Comment on egusphere-2023-47

Accurate root zone soil moisture (RZSM) estimation is essential for precision irrigation (PI) systems that seek to optimize water use efficiency. Large-scale in-situ sensors for direct measurement are costly, while existing satellites lack depth resolution for direct RZSM data. Hence, in-direct RZSM estimation methods are required. Literature illustrates that RZSM at a location is related to changing soil-water-plant characteristics. Therefore, these characteristics can provide auxiliary information on RZSM changes. By leveraging auxiliary information derived from changing soil-water-plant characteristics, this paper enables indirect RZSM estimation at non-sensor locations, effectively addressing the limitations inherent in direct RZSM measurement techniques initially discussed. Compared to existing methods, deep learning (DL) is most suitable for such data associations as they are auto-tuned to extract relative relationships from diverse big data. Among DL models, sequential models are apt for finding these relationships as all these variables are time-series sequences. The transformer neural network (TNN) is the state-of-the-art DL model for analyzing sequences. However, for in-direct RZSM estimation, the data associations within a location and multiple sensor sites with the target need to be found. Conventional TNN cannot incorporate such simultaneous multi-associations, hence, we develop a new TNN model called the hybrid TNN model, which is able to facilitate the capturing of complicated dependencies through thoughtful feature selection and engineering. First, sensor locations exhibiting analogous downscaled 1-km satellite soil moisture (SM) are identified. Next, a dynamic multilayer perceptron (D-MLP) network discerns highly correlated auxiliary-RZSM data, utilizing both ground-based and downscaled satellite data. Following this, the dual attention module identifies essential multi-associations, leveraging selected sensor and target region information. Finally, the Bayesian layer averages multi-location RZSM using the conditional probability generated based on relative relationships to yield the target location RZSM estimate. Our proposed model shows 13.066% better RZSM estimation compared to popular sequential models. The hybrid TNN RZSM estimates are used to monitor root water depletion tolerance levels for optimal PI schedules, which shows 10.846% water and 10.339% cost-saving on our selected sites. Overall the proposed model effectively demonstrates that a more accurate PI predictive algorithm saves water, improves resource conservation, and reduces irrigation costs.

Multi-objective calibration of a hydrologic model using spatially distributed remotely sensed/in-situ soil moisture

Estimation of root zone soil moisture using passive microwave remote sensing: A case study for rice and wheat crops for three states in the Indo-Gangetic basin

Root zone soil moisture (RZSM) serves as a crucial metric for assessing water stored in the soil. Modeling approaches are commonly employed in estimating RZSM. However, modelled RZSM often deviate from true RZSM values due to errors from model input data and parameters. Machine learning methods and data fusion techniques can enhance simulation accuracy. In this study, we conducted a comparative analysis of three methods for RZSM data fusion: random forest (RF), extended triple collocation (ETC), and Bayes Three Cornered Hat (BTCH).Soil moisture observation data from 2018 to 2022 were collected at 2121 sites across China from the China Meteorological Administration (Fig.1). Daily average data were calculated by arithmetically averaging hourly data and used in the analysis. Six RZSM datasets were utilized, including SMAP Level 4, GLDAS-NOAH2.1, GLDAS-Catchment2.2, ERA5, MERRA2, and CRSR. All these data were resampled to 0.25° to maintain the same spatial resolution and were arithmetically averaged as daily averages. Additionally, some parameters related to soil, climate, and vegetation were used to build a machine learning model, specifically a random forest model. Fig. 1 Distribution of soil moisture sites and daily soil moisture (m3/m3) at depths ranging from 0–50 cm across China during the period from 2018 to 2019To investigate the impact of different inputs on the performance of the RF method, three groups of inputs were employed. The specifics of the inputs used for the three methods are outlined in Table 1. The evaluation of the RF method results was carried out using a five-fold cross-validation approach.Model Inputs RFmodel1 NOAH, SMAP, ERA5, MERRA2, CFSR, CLSM, LAI, Soil properties, Meteorological data RFmodel2 NOAH, LAI, Soil properties, Meteorological data RFmodel3 NOAH, SMAP, ERA5, MERRA2, CFSR, CLSM BTCH NOAH, SMAP, ERA5, MERRA2, CFSR, CLSM ETC NOAH, MERRA2, CLSM  The boxplots show RFmodel1 performs best, emphasizing the need for comprehensive information in machine learning models. RFmodel2, superior to RFmodel3, highlights the significance of LAI, soil properties, and meteorological data in RZSM estimation. ETC and BTCH outperform individual RZSM datasets, especially in the absence of true data. The superior performance of ETC over BTCH is attributed to ETC's inputs, namely NOAH, MERRA2, and CLSM, which exhibit better accuracy compared to SMAP, ERA5, and CFSR, the inputs used by BTCH.Fig.2 Boxplots of the Pearson coefficient (R), Root Mean Square Error (RMSE), and bias between in situ root zone soil moisture (RZSM) and its estimates from the three random forest models, Bayes Three Cornered Hat (BTCH), and Extended Triple Collocation (ETC) methodsIn summary, the random forest method outperforms BTCH and ETC in the fusion of root zone soil moisture (RZSM) data, highlighting the importance of including leaf area index (LAI), soil properties, and meteorological data in the construction of the random forest model. Both BTCH and ETC demonstrate utility in enhancing RZSM estimates, making them valuable options when true data is unavailable.

