Search

Accurate information on surface soil moisture (SSM) content at a global scale under different climatic conditions is important for hydrological and climatological applications. Machine learning (ML) based systematic integration of in-situ hydrological measurements, complex environmental and climate data and satellite observation facilitate to generate the best data products to monitor and analyse the exchanges of water, energy and carbon in the Earth system at a proper space-time resolution. This study investigates the estimation of daily SSM using eight optimised ML algorithms and ten ensemble models (constructed via model bootstrap aggregating techniques and five-fold cross-validation). The algorithmic implementations were trained and tested using the international soil moisture network (ISMN) data collected from 1722 stations distributed across the World. The result showed that K-neighbours Regressor (KNR) performs best on "test_random" set, while Random Forest Regressor (RFR) performs best on "test_temporal" and "test_independent-stations". Independent evaluation on novel stations across different climate zones was conducted. For the optimised ML algorithms, the median RMSEs were below 0.1 cm3/cm3. GradientBoosting (GB), Multi-layer Perceptron Regressor (MLPR), Stochastic Gradient Descent Regressor (SGDR), and Random Forest Regressor (RFR) achieved a median r score of 0.6 in twelve, eleven, nine and nine climate zones, respectively, out of fifteen climate zones. The performance of ensemble models improved significantly with the median value of RMSE below 0.075 cm3/cm3 for all climate zones . All voting regressors achieved the r scores of above 0.6 in thirteen climate zones except BSh and BWh because of the sparse distribution of training stations. The metrical evaluation showed that ensemble models can improve the performance of single ML algorithms and achieve more stable results. Based on the results computed for three different test sets, the ensemble model with KNR, RFR and XB performed the best. Overall, our investigation shows that ensemble machine learning algorithms have a greater capability for predicting SSM compared to the optimised, or base ML algorithms, and indicates their huge potential applicability in estimating water cycle budgets, managing irrigation and predicting crop yields. [ABSTRACT FROM AUTHOR]

How to efficiently and physically integrate the soil water dynamics with groundwater flow processes has drawn much attention. We present a coupled soil water-groundwater model, considering the two-way feedback coupling scheme, and verified its performance using two synthetic cases (using the fully 3D variably saturated flow (VSF) model simulations as the 'reference') and one real catchment 25 case (using the groundwater table depth and soil moisture profile measurements). By the cross-validation between the observations and various model simulations, the two-way coupling approach is proven physically accurate and is applicable for large-scale groundwater flow problems. Compared to the simulation by groundwater model alone (i.e., only MODFLOW), the coupling of MODFLOW with one soil column reduced the overestimation of groundwater table simulation (taking the VSF model simulations as the reference). The results were further improved as more soil columns were used to represent the heterogeneous soil watergroundwater interactions. Compared to the HYDRUS-MODFLOW, the two-way coupling approach produces a similar spatial distribution of hydraulic heads while better performs in mimicking the temporal dynamics of groundwater table depth and soil moisture profiles. We attribute the better performance to the different coupling strategies across the soil-water and groundwater interface. It is thus suggested to adopt the two-way feedback coupling scheme, together with the moving phreatic boundary and multi-scale water balance analysis, to maintain physical consistency and reduce coupling errors. The realistic implementation of the vadose zone processes (with STEMMUS), coupling approach, and spatiotemporal heterogeneity of soil water-groundwater interactions were demonstrated critical to accurately represent an integrated soil water-groundwater system. The developed STEMMUS-MODFLOW model can be further equipped with different complexities of soil physics (e.g., coupled soil water and heat transfer, freeze-thaw, airflow processes), surface hydrology (snowfall, runoff), soil and plant biogeochemical processes, towards an integrated "from bedrock to atmosphere" modeling framework. [ABSTRACT FROM AUTHOR]