Assessment Of ETo Estimation Methods (FAO-PM Temperature AndHargreaves-Samani Methods) In Different Climates Of Inner Mongolia

The FAO Penman-Monteith (PM-ETo) method is often recommended for reference evapotranspiration (ETo) computation across the world. However, in spite of the air temperature, it also requires solar duration, relative humidity and wind speed which are always not available or of questionable quality. Thus, other ETo calculation approaches such as Hargreaves-Samani (HS) and PM temperature (PMT) methods requiring only temperature data have emerged. Unfortunately, contradictory results were obtained when these two temperature based equations were applied in different climates and world regions.Firstly, the UNEP aridity index was applied to the regionalization of climatic zones in Inner Mongolia. Herein, daily averages of the PM-ETo demanded full datasets, covering the period of1981to2012in50surface meteorological stations over Inner Mongolia were utilized for validation of HS and PMT methods with selections of the best fitted kRs against the reference PM-ETo. In PMT estimation, a correction on temperature which is used to estimate Tdew, as well as another set of kRs instead of the correction was conducted with an objective to see whether the adjustments were needed in various climatic sub-regions.It is detected that all the three sets of kRs follow a decreasing order from hyper-arid to moist sub-humid climates, the temperature correction involved PMT-kRs are similar to the HS ones except for the stations with high mean annual wind speed and extreme ones. But the no adjustments contained PMT-kRs are relatively higher than the HS ones in hyper-arid and arid areas, while been close to the HS ones in semi-arid and dry sub-humid climates and, smaller than HS-kRs in moist sub-humid climates. Results for the performances of HS and PMT methods indicate:(1) PMT is preferable in hyper-arid zones with the precondition of Tmin correction,(2) in arid climates, both HS and PMT methods are suitable for ETo estimation but, HS is highly recommended since it is easier to compute and requires no correction on Tmin,(3) for semi-arid climates, PMT is more advantageous than HS and the temperature adjustments in PMT are not necessary,(4) in dry sub-humid areas, PMT provides the highest accuracy on ETo estimates and no temperature corrections are needed,(5) in moist sub-humid regions, PMT is more acceptable than HS and the adjustments on temperature are indispensable.Additionally, the role of winds in computing ETo was examined since weak correlations were observed at windy stations. Results illustrated that the extreme winds, as well as the temporal and spatial variability of this parameter are responsible for the detected disagreements between PMT and PM-ETo estimates. And the deviations between HS and PM-ETo estimates are possibly due to not considering the wind term in HS equation.The results for PCA applying to PM-ETo, HS-ETo and PMT-ETo variables reveal that:(1) the three methods rotated PC1depicts a similar decreasing order from west and northwest to east and southeast regions, the rotated PC2increased from north and northeast to south and southeast of the study area,(2) the spatial patterns of these three approaches associated mean annual and seasonal ETo are proved to be consistent with each other. Thus, both HS and PMT methods are sufficient for ETo computation with limited data in all the climates of Inner Mongolia.