An official website of the United States government
Abstract Most environmental data come from a minority of well‐monitored sites. An ongoing challenge in the environmental sciences is transferring knowledge from monitored sites to unmonitored sites. Here, we demonstrate a novel transfer‐learning framework that accurately predicts depth‐specific temperature in unmonitored lakes (targets) by borrowing models from well‐monitored lakes (sources). This method, meta‐transfer learning (MTL), builds a meta‐learning model to predict transfer performance from candidate source models to targets using lake attributes and candidates' past performance. We constructed source models at 145 well‐monitored lakes using calibrated process‐based (PB) modeling and a recently developed approach called process‐guided deep learning (PGDL). We applied MTL to either PB or PGDL source models (PB‐MTL or PGDL‐MTL, respectively) to predict temperatures in 305 target lakes treated as unmonitored in the Upper Midwestern United States. We show significantly improved performance relative to the uncalibrated PB General Lake Model, where the median root mean squared error (RMSE) for the target lakes is 2.52°C. PB‐MTL yielded a median RMSE of 2.43°C; PGDL‐MTL yielded 2.16°C; and a PGDL‐MTL ensemble of nine sources per target yielded 1.88°C. For sparsely monitored target lakes, PGDL‐MTL often outperformed PGDL models trained on the target lakes themselves. Differences in maximum depth between the source and target were consistently the most important predictors. Our approach readily scales to thousands of lakes in the Midwestern United States, demonstrating that MTL with meaningful predictor variables and high‐quality source models is a promising approach for many kinds of unmonitored systems and environmental variables.
more »
« less
The distribution of depth, volume, and basin shape for lakes in the conterminous United States
Abstract Depth regulates many attributes of aquatic ecosystems, but relatively few lakes are measured, and existing datasets are biased toward large lakes. To address this, we used a large dataset of maximum (Zmax;n = 16,831) and mean (Zmean;n = 5,881) depth observations to create new depth models, focusing on lakes < 1,000 ha. We then used the models to characterize patterns in lake basin shape and volume. We included terrain metrics, water temperature and reflectance, polygon attributes, and other predictors in a random forest model. Our final models generally outperformed existing models (Zmax; root mean square error [RMSE] = 8.0 m andZmean; RMSE = 3.0 m). Our models show that lake depth followed a Pareto distribution, with 2.8 orders of magnitude fewer lakes for an order of magnitude increase in depth. In addition, despite orders of magnitude variation in surface area, most size classes had a modal maximum depth of ~ 5 m. Concave (bowl‐shaped) lake basins represented 79% of all lakes, but lakes were more convex (funnel‐shaped) as surface area increased. Across the conterminous United States, 9.8% of all lake water was within the top meter of the water column, and 48% in the top 10 m. Excluding the Laurentian Great Lakes, we estimate the total volume in the conterminous United States is 1,057–1,294 km3, depending on whetherZmaxorZmeanwas modeled. Lake volume also exhibited substantial geographic variation, with high volumes in the upper Midwest, Northeast, and Florida and low volumes in the southwestern United States.
