Dataset Information

Using Knowledge-Guided Machine Learning To Assess Patterns of Areal Change in Waterbodies across the Contiguous United States.

ABSTRACT: Lake and reservoir surface areas are an important proxy for freshwater availability. Advancements in machine learning (ML) techniques and increased accessibility of remote sensing data products have enabled the analysis of waterbody surface area dynamics on broad spatial scales. However, interpreting the ML results remains a challenge. While ML provides important tools for identifying patterns, the resultant models do not include mechanisms. Thus, the "black-box" nature of ML techniques often lacks ecological meaning. Using ML, we characterized temporal patterns in lake and reservoir surface area change from 1984 to 2016 for 103,930 waterbodies in the contiguous United States. We then employed knowledge-guided machine learning (KGML) to classify all waterbodies into seven ecologically interpretable groups representing distinct patterns of surface area change over time. Many waterbodies were classified as having "no change" (43%), whereas the remaining 57% of waterbodies fell into other groups representing both linear and nonlinear patterns. This analysis demonstrates the potential of KGML not only for identifying ecologically relevant patterns of change across time but also for unraveling complex processes that underpin those changes.

SUBMITTER: Wander HL

PROVIDER: S-EPMC10956424 | biostudies-literature | 2024 Mar

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Using Knowledge-Guided Machine Learning To Assess Patterns of Areal Change in Waterbodies across the Contiguous United States.

Wander Heather L HL Farruggia Mary Jade MJ La Fuente Sofia S Korver Maartje C MC Chapina Rosaura J RJ Robinson Jenna J Bah Abdou A Munthali Elias E Ghosh Rahul R Stachelek Jemma J Khandelwal Ankush A Hanson Paul C PC Weathers Kathleen C KC

Environmental science & technology 20240306 11

Lake and reservoir surface areas are an important proxy for freshwater availability. Advancements in machine learning (ML) techniques and increased accessibility of remote sensing data products have enabled the analysis of waterbody surface area dynamics on broad spatial scales. However, interpreting the ML results remains a challenge. While ML provides important tools for identifying patterns, the resultant models do not include mechanisms. Thus, the "black-box" nature of ML techniques often la ...[more]

PMID: 38446785

Similar Datasets

Project description:Wetlands provide key functions in the landscape from improving water quality, to regulating flows, to providing wildlife habitat. Over half of the wetlands in the contiguous United States (CONUS) have been converted to agricultural and urban land uses. However, over the last several decades, research has shown the benefits of wetlands to hydrologic, chemical, biological processes, spurring the creation of government programs and private initiatives to restore wetlands. Initiatives tend to focus on individual wetland creation, yet the greatest benefits are achieved when strategic restoration planning occurs across a watershed or multiple watersheds. For watershed-level wetland restoration planning to occur, informative data layers on potential wetland areas are needed. We created an indicator of potential wetland areas (PWA), using nationally available datasets to identify characteristics that could support wetland ecosystems, including: poorly drained soils and low-relief landscape positions as indicated by a derived topographic data layer. We compared our PWA with the National Wetlands Inventory (NWI) from 11 states throughout the CONUS to evaluate their alignment. The state-level percentage of NWI-designated wetlands directly overlapping the PWA ranged from 39 to 95%. When we included NWI that was immediately adjacent to the overlapping NWI, our range of correspondence to NWI ranged from 60 to 99%. Wetland restoration is more likely on certain landscapes (e.g., agriculture) than others due to the lack of substantive infrastructure and the potential for the restoration of hydrology; therefore, we combined the National Land Cover Dataset (NLCD) with the PWA to identify potentially restorable wetlands on agricultural land (PRW-Ag). The PRW-Ag identified a total of over 46 million ha with the potential to support wetlands. The largest concentrations of PRW-Ag occurred in the glaciated corn belt of the upper Mississippi River from Ohio to the Dakotas and in the Mississippi Alluvial Valley. The PRW-Ag layer could assist land managers in identifying sites that may qualify for enrollment in conservation programs, where planners can coordinate restoration efforts, or where decision makers can target resources to optimize the services provided across a watershed or multiple watersheds.

Dataset Information

Using Knowledge-Guided Machine Learning To Assess Patterns of Areal Change in Waterbodies across the Contiguous United States.

Publications

Using Knowledge-Guided Machine Learning To Assess Patterns of Areal Change in Waterbodies across the Contiguous United States.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets