INTERNATIONAL JOURNAL OF CLIMATOLOGY, VOL. 17, 55–66 (1997) REGIONAL INDICES OF CLIMATE VARIATION: TEMPERATURE AND RAINFALL IN QUEENSLAND, AUSTRALIA J. M. LOUGH Australian Institute of Marine Science, PMB 3, Townsville MC, Queensland 4810, Australia email: email@example.com Received 2 November 1995 Revised 27 March 1996 Accepted 20 May 1996 ABSTRACT Indices of temperature and rainfall are developed for Queensland, Australia as an aid to monitoring climate variation and change. The rainfall index is based on 17 stations and extends from 1890. The temperature index is based on six stations and extends from 1910. These series can be updated promptly and although based on a small number of stations appear to reflect the major climate variations in Queensland over the past century. There is no significant trend towards wetter or drier conditions in Queensland in either summer or winter. Although the State has recently suffered from an extended drought, this dry period does not appear to be particularly unusual in the 105–year record. In contrast, average and minimum temperatures and the daily temperature range show significant trends since 1910 in both summer and winter. Minimum temperatures have increased and the daily temperature range has decreased, especially in the early winter season, April–May. Int. J. Climatol., Vol. 17, 55–66 (1977) (No. of figures: 6 KEY WORDS: No. of tables: 2 No. of refs: 21) Queensland; Australia; trend; rainfall; temperature; regional indices; ENSO. INTRODUCTION There is growing awareness of the possible consequences of global climate change. Yet to many people the concern is how climate may vary or is varying at the regional level. The enhanced greenhouse scenario suggests that temperatures in Australia may rise by 1–2 C, summer rainfall may increase, and the frequency of highrainfall and flooding events may also increase (Whetton, 1993; Whetton et al., 1993). Such a scenario is difficult to grasp for those living in Queensland and parts of eastern Australia which have been subject to a ‘severe and persistent’ drought since 1991 (Bureau of Meteorology, 1995). This extended drought is largely a result of a prolonged El Niño–Southern Oscillation (ENSO) event. Observational studies, however, lend some support to these projected changes. Over the period 1910 to 1988, summer rainfall appears to have increased over much of eastern Australia (Nicholls and Lavery, 1992.) This increase occurred rather abruptly about 1950, confirming earlier findings (e.g. Pittock, 1975). There is also evidence that annual rainfall intensity and the frequency of heavy rainfall events have increased over tropical Australia over the period 1910 to 1989 (Suppiah and Hennessy, 1996). Observational studies also provide evidence of temperature increases over Australia (e.g. Jones et al., 1990; Plummer, 1991). These temperature changes have been greater for minimum than maximum temperatures with a consequent decline in the daily temperature range (DTR) in recent decades (Plummer et al., 1994), matching trends in DTR found in other parts of the world (Karl et al., 1993). Some of these trends in Australian climate over recent decades have also been identified in climate regions of the south-west Pacific (Salinger et al., 1995). The purpose of the present study is twofold: (i) development of indices of temperature and rainfall variations in Queensland (which encompasses 26 per cent of the land area of Australia) based on a small number of stations which can be readily up-dated and, therefore, monitored for change; and (ii) analysis of these Queensland climate indices for evidence of climate variation or change, particularly within the past 10 to 20 years. The spatial and temporal variability of Queensland rainfall (Lough, 1991, 1993) and temperatures (Lough, 1995) have been # CCC 0899-8418/97/010055-12 1997 by the Royal Meteorological Society 56 J. M. LOUGH examined previously using a relatively large number of stations. These analyses were based on data only through the mid 1980s. The present study uses data through the Southern Hemisphere summer of 1994–1995. We can, therefore, consider how unusual the recent drought in Queensland has been and whether the trends towards higher minimum temperatures and lower DTR (Lough, 1995) have continued through the early 1990s. DATA Rainfall