Geology topics

Samantha K. Oliver

Publications and source records attributed to Samantha K. Oliver.

36 records · Page 2Linked to original sources

USGS2021

Graph-based reinforcement learning for active learning in real time: An application in modeling river networks

Effective training of advanced ML models requires large amounts of labeled data, which is often scarce in scientific problems given the substantial human labor and material cost to collect labeled data. This poses a challenge on determining when and where we should deploy measuring instruments (e.g., in-situ sensors) to collect labeled data efficiently. This problem differs from traditional pool-based active learning settings in that the labeling decisions have to be made immediately after we observe the input data that come in a time series. In this paper, we develop a real-time active learning method that uses the spatial and temporal contextual information to select representative query samples in a reinforcement learning framework. To reduce the need for large training data, we further propose to transfer the policy learned from simulation data which is generated by existing physics-based models. We demonstrate the effectiveness of the proposed method by predicting streamflow and water temperature in the Delaware River Basin given a limited budget for collecting labeled data. We further study the spatial and temporal distribution of selected samples to verify the ability of this method in selecting informative samples over space and time.

Conference Paper↗

USGS2021

Inconsistent browning of northeastern U.S. lakes despite increased precipitation and recovery from acidification

Multiple studies have reported widespread browning of Northern Hemisphere lakes. Most examples are from boreal lakes that have experienced limited human influence, and browning has alternatively been attributed to changes in atmospheric deposition, climate, and land use. To determine the extent and possible causes of browning across a more geographically diverse region, we examined watercolor and dissolved organic carbon (DOC) time series in hundreds of northeastern U.S. lakes. The majority of lakes have increased in both DOC and color, but there were neither coherent spatial patterns in trends nor relationships with previously reported drivers. Color trends were more variable than DOC trends, and DOC and color trends were not strongly correlated, indicating a cause other than or in addition to increased loading of terrestrial carbon. Browning may be pronounced in regions where climate and atmospheric deposition are dominant drivers but muted in more human-dominated landscapes with a limited extent of organic soils where other disturbances predominate.

Connecticut, Massachusetts, Maine, Michigan, Minne↗

USGS2021

Partial differential equation driven dynamic graph networks for predicting stream water temperature

This paper presents a physics-guided machine learning approach that incorporates partial differential equations (PDEs) in a graph neural network model to improve the prediction of water temperature in river networks. The standard graph neural network model often uses pre-defined edge weights based on distance or similarity measures. Such static graph structure can be limited in capturing multiple processes in a physical system that interact and evolve over time. The limitation to represent underlying physical processes can severely affect the performance of the predictive model, especially when we have access to limited training data. To better capture the dynamic interactions among multiple segments in a river network, we built a dynamic graph model, where the graph structure is driven by the PDE that describes underlying physical processes. We further combine the dynamic graph structure and the recurrent layers to model temporal dependencies and improve the prediction. We demonstrate the effectiveness of the proposed method in a subnetwork of the Delaware River Basin. In particular, we show that the proposed method outperforms existing physics-based and machine learning models in temperature prediction using sparse observation data for training. The proposed method has also been shown to produce better performance when generalized to different seasons.

Conference Paper↗

USGS2021

Heterogeneous stream-reservoir graph networks with data assimilation

Accurate prediction of water temperature in streams is critical for monitoring and understanding biogeochemical and ecological processes in streams. Stream temperature is affected by weather patterns (such as solar radiation) and water flowing through the stream network. Additionally, stream temperature can be substantially affected by water releases from man-made reservoirs to downstream segments. In this paper, we propose a heterogeneous recurrent graph model to represent these interacting processes that underlie stream-reservoir networks and improve the prediction of water temperature in all river segments within a network. Because reservoir release data may be unavailable for certain reservoirs, we further develop a data assimilation mechanism to adjust the deep learning model states to correct for the prediction bias caused by reservoir releases. Our evaluation for the Delaware River Basin has demonstrated the superiority of our proposed method over multiple existing methods. We have extensively studied the effect of the data assimilation mechanism under different scenarios.

