Deep learning for water quality.

Understanding and predicting the quality of inland waters are challenging, particularly in the context of intensifying climate extremes expected in the future. These challenges arise partly due to complex processes that regulate water quality, and arduous and expensive data collection that exacerbate the issue of data scarcity. Traditional process-based and statistical models often fall short in predicting water quality. In this Review, we posit that deep learning represents an underutilized yet promising approach that can unravel intricate structures and relationships in high-dimensional data. We demonstrate that deep learning methods can help address data scarcity by filling temporal and spatial gaps and aid in formulating and testing hypotheses via identifying influential drivers of water quality. This Review highlights the strengths and limitations of deep learning methods relative to traditional approaches, and underscores its potential as an emerging and indispensable approach in overcoming challenges and discovering new knowledge in water-quality sciences.

Improving ecosystem health in highly altered river basins: a generalized framework and its application to the Mississippi-Atchafalaya River Basin.

Continued large-scale public investment in declining ecosystems depends on demonstrations of "success". While the public conception of "success" often focuses on restoration to a pre-disturbance condition, the scientific community is more likely to measure success in terms of improved ecosystem health. Using a combination of literature review, workshops and expert solicitation we propose a generalized framework to improve ecosystem health in highly altered river basins by reducing ecosystem stressors, enhancing ecosystem processes and increasing ecosystem resilience. We illustrate the use of this framework in the Mississippi-Atchafalaya River Basin (MARB) of the central United States (U.S.), by (i) identifying key stressors related to human activities, and (ii) creating a conceptual ecosystem model relating those stressors to effects on ecosystem structure and processes. As a result of our analysis, we identify a set of landscape-level indicators of ecosystem health, emphasizing leading indicators of stressor removal (e.g., reduced anthropogenic nutrient inputs), increased ecosystem function (e.g., increased water storage in the landscape) and increased resilience (e.g., changes in the percentage of perennial vegetative cover). We suggest that by including these indicators, along with lagging indicators such as direct measurements of water quality, stakeholders will be better able to assess the effectiveness of management actions. For example, if both leading and lagging indicators show improvement over time, then management actions are on track to attain desired ecosystem condition. If, however, leading indicators are not improving or even declining, then fundamental challenges to ecosystem health remain to be addressed and failure to address these will ultimately lead to declines in lagging indicators such as water quality. Although our model and indicators are specific to the MARB, we believe that the generalized framework and the process of model and indicator development will be valuable in an array of altered river basins.

Climate Change Will Impact Surface Water Extents and Dynamics Across the Central United States.

Climate change is projected to impact river, lake, and wetland hydrology, with global implications for the condition and productivity of aquatic ecosystems. We integrated Sentinel-1 and Sentinel-2 based algorithms to track monthly surface water extent (2017-2021) for 32 sites across the central United States (U.S.). Median surface water extent was highly variable across sites, ranging from 3.9% to 45.1% of a site. To account for landscape-based differences (e.g., water storage capacity, land use) in the response of surface water extents to meteorological conditions, individual statistical models were developed for each site. Future changes to climate were defined as the difference between 2006-2025 and 2061-2080 using MACA-CMIP5 (MACAv2-METDATA) Global Circulation Models. Time series of climate change adjusted surface water extents were projected. Annually, 19 of the 32 sites under RCP4.5 and 22 of the 32 sites under RCP8.5 were projected to show an average decline in surface water extent, with drying most consistent across the southeast central, southwest central, and midwest central U.S. Projected declines under surface water dry conditions at these sites suggest greater impacts of drought events are likely in the future. Projected changes were seasonally variable, with the greatest decline in surface water extent expected in summer and fall seasons. In contrast, many north central sites showed a projected increase in surface water in most seasons, relative to the 2017-2021 period, likely attributable to projected increases in winter and spring precipitation exceeding increases in projected temperature.

