David Watkins
David Watkins is a machine learning engineer in the Data Science Branch of the Water Mission Area Integrated Information Dissemination Division.
David is focused on accelerating machine learning model development and building systems for hydrologic forecasting in the cloud. In 2021 he helped deliver the WMA’s first operational forecast system, forecasting water temperature in the Delaware River Basin for reservoir operators.
Education and Certifications
M.S. in Geophysics from the University of Wisconsin-Madison
Science and Products
Data-Driven Drought Prediction Project Model Outputs for Select Spatial Units within the Conterminous United States
This metadata record describes model outputs and supporting model code for the Data-Driven Drought Prediction project of the Water Resources Mission Area Drought Program. The data listed here include outputs of multiple machine learning model types for predicting hydrological drought at select locations within the conterminous United States. The child items referenced below correspond to different
Data to support near-term forecasts of stream temperature using process-guided deep learning and data assimilation
This data release contains the forcings and outputs of 7-day ahead maximum water temperature forecasting models that made real-time predictions in the Delaware River Basin during 2021. The model is driven by weather forecasts and observed reservoir releases and produces maximum water temperature forecasts for the issue day (day 0) and 7 days into the future (days 1-7) at five sites. This data rele
Predicting water temperature in the Delaware River Basin
Daily temperature predictions in the Delaware River Basin (DRB) can inform decision makers who can use cold-water reservoir releases to maintain thermal habitat for sensitive fish and mussel species. This data release supports a variety of flow and water temperature modeling efforts and provides the inputs and outputs of both machine learning and process-based modeling methods across 456 river rea
Evaluating deep learning architecture and data assimilation for improving water temperature forecasts at unmonitored locations
Deep learning (DL) models are increasingly used to forecast water quality variables for use in decision making. Ingesting recent observations of the forecasted variable has been shown to greatly increase model performance at monitored locations; however, observations are not collected at all locations, and methods are not yet well developed for DL models for optimally ingesting recent observations
Authors
Jacob Aaron Zwart, Jeremy Alejandro Diaz, Scott Douglas Hamshaw, Samantha K. Oliver, Jesse Cleveland Ross, Margaux Jeanne Sleckman, Alison P. Appling, Hayley Corson-Dosch, Xiaowei Jia, Jordan S Read, Jeffrey M Sadler, Theodore Paul Thompson, David Watkins, Elaheh (Ellie) White
Regional streamflow drought forecasting in the Colorado River Basin using Deep Neural Network models
Process-based, large-scale (e.g., conterminous United States [CONUS]) hydrologic models have struggled to achieve reliable streamflow drought performance in arid regions and for low-flow periods. Deep learning has recently seen broad implementation in streamflow prediction and forecasting research projects throughout the world with performance often equaling or exceeding that of process-based mode
Authors
Scott Douglas Hamshaw, Phillip J. Goodling, Konrad Hafen, John C. Hammond, Ryan R. McShane, Roy Sando, Apoorva Ramesh Shastry, Caelan E. Simeone, David Watkins, Elaheh (Ellie) White, Michael Wieczorek
Near-term forecasts of stream temperature using deep learning and data assimilation in support of management decisions
Deep learning (DL) models are increasingly used to make accurate hindcasts of management-relevant variables, but they are less commonly used in forecasting applications. Data assimilation (DA) can be used for forecasts to leverage real-time observations, where the difference between model predictions and observations today is used to adjust the model to make better predictions tomorrow. In this us
Authors
Jacob Aaron Zwart, Samantha K. Oliver, William Watkins, Jeffrey Michael Sadler, Alison P. Appling, Hayley Corson-Dosch, Xiaowei Jia, Vipin Kumar, Jordan Read
Physics-guided neural networks (PGNN): An application in lake temperature modeling
This chapter introduces a framework for combining scientific knowledge of physics-based models with neural networks to advance scientific discovery. It explains termed physics-guided neural networks (PGNN), leverages the output of physics-based model simulations along with observational features in a hybrid modeling setup to generate predictions using a neural network architecture. Data science ha
Authors
Arka Daw, Anuj Karpatne, William Watkins, Jordan Read, Vipin Kumar
Process-guided deep learning predictions of lake water temperature
The rapid growth of data in water resources has created new opportunities to accelerate knowledge discovery with the use of advanced deep learning tools. Hybrid models that integrate theory with state‐of‐the art empirical techniques have the potential to improve predictions while remaining true to physical laws. This paper evaluates the Process‐Guided Deep Learning (PGDL) hybrid modeling framework