IEEE International Conference on Data Mining (ICDM↗

USGS2020

Exploring the exceptional performance of a deep learning stream temperature model and the value of streamflow data

Stream water temperature ( T s ) is a variable of critical importance for aquatic ecosystem health. T s is strongly affected by groundwater-surface water interactions which can be learned from streamflow records, but previously such information was challenging to effectively absorb with process-based models due to parameter equifinality. Based on the long short-term memory (LSTM) deep learning architecture, we developed a basin-centric lumped daily mean T s model, which was trained over 118 data-rich basins with no major dams in the conterminous United States, and showed strong results. At a national scale, we obtained a median root-mean-square error of 0.69°C, Nash–Sutcliffe model efficiency coefficient of 0.985, and correlation of 0.994, which are marked improvements over previous values reported in literature. The addition of streamflow observations as a model input strongly elevated the performance of this model. In the absence of measured streamflow, we showed that a two-stage model could be used, where simulated streamflow from a pre-trained LSTM model ( Q sim ) still benefited the T s model even though no new information was brought directly into the inputs of the T s model. The model indirectly used information learned from streamflow observations provided during the training of Q sim , potentially to improve internal representation of physically meaningful variables. Our results indicate that strong relationships exist between basin-averaged forcing variables, catchment attributes, and T s that can be simulated by a single model trained by data on the continental scale.

Environmental Research Letters↗

USGS2020

Primary sources of polycyclic aromatic hydrocarbons to streambed sediment in Great Lakes tributaries using multiple lines of evidence

Polycyclic aromatic hydrocarbons (PAHs) are among the most widespread and potentially toxic contaminants in Great Lakes (USA/Canada) tributaries. The sources of PAHs are numerous and diverse, and identifying the primary source(s) can be difficult. The present study used multiple lines of evidence to determine the likely sources of PAHs to surficial streambed sediments at 71 locations across 26 Great Lakes Basin watersheds. Profile correlations, principal component analysis, positive matrix factorization source‐receptor modeling, and mass fractions analysis were used to identify potential PAH sources, and land‐use analysis was used to relate streambed sediment PAH concentrations to different land uses. Based on the common conclusion of these analyses, coal‐tar–sealed pavement was the most likely source of PAHs to the majority of the locations sampled. The potential PAH‐related toxicity of streambed sediments to aquatic organisms was assessed by comparison of concentrations with sediment quality guidelines. The sum concentration of 16 US Environmental Protection Agency priority pollutant PAHs was 7.4–196 000 µg/kg, and the median was 2600 µg/kg. The threshold effect concentration was exceeded at 62% of sampling locations, and the probable effect concentration or the equilibrium partitioning sediment benchmark was exceeded at 41% of sampling locations. These results have important implications for watershed managers tasked with protecting and remediating aquatic habitats in the Great Lakes Basin.

Indiana, Michigan, Minnesota, New York, Ohio, Wisc↗

USGS2019

Process-guided deep learning predictions of lake water temperature

The rapid growth of data in water resources has created new opportunities to accelerate knowledge discovery with the use of advanced deep learning tools. Hybrid models that integrate theory with state‐of‐the art empirical techniques have the potential to improve predictions while remaining true to physical laws. This paper evaluates the Process‐Guided Deep Learning (PGDL) hybrid modeling framework with a use‐case of predicting depth‐specific lake water temperatures. The PGDL model has three primary components: a deep learning model with temporal awareness (long short‐term memory recurrence), theory‐based feedback (model penalties for violating conversation of energy), and model pretraining to initialize the network with synthetic data (water temperature predictions from a process‐based model). In situ water temperatures were used to train the PGDL model, a deep learning (DL) model, and a process‐based (PB) model. Model performance was evaluated in various conditions, including when training data were sparse and when predictions were made outside of the range in the training data set. The PGDL model performance (as measured by root‐mean‐square error (RMSE)) was superior to DL and PB for two detailed study lakes, but only when pretraining data included greater variability than the training period. The PGDL model also performed well when extended to 68 lakes, with a median RMSE of 1.65 °C during the test period (DL: 1.78 °C, PB: 2.03 °C; in a small number of lakes PB or DL models were more accurate). This case‐study demonstrates that integrating scientific knowledge into deep learning tools shows promise for improving predictions of many important environmental variables.

