Spatiotemporal Analysis of Reference Evapotranspiration in Arid, Semiarid, Mediterranean and Very Humid Climates Considering Developed Models and Lysimeter Measurements

Abstract

This study aims to present a spatial and temporal analysis of the reference evapotranspiration (ETo) based on the climate change alarms and collected lysimeter data. To this end, 50 years (1961–2010) meteorological data of 18 regions in Iran with various climates (arid, semiarid, Mediterranean and very humid) as well as recorded lysimeter measurements in these regions based on other researchers’ works, were collected. For estimating the ETo, 5 temperature-based, 5 radiation-based and 5 mass transfer–based models were selected with respect to better performance of them in different climates on the basis of past investigations and the results were compared with the Food and Agriculture Organization of the United Nations (FAO) Penman-Monteith (FPM). In addition, two new models were developed by replacing temperature difference with saturated vapour pressure deficit in Hargreaves-Samani (HS) equation named modified Hargreaves-Samani 3 (MHS3) and by replacing mean temperature with minimum temperature in Blaney-Criddle (BC) equation only for arid and semiarid regions named modified Blaney-Criddle (MBC). The results indicate the BC (Abtew (Ab)) is the superior model to estimate the ETo in arid (semiarid) region. While, modified Hargreaves-Samani 2 (MHS2) represents the best performance in Mediterranean and very humid regions. In addition, a spatiotemporal analysis of the best models for each region indicates that in some cases, error of the estimation has increased in recent 25 years which alarms a climate change for the region. As a result, all of the models present their best performance in the condition for which the variations of the meteorological parameters are less than 20% (with the exception of Jensen-Haise (JH), Makkink (Mk) and World Meteorological Organization (WMO) only for minimum temperature and wind speed) during 1961–1985 to 1986–2010. Meanwhile, the developed model (MBC) could estimate the ETo more accurate (RMSE = 0.69 mm day−1 and 0.77 mm day−1, respectively) in arid and semiarid regions. The lysimeter measurements confirm the reliability of the results for arid, semiarid and Mediterranean climates with a good R2 > 0.9000 (average) and with a fairly R2 = 0.7804 for very humid regions.