Wetland Flowpaths Mediate Nitrogen and Phosphorus Concentrations across the Upper Mississippi River Basin.

Eutrophication, harmful algal blooms, and human health impacts are critical environmental challenges resulting from excess nitrogen and phosphorus in surface waters. Yet we have limited information regarding how wetland characteristics mediate water quality across watershed scales. We developed a large, novel set of spatial variables characterizing hydrological flowpaths from wetlands to streams, that is, "wetland hydrological transport variables," to explore how wetlands statistically explain the variability in total nitrogen (TN) and total phosphorus (TP) concentrations across the Upper Mississippi River Basin (UMRB) in the United States. We found that wetland flowpath variables improved landscape-to-aquatic nutrient multilinear regression models (from R2 = 0.89 to 0.91 for TN; R2 = 0.53 to 0.84 for TP) and provided insights into potential processes governing how wetlands influence watershed-scale TN and TP concentrations. Specifically, flowpath variables describing flow-attenuating environments, for example, subsurface transport compared to overland flowpaths, were related to lower TN and TP concentrations. Frequent hydrological connections from wetlands to streams were also linked to low TP concentrations, which likely suggests a nutrient source limitation in some areas of the UMRB. Consideration of wetland flowpaths could inform management and conservation activities designed to reduce nutrient export to downstream waters.

Mapping global non-floodplain wetlands.

Non-floodplain wetlands - those located outside the floodplains - have emerged as integral components to watershed resilience, contributing hydrologic and biogeochemical functions affecting watershed-scale flooding extent, drought magnitude, and water-quality maintenance. However, the absence of a global dataset of non-floodplain wetlands limits their necessary incorporation into water quality and quantity management decisions and affects wetland-focused wildlife habitat conservation outcomes. We addressed this critical need by developing a publicly available "Global NFW" (Non-Floodplain Wetland) dataset, comprised of a global river-floodplain map at 90 m resolution coupled with a global ensemble wetland map incorporating multiple wetland-focused data layers. The floodplain, wetland, and non-floodplain wetland spatial data developed here were successfully validated within 21 large and heterogenous basins across the conterminous United States. We identified nearly 33 million potential non-floodplain wetlands with an estimated global extent of over 16×106 km2. Non-floodplain wetland pixels comprised 53% of globally identified wetland pixels, meaning the majority of the globe's wetlands likely occur external to river floodplains and coastal habitats. The identified global NFWs were typically small (median 0.039 km2), with a global median size ranging from 0.018-0.138 km2. This novel geospatial Global NFW static dataset advances wetland conservation and resource-management goals while providing a foundation for global non-floodplain wetland functional assessments, facilitating non-floodplain wetland inclusion in hydrological, biogeochemical, and biological model development. The data are freely available through the United States Environmental Protection Agency's Environmental Dataset Gateway (https://gaftp.epa.gov/EPADataCommons/ORD/Global_NonFloodplain_Wetlands/, last access: 24 May 2023) and through https://doi.org/10.23719/1528331 (Lane et al., 2023a).

Human alterations of the global floodplains 1992-2019.

Floodplains provide critical ecosystem services; however, loss of natural floodplain functions caused by human alterations increase flood risks and lead to massive loss of life and property. Despite recent calls for improved floodplain protection and management, a comprehensive, global-scale assessment quantifying human floodplain alterations does not exist. We developed the first publicly available global dataset that quantifies human alterations in 15 million km2 floodplains along 520 major river basins during the recent 27 years (1992-2019) at 250-m resolution. To maximize the reuse of our dataset and advance the open science of human floodplain alteration, we developed three web-based programming tools supported with tutorials and step-by-step audiovisual instructions. Our data reveal a significant loss of natural floodplains worldwide with 460,000 km2 of new agricultural and 140,000 km2 of new developed areas between 1992 and 2019. This dataset offers critical new insights into how floodplains are being destroyed, which will help decision-makers to reinforce strategies to conserve and restore floodplain functions and habitat.