Seventeen stations with rainfall records extending back to at least 1890 were obtained from the Australian Bureau of Meteorology (Figure 1). The choice of stations was based on quality of record, site location (to give as extensive coverage of Queensland as possible), and that the monthly station rainfall was reported regularly in the Monthly Weather Review, Queensland (published by the Australian Bureau of Meteorology). This allows the data series to be regularly and promptly updated. Monthly rainfall at each station was totalled for the following seasonal groupings: October–November (ON), December–January (DJ), February–March (FM), April–May (AM), June–July (JJ), August–September (AS), October–March (summer) and April–September (winter). The rainfall series are positively skewed so that average rainfall can be a poor measure of the most commonly occurring rainfall total. For example, annual (October–September) median rainfall at the 17 stations ranges from 275 to 1693 mm, which is, on average, 5 per cent less than average rainfall (6 per cent less for summer and 15 per cent less for winter). Interannual variability (the standard deviation expressed as a percentage of the mean) is also high, 37 per cent for the year, 42 per cent for summer and 68 per cent in winter. Most of the annual rainfall at the 17 stations occurs during the Southern Hemishphere summer monsoon, on average 78 per cent in summer and 22 per cent in winter. The skewness and interannual variability of rainfall totals become more marked when the Figure 1. Location of 17 rainfall stations (circles) and 6 temperature stations (triangles) used to develop regional indices. Georgetown is used for both rainfall and temperature. CLIMATE INDICES FOR QUEENSLAND 57 2-month seasonal totals are considered. To allow for this the individual station series were converted to percentage rank anomalies (see Ropelewski and Halpert, 1987). For each seasonal, station rainfall time series, the complete record was ranked from 1 to n, where n was the total number of years. The ranked series were normalized by the number of years and multiplied by 100. The series were then expressed as departures from the mean value of 50. This procedure allows comparison of rainfall series with different means and variances. The transformed series for each season were then averaged for the 17 stations to give seasonal indices for the period 1890 (1891 for summer) to 1994 (1995 for summer). Temperature Six stations with monthly maximum and minimum temperatures extending back to at least 1910 were obtained from the Australian Bureau of Meteorology (Figure 1). The choice of stations was based on quality of record, site location (to give as extensive coverage of Queensland as possible), lack of urban influence (see Lough, 1995) and that monthly maximum and minimum temperatures were reported regularly in the Monthly Weather Review, Queensland (published by the Australian Bureau of Meteorology). Again, allowing the data series to be regularly and promptly updated. One of the six stations, Longreach changed observing sites in the late 1960s. Comparison of the period of overlap between measurements made at the old and new sites provided monthly correction factors (c.a. ÿ011 to ÿ013 C) which were applied to the old record. Monthly average, maximum, minimum, and daily temperature range (DTR) were averaged over the six stations for the period January 1910 to March, 1995. The combined station series were then averaged for the six, 2-month and two, 6-month seasons, as for rainfall, and expressed as anomalies from the mean over the whole record period. Southern Oscillation Index El Niño–Southern Oscillation events have a major impact on climate in Queensland. The Tahiti–Darwin standardized index of the Southern Oscillation was used as a measure of this phenomenon. The series was averaged for the six, 2-month and two, 6-month seasons, as for rainfall and temperature. Analyses Changes in the series over time were assessed with linear trend analysis. Comparisons between series were made using correlation analysis. Anomalies for the 10- and 20-year periods were assessed for significant differences from the long-term mean using Cramer’s test (Mitchell et al., 1966). RESULTS Rainfall variations: 1890–1995 The Queensland summer rainfall index (Figure 2(a)), based on 