You are currently viewing a new version of our website. To view the old version click .
MDPIMDPI
Search all articles
Water
Download PDF
  • Article
  • Open Access

11 May 2024

Vulnerability Assessment of Groundwater Influenced Ecosystems in the Northeastern United States

,
and
1
Department of Wildlife, Fisheries and Conservation Biology, University of Maine, 5755 Nutting Hall, Orono, ME 04469, USA
2
U.S. Geological Survey, Maine Cooperative Fish and Wildlife Research Unit, Orono, ME 04469, USA
3
School of Earth and Climate Sciences, University of Maine, Orono, ME 04469, USA
*
Author to whom correspondence should be addressed.
Water2024, 16(10), 1366;https://doi.org/10.3390/w16101366
This article belongs to the Special Issue Aquatic Ecosystems Health Assessment Using Biological and Geospatial Analyses
Version Notes
Order Reprints

Abstract

Groundwater-influenced ecosystems (GIEs) are increasingly vulnerable due to groundwater extraction, land-use practices, and climate change. These ecosystems receive groundwater inflow as a portion of their baseflow or water budget, which can maintain water levels, water temperature, and chemistry necessary to sustain the biodiversity that they support. In some systems (e.g., springs, seeps, fens), this connection with groundwater is central to the system’s integrity and persistence. Groundwater management decisions for human use often do not consider the ecological effects of those actions on GIEs. This disparity can be attributed, in part, to a lack of information regarding the physical relationships these systems have with the surrounding landscape and climate, which may influence the environmental conditions and associated biodiversity. We estimate the vulnerability of areas predicted to be highly suitable for the presence of GIEs based on watershed (U.S. Geological Survey Hydrologic Unit Code 12 watersheds: 24–100 km2) and pixel (30 m × 30 m pixels) resolution in the Atlantic Highlands and Mixed Wood Plains EPA Level II Ecoregions in the northeastern United States. We represent vulnerability with variables describing adaptive capacity (topographic wetness index, hydric soil, physiographic diversity), exposure (climatic niche), and sensitivity (aquatic barriers, proportion urbanized or agriculture). Vulnerability scores indicate that ~26% of GIEs were within 30 m of areas with moderate vulnerability. Within these GIEs, climate exposure is an important contributor to vulnerability of 40% of the areas, followed by land use (19%, agriculture or urbanized). There are few areas predicted to be suitable for GIEs that are also predicted to be highly vulnerable, and of those, climate exposure is the most important contributor to their vulnerability. Persistence of GIEs in the northeastern United States may be challenged as changes in the amount and timing of precipitation and increasing air temperatures attributed to climate change affect the groundwater that sustains these systems.

1. Introduction

Practices that promote ecosystem integrity and persistence are at the forefront of natural resource management in the face of climate change and anthropogenic activity [1,2,3]. Ecosystem modifications or loss have contributed to the recent worldwide species decline, described as the sixth great mass extinction event [4,5]. Consequently, efforts to mitigate loss of ecosystems and the species they support have been prioritized [6,7]. Freshwater ecosystems are some of the most threatened systems in the world, owing to anthropogenic effects on water quantity and quality [8,9,10]. Threats to these systems are diverse and occur at various spatial scales, which make their conservation and management especially challenging [11]. Frameworks that can help identify where freshwater ecosystems are threatened, and that consider groundwater connectivity, could inform proactive management and mitigation to enhance the resilience of these systems.
Freshwater ecosystems directly influenced by groundwater contributions (hereafter termed groundwater-influenced ecosystems or GIEs) can occur in a wide variety of forms (e.g., fens, springs, rivers, lakes). The environmental conditions in these ecosystems reflect their geological setting, climate conditions, and landscape position [12,13]. GIEs can stay saturated during times of drought [14] and can act as carbon capture or carbon runoff zones [15,16]. Additionally, groundwater-influenced systems can improve water quality in surface water systems by providing summertime cold water habitats for aquatic species, cooling summer water temperatures, supporting micro-fauna that promote the breakdown of contaminants [17], and by regulating nutrient cycling and decomposition of organic matter [18].
Shallow groundwater systems are inherently vulnerable to a variety of human practices, such as land use (e.g., farming, urbanization), which can create pollution that degrades ecosystems (e.g., contaminated runoff) and can result in excessive water use (e.g., irrigation, [19]). Disruptions in groundwater flow timing, volume, temperature, and composition can affect the integrity and persistence of GIEs [20,21,22]. For example, concentrated groundwater extraction from private and municipal wells in Oregon threatened >18% of GIE clusters (watersheds containing two or more types of groundwater-dependent ecosystems), while 70% of GIE clusters were threatened by groundwater contamination [21]. Additionally, GIEs are vulnerable to reduced precipitation and increased evapotranspiration attributed to climate change, leading to reduced groundwater recharge and increased groundwater withdrawals [22,23,24]. GIEs are connected to groundwater by local and regional flow paths that determine the sources of water that discharge to a GIE [19,25,26]. Local groundwater flow paths are more sensitive to changes in climate (i.e., precipitation, air temperature, evapotranspiration) than regional groundwater flow paths [27]. Additionally, upslope groundwater recharge areas connect GIEs to the surrounding landscape and watershed processes and affect the duration and amount of groundwater received by GIEs [25].
Groundwater extraction for anthropogenic needs is expected to increase with increasing human populations in the northeastern United States (U.S.) [28]. As one of the most densely populated areas of the U.S., the region’s landscapes are intensively modified with agriculture and urbanization [29], which may amplify effects of climate change by increasing surface temperatures [30,31]. These increased temperatures may affect groundwater recharge and water table depth [22], which could affect the occurrence, distribution, and condition of GIEs. Further, coastal aquifers are particularly vulnerable to groundwater extraction [28]. The northeastern U.S. includes >28,000 km of coastline where coastal aquifers and their associated GIEs may be at risk from this extraction [28]. A growing human population can further lead to increases in pollution, which can alter water chemistry, creating additional threats to a region’s GIEs [32].
Vulnerability can be defined as the degree to which a system is susceptible to, and unable to cope with, the combined effects of climate change and anthropogenic modifications [33]. Magness et al. [34] calculated vulnerability by combining estimates of exposure, system sensitivity, and the adaptive capacity of the system. Exposure is estimated by quantifying factors attributable to climate change, such as changes in air temperature and precipitation [34,35]. Factors that affect a system’s survival, persistence, fitness, or regeneration, such as land use, provide an aggregated estimate of sensitivity. Factors that promote adaptation responses, such as protected areas managed for conservation that can sustain ecosystem integrity, provide an aggregated estimate of adaptive capacity [34]. Within this framework, system vulnerability is estimated spatially by summing data layers representing the variables contributing to exposure, adaptive capacity, and sensitivity into a relative vulnerability score. This approach to estimating vulnerability provides a systematic and hypothesis-driven framework to examine factors influencing GIEs’ potential vulnerability.
We conducted a vulnerability assessment of areas predicted in the northeastern U.S. to be highly suitable for GIEs [36], using the Magness et al. [34] framework to identify vulnerable GIEs and watersheds. We estimated vulnerability at two spatial scales: 30 m pixels and U.S. Geological Survey Hydrologic Unit Code (HUC) 12 watersheds (24–100 km2). We identified areas predicted to be highly suitable for GIE occurrence [36] that also were predicted to be highly vulnerable, and we identified the input variables (i.e., exposure, sensitivity, adaptive capacity) with the most substantial contributions to GIE vulnerability. We further evaluated predicted vulnerability of areas in current conservation management that have predicted high suitability for GIE occurrence, as well as the vulnerability of watersheds surrounding those GIEs. By identifying the variables with the greatest contributions to the vulnerability scores, our results could inform management and conservation of groundwater-influenced ecosystems in the northeastern U.S.