In this paper, L-band microwave observations were assimilated using the ensemble Kalman filter to improve root-zone soil moisture (RZSM) estimates from a coupled soil vegetation atmosphere transfer (SVAT)-vegetation model linked to a forward microwave model. Simultaneous state-parameter updates were performed by assimilating both synthetic and field observations during a growing season of sweet corn every three days, matching the temporal coverage of observations from the Soil Moisture and Ocean Salinity and Soil Moisture Active Passive missions. The sensitivities of parameters to the states were investigated using the information-theoretic measure of conditional entropy. Among the soil parameters, the pore-size index (λ) was the most sensitive to brightness temperatures ( <i xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">T</i> <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">B</sub> ) during the early and midgrowth stages, while porosity (φ) was the most sensitive to <i xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">T</i> <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">B</sub> during the reproductive stage. In the microwave model, the soil roughness parameters, root mean square (RMS) height ( <i xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">r</i> ), and correlation length ( <i xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">l</i> ) were the most sensitive during the early and mid stages, while the vegetation regression parameter ( <i xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">b</i> ) was the most sensitive during the reproductive stage. In the synthetic experiment, assimilation of <i xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">T</i> <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">B</sub> provided RMS error reductions in RZSM estimates of 70% compared to open loop estimates. Minimal variations in performance were observed across different stages of the season during the synthetic experiment. However, when field observations of <i xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">T</i> <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">B</sub> were assimilated, significant differences in RZSM estimates were observed during different growth stages. Maximum RMS difference (RMSD) reductions in RZSM estimates of 33.3% were observed compared to open loop estimates during the early stages, while improvements of 4.8% and 16.7% were observed in the mid- and reproductive stages, respectively. Further analyses of assimilation with field observations also suggest some improvements in the SVAT model are needed for moisture transport immediately following the precipitation/irrigation events. In the microwave model, the linear vegetation formulation for estimating canopy opacity, parameterized by <i xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">b</i> , was inadequate in capturing the complexities in <i xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">T</i> <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">B</sub> during stages of high vegetative and reproductive growth rates.

Root-zone soil moisture estimation from assimilation of downscaled Soil Moisture and Ocean Salinity data

Quantification of root-zone soil moisture (RZSM) is crucial for agricultural applications and soil sciences. RZSM impacts processes such as vegetation transpiration and water percolation. Surface soil moisture (SSM) can be assessed through active and passive microwave remote sensing methods, but no current sensor enables direct RZSM retrieval. Spatial maps of RZSM can be retrieved via proxy observations (vegetation stress, water storage change, and surface soil moisture) or via land surface model predictions. In this study, we investigated the combination of surface soil moisture information with process-based inferred features involving artificial neural networks (ANNs). We considered the infiltration process through the soil water index (SWI) computed with a recursive exponential filter and the evaporation process through the evaporative efficiency computed based on a MODIS remote sensing dataset and simplified analytical model, while vegetation growth was expressed through normalized difference vegetation index (NDVI) time series. Several ANN models with different sets of features were developed. Training was conducted considering in situ stations distributed several areas worldwide characterized by different soil and climate patterns of the International Soil Moisture Network (ISMN), and testing was applied to stations of the same data hosting facility. The results indicate that the integration of process-based features into ANN models increased the overall performance over the reference model level in which only SSM features were considered. In arid and semi-arid areas, for instance, performance enhancement was observed when the evaporative efficiency was integrated into the ANN models. To assess the robustness of the approach, the trained models were applied on observation sites in Tunisia, Italy and South-India that are not part of ISMN. The results reveal that joint use of surface soil moisture, evaporative efficiency, NDVI and recursive exponential filter represented the best alternative for more accurate predictions in the case of Tunisia, where the mean correlation of the predicted RZSM based on SSM only sharply increased from 0.443 to 0.801 when process-based features were integrated into the ANN models in addition to SSM. However, process-based features have no to little added value in temperate to tropical conditions.