National hydrologic connectivity classification links wetlands with stream water quality.

Wetland hydrologic connections to downstream waters influence stream water quality. However, no systematic approach for characterizing this connectivity exists. Here using physical principles, we categorized conterminous US freshwater wetlands into four hydrologic connectivity classes based on stream contact and flowpath depth to the nearest stream: riparian, non-riparian shallow, non-riparian mid-depth and non-riparian deep. These classes were heterogeneously distributed over the conterminous United States; for example, riparian dominated the south-eastern and Gulf coasts, while non-riparian deep dominated the Upper Midwest and High Plains. Analysis of a national stream dataset indicated acidification and organic matter brownification increased with connectivity. Eutrophication and sedimentation decreased with wetland area but did not respond to connectivity. This classification advances our mechanistic understanding of wetland influences on water quality nationally and could be applied globally.

River Basin Simulations Reveal Wide-Ranging Wetland-Mediated Nitrate Reductions.

River basin-scale wetland restoration and creation is a primary management option for mitigating nitrogen-based water quality challenges. However, the magnitude of nitrogen reduction that will result from adding wetlands across large river basins is uncertain, partly because the areal extent, location, and physical and functional characteristics of the wetlands are unknown. We simulated over 3600 wetland restoration scenarios across the ∼450,000 km2 Upper Mississippi River Basin (UMRB) depicting varied assumptions for wetland areal extent, physical and functional characteristics, and placement strategy. These simulations indicated that restoring wetlands will reduce local nitrate yields and nitrate loads at the UMRB outlet. However, the projected magnitude of nitrate reduction varied widely across disparate scenario assumptions─e.g., restoring 4500 km2 of wetlands (i.e., 1% of UMRB area) decreased mean annual nitrate loads at the UMRB outlet between 3 and 42%. Higher magnitude nitrate reductions correlated with best-case assumptions, particularly for characteristics controlling nitrate loading rates to the wetlands. These results show that simplified claims about basin-scale wetland-mediated water quality improvements discount the breadth of possible wetland impacts across disparate wetland physical and functional conditions and highlight a need for greater clarity regarding the likelihood of these conditions at river basin scales.

Advancing Watershed Legacy Nitrogen Modeling to Improve Global Water Quality.

Despite widespread implementation of watershed nitrogen reduction programs across the globe, nitrogen levels in many surface waters remain high. Watershed legacy nitrogen storage, i.e., the long-term retention of nitrogen in soils and groundwater, is one of several explanations for this lack of progress. However scientists and water managers are ill-equipped to estimate how legacy nitrogen moderates in-stream nitrogen responses to land conservation practices, largely because modeling tools and associated long-term monitoring approaches to answering these questions remain inadequate. We demonstrate the need for improved watershed models to simulate legacy nitrogen processes and offer modeling solutions to support long-term nitrogen-based sustainable land management across the globe.

A global review of beaver dam impacts: Stream conservation implications across biomes.

Beaver are recolonizing previously occupied regions, expanding into new territories, and increasingly being introduced and protected for stream conservation and restoration across numerous biomes. However, beaver dam effects on the physical, chemical, and biological characteristics of streams may vary within and among biomes. A comprehensive review of these impacts is lacking. The goals of this review were to: 1) summarize the distribution of studies by biome on beaver dam effects related to channel morphology, hydrology, water quality, and aquatic biota, as well as on beaver habitat selection, 2) summarize the extent to which beaver dam impacts have been consistent within and among biomes, and 3) share testable hypotheses regarding beaver impacts within understudied biomes. We quantify the directionality of beaver dam impacts from 267 peer-reviewed studies. Results show that the majority of studies have been completed within temperate forest environments and that many biomes are understudied. Across biomes, beaver preferred sites for dam development characterized by relatively low gradients and unconfined reaches with small drainage areas. Overall, parameters related to stream morphology and hydrology showed relatively consistent responses to beaver dams within and among biomes, yet water quality and biotic responses were variable among biomes. Responses also varied by parameter within water quality and biotic impact categories. The findings of this study can be useful for stream conservation and restoration efforts that introduce or protect beaver. Additional studies are needed within arid and cold biomes historically occupied by beaver and in novel biomes where beaver populations are currently expanding.

