cm
volumetric soil moisture and observed average AMSR-E
temperature. Average brightness temperature,
T B( a v ) ,
10.7 GHz brightness
is defined as (TBv + TBH)/2, H is
horizontally-polarized TB and V is vertically-polarized TB. This value has been shown to
correlate well with near surface soil moisture (Bolten et al. 2004) and is used in this
analysis for the TB vs. soil moisture comparisons and soil moisture retrieval algorithm. In
this case, the negative correlation resulted in an R2 o f 0.44 for all common in situ/AM&EE overpasses.
68
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
3.2.2. Polarimetric Scanning Radiometer Instrument (PSR)
The PSR C/X is a passive airborne microwave imaging radiometer system
developed for the purpose o f obtaining high-resolution multi-band polarimetric emission
imagery and is an upgraded version of the previously successful PSR/C scanhead used
during SGP99 (Piepmeier & Gasiewski 2001, Jackson et al. 2002). Consisting of fully
polarimetric C- and X-band radiometers inside a standardized gimbal-mounted PSR
scanhead drum, PSR C/X provides simultaneous vertical and horizontal polarized
measurements within four adjacent frequency bands as listed in Table 3.2. Channel center
frequencies for the C and X-bands are 6.00, 6.50, 6.925, 7.325, 10.65, 10.68, and 10.71
GHz. The 3-dB beamwidths are 10° for the 6-7 GHz channels and 7° for the 10 GHz
channels. The 6.925 GHz channel has not been used in this analysis due to significant
effects of radio frequency interference, however, the 7.325 and 10.71 GHz channel data
have been run through a radio frequency inference mitigation filter as described in
section 3.4 and are applied in this study.
During SMEX02, the PSR C/X scanhead was installed on the NASA P-3B
aircraft and operated in a conical scanning mode at an incidence angle of 55°. The
instrument was flown over two watershed and four regional flight lines at altitudes of
-1500 m and -8000 m respectively. These flight patterns were performed for ten days of
the study (June 25, 27, 29 and July 1, 4, 8, 9, 10, 11, 12), seven o f which correspond to
AMSR-E ascending overpasses and ground sampling. Flown at 8000 m, the PSR C/X
provided a swath width o f 25.5 km with and an average footprint size o f 2.3 km for the
10.7 GHz channel. This analysis uses the high altitude data exclusively. A study by
69
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
(Bindlish et al. 2005) utilizes both the high and low altitude PSR data during SMEX02
and focuses more on addressing the general trends in the observed and predicted moisture
values for the different scales.
Figure 3.3 shows the spatial distribution o f mapped PSR C/X 7.3 GHz
horizontally-polarized TB for 10 days of the study. Patterns o f brightness temperatures are
consistent with observed precipitation during the study, specifically on July 8th and 11th.
These precipitation events resulted in a decrease in horizontally-polarized brightness
temperature of -10-15 K on average. Stationary cold features such as in the southwest
comer correspond to water bodies in the region.
3.2.3. Radiative transfer
The retrieval algorithm used in this study combines observed brightness
temperatures and estimated emission from a radiative transfer (RT) model using surface
parameters collected in situ or from operationally-derived products. Elements o f the
model have been discussed in (Mo et al. 1982, Kerr & Njoku 1990, Njoku & Entekhabi
1996, Njoku & Li 1999). Variations o f this model have been used to analyze ground,
aircraft and satellite data, and have been shown to be valid for frequencies in the range 120 GHz (Bolten et al. 2003, Narayan et al. 2004). When applied to this study, the
composition o f the soil (moisture content, bulk density, sand and clay mass fractions) is
used to determine the dielectric constant and resulting emissivity from the effective soil
moisture depth using the empirical mixing model and calculations. The emitting soil
depth is assumed to be roughly one-tenth of a wavelength in the medium (i.e. - 0.4 cm
depth for the 7.3 GHz channel used in this study) (Ulaby 1986). The Fresnel equations
70
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
are used to estimate surface reflectivities for each polarization and incidence angle over
this surface (Wang & Schmugge 1980, Dobson et al. 1985). The resulting reflectivities
are then modified for rough surfaces by h, Q as shown in (Choudhury 1979, Njoku & Li
1999). Vegetation is represented in the model as a single scattering layer above the
emitting soil surface. The opacity is related to the vegetation water content wc, by:
bwr
Z° = cos a0
where 6 is the incidence angle and the parameter b is approximately proportional to
frequency and depends weakly on vegetation type and canopy structure at low
frequencies, (Jackson & Schmugge 1991, LeVine & Karam 1996, Wigneron et al. 2004).
Values of the b parameter were assigned that satisfied the calibration routines and were
consistent with values found in literature (Van de Griend & Wigneron 2004). The
brightness temperature at the top o f the vegetation layer is calculated as a function of the
aforementioned soil brightness temperature and reflectivity with the addition of
vegetation opacity rc, vegetation single-scattering albedo cop, and vegetation effective
temperature Tce:
TBp = T s {1 - rp)e~'< + Tce(1 - r»)(l - ^
)(1 + rpe ^ )
where p is polarization and Ts is the effective soil temperature (in this study, the average
of the 1 and 5 cm soil temperatures is used due to noise in the 1 cm data). Tce values were
assumed equivalent to the 6 ft elevation air temperature data taken from the Natural
Resources Conservation Service (NRCS) Soil Climate Analysis Network (SCAN) site
located within the study area during the hour of AMSR-E overpass (Jackson & Williams
71
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
2004). Since roughness characteristics of the area were not expected to change
significantly throughout the duration of the experiment; the parameters h and Q were
determined once and held constant for all time steps in the algorithm, but calibrated
individually for com and soy fields in each frequency.
Calibration o f the model was performed on a field basis for each instrument using
two days o f the study (July 4th and 10th for PSR C/X and July 6th and 10th for AMSR-E
calibration). The calibration technique was based on an iterative least squares calculation
of the forward model. The range o f model variables h, Q and b were assigned to the
possible values stated in the available literature. Next, observed values o f moisture,
vegetation and temperature were input to the model. The mean model variables that gave
the best agreement between observed and simulated brightness temperatures for each
location were assigned to the remaining data.
The single-scattering albedo also shows dependence on vegetation water content.
However, the effect o f cop on resulting brightness temperatures within these frequencies is
thought to be small compared to the effect on tCj therefore it was set to zero in the model
for all computations in an attempt to reduce further uncertainty and emphasize the surface
roughness and vegetation effects (Bindlish et al. 2003). A summary o f model parameters
is listed in Table 3.3.
An iterative, least-squares-minimization method is employed in the retrieval
algorithm. In the iterative procedure, the volumetric soil moisture is adjusted to minimize
the weighted-sum o f squared differences between measured and computed average
brightness temperatures using a binary search method for the range o f possible soil
moisture values in the 0-6 cm soil column. The estimated moisture value for that field-
72
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
site location is that which satisfies the minimization calculation. This value represents the
mean field-scale soil moisture content and is up-scaled to footprint and regional-scales
for comparison with PSR C/X and AMSR-E brightness temperatures and in situ
footprint-scale and regional-scale soil moisture.
3.2.5. Co-location and comparison techniques
Upscaling was performed based on the assumption that the regional soil moisture
values provide a single average value of the regional soil moisture for the entire SMEX02
region. The experiment was designed so that the regional field sites would to cover two
AMSR-E footprints. Therefore, mean values of the 47 regional field sites were used as
footprint average values in this study. Spatial averaging was used as an upscaling
technique for comparison o f the spatially-averaged TB, and observed and estimated
volumetric soil moisture. Many efforts have been performed to determine the most
effective method o f representing such a large (-5000 km2) area with point-scale
observations (Hu & Islam 1997, Cosh et al. 2004, Moran et al. 2004). The current method
assumes that the aggregated model outputs should maintain spatial information pertinent
to both the local scale, i.e., field (800 m) and representative larger regional area (1000
km). In addition, this averaging technique assumes a scaling linearity to simulate the
averaging effect o f the large satellite footprint.
The experiment design of SMEX02 allows three scaling methods o f soil moisture
validation
and
comparison with AMSR-E
and PSR
C/X
observed brightness
temperatures; field-scale (-800 m), AMSR-E footprint-scale (-5 0 km), and domain-scale
(100 km). Different methods of co-location and comparison are needed to analyze the
73
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
finer resolution (~2.3 km) PSR C/X footprints and the larger AMSR-E footprints. In
addition, forward modeling and soil moisture retrieval using these different scaled
brightness temperature products required a combination o f distributed modeling and
spatial averaging of the input and output variables.
To allow accurate footprint-scale comparisons o f the environmental variables and
instrument observations, all ground data was upscaled to a 2 x 1 grid for the entire
SMEX02 region. Both the AMSR-E and PSR C/X data were gridded to the same 2 x 1
dimension. The mean of the ascending AMSR-E footprint values within each grid was
calculated for each overpass during the study. It is assumed that the two AMSR-E grids
represent the entire regional SMEX02 area and that the parameter values within each
footprint have a low variance o f soil moisture distribution. AMSR-E TB were then
analyzed against the mean, standard deviation and range o f the PSR C/X brightness
temperatures which fell within each ( 2 x 1 ) grid domain. The same statistics were also
calculated for the AMSR-E footprints that fell entirely within the entire SMEX02
domain. Approximately 12 AMSR-E footprint centers were located entirely within the
regional domain due to the high sampling rate.
AMSR-E brightness temperatures were simulated using the ground parameter
values that were located within each footprint domain. The mean o f the resulting fieldscale simulated brightness temperatures was calculated to represent the entire footprint. A
similar distributed approach was used for the inversion technique. Simulated field-scale
brightness temperatures were then compared with the observed single footprint AMSR-E
brightness temperature.
74
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Spatial averaging was performed for the PSR C/X brightness temperatures and
estimated moisture. Field-scale PSR C/X simulated brightness temperatures were used to
estimate the soil moisture for each field. For comparison with AMSR-E Tb, each fieldscale brightness temperature and corresponding estimated soil moisture were averaged to
the 2 x 1 grid domain. Footprint-scale soil moisture comparisons were calculated by
averaging the in situ and PSR C/X estimated co-located soil moisture values that fell
within each AMSR-E grid domain.
Both o f these methods assume that field collected soil moisture represents the full
spatial coverage of the AMSR-E footprint, and that the field sampling provides
independent (no field/footprint/regional bias) sampling errors. Independent sampling
errors are seen when analyzing the field-scale soil moisture results. A comparison of the
predicted and observed field-scale soil moisture values result in an increase in deviation
from the spatial average (soil moisture mean) as shown by the increase in root mean
square error (RMSE) and decrease in R2. However, the level o f agreement suggests some
field-scale stability. These results are discussed in Section 3.3.2.
3.3. Results
Comparisons o f observations and soil moisture estimates were performed at the
footprint-scale for all days having both AMSR-E overpass and PSR C/X flights. Inter­
instrument comparison was possible for seven days o f the study (June 25, July 1, 4, 8, 9,
10 and 11). Soil moisture estimation was possible for five days o f the study, those days
which had instrument observations, in situ sampling and were not used for the retrieval
calibration (June 25, July 1, 8, 9, and 11). Unless otherwise noted, analysis was
75
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
performed at the footprint-scale (50 km). Daily footprint-scale statistics o f sensor and
surface parameters are listed in Table 3.4. The close proximity ( - 1 0 km) o f the AMSR-E
footprint centers resulted in an over-sampling of the region covered by one footprint,
noted by the very low range and standard deviation o f ASMR-E 10.7 GHz brightness
temperatures. The AMSR-E 6.9 GHz channels show a higher range and standard
deviation than the 10.7 GHz channel, due to radio frequency interference (RFI) discussed
in section 3.4. The range o f vegetation water content increased due to the differences in
maximum vegetation water content between the soybean -1.25 kg/m2 and com of -6
kg/m2 fields, (i.e. - range from 3.69 - 5.57 kg m2 from June 25 to July 11). The influence
of the rainfall events can be observed by the increase in volumetric soil moisture (both 01 cm and 0-6 cm) and a decrease in mean footprint-scale soil temperature o f 30.08 C
(drier) to 20.67 C (wetter) on July 10-11. A mean decrease in brightness temperature was
observed in all channels from July 9th to July 10th o f 7.31 K. Also, the heterogeneity the
rainfall increased the range and standard deviation of observed TB in all channels during
these days.
3.3.1. AM SR-E 10.65 GHz observations and simulations
The gridded AMSR-E and modeled TB were compared with the in situ field data.
Regional effects of soil moisture change on brightness temperatures were calculated for
each day of the study. The mean o f the AMSR-E 10.65 GHz horizontally-polarized TB
was compared with the model outputs for those days. Footprint-scale observed and
simulated average brightness temperatures are shown in Figure 3.4. The range of the
simulated brightness temperatures is -2 5 K compared to the observed brightness
76
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
temperature range o f -1 5 K. Simulated and observed ASMR-E footprint-scale horizontal
and vertical brightness temperatures, soil moisture, and temperature are shown in Figure
3.5. The simulated 10.65 GHz H-pol TB agrees well with the observed 10.65 GHz H-pol
T b for the first part o f the experiment, June 25 to July 4. After July 4, the observed and
simulated TB are less correlated. The simulated H-pol TB decreases more rapidly (280 K
to 275 K) from July 4th to July 7th than the observed H-pol TB (280 K to 275 K)
corresponding to the increase in footprint-scale volumetric soil moisture (0.12-0.26
m3/m3). The simulated H-pol TB falls from 290 K to 265 K and the observed H-pol TB
falls from. One explanation for this discrepancy could be that the vegetation water
content is attenuating the observed signal and the model does no properly account for this
effect.
3.3.2. PSR C /X observations and simulations
Figure 3.6 shows the mean footprint-scale estimated and observed average PSR
7.3 GHz brightness temperatures for 5 days o f the study. The model has a better
agreement for wetter observations. A strong equivalence is observed, giving an R2 value
of 0.88 and a root mean square error of 8.50 K. Similar results are observed when
comparing the PSR C/X 10.7 GHz modeled and observed brightness temperatures, with a
slightly lower (R2 = 0.84 and RMSE = 8.09 K) (Figure 3.7). Similar ranges o f observed
and simulated PSR 10.7 GHz TB are seen (20 K and 30 K respectively) with distinctly
better R2 in the PSR simulations than AMSR-E. Drier regions result in an overestimation
of average brightness temperature in both channels; the opposite is true for the lower
brightness temperatures (wetter regions). The soil moisture influence is illustrated in
77
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Figures 3.8-3.9. Both the estimated vertical and horizontal polarizations are higher during
June 25th through July 7th than July 8th through 12th. However, the general pattern during
the wetter portion of the experiment corresponds to lower TB, which implies an
underestimation o f wc by the RT model calibration and/or unusually large effects of
higher wc on observed PSR C/X TB.
