Multi-mission satellite remote sensing data for improving land hydrological models via data assimilation

Area of Science:

• Earth and hydrological sciences
• Satellite remote sensing applications
• Land surface modeling

Background:

• Hydrological model predictions are limited by erroneous input data, uncertain forcings, and parameter uncertainties.
• Satellite remote sensing provides abundant datasets crucial for improving model estimates and parameters.
• Multi-mission satellite data offers opportunities to enhance land surface model accuracy.

Purpose of the Study:

• To improve hydrological model predictions using multi-mission satellite remote sensing data.
• To estimate model parameters through an advanced data assimilation technique.
• To evaluate the performance of multivariate data assimilation in improving model accuracy.

Main Methods:

• Utilized soil moisture (SMOS, AMSR-E), terrestrial water storage (GRACE), and leaf area index (AVHRR) datasets.
• Employed the unsupervised weak constrained ensemble Kalman filter (UWCEnKF) for parameter estimation.
• Applied a dual scheme with interactive EnKF filters and water balance constraint enforcement.

Main Results:

• Simultaneous assimilation of multiple satellite products with parameter estimation significantly improved model predictions.
• Demonstrated substantial reductions in groundwater RMSE (32% during estimation, 14% during forecast).
• Showcased increases in soil moisture correlation (from 0.66 to 0.85 during estimation, 0.69 to 0.78 during forecast).

Conclusions:

• Multivariate data assimilation combined with parameter estimation is highly effective for enhancing hydrological models.
• The UWCEnKF method successfully improved both model states and parameters, leading to better predictive accuracy.
• Findings highlight the value of integrating diverse satellite observations for robust land surface modeling and water resource management.