Evaluation of the FAO Proposed Models for Reference Crop Evapotranspiration (ETo) Estimation in Hajiabad Watershed of Hormozgan Province

Introduction and Objective
Due to droughts occurred in recent years, available water resources in southern basins and watersheds of Iran especially those are in Hormozgan Province has been drastically decreased which may not only change their ecosystem but also make serious problem for the country’s food security.  In addition to find new water resources, effective use of available water is the best choice to tackle water deficit which in turn, requires accurate estimation of evapotranspiration or water needs of plants. Crop evapotranspiration is usually calculated using the two-step FAO method. In this method, first the reference crop evapotranspiration (ET₀), representing effect of weather on crop evapotranspiration is estimated by models using weather variables and then, multiplied by crop coefficient (Kc) which reflects crop characteristics on evapotranspiration. A review on the research conducted around the world revealed that each model may have different accuracies and performances at different watersheds or locations. Subsequently, using these models without testing their accuracy and relevance at each watershed may increase error in estimating (ET₀) and hence crop water requirements which may complicate any planning for the watershed in the future. Therefore, this research was conducted to evaluate the accuracy and performance of the FAO proposed models at Haji Abad watershed of Hormozgan Province.
Materials and Methods
In this research which was conducted in Haji Abad watershed of Hormozgan for four years, reference crop(grass) evapotranspiration was measured weekly using a drainable lysimeter. In the first year, a large square drainable lysimeter (3 m wide ×3 m long × 1.9 m tall) was installed in the center of a 2-ha agricultural field which was representative of the watershed. In the next third years, grass (Cynodon dactylon L.) was grown inside the lysimeter and in the 900 m² surrounding field. At full cover and when grass height reached 8 cm, evapotranspiration from grass inside the lysimeter was measured weekly by the water balance method. Reference crop evapotranspiration was also estimated using the FAO models: Adjusted penman and Blaney- Criddle, Radiation, Evaporation pan and Penman-Monteith. The estimated (ET₀) values from the FAO models were compared with the data measured from lysimeter using linear regression, root mean square error and error percentage in estimating seasonal (ET₀) and subsequently, the most appropriate models were recommended for the watershed.
Results and Discussion
Seasonal grass evapotranspiration measured from the lysimeter was 2729.8 mm and estimated by Adjusted penman and Blaney- Criddle, Penman-Monteith, Radiation and Evaporation pan 3405, 2695.2, 2868, 2789.2 and 2054.6 mm, respectively. The FAO adjusted Penman, Penman-Monteith and Radiation models overestimated ETo by as much as 24.7, 5.1 and 2.2% respectively, whereas, the FAO evaporation pan and Blaney- Criddle underestimated it by as much as 24.7 and 1.3%, respectively. Comparing weekly (ET₀) data showed that adjusted penman model overestimated (ET₀) but the FAO evaporation pan model underestimated it for most of the periods. FAO radiation model had tendency to overestimate at ETo<7 mmday-1 (normally in Fall and Winter) but underestimate at (ET₀)>10 mmday-1 (normally in Summer).
Conclusion and Suggestions 
According to the results, the FAO Blaney-Criddle and Penman-Monteith models had respectively higher accordance and homogeneity with the real data measured from lysimeter and can predict (ET₀) with higher accuracy than the other tested models. Therefore, these models are recommended, respectively as the most appropriate models to estimate (ET₀) in Haji Abad Watershed and the areas having the same climate. Furthermore, the FAO Blaney-Criddle model needs fewer weather parameters and hence, may be used in the areas with limited climatic data.