Search

We introduce a new diagnosis tool that is well suited to analyzing simulation results over large samples of watersheds. It consists of a modification of the classical Taylor diagram to simultaneously visualize several error components (based on bias, standard deviation or squared errors) that are commonly used in efficiency criteria (such as the Nash–Sutcliffe efficiency (NSE) or the Kling–Gupta efficiency (KGE)) to evaluate hydrological model performance. We propose a methodological framework that explicitly links the graphical and numerical evaluation approaches, and show how they can be usefully combined to visually interpret numerical experiments conducted on large datasets. The approach is illustrated using results obtained by testing two rainfall-runoff models on a sample of 2050 watersheds from 8 countries and calibrated with two alternative objective functions (NSE and KGE). The assessment tool clearly highlights well-documented problems related to the use of the NSE for the calibration of rainfall-runoff models, which arise due to interactions between the ratio of simulated to observed standard deviations and the correlation coefficient. We also illustrate the negative impacts of classical mathematical transformations (square root) applied to streamflow when employing NSE and KGE as metrics for model calibration.

Process-based modelling offers interpretability and physical consistency in many domains of geosciences but struggles to leverage large datasets efficiently. Machine-learning methods, especially deep networks, have strong predictive skills yet are unable to answer specific scientific questions. In this Perspective, we explore differentiable modelling as a pathway to dissolve the perceived barrier between process-based modelling and machine learning in the geosciences and demonstrate its potential with examples from hydrological modelling. ‘Differentiable’ refers to accurately and efficiently calculating gradients with respect to model variables or parameters, enabling the discovery of high-dimensional unknown relationships. Differentiable modelling involves connecting (flexible amounts of) prior physical knowledge to neural networks, pushing the boundary of physics-informed machine learning. It offers better interpretability, generalizability, and extrapolation capabilities than purely data-driven machine learning, achieving a similar level of accuracy while requiring less training data. Additionally, the performance and efficiency of differentiable models scale well with increasing data volumes. Under data-scarce scenarios, differentiable models have outperformed machine-learning models in producing short-term dynamics and decadal-scale trends owing to the imposed physical constraints. Differentiable modelling approaches are primed to enable geoscientists to ask questions, test hypotheses, and discover unrecognized physical relationships. Future work should address computational challenges, reduce uncertainty, and verify the physical significance of outputs.

This study leverages a Random Forest model to predict flood hazard in Arizona, New Mexico, Colorado, and Utah, focusing on enhancing sustainability in flood management. Utilizing the National Flood Hazard Layer (NFHL), an intricate flood map of Arizona was generated, with the Random Forest Classification algorithm assessing flood hazard for each grid cell. Weather variable predictions from TerraClimate were integrated with NFHL classifications and Digital Elevation Model (DEM) analyses, providing a comprehensive understanding of flood dynamics. The research highlights the model's capability to predict flood hazard in areas lacking NFHL classifications, thereby supporting sustainable flood management by elucidating weather's influence on flood hazard. This approach aligns with sustainable development goals by aiding in resilient infrastructure design and informed urban planning, reducing the impact of floods on communities. Despite recognizing constraints such as input data precision and the model's potential limitations in capturing complex variable interactions, the methodology offers a robust framework for flood hazard evaluation in other regions. Integrating diverse data sources, this study presents a valuable tool for decision-makers, supporting sustainable practices, and enhancing the resilience of vulnerable regions against flood hazards. This integrated approach underscores the potential of advanced modeling techniques in promoting sustainability in environmental hazard management. [ABSTRACT FROM AUTHOR]

