scholarly journals Modelling of the shallow water table at high spatial resolution using random forests

2019 ◽  
Vol 23 (11) ◽  
pp. 4603-4619 ◽  
Author(s):  
Julian Koch ◽  
Helen Berger ◽  
Hans Jørgen Henriksen ◽  
Torben Obel Sonnenborg
Keyword(s):  
Shallow Water ◽  
Water Table ◽  
Spatial Resolution ◽  
Random Forests ◽  
Shallow Groundwater ◽  
Coefficient Of Determination ◽  
Shallow Water Table ◽  
Simple Method ◽  
Physically Based ◽  
Physically Based Modelling

Abstract. Machine learning provides great potential for modelling hydrological variables at a spatial resolution beyond the capabilities of physically based modelling. This study features an application of random forests (RF) to model the depth to the shallow water table, for a wintertime minimum event, at a 50 m resolution over a 15 000 km2 domain in Denmark. In Denmark, the shallow groundwater poses severe risks with respect to groundwater-induced flood events, affecting both urban and agricultural areas. The risk is especially critical in wintertime, when the shallow groundwater is close to terrain. In order to advance modelling capabilities of the shallow groundwater system and to provide estimates at the scales required for decision-making, this study introduces a simple method to unify RF and physically based modelling. Results from the national water resources model in Denmark (DK-model) at a 500 m resolution are employed as covariates in the RF model. Thus, RF ensures physical consistency at a coarse scale and fully exhausts high-resolution information from readily available environmental variables. The vertical distance to the nearest water body was rated as the most important covariate in the trained RF model followed by the DK-model. The evaluation test of the trained RF model was very satisfying with a mean absolute error of 76 cm and a coefficient of determination of 0.56. The resulting map underlines the severity of groundwater flooding risk in Denmark, as the average depth to the shallow groundwater is 1.9 m and approximately 29 % of the area is characterized as having a depth of less than 1 m during a typical wintertime minimum event. This study brings forward a novel method for assessing the spatial patterns of covariate importance of the RF predictions that contributes to an increased interpretability of the RF model. Quantifying the uncertainty of RF models is still rare for hydrological applications. Two approaches, namely random forests regression kriging (RFRK) and quantile regression forests (QRF), were tested to estimate uncertainties related to the predicted groundwater levels.

scholarly journals Modelling of the shallow water table at high spatial resolution using Random Forests

2019 ◽  
Author(s):  
Julian Koch ◽  
Helen Berger ◽  
Hans Jørgen Henriksen ◽  
Torben Obel Sonnenborg
Keyword(s):  
Shallow Water ◽  
Water Table ◽  
Spatial Resolution ◽  
Random Forests ◽  
Shallow Groundwater ◽  
Coefficient Of Determination ◽  
Shallow Water Table ◽  
Simple Method ◽  
Physically Based ◽  
Physically Based Modelling

Abstract. Machine learning provides a great potential to model hydrological variables at a spatial resolution beyond the capabilities of traditional physically-based modelling. This study features an application of Random Forests (RF) to model the depth to the shallow water table, for a wintertime minimum event, at 50 m resolution over a 15,000 km2 large domain in Denmark. In Denmark, the shallow groundwater poses severe risks of groundwater induced flood events affecting both, urban and agricultural areas. The risk is especially critical in wintertime, when the shallow groundwater is close to terrain. In order to advance modelling capabilities of the shallow groundwater system and to provide estimates at scales required for decision making, this study introduces a simple method to unify RF and physically-based modelling. Results from the national water resources model in Denmark (DK-model) at 500 m resolution are employed as covariate in the RF model. Thereby, RF ensures physical consistency at coarse scale and fully exhausts high-resolution information from readily available environmental variables. The vertical distance to the nearest waterbody was rated the most important covariate in the trained RF model followed by the DK-model. The validation test of the trained RF model was very satisfying with a mean absolute error of 79 cm and a coefficient of determination of 0.55. The resulting map underlines the severity of groundwater flooding risk in Denmark, as the average depth to the shallow groundwater is 1.9 m and approximately 29 % of the area is characterised with a depth less than 1 m during a typical wintertime minimum event. This study brings forward a novel method to assess the spatial patterns of covariate importance of the RF predictions which contributes to an increased interpretability of the RF model. Quantifying uncertainty of RF models is still rare for hydrological applications. Two approaches, namely Random Forests Regression Kriging (RFRK) and Quantile Regression Forests (QRF) were tested to estimate uncertainties related to the predicted groundwater levels. This study argues that the uncertainty sources captured by RFRK and QRF can be considered independent and hence, they can be combined to a total variance through simple uncertainty propagation.