Minnesota, Wisconsin↗

USGS2019

Spatial and temporal variation of ecosystem properties at macroscales

Although spatial and temporal variation in ecological properties has been well-studied, crucial knowledge gaps remain for studies conducted at macroscales and for ecosystem properties related to material and energy. We test four propositions of spatial and temporal variation in ecosystem properties within a macroscale (1000 km's) extent. We fit Bayesian hierarchical models to thousands of observations from over two decades to quantify four components of variation – spatial (local and regional) and temporal (local and coherent); and to model their drivers. We found strong support for three propositions: (1) spatial variation at local and regional scales are large and roughly equal, (2) annual temporal variation is mostly local rather than coherent, and, (3) spatial variation exceeds temporal variation. Our findings imply that predicting ecosystem responses to environmental changes at macroscales requires consideration of the dominant spatial signals at both local and regional scales that may overwhelm temporal signals.

Ecology Letters↗

USGS2019

Widespread loss of lake ice around the Northern Hemisphere in a warming world

Ice provides a range of ecosystem services—including fish harvest , cultural traditions , transportation , recreation and regulation of the hydrological cycle —to more than half of the world’s 117 million lakes. One of the earliest observed impacts of climatic warming has been the loss of freshwater ice , with corresponding climatic and ecological consequences . However, while trends in ice cover phenology have been widely documented , a comprehensive large-scale assessment of lake ice loss is absent. Here, using observations from 513 lakes around the Northern Hemisphere, we identify lakes vulnerable to ice-free winters. Our analyses reveal the importance of air temperature, lake depth, elevation and shoreline complexity in governing ice cover. We estimate that 14,800 lakes currently experience intermittent winter ice cover, increasing to 35,300 and 230,400 at 2 and 8 °C, respectively, and impacting up to 394 and 656 million people. Our study illustrates that an extensive loss of lake ice will occur within the next generation, stressing the importance of climate mitigation strategies to preserve ecosystem structure and function, as well as local winter cultural heritage.

Nature Climate Change↗

USGS2018

Strategies for effective collaborative manuscript development in interdisciplinary science teams

Science is increasingly being conducted in large, interdisciplinary teams. As team size increases, challenges can arise during manuscript development, where achieving one team goal (e.g., inclusivity) may be in direct conflict with other goals (e.g., efficiency). Here, we present strategies for effective collaborative manuscript development that draw from our experiences in an interdisciplinary science team writing collaborative manuscripts for six years. These strategies are rooted in six guiding principles that were important to our team: to create a transparent, inclusive, and accountable research team that promotes and protects team members who have less power to influence decision‐making while fostering creativity and productivity. To help alleviate the conflicts that can arise in collaborative manuscript development, we present the following strategies: understand your team composition, create an authorship policy and discuss authorship early and often, openly announce manuscript ideas, identify and communicate the type of manuscript and lead author management style, and document and describe authorship contributions. These strategies can help reduce the probability of group conflict, uphold individual and team values, achieve fair authorship practices, and increase science productivity.

Ecosphere↗

USGS2018

Prediction of lake depth across a 17-state region in the United States

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 (Z max ) 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 Z max . 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 ( R 2 = 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.

