Development of high-resolution UKCIP02-based climate change scenarios in the UK

Analysis of possible impacts of climate change on agriculture, based on process-based simulation models, which use daily weather as their input, requires climate change scenarios with high spatial and temporal resolutions. Despite improvements in the performance of global and regional climate models, direct daily outputs from them are not suitable for such analysis. A methodology for construction of daily site-specific climate scenarios, based on a stochastic weather generator, is described. Initially the LARS-WG stochastic weather generator, used in our study, was calibrated for current climate with observed daily data. Then its parameters were adjusted for climate change, using the output from UKCIP02 projections, presented as changes in monthly mean climatic variables between the control run and future scenarios. To be able to generate scenarios at any given location in the UK, parameters of LARS-WG, computed for locations with long historical weather records, were interpolated over the UK. Distributions for climatic variables were interpolated locally and then modified by globally interpolated mean values to account for the effect of topography. As illustrations, daily UKCIP02-based scenarios were generated and used to calculate various weather extreme events and impact of climate change on wheat growth. Under a warmer climate, extreme statistics related to temperature, such as heat-waves, are likely to increase substantially in magnitude and frequency. Two impact statistics for wheat, i.e. drought stress index and probability of an episode of hot temperature after anthesis, were analysed. Despite higher temperature and lower summer precipitation for the 2080HI scenario, the relative impact on yield due to drought stress is smaller for 2080HI than for the baseline climate, because of the ability of wheat to mature early in a warmer climate avoiding summer heat and drought stress.