scholarly journals High Resolution Water Table Modeling of the Shallow Groundwater Using a Knowledge-Guided Gradient Boosting Decision Tree Model

2021 ◽  
Vol 3 ◽  
Author(s):  
Julian Koch ◽  
Jane Gotfredsen ◽  
Raphael Schneider ◽  
Lars Troldborg ◽  
Simon Stisen ◽  
...  
Keyword(s):  
High Resolution ◽  
Water Table ◽  
Spatial Resolution ◽  
Shallow Groundwater ◽  
Water Table Depth ◽  
Training Data ◽  
Gradient Boosting ◽  
Detailed Knowledge ◽  
Mitigation And Adaptation ◽  
Physically Based

Detailed knowledge of the uppermost water table representing the shallow groundwater system is critical in order to address societal challenges that relate to the mitigation and adaptation to climate change and enhancing climate resilience in general. Machine learning (ML) allows for high resolution modeling of the water table depth beyond the capabilities of conventional numerical physically-based hydrological models with respect to spatial resolution and overall accuracy. For this, in-situ well and proxy observations are used as training data in combination with high resolution covariates. The objective of this study is to model the depth of the uppermost water table for a typical summer and winter condition at 10 m spatial resolution over entire Denmark (43,000 km2). CatBoost, a state of the art implementation of gradient boosting decision trees, is employed in this study to model the water table depth and the associated uncertainties. The groundwater domain has not been the most prominent field of applications of recent hydrological ML advances due to the lack of big data. This study brings forward a novel knowledge-guided ML framework to overcome this limitation by integrating simulation results from a physically-based groundwater flow model. The simulation data are utilized to (1) identify wells that represent the uppermost water table, (2) augment missing training data by accounting for simulated water level seasonality, and (3) expand the list of covariates. The curated training dataset contains around 13,000 wells, 19,000 groundwater proxy observations at lakes, streams and coastline as well as 15 covariates. Cross validation attests that the ML model generalizes well with a mean absolute error of around 115 cm considering solely well observations and a MAE of <50 cm taking also the proxy observations into consideration. Quantile regression is applied to estimate confidence intervals and the estimated uncertainty is largest for moraine clay soils that are characterized with a distinct geological heterogeneity. This study highlights a novel research avenue of knowledge-guided ML for the groundwater domain by efficiently supporting a ML model with a physically-based hydrological model to predict the depth of the water table at unprecedented spatial detail and accuracy.

scholarly journals Modelling of shallow water table dynamics using conceptual and physically based integrated surface-water–groundwater hydrologic models

2019 ◽  
Vol 23 (5) ◽  
pp. 2245-2260 ◽  
Author(s):  
Mohammad Bizhanimanzar ◽  
Robert Leconte ◽  
Mathieu Nuth
Keyword(s):  
Shallow Water ◽  
Groundwater Recharge ◽  
Water Table ◽  
Soil Column ◽  
Three Dimensional ◽  
Vertical Plane ◽  
Dimensional Case ◽  
Shallow Water Table ◽  
Mike She ◽  
Physically Based

Abstract. We present a new conceptual scheme of the interaction between unsaturated and saturated zones of the MOBIDIC (MOdello Bilancio Idrologico DIstributo e Continuo) hydrological model which is applicable to shallow water table conditions. First, MODFLOW was coupled to MOBIDIC as the physically based alternative to the conceptual groundwater component of the MOBIDIC–MODFLOW. Then, assuming a hydrostatic equilibrium moisture profile in the unsaturated zone, a dynamic specific yield that is dependent on the water table level was added to MOBIDIC–MODFLOW, and calculation of the groundwater recharge in MOBIDIC was revisited using a power-type equation based on the infiltration rate, soil moisture deficit, and a calibration parameter linked to the initial water table depth, soil type, and rainfall intensity. Using the water table fluctuation (WTF) method for a homogeneous soil column, the parameter of the proposed groundwater recharge equation was determined for four soil types, i.e. sand, loamy sand, sandy loam, and loam under a pulse of rain with different intensities. The fidelity of the introduced modifications in MOBIDIC–MODFLOW was assessed by comparison of the simulated water tables against those of MIKE SHE, a physically based integrated hydrological modelling system simulating surface and groundwater flow, in two numerical experiments: a two-dimensional case of a hypothetical watershed in a vertical plane (constant slope) under a 1 cm d−1 uniform rainfall rate and a quasi-real three-dimensional watershed under 1 month of a measured daily rainfall hyetograph. The comparative analysis confirmed that the simplified approach can mimic simple and complex groundwater systems with an acceptable level of accuracy. In addition, the computational efficiency of the proposed approach (MIKE SHE took 180 times longer to solve the three-dimensional case than the MOBIDIC–MODFLOW framework) demonstrates its applicability to real catchment case studies.