2. Methods

2.1. Study Area

Our study extent spanned two EPA Level II ecoregions (Atlantic Highlands and Mixed Wood Plains; [37]; source: https://www.epa.gov/eco-research/ecoregion-download-files-region; accessed on 27 January 2021) with similar physical and biological conditions in portions of nine northeastern U.S. states (Connecticut, Massachusetts, Maine, New Hampshire, New Jersey, New York, Pennsylvania, Rhode Island, Vermont) (Figure 1). We combined the ecoregions into one extent for our analysis. Suitability at a 30 m resolution for GIEs within our study area was obtained from ensembled correlative distributions models created for the two EPA ecoregions [36]. These ensembled models used known locations of GIEs and environmental variables representing topography, geology, and vegetation to predict suitability for GIEs across the ecoregions.
Figure 1. Environmental Protection Agency (EPA) Level II Ecoregions [37] (Atlantic Highlands, Mixed Woods) in the northeastern United States (source: https://www.epa.gov/eco-research/ecoregions-north-america; accessed on 27 January 2022). State abbreviations: Connecticut (CT), Delaware (DE), Maine (ME), Maryland (MD), Massachusetts (MA), New Hampshire (NH), New Jersey (NJ), New York (NY), Ohio (OH), Pennsylvania (PA), Rhode Island (RI), Vermont (VT), Virginia (VA), West Virginia (WV), District of Columbia (DC).

2.2. Vulnerability Framework

We used the conceptual framework of Magness et al. [34] to estimate vulnerability of pixels in the study extent by calculating sensitivity (i.e., degree to which ecosystem survival, persistence, fitness, or regeneration may be affected by stressors), adaptive capacity (i.e., capacity of an ecosystem to cope with stressors, including adaptation responses), and exposure (i.e., effect of climate change experienced by the locale). Vulnerability was summed as (1) adaptive capacity − sensitivity = resilience and (2) exposure − resilience = vulnerability (Figure 2). We calculated vulnerability at the pixel (30 m × 30 m) and watershed extents [mean vulnerability score of pixels within HUC12 watersheds (24–100 km2; 4185 HUC12 watersheds in study area)].
Figure 2. Vulnerability conceptual framework [34] and the framework used to estimate groundwater-influenced ecosystem (GIE) vulnerability across our study area. Orange boxes in the conceptual framework represent the core components of estimating vulnerability (tan boxes). Source data for the variables are listed in Table 1. Abbreviations: Topographic Wetness Index (TWI).

