Janet Rice Barclay

Janet Rice Barclay

Auto-generated profile, based on publicly available information

Janet Rice Barclay

New England Water Science Center U.S. Geological Survey East Hartford CT USA

Research Focus & Metrics

Academic Career Length

12Years (2015-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
4 April 2023 in Water Resources Research
Verified
Journal Article
Open access
STSimon Topp
Simon Topp
JBJanet Barclay
Janet Barclay
JDJeremy Diaz
Jeremy Diaz
ASAlexander Sun
Alexander Sun
Xiaowei JiaXiaowei Jia
Xiaowei Jia
AAAlison Appling
Alison Appling
Stream temperature is a fundamental control on ecosystem health. Recent efforts incorporating process guidance into deep learning models for predicting stream temperature have been shown to outperform existing statistical and physical models. This performance is in part because deep learning architectures can actively learn spatiotemporal relationships that govern how water and energy propagate through a river network. However, exploration of how spatiotemporal awareness and process guidance influence a model's generalizability under shifting environmental conditions such as climate change is limited. Here, we use Explainable Artificial Intelligence (XAI) to interrogate how differing deep learning architectures affect a model's learned spatial and temporal dependencies, and how those learned dependencies affect a model's ability to maintain high accuracy when applied to unseen environmental conditions. Using the Delaware River Basin in the northeastern United States as a test case, we compare two spatiotemporally aware process‐guided deep learning models for predicting stream temperature (a recurrent graph convolution network—RGCN, and a temporal convolution graph model—Graph WaveNet). Both models achieve equally high predictive performance when testing data are well represented in the training data (test root mean squared errors of 1.64°C and 1.65°C); however, Graph WaveNet significantly outperforms RGCN in 4 out of 5 experiments where test partitions represent different types of unseen environmental conditions. XAI results show that the architecture of Graph WaveNet leads to learned spatial relationships with greater fidelity to physical processes, and that this fidelity improves the generalizability of the model when applied to shifting and/or unseen environmental conditions.

13 October 2020 in Water Resources Research
Verified
Journal Article
Open access
PDF file
JBJ. R. Barclay
J. R. Barclay
JSJ. J. Starn
J. J. Starn
MBMartin A. Briggs
Martin A. Briggs
AHA. M. Helton
A. M. Helton
Groundwater discharge zones connect aquifers to surface water, generating baseflow and serving as ecosystem control points across aquatic ecosystems. The influence of groundwater discharge on surface flow connectivity, fate and transport of contaminants and nutrients, and thermal habitat depends strongly on hydrologic characteristics such as the spatial distribution, age, and depth of source groundwater flow paths. Groundwater models have the potential to predict spatial discharge characteristics within river networks, but models are often not evaluated against these critical characteristics and model equifinality with respect to discharge processes is a known challenge. We quantify discharge characteristics across a suite of groundwater models with commonly used frameworks and calibration data. We developed a base model (MODFLOW‐NWT) for a 1,570‐km2 watershed in the northeastern United States and varied the calibration data, control of river‐aquifer exchange directionality, and resolution. Most models (n = 11 of 12) fit similarly to calibration metrics, but patterns in discharge location, flow path depth, and subsurface travel time varied substantially. We found (1) a 15% difference in the percent of discharge going to first‐order streams, (2) threefold variations in flow path depth, and (3) sevenfold variations in the subsurface travel times among the models. We recalibrated three models using a synthetic discharge location data set. Calibration with discharge location data reduced differences in simulated discharge characteristics, suggesting an approach to improved equifinality based on widespread field‐based mapping of discharge zones. Our work quantifying variation across common modeling approaches is an important step toward characterizing and improving predictions of groundwater discharge characteristics.

22 December 2023 in Water Resources Research
Verified
Journal Article
Open access
JBJ. R. Barclay
J. R. Barclay
STSimon Topp
Simon Topp
LKL. E. Koenig
L. E. Koenig
MSM. J. Sleckman
M. J. Sleckman
AAAlison Appling
Alison Appling
Although groundwater discharge is a critical stream temperature control process, it is not explicitly represented in many stream temperature models, an omission that may reduce predictive accuracy, hinder management of aquatic habitat, and decrease user confidence. We assessed the performance of a previously‐described process‐guided deep learning model of stream temperature in the Delaware River Basin (USA). We found lower accuracy (root mean square error [RMSE] of 1.71 versus 1.35°C) and stronger seasonal bias (absolute mean monthly bias of 1.06 vs. 0.68°C) for reaches primarily influenced by deep groundwater as compared to atmospheric conditions. We then tested four approaches for improving groundwater process representation: (a) a custom loss function leveraging the unique patterns of air and water temperature coupling characteristic of different temperature drivers, (b) inclusion of additional groundwater‐relevant catchment attributes, (c) incorporation of additional process model outputs, and (d) a composite model. The custom loss function and the additional attributes significantly improved the predictive accuracy in groundwater‐dominated reaches (RMSE of 1.37 and 1.26°C) and reduced the seasonal bias (absolute mean monthly bias of 0.44 and 0.48°C), but neither approach could identify holdout groundwater reaches. Variable importance analysis indicates the custom loss function nudges the model to use the existing inputs more efficiently, whereas with the added features the model relies on a broader suite of inputs. This analysis is a substantial step toward more accurately representing groundwater discharge processes in stream temperature models and will improve predictive accuracy and inform habitat management.

3 November 2016 in Environmental Management
Verified
Journal Article
Closed access
PDF file
LDLindsay M. Dreiss
Lindsay M. Dreiss
JHJan-Michael Hessenauer
Jan-Michael Hessenauer
LNLucas R. Nathan
Lucas R. Nathan
KOKelly M. O’connor
Kelly M. O’connor
MLMarjorie R. Liberati
Marjorie R. Liberati
AMAnita T. Morzillo
Anita T. Morzillo

1 December 2016 in Journal of Environmental Management
Verified
Journal Article
Closed access
JBJanet R. Barclay
Janet R. Barclay
HTHannah Tripp
Hannah Tripp
CBChristopher J. Bellucci
Christopher J. Bellucci
GWGlenn Warner
Glenn Warner
AHAshley M. Helton
Ashley M. Helton

1 January 2020 in Open-File Report
Verified
Other
Closed access
PDF file
PLPamela J. Lombard
Pamela J. Lombard
JBJanet R. Barclay
Janet R. Barclay
DMDee-Ann E. McCarthy
Dee-Ann E. McCarthy
First posted December 22, 2020 Director, New England Water Science CenterU.S. Geological Survey10 Bearfoot RoadNorthborough, MA 01532 Below average and infrequent rainfall from May through September 2020 led to an extreme hydrologic drought across much of New England, with some areas experiencing a flash drought, reflecting its quick onset. The U.S. Geological Survey (USGS) recorded record-low streamflow and groundwater levels throughout the region. In September, the U.S. Department of Agriculture (2020) declared Aroostook County in Maine and Hillsborough and Merrimack Counties in New Hampshire as crop disaster areas. By the beginning of October, 166 community water systems and 5 municipalities in New Hampshire, more than 100 municipalities in Massachusetts, and several community water supplies in Connecticut, Maine, and Rhode Island had mandatory water restrictions in place.