Automated Quantification of Surface Water Inundation in Wetlands Using Optical Satellite Imagery

Practical Guide to Measuring Wetland Carbon Pools and Fluxes

Wetlands cover a small portion of the world, but have disproportionate influence on global carbon (C) sequestration, carbon dioxide and methane emissions, and aquatic C fluxes. However, the underlying biogeochemical processes that affect wetland C pools and fluxes are complex and dynamic, making measurements of wetland C challenging. Over decades of research, many observational, experimental, and analytical approaches have been developed to understand and quantify pools and fluxes of wetland C. Sampling approaches range in their representation of wetland C from short to long timeframes and local to landscape spatial scales. This review summarizes common and cutting-edge methodological approaches for quantifying wetland C pools and fluxes. We first define each of the major C pools and fluxes and provide rationale for their importance to wetland C dynamics. For each approach, we clarify what component of wetland C is measured and its spatial and temporal representativeness and constraints. We describe practical considerations for each approach, such as where and when an approach is typically used, who can conduct the measurements (expertise, training requirements), and how approaches are conducted, including considerations on equipment complexity and costs. Finally, we review key covariates and ancillary measurements that enhance the interpretation of findings and facilitate model development. The protocols that we describe to measure soil, water, vegetation, and gases are also relevant for related disciplines such as ecology. Improved quality and consistency of data collection and reporting across studies will help reduce global uncertainties and develop management strategies to use wetlands as nature-based climate solutions.

Supplementary Information

The online version contains supplementary material available at 10.1007/s13157-023-01722-2.

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).

Convolutional neural network for high-resolution wetland mapping with open data: Variable selection and the challenges of a generalizable model

Landscape scale wetland conservation requires accurate, up-to-date wetland maps. The most useful approaches to creating such maps are automated, spatially generalizable, temporally repeatable, and can be applied at large spatial scales. However, mapping wetlands with predictive models is challenging due to the highly variable characteristics of wetlands in both space and time. Currently, most approaches are limited by coarse resolution, commercial data, and geographic specificity. Here, we trained a deep learning model and evaluated its ability to automatically map wetlands at landscape scale in a variety of geographies. We trained a U-Net architecture to map wetlands at 1-meter spatial resolution with the following remotely sensed covariates: multispectral data from the National Agriculture Imagery Program and the Sentinel-2 satellite system, and two LiDAR-derived datasets, intensity and geomorphons. The full model mapped wetlands accurately (94 % accuracy, 96.5 % precision, 95.2 % AUC) at 1-meter resolution. Post hoc model evaluation showed that the model correctly predicted wetlands even in areas that had incorrect label/training data, which penalized the recall rate (90.2 %). Applying the model in a new geography resulted in poor performance (precision = ~80 %, recall = 48 %). However, limited retraining in this geography improved model performance substantially, demonstrating an effective means to create a spatially generalizable model. We demonstrate wetlands can be mapped at high-resolution (1 m) using free data and efficient deep-learning models that do not require manual feature engineering. Including LiDAR and geomorphons as input data improved model accuracy by 2 %, and where these data are unavailable a simpler model can efficiently map wetlands. Given the dynamic nature of wetlands and the important ecosystem services they provide, high-resolution mapping can be a game changer in terms of informing restoration and development decisions.

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.

Creating a Detailed Wetland Inventory with Sentinel-2 Time-Series Data and Google Earth Engine in the Prairie Pothole Region of Canada

Wetlands in the Prairie Pothole Region (PPR) of Canada and the United States represent a unique mapping challenge. They are dynamic both seasonally and year-to-year, are very small, and frequently altered by human activity. Many efforts have been made to estimate the loss of these important habitats but a high-quality inventory of pothole wetlands is needed for data-driven conservation and management of these resources. Typical landcover classifications using one or two image dates from optical or Synthetic Aperture Radar (SAR) Earth Observation (EO) systems often produce reasonable wetland inventories for less dynamic, forested landscapes, but will miss many of the temporary and seasonal wetlands in the PPR. Past studies have attempted to capture PPR wetland dynamics by using dense image stacks of optical or SAR data. We build upon previous work, using 2017–2020 Sentinel-2 imagery processed through the Google Earth Engine (GEE) cloud computing platform to capture seasonal flooding dynamics of wetlands in a prairie pothole wetland landscape in Alberta, Canada. Using 36 different image dates, wetland flood frequency (hydroperiod) was calculated by classifying water/flooding in each image date. This product along with the Global Ecosystem Dynamics Investigation (GEDI) Canopy Height Model (CHM) was then used to generate a seven-class wetland inventory with wetlands classified as areas with seasonal but not permanent water/flooding. Overall accuracies of the resulting inventory were between 95% and 96% based on comparisons with local photo-interpreted inventories at the Canadian Wetland Classification System class level, while wetlands themselves were classified with approximately 70% accuracy. The high overall accuracy is due, in part, to a dominance of uplands in the PPR. This relatively simple method of classifying water through time generates reliable wetland maps but is only applicable to ecosystems with open/non-complex wetland types and may be highly sensitive to the timing of cloud-free optical imagery that captures peak wetland flooding (usually post snow melt). Based on this work, we suggest that expensive field or photo-interpretation training data may not be needed to map wetlands in the PPR as self-labeling of flooded and non-flooded areas in a few Sentinel-2 images is sufficient to classify water through time. Our approach demonstrates a framework for the operational mapping of small, dynamic PPR wetlands that relies on open-access EO data and does not require costly, independent training data. It is an important step towards the effective conservation and management of PPR wetlands, providing an efficient method for baseline and ongoing mapping in these dynamic environments.

