IS OUR CLIMATE CHANGING? A. Bootsma Agriculture and Agri-Food Canada, Research Branch, Centre for Land and Biological Resources Research, Central Experimental Farm, Bldg. #74, Ottawa, Ontario K1A 0C6 Introduction The topic of climate change is receiving much attention in the news media these days. Only a few decades ago, reports of scientists predicting the onset of an ice age were not uncommon. Today, the emphasis has switched to the concern over the potential effects of rising concentrations of Ógreenhouse gasesÓ in the atmosphere on global warming. The issue of climate change is no doubt an important one for agriculture and forestry since climatic variations can have significant impacts on production in both sectors. In this presentation, I would like to focus briefly on several aspects of climate change and variability ranging from some global concerns to some very local site-specific information. Firstly, I would like to make a few comments about climatic change in general, and then present some results of a study of climatic trends at Charlottetown for the past 100 years. Then we will take a brief look at some examples of how information on the variation in climate over space and time at the regional as well as local on-farm scale can be used in crop management decisions. Long Term Climatic Changes Climatic changes have occurred in the past and will likely continue to occur in the future. Figure 1 provides some indication of the global temperature trends that scientists believe have occurred in the past. During the last millennium, a warm period occurred from about 1100 to 1300 AD and what climatologists often call a "little ice age" took place from about 1400 to 1900 AD (Fig. 1c). Although these changes appear relatively small (about + 0.5¡C in average global temperatures), they were large enough to have significant impact on agriculture, especially in Europe, at the time. In the longer term (e.g. past 1 million years) scientists tell us that the earth has experienced a series of ice ages during which global temperatures were as much as 4¡C below the present (Fig. la). At the present time relatively little is known about why past climatic changes occurred. Even less can be said about future changes that might be expected In addition, past changes have mast often occurred over time periods that are much longer than those of concern to a farming enterprise. Therefore, information on long term climatic changes is presently not of much value for use in management decisions by farmers and foresters. Global Warming and the "Greenhouse EffectÓ In recent years much interest and debate has been generated over possible global warming due to the "greenhouse effect". This warming is expected to occur because of increases in concentration of carbon dioxide (CO2) and other "greenhouse gasesÓ in the atmosphere. CO2 concentrations have increased almost 30% above preindustrial levels and continue to increase each year (Environment Canada, 1985), largely due to burning of fossil fuels and clearing of forests. There is considerable uncertainty over the expected effect that rising concentration of greenhouse gases will have on climate. However, most scientists agree that the influence will tend to increase temperatures. Scientific studies have suggested that global temperature increases of about 2 to 5¡C are likely with double CO2 concentrations (which could occur by the year 2050, depending on emission trends). However, these estimates are based on simplified models of the behavior of the atmosphere and may not adequately account for all the complex processes. Many studies have been done to determine if global temperatures have already increased as a result of the present build-up of CO2. Some have concluded that an increase in global temperature of about 0.5¡C has occurred within the past 100 years (Hansen and Lebedeff, 1987). While this is consistent with predictions of greenhouse warming, it is also within the bounds of natural fluctuations and could be due to natural causes. Others have suggested that significant increases due to increasing CO2 have not yet occurred (Maddan and Ramanathan, 1980; Hanson et al., 1989). Because of the present uncertainties associated with global warming and the greenhouse effect, it is premature to begin using present predictions in operational management decisions in agriculture and forestry. 