The finer spatial resolution of the PSR C/X instrument allows the influence of the
spatial distribution o f ecological/vegetation variables to be observed. This is illustrated in
Figures 3.10 and 3.11. The influence of the rainfall event on July 10th is reflected in a
general decrease in both the regional-scale modeled and observed mean brightness
temperatures on July 11th. The southwestern portion o f the region received more rainfall
seen as a southwest-northeast gradient in brightness temperature values. In addition, the
western portion o f the SMEX02 region also had several water bodies, particularly in the
southeast. These water bodies result in significant decreases in PSR C/X and AMSR-E
brightness temperatures. The differences between the observed and simulated H-pol TB is
larger on July 4th (-20 to +20 K). On July 11th, the differences range between -2 0 to +20
K witnessed particularly in the positive. This bias is also seen in the line plots o f 3.5, 3.83.9 and suggests that the greater dynamic range o f the simulated values could be the
result of smaller sensing depth o f AMSR-E.
3.3.3. Comparison o f AMSR-E and PSR C/X
Figure 3.12 illustrates the mean footprint-scale 10.7 GHz horizontally-polarized
AMSR-E and PSR C/X brightness temperatures for June 25, July 1 ,4 , 8, 9, and 11. The
instruments are relatively well correlated (R2= 0.83). Greater deviation was observed in
78
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
the higher brightness temperatures. The deviation at higher brightness temperatures could
be result o f soil moisture heterogeneity within the entire SMEX02 region. As seen in
Table 3.4, the range o f 0-1 cm volumetric soil moisture was much lower during and after
the major rain events (i.e., a decrease in moisture content range o f .26 to .22). Table 3.5
lists the correlation values for the 6.9 and 10.7 GHz channels o f AMSR-E and 7.3 and
10.7 GHz channels o f PSR C/X at both moisture depths. Overall, the PSR C/X instrument
was seen to have the best correlation with soil moisture, both at 0-1 and 0-6 cm depths,
with the highest correlation being R2=0.8 for the mean domain-scale PSR C/X 7.3 GHz
vertically-polarized channel and 0-1 cm volumetric soil moisture content. With the
exception of the PSR C/X 10.7 H channel, both instruments had greater explained
variance with the 0-1 cm volumetric soil moisture (by a very minor margin).
3.4. Soil moisture prediction
A summary of mean domain-scale estimated moisture is presented in Table 3.6
with in situ 0-1 cm and 0-6 cm volumetric soil moisture. The PSR C/X 7.3 GHz channel
provided the best soil moisture prediction overall. Figure 3.13 compares the AMSR 10.7
GHz predicted soil moisture with the observed 0-1 cm soil moisture. Soil predictions
were better in the wetter regions/days of the study. The model has difficulty estimating
lower volumetric soil moisture values (<0.1 m3/m 3). This was also observed when
applying the PSR C/X data in the retrieval algorithm, as seen in Figure 3.14. The
modeled values had a smaller range than the observed 0-1 cm values (-0.2 and .03 for the
modeled and observed respectively). There are a number o f possible explanations for this
difference. This disagreement can possibly be attributed to the extremely dry state o f top
79
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
layer o f soil from June 25th to July 4th, when the mean o f the in situ 0-1 cm soil moisture
was < 0.04 m3/m3 in many fields. These dry conditions exaggerate an inherent weakness
in the model. In addition, errors in the other ancillary data (h, Q, wc, Ts) will have a
greater impact on the modeled brightness in dry areas. Such effects are further illustrated
in Table IV. In addition, when examining estimates for the wetter conditions, the AMSRE retrievals are more accurate. Mean estimates for the SMEX02 domain for the wet areas
resulted in much higher (R2=0.87) explained variance.
The days used in the calibration o f the model for each instrument will determine
the feasibility o f retrieving soil moisture on both wet and dry days. Ideally, the days used
for calibration would contain a range of all conditions observed. Given the small dataset
o f all co-located (space and time) remote sensing and ground truth data, only two days
were used for the calibration.
3.5. Heterogeneity
Analysis o f sub-scale and sub-pixel variability was accomplished examining the
range of observed PSR C/X brightness temperatures within the larger AMSR-E footprints
and SMEX02 domain region averages. Estimated field-scale soil moisture was compared
with footprint-scale and domain-scale soil moisture means. Figure 3.15 illustrates the
range o f observed PSR C/X values within a single AMSR-E footprint for all o f the co­
located days during the study. The correlation between the mean o f the PSR C/X values
and the AMSR-E values is reasonable (R2=0.83) considering the large-scale disparity
between the two instruments. However, the range o f values for PSR C/X is greater than
20 K for some instances. In Figures 3.16-3.17 the field-scale and footprint-scale values
80
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
are plotted against the field 0-1 cm volumetric soil moisture observations. The high
variance in the field-scale data results in very little correlation as seen in Figure 3.16.
When resampled to footprint-scale, a significant increase in R2 is observed (0.05 to 0.70).
This increase in agreement with scaling supports other work showing a decrease in
variability at larger spatial scales (McCabe, 2004).
3.6. Radio Frequency Interference
Notable radio frequency interference (RFI) from anthropogenic sources in the
region was observed in both the AMSR-E and PSR C/X data.
These effects are
manifested by extremely high brightness temperatures, particularly in the ASMR-E 6.9
GHz frequency (Li et al. 2004). The effects o f RFI in the PSR C/X data has been reduced
through the application of a spectral interference correction algorithm using four sub­
bands described in (Gasiewski et al. 2002). This introduced noise is observed in Table 3.4
by the large range and standard deviation for the AMSR 6.9 channels during all days.
Figure 3.18 compares the simulated and observed 6.9 GHz brightness temperatures. With
the exception of two cases, all o f the AMSR-E 6.9 GHz are underestimated. The AMSRE 6.9 GHz and PSR C/X 7.3 GHz channels are relatively close in frequency and we
would expect at least similar trends in the data when observing the same scene. On
average, these values are within ~3 K for both the horizontal and vertical polarizations.
However, when the footprint-scale PSR C/X 7.3 GHz brightness temperatures are
compared to the observed AMSR-E 6.9 GHz brightness temperatures, the ASMR-E 6.9
values exceed the PSR C/X 7.3 GHz values, up to 25 K in some cases.
81
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
3.7. Conclusions and discussion
Results demonstrated the ability of resampled observed and modeled brightness
temperatures to accurately reflect the spatial and temporal soil moisture variability during
the SMEX02 campaign. The observed spatial and temporal soil moisture heterogeneity
observed during SMEX02 was reproduced with some degree o f certainty for the wetter
regions.
The simulated brightness temperatures show a larger range from dry days to wet
days. A mean decrease in AMSR-E brightness temperature o f 18 K was observed for the
10.65 GHz horizontally polarized channels. Decreases in the modeled brightness
temperature were 26 K and 25 K for the 10.65 GHz horizontally and vertically-polarized
channel, respectively. The greater range in modeled brightness temperatures than
observed AMSR-E brightness temperatures could be a result o f sub-footprint
heterogeneity. The modeled brightness temperatures show a distinct bias that may be due
in part to incorrect parameterization and extremely dry soils in the 0-1 cm depth during
the study. Both AMSR-E and PSR C/X data show greater variability during the wetter
days, presumably due to variable rainfall distribution, but also possibly from an increase
in biomass differences between com and soybean during the elapsed period.
Radio frequency interference is an issue that requires continuous monitoring.
Today the C-band is affected by RFI. In the future, it is possible RFI could affect the Lband and other sensors in the Hydros (Entekhabi et al. 2004) and SMOS (Kerr et al.
2001) missions.
82
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
3.8. References
Ashcroft P, Wentz F (2003) AMSR-E/Aqua L2A Global Swath Spatially-Resampled
Brightness Temperatures (Tb) V001, June to July 2002. Boulder, CO, USA:
National Snow and Ice Data Center. Digital media.
Basara JB, Crawford KC (2002) Linear relationships between root-zone soil moisture and
atmospheric processes in the planetary boundary layer. Journal of Geophysical
Research-Atmospheres 107:1-10
Bindlish R, Jackson TJ, Gasiewski AJ, Klein M, Njoku EG (2005) Soil moisture mapping
and AMSR-E validation using the PSR in SMEX02. Remote Sensing of
Environment (In Press)
Bindlish R, Jackson TJ, Wood E, Gao HL, Starks P, Bosch D, Lakshmi V (2003) Soil
moisture estimates from TRMM Microwave Imager observations over the
Southern United States. Remote Sensing of Environment 85:507-515
Bolten JD, Lakshmi V, Njoku EG (2003) Soil moisture retrieval using the passive/active
L- and S-band radar/radiometer. IEEE Transactions on Geoscience and Remote
Sensing 41:2792-2801
Bolten JD, Narayan U, Guijarro L, Lakshmi V (2004) Passive-Active Microwave Remote
Sensing o f Soil Moisture at Both L- and C-band: A Comparison of Two Field
Experiments. Italian Journal of Remote Sensing 30/31:65-86
Brubaker KL, Entekhabi D (1996) Analysis o f feedback mechanisms in land-atmosphere
interaction. Water Resources Research 32:1343-1357
Choudhury BJS, T. J.; Newton, R.W.;Chang, A.T.C. (1979) Effect o f surface roughness
on microwave emission of soils. Journal of Geophysical Research 84:5699-5706
Cosh MH, Jackson TJ, Bindlish R, Prueger JH (2004) Watershed scale temporal and
spatial stability o f soil moisture and its role in validating satellite estimates.
Remote Sensing of Environment 92:427-435
Delworth T, Manabe S (1989) The Influence o f Soil Wetness on near-Surface
Atmospheric Variability. Journal of Climate 2:1447-1462
83
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Dirmeyer PA, Zeng FJ, Duchame A, Morrill JC, Koster RD (2000) The sensitivity of
surface fluxes to soil water content in three land surface schemes. Journal of
Hydrometerology 1:121 - 134
Dobson MC, Ulaby FT, Hallikainen MT, Elrayes MA (1985) Microwave Dielectric
Behavior of Wet Soil .2. Dielectric Mixing Models. IEEE Transactions on
Geoscience and Remote Sensing 23:35-46
Drusch M, Wood EF, Gao H, Thiele A (2004) Soil moisture retrieval during the Southern
Great Plains Hydrology Experiment 1999: A comparison between experimental
remote sensing data and operational products. Water Resources Research 40:-
Entekhabi D, Njoku EG, Houser P, Spencer M, Doiron T, Kim YJ, Smith J, Girard R,
Belair S, Crow W, Jackson TJ, Kerr YH, Kimball JS, Koster R, McDonald KC,
O'Neill PE, Pultz T, Running SW, Shi JC, Wood E, van Zyl J (2004) The
hydrosphere state (Hydros) satellite mission: An earth system pathfinder for
global mapping of soil moisture and land freeze/thaw. IEEE Transactions on
Geoscience and Remote Sensing 42:2184-2195
Fennessy MJ, Shukla J (1999) Impact o f initial soil wetness on seasonal atmospheric
prediction. Journal o f Climate 12:3167-3180
Gao BC (1996) NDWI - A normalized difference water index for remote sensing of
vegetation liquid water from space. Remote Sensing o f the Environment 58:257 266
Gao HL, W ood EF, Drusch M, Crow W, Jackson TJ (2004) Using a microwave emission
model to estimate soil moisture from ESTAR observations during SGP99. Journal
o f Hydrometeorology 5:49-63
Gasiewski AJ, Klein M, Yevgrafov A, Leuskiy V (2002) Interference mitigation in
passive microwave radiometry International Geoscience and Remote Sensing
Symposium, Toronto, Canada
Hollenbeck KJ, Schmugge TJ, Homberger GM, Wang JR (1996) Identifying soil
hydraulic heterogeneity by detection o f relative change in passive microwave
remote sensing observations. Water Resources Research 32:139-148
Hu ZL, Islam S (1997) Scale invariant remote sensing algorithms. IEEE Transactions on
Geoscience and Remote Sensing 35:747-755
84
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Jackson T, Cosh M ( 2003) SMEX02 Vegetation Water Content, Iowa Regional and
Walnut Creek Watershed. Boulder, CO: National Snow and Ice Data Center.
Digital media.
Jackson T, Schmugge J, Engman E (1996) Remote sensing applications to hydrology:
soil moisture. Hydrological Sciences 41:3167 - 3180
Jackson T, Williams S (2004) SMEX02 CODIAC Soil Climate Analysis Network
(SCAN) Data, Iowa. Boulder, CO. University Corporation for Atmospheric
Research, Joint Office for Science Support (UCAR/JOSS). Digital Media.
Jackson TJ, Chen DY, Cosh M, Li FQ, Anderson M, Walthall C, Doriaswamy P, Hunt
ER (2004) Vegetation water content mapping using Landsat data derived
normalized difference water index for com and soybeans. Remote Sensing of
Environment 92:475-482
Jackson TJ, Gasiewski AJ, Oldak A, Klein M, Njoku EG, Yevgrafov A, Christiani S,
Bindlish R (2002) Soil moisture retrieval using the C-band polarimetric scanning
radiometer during the Southern Great Plains 1999 Experiment. IEEE Transactions
on Geoscience and Remote Sensing 40:2151-2161
Jackson TJ, Hsu AY (2001) Soil moisture and TRMM microwave imager relationships in
the Southern Great Plains 1999 (SGP99) Experiment. IEEE Transactions on
Geoscience and Remote Sensing 39:1632-1642
Jackson TJ, Schmugge TJ (1991) Vegetation Effects on the Microwave Emission of
Soils. Remote Sensing o f Environment 36:203-212
Kawanishi T, Sezai T, Ito Y, Imaoka K, Takeshima T, Ishido Y, Shibata A, Miura M,
Inahata H, Spencer RW (2003) The Advanced Microwave Scanning Radiometer
for the Earth Observing System (AMSR-E), NASDA's contribution to the EOS
for global energy and water cycle studies. IEEE Transactions on Geoscience and
Remote Sensing 41:184-194
Kerr YH, Njoku EG (1990) A Semi empirical Model for Interpreting Microwave
Emission from Semiarid Land Surfaces as Seen from Space. IEEE Transactions
on Geoscience and Remote Sensing 28:384-393
85
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Kerr YH, Waldteufel P, Wigneron JP, Martinuzzi JM, Font J, Berger M (2001) Soil
moisture retrieval from space: The Soil Moisture and Ocean Salinity (SMOS)
mission. IEEE Transactions on Geoscience and Remote Sensing 39:1729-1735
Koster R, Milly PCD (1997) The interplay between transpiration and runoff formulations
in land surface schemes used with atmospheric models. Journal of Climate
10:1578-1591
Kustas WP, French AN, Hatfield JL, Jackson TJ, Moran MS, Rango A, Ritchie JC,
Schmugge TJ (2003) Remote sensing research in hydrometeorology.
