effect of climate change on global potato production, The

American Journal of Potato Research, Jul/Aug 2003 by Hijmans, Robert J

In the adapted model, radiation use efficiency (RUE) was made dependent on daily average temperature, following Kooman (1995) and Kooman and Haverkort (1995). RUE is highest (2.9 g MJ^sup -1^ (PAR)) between 15 and 21 C and zero below 2 C and above 34 C, with intermediate values in between. Decrease of RUE at high temperatures is due to increasing respiration (Kooman and Haverkort 1995). Radiation (PAR) above 12 MJ (PAR) m^sup -2^ day^sup -1^ was not considered, to account for light saturation (Kooman and Haverkort 1995).

For each grid cell, the model was run for 12 planting times (with planting at the first day of each month) and for five maturity classes of potato, representing different cultivars with early to late senescence. This was repeated for the heat-tolerant potato. Maturity classes used were 1000, 1200, 1400, 1600, or 1800 Cd, expressed as the temperature sum (thermal time) between emergence and harvest, with a base temperature of 2 C. The optimal planting time for a location (grid cell) was determined after the simulations, selecting the month/cultivar combination that led to the highest yield. Average temperature during the optimal planting time was calculated for each grid cell. To distinguish between the effect of changes in radiation and in temperature, the model was also run for the projected climate, while using radiation data for the current climate.

Current and projected potential yield were compared for two cases: with and without adaptation. Adaptation is narrowly defined as changes in the month of planting or in the maturity class of the cultivar. This is sometimes referred to as "autonomous adaptation" in the sense that these are inexpensive and can be carried out at the farm level (McCarthy et al. 2001). In the case of "without adaptation," potential yield for projected conditions is calculated for the combination of cultivar and month of planting that gave the highest yield for the current climate. In the case of "with adaptation," the highest yield is taken from the 60 (5 cultivars x 12 months) months of planting/cultivar combinations for the projected climate scenarios. Hence, in this latter case, the month of planting and cultivar type in a location (grid cell) can be different for current and future climates.

Maps and Potato Distribution Data

For each grid cell, the mean potential yield (over all climate scenarios) was compared with the yield as calculated for the current climate. Maps were made of changes in potential yield for the "with adaptation" and "without adaptation" cases for the 2010-39 and the 2040-69. Results were also summarized on maps indicating the potential contribution to yield of heat-tolerant potatoes.

The maps only include data for the current areas with potato production according to a global 1-degree grid described by Hijmans (2001) (7004 cells with potato area; 42% of the global land area excluding Antarctica). These potato distribution data were also used to weigh changes in temperature and yield by potato area. Such potato-area-weighted results were obtained by multiplying the grid of potato area with the grids of current and projected temperature and yield, and dividing these by the total global potato area. Thus, in the aggregate results, the weight of an area (grid cell) with, for example, 10,000 ha of potatoes would be twice that of an area with 5,000 ha. Results were summarized for countries with more than 100,000 ha of potato area. Average change in yield and the percentage of grid cells where climate change would lead to higher yields was calculated for these countries.