Abstract: Accurate estimation of reference crop evapotranscpiration(ET0) is critical for computation of crop water requirement. The currently existed numerous ET0 methods, while bring convenience for computation of crop water requirement, also cause confusion in method choice. Knowledge on validity of these methods in different regions is thus the basis for reasonable choice. Using data from six sites in main agricultural areas in China, four commonly used combination methods were assessed with FAO 56 Penman-Monteith (PM) as standard. The results showed that the Pen63 (1963 Penman)，FAO 79 (FAO 1979 Penman) and Kpen (1996 Kimberly Penman) overestimated the daily value of PM, while the FAO 24 (FAO 24 Penman) underestimated it. The average deviation was respectively 0.28, 0.52, 0.14 and -0.17mm×d-1, corresponding to relative deviation of 16.0%, 25.2%, 2.4%, -5.3% and relative root mean squire error of 12.1%, 22.4%, 14.2% and 13.5%. Monthly totals of the Pen63 and FAO 79 were significantly higher than that of the PM, the largest being respectively 12.5mm (10.8%) and 28.2mm (22.6%) occurred in May. Monthly values of the FAO 24 were lower than that of the PM, the largest being 11.4mm (8.1%), but they showed insignificant difference in most months for the southern sites. The Kpen overestimated the PM, the largest being 19.7mm (14.5%), during May through October and underestimated it in other months, and they showed insignificant difference in six months for the southern sites. On basis of yearly total, both the Pen63 and the FAO 79 significantly overestimated the PM by 103.8mm (11.8%) and 191.5mm (21.3%), respectively. Though the FAO 24 significantly underestimated the PM by 60.9mm (6.3%) and the Kpen overestimated it by 50.5mm (5.8%) in general, they showed insignificant difference from the PM at Shapingba and Beijing. Time scale affected the evaluation results. Based on daily and yearly comparison the performance order was respectively Pen63＞FAO 24＞Kpen＞FAO 79 and Kpen＞FAO 24＞ Pen63＞FAO 79. In addition, all methods performed better in humid climates at daily scale, but only FAO 79 and FAO 24 did so at yearly scale. Validity of the four combination methods varied, and Pen63 was the best and the FAO 79 the poorest, suggesting the importance to evaluate the latter before use.

Key words: Reference crop evapotranspiration (ET0), Combination method, FAO 79 penman, Penman-Monteith, FAO 24, Kimberly Penman, Main agricultural area