2.3. Sensitivity

We calculated sensitivity with three variables (Table 1): developed land use, agricultural land use, and aquatic barriers. Agriculture can affect water quality by contributing nutrients (e.g., nitrogen and phosphorus) and pesticides to groundwater ecosystems [21,22]. Urbanization and agriculture can lead to increased groundwater extraction for public water supplies and irrigation [21,22]. We extracted 30 m pixels from urbanized and agricultural land uses by selecting land cover classes (developed, pasture/hay, and cultivated crops) within the National Land Cover Database (NLCD; source: https://www.mrlc.gov/data; 2019; accessed on 10 November 2022) that represented these land-use types. Pixels extracted from these land-use types were given a value between 0.5 and 1, corresponding with the weights assigned by McGarigal et al. [38] (developed-high intensity = 1, developed-medium intensity = 0.8, developed low-intensity = 0.5, pasture/hay = 0.5, cultivated crops = 0.5) and all other pixels within the study area were given a value of 0. Aquatic barriers represent the relative degree to which road–stream crossings and dams potentially impede upstream and downstream movement of water. Aquatic barriers (e.g., roads, dams) can alter the flow and temperature of surface water, which can reduce downstream recharge and decrease the thermal influence of groundwater [20,22]. We represented the effect of aquatic barriers by using an aquatic barriers dataset created for the Northeast [39] with values ranging from 0 (no aquatic barrier effect) to 1 (high aquatic barrier effect). We clipped the data layers to the extent of the study area, resampled the clipped layers to 30 m pixels, and summed the resampled data layers with Geographic Information Systems (GIS) software (ArcGis Pro v. 2.8.0; ESRI; Redland, CA, USA) to estimate sensitivity (Figure 2).
Table 1. Data used to estimate exposure, adaptive capacity, and sensitivity categories for GIEs under current climate and anthropogenic conditions. Variable abbreviations: minimum temperature (tmin), maximum temperature (tmax), precipitation (prcp), mean annual temperature (bio1), isothermality (bio3), mean temperature of driest quarter (bio9), mean temperature of warmest quarter (bio10), annual precipitation (bio12), precipitation of driest month (bio14), precipitation of warmest quarter (bio18), annual evapotranspiration (ET), potential evapotranspiration (PET), evapotranspiration in the growing season (ET-GS; May 15–September 15).

2.4. Adaptive Capacity

We calculated adaptive capacity with three variables: topographic wetness index, physiographic diversity, and hydric soils (Table 1). Topographic wetness index is the relative amount of moisture at any point in the landscape [38] contributed by up-gradient topography, which has been shown to be positively associated with GIEs in the northeastern U.S. [40]. Physiographic diversity is an estimated index of physiographic types [41]. Hydric soil affects landscape suitability for groundwater-influenced systems [36] and is represented as percentage of hydric soil in the gridded soil survey geographic database for the conterminous United States [42]. We clipped the data layers to the study area extent, resampled the clipped data to 30 m pixels, and summed the resampled data to estimate adaptive capacity (ArcGIS Pro v. 2.8, Redland, CA, USA) (Figure 2).

2.5. Exposure

We calculated climate exposure with climatic niche models (CNMs). These models use current geographic distribution data for ecosystems or species to infer climatic environmental requirements [43]. This modeling technique has been used to predict species or ecosystem range shifts under current and projected climate scenarios [44,45,46]. We used CNMs to model the current climatic niche of GIEs to calculate climate exposure of these ecosystems.

2.6. Geographic Distribution Data

We trained our CNM models by compiling location data for 3168 GIEs that were field-verified during 1981–2020 by state Natural Heritage Programs (see [36] for geographic distribution data sources). We reduced spatial autocorrelation in the dataset by removing occurrences within 2 km of other recorded locations, which removed 296 locations from the dataset. We reduced the effect of spatial sorting bias (SSB) with point-wise distance sampling that produced a subsample with SSB = 1. We used the final SSB-reduced dataset (1690 locations) for training the CNMs.

2.7. Climate Variables

We selected climate variables to include in our CNMs by reviewing literature describing effects of climate change on GIEs [19,22,28,47,48] (Table 1). We included ten climate variables in the CNMs that measured or estimated temperature, precipitation, and annual evapotranspiration (Table 1). We estimated seven bioclimatic variables (Table 1) with monthly Daymet V4 Daily Surface Weather and Climatogical summaries (1 km resolution; [49]). Maximum temperature (tmax), minimum temperature (tmin), and precipitation (prcp) during 1980–2019 were used to calculate the seven bioclimatic variables using the ‘biovars’ function within the ‘dismo’ R (v1.4.1106) package [50]. We estimated evapotranspiration (ET), potential evapotranspiration (PET), and evapotranspiration in the growing season (ET-GS) from MOD16A2 V.6 Terra Net Evapotranspiration 8-day Global dataset (https://lpdaac.usgs.gov/products/mod16a2gfv006/; accessed on 25 January 2023) within a web-available mapping system (Google Earth Engine, GEE; https://earthengine.google.com/; [51]) to obtain average annual rates during 2001–2019. All climate data layers were clipped to the extent of our study area and resampled to 1 km pixels if the source data were not 1 km resolution. We compared all variables with a Pearson correlation test and determined that none of the climatic variables were highly correlated (R2 < 0.60).

2.8. CNM Development and Evaluation

We selected two climatic niche modeling methods, maximum entropy (Maxent) and generalized additive models (GAM), which have been used frequently and have outperformed other climatic niche modeling methods [43,52,53,54]. No single modeling approach will perform best in every scenario [55], thus we developed two statistically contrasting models and integrated the predictions by calculating the mean suitability score for each pixel. Maxent is a machine learning modeling method that estimates suitability by finding the distribution that achieves maximum entropy given the environmental conditions at occurrence locations [56]. Generalized additive models smooth data to fit non-linear functions with non-parametric distributions. We developed our CNMs within the ‘dismo’ (Maxent; [50]) and ‘mgcv’ (GAM; [57]) packages in R (version 1.4.1106). We generated 20,000 pseudo-absence locations by randomly sampling across the study area extent. We partitioned occurrence (1690) and pseudo-absence data (20,000) into training (1352 presences and 16,902 background points) and testing (338 presences and 4225 background points) datasets with a K-fold cross validation with five folds. We evaluated CNM prediction accuracy with five metrics: area under the curve (AUC) estimates, Cohens Kappa statistic, sensitivity rates, specificity rates, and the true skill statistic (TSS). We evaluated CNM performance with each metric with thresholds to determine if the CNM was informative: AUC ≥ 0.70 [58], Cohens Kappa ≥ 0.50 [59], sensitivity and specificity rates ≥ 0.70, and TSS ≥ 0.50 [60].

