Abstract:
With the increasing pressure of agriculture on water and land, achieving high water
productivity is essential. Models such as FAO’s Water Productivity Portal (WaPOR)
aim to estimate water productivity by providing yield and evapotranspiration. In this
study, the aim was to validate the yield product of WaPOR in the Beqaa Valley,
Lebanon. The study was focused in two main fields. Yield of Potato and wheat, planted
during 2017-2018, was validated in one field. In addition, the yields of barley, vetch,
barley/vetch mixed fields and vetch/oat mixed fields, planted during 2012-2019,were
validated in the other field by comparison against farmer reported yields. Statistical
Indicators such as percentage relative error (RE), Root Mean Square Error (RMSE),R2
,
correlation (r), and bias were used for this validation.
Wheat yield showed better results at a resolution of 30 m than that at a 100 m where
recorded RE% of 1.14% < 20%, R2 of 0.38, RMSE of 0.71 ton/ha, r of 0.61 and a bias
of 0.68 ton/ha versus an RE%|-12.43%| <20%, R2 of 0.38, RMSE of 1.2 ton/ha, r of
0.62, and a bias of 0.73 ton/ha respectively. For the other crops, level consistency could
not be tested for since not all crops where equally identified at the different levels.
Potato yield was considerably accurate at a 100m resolution with RE% of
19.55%<20%, an R2 of 0.22 and RMSE of 9.31 ton/ha, r of 0.47, and a bias of 2.39
ton/ha. As for barley,vetch,barley/vetch and vetch/oat mixed fields results were
considered to be poor with a RE% > 20% for all crops at both levels 2 and 3. This
inaccuracy in crop yield estimations was attributed to inaccuracy in farmer reported
yields which was detected in both fields, the standardized LUE max, harvest Index and
moisture content.