Historical soil maps with a reference to redoximorphic features as basis for ecohydrological modelling.

2021 ◽  
Author(s):  
Alla Yurova ◽  
Daniil Kozlov ◽  
Maria Smirnova ◽  
Pavel Fil
Keyword(s):  
Shallow Water ◽  
Water Table ◽  
Expert Knowledge ◽  
Shallow Groundwater ◽  
Wave Form ◽  
Kinematic Wave ◽  
State Variables ◽  
Soil Maps ◽  
Shallow Water Table ◽  
Redoximorphic Features

<p>Historical soil maps with a reference to profile redoximorphic features have obvious utility for ecohydrological modelling. That is particularly pertinent in areas with shallow water tables where catchments have both dry and moist parts, latter due to moisture source from either upper or low boundary. However, there is no convenient method for converting maps to hydrological model state variables. Here we propose that the steady state continuity equation in kinematic wave form can be parameterized using expert knowledge of the typical water table depth (WTD) for soils with different hydromorphy degrees based on redoximorphic features. To test the approach, six hillslope-based computational units (catenas) were obtained, for use in simulations, by automated of the Samovetc and Izberdey catchments in the Tambov region (Russia) using lumpR software. Five of the six catenas began at poorly drained flat upslope positions with soils with various degrees of saturation by shallow groundwater and one began at a well-drained upslope position. We guided parameterization of the kinematic wave model by critical range of the WTD known to correspond to each soil group on historical soil map of hydromorphy degree. Application of expert knowledge in this manner alone yielded a broad range of possible WTD values (e.g. 1.5-5 m for a semi-hydromorphic soil) for each soil entity, but linking a catena by the fundamental physical constraint of flow continuity enabled narrowing of the range to 0.2-1 m thereby reducing it by ca. 80%. We further tested the shallow water table approximation in the WASA-SED ecohydrological model based on catenary approach to simulate soil moisture profiles referring explicitly to soil groups of different hydromorphy degree and distinguish stagnic and gleyic regimes of waterlogging. The results show that the approach could substantially improve crop and water management precision.</p>

Soil salinization at a side-hill seep and closed basin in southern Alberta

1993 ◽  
Vol 73 (2) ◽  
pp. 209-222 ◽  
Author(s):  
J. J. Miller ◽  
G. J. Beke ◽  
S. Pawluk
Keyword(s):  
Shallow Water ◽  
Water Table ◽  
Shallow Groundwater ◽  
Soil Salinization ◽  
Upward Movement ◽  
Soluble Salts ◽  
Shallow Water Table ◽  
Soil Extracts ◽  
Closed Basin ◽  
Southern Alberta

One saline soil in a side-hill seep and another in a closed basin in southern Alberta were investigated using hydrological, chemical and mineralogical techniques to compare the nature of soil salinization. The morphology of the saline Gleyed Regosolic soil in the seep indicated strong upward movement of water and soluble salts. This was evident from the presence of mottles in the soil solum, a shallow water table (≤ 2.69 m), and high EC (2.8–32.5 dS m−1) and SARt (4.8–55.5) values in the shallow groundwater and soil. High soluble Mg/Ca ratios (2.1) in the soil extract of the Ccasa3 horizon and the presence of Mg-calcites (4–5 mol %) indicated restricted leaching at this site. The morphology of the saline Orthic Dark Brown Chernozemic soil in the closed basin reflected slight upward movement of water and soluble salts. This morphology was consistent with a deeper water table (≥ 2.62 m), and lower EC (2.2–10.0 dS m−1) and SARt (2.9–11.3) values in the shallow groundwater and soil. High Mg/Ca ratios (4.8–7.0) in soil extracts of the C horizons, but no Mg-calcites, indicated greater leaching at this site. We estimated that it would take about 77 yr to salinize the soil at the seep via upward groundwater movement, and 6500 yr to salinize the soil at the closed basin. The high tritium content of the shallow groundwater at both sites suggested that downward (closed basin) and lateral (seep) movement of water to the water table was an important factor contributing to shallow water tables and soil salinization. Key words: Salinization, hydrology, chemistry, mineralogy, side-hill seep, closed basin

scholarly journals Simulation of upward flux from shallow water-table using UPFLOW model

