Comparison of two potato simulation models under climate change. I. Model calibration and sensitivity analyses

To analyse the effects of climate change on potato growth and production, both a simple growth model, POTATOS, and a comprehensive model, NPOTATO, were applied. Both models were calibrated and tested against results from experiments and variety trials in The Netherlands. The sensitivity of model results to different values of weather variables was determined. The comparison of the results from both models indicated the sort of conditions in which model results differed and may become less reliable. The average tuber yield level and the inter-annual yield variation in potato experiments were predicted well by NPOTATO, whereas POTATOS sometimes calculated yields that were too high. The fit between yields observed in variety trials on clay soils and simulated yields from both models was quite good over the last 4 yr of the period 1974–1988. However, in almost all earlier years a considerably lower yield occurred in the trials than was calculated. This yield difference might be caused by factors that were not described by the models (e.g. a change in management). Irrigated tuber yield from both models considerably increased with increases in both solar radiation and atmospheric CO2, and it had its optimum at the present temperatures in Wageningen, The Netherlands. Water-limited yield from both models had a slightly lower temperature optimum, considerably increased with increasing precipitation, atmospheric CO2 and vapour pressure and decreased with increasing wind speed. The main differences between NPOTATO and POTATOS results were the higher evapo-transpiration and, hence, the stronger yield reduction by water limitation from NPOTATO, and with irrigation, the lower yields for present conditions and the weaker and stronger yield increases with increasing radiation and atmospheric CO2, respectively, from NPOTATO compared with those from POTATOS