2.9. Pixel-Scale Vulnerability Calculation

We scaled (0 to 1) 30 m data representing adaptive capacity and sensitivity and averaged the scaled pixel values separately for both. We subtracted the sensitivity pixel values from the adaptive capacity pixel values and scaled (0 to 1) the resulting pixel values to estimate resiliency in each 30 m pixel in the study extent (Figure 2). We used the CNM models, which produced climatic niche suitability scores from 0 (least suitable) to 1 (most suitable), to calculate exposure. We scaled the exposure estimate as 1 − suitability score, assigning the greatest exposure score to the smallest climate niche suitability. We resampled the 1 km exposure raster data to 30 m pixels, and we calculated the vulnerability raster as exposure − resilience, scaling the resultant pixels from 0 (least vulnerable) to 1 (most vulnerable). We classified the scaled vulnerability values into four vulnerability categories: least [0 ≤ value ≤ 0.25], low [0.25 < value ≤ 0.50], moderate [0.50 < value ≤ 0.75], and high [0.75 < value ≤ 1.0].

2.10. Land Ownership

We identified lands in conservation ownership within the study extent with the Protected Areas Database for the United States (PAD-US; https://www.usgs.gov/core-science-systems/science-analytics-and-synthesis/gap/science/protected-areas; accessed on 5 October 2022). The PAD-US database designates 10 land ownership types; we calculated the average vulnerability scores in areas in the Federal, Joint, Non-Governmental Organization, Private, and State ownership types. The PAD-US also assigns the type of conservation practices that occur within each protected area into four categories: (1) managed for biodiversity, where natural disturbance events proceed or are mimicked; (2) managed for biodiversity, where natural disturbance events are suppressed; (3) managed for multiple uses, including extractive (e.g., mining or logging) or off-highway motor vehicles (OHV) use; and (4) no known mandate for biodiversity protection. We calculated the average vulnerability scores in each protected area and then identified the number and total area of protected areas that had moderate [0.50 < value ≤ 0.75] or high vulnerability [0.75 < value ≤ 1.0] for each conservation practice type.

2.11. Landscape Suitability Model Comparison

We created a polygon layer from the ensembled model raster predicting areas that are suitable (80th percentile) for GIEs [36], and we used this polygon layer to extract (mask) pixels from the pixel-scale vulnerability raster to identify GIEs with vulnerability score > 0.50 (indicating vulnerable). We identified areas predicted to be suitable for GIEs that also contained vulnerable pixels (>0.50) that are within the United States Department of Agriculture’s (USDA) recommended 30 m wide conservation buffer around shallow groundwater (source: https://www.fs.usda.gov/nac/buffers/guidelines/1_water_quality/15.html; accessed on 10 March 2023). We calculated the average vulnerability score of pixels occurring within 30 m of each GIE-suitable polygon by overlaying the buffered (30 m) GIE polygons as a mask on the landscape vulnerability raster and then identified GIEs that are highly (value > 0.75) or moderately (0.50 < value ≤ 0.75) vulnerable.
We extracted patches of contiguous pixels (≥10 ha area) with high vulnerability (>0.75), converted the pixels to polygons, and calculated the distance between the edge of each GIE polygon and the nearest high-vulnerability polygon (mean = 54 ha, Standard Deviation = 301.4 ha). This allowed us to determine the number and total area of areas predicted to be suitable for GIEs that are within close proximity (<200 m) to highly vulnerable polygons. We calculated the proportion of GIE areas and highly vulnerable areas (>0.75 vulnerability) within each HUC12 watershed to identify watersheds with the greatest proportion of areas that are both suitable for GIEs and also vulnerable.