The Applicability of LandTrendr to Surface Water Dynamics: A Case Study of Minnesota from 1984 to 2019 Using Google Earth Engine

The means to accurately monitor wetland change over time are crucial to wetland management. This paper explores the applicability of LandTrendr, a temporal segmentation algorithm designed to identify significant interannual trends, to monitor wetlands by modeling surface water presence in Minnesota from 1984 to 2019. A time series of harmonized Landsat and Sentinel-2 data in the spring is developed in Google Earth Engine, and calculated to sub-pixel water fraction. The optimal parameters for modeling this time series with LandTrendr are identified by minimizing omission of known surface water locations, and the result of this optimal model of sub-pixel water fraction is evaluated against reference images and qualitatively. Accuracy of this method is high: overall accuracy is 98% and producer’s and user’s accuracies for inundation are 82% and 88% respectively. Maps summarizing the trendlines of multiple pixels, such as frequency of inundation over the past 35 years, also show LandTrendr as applied here can accurately model long-term trends in surface water presence across wetland types. However, the tendency of omission for more variable prairie pothole wetlands and the under-prediction of inundation for small or emergent wetlands suggests the algorithm will require careful development of the segmented time series to capture inundated conditions more accurately.

AHSWFM: Automated and Hierarchical Surface Water Fraction Mapping for Small Water Bodies Using Sentinel-2 Images

Accurately mapping surface water fractions is essential to understanding the distribution and area of small water bodies (SWBs), which are numerous and widespread. Traditional spectral unmixings based on the linear mixture model require high-quality prior endmember information, and are not appropriate in situations such as dealing with multiple scattering effects. To overcome difficulties with unknown mixing mechanisms and parameters, a novel automated and hierarchical surface water fraction mapping (AHSWFM) for mapping SWBs from Sentinel-2 images was proposed. AHSWFM is automated, requires no endmember prior knowledge and uses self-trained regression using scalable algorithms and random forest to construct relationships between the multispectral data and water fractions. AHSWFM uses a hierarchical structure that divides pixels into pure water, pure land and mixed water-land pixels, and predicts their water fractions separately to avoid overestimating water fractions for pure land pixels and underestimating water fractions for pure water pixels. Results show that using the hierarchical strategy can increase the accuracy in estimating SWB areas. AHSWFM predicted SWB areas with a root mean square error of approximately 0.045 ha in a region using more than 1200 SWB samples that were mostly smaller than 0.75 ha.

Monitoring Post-Flood Recovery of Croplands Using the Integrated Sentinel-1/2 Imagery in the Yangtze-Huai River Basin

The increasingly frequent flooding imposes tremendous and long-lasting damages to lives and properties in impoverished rural areas. Rapid, accurate, and large-scale flood mapping is urgently needed for flood management, and to date has been successfully implemented benefiting from the advancement in remote sensing and cloud computing technology. Yet, the effects of agricultural emergency response to floods have been limitedly evaluated by satellite-based remote sensing, resulting in biased post-flood loss assessments. Addressing this challenge, this study presents a method for monitoring post-flood agricultural recovery using Sentinel-1/2 imagery, tested in three flood-affected main grain production areas, in the middle and lower Yangtze and Huai River, China. Our results indicated that 33~72% of the affected croplands were replanted and avoided total crop failures in summer 2020. Elevation, flood duration, crop rotation scheme, and flooding emergency management affect the post-flood recovery performance. The findings also demonstrate rapid intervention measures adjusted to local conditions could reduce the agricultural failure cost from flood disasters to a great extent. This study provides a new alternative for comprehensive disaster loss assessment in flood-prone agricultural regions, which will be insightful for worldwide flood control and management.

Retrieving dynamics of the surface water extent in the upper reach of Yellow River

Multi-time scale surface water extent (SWE) dynamics are very important to understand climate change impacts on water resources. With Landsat 5/7/8 images and Google Earth Engine (GEE), an improved threshold-based water extraction algorithm and a novel surface water gaps (SWGs) interpolation method based on historical water frequency were applied to build surface water area (SWA, namely SWE without ice) and water body area (WBA, namely SWE with ice) monthly (January 2001-December 2019) and annual (1986-2019) time series in the upper reaches of the Yellow River (UYR). The Mann-Kendall test was used to analyse SWE trends, and the ridge regression was performed to figure out the relative contributions of meteorological factors to SWE dynamics. The pixels with modified normalized difference water index (MNDWI) higher than normalized difference vegetation index (NDVI) or enhanced vegetation index (EVI) were identified as SWE. The mean relative error (MRE) of the SWGs interpolation results was below 10%. At the annual scale, the average SWA and number of lakes over 1 ha showed significant upward trends of 4.4 km² yr^-1 and 7.53 yr^-1, respectively. The monthly WBA increased in summer and autumn while decreased in spring and winter. The maximum freezing and thawing ratios were 53.74% in December and 37.32% in May, respectively. Attribution analysis showed that precipitation and wind speed were the foremost factors dominating the dynamics of annual SWA and monthly WBA, respectively. Our findings confirmed that climatic changes have altered the dynamics of water bodies in the UYR.

Automated Training Data Generation from Spectral Indexes for Mapping Surface Water Extent with Sentinel-2 Satellite Imagery at 10 m and 20 m Resolutions

This study presents an automated methodology to generate training data for surface water mapping from a single Sentinel-2 granule at 10 m (4 band, VIS/NIR) or 20 m (9 band, VIS/NIR/SWIR) resolution without the need for ancillary training data layers. The 20 m method incorporates an ensemble of three spectral indexes with optimal band thresholds, whereas the 10 m method achieves similar results using fewer bands and a single spectral index. A spectrally balanced and randomly generated set of training data based on the index values and optimal thresholds is used to fit machine learning classifiers. Statistical validation compares the 20 m ensemble-only method to the 20 m ensemble method with a random forest classifier. Results show the 20 m ensemble-only method had an overall accuracy of 89.5% (±1.7%), whereas the ensemble method combined with the random forest classifier performed better, with a ~4.8% higher overall accuracy: 20 m method (94.3% (±1.3%)) with optimal spectral index and SWIR thresholds of −0.03 and 800, respectively, and 10 m method (93.4% (±1.5%)) with optimal spectral index and NIR thresholds of −0.01 and 800, respectively. Comparison of other supervised classifiers trained automatically with the framework typically resulted in less than 1% accuracy improvement compared with the random forest, suggesting that training data quality is more important than classifier type. This straightforward framework enables accurate surface water classification across diverse geographies, making it ideal for development into a decision support tool for water resource managers.