17 stations, was very similar (r 0197, 1891– 1986) to the summer rainfall index published previously (Lough 1991). The latter was based on the first principal component of summer rainfall at 39 stations in Queensland, which accounted for 49 per cent of the total variance. This strong similarity between the two indices justified the use of the present (readily updatable) index based on a much smaller number of stations. It also confirmed the earlier finding (Lough, 1991) that interannual rainfall variations tend to be coherent throughout the state. The variability and wet and dry periods of the summer index (Figure 2(a)) also characterized rainfall for the three, 2-month seasons that made up the summer total (Figure 2(b–d)). None of these four series showed significant year-to-year persistence or significant long-term trends, i.e. there was no significant tendency to wetter or drier summer rainfall in Queensland over the 105-year record period. The wettest 10-year summer period, 1971–1980, was also wetter in DJ and FM, and to a lesser extent in the early monsoon season ON. The 1950s were also characterized by wetter summers, especially in ON and FM. The 58 J. M. LOUGH Figure 2. Seasonal rainfall variations in Queensland: (a) summer, (b) October–November, (c) December–January, and (d) February–March. Thick line is series filtered with 10-year Gaussian filter. CLIMATE INDICES FOR QUEENSLAND 59 Figure 3. Seasonal rainfall variations in Queensland: (a) winter, (b) April–May, (c) June–July, and (d) August–September. Thick line is series filtered with 10-year Gaussian filter. 60 J. M. LOUGH Figure 4. Annual (April–March) rainfall variations in Queensland. Thick line is series filtered with 10-year Gaussian filter. driest 10-year summer period, 1897–1906, centred on the extreme drought year of 1902 (see Gibbs and Maher, 1967) and was also evident in the component seasons. Over the most recent 20-year period, 1976–1995, 13 of the 20 summers in Queensland were drier than usual. This was most evident in the central part of the summer season, DJ. In the early part of the season, ON, it was wetter in the 1980s and drier in the 1990s. This contrasted with the latter part of the season, FM, which was drier in the 1980s and wetter in the 1990s. Averaged over the 20-year period, 1976–1995, summer rainfall was close to average and for the 10-year period, 1986–1995, although drier than usual it was not significantly so, and ranked only as the fifth driest 10-year period in the 105-year record of summer rainfall in Queensland. For the recent 10and 20-year periods both early (ON) and late (FM) season rainfall were close to average. Only in the central part of the season (DJ) was the most recent 10-year period significantly drier than usual and ranked second after 1931– 1940. Thus the most recent summer monsoons in Queensland, although dry, do not appear to be particularly unusual when considered in the context of the 105-year record. The dryness has been most consistent in the central part of the wet season, DJ. Winter rainfall in Queensland accounts for 20–35 per cent of the annual rainfall total. The Queensland winter rainfall index (Figure 3(a)) was very similar to the previously published Queensland winter rainfall index (Lough, 1991; r 0197, 1891–1986), which was based on the first unrotated principal component of rainfall at 39 stations. This similarity again demonstrated that a small number of stations captured the regional rainfall variations. None of the winter indices showed significant persistence or significant long-term trends. The wettest 10-year period was 1981–1990 but no 10-year period was significantly below the long-term mean apart from 1968–1977 in JJ. Winter rainfall was close to average over the most recent 10-year and 20-year periods. Recent winter rainfall variations in Queensland did not, therefore, appear to be particularly unusual when examined in the context of the past 105 years. There was a weak tendency for the amount of winter rainfall to be linked directly to the amount of rainfall in the ensuing summer season (r 0125, 1891–1994) but not between summer rainfall and that of the following winter (r 0103, 1891–1994). This relationship can be explained largely by the persistence of ENSO across the Southern Hemisphere spring and the relationship between ENSO and both summer and winter rainfall in Queensland (see below). An ‘annual’ mean rainfall index was formed as the average from April to March (Figure 4) to test whether there were any significant changes in the combined summer and winter series. There were not— the annual series showed no significant trends or persistence and the most recent 10- and 20-year periods were not particularly unusual in the context of the 105-year record. A similar result was obtained when the annual series was based on the calendar year and the October–September (‘water season’) year (not shown). 