3. Results

3.1. GIE and Watershed Vulnerability

Approximately 34% of the study area (334,150 km2) is estimated to be at least moderately vulnerable (Table 2), and 54% of watersheds (representing 53% of the study area) are at least moderately vulnerable (Figure 3). Scaled estimates of adaptive capacity, exposure, and sensitivity are presented in Figure 4. Approximately two thirds of HUC12 watersheds are predicted to contain highly suitable conditions for GIEs, and areas predicted to be suitable for GIEs are relatively small (range: 0.2–1992 ha, mean = 2.4 ha, Standard Deviation = 11.4; [36]). Highly vulnerable areas are also small and vary in size (range: 0.12–31,969 ha, mean = 3.6 ha, Standard Deviation = 68.9). Of the HUC 12 watersheds predicted to contain highly suitable conditions for GIEs, 199 (representing 5% of total watershed area) contain GIEs with moderate vulnerability. Only 0.6% of GIEs (0.6% of total GIE area) are predicted to be highly vulnerable (Table 3) within 30 m of GIE edges. Approximately 26% of pixels with high suitability for GIEs (representing 28% of total GIE area) are within 30 m of land predicted to be at least moderately vulnerable. Areas predicted to be suitable for GIEs and within 30 m of highly vulnerable pixels vary in size (range: 0.8–394 ha, mean = 2.4 ha, Standard Deviation = 10.3) and generally are smaller than the highly vulnerable areas around them (range: 0.1–1217 ha, mean = 52.3 ha, Standard Deviation = 150.9). Of the 195,225 ha of area predicted to be suitable for GIEs that are within 30 m of moderately or highly vulnerable pixels, climate exposure and land-use practices (Table 4) are important drivers of vulnerability. Most (98.5%) areas suitable for GIEs are >200 m from large (≥10 ha) highly vulnerable areas of the landscape (Table 5).
Table 2. Total area (km2) and percentage of area of Hydrologic Unit Code 12 (HUC12) watersheds (24–100 km2) in the study area (Figure 1) and their partitioning by vulnerability score. Groundwater-influenced ecosystems (GIE).
Figure 3. Estimated 30 m pixel and Hydrologic Unit Code (HUC) 12 watershed vulnerability across the northeastern United States study area (Figure 1). Grey lines in the lower frame indicate HUC12 watershed boundaries.
Figure 4. Scaled estimates of landscape adaptive capacity, sensitivity, and exposure calculated to estimate groundwater-influenced ecosystem (GIE) vulnerability in the study area (Figure 1). The conceptual framework for estimating vulnerability [34] is provided in Figure 2.
Table 3. Vulnerability scores within 30 m of areas predicted by [36] to be suitable for groundwater-influenced ecosystem (GIE) occurrence in the northeastern United States study area (Figure 1). Vulnerability score range estimates are cumulative (e.g., ≥0.50 indicates 0.50–1 totals). The conceptual framework for estimating vulnerability [34] is provided in Figure 2.
Table 4. Number and proportion of areas predicted to be suitable for groundwater-influenced ecosystems (GIEs) [36] that are within ≤30 m of areas predicted to be vulnerable owing to sensitivity (land use) and exposure (climate) in the study area (Figure 1). Vulnerability score range estimates are cumulative (e.g., ≥0.50 indicates 0.50–1 totals). Variables combined to estimate sensitivity and exposure components of vulnerability are indicated in Table 1. The conceptual framework for vulnerability [34] is provided in Figure 2.
Table 5. Number and proportion of areas predicted to be suitable for groundwater-influenced ecosystems (GIEs) [36] and occurrences within cumulative distance (m) bands around the GIEs extending out of areas predicted to be highly vulnerable (score ≥ 0.75) to large patches (>10 ha) in the study area (Figure 1). The conceptual framework for the vulnerability [34] score is provided in Figure 2.

3.2. State Scale

New York, Pennsylvania, and Maine contain the greatest total land area that is moderately vulnerable (Figure 5; Table 6). Maine (64.7%), New York (46.2%), and New Hampshire (28%) contain the largest proportions of moderate landscape vulnerability by state (Figure 5; Table 6). The largest total area of watersheds that are at least moderately vulnerable are in New York (911; 51% of state), Maine (831; 76%), and Pennsylvania (517; 32%). Watersheds that are highly vulnerable (>0.75) occur in New York (14; 0.5% of state) and New Jersey (2; 0.1% of state) (Figure 5; Table 7). Climate exposure is the greatest contributor to high vulnerability of lands in Connecticut, and climate exposure and land-use practices are both contributors to moderate and high vulnerability in Maine, Massachusetts, New Hampshire, New Jersey, New York, Pennsylvania, Rhode Island, and Vermont. (Figure 6; Table 8).
Figure 5. Proportion of land area vulnerability predicted in 30 m pixels summarized by states in the study area’s Environmental Protection Agency (EPA) Level II Ecoregions (Atlantic Highlands, Mixed Woods) in the northeastern United States. EPA A Level III ecoregions in the study area are indicated in Figure 1. Values used to create pie charts can be found in Table 6.
Table 6. Area (km2) of land in vulnerability score ranges within the study region by state (Figure 1). Vulnerability score range estimates are cumulative (e.g., ≥0.50 indicates 0.50–1 totals). The conceptual framework for vulnerability [34] is provided in Figure 2.
Table 7. Area (km2) of HUC12 watersheds by vulnerability score range within the study region of each state (Figure 1). Vulnerability score range estimates are cumulative (e.g., ≥0.50 indicates 0.50–1 totals). The conceptual framework for vulnerability [34] is provided in Figure 2.
Figure 6. Proportion of the study area predicted to be highly vulnerable (≥0.75) owing to adaptive capacity, sensitivity (indicated by land use), and exposure (indicated by climate) within the Protected Areas Database for the United States (PAD-US), summarized by management type. High adaptive capacity does not appear on any pie charts because estimates are low or are 0 for those categories Management types: (1) managed for biodiversity—disturbance events proceed or are mimicked, (2) managed for biodiversity—disturbance events suppressed, (3) managed for multiple uses—subject to extractive (e.g., mining or logging) or OHV use, and (4) no known mandate for biodiversity protection (PAD-US Source: https://www.sciencebase.gov/catalog/item/602ffe50d34eb1203115c7ab). Values used to create pie charts can be found in Table 12. Variables combined to estimate sensitivity and exposure are indicated in Table 1.
Table 8. Areas (km2) predicted to be highly vulnerable (≥0.75) owing to adaptive capacity, sensitivity (land use) and exposure (climate) within the study region (Figure 1). Variables combined to estimate adaptive capacity, sensitivity, and exposure are indicated in Table 1. The conceptual framework for the vulnerability [34] score is provided in Figure 2. State abbreviations: Connecticut (CT), Maine (ME), Massachusetts (MA), New Hampshire (NH), New Jersey (NJ), New York (NY), Pennsylvania (PA), Rhode Island (RI), Vermont (VT).