Uncertainties Involved in the Use of Thresholds for the Detection of Water Bodies in Multitemporal Analysis from Landsat-8 and Sentinel-2 Images

Although the single threshold is still considered a suitable and easy-to-do technique to extract water features in spatiotemporal analysis, it leads to unavoidable errors. This paper uses an enumerative search to optimize thresholds over satellite-derived modified normalized difference water index (MNDWI). We employed a cross-validation approach and treated accuracy as a random variable in order to: (a) investigate uncertainty related to its application; (b) estimate non-optimistic errors involving single thresholding; (c) investigate the main factors that affect the accuracy's model, and (d) compare satellite sensors performance. We also used a high-resolution digital elevation model to extract water elevations values, making it possible to remove topographic effects and estimate non-optimistic errors exclusively from orbital imagery. Our findings evidenced that there is a region where thresholds values can vary without causing accuracy loss. Moreover, by constraining thresholds variation between these limits, accuracy is dramatically improved and outperformed the Otsu method. Finally, the number of scenes employed to optimize a single threshold drastically affects the accuracy, being not appropriate using a single scene once it leads to overfitted threshold values. More than three scenes are recommended.

Migration stopover ecology of Cinnamon Teal in western North America

Identifying migration routes and fall stopover sites of Cinnamon Teal (Spatula cyanoptera septentrionalium) can provide a spatial guide to management and conservation efforts, and address vulnerabilities in wetland networks that support migratory waterbirds. Using high spatiotemporal resolution GPS-GSM transmitters, we analyzed 61 fall migration tracks across western North America during our three-year study (2017-2019). We marked Cinnamon Teal primarily during spring/summer in important breeding and molting regions across seven states (California, Oregon, Washington, Idaho, Utah, Colorado, and Nevada). We assessed fall migration routes and timing, detected 186 fall stopover sites, and identified specific North American ecoregions where sites were located. We classified underlying land cover for each stopover site and measured habitat selection for 12 land cover types within each ecoregion. Cinnamon Teal selected a variety of flooded habitats including natural, riparian, tidal, and managed wetlands; wet agriculture (including irrigation ditches, flooded fields, and stock ponds); wastewater sites; and golf and urban ponds. Wet agriculture was the most used habitat type (29.8% of stopover locations), and over 72% of stopover locations were on private land. Relatively scarce habitats such as wastewater ponds, tidal marsh, and golf and urban ponds were highly selected in specific ecoregions. In contrast, dry non-habitat across all ecoregions, and dry agriculture in the Cold Deserts and Mediterranean California ecoregions, was consistently avoided. Resources used by Cinnamon Teal often reflected wetland availability across the west and emphasize their adaptability to dynamic resource conditions in arid landscapes. Our results provide much needed information on spatial and temporal resource use by Cinnamon Teal during migration and indicate important wetland habitats for migrating waterfowl in the western United States.

A Disease Control-Oriented Land Cover Land Use Map for Myanmar

Malaria is a serious infectious disease that leads to massive casualties globally. Myanmar is a key battleground for the global fight against malaria because it is where the emergence of drug-resistant malaria parasites has been documented. Controlling the spread of malaria in Myanmar thus carries global significance, because the failure to do so would lead to devastating consequences in vast areas where malaria is prevalent in tropical/subtropical regions around the world. Thanks to its wide and consistent spatial coverage, remote sensing has become increasingly used in the public health domain. Specifically, remote sensing-based land cover/land use (LCLU) maps present a powerful tool that provides critical information on population distribution and on the potential human-vector interactions interfaces on a large spatial scale. Here, we present a 30-meter LCLU map that was created specifically for the malaria control and eradication efforts in Myanmar. This bottom-up approach can be modified and customized to other vector-borne infectious diseases in Myanmar or other Southeastern Asian countries.

Characterizing Wetland Inundation and Vegetation Dynamics in the Arctic Coastal Plain Using Recent Satellite Data and Field Photos

Arctic wetlands play a critical role in the global carbon cycle and are experiencing disproportionate impacts from climate change. Even though Alaska hosts 65% of U.S. wetlands, less than half of the wetlands in Alaska have been mapped by the U.S. Fish and Wildlife Service National Wetlands Inventory (NWI) or other high-resolution wetlands protocols. The availability of time series satellite data and the development of machine learning algorithms have enabled the characterization of Arctic wetland inundation dynamics and vegetation types with limited ground data input. In this study, we built a semi-automatic process to generate sub-pixel water fraction (SWF) maps across the Coastal Plain of the Arctic National Wildlife Refuge (ANWR) in Alaska using random forest regression and 139 Sentinel-2 images taken in ice-free seasons from 2016 to 2019. With this, we characterized the seasonal dynamics of wetland inundation and explored their potential usage in determining NWI water regimes. The highest levels of surface water expression were detected in June, resulting from seasonal active layer thaw and snowmelt. Inundation was most variable in riverbeds, lake and pond margins, and depressional wetlands, where water levels fluctuate substantially between dry and wet seasons. NWI water regimes that indicate frequent inundation, such as permanently flooded wetlands, had high SWF values (SWF ≥ 90%), while those with infrequent inundation, such as temporarily flooded wetlands, had low SWF values (SWF < 10%). Vegetation types were also classified through the synergistic use of a vegetation index, water regimes, synthetic-aperture radar (SAR) data, topographic data, and a random forest classifier. The random forest classification algorithms demonstrated good performance in classifying Arctic wetland vegetation types, with an overall accuracy of 0.87. Compared with NWI data produced in the 1980s, scrub-shrub wetlands appear to have increased from 91 to 258 km2 over the last three decades, which is the largest percentage change (182%) among all vegetation types. However, additional field data are needed to confirm this shift in vegetation type. This study demonstrates the potential of using time series satellite data and machine learning algorithms in characterizing inundation dynamics and vegetation types of Arctic wetlands. This approach could aid in the creation and maintenance of wetland inventories, including the NWI, in Arctic regions and enable an improved understanding of long-term wetland dynamics.