100-yr Climatic Trends at Charlottetown Charlottetown was one of 5 locations across Canada for which we studied long term (100 yr) trends in climatic parameters that are of importance to agriculture and to some extent forestry (Bootsma, et al., 1993; Bootsma, 1994). Although 17 parameters were studied, time will only allow for a quick look at the most important ones. Figure 2 shows the 100 yr trend in growing season (May-September) precipitation. The graph indicates considerable increase in the variability in precipitation after the mid 1930Õs. We have not had a chance to confirm if a similar trend exists at other Maritime locations. In the later years, increased frequency of high and low precipitation amounts likely caused greater fluctuations in crop yields. In addition to increased variability, there was a slight but noticeable increase in average precipitation over the 100 years. However, 5-yr moving averages cycled irregularly above and below the mean value. Figure 3 shows the 100 yr trend in accumulated growing degree-days (GDD) above 5¡C (heat units) over the growing season. GDD are directly related to temperature and do not appear to be increasing significantly at Charlottetown. In fact, GDD reached a peak in the 1940's, and decreased to the 1970's. The decline in temperatures from the 1940Õs to the 1970Õs is a phenomena which has been observed by others as well throughout eastern Canada (Berry, 1991) and the whole northern hemisphere (Jones, et al., 1986). Other significant trends observed in the Charlottetown data over the 100 years (not shown) were slightly earlier dates of first fall frost and an increase in winter snowfall amount. Variables for which significant time trends could not be detected included a moisture stress index for the growing season of forage crops, date of last spring frost, seasonal Corn Heat Units, growing season start and end dates, potential evapotranspiration, and mean temperatures for January and July. The results of this study further confirm that there is little direct evidence that global warming is having a significant effect on climate in the Maritimes. However, it is possible that the effects of greenhouse warming have been offset by other factors causing cooling. Because of the uncertainty, information on long-term climatic trends and global warming is presently not very useful in planning and decision-making for farmers and foresters. This leads us to conclude that the best estimate of what our average climate and its variability will be like in the next 20 to 30 years or so, is the record we have of the past 20 to 30 years or more, which is the view traditionally held by climatologists. From the past records we can know how much the climate varies from year to year and from one location to another (sometimes even at the farm level, e.g. with respect to frost dates and night time minimum temperature). While it is impossible to use climatology to predict exactly when an extreme year will occur, it is possible to use the information to estimate the frequency and extent of extreme climatic events. The following sections provide a few examples of how this knowledge of climate variability over time and space can be used to provide information for use in farm management decisions. Corn Heat Unit Variability in the Maritimes A study of the heat units (CHU) available for corn production in the Maritimes has recently been completed, mainly using data from 37 climate stations in the region for the 1966-85 (30-year) period (Bootsma, 1991; Bootsma, et al., 1992). I would like to briefly summarize the main results of this study and indicate how this information can be used in management decisions on the farm. CHU available for grain corn production were calculated on a daily basis using daily maximum and minimum air temperatures and then summed up for each year from an estimated seeding date to the date of first killing frost (-2¡C) in the fall. Average values for the 1966-'85 period (Figure 4) range from over 2600 CHU in the Annapolis valley near Kentville and the Saint John River Valley below Fredericton, to less than 2200 CHU in northern New Brunswick and southeastern Nova Scotia. CHU available for silage corn production (not shown) are generally about 150 CHU lower than those available for grain corn due to earlier fall cut-off dates. Average CHU ratings are useful for comparing corn production potential of different areas and identifying which hybrids are most suited for a given area. Because of the relatively cool, short growing season in the region, areas of highest corn growing potential are those with greatest amount of heat units. Heat units are mainly important in promoting earlier maturity or higher dry matter content in silage or grain at harvest. It should be noted that, for unknown reasons, CHU are less effective in maturing corn on P.E.I. and possibly other coastal regions than on the mainland. This can be accounted for by adding about 150 CHU to the hybrid ratings (not shown) in these areas or subtracting 150 CHU from the average CHU rating in Figure 4. At present, recommended corn hybrids are available which have good potential for shelled grain corn in areas with an average of about 2500 CHU or more (2650 in P.E.I.). High moisture grain or cribbed ear corn can be produced in most years in areas with 2300 CHU or more (2450 in P.E.I.) using the very earliest available hybrids. Management decisions such as whether or not to grow grain corn or which hybrid