Photogrammetric Engineering and Remote Sensing 69:631-646
Lee KH, Anagnostou EN (2004) A combined passive/active microwave remote sensing
approach for surface variable retrieval using Tropical Rainfall Measuring Mission
observations. Remote Sensing of Environment 92:112-125
LeVine DM, Karam MA (1996) Dependence of attenuation in a vegetation canopy on
frequency and plant water content. IEEE Transactions on Geoscience and Remote
Sensing 34:1090-1096
Li L, Njoku EG, Im E, Chang PS, Germain KS (2004) A preliminary survey o f radio­
frequency interference over the US in Aqua AMSR-E data. IEEE Transactions on
Geoscience and Remote Sensing 42:380-390
McCabe MF, Gao BH, Wood EF (2004) An evaluation o f AMSR-E derived soil moisture
retrievals using ground based, airborne and ancillary data during SMEX02.
Journal o f Hydrometeorology
McCabe MF, Wood EF, Gao BH (2005) Initial soil moisture retrievals from AMSR-E:
Multiscal comparison using in situ data and rainfall patterns over Iowa.
Geophysical Research Letters 32
Miller DA, White RA (1998) A Conteminous United States Multi-Layer Soil
Characteristics Data Set for Regional Climate and Hydrology Modeling. Earth
Interactions 2
Mo T, Choudhury BJ, Schmugge TJ, Wang JR, Jackson TJ (1982) A Model for
Microwave Emission from Vegetation-Covered Fields. Journal o f Geophysical
Research-Oceans and Atmospheres 87:1229-1237
86
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Montaldo N, Albertson JD (2003) Temporal dynamics of soil moisture variability: 2.
Implications for land surface models. Water Resources Research 39
Moran MS, Peters-Lidard CD, Watts JM, McElroy S (2004) Estimating soil moisture at
the watershed scale with satellite-based radar and land surface models. Canadian
Journal o f Remote Sensing 30:805-826
Narayan U, Lakshmi V, Njoku EG (2004) Retrieval o f soil moisture from passive and
active L/S band sensor (PALS) observations during the Soil Moisture Experiment
in 2002 (SMEX02). Remote Sensing of Environment 92:483-496
Njoku E (2004) Soil moisture field experiment special issue. Remote Sensing of
Environment 92:425-426
Njoku EG, Entekhabi D (1996) Passive microwave remote sensing of soil moisture.
Journal o f Hydrology 184:101-129
Njoku EG, Jackson TJ, Lakshmi V, Chan TK, Nghiem SV (2003) Soil moisture retrieval
from AMSR-E. IEEE Transactions on Geoscience and Remote Sensing 41:215229
Njoku EG, Li L (1999) Retrieval of land surface parameters using passive microwave
measurements at 6-18 GHz. IEEE Transactions on Geoscience and Remote
Sensing 37:79-93
Njoku EG, Wilson WJ, Yueh SH, Dinardo SJ, Li FK, Jackson TJ, Lakshmi V, Bolten J
(2002) Observations o f soil moisture using a passive and active low-frequency
microwave airborne sensor during SGP99. IEEE Transactions on Geoscience and
Remote Sensing 40:2659-2673
Njoku EGK, N. (1977) Theory for passive microwave remote sensing o f near-surface soil
moisture. Journal o f Geophysical Research 82
Pielke RA (2001) Influence of the spatial distribution of vegetation and soils on the
prediction o f cumulus convective rainfall. Reviews of Geophysics 39:151-177
Piepmeier JR, Gasiewski AJ (2001) High-resolution passive polarimetric microwave
mapping o f ocean surface wind vector fields. IEEE Transactions on Geoscience
and Remote Sensing 39:606-622
87
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Schmugge T, Jackson TJ (1994) Mapping Surface Soil-Moisture with Microwave
Radiometers. Meteorology and Atmospheric Physics 54:213-223
Shibata A, Imaoka K, Koike T (2003) AMSR/AMSR-E level 2 and 3 algorithm
developments and data validation plans o f NASD A. IEEE Transactions on
Geoscience and Remote Sensing 41:195-203
Ulaby FTM, R.K., Fung, A.K. (1986) Microwave remote sensing: Active and passivevoume scattering and emission theory, advanced systems and applications.
Dedham, MA: Artech Flouse.
Van de Griend AA, Wigneron JP (2004) The b-factor as a function of frequency and
canopy type at h-polarization. IEEE Transactions on Geoscience and Remote
Sensing 42:786-794
Wang JR, Engman ET, Mo T, Schmugge TJ, Shiue JC (1987) The Effects of SoilMoisture, Surface-Roughness, and Vegetation on L-Band Emission and
Backscatter. IEEE Transactions on Geoscience and Remote Sensing 25:825-833
Wang JR, Schmugge TJ (1980) An empirical model for the complex dielectric
permittivity o f soil as a function o f water content. IEEE Transactions on
Geoscience and Remote Sensing GE-23:35-46
Wigneron JP, Calvet JC, Pellarin T, Van de Griend AA, Berger M, Ferrazzoli P (2003)
Retrieving near-surface soil moisture from microwave radiometric observations:
current status and future plans. Remote Sensing o f Environment 85:489-506
Wigneron JP, Parde M, Waldteufel P, Chanzy A, Kerr Y, Schmidl S, Skou N (2004)
Characterizing the dependence o f vegetation model parameters on crop structure,
incidence angle, and polarization at L-band. IEEE Transactions on Geoscience
and Remote Sensing 42:416-425
88
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Center Frequencies (GHz)
Bandwidth (MHz)
Sensitivity (K)
Mean Spatial Resolution (km)
IFO V (km x km)
Sampling Rate (km x km)
Integration Tim e (msec)
Main Beam Efficiency (%)
Beamwidth (degrees)
6.925
350
0.3
56
74x43
1 0 x 10
2.6
95.3
2.2
10.65
100
0.6
38
51 x 30
1 0 x 10
2.6
95
1.4
18.7
200
0.6
21
27 x 16
1 0 x 10
2.6
96.3
0.8
23.8
400
0.6
24
31 x 18
1 0 x 10
2.6
96.4
0.9
36.5
1000
0.6
12
14x8
1 0 x 10
2.6
95.3
0.4
Table 3.1. AMSR-E instrument characteristics.
89
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
89
3000
1.1
5.4
6x4
5x5
1.3
96
0.18
Center Frequencies (GHz)
Bandwidth (MHz)
Beamwidth (degrees)
5.97
330
10
6.47
330
10
6.92
350
10
7.32
350
10
Conical Scanning
Incidence Angle (deg)
45
Flight Altitude (km)
7.3
10.71
200
7
10.69
20
7
10.65
1900
7
Polarization (polarimetric)
V and H, 3rd and 4th Stoke's Parameters in 6.75-7.10 GHz
Resolution (km) min,max
@ 6 / 1 0 GHz
0.8/0.5, 2.8/2.0
Table 3.2. PSR Instrument characteristics for SMEX02 region flights.
90
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
(a) Media & Sensor Parameters
Vegetation:
Single scattering albedo, co
Opacity coefficient, b
|
AMSR
0
6.9 GHz
0.15(com), 0.2(soy)
10.7 GHz
0.15(com), 0.25(soy)
|
PSR C/X
0
7.3 and 10.7 GHz
0.15(com), 0.25(soy)
Soil:
Roughness parameter, h
Polarization mixing parameter, Q
Bulk density (g cm'3), bd
Sand and clay mass fractions, c & s
Sensor:
Viewing angle PSR C/X AMSR (deg)
Frequency (GHz)
Polarization
(b) M edia Variables
Land Surface:
Volumetric soil moisture, m v (m3 m"3)
Vegetation water content, w c (kg m'2)
Vegetation temperature, Tee (C)
Surface temperature, Ts (C)
6.9 GHz
0(com), 0.2(soy)
10.7 GHz
0(com), 0.2(soy)
0
7.3 and 10.7 GHz
0.1 (com), 0.2(soy)
0
in situ
in situ
CONUS-SOIL dataset
CONUS-SOIL dataset
55
6.9, 10.7
55
7.3, 10.7
H, V
H ,V
in situ
in situ
in situ , LandSat TM,
in situ , LandSat TM,
ETM+
in situ, SCAN site air
temperature
in situ, SCAN site air
in situ
in situ
ETM+
temperature
Table 3.3. Radiative transfer model inputs for AMSR and PSR C/X simulations.
91
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
25-Jun
1-Jul
4-Jul
8-Jul
9-Jul
10-Jul
11-Jul
m ean
2 9 6 .4 3
290 .0 0
292.81
287.77
294.99
281 67
281.81
range
29 .9 8
14.79
22.81
8.49
22.23
13.39
34.60
stdev
9.1 2
4.95
8.00
3.01
7.78
4.77
11.98
A M SR 6.9H (K)
AMSR 6.9V (K)
m ean
30 4 .8 2
301.23
305.29
297.88
302.49
292.22
285.68
range
18.36
15.14
13.14
11.01
23.47
14.45
20.18
stdev
5.73
4.92
4.33
3.72
8.76
4.48
7 14
264.79
AMSR 10.7H (K)
m ean
279.41
284 .0 4
281.13
283.18
281.89
2 7 4.58
range
0.67
0.32
2.41
0.61
0.62
1.29
2.76
stdev
0 .2 0
0.09
0.70
0.24
0.17
0.40
0.87
m ean
2 8 9 .2 4
290.42
287.83
2 8 8.97
288.25
280.77
272.03
range
0.54
0.60
2.20
0.91
0.86
1.53
2.52
stdev
0.17
0.21
0.69
0.33
0.28
0.51
0.88
2 7 5 .8 0
27 9 .3 0
279.14
275.74
278.68
270.01
264.52
15.68
AMSR 10.7 V (K)
.......
PSR 7 .3 H (K)
m ean
.
range
9.95
9.97
29.12
11.41
10.12
10.88
stdev
2.52
2.74
5.03
2.97
2.53
2 31
3.66
P SR 7 .3 V (K)
.
m ean
2 8 7 .0 0
287 .3 3
286.86
2 8 4.16
286.11
2 7 7.99
272.19
range
12.09
10.67
21.06
8.32
8.08
6.78
14.80
stdev
2 40
2.84
3.78
2.32
2.18
1.89
3.40
m ean
2 78 24
28 0 .0 4
282.03
277.90
280.98
2 7 3.52
268.01
range
15 26
15.72
23.82
11.57
10.88
10.22
16.36
stdev
3 .6 5
3.94
4.31
3.21
2.58
2.33
3.41
m ean
2 8 6 .6 8
286.53
287.69
2 8 4.26
284.26
2 7 9.87
273.22
range
12 32
15.03
22.46
15.65
12.76
9.19
13.72
stdev
3.11
4.26
4.06
3.28
2.61
2.48
3.16
0.04
0.04
0.U6
0.22
0.19
0.34
0.34
P SR 10.7 H (K)
PSR 10.7 V (K)
m„ 0-1 cm (m1’ m"’)
m ean
~
range
0.12
0.10
0.24
0.42
0.38
0.26
0.22
stdev
0.02
0.02
0.05
0.10
0.10
0.06
0.06
0.29
-
m„ 0-6 cm (m J m'J)
m ean
0.16
0.13
0.14
0.24
0.22
0.29
range
0.19
0.23
0.19
0.30
0.26
0.17
0.15
stdev
0.0 4
0.05
0.05
0.06
0.06
0.04
0.04
w t (k g m * )
m ean
1.53
2.34
2.59
2.90
3.09
3.18
3.27
range
3.69
4 .4 3
4.79
5.22
5.33
5.45
5.57
stdev
1.25
1.77
1.83
1.99
2.02
2.08
2.13
Ts 5 cm (C)
m ean
34 .8 7
34.02
32.83
30.01
30.08
2 6 .1 5
20.67
range
17.50
16.30
18.70
13.70
12.20
10.50
4.05
stdev
4 .0 6
4 .7 5
5.03
3.06
3.33
2.14
0.97
Table 3.4. Observed SMEX02 regional mean daily statistics for instrument and in situ
data.
92
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
AMSR 10.7 GHz II
AMSR 10.7 G i l / Y
\ M S R 6 G I I / II
XMSR6GH/ \
I'SK 10.7 ( i l l / II
PSR 10.7 ( i l l / V
PSR 7.3 GHz H
PSR 7.3 GHz V
mv 0-1 cin
mv 0-6 cm
R2
0.409
0.462
0.218
0.279
0.691
0.767
0.701
0.800
R2
©
<1
N>
Observed TB
0.389
0.157
0.212
0.695
0.752
0.694
0.769
Table 3.5. R2 o f observed TB and near surface soil moisture.
93
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
in situ m, 0-1 cm
Estimated mv
AMSR 10.7 ( i ll/
PSR 10.7 (JM/
PSR 7.3 ( i l l ,
m situ m* 0-6 cm
R2
RMSK (in' m ')
Bias (ni1m 'i
R2
0.436
0.711
0.738
0.127
0.099
0.102
-0.093
-0.073
-0.078
0.353
0.705
0.722
RMSE (m 'm '3) Bias (m3 m'J)
0.066
0.040
0.041
-0.048
-0.028
-0.033
Table 3.6. Correlation statistics of estimated and observed AMSR footprint volumetric
soil moisture.
94
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
IO W A
SMEX02 Region
'Des Moines
SO ton
Figure 3.1. Location o f SMEX02 region.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
290
R2 = 0.44
t
285 j
♦
' .
♦♦
♦
AMSR
Average 10.7 GHz Tj
♦
♦
'
280 j
'
.
^
♦
^
♦
♦
'
♦
275
270
265
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
V olum etric Soil M oisture (0-1 cm )
Figure 3.2. Mean observed AMSR 10.7 GHz average TB vs. observed 0-1 cm volumetric
soil moisture.