Inland Waters↗

USGS2018

Consequences of gas flux model choice on the interpretation of metabolic balance across 15 lakes

Ecosystem metabolism and the contribution of carbon dioxide from lakes to the atmosphere can be estimated from free-water gas measurements through the use of mass balance models, which rely on a gas transfer coefficient ( k ) to model gas exchange with the atmosphere. Theoretical and empirically based models of k range in complexity from wind-driven power functions to complex surface renewal models; however, model choice is rarely considered in most studies of lake metabolism. This study used high-frequency data from 15 lakes provided by the Global Lake Ecological Observatory Network (GLEON) to study how model choice of k influenced estimates of lake metabolism and gas exchange with the atmosphere. We tested 6 models of k on lakes chosen to span broad gradients in surface area and trophic states; a metabolism model was then fit to all 6 outputs of k data. We found that hourly values for k were substantially different between models and, at an annual scale, resulted in significantly different estimates of lake metabolism and gas exchange with the atmosphere.

Inland Waters↗

USGS2017

Evidence for regional nitrogen stress on chlorophyll a in lakes across large landscape and climate gradients

Nitrogen (N) and phosphorus (P) commonly stimulate phytoplankton production in lakes, but recent observations from lakes from an agricultural region suggest that nitrate may have a subsidy‐stress effect on chlorophyll a (Chl a ). It is unclear, however, how generalizable this effect might be. Here, we analyzed a large water quality dataset of 2385 lakes spanning 60 regions across 17 states in the Northeastern and Midwestern U.S. to determine if N subsidy‐stress effects on phytoplankton are common and to identify regional landscape characteristics promoting N stress effects in lakes. We used a Bayesian hierarchical modeling framework to test our hypothesis that Chl a –total N (TN) threshold relationships would be common across the central agricultural region of the U.S. (“the Corn Belt”), where lake N and P concentrations are high. Data aggregated across all regions indicated that high TN concentrations had a negative effect on Chl a in lakes with concurrent high total P. This large‐scale pattern was driven by relationships within only a subset of regions, however. Eight regions were identified as having Chl a –TN threshold relationships, but only two of these regions located within the Corn Belt clearly demonstrated this subsidy‐stress relationship. N stress effects were not consistent across other intense agricultural regions, as we hypothesized. These findings suggest that interactions among regional land use and land cover, climate, and hydrogeology may be important in determining the synergistic conditions leading to N subsidy‐stress effects on lake phytoplankton.

Limnology and Oceanography↗

USGS2017

LAGOS-NE: a multi-scaled geospatial and temporal database of lake ecological context and water quality for thousands of US lakes

Understanding the factors that affect water quality and the ecological services provided by freshwater ecosystems is an urgent global environmental issue. Predicting how water quality will respond to global changes not only requires water quality data, but also information about the ecological context of individual water bodies across broad spatial extents. Because lake water quality is usually sampled in limited geographic regions, often for limited time periods, assessing the environmental controls of water quality requires compilation of many data sets across broad regions and across time into an integrated database. LAGOS-NE accomplishes this goal for lakes in the northeastern-most 17 US states. LAGOS-NE contains data for 51 101 lakes and reservoirs larger than 4 ha in 17 lake-rich US states. The database includes 3 data modules for: lake location and physical characteristics for all lakes; ecological context (i.e., the land use, geologic, climatic, and hydrologic setting of lakes) for all lakes; and in situ measurements of lake water quality for a subset of the lakes from the past 3 decades for approximately 2600–12 000 lakes depending on the variable. The database contains approximately 150 000 measures of total phosphorus, 200 000 measures of chlorophyll, and 900 000 measures of Secchi depth. The water quality data were compiled from 87 lake water quality data sets from federal, state, tribal, and non-profit agencies, university researchers, and citizen scientists. This database is one of the largest and most comprehensive databases of its type because it includes both in situ measurements and ecological context data. Because ecological context can be used to study a variety of other questions about lakes, streams, and wetlands, this database can also be used as the foundation for other studies of freshwaters at broad spatial and ecological scales.

