Multiscale Assimilation of Sentinel and Landsat Data for Soil Moisture and Leaf Area Index Predictions Using an Ensemble-Kalman-Filter-Based Assimilation Approach in a Heterogeneous Ecosystem

Abstract

Data assimilation techniques allow researchers to optimally merge remote sensing observations in ecohydrological models, guiding them for improving land surface fluxes predictions. Presently, freely available remote sensing products, such as those of Sentinel 1 radar, Landsat 8 sensors, and Sentinel 2 sensors, allow the monitoring of land surface variables (e.g., radar backscatter for soil moisture and the normalized difference vegetation index (NDVI) and for leaf area index (LAI)) at unprecedentedly high spatial and time resolutions, appropriate for heterogeneous ecosystems, typical of semiarid ecosystems characterized by contrasting vegetation components (grass and trees) competing for water use. A multiscale assimilation approach that assimilates radar backscatter and grass and tree NDVI in a coupled vegetation dynamic–land surface model is proposed. It is based on the ensemble Kalman filter (EnKF), and it is not limited to assimilating remote sensing data for model predictions, but it uses assimilated data for dynamically updating key model parameters (the ENKFdc approach), including saturated hydraulic conductivity and grass and tree maintenance respiration coefficients, which are highly sensitive parameters of soil–water balance and biomass budget models, respectively. The proposed EnKFdc assimilation approach facilitated good predictions of soil moisture, grass, and tree LAI in a heterogeneous ecosystem in Sardinia for a 3-year period with contrasting hydrometeorological (dry vs. wet) conditions. Contrary to the EnKF-based approach, the proposed EnKFdc approach performed well for the full range of hydrometeorological conditions and parameters, even assuming extremely biased model conditions with very high or low parameter values compared with the calibrated (“true”) values. The EnKFdc approach is crucial for soil moisture and LAI predictions in winter and spring, key seasons for water resources management in Mediterranean water-limited ecosystems. The use of ENKFdc also enabled us to predict evapotranspiration and carbon flux well, with errors of less than 4% and 15%, respectively; such results were obtained even with extremely biased initial model conditions.

A Soil Moisture Prediction Model, Based on Depth and Water Balance Equation: A Case Study of the Xilingol League Grassland

“…Compared to predicting the soil moisture using methods, such as machine learning [52][53][54], the model can explain the variables and the model principles with a theoretical significance. Compared to the monitoring and prediction of the soil moisture using remote sensing data [55][56][57], the model is not only able to predict the soil moisture at the surface, but also at different depths, based on the inclusion of the depth coefficients. The experimental results showed that the model proposed in this study was able to provide a higher prediction accuracy for the soil moisture prediction at 40 cm, 100 cm and 200 cm, compared to the soil moisture prediction model, based on the seasonal ARIMA model.…”

Assimilation of Remotely Sensed Leaf Area Index for Improving Land Surface Simulation Performance at a Global Scale

“…For instance, replacing default LAI values with field observations or remote sensing products has been shown to improve simulations of momentum and trace gas exchanges [20][21][22][23]. Additionally, assimilating other variables, such as brightness temperature, radar backscatter, and NDVI for grass and trees, has also been effective in refining LAI [24,25]. Furthermore, the assimilation of canopy parameters, including LAI), and/or soil moisture (SM), has led to improved model performance in several areas.…”

“…Furthermore, the assimilation of canopy parameters, including LAI), and/or soil moisture (SM), has led to improved model performance in several areas. These areas include the simulation of the carbon cycle [26][27][28][29], evapotranspiration (ET) [13,30], hydrological processes [31][32][33], agriculture dynamics [34][35][36], ecosystem functions [24,37], and seasonal temperature predictions [38].…”