CLIMATIC CHANGE IN RELATION TO SUSTAINABLE AGRICULTURE IN CANADA * A. Bootsma, D. Anderson and D. Chaput Agriculture Canada, Research Branch, Centre for Land and Biological Resources Research, Land Resource Division, Central Experimental Farm Bldg. #74, Ottawa, Ontario, Canada K1A 0C6. INTRODUCTION The question may be asked, 'Is our climate sufficiently stable to sustain agricultural production in the long term, or are we headed for economic and ecological problems due to climatic change induced by human activity?'. To answer this question, we examined the yearly variability and the long term trends of selected agroclimatic variables over the past century at 5 locations in Canada (see Figure 1 and Table 1). Stations were selected because of their geographic location and the length and consistency of their long term record. All 5 stations were located at agricultural Research Stations maintained by Agriculture Canada. Most stations had at least 100 years of record and were considered relatively free of urban influences. METHODOLOGY Selected agroclimatic variables were computed on a yearly basis from the daily climatological record (Table 2). These variables were selected because of their considerable importance to agriculture. Simple linear correlation coefficients (r) were computed for each variable with time (years). Five-year moving averages were plotted in relation to the long term mean to identify if any trends were evident in the data. Cumulative frequency distributions (probabilities) were plotted and significance test were applied to determine if distributions for the last 30 years differed from the long term. RESULTS Correlation Coefficients Correlation coefficients (Table 3) indicate that spring frosts have become earlier and fall frosts later at the 3 locations in western Canada. These locations also show strong evidence of warming during the growing season over the past 100 years (highly significant positive r value for GDD). This warming trend was not evident for the 2 stations in eastern Canada. Agassiz was the only station with a significant decrease (negative r) in precipitation (PREC), while Charlottetown showed an increase. 5-Year moving averages Yearly values and 5-year moving averages for each variable are shown in Figure 2 for Ottawa and Figure 3 for Indian Head. These stations typify results for eastern Canada and the western prairie region, respectively. Results for Ottawa show the highly random nature of the data with a clear lack of long term trends. This is consistent with the lack of significant r values (Table 3). Results for Indian Head clearly show increasing GDD over the past 100 years and slight trends in earlier spring and later fall frost dates. Cumulative frequency distributions Cumulative frequencies (probabilities) for the last 30 years are compared to the long term (~100 year) values in Figures 4 and 5 for Ottawa and Indian Head respectively. Results for Ottawa indicate spring frosts have tended to be 5-7 days earlier in the last 30 years, except at extremely high and low probability values. GDD have increased significantly at low probability levels for the 1961-90 period. The other variables indicate little change at Ottawa. At Indian Head (Figure 5) spring frosts averaged about 5 days earlier during the last 30 years. A dramatic increase in GDD was observed for 1961-90. On average, GDD increased by 86 units (about 5%) over the season, which is equivalent to an average increase in temperature of 0.5¡C. Other variables showed little change at Indian Head. Statistical comparisons of frequency distributions The Wilcoxon rank test was used to determine if frequency distributions for the period 1960-198~ were statistically significantly different from the period before 1960 for all 5 locations. This test does not assume that values are normally distributed. Results are shown in terms of the significance of Z- values (Table 4). This test further confirms that statistically, the increase in GDD on the Canadian prairies is highly significant. It also indicates a significant trend to earlier dates of last spring frost (SF) at 3 locations. Charlottetown is the only location showing a significant change (increase) in precipitation. CONCLUSIONS Eastern Canada Relatively small changes in agro-climate have occurred. Climate change not a great concern for agriculture. Present agricultural practices not adversely affected. Prairie Region Significant shifts in climate, notably a warming trend during the growing season. Increased drought stress with higher potential evapotranspiration. Adjustments in management practices may be necessary if change continues. Possible causes for change: - global warming from 'greenhouse' gases. - increased area cultivated, summerfallowed on prairies. - change in weather circulation patterns (e.g. due to El Nino). * Poster paper presented at Agricultural Institute of Canada, 73rd Annual Conference, Memorial University, St. John's, Newfoundland, Aug. 18-21, 1993.