Connecticut, Iowa, Illinois, Indiana, Massachusett↗

USGS2017

Unexpected stasis in a changing world: Lake nutrient and chlorophyll trends since 1990

The United States (U.S.) has faced major environmental changes in recent decades, including agricultural intensification and urban expansion, as well as changes in atmospheric deposition and climate—all of which may influence eutrophication of freshwaters. However, it is unclear whether or how water quality in lakes across diverse ecological settings has responded to environmental change. We quantified water quality trends in 2913 lakes using nutrient and chlorophyll (Chl) observations from the Lake Multi-Scaled Geospatial and Temporal Database of the Northeast U.S. (LAGOS-NE), a collection of preexisting lake data mostly from state agencies. LAGOS-NE was used to quantify whether lake water quality has changed from 1990 to 2013, and whether lake-specific or regional geophysical factors were related to the observed changes. We modeled change through time using hierarchical linear models for total nitrogen (TN), total phosphorus (TP), stoichiometry (TN:TP), and Chl. Both the slopes (percent change per year) and intercepts (value in 1990) were allowed to vary by lake and region. Across all lakes, TN declined at a rate of 1.1% year −1 , while TP, TN:TP, and Chl did not change. A minority (7%–16%) of individual lakes had changing nutrients, stoichiometry, or Chl. Of those lakes that changed, we found differences in the geospatial variables that were most related to the observed change in the response variables. For example, TN and TN:TP trends were related to region-level drivers associated with atmospheric deposition of N; TP trends were related to both lake and region-level drivers associated with climate and land use; and Chl trends were found in regions with high air temperature at the beginning of the study period. We conclude that despite large environmental change and management efforts over recent decades, water quality of lakes in the Midwest and Northeast U.S. has not overwhelmingly degraded or improved.

Global Change Biology↗

USGS2017

Lake nutrient stoichiometry is less predictable than nutrient concentrations at regional and sub-continental scales

Production in many ecosystems is co-limited by multiple elements. While a known suite of drivers associated with nutrient sources, nutrient transport, and internal processing controls concentrations of phosphorus (P) and nitrogen (N) in lakes, much less is known about whether the drivers of single nutrient concentrations can also explain spatial or temporal variation in lake N:P stoichiometry. Predicting stoichiometry might be more complex than predicting concentrations of individual elements because some drivers have similar relationships with N and P, leading to a weak relationship with their ratio. Further, the dominant controls on elemental concentrations likely vary across regions, resulting in context dependent relationships between drivers, lake nutrients and their ratios. Here, we examine whether known drivers of N and P concentrations can explain variation in N:P stoichiometry, and whether explaining variation in stoichiometry differs across regions. We examined drivers of N:P in ~2,700 lakes at a sub-continental scale and two large regions nested within the sub-continental study area that have contrasting ecological context, including differences in the dominant type of land cover (agriculture vs. forest). At the sub-continental scale, lake nutrient concentrations were correlated with nutrient loading and lake internal processing, but stoichiometry was only weakly correlated to drivers of lake nutrients. At the regional scale, drivers that explained variation in nutrients and stoichiometry differed between regions. In the Midwestern U.S. region, dominated by agricultural land use, lake depth and the percentage of row crop agriculture were strong predictors of stoichiometry because only phosphorus was related to lake depth and only nitrogen was related to the percentage of row crop agriculture. In contrast, all drivers were related to N and P in similar ways in the Northeastern U.S. region, leading to weak relationships between drivers and stoichiometry. Our results suggest ecological context mediates controls on lake nutrients and stoichiometry. Predicting stoichiometry was generally more difficult than predicting nutrient concentrations, but human activity may decouple N and P, leading to better prediction of N:P stoichiometry in regions with high anthropogenic activity.

