|
A Summary of some Results of Research on Potential Impacts of Climate Changeon Agriculture in Eastern Canada |
|
| Andrew Bootsma, Agroclimatologist, Eastern Cereal and Oilseed Research Center (ECORC) | |
|
We have been involved in three projects that were funded by the Climate Change Action Fund in Canada. These involved scientists from Agriculture and Agri-Food Canada, and other outside collaborators. These were relatively short-term projects of 1 to 2 years. Therefore, we have used only one climate scenario and that is from the Canadian General Circulation Model. The results are limited in that sense. The three projects were:
The study of the Impacts of Agriculture Production in the Atlantic Region (Slide #4) looked at three climate indices: crop heat units (CHU) for corn and soybeans, effective growing degree days for barley-a cool season crop, and water deficits (defined as potential evaporation minus the precipitation.) We took the large-scale GCM changes in temperature and precipitation and we interpolated them to a small scale - to a fine grid of about 10-15 kilometers. Then we applied those changes to the observed data, for 1961 to 1990. That 61-90 grid was constructed from station data with the help of a digital elevation model. We calculated the crop heat units for three time periods: 1961 to 90 and for two future time periods, 2010-2039 and 2040-2069. You can see the warming effect in the increase in crop heat units (Slide #5). We also looked at the change in crop heat units between the two time periods. In the 2040-2069 time period we´re looking at increases of about 500-600 crop heat units using this scenario from the Canadian Model. The same with effective growing degree days. (Slide #6) This was more for barley, with a lower base temperature of 5¼C (42¼F). The same trends are seen because it is also a temperature-based unit. For water deficits we saw some increase in the interior and some decreases along the coastal regions. (Slide #7) They vary plus or minus one inch by the year 2055. Not a great big change in water deficits for that part of the world. Then we wanted to know what are the likely impacts of theses changes on crop yield. We looked at historical yield data from corn hybrid trials, ranging from the coolest regions in the Atlantic area (about 2300 CHU) right down to southern Ontario down towards Windsor, where we get up to 3300-3400 corn heat units. We get a very nice relationship between yield and bushels per acre and corn heat units from these field trials (Slide #8). So we can see as the climate warms under existing conditions we get considerably higher yields. So that gives us a good clue of what might happen when there is warming taking place. The scope of this yield increase is about 10 bushels of corn/acre for each 100-heat unit increase. Then we also looked at crop yield statistics from the farm level (Slide #9). A similar relationship between yield and available heat units exists, but there is quite a bit more scatter in the data and you don´t reach quite the same increase in yield potential. The same is true for soybeans (Slide #10). For soybeans, the yield increases are less; on a percentage basis they come close to corn. These data are from soybean variety trials all across eastern Canada, from the warmest regions to the coolest. Again, if we compare the farm statistics there is a lot more scatter in the farm yield data and there is a lower yield trend (Slide #11). That increase in variability and lower trend may be due to a number of reasons. It could be because the yield and the crop heat unit information is less accurate. It could also be because crops may be grown under less optimum soil and/or management conditions at the farm level compared to corn hybrid trials. With barley we see a slightly negative effect with warming temperatures (Slide #12). As effective growing degrees increase, yields in barley seem to be decreasing. This is not a very significant trend. Barley is a cool season crop that tends to do better under cooler climates. Cooler conditions lengthen the growing period and allow for a longer period of dry matter accumulation. When we look at the statistics for farm yields over the range of degree-days in eastern Canada, there is a lot of variability (Slide #13). There is no big trend except perhaps you could say that in the middle regions, between about 1500-1700 growing degree-days in Celsius units, we seem to get less variability in the yield. We also included the water deficits in some of these relationships and it did not explain the changes in yields any better than using only heat units. We don´t think water deficits are important in the equation estimating yields on an average basis for conditions in eastern Canada. Our conclusions from the study were that corn and soybean yields and acreages in the Maritime Provinces of Canada were likely to increase significantly with climate warming. Barley yields were not likely to change significantly and therefore acreages were likely to decrease as producers switch from barley to corn and soybeans. The change in the water deficits was not likely to impact average yield very much. Certainly there are years when you have water deficits that will have a negative impact on yield, but it´s likely going to be much the same across all locations. We constructed a production scenario for the present and future climate by 2055 that indicate a very significant increase in yield and acreage in the corn and soybeans, and a slight increase in yield in barley (because the direct effect of higher CO2 on barley yields would more than offset the negative effect of warmer temperatures), and big increases in corn and soybean production and decrease in barley production. Barley acreage is still significant in our scenario because it is an excellent crop in rotation with potatoes. Comparing the climatic changes from the GCM model to the Great Lakes region, we note that there is slightly more warming in the Great Lakes area than in the Maritime Provinces during the spring period (Slide #16). Summer precipitation shows a slight drop in the Maritimes and a lesser drop in the Great Lakes area (Slide #17); in Wisconsin and Minnesota there are increases in precipitation shown by this Canadian Model output. To put this model into perspective in terms of many of the other GCM model results that are available, the Canadian GCM model output that we used is more or less in the middle of the scatter plot for changes in both temperature and precipitation during the summer period (Slide #18). This scatter plot is for a location near London, Ontario, but plots for southern Michigan, northern Ohio to northern Illinois and for northern New York south of Lake Ontario would be very similar. As you go north and west of the Great Lakes into Wisconsin and Minnesota, scatter plots show precipitation changes in the +0 to 25% range for pretty well