more »
« less
- Award ID(s):
- 2048031
- PAR ID:
- 10480810
- Publisher / Repository:
- Wiley Blackwell (John Wiley & Sons)
- Date Published:
- Journal Name:
- Limnology and Oceanography
- ISSN:
- 0024-3590
- Format(s):
- Medium: X
- Sponsoring Org:
- National Science Foundation
More Like this
-
-
Lake depth is an important characteristic for understanding many lake processes, yet it is unknown for the vast majority of lakes globally. Our objective was to develop a model that predicts lake depth using map-derived metrics of lake and terrestrial geomorphic features. Building on previous models that use local topography to predict lake depth, we hypothesized that regional differences in topography, lake shape, or sedimentation processes could lead to region-specific relationships between lake depth and the mapped features. We therefore used a mixed modeling approach that included region-specific model parameters. We built models using lake and map data from LAGOS, which includes 8164 lakes with maximum depth (Zmax) observations. The model was used to predict depth for all lakes ≥4 ha (n = 42 443) in the study extent. Lake surface area and maximum slope in a 100 m buffer were the best predictors of Zmax. Interactions between surface area and topography occurred at both the local and regional scale; surface area had a larger effect in steep terrain, so large lakes embedded in steep terrain were much deeper than those in flat terrain. Despite a large sample size and inclusion of regional variability, model performance (R2 = 0.29, RMSE = 7.1 m) was similar to other published models. The relative error varied by region, however, highlighting the importance of taking a regional approach to lake depth modeling. Additionally, we provide the largest known collection of observed and predicted lake depth values in the United States.more » « less
-
Abstract As ice sheets load Earth's surface, they produce ice‐marginal depressions which, when filled with meltwater, become proglacial lakes. We include self‐consistently evolving proglacial lakes in a glacial isostatic adjustment (GIA) model and apply it to the Laurentide ice sheet over the last glacial cycle. We find that the locations of modeled lakes and the timing of their disappearance is consistent with the geological record. Lake loads can deflect topography by >10 m, and volumes collectively approach 30–45 cm global mean sea‐level equivalent. GIA increases deglaciation‐phase lake volume up to five‐fold and average along‐ice‐margin depth ≤90 m compared to glaciation‐phase ice volume analogs—differences driven by changes in the position and size of the peripheral bulge. Since ice‐marginal lake depth affects grounding‐line outflow, GIA‐modulated proglacial lake depths could affect ice‐sheet mass loss. Indeed, we find that Laurentide ice‐margin retreat rate sometimes correlates with proglacial lake presence, indicating that proglacial lakes aid glacial collapse.more » « less
-
Abstract This study evaluates the methods of identifying the heightziof the top of the convective boundary layer (CBL) during winter (December and January) over the Great Lakes and nearby land areas using observations taken by the University of Wyoming King Air research aircraft during the Lake-Induced Convection Experiment (1997/98) and Ontario Winter Lake-effect Systems (2013/14) field campaigns. Since CBLs facilitate vertical mixing near the surface, the most direct measurement ofziis that above which the vertical velocity turbulent fluctuations are weak or absent. Thus, we usezifrom the turbulence method as the “reference value” to whichzifrom other methods, based on bulk Richardson number (Rib), liquid water content, and vertical gradients of potential temperature, relative humidity, and water vapor mixing ratio, are compared. The potential temperature gradient method using a threshold value of 0.015 K m−1for soundings over land and 0.011 K m−1for soundings over lake provided the estimates ofzithat are most consistent with the turbulence method. The Ribthreshold-based method, commonly used in numerical simulation studies, underestimatedzi. Analyzing the methods’ performance on the averaging windowzavgwe recommend usingzavg= 20 or 50 m forziestimations for lake-effect boundary layers. The present dataset consists of both cloudy and cloud-free boundary layers, some having decoupled boundary layers above the inversion top. Because cases of decoupled boundary layers appear to be formed by nearby synoptic storms, we recommend use of the more general term, elevated mixed layers. Significance StatementThe depthziof the convective atmospheric boundary layer (CBL) strongly influences precipitation rates during lake-effect snowstorms (LES). However, variousziapproximation methods produce significantly different results. This study utilizes extensive concurrently collected observations by project aircraft during two LES field studies [Lake-Induced Convection Experiment (Lake-ICE) and OWLeS] to assess howzifrom common estimation methods compare with “reference”ziderived from turbulent fluctuations, a direct measure of CBL mixing. For soundings taken both over land and lake; with cloudy or cloud-free conditions, potential temperature gradient (PTG) methods provided the best agreement with the referencezi. A method commonly employed in numerical simulations performed relatively poorly. Interestingly, the PTG method worked equally well for “coupled” and elevated decoupled CBLs, commonly associated with nearby cyclones.more » « less
-
The LAGOS-US LAKE DEPTH v1.0 module (hereafter, called DEPTH) contains in situ measurements of lake depth for a subset of all lakes (n = 17,675) in the conterminous U.S. > 1 ha (3.7% of 479,950) that are in the LAGOS-US LOCUS v1.0 data module (Smith et al. 2021). All 17,675 lakes in DEPTH have a maximum depth value and 6,137 lakes have a mean depth. DEPTH includes approximately 65 data sources obtained from community, government, and university monitoring programs, as well as academic reports and commercial websites. DEPTH includes lake identifiers, lake location, lake area, lake depth (both maximum and mean depth when available), source information, and data flags. The unique lake identifier (lagoslakeid) for all lakes is the same one used in LAGOS-US LOCUS v1.0.more » « less
