Impact of ET and biomass model choices on economic irrigation water productivity in water-scarce basins

Using Economic Irrigation Water Productivity (EIWP) as an indicator for on-farm irrigation decision-making is of utmost importance in addressing the challenges posed by poor policies in managing agricultural water use, particularly in water-scarce basins. Various modeling systems are available for quantifying crop actual evapotranspiration (ETa) and biomass needed for measuring the economic output obtained from each unit of irrigation water utilized. The difference in ET(a )and biomass estimates between the modeling systems could translate into a difference in EIWP outcomes, which influences the basin-wide irrigation water management. In this paper we examine the influence of selecting different ET(a )and biomass models, particularly the hybrid single-source energy balance HSEB, the Global Field-Scale Crop Yield and ET Mapper in Google Earth Engine GYMEE, and FAO's WaPOR V2, on evaluating EIWP on a basin-wide level. The method includes combining remote sensing and economic data to compare variability in ETa, biomass, and EIWP values derived from HSEB, GYMEE, and WaPOR. The approach is demonstrated with field survey data from the upper hydrologic unit of Lebanon's largest catchment, the Litani River Basin, in its three productive districts Baalbak, Zahleh, and West Bekaa for the year 2021. Field-scale mean monthly ET(a )and biomass estimates for all crops, obtained from both models, are very comparable. On a district level, the results reveal a reasonable model agreement for the four crops in the estimation of ET(a )with moderate to strong correlation (0.75 < r < 0.95). WaPOR consistently produces slightly higher mean ET(a )values for potato, wheat, and table grapes when compared to HSEB. Both models reasonably agree when estimating biomass for the four crops with high correlation (r > 0.9). Contrary to the ETa results, the GYMEE model consistently estimates slightly higher mean biomass values for all crops compared to WaPOR. The EIWP values produced by both models consistently indicate that potato holds the highest EIWP across all districts, followed by onion, table grapes, and wheat. The mean district HSEB-GYMEE model derived EIWPs are slightly higher than those derived from the WaPOR model for most crops. For EIWP obtained from HSEB-GYMEE, mean EIWP for potato is 12 times higher than that of wheat. As for that obtained from WaPOR, mean EIWP for potato is 10 times higher than that of wheat. The paper establishes a basis for future research on the application of remote sensing models in addressing water-stressed and socioeconomically challenged basins, with the potential to inform strategic irrigation management decisions based on model selection.