Evaluation of HadCM2 and direct use of daily GCM data in impact assessmentstudies

This paper investigates two important aspects of methods used to explore po ssible effects of climatic changes on agricultural productivity on regional spatial scales. First, an evaluation of precipitation and near surface air temperature in two successive versions of the Hadley Centre General Circul ation Model (GCM) has been performed to consider to what extent GCMs are ca pable of simulating the mean and variability of local climates. This is exp lored by comparing the output of an individual GCM grid box with three stat ion observations. Several ancillary issues associated with the comparisons of observations of daily precipitation and model output that affect the sta tistical results are also discussed. Finally, daily data from the control a nd sulphate runs of the latest Hadley Centre GCM (HadCM2) have been used di rectly as input to the CERES-Wheat model, and the modelled yield distributi on is compared to that produced with the historical data series. Our result s imply that for this particular grid box covering the study region in cent ral France, the daily raw data from HadCM2 experiment can be used directly to assess the potential impact of the greenhouse gas and sulphate aerosol r adiative induced forcings and the associated climatic change on average reg ional winter wheat production. On the other hand, less confidence should be placed on their use regarding the estimation of future agricultural risk a nd variability assessment. Furthermore, a possibly more severe methodologic al problem that has arisen from our study is the inability of CERES-Wheat t o simulate the waterlogging effects of excessive soil water on crop growth and development. Finally, we assess the potential impact of changing climat e on regional winter wheat production by using the daily data from the sulp hate integration up to the end of the 21st century.