96
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Figure 3.3. Spatial distribution of PSR 7.3 GHz Horizontally-polarized TB.
97
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
300
R =0.61
RMSE = 6.47 K
295
g
290
CO
H
N
DC 285 -
0
©
8, 280
F-t
0>)
1
275
w
270 -
■0d
s
1I
Pi
uy
s
<
265 -
260
255
260
265
270
275
280
285
290
295
300
AMSR-E Observed Average 10.7 GHz TB (K)
Figure 3.4. Comparison of mean observed and mean simulated AMSR 10.7 GHz
footprint-scale brightness temperatures.
98
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
320
0.4
310
0.35
300
3
£
1i-H
CD
Q,
E
H
CD
B
bp
ffl
(J
0.3
290
0.25
280
270
0.2
260
0.15
250
0.1
240
0.05
230 -
0
220 6/25 7/1
"s
o
O)
3
'o
*o
on
u
*B
6
3
o
>
7/2 7/4 7/6 7/7 7/8 7/9 7/10 7/11
I Observed Soil Moisture (0-1 cm)
I A M SR Predicted Volumetric Soil Moisture
I Observed Volumetric Soil Moisture (0-6cm )
- Observed AM SR 10.65H GHz TB
■Observed A M SR 10.65V GHz TB
■Estimated 10.65H GHz TB
-Estimated 10.65V GHz TB
- Soil Temperature (5cm )
Figure 3.5. Comparison of simulated AMSR 10.7 GHz horizontal and vertical polarized
Tb with in situ soil moisture and observed TB.
99
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
300 1r~i r i l i i | r r r i 1 11 1' f 'pTT'TTT I i i T~yri~T i i i r i i^
X
/ ■ R squared = 0.88
XX
• RMSE = 8.50 K
^
!
/
xX x
.
/
x JfX
w
/
CD
i— 2 9 0
/
<u
/
X
✓
E
m
■
>
'
/
I
>X
/
N
.
/
I
■
/
ts>
/
2
SO
<
N
/
Kl
'
/X
'
.
/
CD
/
■
■
lL
/
XI
<u
X
X
/
270 '
X
/■ X
/
.
/
X
•
✓ X
■ /
■
!/
2 6 0 f \ 1 1 1 1 I...1 i l l 11 t 1 11 >a l l 1 II 1 1i i i 1 i ,i i i i i i i i
260
270
280
290
Observed PSR 7 ,3 2 GHz Average TB (K)
300
Figure 3.6. Mean estimated vs. observed average PSR 7.3 GHz TB.
100
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
300
00
I— 2 3 0
V
E
CD
>
<
N
I
a
I 1 1 1 1 1 1 1 1 j ! 1 1 r r r t 1 1 1 1 1 1 1 1 i i 1 1 1 1 1 1 1 1 1 1 1 1^,
X
/ R s q u a re d = 0 .8 4
/
X X
RMSE = 8 . 0 9 K
J
x x X
x
r
■
■
CE
C/l
CL
TS
V
/
X
/
/
X
/
/
/
-
m
/
■
2 SO -
o
/
*
.
/
/
/
“
.
-
SX
.
■
.
/
-
/
■
■
_
-
j
/
y
-
X
/
270 —
X
X
.
/
X
*
X
/
■ /
;/
2 6 0 f i 1 1 1 1 1 1 1 1 1 1 1 i i i 1 1 1 i i i i i 1 1J—L111 1 1 1 1 1 1 1
260
270
280
290
Observed PSR 1 0 .7 GHz A verage TB (K )
300
Figure 3.7. Mean estimated vs. observed average PSR 10.7 GHz TB.
101
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
280 ^
t*
<D
H
<u
s
■si
oo
'C
oa
6/25 6/27 6/29 7/1 7/4 7/8 7/9 7/10 7/11 7/12
I Observed Soil Moisture (0-1 cm)
I PSR Predicted Volumetric Soil Moisture
I Observed Volumetric Soil Moisture (0-6cm)
■Observed PSR 10.65H GHz TB
■Observed PSR 10.65V GHz TB
■Estimated 10.65H GHz TB
■Estimated 10.65V GHz TB
- Soil Temperature (5 cm)
Figure 3.8. Comparison o f simulated PSR 10.7 GHz horizontal and vertical polarized Tb
with in situ soil moisture and observed Tb.
102
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
S-i
<
L>
I*
u
H
o>
£
JS
ap
'C
m
6/25 6/27 6/29 7/1 7/4 7/8 7/9 7/10 7/11 7/12
IObserved Soil Moisture (0-1 cm)
IObserved Volumetric Soil Moisture (0-6cm)
■Observed PSR 7.3V GHz TB
•Estimated 7.3V GHz TB
IPSR Predicted Volumetric Soil Moisture
■Observed PSR 7.3H GHz TB
•Estimated 7.3H GHz TB
■Soil Temperature (5 cm)
Figure 3.9. Comparison o f simulated PSR 7.3 GHz horizontal and vertical polarized TB
with in situ soil moisture and observed TB.
103
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Figure 3.10. Spatial distribution o f simulated and observed PSR 7.3 GHz horizontally
polarized T b .
104
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Observed 10.7 GHz
Simulated 10.7 GHz
(Observed - Simulated)
10.7 GHz
July 4th
July Uth
Figure 3.11. Spatial distribution o f simulated and observed PSR 10.7 GHz horizontally
polarized TB.
105
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
R =0.83
260
265
270
275
280
285
290
AMSR 10.7 GHz H (K)
Figure 3.12. Comparison of mean observed domain-scale AMSR 10.7 GHz and PSR 10.7
GHz horizontally polarized TB.
106
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
0.4
R = 0.44
RMSE=0.13
AMSR
10 GHz predicted m
0.35
0.3
♦
0.25
♦
♦
0.2
♦
♦♦
0.15
0.1
0.05
0.1
0.2
0.3
0.4
in situ mv 0-1 cm
Figure 3.13. AMSR 10.7 GHz predicted soil moisture vs. in situ 0-1 cm soil moisture.
107
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
0.4
E
0
1
o
£
o
0
<f>
%
4—*
E
3
O
T>
TT i i i r~i i i | i i i i i r~r tt~[ \ i
:
0 .3 7
.
’
.
.
0.2 _
■
x
/
1
.
.
■
/
XX
X
,
/
/
X
.
.
0.1 -
/
X
*
,
/
/
.
■
■
/
.
■
s
/
/
/
CL
to
CL
„
X
X
.
■
■
ET
R s q u a r e d = 0 .7 3
RMSE = 0 . 1 0
>
<p
i t i i I I | i i I i t i i i xy
/ /
/
'
/
/
/
/
ji
x x
:
/
X
/
X
/
x ✓
.
/
.
■
/
•
/
■
■ /
;/
0.0 f 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 ■ 1 1 1 1 1 1 1 1 1 i i 11
0, 0
0.1
0.2
0.3
Q,
In—Situ V olumetric Soil M oisture 0 - 1 c m
Figure 3.14. PSR 7.3 GHz predicted soil moisture vs. in situ 0-1 cm soil moisture.
108
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
290
R = 0.83
285
g
280
O 275
270
265
260
260
265
270
275
280
285
290
PSR 10.7 GHz H-pol (K)
Figure 3.15. Range o f mean observed footprint-scale PSR 10.7 GHz horizontallypolarized Tb v s . observed AMSR 10.7 GHz horizontally polarized TB.
109
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
3
Q Q
I
I
*
I I
I
I
I I
I
I
I |
I
I I
I
I
I I
I' T " |
I
I
I I
I
f"l I
I
|
I
I
I
I I
I
I
I I
R s q u a r e d = 0 .0 5
230
v
W X X * XXXX
oc 2 7 0
n
S
260
.
I
I.
0.1
0 .2
0 .3
V olum etric Soil M oisture 0 —1 c m
0 .4
Figure 3.16. PSR C/X horizontally polarized TB.
110
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
300
TT TTT I
1 T'T J T " I
T1
'!' '1 T
1
|
1 1 F II"
1 R sq u a re d = 0 .7 0
J. 290
_A
-j
*
X
XX
2 SO
X
^
(X
<
|0
CL
T 1 1 11 1 | 1 11 1 1 1
1
N
o
<
KN
J
X
X
270 7
X X
X
o
>
uU
o>
I
<
X
X
260
o
£
250
l l l l l l l l l l l . i l .....................1 1 1 1 . I
5.0
0.1
0.2
0.3
0.4
Footprint Averaged Volumetric Soil Moisture 0 —1 cm
Figure 3.17. Mean o f PSR C/X horizontally-polarized TB within each AMSR-E
footprints.
Ill
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
320
: R squared = 0.45
: RM5E = 12.4
» 290
•a>c
■i 200
£
<n
cn
<n
270
2 60
260
270
290
300
310
AMSR Observed Average 6.9 GHz Tb (K)
320
Figure 3.18. Observed vs. simulated AMSR 6.9 GHz T b.
112
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
315 1
R =0.52
305
295
&
N
EC
o
285
♦♦
* ♦
oi
on
Oh
275
265
255
255
265
275
285
295
305
315
AMSR 6.9 GHz (K)
Figure 3.19. Observed AMSR 6.9 GHz average TB vs. observed PSR 7.3 GHz average
Tb.
113
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
CHAPTER 4
AMSR-E OBERSERVATIONS DURING THE SOIL MOISTURE
EXPERIMENTS 2004 (SMEX04)
4.1. Introduction
This chapter provides a preliminary assessment o f AM SR-E performance in a
lightly vegetated terrain through the examination o f AM SR-E soil moisture retrievals
during SMEX04. Further analysis in this region will be continued later to determine the
impacts in the AM SR-E retrieval process in a longer time-scale and also in more
topographically diverse regions involved in the study.
The Soil Moisture Experiments in 2004 built on the preceding experiments
(SGP99, SMEX02) by focusing on soil moisture remote sensing in areas o f topography
and varying vegetation conditions. The experiment was designed to coincide with the
North American Monsoon Experiment (NAME). The goal o f NAME was to improve the
prediction o f warm season precipitation in the interior North American continent by
means o f intense monitoring o f climate conditions during the monsoon season (mid-July
to mid-August) (Weiss et al. 2004, Gochis et al. In Press). These studies were designed to
further the principle o f climatological memory provided by the land surface and monitor
114
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
the positive feedback between soil moisture and rainfall (Eltahir 1989, Delworth &
Manabe 1993). In order to accurately assess the role o f the land surface soil moisture,
observations and simulations were necessary. For a complete discussion o f the North
American Monsoon System (NAMS), refer to the North American Monsoon Experiment
(NAME) Science Working Group at [http: //www.cpc.ncep.noaa.gov/ products/ precip/
monsoon/ NAME.htm].
4.2. Methods
4.2.1. SMEX04
This chapter focuses on observations from AMSR-E over the Walnut Gulch
Watershed during SMEX04. The role o f SMEX04 included intense monitoring during the
summer of 2004 with in situ soil moisture networks and in August using manual
sampling from groups o f ground teams, and mapping o f soil moisture from aircraft and
satellite. This experiment extended the framework and methodology for large-scale land
surface characterization demonstrated in SGP99 and SMEX02. The field experiment was
located in southern Arizona and northwestern Mexico. Two regional study sites, each
approximately 50 km by 75 km, in Arizona (AZ) and Sonora (SO), Mexico were
investigated. The two locations were quite different. Arizona has moderate topography
and sparse vegetation, and Sonora has strong topographic gradients and moderate
vegetation (Figure 4.1). The regional relief in the Sonoran region is dominated by
M exico’s largest mountain range, the Sierra Madre Occidental. At the time of this
analysis, processing o f the Sonora region in situ data was not completed. Therefore we
115
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
discuss the Arizona region exclusively. AMSR-E observations over the Sonora region
during SMEX04 will be the topic o f a future study.
The Arizona study area included the Walnut Gulch Experimental Watershed
O
0
(WG), which encompasses 150 km2 in southeastern Arizona, U.S.A. (31 43'N, 110
41'W) near Tombstone, AZ. The dominant vegetation within WG is brush and grass
covered rangeland. This area has relatively low topographic relief; with elevation ranging
from 1250 m to 1585 m. The Walnut Gulch Experimental Watershed soils consist mainly
of limestone alluvial fill. The dominant soil type is well-drained, calcareous, gravelly
loam.
Soil moisture within the watershed was well characterized by 21 electronicmeasuring, digital-recording devices. Nineteen o f these sites were located at raingage
stations (a total o f 83 raingages were in operation during the study). Soil moisture was
recorded using automated hydra probes located 5 cm depth, continuously recorded at 30minute intervals. Three o f the sites have a shallow profile array o f 3 sensors at 5, 15 and
30 cm. Mean watershed values o f the 5 cm depth probes at 30-minute intervals have been
used in this analysis. Walnut Gulch has been the focus of previous microwave remote
sensing validation studies; for further description o f the watershed characteristics and
past field campaigns in the region, the reader is referred to (Kustas et al. 1991, Jackson et
al. 1993).
The annual precipitation within both the AZ and SO regions is dominated by
warm-season convection which is affected by regional topography and surrounding
bodies of seawater. Both regions are noted for high intensity rain events that often lead to
flash flooding in areas o f low infiltration. Multiple localized precipitation events were
116
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
observed during the experiment, which allowed soil moisture wetting and dry-down
patterns to be observed.
4.2.2. Topography
An important component of SMEX04 was the study o f microwave remote sensing
in areas with high topographic gradients. As discussed earlier, soil moisture retrieval
from microwave remote sensing is influenced by many factors such as vegetation, surface
roughness, soil type and spatial resolution. One important land surface characteristic that
influences both the spatial variability o f soil moisture and microwave emission from the
surface is topographic relief. Effects of relief are o f special significance to microwave
radiometry due to the change in moisture patterns (and resulting brightness variation)
with changes in terrain (Albertson & Montaldo 2003).
In addition, variable topography o f land surfaces can result in shadowing of the
surface emission by neighboring terrain, or a net increase in radiation can occur and
enhance the effective emission (Matzler & Standley 2000). These effects are seen on a
range of scales, from micro-topography on a tilled surface to macro-topography such as
the Sierra Madre Occidental in Sonora. However, the main influence o f topography in all
scales is the increase in signal heterogeneity. As discussed in Chapter 3, the issues of
scale and heterogeneity are inter-related. Therefore, in order to best estimate soil moisture
using remote sensing, it is important to understand not only the effects o f topography on
hydrology and climate, but also on surface emission and scattering.
Previous studies have demonstrated the influence o f large-scale topography in
hydrology (Mohanty & Skaggs 2001), ground-based sampling characterization (Jacobs et
117
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
al. 2004), and microwave emission (Charpentier & Groffman 1992, Kim & Barros 2002).
Recently, progress has been achieved with the use o f topography in hydrologic modeling
to include lateral flow mapping and spatial distribution conditioning using remote sensing
(Wigneron et al. 2003, Ivanov et al. 2004, Chen et al. 2005).
In summary, these studies have illustrated that unless remote sensing techniques
can better characterize the influence o f topography (in order to reduce the introduced
error), their usefulness will be limited. The current research extends these previous efforts
by presenting new insight to the influence o f topography on the received brightness
temperatures from AMSR-E and the AMSR-E soil moisture product.
4.2.3. Operational model moisture estimates
Soil moisture estimates from a numerical weather prediction (NWP) model are
used to provide both another basis o f comparison as well as an evaluation o f this model
product. As an alternative to in situ or remotely sensed soil moisture values for the study
o f land-atmosphere interactions, NWP model soil moisture products have proved to be
useful (Robock et al. 1998, Entin et al. 2000). Most widely used are the reanalysis
products, which have global coverage and long time ranges. The soil moisture calculated
by reanalyses depends on the land surface scheme used, the forcing (particularly
precipitation and solar insulation), and the nudging employed. Entin et al. (2000)
provided a review of three well known soil moisture reanalysis products, including a
similar model to the one provided by the European Centre for Medium-Range Weather
Forecasts (ECMWF) used in this analysis.
118
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
This study uses a new operational analysis provided by ECMWF. Model products
include soil moisture data for four different soil layers: 0-7, 7-28, 28-100, and 100-289
cm. Wilting point, field capacity, and saturation are prescribed with 0.171, 0.323, and
0.472 m3 m-3, respectively, and are uniform for the vertical and constant for the entire
globe. The surface scheme within the operational Integrated Forecast System (IFS) is the
Tiled ECMWF Scheme for Surface Exchanges over Land (TESSEL) (van den Hurk et al.
2000). Soil moisture data used for this study are obtained from the operational analysis at
18 and 6 UTC, which correspond approximately to the AMSR-E overpasses at 1:30 PM
and 1:30 AM local Arizona time. The analysis is an Optimal Interpolation system based
on the modelled first guess and analysis increments o f 2 m temperature and relative
humidity (Douville et al. 2000). Soil moisture and screen level parameters are analysed
every six hours before the atmospheric analysis is performed. Soil temperatures at 18
UTC are the short range forecast from 12 UTC lead time. For the SMEX04 period, the
horizontal resolution corresponds to ~ 40 km.
4.2.4. AM SR-E products
Horizontally-polarized brightness temperatures from the 6.9 GHz and 10.7 GHz
AMSR-E (Kawanishi et al. 2003) channels and AMSR-E derived soil moisture values
were compared to in situ and ECMWF soil moisture over the Arizona and Sonora
regions. The Level-3 land surface product (AE_Land3) data were used in this research.
The AE_Land3 data are provided in a resampled Equal-Area Scalable Earth Grid (EASEGrid) format that has been designed for use with global-scale gridded data to facilitate the
surface type classification and retrieval steps. The EASE-Grid data are resampled to a
119
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
global cylindrical 25 km cell spacing. Data from both the ascending (1:30 PM local time)
and descending (1:30 AM local time) orbits are used in this study.
The current AMSR-E soil moisture retrieval algorithm is based on a changedetection approach using the polarization ratio (PR) in 10.7 GHz and 18.7 GHz. The use
of the polarization ratio has been shown to be insensitive to surface temperature, which
can facilitate soil moisture retrieval. The algorithm first computes a parameter to typify
vegetation and roughness using the PR 10.7 GHz and PR 18.7 GHz, plus three empirical
coefficients. Soil moisture is computed using deviation of the PR 10.7 GHz from a
monthly minima o f PR 10.7 GHz for that pixel over an annual cycle. The microwave
emission model and AMSR-E soil moisture algorithm development are described (Njoku
et al. 2003, Njoku et al. 2004).
4.3. Preliminary Results
Figure 4.2 shows AMSR-E volumetric soil moisture for August 5th, 2004 over the
NAME TIER 1 region for NAME. Spatial patterns o f wetness are observed. East-west
gradients o f moisture that are associated with regional topography are evident. Figure 4.3
shows the 1:30 AM overpass comparison results. The mean daily values o f precipitation
and 0-5 cm volumetric soil moisture the Walnut Gulch soil moisture network are plotted
against the ECMWF Layer 1 (0-7 cm) soil moisture estimates and AMSR-E soil moisture
product. It can be seen that the in situ moisture observations have a high correlation with
the precipitation events. The region had seven periods o f rainfall on two consecutive
days. The first major rain event within the watershed occurred on July 12th and was
followed by four days o f consecutive rainfall greater than 1 mm. This increased wetting
120
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
did not allow a full drydown o f the soil and an increase in soil moisture values was seen
in all datasets. These rainfall events are emphasized in this figure by the large increase in
moisture and sustained wetness.
The AMSR-E soil moisture values are well correlated with the precipitation
events and in situ moisture values. The largest response in AMSR-E occurred after the
July 28th rainfall, with a 0.03 cm3 cm'3 increase in moisture from a cumulative
precipitation amount o f 7.6 mm. The range o f the AMSR-E estimates for the SMEX04
period is small, however drying and wetting within the watershed is apparent. Dynamic
ranges for each data set were 0.048 cm3 cm'3, 0.15 cm3 cm'3, and 0.23 cm3 cm'3 for the
AMSR-E, in situ and ECMWF datasets, respectively.
Figure 4.4 shows a similar analysis for the 1:30 PM data. All o f the datasets show
a similar trend with precipitation events, as seen in the 1:30 AM data. However, the in
situ data are generally wetter, which could be attributed to the difference in
sampling/sensing depth o f the datasets. The AMSR-E estimates do not reflect this general
increase. In fact, there is a negative bias and more subtle moisture signal when compared
with the morning overpass. The ECMWF estimates are generally the same for both AM
and PM except for a few small increases in modeled moisture in mid June and July. This
consistency between AM and PM estimates could be a function o f the large time step
used (every six hours).
The effects o f Radio Frequency Interference on the 6.9 GHz channels are
demonstrated in Figure 4.5. When plotted against in situ soil moisture during the 1:30
AM overpass, the 6.9 GHz brightness temperatures are higher than the 10.7 GHz, an
indication that noise has been introduced in the signal. However, the slope o f the 6.9 GHz
121
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
vs. soil moisture function is still greater than the 10.7 GHz channel. This increase in
sensor noise supports findings in previous literature o f RFI effects in the AMSR-E 6.9
GHz channels (Li et al. 2004).
Both the ECMWF and AMSR-E soil moisture estimates are shown to have some
correlation with the Walnut Gulch precipitation events. However, the AMSR-E predicted
moisture are not within the expected range, nor do they represent the actual state well.
The small dynamic range o f AMSR-E values can be attributed to a systematic modeling
or instrument error, which caused a positive bias. As shown in Figure 4.5, the observed
brightness temperatures have a reasonable change (~20 K for 10 % moisture increase at
10.7 GHz) with observed moisture increase, so the moisture inaccuracy is likely due to
modeling error. We would expect some differences between the results o f the three data
sets, from a difference in sensing depth and modeling techniques. However, the expected
result o f the shallow AMSR-E sensing depth is an increase in dynamic range, from the
rapid wetting and drying o f the top soil layer.
The accuracy (correlation with high and low moisture patterns) o f the AMSR-E
predicted moisture is further illustrated in Figure 4.6. This plot compares (ASMR-E soil
moisture - 0.09 cm3 cm'3) and in situ soil moisture for the 1:30 AM time period during
the study. The AM AMSR-E and in situ values exhibited the highest correlation. A shift
o f the systematic bias (i.e., high AMSR-E values) results in a more reasonable correlation
(R2 = 0.54). The better correlation may be due to AMSR-E sensing drier (shallower) soils
than the in situ moisture probes, however it seems that the predicted values never
approach those expected for wet conditions.
122
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
4.4. Discussion
The AMSR-E soil moisture retrieval algorithm was shown to perform poorly in
the low vegetated and low topographic Arizona region. Patterns o f wetting and drying
can be observed in the AMSR-E, ECMWF and in situ soil moisture data in the Walnut
Gulch, AZ study area. Comparisons o f each dataset with daily precipitation values in the
region validate their inter-comparison. The AMSR-E estimates were shown to have a
positive bias o f roughly 0.1 cm3 cm'3 but demonstrated the potential for observing
moisture changes. Even though the small dynamic range of the AMSR-E values seems
misleading, a better agreement is apparent (R2 = 0.54) when the uniform overestimates
3
3
are removed. The range of estimated values was greatest for the ECMWF (0.23 cm c m ' )
compared to the in situ (0.15 cm3 cm'3) and AMSR-E (0.048 cm3 cm '3) estimated values.
Preliminary analysis has shown that AMSR-E retrievals o f soil moisture have
little temporal variability in the Walnut Gulch watershed, AZ. This could be attributed to
various caused such as deficiency in the retrieval process due to problems with
topography and/or inaccurate description of the soil surface and vegetation over the
region. The poor agreement o f the AMSR-E moisture product with observed values
suggests that the model does not adequately represent the surface emission over the
region. However, the observed brightness temperatures are consistent in dynamical range
with those observed in similar frequencies from other instmment and modeled emission
(L aym onetal. 1999).
Retrievals o f soil moisture using AMSR-E brightness temperatures and modeled
microwave emission have been shown to be reasonable in areas having similar vegetation
water content and slightly less topographic relief (Njoku et al. 2003, Bindlish et al. 2005,
123
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
McCabe et al. 2005). Even though the current AMSR-E soil moisture retrieval algorithm
is deemed inefficient, these studies demonstrate that more accurate soil moisture
estimates are possible. This work emphasizes the need for large-scale in situ observations
in various landscapes, especially where the model calibration is in error.
124
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
4.5. References
Albertson JD, Montaldo N (2003) Temporal dynamics o f soil moisture variability: 1.
Theoretical basis. Water Resources Research 39 (10)
Bindlish R, Jackson TJ, Gasiewski AJ, Klein M, Njoku EG (2005) Soil moisture mapping
and AMSR-E validation using the PSR in SMEX02. Remote Sensing of
Environment
Charpentier MA, Groffman PM (1992) Soil-Moisture Variability within Remote-Sensing
Pixels. Journal o f Geophysical Research-Atmospheres 97:18987-18995
Chen JM, Chen X, Ju W, Geng X (2005) Distributed hydrological model for mapping
evapotranspiration using remote sensing inputs. Journal o f Hydrology 305:15-39
Delworth T, Manabe S (1993) Climate Variability and Land-Surface Processes.
Advances in Water Resources 16:3-20
Douville H, Viterbo P, M ahfouf JF, Beljaars ACM (2000) Evaluation o f the Optimum
Interpolation and Nudging Techniques for soil moisture analysis using FIFE data.
Monthly Weather Review 128:1733-1756
Eltahir EAB (1989) A feedback mechanism in annual rainfall, Central Sudan. Journal of
Hydrology 110:323-334
Entin JK, Robock A, Vinnikov KY, Hollinger SE, Liu S, Namkhai A (2000) Temporal
and spatial scales o f observed soil misture varaitions in the extratropics. Journal
o f Geophysical Research 105:11865-11877
Gochis DJ, Brito-Castillo L, Shuttleworth WJ (In Press) Hydroclimatology o f the North
American Monsoon region in northwest Mexico. Journal o f Hydrology In Press,
Corrected Proof
Ivanov VY, Vivoni ER, Bras RL, Entekhabi D (2004) Preserving high-resolution surface
and rainfall data in operational-scale basin hydrology: a fully-distributed
physically-based approach. Journal o f Hydrology 298:80-111
125
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Jackson TJ, Le Vine DM, Griffis AJ, Goodrich DC, Schmugge TJ, Swift CT, Oneill P
(1993) Soil moisture and rainfall estimation over a semiarid environment with the
ESTAR microwave radiometer. IEEE Transactions on Geoscience and Remote
Sensing 31:836-846
Jacobs JM, Mohanty BP, Hsu E-C, Miller D (2004) SMEX02: Field scale variability,
time stability and similarity o f soil moisture. Remote Sensing o f Environment
92:436-446
Kawanishi T, Sezai T, Ito Y, Imaoka K, Takeshima T, Ishido Y, Shibata A, Miura M,
Inahata H, Spencer RW (2003) The Advanced Microwave Scanning Radiometer
for the Earth Observing System (AMSR-E), NASDA's contribution to the EOS
for global energy and water cycle studies. IEEE Transactions on Geoscience and
Remote Sensing 41:184-194
Kim G, Barros AP (2002) Space-time characterization of soil moisture from passive
microwave remotely sensed imagery and ancillary data. Remote Sensing of
Environment 81:393-403
Kustas WP, Goodrich DC, Moran MS, Amer SA, Bach LB, Blanford JH, Chehbouni A,
Claassen H, Clements WE, Doraiswamy PC, Dubois P, Clarke TR, Daughtry
CST, Gellman DI, Grant TA, Hipps LE, Huete AR, Humes KS, Jackson TJ,
Keefer TO, Nichols WD, Parry R, Perry EM, Pinker RT, Pinter PJ, Qi J, Riggs
AC, Schmugge TJ, Shutko AM, Stannard DI, Swiatek E, Vanleeuwen JD, Vanzyl
J, Vidal A, Washbume J, Weltz MA (1991) An Interdisciplinary Field-Study of
the Energy and Water Fluxes in the Atmosphere-Biosphere System over Semiarid
Rangelands - Description and Some Preliminary-Results. Bulletin o f the
American Meteorological Society 72:1683-1705
Laymon CA, Crosson WL, Soman VV, Belisle WR, Coleman TL, Fahsi A, Manu A,
Senwo ZN, Tsegaye TD, O'Neill PE, Jackson TJ (1999) Huntsville'96: an
experiment in ground-based microwave remote sensing o f soil moisture.
International Journal o f Remote Sensing 20:823-828
Li L, Njoku EG, Im E, Chang PS, Germain KS (2004) A preliminary survey o f radio­
frequency interference over the US in Aqua AMSR-E data. IEEE Transactions on
Geoscience and Remote Sensing 42:380-390
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Matzler C, Standley A (2000) Relief effects for passive microwave remote sensing.
International Journal o f Remote Sensing 21:2403-2412
McCabe MF, Wood EF, Gao BH (2005) Initial soil moisture retrievals from AMSR-E:
Multiscal comparison using in situ data and rainfall patterns over Iowa.
Geophysical Research Letters 32
Mohanty BP, Skaggs TH (2001) Spatio-temporal evolution and time-stable
characteristics o f soil moisture within remote sensing footprints with varying soil,
slope, and vegetation. Advances in Water Resources 24:1051-1067
Njoku E, Chan T, Crosson WL, Limaye A (2004) Evaluation o f the AMSRE-E data
calibration over land. Italian Journal o f Remote Sensing 30/31:19-38
Njoku EG, Jackson TJ, Lakshmi V, Chan TK, Nghiem SV (2003) Soil moisture retrieval
from AMSR-E. IEEE Transactions on Geoscience and Remote Sensing 41:215229
Robock A, Schlosser CA, Vinnikov KY, Speranskaya NA, Entin JK, Qiu S (1998)
Evaluation o f the AMIP soil moisture simulations. Global and Planetary Change
19:181-208
van den Hurk B, Viterbo P, Belajaars A, Betts A (2000) Offline validation for the ERA40
surface scheme. ECMWF Tech Memo 295:42
Weiss JL, Gutzler DS, Allred Coonrod JE, Dahm CN (2004) Seasonal and inter-annual
relationships between vegetation and climate in central New Mexico, USA.
Journal o f Arid Environments 57:507-534
Wigneron JP, Calvet JC, Pellarin T, Van de Griend AA, Berger M, Ferrazzoli P (2003)
Retrieving near-surface soil moisture from microwave radiometric observations:
current status and future plans. Remote Sensing o f Environment 85:489-506
127
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
1ir30'0"W
110°30'0"W
109°30'0"W
33°0'0"N
-33°0'0"N
32°30'0"N
32°30'0"N
32°0'0"N
32°0'0"N
¥
3r30'0"N
-31°30'0"N
31°0'0"N
5j-3r0'0"N
30°30'0"N
30°30'0"N
30°0'0"N
•30°0'0"N
29° 30' 0"N
111°30'0"W
110°30'0"W
109°30'0"W
E levation
WAHigh: 3000
Low: 500
Figure 4.1. Digital Elevation Map o f SMEX04 study areas.
128
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
I
30
« ry |
-114
-112
-110
-108
-100
Longitude
Figure 4.2. Mapped AMSR-E estimated volumetric soil moisture for August 5th, 2004
over NAME TIER 1 region.
129
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
ECMWF VSM
16
AMSR-E VSM
In situ VSM
0 .3
Precipitation
precipitation (mm)
12
0.2
o
o
n>
E
D
co
o
>
0.1
6 /1 2 /0 4
7 /7 /0 4
8 /1 /0 4
8 /2 6 /0 4
9 /2 0 /0 4
Figure 4.3. AMSR-E 1:30 AM estimates o f soil moisture for Walnut Gulch Watershed,
AZ plotted with mean in-situ precipitation and soil moisture, and ECMWF estimated
volumetric soil moisture.
130
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
16
ECMWF VSM
AMSR-E VSM
In situ VSM
Precipitation
precipitation (mm)
12
«
o
0
E
D
o
>
0
6 /1 2 /0 4
7 /7 /0 4
8 /1 /0 4
8 /2 6 /0 4
9 /2 0 /0 4
Figure 4.4. AMSR-E 1:30 PM estimates o f soil moisture for Walnut Gulch Watershed,
AZ plotted with mean in-situ precipitation and soil moisture, and ECMWF estimated
volumetric soil moisture.
131
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
300 — |
•
•
• A M SR -E 1 0 .7 G H z (H-pol)
Y = -9 3 .0 2 * X + 2 6 4 .3 5
•
•
• A SM R -E 6 .9 G H z (H-pol)
Y = -1 4 2 .1 7 * X + 2 7 8 .3 8
280
£
* •
•
••• • •
'• *
*
•
•
•
• •
•
•
w
£
•
•
w
••
•
260
•
• *. *
• ••. . ..
•• •
240
0 .0 4
0 .0 8
In Situ VSM (cm 3/c m 3)
0.12
0 .1 6
Figure 4.5. AMSR-E 6.9 and 10.7 GHz vs. in situ soil moisture for 1:30 AM overpass for
Walnut Gulch Watershed, AZ.
132
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
0.16
Y = 0.31 *X + 0.011
R2 = 0.54
_
0.12
E
o
E
0)
3
to
O
-
0.08
A
A
>
HI
OL
OT
2
.A
A
0.04
AA
<
0
0
0.04
0.08
In Situ Volum etric
0.12
0.16
Soil M oisture (cm 3/ cm 3)
Figure 4.6. 1:30 AM overpass AMSR-E soil moisture (- 0.09 cm3/cm3) vs. in situ soil
moisture for the Walnut Gulch Watershed from Junel to Sept 30th, 2004.
133
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
CHAPTER 5
SUMMARY
The lack o f a reliable, continuous global soil moisture data has spurred research in
the field of microwave remote sensing. Such desired soil moisture values would provide
quantitative near-surface moisture estimates at the temporal and spatial scales needed for
biophysically based soil-vegetation-atmosphere transfer (SVAT) models (Wigneron et al.
1999, Boegh et al. 2004, Moran et al. 2004), and other hydrological applications
(Choudhury 1991, Jackson et al. 1996, Njoku & Entekhabi 1996, Schmugge et al. 2002).
Currently, the desired frequency and spatial scales are unattainable on space-borne
platforms. However, previous research has shown that large-scale validation experiments
involving aircraft and satellite-based instruments are essential for accurate soil moisture
retrievals and improving instrument design. O f particular interest is the influence of
microwave frequency and spatial resolution on microwave remote sensing o f soil
moisture.
In summary, this dissertation addresses the following research questions:
(1) Does lower frequency outweigh higher spatial resolution in applications o f
p a s s iv e m icro w a ve radiom etry f o r so il m oistu re sen sitivity?
(2) Do measurements o f microwave brightness using L-, C-, and X-band
provide reliable estimates o f soil moisture under varying land use/land cover,
climatic and topographic regions?
135
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
The discussed work draws much of its substantive methodology from previous
validation experiments and algorithm calibration techniques, to better understand the
behavior o f microwave brightness temperatures at L-, C- and X-band over many different
land conditions. Co-existing satellite, aircraft, and ground measurements from three
large-scale field experiments were used to investigate these questions in a unique fashion.
The synergistic application o f radiative transfer theory, modeled soil moisture, in situ
field measurements, and remote sensing observations presented unique opportunities for
multi-frequency microwave remote sensing analysis.
In Chapter 2, L- and C-band remote sensing observations obtained during SGP99
were compared. This study was one of the first to compare large-scale observations at
these frequencies and spatial resolutions. The large contrast in spatial resolution o f the
two instruments allowed a comprehensive investigation of sub-footprint heterogeneity
over the field area. Changes in soil moisture, vegetation, and temperature were observed
at all scales and frequencies used in the study. The horizontally-polarized L-band (1.4
GHz) data were best correlated with soil moisture (R2=0.83). However, significant
decreases in brightness temperatures were observed from both instruments, i.e., 85 K and
43 K were observed for the PALS 1.4 GHz and PSR/C 6.9 GHz horizontally-polarized
channels, respectively. These results conclude that the data are compatible, and
demonstrate soil moisture sensitivity on a range o f spatial scales and frequencies.
When applying the observed field data to a radiative transfer model, the average
absolute difference between estimated and measured brightness temperature was less than
5 K for most cases. The fact that the algorithm performed well for the SGP99 region
indicated that the surface roughness and vegetation effects were correctly represented in
136
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
the retrieval, which validated the model and its application to the passive microwave data
used in this study.
Resampling o f the PALS horizontally-polarized L-band Tb to the PSR/C-footprint
resolution resulted in slightly lower soil moisture sensitivity, i.e., to -2.58 from -3.53 at
field-scale. However, the lower spatial resolution L-band data remained more sensitive
to the observed moisture than the same resolution C-band data. These results demonstrate
that it is in fact frequency, rather than spatial resolution, which has the greatest influence
on soil moisture sensitivity.
Chapter 3 applied similar techniques discussed in Chapter 2 to both satellite
(AMSR-E) and airborne (PSR/C) observations from SMEX02. It was necessary to
aggregate the in situ field data and PSR/C data for comparison with the larger satellite
observations. A mean decrease in AMSR-E brightness temperature o f 18 K and 26 K was
seen for the observed and modeled 10.65 GHz horizontally-polarized channels,
respectively. These observations demonstrated the ability o f observed and modeled
satellite-scale C-band brightness temperatures to accurately reflect the spatial and
temporal soil moisture variability during the SMEX02 campaign. The correlation
between soil moisture and predicted values was best for the higher resolution PSR/C data
and for wetter (> 0.1 cm3/cm3) conditions. Estimated moisture values for 0-6 cm soil
moisture was 0.04 m3/m3 and 0.07 m3/m3 for the PSR/C and AMSR-E, respectively.
These findings indicate that we are able to use mean values o f soil moisture,
vegetation, and other surface variables to calculate expected brightness temperatures at
C- and X-band frequencies and satellite-scales. Another important conclusion from this
work is that both the modeled and observed brightness temperatures are relatively scale-
137
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
invariant. This supports previous literature on aggregation effects in microwave
radiometry (Bindlish & Barros 2002) and also provides further motivation for future
satellite missions involving L- and C-band technologies such as the planned Hydros
(Entekhabi et al. 2004) and SMOS (Kerr et al. 2001) missions. More importantly, we
estimated near-surface soil moisture within an acceptable accuracy (R2=0.87 for wetter
regions) using the AMSR-E instrument. However, as shown by the poor soil moisture
retrievals in dry areas during SMEX04, further validation over areas with different
vegetation types, soils and more extreme topography are required to improve retrieval
confidence.
Quantitative estimates o f soil moisture using the AMSR-E provided data have
required routine radiometric data calibration and validation using comparisons o f satellite
observations, extended targets and field campaigns such as SMEX02 and SMEX04. Yet
there remain many questions regarding the performance o f the AMSR-E standard product
soil moisture algorithm (Njoku et al. 2003). This algorithm is currently being updated and
calibrated by multiple principle investigators (Shibata et al. 2003, Njoku et al. 2004). The
currently applied NASA EOS Aqua ASMR-E soil moisture algorithm is based on a
change detection approach using polarization ratios (PR) o f the calibrated AMSR-E
channel brightness temperatures. To date, the accuracy o f the soil moisture algorithm has
been investigated on short time scales during field campaigns such as the 2002 Soil
Moisture Experiments (SMEX02) (Bindlish et al. 2005, McCabe et al. 2005). Results
have indicated self-consistency and calibration stability o f the observed brightness
temperatures; however the performance of the moisture retrieval algorithm has been less
than satisfactory. The primary objective o f Chapter 4 was to evaluate the current version
138
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
o f the AMSR-E soil moisture product during the SMEX04 period over the Walnut Gulch
Experimental W atershed (150 km2) near Tombstone, AZ. This preliminary study assessed
soil moisture algorithm performance by comparing temporal variations o f moisture
estimates with mean values of in situ watershed soil moisture and precipitation values.
Further comparisons were made with a standard soil dataset from the European Centre
for Medium-Range Weather Forecasts (ECMWF).
The AMSR-E derived soil moisture were shown to have a positive bias of roughly
0.1 cm3 cm'3 but demonstrated the potential for observing moisture changes in the lightlyvegetated Arizona region. Results indicated a much smaller dynamical range in the
AMSR-E data compared to the observed and ECMWF modeled soil moisture for the
watershed, 0.048 cm cm' ,0.15 cm c m ', and 0.23 cm cm' respectively. Some o f the
difference in values can be attributed to the different sensing depths o f the three datasets.
Also, incorrect portrayal o f soil type by the models, especially in the gravelly soils
located in Walnut Gulch, could have caused erroneous results. It is likely that the
overestimated AMSR-E values and small dynamical range is due to inadequate model
calibration, and/or perhaps from incorrect retrieval technique. Shortfalls may be found in
the change detection procedure used in the current algorithm. For example, it is possible
that the monthly time-step used as reference may be too long, i.e., monthly vegetation
growth and climatological changes could be causing error.
When considering the problems identified by the current investigation, many
possibilities for future work became clear. There is still a need for a definitive modeling
technique that can accurately incorporate satellite observations and ground observations,
either through assimilation o f network data or remotely-sensed variables, but most
139
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
importantly, fo r a variety o f surface and vegetation conditions. Thus, more observational
studies are needed, particularly in areas that are prone to estimate errors. An important
improvement to the existing field campaign strategy would be to increase their duration
so that more vegetation and climate conditions can be observed. However, a more
realistic approach would be to include automated networks such as the Global Soil
Moisture Data Bank (Robock et al. 2000) in future validation studies. This is in fact
becoming more common, as these data are becoming more readily available and have
been shown useful in data assimilation studies. These network data are cost efficient,
available in long time-scales, and are not as prone to human error as in situ field
measurements. Still, validation experiments are needed in areas where these networks do
not exist and more intense coverage is required.
140
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
5.1. References
Bindlish R, Barros AP (2002) Subpixel variability of remotely sensed soil moisture: An
inter-comparison study o f SAR and ESTAR. IEEE Transactions on Geoscience
and Remote Sensing 40:326-337
Bindlish R, Jackson TJ, Gasiewski AJ, Klein M, Njoku EG (2005) Soil moisture mapping
and AMSR-E validation using the PSR in SMEX02. Remote Sensing of
Environment (In Press)
Boegh E, Thorsen M, Butts MB, Hansen S, Christiansen JS, Abrahamsen P, Hasager CB,
Jensen NO, van der Keur P, Refsgaard JC, Schelde K, Soegaard H, Thomsen A
(2004) Incorporating remote sensing data in physically based distributed agrohydrological modelling. Journal o f Hydrology 287:279-299
Choudhury BJ (1991) Passive Microwave Remote-Sensing Contribution to Hydrological
Variables. Surveys in Geophysics 12:63-84
Entekhabi D, Njoku EG, Houser P, Spencer M, Doiron T, Kim YJ, Smith J, Girard R,
Belair S, Crow W, Jackson TJ, Kerr YH, Kimball JS, Koster R, McDonald KC,
O'Neill PE, Pultz T, Running SW, Shi JC, Wood E, van Zyl J (2004) The
hydrosphere state (Hydros) satellite mission: An earth system pathfinder for
global mapping o f soil moisture and land freeze/thaw. IEEE Transactions on
Geoscience and Remote Sensing 42:2184-2195
Jackson TJ, Schmugge J, Engman ET (1996) Remote sensing applications to hydrology:
Soil moisture.