Multi-Source EO for Dynamic Wetland Mapping and Monitoring in the Great Lakes Basin

Wetland managers, citizens and government leaders are observing rapid changes in coastal wetlands and associated habitats around the Great Lakes Basin due to human activity and climate variability. SAR and optical satellite sensors offer cost effective management tools that can be used to monitor wetlands over time, covering large areas like the Great Lakes and providing information to those making management and policy decisions. In this paper we describe ongoing efforts to monitor dynamic changes in wetland vegetation, surface water extent, and water level change. Included are assessments of simulated Radarsat Constellation Mission data to determine feasibility of continued monitoring into the future. Results show that integration of data from multiple sensors is most effective for monitoring coastal wetlands in the Great Lakes region. While products developed using methods described in this article provide valuable management tools, more effort is needed to reach the goal of establishing a dynamic, near-real-time, remote sensing-based monitoring program for the basin.

Malaria Exposure in Ann Township, Myanmar, as a Function of Land Cover and Land Use: Combining Satellite Earth Observations and Field Surveys

Despite progress toward malaria elimination in the Greater Mekong Subregion, challenges remain owing to the emergence of drug resistance and the persistence of focal transmission reservoirs. Malaria transmission foci in Myanmar are heterogeneous and complex, and many remaining infections are clinically silent, rendering them invisible to routine monitoring. The goal of this research is to define criteria for easy-to-implement methodologies, not reliant on routine monitoring, that can increase the efficiency of targeted malaria elimination strategies. Studies have shown relationships between malaria risk and land cover and land use (LCLU), which can be mapped using remote sensing methodologies. Here we aim to explain malaria risk as a function of LCLU for five rural villages in Myanmar's Rakhine State. Malaria prevalence and incidence data were analyzed through logistic regression with a land use survey of ~1,000 participants and a 30-m land cover map. Malaria prevalence per village ranged from 5% to 20% with the overwhelming majority of cases being subclinical. Villages with high forest cover were associated with increased risk of malaria, even for villagers who did not report visits to forests. Villagers living near croplands experienced decreased malaria risk unless they were directly engaged in farm work. Finally, land cover change (specifically, natural forest loss) appeared to be a substantial contributor to malaria risk in the region, although this was not confirmed through sensitivity analyses. Overall, this study demonstrates that remotely sensed data contextualized with field survey data can be used to inform critical targeting strategies in support of malaria elimination.

Meta-Analysis of Wetland Classification Using Remote Sensing: A Systematic Review of a 40-Year Trend in North America

North America is covered in 2.5 million km2 of wetlands, which is the remainder of an estimated 56% of wetlands lost since the 1700s. This loss has resulted in a decrease in important habitat and services of great ecological, economic, and recreational benefits to humankind. To better manage these ecosystems, since the 1970s, wetlands in North America have been classified with increasing regularity using remote sensing technology. Since then, optimal methods for wetland classification by numerous researchers have been examined, assessed, modified, and established. Over the past several decades, a large number of studies have investigated the effects of different remote sensing factors, such as data type, spatial resolution, feature selection, classification methods, and other parameters of interest on wetland classification in North America. However, the results of these studies have not yet been synthesized to determine best practices and to establish avenues for future research. This paper reviews the last 40 years of research and development on North American wetland classification through remote sensing methods. A meta-analysis of 157 relevant articles published since 1980 summarizes trends in 23 parameters, including publication, year, study location, application of specific sensors, and classification methods. This paper also examines is the relationship between several remote sensing parameters (e.g., spatial resolution and type of data) and resulting overall accuracies. Finally, this paper discusses the future of remote sensing of wetlands in North America with regard to upcoming technologies and sensors. Given the increasing importance and vulnerability of wetland ecosystems under the climate change influences, this paper aims to provide a comprehensive review in support of the continued, improved, and novel applications of remote sensing for wetland mapping across North America and to provide a fundamental knowledge base for future studies in this field.

Volume Variations of Small Inland Water Bodies from a Combination of Satellite Altimetry and Optical Imagery

In this study, a new approach for estimating volume variations of lakes and reservoirs using water levels from satellite altimetry and surface areas from optical imagery is presented. Both input data sets, namely water level time series and surface area time series, are provided by the Database of Hydrological Time Series of Inland Waters (DAHITI), developed and maintained by the Deutsches Geodätisches Forschungsinsitut der Technischen Universität München (DGFI-TUM). The approach is divided into three parts. In the first part, a hypsometry model based on the new modified Strahler approach is computed by combining water levels and surface areas. The hypsometry model describes the dependency between water levels and surface areas of lakes and reservoirs. In the second part, a bathymetry between minimum and maximum surface area is computed. For this purpose, DAHITI land-water masks are stacked using water levels derived from the hypsometry model. Finally, water levels and surface areas are intersected with the bathymetry to estimate a time series of volume variations in relation to the minimum observed surface area. The results are validated with volume time series derived from in-situ water levels in combination with bathymetric surveys. In this study, 28 lakes and reservoirs located in Texas are investigated. The absolute volumes of the investigated lakes and reservoirs vary between 0.062 km 3 and 6.041 km 3 . The correlation coefficients of the resulting volume variation time series with validation data vary between 0.80 and 0.99. Overall, the relative errors with respect to volume variations vary between 2.8% and 14.9% with an average of 8.3% for all 28 investigated lakes and reservoirs. When comparing the resulting RMSE with absolute volumes, the absolute errors vary between 1.5% and 6.4% with an average of 3.1%. This study shows that volume variations can be calculated with a high accuracy which depends essentially on the quality of the used water levels and surface areas. In addition, this study provides a hypsometry model, high-resolution bathymetry and water level time series derived from surface areas based on the hypsometry model. All data sets are publicly available on the Database of Hydrological Time Series of Inland Waters.

