A SERIAL APPROACH TO LOCAL STOCHASTIC WEATHER MODELS

被引:338
作者
RACSKO, P [1 ]
SZEIDL, L [1 ]
SEMENOV, M [1 ]
机构
[1] ACAD SCI USSR, INST ATMOSPHER PHYS, MATH ECOL LAB, MOSCOW 109017, USSR
关键词
D O I
10.1016/0304-3800(91)90053-4
中图分类号
Q14 [生态学(生物生态学)];
学科分类号
071012 ; 0713 ;
摘要
Crop growth simulation models have been studied and constructed by many authors, including the authors of this paper. It was realized that the available local time series of the weather parameters are not numerous or long enough for a good statistical identification of the model's parameters, and they are too short if one aims at generating the multidimensional probability distribution of the output parameters for the stochastic input variables. Thus it was decided to construct a stochastic weather 'generator' that provides as long a time series as is necessary and as many repetitions as the simulation experiments require. The weather 'generator' must, of course, be statistically identical to the observed time series. The paper gives a description of the weather generator developed by the authors. The generator contains a stochastic weather model based on a new approach to weather data analysis and also a computer program package that carries out the tuning of the model on real time series and generates the random weather processes. The model was identified and applied to two geographical locations in Hungary.
引用
收藏
页码:27 / 41
页数:15
相关论文
共 3 条
[1]  
Feller W., 1970, INTRO PROBABILITY TH
[2]   ANALYSIS OF MATHEMATICAL PRINCIPLES IN CROP GROWTH SIMULATION-MODELS [J].
RACSKO, P ;
SEMENOV, M .
ECOLOGICAL MODELLING, 1989, 47 (3-4) :291-302
[3]   STOCHASTIC SIMULATION OF DAILY PRECIPITATION, TEMPERATURE, AND SOLAR-RADIATION [J].
RICHARDSON, CW .
WATER RESOURCES RESEARCH, 1981, 17 (01) :182-190