Limnology and Oceanography Bulletin Volume 31 Number 1 November 2021 1–29

Depth is a fundamental property of lakes, essential for understanding and quantifying a range of biological, chemical, and physical processes of individual lakes as well as the collective functioning of the millions of lakes across the globe. Despite this importance, lake depth is rarely available for large numbers of lakes across the broad scales of regions and continents. We describe a new open-access dataset, LAGOS-US DEPTH, which contains 17,675 maximum depths and 6,137 mean depths for lakes in the conterminous United States. These data represent one of the largest compilations of lake depths in the world and are connected to other data products through the LAGOS-US research platform. Here, we describe characteristics of lake maximum depth across the conterminous United States and identify gaps in data coverage that we encourage other researchers to fill by linking their own depth datasets for lakes within the LAGOS-US spatial footprint. Extending this open-access effort to include other regions and countries across the globe to build a more comprehensive and representative database of lake depth would better position scientists to quantify and articulate the critical roles that lakes play globally in biogeochemical cycling, maintaining biodiversity, and in maintaining the many ecosystem services that lakes provide. Getting our feet wet The depth of a lake is one of its most fundamental properties informing scientists on the biology, chemistry, and physical properties of lakes across the world, as well as anglers searching for the best place to fish. When scientists study individual lakes, lake depth is probably one of the first pieces of information acquired, and is relatively straightforward to measure. However, freshwater scientists are increasingly studying hundreds to millions of lakes across regions, continents, and the globe to examine such pressing issues as the estimated contribution of lakes to global carbon cycles or the projected trajectory of lake eutrophication and subsequent effects on the ecosystem services that lakes provide. Although lake depth is known for the world’s largest (and often deepest) lakes, these lakes represent a tiny fraction of lakes globally. One might think that lake depth could be estimated from the depth of nearby lakes, from a lake’s origin (e.g., glacial vs. volcanic), or from the nearby terrain of the land. However, none of these factors has been shown to accurately predict lake depth; in fact, models designed to predict lake depth from these types of factors have generated lake depth predictions with very large uncertainties that do little to improve our understanding of lake depth or to inform the use of lake depth to predict other important factors (Oliver et al. 2016; Stachelek et al. In press). Thus, we are left with having to measure lake depth directly rather than predicting it from easily acquired data sources. Perhaps ironically, models can more easily predict lake ecosystem properties such as water temperature and water clarity than depth itself, but depth is often a key predictor variable that limits the scope and accuracy of predictions (McCullough et al. 2012; Willard et al. 2021). One reason for the lack of comprehensive lake depth data is that current and emerging technologies do not allow for the remote measurement of lake depth using satellite, aerial, or drone sensors; thus, lake depth can only be quantified by going out to the lake and taking measurements in a boat, which can be labor-intensive and requires direct access to the lake that is not always available. In addition, when lake depth measurements are taken, they are often done on an individual lake or on a regional basis by government agencies, researchers, private entities such as fishing groups, or individual lake property owners. Thus, the data that do exist are often stored in disparate places and not always easily accessible. Therefore, to increase the availability of lake depth data for scientists (and lake users) at broader spatial scales, datasets need to be compiled from a variety of different sources, an effort now more feasible thanks to web-based access to data and increasing open data practices by scientists and government agencies. In this article, we describe our approach to compile lake depth data into a nationalscale, publicly accessible database, LAGOSUS DEPTH, with maximum and mean depth data for, respectively, 17,675 and 6,137 lakes and reservoirs across the lower 48 U.S. states (Stachelek et al. 2021). We believe this is the largest compilation of maximum lake depth, particularly across this large of a geographic extent. We chose this footprint to match LAGOS-US LOCUS, another open-access dataset of the location, © 2022 The Authors. Limnology and Oceanography Bulletin published by Wiley Periodicals LLC on behalf of Association for the Sciences of Limnology and Oceangraphy. february 2022 1 identifiers, and physical characteristics