Isolating Anthropogenic Wetland Loss by Concurrently Tracking Inundation and Land Cover Disturbance across the Mid-Atlantic Region, U.S

Global trends in wetland degradation and loss have created an urgency to monitor wetland extent, as well as track the distribution and causes of wetland loss. Satellite imagery can be used to monitor wetlands over time, but few efforts have attempted to distinguish anthropogenic wetland loss from climate-driven variability in wetland extent. We present an approach to concurrently track land cover disturbance and inundation extent across the Mid-Atlantic region, United States, using the Landsat archive in Google Earth Engine. Disturbance was identified as a change in greenness, using a harmonic linear regression approach, or as a change in growing season brightness. Inundation extent was mapped using a modified version of the U.S. Geological Survey's Dynamic Surface Water Extent (DSWE) algorithm. Annual (2015-2018) disturbance averaged 0.32% (1095 km2 year-1) of the study area per year and was most common in forested areas. While inundation extent showed substantial interannual variability, the co-occurrence of disturbance and declines in inundation extent represented a minority of both change types, totaling 109 km2 over the four-year period, and 186 km2, using the National Wetland Inventory dataset in place of the Landsat-derived inundation extent. When the annual products were evaluated with permitted wetland and stream fill points, 95% of the fill points were detected, with most found by the disturbance product (89%) and fewer found by the inundation decline product (25%). The results suggest that mapping inundation alone is unlikely to be adequate to find and track anthropogenic wetland loss. Alternatively, remotely tracking both disturbance and inundation can potentially focus efforts to protect, manage, and restore wetlands.

Climate and human water use diminish wetland networks supporting continental waterbird migration

Migrating waterbirds moving between upper and lower latitudinal breeding and wintering grounds rely on a limited network of endorheic lakes and wetlands when crossing arid continental interiors. Recent drying of global endorheic water stores raises concerns over deteriorating migratory pathways, yet few studies have considered these effects at the scale of continental flyways. Here, we investigate the resiliency of waterbird migration networks across western North America by reconstructing long-term patterns (1984-2018) of terminal lake and wetland surface water area in 26 endorheic watersheds. Findings were partitioned regionally by snowmelt- and monsoon-driven hydrologies and combined with climate and human water-use data to determine their importance in predicting surface water trends. Nonlinear patterns of lake and wetland drying were apparent along latitudinal flyway gradients. Pervasive surface water declines were prevalent in northern snowmelt watersheds (lakes -27%, wetlands -47%) while largely stable in monsoonal watersheds to the south (lakes -13%, wetlands +8%). Monsoonal watersheds represented a smaller proportion of total lake and wetland area, but their distribution and frequency of change within highly arid regions of the continental flyway increased their value to migratory waterbirds. Irrigated agriculture and increasing evaporative demands were the most important drivers of surface water declines. Underlying agricultural and wetland relationships however were more complex. Approximately 7% of irrigated lands linked to flood irrigation and water storage practices supported 61% of all wetland inundation in snowmelt watersheds. In monsoonal watersheds, small earthen dams, meant to capture surface runoff for livestock watering, were a major component of wetland resources (67%) that supported networks of isolated wetlands surrounding endorheic lakes. Ecological trends and human impacts identified herein underscore the importance of assessing flyway-scale change as our model depictions likely reflect new and emerging bottlenecks to continental migration.

Land-Cover Changes to Surface-Water Buffers in the Midwestern USA: 25 Years of Landsat Data Analyses (1993-2017)

To understand the timing, extent, and magnitude of land use/land cover (LULC) change in buffer areas surrounding Midwestern US waters, we analyzed the full imagery archive (1982-2017) of three Landsat footprints covering ~100,000 km2. The study area included urbanizing Chicago, Illinois and St. Louis, Missouri regions and agriculturally dominated landscapes (i.e., Peoria, Illinois). The Continuous Change Detection and Classification algorithm identified 1993-2017 LULC change across three Landsat footprints and in 90 m buffers for ~110,000 surface waters; waters were also size-binned into five groups for buffer LULC change analyses. Importantly, buffer-area LULC change magnitude was frequently much greater than footprint-level change. Surface-water extent in buffers increased by 14-35x the footprint rate and forest decreased by 2-9x. Development in buffering areas increased by 2-4x the footprint-rate in Chicago and Peoria area footprints but was similar to the change rate in the St. Louis area footprint. The LULC buffer-area change varied in waterbody size, with the greatest change typically occurring in the smallest waters (e.g., <0.1 ha). These novel analyses suggest that surface-water buffer LULC change is occurring more rapidly than footprint-level change, likely modifying the hydrology, water quality, and biotic integrity of existing water resources, as well as potentially affecting down-gradient, watershed-scale storages and flows of water, solutes, and particulate matter.

Mapping Forested Wetland Inundation in the Delmarva Peninsula, USA Using Deep Convolutional Neural Networks

The Delmarva Peninsula in the eastern United States is partially characterized by thousands of small, forested, depressional wetlands that are highly sensitive to weather variability and climate change, but provide critical ecosystem services. Due to the relatively small size of these depressional wetlands and their occurrence under forest canopy cover, it is very challenging to map their inundation status based on existing remote sensing data and traditional classification approaches. In this study, we applied a state-of-the-art U-Net semantic segmentation network to map forested wetland inundation in the Delmarva area by integrating leaf-off WorldView-3 (WV3) multispectral data with fine spatial resolution light detection and ranging (lidar) intensity and topographic data, including a digital elevation model (DEM) and topographic wetness index (TWI). Wetland inundation labels generated from lidar intensity were used for model training and validation. The wetland inundation map results were also validated using field data, and compared to the U.S. Fish and Wildlife Service National Wetlands Inventory (NWI) geospatial dataset and a random forest output from a previous study. Our results demonstrate that our deep learning model can accurately determine inundation status with an overall accuracy of 95% (Kappa = 0.90) compared to field data and high overlap (IoU = 70%) with lidar intensity-derived inundation labels. The integration of topographic metrics in deep learning models can improve the classification accuracy for depressional wetlands. This study highlights the great potential of deep learning models to improve the accuracy of wetland inundation maps through use of high-resolution optical and lidar remote sensing datasets.

