Interannual variability characteristics and simulated crop response of four stochastic weather generators

Many stochastic weather generators commonly used with crop models tend to u nder predict, interannual variability of climate and, as a result, distort distributions of crop simulation results. We examine the ability of four st ochastic weather generators, WeatherMan, MARKSIM, WM2 and LARS-WG, to repro duce interannual. variability of monthly climate and crop simulation result s. Comparisons were based on bias and RMSE of means and standard deviations of monthly precipitation totals, frequencies of wet days, mean daily tempe rature extremes for 12 long-term weather data sets; and yields and dates of anthesis and harvest maturity of three crops under a total of 10 scenarios simulated with dynamic crop models. Evaluation also considered statistical tests of equality of distributions of the same variables between observed and generated weather data sets. The generators generally reproduced climat ic means well, except for MARKSIM which showed positive bias of mean monthl y rainfall totals and frequencies. WM2 and LARS-WG showed substantial negat ive bias of interannual variability of monthly precipitation totals. As a r esult of stochastic resampling of wet-day frequencies, MARKSIM and WM2 show ed little negative variability bias of monthly precipitation totals, but po sitive variability bias of monthly wet-day frequencies. WeatherMan, MARKSIM and LARS-WG under represented variability of monthly mean temperatures, wh ereas WM2 showed positive but smaller mean temperature variability bias. WM 2 reproduced simulated distributions of yield and harvest maturity dates be tter than the other generators. Results for simulated anthesis dates were n ot consistent among generators. This study supports the need for some form of low-frequency variability correction in stochastic weather generators fo r applications in which reproducing interannual variability is important. R esults generally favor WM2 over the other generators tested for application s in which variability of simulated yields is of primary importance. (C) 20 01 Elsevier Science B.V. All rights reserved.