to select are often best made using additional information on how often and how far yearly CHU sums differ from the average value. Figure 5 indicates the probability of CHU being above or below the average values shown in Figure 4. For example there is a 5% probability (1 year in 20) that CHU are 300 units or more below the average. Similarly, 1 year in 10 (10% probability), CHU are 220 units or more above average. Normally, it is probably appropriate to base decisions on the 10 to 20% probability level, but the user can select whatever risk level is appropriate. For example, if a producer wants to ensure that a grain hybrid reaches maturity (35% moisture in the grain) before frost 9 years out of 10, it will be necessary to grow a hybrid requiring 240 CHU less than the average CHU rating in the area (1 year in 10 CHU are more than 240 units below the average). Information on the probability that certain CHU sums are or are not reached can provide the grower with an estimate of the risk of not reaching maturity before frost for a given hybrid. Although the CHU system has its limitations (for example, local factors can affect availability of CHU; hybrid requirements can vary between locations and seasons), nevertheless, it is a good example of how information on past climate can be used to help make appropriate management decisions on the production of corn and other warm-season crops at the farm level. Similar information has also been generated for growing degree-days above 0, 5 and 10¡C base temperatures for the Atlantic region (Gordon and Bootsma, 1993), which may be more applicable to other crops such as beans, peas, forages and selected vegetables. Frost Risk on the Farm One of the most important climatic ÒchangesÓ that producers of cost-sensitive crops experience on P.E.I. and elsewhere in the Maritimes is that of local variation in nighttime minimum temperatures on clear calm nights in areas of hilly terrain or near the coast. Figure 6 is a typical example of minimum temperatures observed in the hilly area of southeastern PEI on a clear, calm night in autumn. Differences of 6¡C or more in temperature can occur over relatively short horizontal distances on slopes with elevation differences of 30 meters or more. These temperature gradients result in significant differences in dates of occurrence of last spring and first fall frost at the local, farm-field level. In these situations long term minimum temperature records at the nearest climate station are not of much direct value in assessing spring and fall frost risk on the farm. Similarly, a minimum temperature forecast for the nearest synoptic weather station (Charlottetown A) cannot adequately predict on-farm temperature conditions on nights when frost may be expected. Both accurate frost risk assessments and minimum temperature forecasts at the field level can be useful to producers of frost sensitive crops for management decisions such as planting and harvesting dates, field selection, and active frost prevention (e.g. sprinkler irrigation). Based on numerous night-time minimum temperature measurements made on P.E.I. in the mid 1970's (primarily on tobacco farms in the southeastern region), detailed frost risk maps have been prepared of selected areas of P.E.I. Figure 7 is an example of a typical frost map of a farm located in hilly terrain. The isolines represent the average date of first fall frost, ranging in this case from Sept. 15 to Oct. 9. Areas between isolines were assigned frost classes with maximum range from A1 to D2, with A1 representing the safest area from the standpoint of frost and D2 having the greatest risk Table 1 illustrates spring and fall frost dates at 10% and 50% risk level for each frost class. The 50% risk level corresponds to the average frost dates. The 10% risk level indicates the date on or after which in spring and on or before which in fall a frost may be expected 1 year in 10. Frost dates typically differ by as much as 4 to 5 weeks between extreme classes (e.g. A2 and D1). The dates apply to a minimum temperature threshold of 0¡C at a height of 1 to 1.5 metres above ground. Dates are also available for lower threshold temperatures (e.g. -1 or -2¡C). Under some conditions, temperatures near ground level may be several degrees colder than at 1.5 metres and therefore "ground frosts" may occur later in spring and earlier in fall than shown in Table 1. The typical minimum temperature differences that may occur between the Charlottetown Airport weather site and on-farm locations on clear calm nights are also indicated in Table 1. Temperatures in some low-lying areas (e.g. frost classes D1 and D2) may be more than 6¡C colder. This information can help growers predict on-farm minimum temperatures from a weather forecast for the nearest synoptic weather station (Charlottetown). On nights with partial cloud or with wind, temperature differences will be less, but can still be