Automatic Inundation Mapping Using Sentinel-2 Data Applicable to Both Camargue and Doñana Biosphere Reserves

Flooding periodicity is crucial for biomass production and ecosystem functions in wetland areas. Local monitoring networks may be enriched by spaceborne derived products with a temporal resolution of a few days. Unsupervised computer vision techniques are preferred, since human interference and the use of training data may be kept to a minimum. Recently, a novel automatic local thresholding unsupervised methodology for separating inundated areas from non-inundated ones led to successful results for the Doñana Biosphere Reserve. This study examines the applicability of this approach to Camarque Biosphere Reserve, and proposes alternatives to the original approach to enhance accuracy and applicability for both Camargue and Doñana wetlands in a scientific quest for methods that may serve accurately biomes at both protected areas. In particular, it examines alternative inputs for automatically estimating thresholds while applying various algorithms for estimating the splitting thresholds. Reference maps for Camargue are provided by local authorities, and generated using Sentinel-2 Band 8A (NIR) and Band 12 (SWIR-2). The alternative approaches examined led to high inundation mapping accuracy. In particular, for the Camargue study area and 39 different dates, the alternative approach with the highest overall Kappa coefficient is 0.84, while, for the Doñana Biosphere Reserve and Doñana marshland (a subset of Doñana Reserve) and 7 different dates, is 0.85 and 0.94, respectively. Moreover, there are alternative approaches with high overall Kappa for all areas, i.e., at 0.79 for Camargue, over 0.91 for Doñana marshland, and over 0.82 for Doñana Reserve. Additionally, this study identifies the alternative approaches that perform better when the study area is extensively covered by temporary flooded and emergent vegetation areas (i.e., Camargue Reserve and Doñana marshland) or when it contains a large percentage of dry areas (i.e., Doñana Reserve). The development of credible automatic thresholding techniques that can be applied to different wetlands could lead to a higher degree of automation for map production, while enhancing service utilization by non-trained personnel.

Non-floodplain Wetlands Affect Watershed Nutrient Dynamics: A Critical Review

Wetlands have the capacity to retain nitrogen and phosphorus and are thereby often considered a viable option for improving water quality at local scales. However, little is known about the cumulative influence of wetlands outside of floodplains, i.e., non-floodplain wetlands (NFWs), on surface water quality at watershed scales. Such evidence is important to meet global, national, regional, and local water quality goals effectively and comprehensively. In this critical review, we synthesize the state of the science about the watershed-scale effects of NFWs on nutrient-based (nitrogen, phosphorus) water quality. We further highlight where knowledge is limited in this research area and the challenges of garnering this information. On the basis of previous wetland literature, we develop emerging concepts that assist in advancing the science linking NFWs to watershed-scale nutrient conditions. Finally, we ask, "Where do we go from here?" We address this question using a 2-fold approach. First, we demonstrate, via example model simulations, how explicitly considering NFWs in watershed nutrient modeling changes predicted nutrient yields to receiving waters-and how this may potentially affect future water quality management decisions. Second, we outline research recommendations that will improve our scientific understanding of how NFWs affect downstream water quality.

Integrating LiDAR data and multi-temporal aerial imagery to map wetland inundation dynamics using Google Earth Engine

The Prairie Pothole Region of North America is characterized by millions of depressional wetlands, which provide critical habitats for globally significant populations of migratory waterfowl and other wildlife species. Due to their relatively small size and shallow depth, these wetlands are highly sensitive to climate variability and anthropogenic changes, exhibiting inter- and intra-annual inundation dynamics. Moderate-resolution satellite imagery (e.g., Landsat, Sentinel) alone cannot be used to effectively delineate these small depressional wetlands. By integrating fine spatial resolution Light Detection and Ranging (LiDAR) data and multi-temporal (2009-2017) aerial images, we developed a fully automated approach to delineate wetland inundation extent at watershed scales using Google Earth Engine. Machine learning algorithms were used to classify aerial imagery with additional spectral indices to extract potential wetland inundation areas, which were further refined using LiDAR-derived landform depressions. The wetland delineation results were then compared to the U.S. Fish and Wildlife Service National Wetlands Inventory (NWI) geospatial dataset and existing global-scale surface water products to evaluate the performance of the proposed method. We tested the workflow on 26 watersheds with a total area of 16,576 km2 in the Prairie Pothole Region. The results showed that the proposed method can not only delineate current wetland inundation status but also demonstrate wetland hydrological dynamics, such as wetland coalescence through fill-spill hydrological processes. Our automated algorithm provides a practical, reproducible, and scalable framework, which can be easily adapted to delineate wetland inundation dynamics at broad geographic scales.

The potential role of very high-resolution imagery to characterise lake, wetland and stream systems across the Prairie Pothole Region, United States

Aquatic features critical to watershed hydrology range widely in size from narrow, shallow streams to large, deep lakes. In this study we evaluated wetland, lake, and river systems across the Prairie Pothole Region to explore where pan-sharpened high-resolution (PSHR) imagery, relative to Landsat imagery, could pro-vide additional data on surface water distribution and movement, missed by Landsat. We used the monthly Global Surface Water (GSW) Landsat product as well as surface water derived from Landsat imagery using a matched filtering algorithm (MF Landsat) to help consider how including partially inundated Landsat pixels as water influenced our findings. The PSHR outputs (and MF Landsat) were able to identify ~60-90% more surface water interactions between waterbodies, relative to the GSW Landsat product. However, regardless of Landsat source, by doc-umenting many smaller (<0.2 ha), inundated wetlands, the PSHR outputs modified our interpretation of wetland size distribution across the Prairie Pothole Region.

Modeling Connectivity of Non-floodplain Wetlands: Insights, Approaches, and Recommendations

Representing hydrologic connectivity of non-floodplain wetlands (NFWs) to downstream waters in process-based models is an emerging challenge relevant to many research, regulatory, and management activities. We review four case studies that utilize process-based models developed to simulate NFW hydrology. Models range from a simple, lumped parameter model to a highly complex, fully distributed model. Across case studies, we highlight appropriate application of each model, emphasizing spatial scale, computational demands, process representation, and model limitations. We end with a synthesis of recommended "best modeling practices" to guide model application. These recommendations include: (1) clearly articulate modeling objectives, and revisit and adjust those objectives regularly; (2) develop a conceptualization of NFW connectivity using qualitative observations, empirical data, and process-based modeling; (3) select a model to represent NFW connectivity by balancing both modeling objectives and available resources; (4) use innovative techniques and data sources to validate and calibrate NFW connectivity simulations; and (5) clearly articulate the limits of the resulting NFW connectivity representation. Our review and synthesis of these case studies highlights modeling approaches that incorporate NFW connectivity, demonstrates tradeoffs in model selection, and ultimately provides actionable guidance for future model application and development.

Automated Extraction of Consistent Time-Variable Water Surfaces of Lakes and Reservoirs Based on Landsat and Sentinel-2

In this study, a new approach for the automated extraction of high-resolution time-variable water surfaces is presented. For that purpose, optical images from Landsat and Sentinel-2 are used between January 1984 and June 2018. The first part of this new approach is the extraction of land-water masks by combining five water indexes and using an automated threshold computation. In the second part of this approach, all data gaps caused by voids, clouds, cloud shadows, or snow are filled by using a long-term water probability mask. This mask is finally used in an iterative approach for filling remaining data gaps in all monthly masks which leads to a gap-less surface area time series for lakes and reservoirs. The results of this new approach are validated by comparing the surface area changes with water level time series from gauging stations. For inland waters in remote areas without in situ data water level time series from satellite altimetry are used. Overall, 32 globally distributed lakes and reservoirs of different extents up to 2482.27 km 2 are investigated. The average correlation coefficients between surface area time series and water levels from in situ and satellite altimetry have increased from 0.611 to 0.862 after filling the data gaps which is an improvement of about 41%. This new approach clearly demonstrates the quality improvement for the estimated land-water masks but also the strong impact of a reliable data gap-filling approach. All presented surface area time series are freely available on the Database of Hydrological Time Series of Inland (DAHITI).

Preface: Remote Sensing for Flood Mapping and Monitoring of Flood Dynamics

This Special Issue is a collection of papers that focus on the use of remote sensing data and describe methods for flood monitoring and mapping. These articles span a wide range of topics; present novel processing techniques and review methods; and discuss limitations and challenges. This preface provides a brief overview of the content.

Improved Automated Detection of Subpixel-Scale Inundation—Revised Dynamic Surface Water Extent (DSWE) Partial Surface Water Tests

In order to produce useful hydrologic and aquatic habitat data from the Landsat system, the U.S. Geological Survey has developed the “Dynamic Surface Water Extent” (DSWE) Landsat Science Product. DSWE will provide long-term, high-temporal resolution data on variations in inundation extent. The model used to generate DSWE is composed of five decision-rule based tests that do not require scene-based training. To allow its general application, required inputs are limited to the Landsat at-surface reflectance product and a digital elevation model. Unlike other Landsat-based water products, DSWE includes pixels that are only partially covered by water to increase inundation dynamics information content. Previously published DSWE model development included one wetland-focused test developed through visual inspection of field-collected Everglades spectra. A comparison of that test’s output against Everglades Depth Estimation Network (EDEN) in situ data confirmed the expectation that omission errors were a prime source of inaccuracy in vegetated environments. Further evaluation exposed a tendency toward commission error in coniferous forests. Improvements to the subpixel level “partial surface water” (PSW) component of DSWE was the focus of this research. Spectral mixture models were created from a variety of laboratory and image-derived endmembers. Based on the mixture modeling, a more “aggressive” PSW rule improved accuracy in herbaceous wetlands and reduced errors of commission elsewhere, while a second “conservative” test provides an alternative when commission errors must be minimized. Replication of the EDEN-based experiments using the revised PSW tests yielded a statistically significant increase in mean overall agreement (4%, p = 0.01, n = 50) and a statistically significant decrease (11%, p = 0.009, n = 50) in mean errors of omission. Because the developed spectral mixture models included image-derived vegetation endmembers and laboratory spectra for soil groups found across the US, simulations suggest where the revised DSWE PSW tests perform as they do in the Everglades and where they may prove problematic. Visual comparison of DSWE outputs with an unusual variety of coincidently collected images for locations spread throughout the US support conclusions drawn from Everglades quantitative analyses and highlight DSWE PSW component strengths and weaknesses.

Rapid Assessment of Ecological Integrity for LTER Wetland Sites by Using UAV Multispectral Mapping

Long-term ecological research (LTER) sites need a periodic assessment of the state of their ecosystems and services in order to monitor trends and prevent irreversible changes. The ecological integrity (EI) framework opens the door to evaluate any ecosystem in a comparable way, by measuring indicators on ecosystem structure and processes. Such an approach also allows to gauge the sustainability of conservation management actions in the case of protected areas. Remote sensing (RS), provided by satellite, airborne, or drone-borne sensors becomes a very synoptic and valuable tool to quickly map isolated and inaccessible areas such as wetlands. However, few RS practical indicators have been proposed to relate to EI indicators for wetlands. In this work, we suggest several RS wetlands indicators to be used for EI assessment in wetlands and specially to be applied with unmanned aerial vehicles (UAVs). We also assess the applicability of multispectral images captured by UAVs over two long-term socio-ecological research (LTSER) wetland sites to provide detailed mapping of inundation levels, water turbidity and depth as well as aquatic plant cover. We followed an empirical approach to find linear relationships between UAVs spectral reflectance and the RS indicators over the Doñana LTSER platform in SW Spain. The method assessment was carried out using ground-truth data collected in transects. The resulting empirical models were implemented for Doñana marshes and can be applied for the Braila LTSER platform in Romania. The resulting maps are a very valuable input to assess habitat diversity, wetlands dynamics, and ecosystem productivity as frequently as desired by managers or scientists. Finally, we also examined the feasibility to upscale the information obtained from the collected ground-truth data to satellite images from Sentinel-2 MSI using segments from the UAV multispectral orthomosaic. We found a close multispectral relationship between Parrot Sequoia and Sentinel-2 bands which made it possible to extend ground-truth to map inundation in satellite images.

