Improved evapotranspiration modeling by incorporating seasonal irrigation into a land surface model

Irrigation is very important for maintaining the agricultural production and sustaining the increasing population of India. The irrigation requirement can be estimated with land surface models. The irrigation modeling tends to be affected by various uncertainties related to the spatiotemporal uncertainty of the irrigation map, irrigation frequency and the irrigation factor. The irrigation requirement and hydrological fluxes over India were reconstructed by six simulation experiments with the Community Land Model (CLM) from the year of 2010 to 2020. The multiple simulation scenarios showed that using a season-specific irrigation map resulted in a higher transpiration-total evapotranspiration ratio (T/ET) in the pre-monsoon season compared to other irrigation strategies, implying higher irrigation efficiency. The remote sensing based evapotranspiration products were used to compare with simulated model results, showing a similar increasing ET-trend in the pre-monsoon season as the irrigation induced CLM. More accurate temporal information of irrigation will result in modeled evapotranspiration closer to the spatiotemporal pattern of evapotranspiration observed by remote sensing. In addition, a higher spatial resolution or improved sub-grid heterogeneity in land surface modeling could improve the estimation of soil water deficit and irrigation requirement.