effect of climate change on global potato production, The

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

All these studies on potato were conducted for small regions at high latitudes and their results are difficult to extrapolate to other regions. In this paper, some of the possible effects of climate change on potato production are studied at the global level. A simulation model was used to calculate potential potato yield for the current climate and for projected future climates in 2010-2039 and 2040-2069, using seven climate scenarios from five different climate models. The goal of the paper is to identify regions where there is likely to be a strong decline in productivity due to the increase in temperature, and to determine the extent to which heat-tolerant potato cultivars would be useful to mitigate the effect of climate change in those regions. Only the effect of changes in temperature and solar radiation was considered and not the effect of changes in rainfall, of increased levels of atmospheric CO2, or of increased ultraviolet radiation. The results are presented in relation to the current global distribution of the potato crop.

MATERIALS AND METHODS

Climate Data

Average monthly climate data for 1961-1990 (hereinafter referred to as "current climate"), and for climate change from that period to 2010-2039 and to 2040-2069 were used. For the current climate, data from New et al. (1999) were used; for the projected future climate, seven scenarios from five climate models were used (Table 1). The projected change in climate was superimposed on the current climate to create projected climate surfaces for 2010-2039 and 2040-2069 (further refereed to as 2010-39 and 2040-69, respectively). Daily temperature and radiation data were derived by linear interpolation between the monthly averages. All datasets were resampled (statistically disaggregated by interpolation) to a 1 by 1 degree resolution, using IDRISI software (Clark Labs, Worcester, MA, USA) (the data for the current climate were available at a 0.5 by 0.5 degree resolution and were aggregated).

Global average temperatures for the current climate and for all scenarios were calculated for terrestrial cells only (16,862 cells), without considering Antarctica, taking the size of each square degree grid cell into account (the size of the one square degree cells decreases with increasing latitude).

Simulation Model

The LINTUL simulation model as described by Stol et al. (1991) and Van Keulen and Stol (1995) was used to calculate potential yield. The model has a temperature-dependent development of the canopy (green ground cover). Biomass production is the product of the fraction green ground cover, incident solar radiation and radiation use efficiency.

Stol et al. (1991) estimated tuber yield as a temperature-dependent percentage of the total biomass accumulated during the growing season. In this study, however, the relative allocation of biomass to tubers was calculated on a daily basis. Relative allocation to the tubers is initially 0%. After a thermal time threshold is reached, relative allocation starts increasing linearly with thermal time until the next threshold after which 100% of new biomass goes to the tubers. The values of these parameters were estimated so that harvest index of a mature crop is 80% under normal circumstances (no frost or heat stress), as in the original model. This procedure avoids overestimating yield for a crop with prematurely killed foliage due to frost, or for a crop for which the end of the growing season is very warm, and hence has a lower harvest index than would be expected from the average temperature during the growing season. The absolute allocation of biomass to the tubers is also dependent on daily average temperature: it decreases above 15 C and becomes zero at an average temperature above 28 C (Stol et al. 1991). A heat-tolerant potato cultivar was defined by shifting this curve two degrees.