Hydrological Sciences Joumal-Joumal Des Sciences
Hydrologiques 41:517-530
Kerr YH, Waldteufel P, Wigneron JP, Martinuzzi JM, Font J, Berger M (2001) Soil
moisture retrieval from space: The Soil Moisture and Ocean Salinity (SMOS)
mission. IEEE Transactions on Geoscience and Remote Sensing 39:1729-1735
McCabe MF, Wood EF, Gao BH (2005) Initial soil moisture retrievals from AMSR-E:
Multiscal comparison using in situ data and rainfall patterns over Iowa.
Geophysical Research Letters 32
Moran MS, Peters-Lidard CD, Watts JM, McElroy S (2004) Estimating soil moisture at
the watershed scale with satellite-based radar and land surface models. Canadian
Journal o f Remote Sensing 30:805-826
141
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Njoku E, Chan T, Crosson WL, Limaye A (2004) Evaluation o f the AMSRE-E data
calibration over land. Italian Journal o f Remote Sensing 30/31:19-38
Njoku EG, Entekhabi D (1996) Passive microwave remote sensing o f soil moisture.
Journal o f Hydrology 184:101-129
Njoku EG, Jackson TJ, Lakshmi V, Chan TK, Nghiem SV (2003) Soil moisture retrieval
from AMSR-E. IEEE Transactions on Geoscience and Remote Sensing 41:215229
Robock A, Vinnikov KY, Srinivasan G, Entin JK, Hollinger SE, Speranskaya NA, Liu
SX, Namkhai A (2000) The Global Soil Moisture Data Bank. Bulletin of the
American Meteorological Society 81:1281-1299
Schmugge TJ, Kustas WP, Ritchie JC, Jackson TJ, Rango A (2002) Remote sensing in
hydrology. Advances in Water Resources 25:1367-1385
Shibata A, Imaoka K, Koike T (2003) AMSR/AMSR-E level 2 and 3 algorithm
developments and data validation plans of NASD A. IEEE Transactions on
Geoscience and Remote Sensing 41:195-203
Wigneron JP, Calvet JC, Olioso A, Chanzy A, Bertuzzi P (1999) Estimating the rootzone soil moisture from the combined use o f time series o f surface soil moisture
and SVAT modelling. Physics and Chemistry o f the Earth Part B-Hydrology
Oceans and Atmosphere 24:837-843
142
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
CM*
Albertson JD, Montaldo N (2003) Temporal dynamics of soil moisture variability: 1.
Theoretical basis. Water Resources Research 39: Art. No. 1275
Ashcroft P, Wentz F (2003) AMSR-E/Aqua L2A Global Swath Spatially-Resampled
Brightness Temperatures (Tb) V001, June to July 2002. Boulder, CO, USA:
National Snow and Ice Data Center. Digital media.
Basara JB, Crawford KC (2002) Linear relationships between root-zone soil moisture and
atmospheric processes in the planetary boundary layer. Journal o f Geophysical
Research-Atmospheres 107:1-18
Bindlish R, Jackson TJ, Gasiewski AJ, Klein M, Njoku EG (2005) Soil moisture mapping
and AMSR-E validation using the PSR in SMEX02. Remote Sensing of
Environment (In Press)
Bindlish R, Barros AP (2002) Subpixel variability o f remotely sensed soil moisture: An
inter-comparison study o f SAR and ESTAR. IEEE Transactions on Geoscience
and Remote Sensing 40:326-337
Bindlish R, Jackson TJ, Wood E, Gao E1L, Starks P, Bosch D, Lakshmi V (2003) Soil
moisture estimates from TRMM Microwave Imager observations over the
Southern United States. Remote Sensing o f Environment 85:507-515
Bindlish R, Kustas WP, French AN, Diak GR, Mecikalski JR (2001) Influence of near­
surface soil moisture on regional scale heat fluxes: Model results using
microwave remote sensing data from SGP97. IEEE Transactions on Geoscience
and Remote Sensing 39:1719-1728
Boegh E, Thorsen M, Butts MB, Hansen S, Christiansen JS, Abrahamsen P, Hasager CB,
Jensen NO, van der Keur P, Refsgaard JC, Schelde K, Soegaard H, Thomsen A
(2004) Incorporating remote sensing data in physically based distributed agrohydrological modelling. Journal o f Hydrology 287:279-299
143
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Bolten JD, Lakshmi V, Njoku EG (2003b) Soil moisture retrieval using the passive/active
L- and S-band radar/radiometer. IEEE Transactions on Geoscience and Remote
Sensing 41:2792-2801
Bolten JD, Narayan U, Guijarro L, Lakshmi V (2004) Passive-Active Microwave Remote
Sensing o f Soil Moisture at Both L- and C-band: A Comparison o f Two Field
Experiments. Italian Journal o f Remote Sensing 30/31:65-86
Brubaker KL, Entekhabi D (1996a) Analysis o f feedback mechanisms in landatmosphere interaction. Water Resources Research 32:1343-1357
Brubaker KL, Entekhabi D (1996b) Asymmetric recovery from wet versus dry soil
moisture anomalies. Journal of Applied Meteorology 35:94-109
Burke EJ, Bastidas LA, Shuttleworth WJ (2002) Exploring the potential for multipatch
soil-moisture retrievals using multiparameter optimization techniques. IEEE
Transactions on Geoscience and Remote Sensing 40:114-120
Chang JT, Wetzel PJ (1991) Effects of Spatial Variations o f Soil-Moisture and
Vegetation on the Evolution of a Prestorm Environment - a Numerical CaseStudy. Monthly Weather Review 119:1368-1390
Charpentier MA, Grofffnan PM (1992) Soil-Moisture Variability within Remote-Sensing
Pixels. Journal o f Geophysical Research-Atmospheres 97:18987-18995
Chen JM, Chen X, Ju W, Geng X (2005) Distributed hydrological model for mapping
evapotranspiration using remote sensing inputs. Journal o f Hydrology 305:15-39
Choudhury BJ (1991) Passive Microwave Remote-Sensing Contribution to Hydrological
Variables. Surveys in Geophysics 12:63-84
Choudhury BJS, T. J.; Newton, R.W.;Chang, A.T.C. (1979) Effect o f surface roughness
on microwave emission o f soils. Journal o f Geophysical Research 84:5699-5706
Cosh MH, Jackson TJ, Bindlish R, Prueger JH (2004) Watershed scale temporal and
spatial stability o f soil moisture and its role in validating satellite estimates.
Remote Sensing o f Environment 92:427-435
144
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Delworth T, Manabe S (1989) The Influence o f Soil Wetness on near-Surface
Atmospheric Variability. Journal o f Climate 2:1447-1462
Delworth T, Manabe S (1993) Climate Variability and Land-Surface Processes.
Advances in Water Resources 16:3-20
Dirmeyer PA, Zeng FJ, Duchame A, Morrill JC, Koster RD (2000) The sensitivity of
surface fluxes to soil water content in three land surface schemes. Journal of
Hydrometerology 1:121-134
Dobson MC, Ulaby FT (1986) Active Microwave Soil-Moisture Research. IEEE
Transactions on Geoscience and Remote Sensing 24:23-36
Dobson MC, Ulaby FT, Hallikainen MT, Elrayes MA (1985) Microwave Dielectric
Behavior of Wet Soil .2. Dielectric Mixing Models. IEEE Transactions on
Geoscience and Remote Sensing 23:35-46
Douville H, Viterbo P, M ahfouf JF, Beljaars ACM (2000) Evaluation o f the Optimum
Interpolation and Nudging Techniques for soil moisture analysis using FIFE data.
Monthly Weather Review 128:1733-1756
Drusch M, Wood EF, Gao H, Thiele A (2004) Soil moisture retrieval during the Southern
Great Plains Hydrology Experiment 1999: A comparison between experimental
remote sensing data and operational products. Water Resources Research 40:1-10
Drusch M, Wood EF, Simmer C (1999) Up-scaling effects in passive microwave remote
sensing: ESTAR 1.4 GHz measurements during SGP '97. Geophysical Research
Letters 26:879-882
Eltahir EAB (1989) A feedback mechanism in annual rainfall, Central Sudan. Journal of
Hydrology 110:323-334
Entekhabi D, Njoku EG, Houser P, Spencer M, Doiron T, Kim YJ, Smith J, Girard R,
Belair S, Crow W, Jackson TJ, Kerr YH, Kimball JS, Koster R, McDonald KC,
O'Neill PE, Pultz T, Running SW, Shi JC, Wood E, van Zyl J (2004) The
hydrosphere state (Hydros) satellite mission: An earth system pathfinder for
global mapping o f soil moisture and land freeze/thaw. IEEE Transactions on
Geoscience and Remote Sensing 42:2184-2195
145
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Entin JK, Robock A, Vinnikov KY, Hollinger SE, Liu S, Namkhai A (2000) Temporal
and spatial scales of observed soil misture varaitions in the extratropics. Journal
o f Geophysical Research 105:11865-11877
Eom HJ (1992) A Thermal Microwave Emission Model for Row-Structured Vegetation.
International Journal o f Remote Sensing 13:2975-2982
Famiglietti JS, Devereaux JA, Laymon CA, Tsegaye T, Houser PR, Jackson TJ, Graham
ST, Rodell M, van Oevelen PJ (1999) Ground-based investigation o f soil moisture
variability within remote sensing footprints during the Southern Great Plains 1997
(SGP97) Hydrology Experiment. Water Resources Research 35:1839-1851
Fennessy MJ, Shukla J (1999) Impact of initial soil wetness on seasonal atmospheric
prediction. Journal o f Climate 12:3167-3180
Fetter CW (1994) Applied Hydrologeology, Vol. Macmillan College Publishing
Company, New York
Galantowicz JF, Entekhabi D, Njoku EG (2000) Estimation o f soil-type heterogeneity
effects in the retrieval o f soil moisture from radiobrightness. IEEE Transactions
on Geoscience and Remote Sensing 38:312-316
Gao BC (1996) NDWI - A normalized difference water index for remote sensing of
vegetation liquid water from space. Remote Sensing o f the Environment 58:257 266
Gao HL, Wood EF, Drusch M, Crow W, Jackson TJ (2004) Using a microwave emission
model to estimate soil moisture from ESTAR observations during SGP99. Journal
o f Hydrometeorology 5:49-63
Gasiewski AJ, Klein M, Yevgrafov A, Leuskiy V (2002) Interference mitigation in
passive microwave radiometry International Geoscience and Remote Sensing
Symposium, Toronto, Canada
Gochis DJ, Brito-Castillo L, Shuttleworth WJ (In Press) Hydroclimatology o f the North
American Monsoon region in northwest Mexico. Journal o f Hydrology In Press,
Corrected Proof
146
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Guha A, Lakshmi V (2002) Sensitivity, spatial heterogeneity, and scaling o f C-band
microwave brightness temperatures for land hydrology studies. IEEE
Transactions on Geoscience and Remote Sensing 40:2626-2635
Gulledge J, Schimel JP (1998) Moisture control over atmospheric CH4 consumption and
C 02 production in diverse Alaskan soils. Soil Biology & Biochemistry 30:11271132
Hallikainen MT, Ulaby FT, Dobson MC, Elrayes MA, W u LK (1985) Microwave
Dielectric Behavior of Wet Soil .1. Empirical-Models and ExperimentalObservations. IEEE Transactions on Geoscience and Remote Sensing 23:25-34
Hollenbeck KJ, Schmugge TJ, Homberger GM, Wang JR (1996) Identifying soil
hydraulic heterogeneity by detection of relative change in passive microwave
remote sensing observations. Water Resources Research 32:139-148
Hu ZL, Islam S (1997a) Effects o f spatial variability on the scaling o f land surface
parameterizations. Boundary-Layer Meteorology 83:441-461
Hu ZL, Islam S (1997b) Scale invariant remote sensing algorithms. IEEE Transactions on
Geoscience and Remote Sensing 35:747-755
Ivanov VY, Vivoni ER, Bras RL, Entekhabi D (2004) Preserving high-resolution surface
and rainfall data in operational-scale basin hydrology: a fully-distributed
physically-based approach. Journal o f Hydrology 298:80-111
Jackson T, Cosh M ( 2003) SMEX02 Vegetation Water Content, Iowa Regional and
Walnut Creek Watershed. Boulder, CO: National Snow and Ice Data Center.
Digital media.
Jackson T, Schmugge J, Engman E (1996a) Remote sensing applications to hydrology:
soil moisture. Hydrological Sciences 41:3167-3180
Jackson T, Williams S (2004) SMEX02 CODIAC Soil Climate Analysis Network
(SCAN) Data, Iowa. Boulder, CO. University Corporation for Atmospheric
Research, Joint Office for Science Support (UCAR/JOSS). Digital Media.