7 Temperature variations: 1910–1995 As found for rainfall, the temperature indices based on only six stations appeared to represent state-wide temperature variations. The correlations, over the period 1911–1987, between the present indices and a 13-station CLIMATE INDICES FOR QUEENSLAND 61 Figure 5. Variations of summer temperatures in Queensland: (a) average, (b) maximum, (c) minimum, and (d) daily temperature range. Thick line is series filtered with 10-year Gaussian filter. Dashed line is linear trend. 62 J. M. LOUGH Figure 6. Variations of winter temperatures in Queensland: (a) average, (b) maximum, (c) minimum, and (d) daily temperature range. Thick line is series filtered with 10-year Gaussian filter. Dashed line is linear trend. 63 CLIMATE INDICES FOR QUEENSLAND index published previously (Lough, 1995) ranged from 0194 for DTR in summer to 0198 for minimum temperatures in winter. Average temperatures in Queensland have risen in both summer (Figure 5(a)) and winter (Figure 6(a)). The trend towards warmer average temperatures was significant at all times of year apart from mid- and late winter (JJ and AS) and was greatest in magnitude in early winter (AM, Table I). Average temperatures over the most recent 10-year period were the warmest on record in winter, AM, JJ, and ON, and were significantly above the mean at all times of year except late summer (FM) and late winter (AS). Average Queensland temperatures in early winter (AM) were 111 C warmer than the long-term mean in this recent period. Over the most recent 20-year period, average temperatures were the warmest on record at all times of year except late winter (AS) and were significantly above the mean in AM, DJ, winter, and summer. Maximum temperatures have warmed only slightly in both summer (Figure 5(b)) and winter (Figure 6(b)). The trend was not, however, significant at any time of year (Table I). The most recent 10-year period was the warmest on record (although not significantly so) in mid-summer (DJ) and early winter (AM). Maximum temperatures were the warmest on record for the most recent 20-year period (although not significantly so) in late summer (FM) and early winter (AM). Minimum temperatures in Queensland warmed in both summer (Figure 5(c)) and winter (Figure 6(c)). The trend towards higher temperatures was significant at all times of year except mid-winter (JJ) and was greatest in early winter (AM, see Table I). Over the most recent 10-year period, minimum temperatures were significantly warmer than the long-term mean at all times of year. The warmest 10-year period on record was, however, either 1982–1991 or 1984–1993 (depending on the season) due to slightly less extreme minimum temperatures in 1994 and 1995. The most recent 20-year period was significantly warmer than the long-term mean at all times of year except mid-winter (JJ) and was warmest on record apart from late summer (FM) and late winter (AS). Minimum temperatures in early winter (AM) over this 20-year period averaged 1 C above the long-term mean. The DTR in Queensland decreased in summer (Figure 5(d)) and winter (Figure 6(d)). The trend was significant in all seasons except DJ, FM, and JJ (Table I), and the largest decrease occurred in early winter (AM). The DTR was significantly below the long-term mean during the most recent 10-year period in ON, AM, JJ and winter. This was not the most extreme 10-year period on record (1983–1992) because of a slight rise in DTR in 1994 and 1995. Over the most recent 20-year period, DTR was significantly below the mean in ON, AM, winter and summer. In winter and early summer (ON) this was the most extreme 20-year period on record. For the other seasons, the most extreme 20-year period was 1974–1993. The relationship between the DTR