The Lake-Catchment (LakeCat) Dataset: Characterizing landscape features for lake basins within the conterminous USA
Citation:
Hill, R., M. Weber, R. Debbout, S. Leibowitz, AND Tony Olsen. The Lake-Catchment (LakeCat) Dataset: Characterizing landscape features for lake basins within the conterminous USA. Freshwater Science. The Society for Freshwater Science, Springfield, IL, 37(2):208-221, (2018).
Impact/Purpose:
We developed an extensive dataset of landscape metrics for 378,088 lakes and their associated catchments within the conterminous USA: The Lake-Catchment (LakeCat) Dataset. This dataset summarizes nearby and upslope landscape features for these lakes, and includes both natural (e.g., climate, soils, and lithology) and anthropogenic (e.g., urbanization, agriculture, and dams) landscape features. The dataset will parallel and compliment a recently published dataset of watershed characteristics for streams, the StreamCat Dataset, in the numbers and types of watershed metrics. StreamCat is beginning to be widely used by university, government, and NGO researchers and managers and it is anticipated that LakeCat will provide similar benefits. At publication, the dataset will contain at least 170 landscape metrics, but more will be added as they become available. These data will be made available to the public for download and will greatly reduce the specialized geospatial expertise needed by researchers and managers to extract an extensive suite of watershed characteristics for lakes of interest. This paper provides a detailed description of the development and main features of LakeCat. The final publication will point to the URL home of the dataset (https://www.epa.gov/national-aquatic-resource-surveys/lakecat), which will also contain extensive metadata. All code used to develop the LakeCat Dataset will be made publically available through a GitHub website (https://github.com/USEPA/LakeCat) to ensure transparency of methods. In addition, the paper provides an illustration (with scripting code) of how LakeCat can be used. This illustration modeled eutrophication based on samples from the National Lakes Assessment. It then predicted the probability of eutrophication for 297,071 unsampled lakes across the conterminous US. The map of lake-specific predicted probabilities of eutrophication produced by this illustration could provide an important tool for states and managers to focus monitoring and sampling efforts. LakeCat data may also be of use to Office of Water for a number of applications, including modeling reference condition for National Lakes Assessment sample sites.
Description:
Natural and human-related landscape features influence the ecology and water quality within lakes. It is critical, therefore, to quantify landscape features in a hydrologically meaningful way to effectively manage these important ecosystems. Such summaries of the landscape are often done through the delineation of watershed boundaries of individual lakes. However, there are many technical challenges associated with delineating hundreds or thousands of lake watersheds at broad spatial extents (e.g., conterminous US). These difficulties can limit the application of such analyses to new, unsampled locations. In this paper, we present the development of a dataset of watershed features for 378,088 lakes within the conterminous US called the Lake-Catchment (LakeCat) Dataset (https://www.epa.gov/national-aquatic-resource-surveys/lakecat). We describe the methods and processes we used to (1) delineate a large number of lake catchments, (2) hydrologically connect nested lake catchments, and (3) quantify of over 200 geospatial metrics to produce full watershed summaries that describe both natural (e.g., soils, geology, climate, and land cover) and anthropogenic (e.g., urbanization, agriculture, and mines) features. In addition, we illustrate how this dataset can be used in an analysis that can provide ecological and management insights. For this illustration, we developed a random forest model to predict the probability of lake eutrophication by combining LakeCat with data from US EPA’s National Lakes Assessment (NLA). This model correctly predicted the trophic state of 72% of the NLA lakes and we applied the model to predict the probability of eutrophication at 297,071 unsampled lakes across the conterminous US, i.e., those lakes within the NLA’s sampling frame. The large suite of landscape metrics that are summarized within LakeCat could be used to improve analyses of lakes at broad spatial extents, improve the applicability of analyses to new, unsampled lakes, and ultimately improve the management of these important ecosystems.