estimated from forecast information (Bootsma, 1980). Since temperatures near ground level may be lower than at 1.5 metres, additional measurements may be needed to establish minimum temperature relationships for low-growing crops such as strawberries. Frost risk maps have been prepared for most of the hilly areas of central and southeastern P.E.I. The information has been overlaid on 1:10,000 scale contoured orthophoto maps. Map sheets that have been completed are indicated in Fig. 8. Growers wishing to obtain a copy of the map(s) covering their farm area can contact Agriculture and Agri-Food Canada, Soil Survey Office or the P.E.I. Department of Agriculture, Fisheries and Forestry in Charlottetown. A less detailed frost map showing estimates of spring and fall frost dates for all of P.E.I. at a scale of 1:125,000 is also available (Bootsma, 1980). Conclusions Climate is one the basic natural resources available to agriculture and forestry, and as such, climatic change can have significant impact on the productivity of both sectors. Changes in climate in the past have often occurred gradually and over longer time frames than are of interest to producers. A study of 100 yr climatic trends at Charlottetown showed no discernible warming trend. The impact of rising concentrations of "greenhouse gases" on the Maritime climate as well as elsewhere on the globe are presently still not very clear, although warming is anticipated. Consequently, information on long term climatic changes that have taken place in the past and estimates of potential future changes that may be expected due to the greenhouse effect are presently not of much help to farmers and foresters in making management decisions. Several examples have been presented illustrating information on recent (e.g. last 30 yr period) climate variability over time and space that can be useful to farm managers in decision making. Information on heat units available for corn production in the Maritime region based on a recent 30-yr period can help identify areas most suited for corn production, select suitable hybrids and assess the level of risk involved. Detailed estimates of the risk of spring and fall frost and nighttime minimum temperature at the farm field level can help in management decisions for frost- sensitive crops and in providing on-farm minimum temperature forecasts. A number of publications on these topics indicated in the reference section are available upon request. Presentation made at Seminar entitled: ÒWeather or NotÓ - Increased Profits and Production in Farming and Forestry through the Effective Use of Weather and Climate Information, sponsored by the PEI Climate Advisory Committee, North River, P.E.I. April 7, 1994. CLBRR Contribution No. 94-27. References Berry, M.O. 1991. Recent temperature trends in Canada. The Operational Geographer 9:9-13. Bootsma, A. 1980. Frost risk survey of Prince Edward Island. P.E.I. Dept. of Agric. and Forestry, Charlottetown, and Land Resource Research Institute, Agriculture Canada, Ottawa, 35 pp. Bootsma, A. 1991. Risk analyses of heat units available for corn production in the Maritime provinces. Agric. Canada, Res. Branch, Tech. Bull. 1991-8E, 49 pp. Bootsma, A. 1994. Long term (100 yr) climatic trends for agriculture at selected locations in Canada. Climatic Change (in press). Bootsma, A., Anderson, D. and Chaput, D. 1993. Climatic change in relation to sustainable agriculture in Canada. Agric. Canada, Centre for Land and Biological Resources Research, Ottawa, Ont. Poster paper presented at Agriculture Institute of Canada, 73rd Annual Conference, Memorial University, St. John's, Nfld., Aug. 18-21, 1993, 10 pp. Bootsrna, A., Gordon, R., Read, G. and Richards, W.G. 1992. Heat units for corn in the Maritime provinces. Atlantic Committee on Agrometeorology Publ. No. ACA 92-1, 8 pp. Environment Canada. 1985. Understanding CO2 and Climate. Canadian Climate Centre, Atmospheric Environment Service, Downsview, Ont. Annual Report 1984, 18 pp. Gordon, R. and Bootsma, A. 1993. Risk analyses of growing degree-days in Atlantic Canada. Agric. Canada, Research Branch, Centre for Land and Biological Resources Research, Tech. Bull. 1993-SE 147 pp. Hansen, J. and Lebedeff, S. 1987. Global trends of measured surface air temperature. J. Geophys. Res. 92:1334513372. Hanson, K., Maul, G.A. and Karl T.R. 1989. Are atmospheric "greenhouse" effects apparent in the climatic record of the contiguous U.S. [1895-1987]? Geophysical Research Letters 16(1):49-52. Hengeveld, H. 1991. Understanding Atmospheric Change. Environment Canada, Atmospheric Environment Service, SOE Report No. 91-2, 68 pp. Jones, P.D., Raper, S., Bradley, R., Diaz, H., Kelly, P. and Wigley, T. 1986. Northern hemisphere surface air temperature variations: 1851-1984. J. Clim. Appl. Meteorol. 25:161-179. Madden, R.A. and Ramanathan, V. 1980. Detecting climate change due to increasing carbon dioxide. Science 209:763-768.