and maximum and minimum temperatures varied seasonally (see Lough, 1995). In summer over the period 1911–1994, DTR was related most closely to maximum temperatures (r 0177, significant at the 5 per cent level) than minimum temperatures (r 0100). Thus a decrease in DTR in summer would appear to be due largely to a decrease in maximum temperatures. In winter, DTR was related most closely to minimum temperatures (r ÿ0175, significant at the 5 per cent level) than maximum temperatures (r 0125, significant at the 5 per cent level). Thus a decrease in DTR in winter would appear to be due largely to an increase in minimum temperatures and a lesser decrease of maximum temperatures. These seasonally dependent Table I. Trend ( C per decade) for Queensland temperature indices. Italic indicates trend significant at the 5 per cent level. Average ON DJ FM Summer AM JJ AS Winter 0107 0108 0108 0107 0114 0105 0106 0108 Maximum 0103 0106 0105 0104 0105 0102 0102 0103 Minimum 0112 0109 0111 0110 0123 0108 0111 0114 DTR 7 0 09 7 0 03 7 0 07 7 0 07 7 0 19 7 0 05 7 0 09 7 0 11 1 1 1 1 1 1 1 1 64 J. M. LOUGH relationships between DTR and maximum and minimum temperatures were evident when each of the 2-month series were examined and were also stable over time. Linkages between rainfall, temperature, and El Niño–Southern Oscillation Relationships between summer and winter Queensland rainfall, temperatures and the SOI were examined over the entire common length of record and, to test for stability of the relationships over time, for three, 20-year sub periods; 1911–1930, 1931–1950, and the most recent 20-year period, 1975–1995. The period 1931–1950 was included as this has been noted in many studies because many of the teleconnections between ENSO and regional climate anomalies (including Queensland, see Lough, 1991, 1995) appeared to weaken (Troup, 1965; Trenberth and Shea, 1987; Elliott and Angell,1988; Gu and Philander, 1995). Queensland rainfall was significantly related to the SOI in both summer and winter (Table II) with high rainfall associated with anti-ENSO events and low rainfall with ENSO events. In summer, the strength of the relationship weakened dramatically from 1931 to 1950 and in winter the relationship weakened from 1911 to 1930. The relationship between Queensland summer and winter rainfall and the SOI was significant in the most recent 20year period. Summer rainfall in Queensland varied inversely with the four Queensland temperature series (Table II). Lower rainfall in summer (associated with reduced cloud and enhanced surface radiation) was linked to higher average, maximum, and minimum temperatures and an increase in the DTR. These relationships were stable over time. The relatively weaker relationship between rainfall and summer minimum temperatures was not significant from 1931 to 1950. The summer temperature series were significantly correlated with the SOI. The magnitude and significance of the relationship was stable between the first and last parts of the record only for maximum temperatures and DTR. Winter rainfall in Queensland was not related significantly to average temperatures. Winter rainfall was related inversely to maximum temperatures and DTR and related directly to minimum temperatures. Lower winter rainfall was linked to higher maximum and lower minimum temperatures and an increase in DTR. The relationship between winter rainfall and maximum temperatures was of greater magnitude over the most recent 20-year period than in the first part of the record. The relationship between minimum winter temperatures and rainfall was not significant over the most recent 20-year period. Despite these differences, the relationship between Queensland winter rainfall and DTR has remained stable and significant over the periods examined. Only winter DTR showed a weak, but significant, inverse relationship with the SOI in winter. Thus, DTR in both summer and winter varied inversely with Queensland rainfall and this relationship was stable over time (lower DTR and higher rainfall). In summer, the relationship between DTR and rainfall was linked more closely to changes in maximum than minimum temperatures. In winter, the DTR–rainfall relationship was inverse with maximum temperatures and related directly to minimum temperatures (i.e. decrease in DTR with higher rainfall associated