A daily weather generator for use in climate change studies

This paper describes the development of a weather generator for usein climate impact assessments of agricultural and water systemmanagement. The generator produces internally consistent series ofmeteorological variables including: rainfall, temperature,humidity, wind, sunshine, as well as derivation of potentialevapotranspiration. The system produces series at a daily timeresolution, using two stochastic models in series: first, forrainfall which produces an output series which is then used for asecond model generating the other variables dependent on rainfall.The series are intended for single sites defined nationally acrossthe UK at a 5km resolution, but can be generated to berepresentative across small catchments (2).Scenarios can be generated for the control period (1961-1990) basedon observed data, as well as for the UK Climate Impacts Programme(UKCIP02) scenarios for three time slices (2020s, 2050s and 2080s).Future scenarios are generated by fitting the models toobservations which have been perturbed by application of changefactors derived from the UKCIP02 mean projected changes in thatvariable. These change factors are readily updated, as newscenarios become available, and with suitable calibration data theapproach could be extended to any geographical region.