A workflow to address pitfalls and challenges in applying machine learning models to hydrology

Gharib, Amr; Davies, Evan

doi:10.1016/j.advwatres.2021.103920

Cited by 34 publications

(10 citation statements)

References 71 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In hydrological forecasting, all principal ML approaches have recently been used, ranging from artificial neural networks (ANNs) to deep learning models, such as convolutional neural networks (CNNs), long short-term memory (LSTM) networks, support vector machines (SVMs and SVRs), decision trees, and random forests [8,10,21,67]. Each of the approaches has its specific characteristics, justifying its use for flood modeling in the given context [69,70]. The LSTM and ANN models have proven their ability to capture temporal dependencies and flexibility and provide simulation over long time periods; however, they typically require complex and large datasets for training [7,68].…”

Section: Discussionmentioning

confidence: 99%

Flood Simulations Using a Sensor Network and Support Vector Machine Model

Langhammer

2023

Water

View full text Add to dashboard Cite

This study aims to couple the support vector machine (SVM) model with a hydrometeorological wireless sensor network to simulate different types of flood events in a montane basin. The model was tested in the mid-latitude montane basin of Vydra in the Šumava Mountains, Central Europe, featuring complex physiography, high dynamics of hydrometeorological processes, and the occurrence of different types of floods. The basin is equipped with a sensor network operating in headwaters along with the conventional long-term monitoring in the outlet. The model was trained and validated using hydrological observations from 2011 to 2021, and performance was assessed using metrics such as R2, NSE, KGE, and RMSE. The model was run using both hourly and daily timesteps to evaluate the effect of timestep aggregation. Model setup and deployment utilized the KNIME software platform, LibSVM library, and Python packages. Sensitivity analysis was performed to determine the optimal configuration of the SVR model parameters (C, N, and E). Among 125 simulation variants, an optimal parameter configuration was identified that resulted in improved model performance and better fit for peak flows. The sensitivity analysis demonstrated the robustness of the SVR model, as different parameter variations yielded reasonable performances, with NSE values ranging from 0.791 to 0.873 for a complex hydrological year. Simulation results for different flood scenarios showed the reliability of the model in reconstructing different types of floods. The model accurately captured trend fitting, event timing, peaks, and flood volumes without significant errors. Performance was generally higher using a daily timestep, with mean metric values R2 = 0.963 and NSE = 0.880, compared to mean R2 = 0.913 and NSE = 0.820 using an hourly timestep, for all 12 flood scenarios. The very good performance even for complex flood events such as rain-on-snow floods combined with the fast computation makes this a promising approach for applications.

show abstract

Section: Discussionmentioning

confidence: 99%

Flood Simulations Using a Sensor Network and Support Vector Machine Model

Langhammer

2023

Water

View full text Add to dashboard Cite

show abstract

“…In regression and classification, a best practice is to use k‐fold cross‐validation that partitions available data into k‐sets and iteratively trains the model using data from each set for testing, which leads to better model generalizability (Bergmeir & Benítez, 2012). Common pitfalls in model design are including excess irrelevant or redundant variables as inputs, variable selection bias (i.e., using the same data for training and inputs), resubstitution validation (i.e., testing the model with training data), use of inconsistent cross‐validation and resampling procedures across model architectures being implemented, and data leakage (e.g., using testing data for model training or hyperparameter optimization, or pre‐processing the entire dataset prior to splitting the data into cross‐validation folds), which can lead to overfitting (Gharib & Davies, 2021; Zhang, 2007). These pitfalls can be avoided by understanding the details and limitations of the models being implemented, following best practices, and using robust ML workflows (Gharib & Davies, 2021; Zhang, 2007).…”

Section: Opportunities For Advancement Of Water Quality MLmentioning

confidence: 99%

“…These pitfalls can be avoided by understanding the details and limitations of the models being implemented, following best practices, and using robust ML workflows (Gharib & Davies, 2021;Zhang, 2007).…”

Section: State-of-the-art Machine Learning In River Water Quality Modelsmentioning

confidence: 99%

Can machine learning accelerate process understanding and decision‐relevant predictions of river water quality?

Varadharajan

Appling

Arora

et al. 2022

Hydrological Processes

View full text Add to dashboard Cite

The global decline of water quality in rivers and streams has resulted in a pressing need to design new watershed management strategies. Water quality can be affected by multiple stressors including population growth, land use change, global warming, and extreme events, with repercussions on human and ecosystem health. A scientific understanding of factors affecting riverine water quality and predictions at local to regional scales, and at sub‐daily to decadal timescales are needed for optimal management of watersheds and river basins. Here, we discuss how machine learning (ML) can enable development of more accurate, computationally tractable, and scalable models for analysis and predictions of river water quality. We review relevant state‐of‐the art applications of ML for water quality models and discuss opportunities to improve the use of ML with emerging computational and mathematical methods for model selection, hyperparameter optimization, incorporating process knowledge into ML models, improving explainablity, uncertainty quantification, and model‐data integration. We then present considerations for using ML to address water quality problems given their scale and complexity, available data and computational resources, and stakeholder needs. When combined with decades of process understanding, interdisciplinary advances in knowledge‐guided ML, information theory, data integration, and analytics can help address fundamental science questions and enable decision‐relevant predictions of riverine water quality.

show abstract

“…The success of such approaches is due as well to the mentioned increasing data availability and to the complexity of hydrological phenomena, which are difficult to model with linear or simple non linear statistical methods. For a full overview on the use of ML methods in hydrology, the reader could refer to the following recent papers: Zounemat-Kermani et al ( 2021), Gharib and Davies (2021), Rajaee et al (2020), Tyralis et al (2021).…”

Section: Introductionmentioning

confidence: 99%

Feature importance measures to dissect the role of sub-basins in shaping the catchment hydrological response: a proof of concept

Cappelli

Tauro

Apollonio

et al. 2022

Stoch Environ Res Risk Assess

View full text Add to dashboard Cite

Understanding the response of a catchment is a crucial problem in hydrology, with a variety of practical and theoretical implications. Dissecting the role of sub-basins is helpful both for advancing current knowledge of physical processes and for improving the implementation of simulation or forecast models. In this context, recent advancements in sensitivity analysis tools could be worthwhile for bringing out hidden dynamics otherwise not easy to distinguish in complex data driven investigations. In the present work seven feature importance measures are described and tested in a specific and simplified proof of concept case study. In practice, simulated runoff time series are generated for a watershed and its inner 15 sub-basins. A machine learning tool is calibrated using the sub-basins time series for forecasting the watershed runoff. Importance measures are applied on such synthetic hydrological scenario with the aim to investigate the role of each sub-basin in shaping the overall catchment response. This proof of concept offers a simplified representation of the complex dynamics of catchment response. The interesting result is that the discharge at the catchment outlet depends mainly on 3 sub-basins that are consistently identified by alternative sensitivity measures. The proposed approach can be extended to real applications, providing useful insights on the role of each sub-basin also analyzing more complex scenarios.

show abstract

A workflow to address pitfalls and challenges in applying machine learning models to hydrology

Cited by 34 publications

References 71 publications

Flood Simulations Using a Sensor Network and Support Vector Machine Model

Flood Simulations Using a Sensor Network and Support Vector Machine Model

Can machine learning accelerate process understanding and decision‐relevant predictions of river water quality?

Feature importance measures to dissect the role of sub-basins in shaping the catchment hydrological response: a proof of concept

Contact Info

Product

Resources

About