3.3. Vulnerability of Protected Areas

State-owned lands account for the most land area with moderate or high vulnerability (Table 9). The majority (69%) of protected areas that are highly vulnerable have no mandate for biodiversity conservation (category 4), and 25% are managed for multiple uses (category 3; Table 10). Approximately 18% of the area predicted to be suitable for GIEs occurs in lands mapped in the PAD-US database, with the most (6.8%) occurring in the “managed for biodiversity with natural disturbance events suppressed” (category 2) and “managed for multiple uses” (7.7%) conservation types (Table 11). Climate exposure and land-use practices both contribute to high vulnerability in management categories 2, 3, and 4 (Figure 7; Table 12). For all land ownership types, vulnerability can be attributed to effects of both land use and climate exposure (Figure 8; Table 13). The majority of GIE area that has moderate or high vulnerability occurs within PAD-US protected areas in management categories 3 and 4 (Table 11).
Table 9. Distribution of landscape vulnerability scores calculated in 30 m pixels summarized by land ownership type in the study area (Figure 1). Vulnerability score range estimates are cumulative (e.g., ≥0.50 indicates 0.50–1 totals). The conceptual framework for the vulnerability [34] score is provided in Figure 2. Ownership types are described in the Protected Areas Database for the United States (PAD-US; https://www.usgs.gov/core-science-systems/science-analytics-and-synthesis/gap/science/protected-areas; accessed on 5 October 2022).
Table 10. Distribution of landscape vulnerability scores calculated in 30 m pixels and summarized by land management type in our study area (Figure 1). Vulnerability score ranges estimates are cumulative (e.g., ≥0.50 indicates 0.50–1 totals). The conceptual framework for the vulnerability [34] score is provided in Figure 2. Management types are described in the Protected Areas Database for the United States (PAD-US; https://www.usgs.gov/core-science-systems/science-analytics-and-synthesis/gap/science/protected-areas; accessed on 5 October 2022) and include (1) managed for biodiversity—disturbance events proceed or are mimicked, (2) managed for biodiversity—disturbance events suppressed, (3) managed for multiple uses—subject to extractive (e.g., mining or logging) or OHV use, and (4) no known mandate for biodiversity protection.
Table 11. Groundwater-influenced ecosystem (GIE) area within the moderate and high landscape vulnerability lands within the Protected Areas Database for the United States (PAD-US), summarized by management type. Management types: (1) managed for biodiversity—natural disturbance events proceed or are mimicked, (2) managed for biodiversity—disturbance events are suppressed, (3) managed for multiple uses—subject to extractive (e.g., mining or logging) or off-highway vehicle use, and (4) no known mandate for biodiversity protection (Source: https://www.sciencebase.gov/catalog/item/602ffe50d34eb1203115c7ab).
Figure 7. Proportion of land area predicted to be highly vulnerable (≥0.75) owing to adaptive capacity, sensitivity (indicated by land use), and exposure (indicated by climate) within states in Environmental Protection Agency (EPA) Level II Ecoregions (Atlantic Highlands, Mixed Woods) in the northeastern United States. High adaptive capacity and High sensitivity do not appear on any pie charts because estimates are low or are 0 for those categories. EPA Level III ecoregions in the study area are indicated in Figure 1. Values used to create pie charts can be found in Table 8. Variables combined to estimate sensitivity and exposure are indicated in Table 1.
Table 12. Variables contributing to highly vulnerable areas within the Protected Areas Database for the United States (PAD-US), summarized by management type. Management types: (1) managed for biodiversity—natural disturbance events proceed or are mimicked, (2) managed for biodiversity—disturbance events are suppressed, (3) managed for multiple uses—subject to extractive (e.g., mining or logging) or off-highway vehicle use, and (4) no known mandate for biodiversity protection (Source: https://www.sciencebase.gov/catalog/item/602ffe50d34eb1203115c7ab).
Figure 8. Proportion of the study area predicted to be highly vulnerable (≥0.75) owing to adaptive capacity, sensitivity (indicated by land use), and exposure (indicated by climate) within the Protected Areas Database for the United States (PAD-US), summarized by land ownership type. High adaptive capacity and High sensitivity do not appear on any pie charts because estimates are low or are 0 for those categories. Ownership types are described in the Protected Areas Database for the United States (PAD-US; https://www.usgs.gov/core-science-systems/science-analytics-and-synthesis/gap/science/protected-areas). Non-Governmental Organization (NGO). Values used to create pie charts can be found in Table 13. Variables combined to estimate sensitivity and exposure are indicated in Table 1.
Table 13. Variables contributing to highly vulnerable areas by land ownership types in the study area (Figure 1). The conceptual framework for the vulnerability [34] score is provided in Figure 2. Ownership types are described in the Protected Areas Database for the United States (PAD-US; https://www.usgs.gov/core-science-systems/science-analytics-and-synthesis/gap/science/protected-areas). Non-Governmental organization (NGO).

3.4. Climatic Niche Models

The Maxent and GAM CNMs, which represent exposure in the vulnerability equation, have large evaluation metrics for AUC, TSS, sensitivity, and specificity, but have small Kappa statistics (Table 14). The Maxent CNM outperformed the GAM CNM (Table 14). Precipitation in the warmest quarter, annual precipitation, precipitation in the driest month, and mean temperature in the driest quarter are the most influential climatic variables in the Maxent CNM (Table 15). Precipitation in the driest month, precipitation in the warmest quarter, annual mean temperature, and mean temperature in the warmest quarter are the most influential climatic variables in the GAM CNM (Table 15).
Table 14. Evaluation metrics for generalized additive model (GAM) and Maxent climatic niche models. Abbreviations: area under the curve (AUC), true skill statistic (TSS), Cohen’s Kappa statistic (Kappa).
Table 15. Pearson correlation and area under the curve (AUC) estimated relative variable importance (percent) of climatic variables used to estimate generalized additive model (GAM) and Maxent climatic niche models. Source data for the variables are provided in Table 1.