The importance of soil moisture is recognized globally because it controls hydrological processes that are relevant to agriculture and climate studies. Currently, estimation of root zone soil moisture is largely accomplished using physical models, which are based on flow and transport equations. However, with the complexity of the processes operating in the vadose zone as well as their interactions with each other, parameterizing all the relevant processes is quite a challenge. This complexity is further enhanced by spatio-temporal variability in soil and vegetation properties which demand model parameters to be dynamic. Alternatively, purely data-based methods for root zone soil moisture estimation are still limited despite the growing availability of datasets from networks established within the last decade. Currently, these datasets are used largely for calibration and validation of physical models and retrieval methods from satellites. In this study, we explored the utility of Random Forest (RF) as an approach for predicting and forecasting daily root zone soil moisture from selected stations in the Raam and Twente network. We trained a single RF using meteorological datasets, soil type, land cover type, and LAI as predictor variables. The model was also tuned in order to obtain the optimal hyperparameters (mtry and ntree) and number of training samples. A comparison with model simulation results using Hydrus-1D was also performed. Our results show that RF can accurately predict and forecast root zone soil moisture at the study sites based on RMSE of 0.02 – 0.12 m3m-3, in comparison with Hydrus-1D simulations having RMSE of 0.05-0.22 m3m-3. However, poor results were obtained for saturated water conditions. In addition, 5-95% RF prediction intervals become wider at saturated water conditions for some sites, which indicates higher prediction and forecast uncertainties. RF can be used for root zone soil moisture estimation, especially at data poor regions where information on soil hydraulic parameters are sparse or lacking. It can also be used for estimating missing values at gaps in time series datasets.

Enhanced root zone soil moisture monitoring using multitemporal remote sensing data and machine learning techniques

The NASA Soil Moisture Active Passive (SMAP) mission Level-4 Soil Moisture (L4_SM) product provides global, 9-km resolution, 3-hourly surface (0-5 cm) and root-zone (0-100 cm) soil moisture from April 2015 to present with a mean latency of 2.5 days from the time of observation.&amp;#160; The product is based on the assimilation of SMAP L-band (1.4 GHz) brightness temperature (Tb) observations into the NASA Catchment land surface model as the model is driven with observations-based precipitation forcing.&amp;#160;In this presentation, we describe three recent improvements in the L4_SM algorithm.&amp;#160; First, satellite- and gauge-based precipitation from the NASA Integrated Multi-satellitE Retrievals for the Global Precipitation Measurement mission (IMERG) are used in two ways: (i) The climatology to which all L4_SM precipitation forcing inputs are rescaled is based on IMERG-Final (Version 06B) data, replacing the Global Precipitation Climatology Project v2.2 data used in previous L4_SM versions, and (ii) the precipitation forcing outside of North America and the high latitudes is corrected to match the daily totals from IMERG, replacing the gauge-only, daily product or uncorrected weather analysis precipitation used there in earlier L4_SM versions.&amp;#160; Second, the Catchment model now includes the recently developed PEATCLSM hydrology module for peatlands and uses an updated global map of peatlands.&amp;#160; Third, revised parameters are used in the L-band radiative transfer model that converts the simulated soil moisture and temperature estimates into Tb predictions for use in the radiance-based L4_SM analysis.&amp;#160; Specifically, climatological parameters for the scattering albedo, soil roughness, and (seasonally-varying) vegetation opacity were derived from the SMAP Level-2 radiometer soil moisture retrieval product.&amp;#160; &amp;#160;The revised precipitation inputs result in considerably improved anomaly time series correlation skill of L4_SM surface soil moisture in South America, Africa, Australia, and parts of East Asia.&amp;#160; Particularly large improvements are seen in central Australia and Myanmar, where the quality of the gauge-only precipitation product used in earlier L4_SM versions was particularly poor. &amp;#160;In peatlands, the dynamics of water table depth, surface soil moisture and evapotranspiration are considerably improved when evaluated against in situ measurements.&amp;#160; Moreover, the time series correlation of surface and root-zone soil moisture vs. in situ measurements is slightly improved, owing to the improved annual cycle phasing of the Level-2 derived vegetation opacity parameters.&amp;#160; Collectively, these improvements are also manifested in smaller Tb observation-minus-forecast residuals.