2013 ◽  
Vol 3 ◽  
pp. 123-127
Author(s):  
M.H. Ali ◽  
I. Abustan ◽  
S. Islam
Keyword(s):  
Shallow Water ◽  
Water Table ◽  
Capillary Rise ◽  
Coefficient Of Determination ◽  
Reliable Estimate ◽  
Graphical Display ◽  
Upward Movement ◽  
Irrigation Requirement ◽  
Shallow Water Table ◽  
Upward Flux

The upward movement of water by capillary rise from shallow water-table to the root zone is an important incoming flux. For determining exact amount of irrigation requirement, estimation of capillary flux or upward flux is essential. Simulation model can provide a reliable estimate of upward flux under variable soil and climatic conditions. In this study, the performance of model UPFLOW to estimate upward flux was evaluated. Evaluation of model performance was performed with both graphical display and statistical criteria. In distribution of simulated capillary rise values against observed field data, maximum data points lie around the 1:1 line, which means that the model output is reliable and reasonable. The coefficient of determination between observed and simulated values was 0.806 (r = 0.93), which indicates a good inter-relation between observed and simulated values. The relative error, model efficiency, and index of agreement were found as 27.91%, 85.93% and 0.96, respectively. Considering the graphical display of observed and simulated upward flux and statistical indicators, it can be concluded that the overall performance of the UPFLOW model in simulating actual upward flux from a crop field under variable water-table condition is satisfactory. Thus, the model can be used to estimate capillary rise from shallow water-table for proper estimation of irrigation requirement, which would save valuable water from over-irrigation.

scholarly journals Modelling of shallow water table dynamics using conceptual and physically based integrated surface water-groundwater hydrologic models

2018 ◽  
Author(s):  
Mohammad Bizhanimanzar ◽  
Robert Leconte ◽  
Mathieu Nuth
Keyword(s):  
Shallow Water ◽  
Groundwater Recharge ◽  
Water Table ◽  
Sandy Loam ◽  
Soil Column ◽  
Vertical Plane ◽  
Specific Yield ◽  
Shallow Water Table ◽  
Mike She ◽  
Physically Based

Abstract. We present a new conceptual scheme of the interaction between unsaturated and saturated zones of the MOBIDIC (Modello Bilancio Idrologico DIstributo e Continuo) hydrological model which is applicable to shallow water table conditions. First, a hydrostatic equilibrium moisture profile was assumed for simulating changes in water table levels. This resulted in a water table based expression of specific yield, which was included in the coupled MOBIDIC-MODFLOW modelling framework for capturing shallow water tables fluctuations. Second, the groundwater recharge was defined using a power type equation based on infiltration rate, soil moisture deficit and a calibration parameter linked to initial water table level, soil type and rainfall intensity. Using the Water Table Fluctuation (WTF) method, the water table rise for a homogeneous soil column under a pulse of rain with different intensities (up to 30 mm/day) the parameter of the proposed groundwater recharge equation was determined for four soil types i.e., sand, loamy sand, sandy loam and loam. The simulated water table levels were compared against those simulated by MIKE-SHE, a physically based integrated hydrological modelling system simulating surface and groundwater flow. Two numerical experiments were carried out: a two-dimensional case of a hypothetical watershed in a vertical plane (constant slope) under a 1 cm/day uniform rainfall rate, and a quasi-real three dimensional watershed under one month of measured daily rainfall hyetograph. The comparative analysis confirmed that the simplified approach can mimic simple and complex groundwater systems with an acceptable level of accuracy. In addition, the computational efficiency of the proposed approach (MIKE-SHE took 180 times longer to solve the 3D case than the MOBIDIC-MODFLOW framework) demonstrates its applicability to real catchment case studies.

Sign in / Sign up

Export Citation Format

Share Document