CLIGEN天气发生器在中国东北三省模拟温度的适用性评价

CLIGEN（Climate Generator）是美国农业部针对本土开发用于提供长序列的逐日气象数据的模型，是否适用于中国东北三省并没有进行过评估。本文利用1981-2015年东北三省15个气象站逐日气象数据评估CLIGEN模拟的温度数据，对东北三省72个站点实测值和模拟值的极端最高温度、极端最低温度、≥ 35℃高温日数、≤ 0℃低温日数进行评估。结果表明：CLIGEN对日平均最高温度、最低温度模拟效果较好，平均绝对百分比误差MAPE（mean absolute percent error）和有效系数（E）中，日最高温度的MAPE为0.33%-14.74%，日最低温度的MAPE大多小于10%；有效系数E的值也非常接近，为0.9880-0.9999；CLIGEN对日较差均值的模拟效果较好，对日较差标准差的模拟效果一般，不能较好地模拟温度的连续性和渐变性；CLIGEN对极端温度和日数的评估只有低于0℃的日数模拟效果比较理想，从差值空间分布看来，吉林和辽宁西部地区是差值较大和较多的地区。

Applicability evaluation of the GLIGEN weather generator on the simulation of temperature in three provinces of Northeast China

The CLIGEN (Climate Generator) is a model developed by the U.S.Department of Agriculture to provide daily meteorology data with long sequences.Whether it is applicable to the three provinces i.e.Heilongjiang,Jilin,and Liaoning of Northeast China has not been evaluated.In this paper,the temperature simulated by the CLIGEN was evaluated by daily meteorological data from 15 weather stations in Northeast China from 1981 to 2015.The extreme maximum and minimum temperatures,the numbers of days with temperature ≥ 35℃ and temperature ≤ 0℃ were evaluated at 72 sites in Northeast China.The results show that the CLIGEN performs very well in the simulations of daily average maximum and minimum temperatures,and the MAPE (mean absolute percent error) and effective coefficient (E) of 12 months range from 0.33% to 14.74%.The MAPE of the minimum temperature is basically within 10%.The effective coefficient E value is also very close ranging from 0.9880 to 0.9999.The CLIGEN has a good simulation effect on the mean value of daily temperature range,but not on the standard deviation of the daily temperature range.The features of continuity and gradual change in temperature fail to be simulated by the CLIGEN.In the aspects of simulation of extreme temperature and the number of days,only the number of days with the temperatures lower than 0℃ is simulated well.From the spatial distribution of temperature difference,it shows that the larger and more values are distributed in western parts of Jilin and Liaoning provinces.This paper provides a preliminary study to evaluate the application of the CLIGEN on the simulation of temperature over three provinces in Northeast China to meet relevant research needs.