CMWSim: Development and evaluation of probability-based weather generating software for crop growth simulation

Abstract

Daily weather records often play an important role in applications of crop growth simulation models. Weather generating software is essentially important when evaluating long term time series to generate daily weather records. CMWSim is stochastic weather generating software we developed as a sub model of the MAIZESim maize growth simulation model and it generates daily precipitation, maximum and minimum temperature. The design of the CMWSim was based on a probabilistic simulation approach which adapted from the Markov chain analysis. CMWSim simulates (I) the precipitation occurrence by using transitional probability matrices of first-order two-state Markov chain, (II) the precipitation amounts by using a two-parameter gamma distribution and (III) the temperature values (maximum and minimum) by using a first-order auto-regressive model with a conditional scheme. The system is implemented as a user-friendly Windows-based software program. Simulated data are statistically analyzed and compared with observed data to evaluate the system. The results show that all comparisons are significant and precipitation patterns are well fitted with observed patterns. Simulated precipitation was slightly over-estimated at higher intensity periods (July - September), but was statistically acceptable. Maximum and minimum temperatures are also well simulated by the model and there were no misleading results. CMWSim generates weather records in acceptable significance to be used as inputs of crop growth modelling. Since this development completely used the object oriented programming concept (reusable package), it could also be able to use (integrate) with any other decision support tool which uses weather records.