of 479,950 lakes greater than 1 ha in surface area and their watersheds in the conterminous United States that our research group has compiled (Cheruvelil et al. 2021; Smith et al. 2021). Our dataset provides maximum depth data on 3.7% of the lakes in this footprint, which is still a small fraction, but far better than any other available source of lake depths at this broad of a spatial extent. We use principles of open science to make these data available (Soranno et al. 2015), so any researcher can use our data to study this important lake property at an unprecedented spatial scale. We also hope this work will encourage scientists in other regions of the world to do the same, particularly for smaller lakes often overlooked in global datasets. Until new technologies emerge for the remote measurement of lake depth, these types of coordinated and open access compilation efforts will be the only way to improve continental and global estimates of important lake properties that rely on lake depth data, such as carbon and eutrophication, to name a few. Uncovering hidden depths: the LAGOS-US DEPTH database Our database contains depth data for an unprecedented number of lakes compiled manually from web data sources (e.g., state agency websites, fishing map databases). To focus our search, we first selected those lakes of the 479,950 LAGOS-US LOCUS lake population that had a known water quality sample taken at any point in the last 30 years. Then, for each lake in this subset, we searched manually for lake depth data. As a result, our sample selection was not random but is based largely on the population of lakes sampled for water quality. We describe more detailed methods in the LAGOS-US DEPTH User Guide that is available on EDI along with the database and associated metadata (Stachelek et al. 2021). For fuller use of this database, we recommend downloading and connecting to our research team’s recent data product, LAGOS-US LOCUS (Cheruvelil et al. 2021; Smith et al. 2021), which enables the linking of these lake depth data to source datasets via the common unique lake identifier, lagoslakeid, includes Geographic Information System (GIS) layers with lake polygons, and provides data on geometry, location, and other lake properties. The depth dataset includes data for both lakes and reservoirs representing a wide range of environmental conditions across the conterminous United States and across a broad surface area range. Here, we focus on maximum depth to describe some highlights of the patterns of lake depth across broad scales. In addition, we give an example of how connecting LAGOS-US DEPTH with other LAGOS-US data products (in this case a dataset on reservoir classification) provides a powerful source of data for addressing a wide variety of research questions. For example, these data would serve as a useful training dataset for future models predicting depth for the remaining 96.3% lakes in the conterminous United States that are not included in this data product. Getting to the bottom of lake depth data availability We were not surprised to find that the lakes that have been sampled for water quality and also have a lake depth measure are skewed towards larger lakes (Fig. 1), a bias previously documented by Stanley et al. (2019). The median surface area of lakes with depth values is 50 ha, vs. the median area of all lakes in LAGOS-US LOCUS which is 5 ha. Nevertheless, the lakes with depth values ranged in surface area from 1 to 338,950 ha, a span nearly extending to the full lake area range in LAGOS-US. Such a broad gradient will be valuable for future use of the dataset for building predictive models of lake depth across all size classes. Furthermore, above the size range of 100 ha, we have a very good representation of lake depth values. Our dataset is actually quite complete for larger lakes and less so for moderately sized lakes; small lakes are very undersampled. Regional differences in the number and percentage of lakes with lake depth data were related to the variation among U.S. states in both the absolute number of lakes and the availability of depth data (Fig. 2). For example, lake-dense Minnesota has the greatest number of lakes with maximum depth of any state, but Maine and New Hampshire have the greatest percent of their lake populations with lake maximum depth at 31%. All of these states have very accessible data for their lakes (e.g., http://www.dnr.state.mn.us/lakefind/ for Minnesota, and https://www.maine.gov/dep/ FIG. 1. The kernal density distribution of maximum depth values by lake surface area for lakes lacking maximum depth (blue color) and for lakes with maximum depth values (sand color). Vertical lines indicate median values for each group. FIG. 2. Maps of the 48 U.S. states that depict (a) the number of LAGOS-US lakes in each state (plotted on log10 scale); (b) the number of lakes with maximum depth values in each state (plotted on log10 scale); and (c) the percent of LAGOS-US lakes in a state with maximum depth values i