with lower maximum and higher minimum temperatures). Table II. Correlations between Queensland temperature and rainfall indices and the Southern Oscillation Index. Italic indicates coefficients significant at the 5 per cent level. Summer Winter 1911–1995 1911–1930 1931–1950 1975–1995 RF and SOI RF and AVT RF and MAX RF and MIN RF and DTR AVT and SOI MAX and SOI MIN and SOI DTR and SOI 7 0 63 7 0 75 7 0 30 7 0 72 7 0 45 7 0 53 7 0 23 7 0 49 0159 1 1 1 1 1 1 1 1 7 0 76 7 0 81 7 0 57 7 0 76 7 0 57 7 0 66 7 0 35 7 0 69 0180 1 1 1 1 1 1 1 1 7 0 63 7 0 75 7 0 15 7 0 79 7 0 01 7 0 01 0 13 7 0 09 0116 1 1 1 1 1 1 1 1 7 0 77 7 0 85 7 0 54 7 0 68 7 0 33 7 0 52 7 0 02 7 0 70 1910–1994 1911–1930 1931–1950 1975–1995 0162 1 1 1 1 1 1 1 1 7 7 7 7 0154 0114 0134 0143 0172 0111 0109 0121 0131 7 7 7 0136 0124 0126 0154 0184 0114 0105 0118 0120 7 7 7 0156 0119 0123 0153 0176 0141 0107 0163 0158 7 0 04 7 0 58 0 30 7 0 70 0 12 7 0 10 0 22 7 0 31 0158 1 1 1 1 1 1 1 1 CLIMATE INDICES FOR QUEENSLAND 65 DISCUSSION AND CONCLUSIONS The first goal of this study was to determine if rainfall and temperature variations in Queensland over the past century could be described by relatively simple indices. Indices of rainfall based on only 17 stations and temperature based on only 6 stations closely match interannual variations based on greater spatial coverage of the State developed previously (Lough, 1991, 1995). Both rainfall and temperature appear to vary coherently over much of the State and a relatively small number of stations are necessary to capture that variability. The indices developed here can be up-dated promptly, and therefore provide useful tools for monitoring climate variation and change in this part of Australia. The second goal of the study was to determine whether recent climate variations in Queensland (through the Southern Hemisphere summer of 1994–1995) are in anyway unusual in the context of the 105-year rainfall and 85-year temperature records. Summer monsoon rainfall in Queensland has been below average over the past decade but this does not appear to be outside the range of variability found over the past century. The 1950s and 1970s experienced wetter summer monsoons, which could account for reports of an increasing trend in eastern Australia rainfall since about 1950 (e.g. Pittock, 1975; Nicholls and Lavery, 1992). There is, however, no significant trend towards wetter or drier conditions in Queensland over the past century either during the summer monsoon or the winter dry season. In contrast, temperatures in Queensland show significant changes over time. Average and minimum temperatures have increased significantly in both summer and winter and have been accompanied by a significant decrease in the daily temperature range (DTR), as found in other parts of the world (e.g. Karl et al., 1993). The changes in average and minimum temperatures and DTR have been of greatest magnitude in early winter (April– May). The most recent 10- and 20-year periods have witnessed, overall, the highest minimum and average temperatures and lowest DTR over the period since 1910. There is a weak indication that temperatures were less extreme in 1994 and 1995, but there is insufficient data to assess whether this represents a reversal of the major temperature trends. The El Niño–Southern Oscillation (ENSO) is a major control on Queensland rainfall variations in both summer and winter. The strength of this relationship has been maintained over the most recent 20 years. The period 1931– 1950, as reported elsewhere, is characterized by a breakdown in the relationship between ENSO and Queensland summer rainfall. The relationship between ENSO and Queensland winter rainfall broke down from 1911 to 1930. The daily temperature range (DTR) is significantly inversely related to Queensland rainfall in both summer and winter. Interannually, lower DTR is associated with higher rainfall and vice versa. The relationship between DTR and Queensland rainfall has been high and stable over the record period. Summer rainfall variations are related more closely to maximum than minimum temperatures, with higher temperatures associated with lower rainfall. Lower rainfall in winter tends to be linked with higher maximum and lower minimum temperatures. These relationships also appear to be relatively stable over time. The