Vulnerable Waters are Essential to Watershed Resilience.

Watershed resilience is the ability of a watershed to maintain its characteristic system state while concurrently resisting, adapting to, and reorganizing after hydrological (for example, drought, flooding) or biogeochemical (for example, excessive nutrient) disturbances. Vulnerable waters include non-floodplain wetlands and headwater streams, abundant watershed components representing the most distal extent of the freshwater aquatic network. Vulnerable waters are hydrologically dynamic and biogeochemically reactive aquatic systems, storing, processing, and releasing water and entrained (that is, dissolved and particulate) materials along expanding and contracting aquatic networks. The hydrological and biogeochemical functions emerging from these processes affect the magnitude, frequency, timing, duration, storage, and rate of change of material and energy fluxes among watershed components and to downstream waters, thereby maintaining watershed states and imparting watershed resilience. We present here a conceptual framework for understanding how vulnerable waters confer watershed resilience. We demonstrate how individual and cumulative vulnerable-water modifications (for example, reduced extent, altered connectivity) affect watershed-scale hydrological and biogeochemical disturbance response and recovery, which decreases watershed resilience and can trigger transitions across thresholds to alternative watershed states (for example, states conducive to increased flood frequency or nutrient concentrations). We subsequently describe how resilient watersheds require spatial heterogeneity and temporal variability in hydrological and biogeochemical interactions between terrestrial systems and down-gradient waters, which necessitates attention to the conservation and restoration of vulnerable waters and their downstream connectivity gradients. To conclude, we provide actionable principles for resilient watersheds and articulate research needs to further watershed resilience science and vulnerable-water management.

Headwater streams and inland wetlands: Status and advancements of geospatial datasets and maps across the United States.

Headwater streams and inland wetlands provide essential functions that support healthy watersheds and downstream waters. However, scientists and aquatic resource managers lack a comprehensive synthesis of national and state stream and wetland geospatial datasets and emerging technologies that can further improve these data. We conducted a review of existing United States (US) federal and state stream and wetland geospatial datasets, focusing on their spatial extent, permanence classifications, and current limitations. We also examined recent peer-reviewed literature for emerging methods that can potentially improve the estimation, representation, and integration of stream and wetland datasets. We found that federal and state datasets rely heavily on the US Geological Survey's National Hydrography Dataset for stream extent and duration information. Only eleven states (22%) had additional stream extent information and seven states (14%) provided additional duration information. Likewise, federal and state wetland datasets primarily use the US Fish and Wildlife Service's National Wetlands Inventory (NWI) Geospatial Dataset, with only two states using non-NWI datasets. Our synthesis revealed that LiDAR-based technologies hold promise for advancing stream and wetland mapping at limited spatial extents. While machine learning techniques may help to scale-up these LiDAR-derived estimates, challenges related to preprocessing and data workflows remain. High-resolution commercial imagery, supported by public imagery and cloud computing, may further aid characterization of the spatial and temporal dynamics of streams and wetlands, especially using multi-platform and multi-temporal machine learning approaches. Models integrating both stream and wetland dynamics are limited, and field-based efforts must remain a key component in developing improved headwater stream and wetland datasets. Continued financial and partnership support of existing databases is also needed to enhance mapping and inform water resources research and policy decisions.

Wetland restoration yields dynamic nitrate responses across the Upper Mississippi river basin.