3.5. Discussion

Landscape vulnerability estimates revealed that nearly a third of the study area was predicted to have at least moderate vulnerability, and nearly 11% of the area predicted to be suitable for GIEs in the study area was predicted to be at least moderately vulnerable. GIEs receive water from direct precipitation, as well as overland and subsurface flows, and the quantity, timing, and quality of these flows can be affected by conditions in the landscape surrounding the GIE [19,25]. Conservation measures, such as riparian buffers to protect stream water quality, may be an effective approach for protecting resources that are important for GIEs, particularly if the size of the buffer reflects the conditions in the landscape surrounding a GIE [61,62]. However, the length of the groundwater flow path from the GIE to upslope areas can vary, and conservation buffers that do not account for that variation may not meet ecological requirements for all GIEs. Varying buffer sizes to reflect landscape conditions around focal ecosystems has precedence in best management practices to protect water quality for wetland and riparian conservation [61,62].
The contrast between pixel vulnerability and vulnerability summarized at the HUC12 watershed scale illuminates how scale can influence our understanding about how aquatic resources may be affected by environmental conditions in the surrounding landscape [63,64,65,66]. Approximately 770,000 ha of land area (2.4%) in the study region was predicted to be highly vulnerable at the pixel scale. However, when scaled to the watershed, only 0.3% of the watersheds were predicted to be highly vulnerable. Additionally, 26% of GIEs had moderately vulnerable pixels within 30 m, but 19% of GIEs occurred in watersheds that were moderately vulnerable. Less than 1% of the study area’s moderately or highly vulnerable HUC12 watersheds contained GIEs, and only 1.6% of GIEs occurred near (<200 m) large patches of highly vulnerable areas.
The northeast is warming faster than any other region in the continental United States [67] and continued increases in average annual air temperature and shifts in precipitation patterns could contribute to increased vulnerability of the region’s GIEs to climate exposure. Predicted changes in precipitation amount and frequency that affect water cycling, coupled with longer and warmer growing seasons, could alter the contribution of groundwater discharge to the region’s GIEs. Exposure, represented by climate, was the most important variable in Maine’s highly vulnerable watersheds. Since 1900, Maine’s average annual air temperature has increased 1.9 °C, and the length of growing seasons also has increased by 14 days [68], which may contribute to the estimated greater climate exposure. Nearly half (40%) of the moderately vulnerable GIEs were located within 30 m of areas predicted to be vulnerable to climate exposure. The northeastern United States is projected to have shorter, warmer winters [69]; increases in extreme precipitation and timing between rain events [70]; and longer, warmer growing seasons [70,71]. These projected changes in climate could alter the magnitude and timing of the spring freshet, increase evapotranspiration, increase runoff, and reduce infiltration. Our climatic niche model, which represented exposure, indicated that precipitation in the warmest quarter, annual precipitation, and precipitation in the driest month were the most influential climatic variables affecting suitability of areas in the landscape for GIE occurrence. Precipitation patterns that decrease warm-season (i.e., growing season) precipitation and increase cool-season (fall or winter) precipitation have been observed to increase groundwater temperatures [72]. Predicted increases in precipitation intensity can also lead to more surface run-off and, thus, alter the location and amount of groundwater recharge [73]. Changes in precipitation timing and frequency could play a large role in increasing climate exposure of GIEs in the Northeast.
Land-use-induced sensitivity contributed to highly vulnerable areas in nearly 1% of GIEs. Highly vulnerable areas in more than half of the states in the region were sensitive owing to land-use practices. Land-use practices can affect GIEs by lowering groundwater levels through groundwater extraction in watersheds [19,21]. The high vulnerability of watersheds in New York (5%) can be attributed primarily to sensitivity variables associated with land use. Urban development and agricultural land uses have been observed to reduce groundwater recharge, which has led to altered hydrological dynamics in the region [74]. Conversion to developed or agricultural land-use types could alter the amount and location of groundwater recharge, which may affect GIEs in the region. Additionally, land use may restrict the adaptive capacity of the region’s GIEs to respond to effects of the changing climate on groundwater.
Effects of climate and land use can be both additive and synergistic [75]. For example, agriculture is a leading cause of aquatic ecosystem impairment in the United States due to excessive nutrients in surface water runoff [76]. Additionally, prolonged periods of drought can lead to increased groundwater extraction for agriculture [22]. A combination of moderate climate exposure and moderate sensitivity to land-use practices contributed to high vulnerability scores of watersheds in our study area, where 14% of the land area is agriculture and 23% is developed lands. Land use and climate change have been observed to act synergistically to inhibit water retention [77], reduce water yield in river basins [78], and alter water cycles across landscapes [79], which could pose threats to GIEs and their persistence.
Land in the protected areas database that was highly vulnerable was at risk primarily owing to climate exposure, similar to the observations of [34], where climate exposure was the main contributing factor to vulnerability in approximately half of the National Wildlife Refuges across the United States. Private and jointly owned lands that contained highly vulnerable areas were primarily attributable to high sensitivity caused by land-use practices. The majority of these lands in our study area also had a no mandate for biodiversity conservation practice designation. In addition, the majority (62%) of the study area’s protected lands in the PADUS database classified as being moderately or highly vulnerable had no protection conservation designation (level 4), which may indicate that groundwater ecosystems in these areas (2.5% of total GIEs) may not be managed for conservation.
Landscape management of GIEs may be challenging, owing to the unique potential threats from watershed activities and ecological requirements of individual GIEs [14]. An adaptive management framework to conserve GIEs could address these challenges. As new information is acquired, an adaptive management approach provides an iterative process that allows uncertainties in cause and effect and ecological responses to be considered and addressed [14,80]. This iterative approach could provide opportunities to incorporate new spatial data in the vulnerability analysis to revise or update results. Land management of vulnerable landscapes and GIE conservation practices have largely been conducted in state-owned or federally owned landscapes where practitioners develop and apply land management actions [81]. Increasing demand for public drinking water and irrigation has led to widespread groundwater over-extraction and contamination [21,23], affecting human health and ecological services [82,83]. Few management practices can directly moderate variables contributing to exposure, such as increases in average annual temperatures and the seasonality of precipitation. However, reducing groundwater extraction by modifying agriculture irrigation practices and creating more sustainable municipal water-use practices can directly benefit GIEs [22]. Examples of potential management practices that could help maintain or improve GIE persistence and integrity in the landscape include: reducing the use of pesticides in agricultural lands to improve water quality [21], prioritizing the acquisition of lands with high geodiversity [84] and adaptive capacity [1] to enhance resilience, mitigate disturbances to natural recharge areas to maintain water quantities, and restoration management practices that directly restore degraded GIEs and watersheds.
Our approach to modeling GIE vulnerability is a hypothesis- and data-driven framework that explores the contributions of climate exposure, sensitivity caused by land use, and the adaptive capacity of the landscape on GIEs. The methods used are replicable and easily interpreted and can be applied in a wide range of geographic regions and for various ecosystem types. Despite these strengths, our approach is not encompassing of all the potential threats to GIEs in the Northeast. With the high human population density of the northeastern United States, spatial data on groundwater extraction rates could be an important contributing parameter to GIE vulnerability. Likewise, groundwater contamination is also a global problem that has a significant impact on human health and ecological services not included explicitly here [82,83] and that could have negative impacts on GIE vulnerability. Our results provide insight into vulnerable watersheds and sites, however we did not quantify the potential impacts of highly vulnerable upstream areas or watersheds on downstream locations, which may underestimate the vulnerability of those locations. To date, no such spatial data exist for the Northeast that describe the full extent of groundwater extraction and contamination. The vulnerability of GIEs in the northeastern U.S. to groundwater extraction and contamination due to increasing population demands and irrigation for agriculture is an area for future research. Temporal scales could also be included in assessing the vulnerability of GIEs in the Northeast, as climate factors have been observed to have significant seasonal characteristics [85] that could drive climate exposure.
Understanding the effects of climate change and anthropogenic disturbances on ecosystems has accelerated the development of methods to assess the ability of a system to cope with change [34,86,87,88]. Our analysis of landscape, watershed, and GIE vulnerability in the northeastern U.S. reveals contributing factors to vulnerability for individual sites, which could inform the conservation and prioritization of these systems. Our analysis indicates that the majority of GIEs in the region do not occur in currently vulnerable areas. However, those that are highly vulnerable are mainly vulnerable to climate exposure. This presents a challenge to maintaining the integrity and persistence of these GIEs in the northeastern United States, as climate change effects are projected to increase in the region and in landscapes across the world [67].

