Stochastic Simulation Theory,Validation and Improvement for SWAT 2000 Climate Simulator

张东,张万昌

DOI: https://doi.org/10.3321/j.issn:1007-7588.2004.04.005

2004-01-01

Abstract:The Objective of this article is to improve the simulation efficiency of the climate simulator which is used to simulate maximum, minimum temperature and solar radiation as the input climate driver forces for SWAT2000 distributed hydrological model. Firstly the stochastic theory of climate simulator was described in detail. And then, an analysis was made to the full serial weather data in 7 calibration stations located on Hexi corridor, Gansu, China, such as Jiuquan and Zhangye. It indicated that the correlation coefficients for observed monthly mean maximum, minimum temperatures versus those related mean temperatures were above (0.99) and their standard deviations versus those related ones were all greater than 0.89. Based on these conclusions, a methodology to calculate monthly mean maximum, minimum temperature and their corresponding standard deviations where similar weather conditions do exist was developed. The monthly statistical data could be obtained by adding its monthly mean air temperature to a certain increments and they had a fairy well precision. Using these data as the original climate simulator input, a calibration was made to the 7 calibrated stations. It showed there had some abnormal surges in the simulation results and even some logical errors in them such as the daily maximum temperature lower than daily minimum one. They were caused by the stochastic-noise-matrix to the climate simulator. By using daily mean air temperature to correct the logical errors in the simulated daily maximum and minimum temperatures and by calculating 5-day overlapping mean on them, the correlation coefficients increased from 0.85 to 0.90 and 0.89 to 0.94 respectively by the comparison between simulated results and observed ones. The precision of simulated weather data was fairy improved. And lastly, the temperature and solar radiation data for Qilian, Zhamashike and Yingluoxia weather stations were simulated by the modified climate simulator. Putting them into SWAT hydrological model as the input parameters, the coefficient for the simulated hydrological results with those observed ones reached 0.82, which meet the requirements of the data precision for the distributed hydrological model.