Authors
Jordan S. Read, Xiaowei Jia, Jared Willard, Alison P. Appling, Jacob Aaron Zwart, Samantha K. Oliver, Anuj Karpatne, Gretchen J. A. Hansen, Paul C. Hanson, William Watkins, Michael Steinbach, Vipin Kumar
Science and Products
- Data
Data-Driven Drought Prediction Project Model Outputs for Select Spatial Units within the Conterminous United States
This metadata record describes model outputs and supporting model code for the Data-Driven Drought Prediction project of the Water Resources Mission Area Drought Program. The data listed here include outputs of multiple machine learning model types for predicting hydrological drought at select locations within the conterminous United States. The child items referenced below correspond to differentData to support near-term forecasts of stream temperature using process-guided deep learning and data assimilation
This data release contains the forcings and outputs of 7-day ahead maximum water temperature forecasting models that made real-time predictions in the Delaware River Basin during 2021. The model is driven by weather forecasts and observed reservoir releases and produces maximum water temperature forecasts for the issue day (day 0) and 7 days into the future (days 1-7) at five sites. This data relePredicting water temperature in the Delaware River Basin
Daily temperature predictions in the Delaware River Basin (DRB) can inform decision makers who can use cold-water reservoir releases to maintain thermal habitat for sensitive fish and mussel species. This data release supports a variety of flow and water temperature modeling efforts and provides the inputs and outputs of both machine learning and process-based modeling methods across 456 river rea - Publications
Evaluating deep learning architecture and data assimilation for improving water temperature forecasts at unmonitored locations
Deep learning (DL) models are increasingly used to forecast water quality variables for use in decision making. Ingesting recent observations of the forecasted variable has been shown to greatly increase model performance at monitored locations; however, observations are not collected at all locations, and methods are not yet well developed for DL models for optimally ingesting recent observationsAuthorsJacob Aaron Zwart, Jeremy Alejandro Diaz, Scott Douglas Hamshaw, Samantha K. Oliver, Jesse Cleveland Ross, Margaux Jeanne Sleckman, Alison P. Appling, Hayley Corson-Dosch, Xiaowei Jia, Jordan S Read, Jeffrey M Sadler, Theodore Paul Thompson, David Watkins, Elaheh (Ellie) WhiteRegional streamflow drought forecasting in the Colorado River Basin using Deep Neural Network models
Process-based, large-scale (e.g., conterminous United States [CONUS]) hydrologic models have struggled to achieve reliable streamflow drought performance in arid regions and for low-flow periods. Deep learning has recently seen broad implementation in streamflow prediction and forecasting research projects throughout the world with performance often equaling or exceeding that of process-based modeAuthorsScott Douglas Hamshaw, Phillip J. Goodling, Konrad Hafen, John C. Hammond, Ryan R. McShane, Roy Sando, Apoorva Ramesh Shastry, Caelan E. Simeone, David Watkins, Elaheh (Ellie) White, Michael WieczorekNear-term forecasts of stream temperature using deep learning and data assimilation in support of management decisions
Deep learning (DL) models are increasingly used to make accurate hindcasts of management-relevant variables, but they are less commonly used in forecasting applications. Data assimilation (DA) can be used for forecasts to leverage real-time observations, where the difference between model predictions and observations today is used to adjust the model to make better predictions tomorrow. In this usAuthorsJacob Aaron Zwart, Samantha K. Oliver, William Watkins, Jeffrey Michael Sadler, Alison P. Appling, Hayley Corson-Dosch, Xiaowei Jia, Vipin Kumar, Jordan ReadPhysics-guided neural networks (PGNN): An application in lake temperature modeling
This chapter introduces a framework for combining scientific knowledge of physics-based models with neural networks to advance scientific discovery. It explains termed physics-guided neural networks (PGNN), leverages the output of physics-based model simulations along with observational features in a hybrid modeling setup to generate predictions using a neural network architecture. Data science haAuthorsArka Daw, Anuj Karpatne, William Watkins, Jordan Read, Vipin KumarProcess-guided deep learning predictions of lake water temperature
The rapid growth of data in water resources has created new opportunities to accelerate knowledge discovery with the use of advanced deep learning tools. Hybrid models that integrate theory with state‐of‐the art empirical techniques have the potential to improve predictions while remaining true to physical laws. This paper evaluates the Process‐Guided Deep Learning (PGDL) hybrid modeling frameworkAuthorsJordan S. Read, Xiaowei Jia, Jared Willard, Alison P. Appling, Jacob Aaron Zwart, Samantha K. Oliver, Anuj Karpatne, Gretchen J. A. Hansen, Paul C. Hanson, William Watkins, Michael Steinbach, Vipin Kumar