cause of the decline in the daily temperature range (DTR) in Queensland is unclear. Interannually, a lower DTR in summer is related more closely to a decrease in maximum than to any change in minimum temperatures and is also linked to higher rainfall. In winter a decrease in DTR is related more closely to an increase in minimum temperatures and is also linked to higher rainfall. Yet, the most recent 10-year period, when DTR in Queensland has been at its lowest, has been characterized by lower than average rainfall. ACKNOWLEDGEMENT Thanks to Barry Tobin for preparing Figure 1. REFERENCES Bureau of Meteorology 1995. Annual Climate Summary, Climate Analysis Section, National Climate Centre, Melbourne, Australia, Department of Environment, Sports and Territories, 10 pp. Elliott, W. P. And Angell, J. K. 1988. ‘Evidence for changes in Southern Oscillation relationships during the last 100 years’, J. Climate, 1, 729–737. Gibbs, W. J. And Maher, J. V. 1967. ‘Rainfall deciles as drought indicators’. Commonwealth Bur. Meteorol. Bull., 48, 33 pp. 66 J. M. LOUGH Gu, D. and Philander, S. G. H. 1995. ‘Secular changes of annual and interannual variability in the tropics during the past century’, J. Climate, 8, 864–876. Jones, P. D., Groisman, P. Ya., Coughlan, M., Plummer, N., Wang, W. C. and Karl, T. R. 1990. ‘Assessment of urbanization effects in time series of surface air temperature over land’, Nature, 347, 169–172. Karl, T. R., Jones, P. D., Knight, R. W., Kukla, G., Plummer, N., Razuvayev, V., Gallo, K. P., Lindesay, J., Charlson, R. J. and Peterson, T. C. 1993. ‘A new perspective on recent global warming: asymmetric trends of daily maximum and minimum temperatures’, Bull. Am. Meteorol. Soc., 74, 1007–1024. Lough, J. M. 1991. ‘Rainfall variations in Queensland, Australia: 1891–1986’, Int. J. Climatol., 11, 745–768. Lough, J. M. 1993. ‘Variations of some seasonal rainfall characteristics in Queensland, Australia: 1921–1987’, Int.J. Climatol.,13, 391–409. Lough J. M. 1995. ‘Temperature variations in a tropical–subtropical environment: Queensland, Australia, 1910–1987’, Int. J. Climatol.,15, 77–95. Mitchell, J. M., Dzerdzeevskii, B., Flohn, H., Hofmeyer, W. L., Lamb, H. H., Rao, K. N., and Wallen, C. C. 1966. Climatic Change, WMO Technical Note No. 79, WMO No. 195, TP100, WMO, Geneva, 79 pp. Nicholls, N. And Lavery, B., 1992. ‘Australian rainfall trends during the twentieth century’, Int. J. Climatol., 12, 153–163. Pittock, A. B. 1975. ‘Climatic change and the patterns of variation in Australian rainfall’, Search, 6, 498–504. Plummer, N. 1991. ‘Annual mean temperature anomalies over eastern Australia’, Bull. Aust. Meteorol. Ocean. Soc., 4, 42–44. Plummer, N., Lin, Z. and Torok, S. 1994. ‘Recent changes in the diurnal temperature range over Australia’, in Kukla, G., Karl, T. R. and Riches, M. R. (Eds), Asymetric change of daily temperature range. Proceedings of the International MINIMAX Workshop held under the auspices of NOAA National Environmental Watch and the DOE Global Change Research Program, 27–30 September, 1993, College Park, 025, US Department of Energy, pp. 127–143. Ropelewski, C. F. and Halpert, M. S. 1987. ‘Global and regional scale precipitation patterns associated with the El Niño/Southern Oscillation’, Mon. Weath. Rev., 115, 1606–1626. Salinger, M. J., Basher, R. E., Fitzharris, B. B., Hay, J. E., Jones, P. D., Macveigh, J. P. and Schmidely-Leleu, I. 1995. ‘Climate trends in the south-west Pacific’, Int. J. Climatol., 15, 285–302. Suppiah, R. and Hennessy, K. J. (1996). ‘Trends in the intensity and frequency of heavy rainfall in tropical Australia and links with the Southern Oscillation’, Aust. Meteorol. Mag., 45, 1–17. Trenberth, K. E. and Shea, D. J. 1987. ‘On the evolution of the Southern Oscillation’, Mon. Weath. Rev., 115, 3078–3096. Troup, A. J. 1965. ‘The Southern Oscillation’, Q.J.R. Meteorol. Soc., 91, 490–506. Whetton, P. H. 1993. ‘Assessing possible climate change in the Australian region due to the enhanced greenhouse effect’, Preprints, Fourth International Conference on Southern Hemisphere Meteorology and Oceanography, March 1993, Hobart, pp. 355–356. Whetton, P. H., Fowler, A. M., Haylock, M. R. and Pittock, A. B. 1993. ‘Implications of climate change due to the enhanced greenhouse effect on floods and droughts in Australia’, Clim. Change, 25, 289–317.