Ecological Applications↗

USGS2015

The ecology of methane in streams and rivers: Patterns, controls, and global significance

Streams and rivers can substantially modify organic carbon (OC) inputs from terrestrial landscapes, and much of this processing is the result of microbial respiration. While carbon dioxide (CO 2 ) is the major end-product of ecosystem respiration, methane (CH 4 ) is also present in many fluvial environments even though methanogenesis typically requires anoxic conditions that may be scarce in these systems. Given recent recognition of the pervasiveness of this greenhouse gas in streams and rivers, we synthesized existing research and data to identify patterns and drivers of CH 4 , knowledge gaps, and research opportunities. This included examining the history of lotic CH 4 research, creating a database of concentrations and fluxes (MethDB) to generate a global-scale estimate of fluvial CH 4 efflux, and developing a conceptual framework and using this framework to consider how human activities may modify fluvial CH 4 dynamics. Current understanding of CH 4 in streams and rivers has been strongly influenced by goals of understanding OC processing and quantifying the contribution of CH 4 to ecosystem C fluxes. Less effort has been directed towards investigating processes that dictate in situ CH 4 production and loss. CH 4 makes a meager contribution to watershed or landscape C budgets, but streams and rivers are often significant CH 4 sources to the atmosphere across these same spatial extents. Most fluvial systems are supersaturated with CH 4 and we estimate an annual global emission of 26.8 Tg CH 4 , equivalent to ~15-40% of wetland and lake effluxes, respectively. Less clear is the role of CH 4 oxidation, methanogenesis, and total anaerobic respiration to whole ecosystem production and respiration. Controls on CH 4 generation and persistence can be viewed in terms of proximate controls that influence methanogenesis (organic matter, temperature, alternative electron acceptors, nutrients) and distal geomorphic and hydrologic drivers. Multiple controls combined with its extreme redox status and low solubility result in high spatial and temporal variance of CH 4 in fluvial environments, which presents a substantial challenge for understanding its larger-scale dynamics. Further understanding of CH 4 production and consumption, anaerobic metabolism, and ecosystem energetics in streams and rivers can be achieved through more directed studies and comparison with knowledge from terrestrial, wetland, and aquatic disciplines.

Ecological Monographs↗

USGS2015

Building a multi-scaled geospatial temporal ecology database from disparate data sources: Fostering open science through data reuse

Although there are considerable site-based data for individual or groups of ecosystems, these datasets are widely scattered, have different data formats and conventions, and often have limited accessibility. At the broader scale, national datasets exist for a large number of geospatial features of land, water, and air that are needed to fully understand variation among these ecosystems. However, such datasets originate from different sources and have different spatial and temporal resolutions. By taking an open-science perspective and by combining site-based ecosystem datasets and national geospatial datasets, science gains the ability to ask important research questions related to grand environmental challenges that operate at broad scales. Documentation of such complicated database integration efforts, through peer-reviewed papers, is recommended to foster reproducibility and future use of the integrated database. Here, we describe the major steps, challenges, and considerations in building an integrated database of lake ecosystems, called LAGOS (LAke multi-scaled GeOSpatial and temporal database), that was developed at the sub-continental study extent of 17 US states (1,800,000 km 2 ). LAGOS includes two modules: LAGOS GEO , with geospatial data on every lake with surface area larger than 4 ha in the study extent (~50,000 lakes), including climate, atmospheric deposition, land use/cover, hydrology, geology, and topography measured across a range of spatial and temporal extents; and LAGOS LIMNO , with lake water quality data compiled from ~100 individual datasets for a subset of lakes in the study extent (~10,000 lakes). Procedures for the integration of datasets included: creating a flexible database design; authoring and integrating metadata; documenting data provenance; quantifying spatial measures of geographic data; quality-controlling integrated and derived data; and extensively documenting the database. Our procedures make a large, complex, and integrated database reproducible and extensible, allowing users to ask new research questions with the existing database or through the addition of new data. The largest challenge of this task was the heterogeneity of the data, formats, and metadata. Many steps of data integration need manual input from experts in diverse fields, requiring close collaboration.

Connecticut, Delaware, Illinois, Indiana, Iowa, Ma↗