147
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Jackson TJ, Chen DY, Cosh M, Li FQ, Anderson M, Walthall C, Doriaswamy P, Hunt
ER (2004) Vegetation water content mapping using Landsat data derived
normalized difference water index for com and soybeans. Remote Sensing of
Environment 92:475-482
Jackson TJ, Gasiewski AJ, Oldak A, Klein M, Njoku EG, Yevgrafov A, Christiani S,
Bindlish R (2002) Soil moisture retrieval using the C-band polarimetric scanning
radiometer during the Southern Great Plains 1999 Experiment. IEEE Transactions
on Geoscience and Remote Sensing 40:2151-2161
Jackson TJ, Hsu AY (2001) Soil moisture and TRMM microwave imager relationships in
the Southern Great Plains 1999 (SGP99) Experiment. IEEE Transactions on
Geoscience and Remote Sensing 39:1632-1642
Jackson TJ, Le Vine DM, Griffis AJ, Goodrich DC, Schmugge TJ, Swift CT, Oneill P
(1993) Soil moisture and rainfall estimation over a semiarid environment with the
ESTAR microwave radiometer. IEEE Transactions on Geoscience and Remote
Sensing 31:836-846
Jackson TJ, Le Vine DM, Hsu AY, Oldak A, Starks PJ, Swift CT, Isham JD, Haken M
(1999) Soil moisture mapping at regional scales using microwave radiometry: The
Southern Great Plains Hydrology Experiment. IEEE Transactions on Geoscience
and Remote Sensing 37:2136-2151
Jackson TJ, Schmugge J, Engman ET (1996b) Remote sensing applications to hydrology:
Soil moisture. Hydrological Sciences Joumal-Joumal Des Sciences
Hydrologiques 41:517-530
Jackson TJ, Schmugge TJ (1989) Passive Microwave Remote-Sensing System for SoilMoisture - Some Supporting Research. IEEE Transactions on Geoscience and
Remote Sensing 27:225-235
Jackson TJ, Schmugge TJ (1991) Vegetation Effects on the Microwave Emission of
Soils. Remote Sensing o f Environment 36:203-212
Jacobs JM, Mohanty BP, Hsu E-C, Miller D (2004) SMEX02: Field scale variability,
time stability and similarity of soil moisture. Remote Sensing o f Environment
92:436-446
148
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Kawanishi T, Sezai T, Ito Y, Imaoka K, Takeshima T, Ishido Y, Shibata A, Miura M,
Inahata H, Spencer RW (2003) The Advanced Microwave Scanning Radiometer
for the Earth Observing System (AMSR-E), NASDA's contribution to the EOS
for global energy and water cycle studies. IEEE Transactions on Geoscience and
Remote Sensing 41:184-194
Kerr YH, Njoku EG (1990) A Semiempirical Model for Interpreting Microwave
Emission from Semiarid Land Surfaces as Seen from Space. IEEE Transactions
on Geoscience and Remote Sensing 28:384-393
Kerr YH, Waldteufel P, Wigneron JP, Martinuzzi JM, Font J, Berger M (2001) Soil
moisture retrieval from space: The Soil Moisture and Ocean Salinity (SMOS)
mission. IEEE Transactions on Geoscience and Remote Sensing 39:1729-1735
Kim G, Barros AP (2002) Space-time characterization o f soil moisture from passive
microwave remotely sensed imagery and ancillary data. Remote Sensing of
Environment 81:393-403
Koster R, Milly PCD (1997) The interplay between transpiration and runoff formulations
in land surface schemes used with atmospheric models. Journal o f Climate
10:1578-1591
Kustas WP, French AN, Hatfield JL, Jackson TJ, Moran MS, Rango A, Ritchie JC,
Schmugge TJ (2003) Remote sensing research in hydrometeorology.
Photogrammetric Engineering and Remote Sensing 69:631-646
Kustas WP, Goodrich DC, Moran MS, Amer SA, Bach LB, Blanford JH, Chehbouni A,
Claassen H, Clements WE, Doraiswamy PC, Dubois P, Clarke TR, Daughtry
CST, Gellman DI, Grant TA, Hipps LE, Huete AR, Humes KS, Jackson TJ,
Keefer TO, Nichols WD, Parry R, Perry EM, Pinker RT, Pinter PJ, Qi J, Riggs
AC, Schmugge TJ, Shutko AM, Stannard DI, Swiatek E, Vanleeuwen JD, Vanzyl
J, Vidal A, Washbume J, Weltz MA (1991) An Interdisciplinary Field-Study of
the Energy and Water Fluxes in the Atmosphere-Biosphere System over Semiarid
Rangelands - Description and Some Preliminary-Results. Bulletin of the
American Meteorological Society 72:1683-1705
Laymon CA, Crosson WL, Jackson TJ, Manu A, Tsegaye TD (2001) Ground-based
passive microwave remote sensing observations o f soil moisture at S-band and Lband with insight into measurement accuracy. IEEE Transactions on Geoscience
and Remote Sensing 39:1844-1858
149
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Laymon CA, Crosson WL, Soman VV, Belisle WR, Coleman TL, Fahsi A, Manu A,
Senwo ZN, Tsegaye TD, O'Neill PE, Jackson TJ (1999) Huntsville'96: an
experiment in ground-based microwave remote sensing o f soil moisture.
International Journal of Remote Sensing 20:823-828
Lee KH, Anagnostou EN (2004) A combined passive/active microwave remote sensing
approach for surface variable retrieval using Tropical Rainfall Measuring Mission
observations. Remote Sensing of Environment 92:112-125
LeVine DM, Karam MA (1996) Dependence o f attenuation in a vegetation canopy on
frequency and plant water content. IEEE Transactions on Geoscience and Remote
Sensing 34:1090-1096
Li L, Njoku EG, Im E, Chang PS, Germain KS (2004) A preliminary survey o f radio­
frequency interference over the US in Aqua AMSR-E data. IEEE Transactions on
Geoscience and Remote Sensing 42:380-390
Matzler C, Standley A (2000) Relief effects for passive microwave remote sensing.
International Journal o f Remote Sensing 21:2403-2412
McCabe MF, Gao BH, Wood EF (2004) An evaluation o f AMSR-E derived soil moisture
retrievals using ground based, airborne and ancillary data during SMEX02.
Journal o f Hydrometeorology
McCabe MF, Wood EF, Gao BH (2005) Initial soil moisture retrievals from AMSR-E:
Multiscal comparison using in situ data and rainfall patterns over Iowa.
Geophysical Research Letters 32
Mettemicht GI, Zinck JA (2003) Remote sensing o f soil salinity: potentials and
constraints. Remote Sensing o f Environment 85:1-20
Michot D, Benderitter Y, Dorigny A, Nicoullaud B, King D, Tabbagh A (2003) Spatial
and temporal monitoring of soil water content with an irrigated com crop cover
using surface electrical resistivity tomography. Water Resources Research 39
Miller DA, White RA (1998) A Conteminous United States Multi-Layer Soil
Characteristics Data Set for Regional Climate and Hydrology Modeling. Earth
Interactions 2
150
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Mo T, Choudhury BJ, Schmugge TJ, Wang JR, Jackson TJ (1982) A Model for
Microwave Emission from Vegetation-Covered Fields. Journal o f Geophysical
Research-Oceans and Atmospheres 87:1229-1237
Mohanty BP, Skaggs TH (2001) Spatio-temporal evolution and time-stable
characteristics of soil moisture within remote sensing footprints with varying soil,
slope, and vegetation. Advances in Water Resources 24:1051-1067
Montaldo N, Albertson JD (2003) Temporal dynamics o f soil moisture variability: 2.
Implications for land surface models. Water Resources Research 39:-
Moran MS, Peters-Lidard CD, Watts JM, McElroy S (2004) Estimating soil moisture at
the watershed scale with satellite-based radar and land surface models. Canadian
Journal o f Remote Sensing 30:805-826
Narayan U, Lakshmi V, Njoku EG (2004) Retrieval o f soil moisture from passive and
active L/S band sensor (PALS) observations during the Soil Moisture Experiment
in 2002 (SMEX02). Remote Sensing o f Environment 92:483-496
Njoku E (2004) Soil moisture field experiment special issue. Remote Sensing of
Environment 92:425-426
Njoku E, Chan T, Crosson WL, Limaye A (2004) Evaluation o f the AMSRE-E data
calibration over land. Italian Journal o f Remote Sensing 30/31:19-38
Njoku EG, Entekhabi D (1996) Passive microwave remote sensing o f soil moisture.
Journal o f Hydrology 184:101-129
Njoku EG, Jackson TJ, Lakshmi V, Chan TK, Nghiem SV (2003) Soil moisture retrieval
from AMSR-E. IEEE Transactions on Geoscience and Remote Sensing 41:215229
Njoku EG, Li L (1999) Retrieval o f land surface parameters using passive microwave
measurements at 6-18 GHz. IEEE Transactions on Geoscience and Remote
Sensing 37:79-93
151
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Njoku EG, Wilson WJ, Yueh SH, Dinardo SJ, Li FK, Jackson TJ, Lakshmi V, Bolten J
(2002) Observations o f soil moisture using a passive and active low-frequency
microwave airborne sensor during SGP99. IEEE Transactions on Geoscience and
Remote Sensing 40:2659-2673
Njoku E (1977) Theory for passive microwave remote sensing o f near-surface soil
moisture. Journal of Geophysical Research 82: 3108-3118
Oldak A, Pachepsky Y, Jackson TJ, Rawls WJ (2002) Statistical properties of soil
moisture images revisited. Journal o f Hydrology 255:12-24
Owe M, Chang A, Golus RE (1988) Estimating Surface Soil-Moisture from Satellite
Microwave Measurements and a Satellite Derived Vegetation Index. Remote
Sensing o f Environment 24:331-345
Owe M, Van de Griend AA, de Jeu R, de Vries JJ, Seyhan E, Engman ET (1999)
Estimating soil moisture from satellite microwave observations: Past and ongoing
projects, and relevance to GCIP. Journal o f Geophysical Research-Atmospheres
104:19735-19742
Peplinski NR, Ulaby FT, Dobson MC (1995) Dielectric-Properties o f Soils in the 0.3-1.3Ghz Range. IEEE Transactions on Geoscience and Remote Sensing 33:803-807
Pielke RA (2001) Influence o f the spatial distribution of vegetation and soils on the
prediction o f cumulus convective rainfall. Reviews of Geophysics 39:151-177
Piepmeier JR, Gasiewski AJ (2001) High-resolution passive polarimetric microwave
mapping o f ocean surface wind vector fields. IEEE Transactions on Geoscience
and Remote Sensing 39:606-622
Robock A, Schlosser CA, Vinnikov KY, Speranskaya NA, Entin JK, Qiu S (1998)
Evaluation o f the AMIP soil moisture simulations. Global and Planetary Change
19:181-208
Robock A, Vinnikov KY, Srinivasan G, Entin JK, Hollinger SE, Speranskaya NA, Liu
SX, Namkhai A (2000) The Global Soil Moisture Data Bank. Bulletin of the
American Meteorological Society 81:1281-1299
152
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Schmugge T (1985) Remote sensing o f soil moisture. In: Burt MGAaTP (ed)
Hydrological Forcasting. John Wiley and Sons, New York, p 604
Schmugge T, Jackson TJ (1994) Mapping Surface Soil-Moisture with Microwave
Radiometers. Meteorology and Atmospheric Physics 54:213-223
Schmugge T, Jackson TJ, Kustas WP, Wang JR (1992) Passive Microwave RemoteSensing of Soil-Moisture - Results from Hapex, Fife and Monsoon-90. Isprs
Journal o f Photogrammetry and Remote Sensing 47:127-143
Schmugge TJ, Kustas WP, Ritchie JC, Jackson TJ, Rango A (2002) Remote sensing in
hydrology. Advances in Water Resources 25:1367-1385
Seuffert G, W ilker H, Viterbo P, Drusch M, M ahfouf JF (2004) The usage o f screen-level
parameters and microwave brightness temperature for soil moisture analysis.
Journal o f Hydrometeorology 5:516-531
Shibata A, Imaoka K, Koike T (2003) AMSR/AMSR-E level 2 and 3 algorithm
developments and data validation plans of NASD A. IEEE Transactions on
Geoscience and Remote Sensing 41:195-203
Shukla J, Mintz Y (1982) Influence o f Land-Surface Evapo-Transpiration on the Earths
Climate. Science 215:1498-1501
Ulaby FT, Razani M, Dobson MC (1983) Effects o f Vegetation Cover on the Microwave
Radiometric Sensitivity to Soil-Moisture. IEEE Transactions on Geoscience and
Remote Sensing 21:51-61
Ulaby FTM, R.K., Fung, A.K. (1986) Microwave remote sensing: Active and passivevoume scattering and emission theory, advanced systems and applications.
Dedham, MA: Artech House.
Van de Griend AA, Wigneron JP (2004) The b-factor as a function of frequency and
canopy type at h-polarization. IEEE Transactions on Geoscience and Remote
Sensing 42:786-794
van den Hurk B, Viterbo P, Belajaars A, Betts A (2000) Offline validation fo the ERA40
surface scheme. ECMWF Tech Memo 295:42
153
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Wang JR, Engman ET, Mo T, Schmugge TJ, Shiue JC (1987) The Effects o f SoilMoisture, Surface-Roughness, and Vegetation on L-Band Emission and
Backscatter. IEEE Transactions on Geoscience and Remote Sensing 25:825-833
Wang JR, Schmugge TJ (1980) An empirical model for the complex dielectric
permittivity o f soil as a function o f water content. IEEE Transactions on
Geoscience and Remote Sensing GE-23:35-46
Wang JR, Shiue JC, Schmugge TJ, Engman ET (1989) Mapping Surface Soil-Moisture
with L-Band Radiometric Measurements. Remote Sensing o f Environment
27:305-312
Weiss JL, Gutzler DS, Allred Coonrod JE, Dahm CN (2004) Seasonal and inter-annual
relationships between vegetation and climate in central New Mexico, USA.
Journal o f Arid Environments 57:507-534
Wigneron JP, Calvet JC, Olioso A, Chanzy A, Bertuzzi P (1999) Estimating the rootzone soil moisture from the combined use o f time series o f surface soil moisture
and SVAT modelling. Physics and Chemistry o f the Earth Part B-Hydrology
Oceans and Atmosphere 24:837-843
Wigneron JP, Calvet JC, Pellarin T, Van de Griend AA, Berger M, Ferrazzoli P (2003)
Retrieving near-surface soil moisture from microwave radiometric observations:
current status and future plans. Remote Sensing o f Environment 85:489-506
Wigneron JP, Parde M, Waldteufel P, Chanzy A, Kerr Y, Schmidl S, Skou N (2004)
Characterizing the dependence o f vegetation model parameters on crop structure,
incidence angle, and polarization at L-band. IEEE Transactions on Geoscience
and Remote Sensing 42:416-425
Wilson WJ, Yueh SH, Dinardo SJ, Chazanoff SL, Kitiyakara A, Li FK, Rahmat-Samii Y
(2001) Passive active L- and S-band (PALS) microwave sensor for ocean salinity
and soil moisture measurements. IEEE Transactions on Geoscience and Remote
Sensing 39:1039-1048
Wu JQ, Zhang RD, Yang JZ (1996) Analysis of rainfall-recharge relationships. Journal of
Hydrology 177:143-160
154
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Документ
Категория
Без категории
Просмотров
0
Размер файла
3 183 Кб
Теги
sdewsdweddes
1/--страниц
Пожаловаться на содержимое документа