4种人工智能模型在江西省参考作物蒸散量计算中的适用性

为了实现气象资料缺失下参考作物蒸散量ET₀的高精度预测,以江西南昌、吉安及龙南站1966-2015年每日最高气温T_max、最低气温T_min、日照时数n、相对湿度RH和2 m高风速u₂作为输入参数,以FAO-56 Penman-Monteith(P-M)公式的计算结果作为对照,建立了6种不同气象要素组合条件下的4种ET₀计算模型,并分别与输入相同数据的经验法计算结果进行了比较.结果表明,在3个站点中,多元自适应回归样条法MARS模型的精度最高,且计算简便,可作为江西省蒸散量模拟的推荐方法.当4种模型的输入数据完整时,模拟精度均达到最高,表明4种模型均可适用于对参考作物蒸散量的模拟;输入数据缺失条件下,各气象要素对智能模型模拟ET₀的影响由大到小按参数排序依次为T_max,T_min,n,RH,u₂.与传统经验公式相比,4种智能模型的ET₀计算结果精度均优于输入相同数据的经验法.

Applicability of four kinds of artificial intelligent models to prediction of reference crop evapotranspiration in Jiangxi province

A highly precise estimate of reference crop evapotranspiration(ET₀)in absence of some meteorological data is on demand. Based on daily maximum and minimum ambient temperatures T_max and T_min, sunshine hours n, relative humidity, RH, and wind speed at 2 m height, u_2, during 1966—2015 in Nanchang, Ji′an and Longnan meteorological stations in Jiangxi province, four artificial intelligent(AI)models for predicting ET₀ are established in terms of different combinations of six meteorological elements by using FAO-56 Penman-Monteith(P-M)formula as standard. The predicted results are compared with those calculated by empirical method. The results show that the MARS model has the highest accuracy in three stations but also its computation procedure is simple; eventually, it is the recommended method for estimating ET₀ in the province. If the input data are complete, four mo-dels can achieve the best accuracy, indicating all the models are applicable to ET₀ prediction. In absence of some input data, the influence of meteorological elements on ET₀ estimation from the most important to the least important is as follows: T_max>T_min>n>RH>u₂. Compared with the traditional empirical formulas, the accuracy of four AI models is better for the same input data.