We investigate the applicability of machine learning technologies to the development of parsimonious, interpretable, catchment‐scale hydrologic models using directed‐graph architectures based on the mass‐conserving perceptron (MCP) as the fundamental computational unit. Here, we focus on architectural complexity (depth) at a single location, rather than universal applicability (breadth) across large samples of catchments. The goal is to discover a minimal representation (numbers of cell‐states and flow paths) that represents the dominant processes that can explain the input‐state‐output behaviors of a given catchment, with particular emphasis given to simulating the full range (high, medium, and low) of flow dynamics. We find that a "HyMod Like" architecture with three cell‐states and two major flow pathways achieves such a representation at our study location, but that the additional incorporation of an input‐bypass mechanism significantly improves the timing and shape of the hydrograph, while the inclusion of bi‐directional groundwater mass exchanges significantly enhances the simulation of baseflow. Overall, our results demonstrate the importance of using multiple diagnostic metrics for model evaluation, while highlighting the need for properly selecting and designing the training metrics based on information‐theoretic foundations that are better suited to extracting information across the full range of flow dynamics. This study sets the stage for interpretable regional‐scale MCP‐based hydrological modeling (using large sample data) by using neural architecture search to determine appropriate minimal representations for catchments in different hydroclimatic regimes. Plain Language Summary: We show that conventional machine learning technologies can be used to develop parsimonious, interpretable, catchment‐scale hydrologic models using the mass‐conserving perceptron (MCP) as a fundamental computational unit. Using data from the Leaf River Basin, we test a variety of minimal, dominant process, representations that can explain the input‐state‐output dynamics of the catchment. Our results demonstrate the importance of using multiple diagnostic metrics for evaluation and comparison of different model architectures, and highlight the importance of choosing (or designing) objective functions for model training that are properly suited to the task of extracting information across the full range of flow dynamics. This depth‐focus study sets the stage for interpretable regional‐scale MCP‐based hydrological modeling (using large sample data) by using neural architecture search to determine appropriate minimal representations for catchments in different hydroclimatic regimes. Key Points: We utilize mass‐conserving perceptron (MCP) directed‐graph architectures to develop concise, interpretable catchment‐scale hydrologic modelsWe focus on model complexity (depth) at a single location, rather than universal applicability (breadth) across large samples of catchmentsThis study set the stage for interpretable MCP‐based modeling to find minimal representations in different hydroclimatic regimes [ABSTRACT FROM AUTHOR]

Using information-theoretic quantities in practical applications with continuous data is often hindered by the fact that probability density functions need to be estimated in higher dimensions, which can become unreliable or even computationally unfeasible. To make these useful quantities more accessible, alternative approaches such as binned frequencies using histograms and k-nearest neighbors (k-NN) have been proposed. However, a systematic comparison of the applicability of these methods has been lacking. We wish to fill this gap by comparing kernel-density-based estimation (KDE) with these two alternatives in carefully designed synthetic test cases. Specifically, we wish to estimate the information-theoretic quantities: entropy, Kullback–Leibler divergence, and mutual information, from sample data. As a reference, the results are compared to closed-form solutions or numerical integrals. We generate samples from distributions of various shapes in dimensions ranging from one to ten. We evaluate the estimators' performance as a function of sample size, distribution characteristics, and chosen hyperparameters. We further compare the required computation time and specific implementation challenges. Notably, k-NN estimation tends to outperform other methods, considering algorithmic implementation, computational efficiency, and estimation accuracy, especially with sufficient data. This study provides valuable insights into the strengths and limitations of the different estimation methods for information-theoretic quantities. It also highlights the significance of considering the characteristics of the data, as well as the targeted information-theoretic quantity when selecting an appropriate estimation technique. These findings will assist scientists and practitioners in choosing the most suitable method, considering their specific application and available data. We have collected the compared estimation methods in a ready-to-use open-source Python 3 toolbox and, thereby, hope to promote the use of information-theoretic quantities by researchers and practitioners to evaluate the information in data and models in various disciplines. [ABSTRACT FROM AUTHOR]