A watershed-scale model for depressional wetland-rich landscapes

Wetlands are often dominant features in low relief, depressional landscapes and provide an array of hydrologically driven ecosystem services. However, contemporary models do not adequately represent the role of spatially distributed wetlands in watershed-scale water storage and flows. Such tools are critical to better understand wetland hydrological, biogeochemical, and biological functions and predict management and policy outcomes at varying spatial scales. To develop a new approach for simulating depressional landscapes, we modified the Soil and Water Assessment Tool (SWAT) model to incorporate improved representations of depressional wetland structure and hydrological processes. Specifically, we refined the model to incorporate: (1) water storage capacity and surface flowpaths of individual wetlands and (2) local wetland surface and subsurface exchange. We utilized this model, termed SWAT-DSF (DSF for Depressional Storage and Flows), to simulate the ~289 km2 Greensboro watershed within the Delmarva Peninsula of the US Coastal Plain. Model calibration and verification used both daily streamflow observations and remotely sensed surface water extent data (ca. 2-week temporal resolution), allowing us to assess model performance with respect to both streamflow and watershed inundation patterns. Our findings demonstrate that SWAT-DSF can successfully replicate distributed wetland processes and resultant watershed-scale hydrology. SWAT-DSF provides improved temporal and spatial characterization of watershed-scale water storage and flows in depressional landscapes, providing a new tool to quantify wetland functions at broad spatial scales.

Depressional wetlands affect watershed hydrological, biogeochemical, and ecological functions

Depressional wetlands of the extensive U.S. and Canadian Prairie Pothole Region afford numerous ecosystem processes that maintain healthy watershed functioning. However, these wetlands have been lost at a prodigious rate over past decades due to drainage for development, climate effects, and other causes. Options for management entities to protect the existing wetlands, and their functions, may focus on conserving wetlands based on spatial location vis-à-vis a floodplain or on size limitations (e.g., permitting smaller wetlands to be destroyed but not larger wetlands). Yet the effects of such management practices and the concomitant loss of depressional wetlands on watershed-scale hydrological, biogeochemical, and ecological functions are largely unknown. Using a hydrological model, we analyzed how different loss scenarios by wetland size and proximal location to the stream network affected watershed storage (i.e., inundation patterns and residence times), connectivity (i.e., streamflow contributing areas), and export (i.e., streamflow) in a large watershed in the Prairie Pothole Region of North Dakota, USA. Depressional wetlands store consequential amounts of precipitation and snowmelt. The loss of smaller depressional wetlands (<3.0 ha) substantially decreased landscape-scale inundation heterogeneity, total inundated area, and hydrological residence times. Larger wetlands act as hydrologic "gatekeepers," preventing surface runoff from reaching the stream network, and their modeled loss had a greater effect on streamflow due to changes in watershed connectivity and storage characteristics of larger wetlands. The wetland management scenario based on stream proximity (i.e., protecting wetlands 30 m and ~450 m from the stream) alone resulted in considerable landscape heterogeneity loss and decreased inundated area and residence times. With more snowmelt and precipitation available for runoff with wetland losses, contributing area increased across all loss scenarios. We additionally found that depressional wetlands attenuated peak flows; the probability of increased downstream flooding from wetland loss was also consistent across all loss scenarios. It is evident from this study that optimizing wetland management for one end goal (e.g., protection of large depressional wetlands for flood attenuation) over another (e.g., protecting of small depressional wetlands for biodiversity) may come at a cost for overall watershed hydrological, biogeochemical, and ecological resilience, functioning, and integrity.

Decision-Tree, Rule-Based, and Random Forest Classification of High-Resolution Multispectral Imagery for Wetland Mapping and Inventory

Efforts are increasingly being made to classify the world’s wetland resources, an important ecosystem and habitat that is diminishing in abundance. There are multiple remote sensing classification methods, including a suite of nonparametric classifiers such as decision-tree (DT), rule-based (RB), and random forest (RF). High-resolution satellite imagery can provide more specificity to the classified end product, and ancillary data layers such as the Normalized Difference Vegetation Index, and hydrogeomorphic layers such as distance-to-a-stream can be coupled to improve overall accuracy (OA) in wetland studies. In this paper, we contrast three nonparametric machine-learning algorithms (DT, RB, and RF) using a large field-based dataset (n = 228) from the Selenga River Delta of Lake Baikal, Russia. We also explore the use of ancillary data layers selected to improve OA, with a goal of providing end users with a recommended classifier to use and the most parsimonious suite of input parameters for classifying wetland-dominated landscapes. Though all classifiers appeared suitable, the RF classification outperformed both the DT and RB methods, achieving OA >81%. Including a texture metric (homogeneity) substantially improved the classification OA. However, including vegetation/soil/water metrics (based on WorldView-2 band combinations), hydrogeomorphic data layers, and elevation data layers to increase the descriptive content of the input parameters surprisingly did not markedly improve the OA. We conclude that, in most cases, RF should be the classifier of choice. The potential exception to this recommendation is under the circumstance where the end user requires narrative rules to best manage his or her resource. Though not useful in this study, continuously increasing satellite imagery resolution and band availability suggests the inclusion of ancillary contextual data layers such as soil metrics or elevation data, the granularity of which may define its utility in subsequent wetland classifications.