中文摘要：

作物系数Kc的合理取值对作物需水量分析计算精度十分关键。以水稻为例,基于江西省灌溉试验中心站及浙江永康灌溉试验站的实测数据,分析了Kc与降雨量及ET0的关系,以及灌溉模式对Kc的影响,以江西及云南省为例,探讨了Kc在省域尺度上的空间变化规律。结果表明,Kc年内变化呈现先增大后减小的趋势,在抽穗开花期达到最大;Kc年际变化存在差异,早晚稻Kc与ET0呈负相关,晚稻Kc与降雨呈弱正相关,但关系均不明显,采用连续3年Kc实测平均值代替各年Kc满足精度要求;节水型灌溉模式下作物系数明显小于淹灌模式;采用FAO修订办法计算了江西省19个站点的Kc并进行了比较,Kcini均为1.05,Kcmid和Kcend的标准差与变异系数均在0.01左右,Kcmid、Kcend最大值与最小值相差3%左右,而云南36个站点的比较表明,Kcini标准差与变异系数在0.03~0.04,Kcmid,Kcend标准差和变异系数在0.02左右,Kcini最大值与最小值相差10%,Kcmid、Kcend最大值与最小值相差在5%左右。因此,对于省域尺度,如果省内气象要素变异不大,某个站点的作物系数可以扩展到整个省内使用,否则需进行修订。

英文摘要：

It is very important for the accuracy of crop water requirement calculation to make reasonable value of Crop coefficient （Kc）. Taking rice for example, the K, was calculated based on the observed data from the Jiangxi irrigation experiment station and the Yongkang irrigation experiment station, and the correlation of Kc to rainfall and reference evapotranspiration （ETo）was ana- lyzed. Meanwhile, the effect of the irrigation mode on the K, was analyzed. Furthermore, the spatial variation characteristics of Kc on a provincial scale were discussed based on the calculated values of that in Jiangxi and Yunnan provinces. The results show that, in a year, the Kc increases first to an maximum in the heading and flowering stage, then it decreases; there are differences among the annual values of K,, and the value of that is negatively correlated with ETo for both early and late rice, whereas the K, of late rice is weakly positively correlated with rainfall. Moreover, we found that the average value of Kc calculated with observed data in 3 years instead of annual value can meet the requirements of accuracy. Additionally, the value of Kc in water--saving irrigation mode is sig- nificantly smaller than that in flooding irrigation mode. The values of K, in 19 stations in Jiangxi province was revised according to the method proposed by FAO, and the results indicate that the Kcini of 19 stations are all 1.05, and the standard deviation and the coefficient of variation of Kcmid and Kcend are about 0. 01, meanwhile, the percentage of maximum and minimum of Kcmid and Kcend is about 30/00. Whereas, the result of the 36 stations in Yunnan shows that： the standard deviation and the coefficient of variation of Kcini varies from 0.03 to 0. 04, and standard deviation and variation coefficient of Kcmid and Kcend are about 0. 02. In addition, the percentage difference of maximum and minimum of Kcini is 10%, and the percentage difference of maximum and minimum of Kcmid and Kcend is about 50//00. Therefore, on a provincial scale,