Several studies have demonstrated the ability of long short-term memory (LSTM) machine-learning-based modeling to outperform traditional spatially lumped process-based modeling approaches for streamflow prediction. However, due mainly to the structural complexity of the LSTM network (which includes gating operations and sequential processing of the data), difficulties can arise when interpreting the internal processes and weights in the model. Here, we propose and test a modification of LSTM architecture that is calibrated in a manner that is analogous to a hydrological system. Our architecture, called "HydroLSTM", simulates the sequential updating of the Markovian storage while the gating operation has access to historical information. Specifically, we modify how data are fed to the new representation to facilitate simultaneous access to past lagged inputs and consolidated information, which explicitly acknowledges the importance of trends and patterns in the data. We compare the performance of the HydroLSTM and LSTM architectures using data from 10 hydro-climatically varied catchments. We further examine how the new architecture exploits the information in lagged inputs, for 588 catchments across the USA. The HydroLSTM-based models require fewer cell states to obtain similar performance to their LSTM-based counterparts. Further, the weight patterns associated with lagged input variables are interpretable and consistent with regional hydroclimatic characteristics (snowmelt-dominated, recent rainfall-dominated, and historical rainfall-dominated). These findings illustrate how the hydrological interpretability of LSTM-based models can be enhanced by appropriate architectural modifications that are physically and conceptually consistent with our understanding of the system. [ABSTRACT FROM AUTHOR]

Several studies have demonstrated the ability of Long Short-Term Memory (LSTM) machine learning based modeling to outperform traditional spatially-lumped process-based modeling approaches for streamflow prediction. However, due mainly to the structural complexity of the LSTM network (which includes gating operations and sequential processing of the data), difficulties can arise when interpreting the internal processes and weights in the model. Here, we propose and test a modification of LSTM architecture that is calibrated in a manner that is analogous to a hydrological system. Our architecture, called HydroLSTM, simulates the sequential updating of the Markovian storage while the gating operation has access to historical information. Specifically, we modify how data is fed to the new representation to facilitate simultaneous access to past lagged inputs and consolidated information, which explicitly acknowledges the importance of trends and patterns in the data. We compare the performance of the HydroLSTM and LSTM architectures using data from 10 hydro-climatically varied catchments. We further examine how the new architecture exploits the information in lagged inputs, for 588 catchments across the USA. The HydroLSTM-based models require fewer cell states to obtain similar performance to their LSTM-based counterparts. Further, the weight patterns associated with lagged input variables are interpretable and consistent with regional hydroclimatic characteristics (snowmelt-dominated, recent rainfall-dominated, and historical rainfall-dominated). These findings illustrate how the hydrological interpretability of LSTM-based models can be enhanced by appropriate architectural modifications that are physically and conceptually consistent with our understanding of the system. [ABSTRACT FROM AUTHOR]

We study the sensitivity of aquifer‐scale estimates of transmissivity (T) and storativity (S) to the variance and correlation length scale of aquifer heterogeneity, when such estimates are obtained by the traditional approach of analyzing pumping test data. We consider both constant‐rate and variable‐rate pumping tests, and a variety of Theis‐based solution methodologies (single‐ and multiple‐observation well methods, and interpreted single value or transient values for T and S parameters) applied in pumping test data analysis. Our results indicate that achieving reliable inference of effective T and S requires that pumping be continued until the radius of the test‐induced cone of depression exceeds a "representative length" that corresponds to ∼15 times than the correlation length scale of the aquifer heterogeneity. Independent of solution type, pumping history, correlation length scale, magnitude of the sill, and location of observation well, the estimates for T will converge toward the geometric mean. For S, the estimation uncertainty is relatively large for observation wells that are close to the pumping well, but diminishes for observation wells located beyond the representative distance, resulting in convergence to the arithmetic mean. Given that these results have practical implications for how pumping tests should be carried out, we present a simple "rule‐of‐thumb" for estimating the correlation length scale of the heterogeneity of an aquifer. Key Points: The T estimates for constant and heterogeneous S aquifers differ for every observation well; instead, the difference for S strongly depends on observation well distancePumping rate and history does not affect the inferred aquifer parameters; however, Theis‐based solution type results in different interpretationsAs a rule‐of‐thumb, temporal variability of T and S diminishes when the depression cone radius exceeds 15 times the correlation length scale plus the observation well distance [ABSTRACT FROM AUTHOR]