all of the GCM models. Q: Roy Black Is it going to rain in the Prairie Provinces? A: No. It´s showing drier conditions. Comparisons with other GCM models from several different countries, for the annual period, show that temperature increase is on the high side in the Canadian model (Slide #19). Precipitation changes are near the midpoint, or perhaps a bit lower than the average for the other models. Again, this scatter plot is for a point near London, Ontario, but would be very similar to plots for many areas around the Great Lakes in the USA. When you move into Wisconsin and Minnesota, however, most GCM´s show precipitation change above the 0% line, but temperature changes are roughly the same until you go further west, where the temperature increases are higher. We took a quick look at average corn yields in the Corn Belt (based on yield information gathered from corn hybrid trials conducted in that area) and how it relates to average corn heat units available (Slide #20). You don´t see that nice relationship that we had in the Canadian data from about 2200-3400 heat units; as you get above 3200 heat units, you don´t get a clear linear trend of increased yield any more. This may be because of increased water deficits, but even in some irrigated trials there isn´t a linear trend (for example, some trial data from Nebraska and Missouri). Some of the scatter in the data may be due to the effect of differences in management, soil conditions, etc. on average yields, but we have not had the time to analyze these results further to try and explain the yield differences that we see. For areas in Ontario near the Great Lakes we see an increase of about 400 heat units by 2010-2039, and 800 corn heat units by 2040-2069 (Slide #21). I would expect areas in the USA near the Great Lakes that presently have 2800 crop heat units to show increases in yield up to the year 2025 or beyond (Slide #22). Areas that are down to about 2400 crop heat units should show increases up to the 2050´s or beyond. How do water deficits change based on this GCM model scenario? These data are for southern Ontario locations (Slide #23). We´re looking at a 1-2 inch increase in water deficits by 2040-2069. This is probably similar for areas in the USA near the Great Lakes region. This increase will not have a big impact on yields in areas that have 7-8 inches of water deficit now (Slide #24). We did not see a trend at those levels in our Canadian data. As you get further south into higher heat unit areas, water deficits may become more important. Corn yields vs. Water Deficits on the Great Plains shows a lot of scatter (Slide #25). Over 10 inches (e.g. Nebraska data) shows a very significant drop in yields. Crop Yield and Variability from the EPIC (Environmental Policy Integrated Climate) Model In this study we simulated annual yields for a baseline in the 1965-1995 period and for 2 x CO2 in the 2040-2060 period for barley, spring wheat and several other crops using the EPIC model. There were 29 locations across the country (Canada) (Slide #27). We will focus on the results from the southern Ontario and northern Manitoba areas, those closest to the Great Lakes. Looking at the changes in monthly mean temperature, January, February and March show large changes up to 11°F (Slide #28). Precipitation changes are not that great, but there are some decreases in July and August. A summary of the results averaged for southern Ontario showed increases in yield of spring wheat, soybeans, and winter wheat (Slide #30). Barley yield did not change very much. The model showed a decrease in corn yield in Ontario, some of which can be alleviated by adding more nitrogen fertilizer to the crop in the model. (These negative yields are a bit suspicious for southern Ontario, as the field trial data tend to suggest yields could still increase, except perhaps in the highest heat unit areas.) The model indicated lower averaged yields in southern Manitoba for barley, spring wheat and canola. The model also indicated that yields would likely be more variable under the 2 x CO2 climate, as indicated by the increases in standard deviations. It should be noted that the EPIC model took into account the direct effect of higher CO2 on crop yields. Impacts of Risk of Winter Damage on Perennial Crops This focused on both forage crops and fruit trees. We developed a suite of climatic indices that would give some indication of what might happen. Indices for forage crops were:
Indices for fruit trees:
The station results for southern Ontario and some locations near Lake Superior were examined (Slide #33). For forage crops we expect some reduced hardening in the fall, due to warmer temperatures during the hardening phase. There would also likely be some loss of hardiness during the winter because of mild periods during the cold. There would be less protection from snow cover during the cold period in the colder regions of Ontario. More heaving and smothering is likely in the colder areas, with less effect in the milder areas because the cold period starts to shrink and in mild areas it becomes insignificant. Overall we expect increased risk damage in most areas of eastern Canada; effects on some areas near the Great Lakes that are very mild now are less certain. For tree fruits we concluded that fall hardening would improve due to shorter day length at first autumn frost. There would be less cold stress in the winter because of fewer temperatures less than 5°F and a higher minimum temperature. More de-hardening is likely in cold areas because of warm temperatures, less in milder regions. Less bud damage due to spring frost in the cold areas, but in some of the milder areas there could be more. This assumes that chilling requirements for dormancy before January 1 is met; otherwise bud burst would be delayed, which would reduce the risk. Overall conclusions: New varieties or species may be possible in current regions. There may be some northward extension of commercial production of tree fruits. More stable production in currently marginal areas is likely, because of lower risk of spring frost damage. Q: George Hubka, Landowner/Crops Farmer Does the area around Essex County get more protection because bodies of water surround it? A: It is a much more moderate winter climate, so we get more production. It would have less snow cover and that´s critical (for some crops like winter wheat). There may be years when lack of snow cover causes problems. Q: Jay Harman, Department of Geography, Michigan State University We heard about tart cherry flowering and the freezing of Grand Traverse Bay, for Atlantic Canada like the Annapolis Valley, do you have any information about changes such as warming indicated by earlier flowering going on there? A: The winter injury data included locations in the Maritime Provinces and the report does cover that. An electronic copy is available (Slide #37). |
