MET&ROLL: THE WEATHER GENERATOR FOR CROP GROWTH MODELLING

Wilks, 1992] four-variate stochastic weather generator [Dubrovsky, 1997]. It is designed to provide synthetic daily weather series representing present and changed climate conditions to be used in crop growth modelling [Zalud et al., 1999; Dubrovsky et al., 1999]. Precipitation (RAIN) is modelled by a first-order two-state Markov chain (occurrence) and Gamma distribution (amount). Standardised deviations of daily extreme temperatures (TMAX and TMIN) and daily sums of global solar radiation (SRAD) are modelled by a first-order trivariate autoregressive (AR1) model. The set of parameters of the generator includes: · 2 parameters of the Markov chain model: probability of wet day occurrence, ð1, and probability that a wet day follows a dry day, ð01 (parameters are defined for each month of the year) · 2 parameters of the Gamma distribution (defined for months) · parameters of the AR1 model:- means and standard deviations of SRAD, TMAX and TMIN (defined for each day of the year, separately for wet and dry days)- two 3́3 matrices (defined for week) Met&Roll may be operated either in an interactive mode or in a command-line mode. The latter approach allows to chain several procedures into a single batch which makes the data processin