Chaudhary et al. (2025) A comprehensive water balance approach for improved assimilation of evapotranspiration estimates derived from soil moisture

Identification

• Journal: Irrigation Science
• Year: 2025
• Date: 2025-11-20
• Authors: D R Chaudhary, Ickkshaanshu Sonkar
• DOI: 10.1007/s00271-025-01048-3

Research Groups

Department of Civil Engineering, Indian Institute of Technology Ropar, Rupnagar, India.

Short Summary

This study develops and evaluates a Comprehensive Water Balance (CWB) model, coupled with an ensemble Kalman filter, to improve daily evapotranspiration (ET) estimation by explicitly accounting for ET-driven percolation, demonstrating superior performance over a Simple Water Balance (SWB) model, especially in coarse-textured soils.

Objective

• To develop and test a Comprehensive Water Balance (CWB) model that explicitly incorporates the influence of evapotranspiration (ET) on vertical flow (percolation) for improved daily ET estimation using synthetic soil moisture data.
• To compare the CWB model's performance with a conventional Simple Water Balance (SWB) model across different soil types and sensor configurations.

Study Configuration

• Spatial Scale: One-dimensional homogeneous soil column of 1.5 meters length, with a root depth of 1 meter.
• Temporal Scale: 50-day crop growth simulation period, with a temporal discretization of 15 minutes for the Richards' equation, and daily ET predictions.

Methodology and Data

• Models used:
  • Comprehensive Water Balance (CWB) model (developed in this study)
  • Simple Water Balance (SWB) model (for comparison)
  • Ensemble Kalman Filter (EnKF) for data assimilation
  • One-dimensional Richards' equation for root zone water dynamics
  • van Genuchten (1980) and Mualem (1976) constitutive relations for soil hydraulic properties.
• Data sources: Synthetic soil moisture data generated from a 50-day numerical crop growth experiment, incorporating white Gaussian noise to simulate sensor error. Soil hydraulic parameters were sourced from Carsel and Parrish (1988).

Main Results

• The CWB model performed better than the SWB model in evapotranspiration (ET) prediction, particularly in coarse-textured soils (Loam Sand, LS).
• In LS soil, the CWB model reduced the mean ET prediction error by 45% and achieved higher accuracy (Nash–Sutcliffe efficiency coefficient (NSE) = 0.918) compared to the SWB model (NSE = 0.727).
• In relatively low-conductance soil (Sandy Loam, SL), both CWB and SWB models performed similarly well (NSE ≈ 0.9).
• ET prediction using eight soil moisture sensors (SMS8 setup) showed high accuracy with a relative measure (RV) close to 1, NSE of 0.9, and a root-mean-square error (RMSE) of approximately 0.3 mm d⁻¹.
• ET prediction using five sensors (SMS5 setup) had slightly lower accuracy (RV > 0.75, NSE > 0.84, RMSE = 0.4 mm d⁻¹), but still provided reliable estimates under normal ET variations.
• Reducing the number of sensors from eight to four increased the RMSE (e.g., from 0.31 mm d⁻¹ to 0.61 mm d⁻¹ in LS soil).
• Fine-textured soils (Sandy Clay Loam, SCL) exhibited greater resilience to sensor observation error, enabling more reliable ET estimates.
• The Ensemble Kalman Filter (EnKF) model achieved convergence with an ensemble size of 150.

Contributions

• Introduction of a novel Comprehensive Water Balance (CWB) model that explicitly accounts for ET-driven percolation, addressing a limitation of previous Simple Water Balance (SWB) models that often neglect ET's influence on vertical fluxes.
• Demonstrates the necessity of incorporating vertical flux effects to avoid ET underestimation, especially in high-conductance soils where percolation remains significant.
• Proposes root water uptake (RWU) rather than soil moisture as the observable for updating within the Ensemble Kalman Filter (EnKF) framework for ET estimation.
• Provides valuable guidance on optimal soil moisture sensor placement strategies (proportional to RWU rate) and acceptable sensor error levels for different soil types to maintain reliable ET predictions.
• Highlights the greater resilience of fine-textured soils to observation noise due to their higher storage capacity, which helps maintain assimilation performance.

Funding

Not applicable (No external funding reported).

Citation

@article{Chaudhary2025comprehensive,
  author = {Chaudhary, D R and Sonkar, Ickkshaanshu},
  title = {A comprehensive water balance approach for improved assimilation of evapotranspiration estimates derived from soil moisture},
  journal = {Irrigation Science},
  year = {2025},
  doi = {10.1007/s00271-025-01048-3},
  url = {https://doi.org/10.1007/s00271-025-01048-3}
}