Wetland restoration is a primary management option for removing surplus nitrogen draining from agricultural landscapes. However, wetland capacity to mitigate nitrogen losses at large river-basin scales remains uncertain. This is largely due to a limited number of studies that address the cumulative and dynamic effects of restored wetlands across the landscape on downstream nutrient conditions. We analyzed wetland restoration impacts on modeled nitrate dynamics across 279 subbasins comprising the ∼0.5 million km2 Upper Mississippi River Basin (UMRB), USA, which covers eight states and houses ∼30 million people. Restoring ∼8,000 km2 of wetlands will reduce mean annual nitrate loads to the UMRB outlet by 12%, a substantial improvement over existing conditions but markedly less than widely cited estimates. Our lower wetland efficacy estimates are partly attributed to improved representation of processes not considered by preceding empirical studies - namely the potential for nitrate to bypass wetlands (i.e., via subsurface tile drainage) and be stored or transformed within the river network itself. Our novel findings reveal that wetlands mitigate surplus nitrogen basin-wide, yet they may not be as universally effective in tiled landscapes and because of river network processing.

Modeling spatially resolved characterization factors for eutrophication potential in life cycle assessment.

PURPOSE: Prior versions of the Tool for Reduction and Assessment of Chemical and other environmental Impacts (TRACI) have recognized the need for spatial variability when characterizing eutrophication. However, the method's underlying environmental models had not been updated to reflect the latest science. This new research provides the ability to differentiate locations with a high level of detail within the USA and provides global values at the country level. METHODS: In previous research (Morelli et al. 2018), the authors reviewed a broad range of domain-specific models and life cycle assessment methods for characterization of eutrophication and ranked these by levels of importance to the field and readiness for further development. The current research is rooted in the decision outcome of Morelli et al. (2018) to separate freshwater and marine eutrophication to allow for the most tailored characterization of each category individually. The current research also assumes that freshwater systems are limited by phosphorus and marine systems are limited by nitrogen. Using a combination of spatial modeling methods for soil, air, and water, we calculate midpoint characterization factors for freshwater and marine eutrophication categories and evaluate the results through a US-based case application. RESULTS AND DISCUSSION: Maps of the nutrient inventories, characterization factors, and overall impacts of the case application illustrate the spatial variation and patterns in the results. The importance of variation in geographic location is demonstrated using nutrient-based activity likelihood categories of agricultural (rural fertilizer), non-agricultural (urban fertilizer), and general (human waste processing). Proximity to large bodies of water, as well as individual hydraulic residence times, was shown to affect the comparative values of characterization factors across the USA. CONCLUSIONS: In this paper, we have calculated and applied finely resolved freshwater and marine eutrophication characterization factors for the USA and country-level factors for the rest of the globe. Additional research is needed to provide similarly resolved characterization factors for the entire globe, which would require expansion of publicly available data and further development of applicable fate and transport models. Further scientific advances may also be considered as computing capabilities become more sophisticated and widely accessible.

The changing face of floodplains in the Mississippi River Basin detected by a 60-year land use change dataset.

Floodplains provide essential ecosystem functions, yet >80% of European and North American floodplains are substantially modified. Despite floodplain changes over the past century, comprehensive, long-term land use change data within large river basin floodplains are limited. Long-term land use data can be used to quantify floodplain functions and provide spatially explicit information for management, restoration, and flood-risk mitigation. We present a comprehensive dataset quantifying floodplain land use change along the 3.3 million km2 Mississippi River Basin (MRB) covering 60 years (1941-2000) at 250-m resolution. We developed four unique products as part of this work, a(n): (i) Google Earth Engine interactive map visualization interface, (ii) Python code that runs in any internet browser, (iii) online tutorial with visualizations facilitating classroom code application, and (iv) instructional video demonstrating code application and database reproduction. Our data show that MRB's natural floodplain ecosystems have been substantially altered to agricultural and developed land uses. These products will support MRB resilience and sustainability goals by advancing data-driven decision making on floodplain restoration, buyout, and conservation scenarios.

Monthly river temperature trends across the US confound annual changes.

