四种参考作物蒸散量综合法在我国农业主产区的适用性

Abstract

参考作物蒸散量(ET0)的准确估算是作物需水量计算的关键。目前的许多ET0方法在为需水量估算带来便利的同时，也造成选择上的困惑。本文利用我国农业主产区六个站点的数据，以FAO 56 Penman- Monteith (PM)为标准，评价了国内外常用的四种综合法的适用性。结果表明，1963 Penman、FAO- ppp-17及1996 Kimberly penman都高估PM，在六站点平均高估0.28 mmd−1、0.52 mmd−1和0.14 mmd−1，相当于16.0%、25.2%和2.4%；而FAO 24 penman低估PM，六站点平均−0.17 mmd−1，相当于−5.3%。总体上，四种综合法与PM间的日值差异显著。依据RMSE(均方根误差)的吻合程度排序为：FAO 24 > 1963 Penman > 1996 Kimberly penman > FAO-ppp-17，但南、北方站点有差异：北方站点1963 Penman与PM最吻合，南方站点FAO 24与PM最吻合。此外，1963 Penman的普适性最好，FAO- ppp-17最差，故应用前对其评价更重要。四种综合法在北方站点表现较差，在南方站点较好，其中在沙坪坝适用性最好，在哈尔滨最差，表明Penman法及其衍生的多种综合法在湿润气候下适用性更好。逐月日ET0比较显示，绝对差值在夏季月份较大、冬季月份较小，但相对差值则是夏季月份较小、冬季月份较大。影响四种综合法与PM差异的关键因子为总辐射、净辐射和相对湿度。 Accurate estimation of reference crop evapotranscpiration (ET0) is critical for computation of crop water requirement. The currently numerous ET0 methods, which bring convenience for computing water requirement, cause confusion in method choice. Using data from six sites in main agricultural areas in China and with FAO 56 Penman-Monteith (PM) as reference, four commonly used combination methods were assessed. Results showed that the 1963 Penman, FAO-ppp-17 and 1996 Kimberly Penman overestimated the PM by 0.28, 0.52 and 0.14 mmd−1 (averaged over six sites), respectively, corresponding to 16.0%, 25.2% and 2.4%. In contrast, the FAO 24 underestimated the PM by 0.17 mmd−1 or −5.3%. Overall, daily ET0 of the four methods differed significantly from that of the PM, and their performance based on RMSE showed the order: FAO 24 > Pen63 > Kpen > FAO-ppp-17, which varied among southern and northern sites: the 1963 Penman was the best at northern sites and the FAO 24 was the best at southern sites. In addition, validity of the 1963 Penman was the best, and that of the FAO-ppp-17 was the poorest, implying the importance to evaluate the latter before use. As a whole, the four methods performed poorer at the northern sites than that at the southern ones, and they gave best performance at Shapingba and poorest at Haerbin, meaning that the Penman and its derivative versions are more applicable in humid climates. Daily comparison at each month indicated that mean bias error (MBE) was larger at summer months and smaller at winter months, but it was opposite for relative error (RMBE). Total solar radiation, net radiation and humidity are the most important factors influencing performance of the four methods.