Improving Land-Surface Model Simulations in Irrigated Areas by Incorporating Soil Moisture–Based Irrigation Estimates in Community Land Model

Currently, land surface models (LSMs) are limited in representing realistic water and energy fluxes owing to the absence of reliable parameterization of irrigation. In this study, a novel method was employed to incorporate irrigation in the Community Land Model (CLM) Version 4.0. Two CLM experiments were set up, designated CLM-default run and CLM-irrigated run. The SM2RAIN algorithm was employed to reproduce the observed precipitation and irrigation using soil moisture (SM) information measured at the Fluxnet sites. The results showed that SM2RAIN reliably reproduced the observed precipitation on a daily timescale ($R\sim 0.70$ for all three sites) but significantly underestimated high-intensity precipitation (bias $\sim 0.5\mathrm{mm}{\mathrm{day}}^{-1}$ for all sites). The bias-corrected SM2RAIN output showed improved representation of observed daily precipitation ($R=0.89$ and 0.86) and monthly irrigation ($R=0.89$ and 0.96) at US-Ne1 and US-Ne2, respectively. The SM2RAIN estimated irrigation was input to CLM as independent forcing data along with other atmospheric forcings. The simulated surface energy fluxes from CLM were compared with eddy covariance–based flux tower observations. The results showed that CLM simulated energy fluxes from the CLM-irrigated run improved the representation of turbulent heat fluxes (latent and sensible). Overall, mean bias decreased by 32% and 64% for sensible and latent heat fluxes, respectively. This indicates that SM2RAIN-estimated irrigation is reliable input data for LSMs that potentially improved model representations of surface energy fluxes, which are important for comprehending the complex interactions between land surface and atmosphere in irrigated areas.

A novel method of estimating actual irrigation amount is presented for incorporating into land surface models (LSMs). The primary goal of the study was more accurate simulation of land surface states and fluxes by better representing agricultural land use. Moreover, this technique may allow numerical weather prediction (NWP) models to more precisely represent land–atmosphere feedback in managed areas, hence improving forecast skill. It was demonstrated in this study that using the novel irrigation method in LSMs can reliably represent surface water and energy fluxes. The findings of this study are very relevant to regional-to-global–scale water and energy cycle research, which has struggled to quantify the consequences of agricultural management practices like irrigation in the past. It is anticipated that improved representation of managed lands will make it possible to provide better weather and climate forecasting when the new irrigation plan is incorporated in NWP models.