Climate variations and human modifications of the water cycle continue to alter the Earth's surface water and energy exchanges. It is therefore critical to ascertain how these changes impact water quality and aquatic ecosystem habitat metrics such as river temperatures. Though river temperature trend analyses exist in the literature, studies on seasonal trends in river temperatures across large spatial extents, e.g. the contiguous United States (US), are limited. As we show through both annual and monthly trend analyses for 20 year (n = 138 sites) and 40 year (n = 40 sites) periods, annual temperature trends across the US mask extensive monthly variability. While most sites exhibited annual warming trends, these annual trends obscured sub-annual cooling trends at many sites. Monthly trend anomalies were spatially organized, with persistent regional patterns at both reference and human-impacted sites. The largest warming and cooling anomalies happened at human impacted sites and during summer months. Though our analysis points to coherence in trends as well as the overall impact of human activity in driving these patterns, we did not investigate the impact of river temperature observation accuracy on reported trends, an area needed for future work. Overall, these patterns emphasize the need to consider sub-annual behavior when managing the ecological impacts of river temperature throughout lotic networks.

K in an Urban World: New Contexts for Hydraulic Conductivity.

Hydraulic conductivity (K) is a key hydrologic parameter widely recognized to be difficult to estimate and constrain, with little consistent assessment in disturbed, urbanized soils. To estimate K, it is either measured, or simulated by pedotransfer functions, which relate K to easily measured soil properties. We measured K in urbanized soils by double-ring infiltrometer (K dring), near-saturated tension infiltrometry (K minidisk), and constant head borehole permeametry (K borehole), along with other soil properties across the major soil orders in 12 United States cities. We compared measured K with that predicted from the pedotransfer function, ROSETTA. We found that regardless of soil texture, K dring was consistently larger than K minidisk; with the latter having slightly less sample variance. K borehole was dependent upon specific subsurface conditions, and contrary to common expectations, did not always decrease with depth. Based on either soil textural class, or percent textural separates (sand, silt clay), ROSETTA did not accurately predict measured K for surface nor subsurface soils. We go on to discuss how K varies in urban landscapes, the role of measurement methods and artifacts in the perception of this metric, and implications for hydrologic modeling. Overall, we aim to inspire consistency and coherence when addressing K-related challenges in sustainable urban water management.

Surface Depression and Wetland Water Storage Improves Major River Basin Hydrologic Predictions.

Surface water storage in small yet abundant landscape depressions-including wetlands and other small waterbodies-is largely disregarded in conventional hydrologic modeling practices. No quantitative evidence exists of how their exclusion may lead to potentially inaccurate model projections and understanding of hydrologic dynamics across the world's major river basins. To fill this knowledge gap, we developed the first-ever major river basin-scale modeling approach integrating surface depressions and focusing on the 450,000-km2 Upper Mississippi River Basin (UMRB) in the United States. We applied a novel topography-based algorithm to estimate areas and volumes of ~455,000 surface depressions (>1 ha) across the UMRB (in addition to lakes and reservoirs) and subsequently aggregated their effects per subbasin. Compared to a "no depression" conventional model, our depression-integrated model (a) improved streamflow simulation accuracy with increasing upstream abundance of depression storage, (b) significantly altered the spatial patterns and magnitudes of water yields across 315,000 km2 (70%) of the basin area, and (c) provided realistic spatial distributions of rootzone wetness conditions corresponding to satellite-based data. Results further suggest that storage capacity (i.e., volume) alone does not fully explain depressions' cumulative effects on landscape hydrologic responses. Local (i.e., subbasin level) climatic and geophysical drivers and downstream flowpath-regulating structures (e.g., reservoirs and dams) influence the extent to which depression storage volume in a subbasin causes hydrologic effects. With these new insights, our study supports the integration of surface depression storage and thereby catalyzes a reassessment of current hydrological modeling and management practices for basin-scale studies.