An Ensemble Kalman Filter (EnKF)-based assimilation algorithm was implemented to estimate root-zone soil moisture (RZSM) using a Soil-Vegetation-Atmosphere Transfer (SVAT) model during a complete growing season of corn in Central Mexico. Synthetic and field soil moisture (SM) observations and NASA SMAP SM retrievals were used to understand the effect of vertically spatial updates and uncertainties in meteorological forcings on RZSM estimates. Assimilation of RZSM every 3 days using SM observations at 4 depths lowered the averaged standard deviation (ASD) and the root mean square error (RMSE) by 60 % and 50 %, respectively, compared to the open-loop ASD. The assimilation of synthetic SM at the top 0-5 cm obtained RZSM closer to observations compared to THEXMEX-18 SM measurements and SMAP SM retrievals. Differences between EnKF estimates and SM observations and SMAP SM retrievals are mainly due to misrepresentation of vegetation conditions. The results improved SM estimates up to 10-cm depth using SMAP SM retrievals; however, additional studies are needed to improve SM at deeper layers. The implemented methodology can estimate SM at the top 10 cm of the soil every 3 days to mitigate the impact of the climate change on agricultural production over rainfed areas, particularly in developing countries.

High-resolution surface and rootzone soil moisture over US cropland: A novel framework assimilating multi-source remote sensing data, machine learning, and the Layered Green and Ampt Infiltration with Redistribution model

Inversion of soil hydraulic properties from the DEnKF analysis of SMOS soil moisture over West Africa

Land surface models (LSMs) are limited in their ability to reproduce observed soil moisture partially due to uncertainties in model parameters. Here we conduct a variance-based sensitivity analysis to quantify the relative contribution of different model parameters and their interactions to the uncertainty in the surface and root-zone soil moisture in the Community Land Model 5.0 (CLM5). We focus on soil-texture-related parameters (porosity, saturated matric potential, saturated hydraulic conductivity, shape parameter of soil-water retention model) and organic matter fraction. A Gaussian process emulator is constructed based on CLM5 simulations and used to estimate soil moisture across the five-dimensional parameter space for sensitivity analysis. The procedure is demonstrated for four seasons across various U.S. sites of distinct soil and vegetation types. We find that the emulator captures well the CLM5 behavior across the parameter space for different soil textures and seasons. The uncertainties of surface and root-zone soil moisture are dominated by the uncertainties in porosity and shape parameter with negligible parametric interactions. However, relative importance of porosity versus shape parameter varies with soil textures (sites), depths (surface versus root zone), and seasons. At most of the sites, surface soil moisture uncertainty is attributed largely to shape parameter uncertainty, while porosity uncertainty is more important for the root-zone soil moisture uncertainty. All individual parameter and interaction effects demonstrate less variability across different soil textures and seasons for root zone than for surface soil moisture. These results provide scientific guidance to prioritize reducing the uncertainty of sensitive parameters for improving soil moisture modeling with CLM.

High-resolution spatial–temporal root zone soil moisture (RZSM) information collected at different scales is useful for a variety of agricultural, hydrologic, and climate applications. RZSM can be estimated using remote sensing, empirical equations, or process-based simulation models. Machine learning (ML) approaches for evaluating RZSM across numerous spatial–temporal scales are less generalizable than process-based models. However, data-driven ML approaches offer a unique opportunity to develop complex models of soil moisture without making assumptions about the processes governing soil water dynamics in a given study region. In this study, comparisons were made between two models, pySEBAL and EFSOIL, which were based on evaporation fraction (EF) and soil properties, and a data-driven model based on the Random Forest (RF) ensemble algorithm. These approaches were evaluated to demonstrate their capabilities for RZSM estimation. The EF obtained from Landsat images was used after validation with eddy covariance measurements as the major input to all three models, along with other meteorological and soil physical properties. The RF model was trained using in situ soil moisture data from Time Domain Reflectometry (TDR) sensors installed in a vineyard from 2018 to 2020. The predictor variables comprised of meteorological, soil properties, EF, and a vegetation index. The results reveal that there was a strong correlation between the in situ measured soil moisture and the RF predicted soil moisture at all sensor locations. Due to the complexity of the physical processes involved in soil water flow, the empirical models pySEBAL and EFSOIL were unable to reliably predict RZSM values at all monitored locations. The high RZSM predicted by pySEBAL demonstrated the presence of possible bias in the model’s algorithm used to estimate soil moisture. We also demonstrated that ML based on the RF algorithm may be used to predict spatially distributed RZSM when a few soil moisture ground measurements are combined with remote sensing to produce soil moisture maps.