Author Contributions

S.D.S., A.S.R. and C.S.L. conceived the ideas and designed methodology; S.D.S. collected the data; S.D.S. analyzed the data; and S.D.S. led the writing of the manuscript. All authors have read and agreed to the published version of the manuscript.

Funding

This project was funded by the University of Maine and Maine Department of Inland Fisheries and Wildlife through the Cooperative Agreement with the U.S. Geological Survey, Maine Cooperative Fish and Wildlife Research Unit, Maine Agricultural and Forest Experiment Station, and the U.S. Geological Survey Science Support Program.

Data Availability Statement

Data layers created during this analysis are available at https://www.sciencebase.gov/catalog/item/660ab004d34e4df16bd5898f.

Acknowledgments

The authors thank the anonymous reviewers and B. Boxler for review of the manuscript. Data layers created during this analysis are available at https://www.sciencebase.gov/catalog/item/660ab004d34e4df16bd5898f [89]. Any use of trade, firm, or product names is for descriptive purposes only and does not imply endorsement by the U.S. Government.

Conflicts of Interest

We have no competing interests to disclose.

References

  1. Lindenmayer, D.; Hobbs, R.J.; Montague-Drake, R.; Alexandra, J.; Bennett, A.; Burgman, M.; Cale, P.; Calhoun, A.; Cramer, V.; Cullen, P. A Checklist for Ecological Management of Landscapes for Conservation. Ecol. Lett. 2008, 11, 78–91. [Google Scholar] [CrossRef] [PubMed]
  2. Folke, C.; Carpenter, S.R.; Walker, B.; Scheffer, M.; Chapin, T.; Rockström, J. Resilience Thinking: Integrating Resilience, Adaptability and Transformability. Ecol. Soc. 2010, 15, 20. [Google Scholar] [CrossRef]
  3. Wurtzebach, Z.; Schultz, C. Measuring Ecological Integrity: History, Practical Applications, and Research Opportunities. BioScience 2016, 66, 446–457. [Google Scholar] [CrossRef]
  4. Mccallum, M. Vertebrate Biodiversity Losses Point to Sixth Mass Extiction. Biodivers. Conserv. 2015, 24, 2497–2519. [Google Scholar] [CrossRef]
  5. Cowie, R.H.; Bouchet, P.; Fontaine, B. The Sixth Mass Extinction: Fact, Fiction or Speculation? Biol. Rev. 2022, 97, 640–663. [Google Scholar] [CrossRef]
  6. Brooks, T.M.; Mittermeier, R.A.; Da Fonseca, G.A.B.; Gerlach, J.; Hoffmann, M.; Lamoreux, J.F.; Mittermeier, C.G.; Pilgrim, J.D.; Rodrigues, A.S.L. Global Biodiversity Conservation Priorities. Science 2006, 313, 58–61. [Google Scholar] [CrossRef] [PubMed]
  7. Wilson, K.A.; McBride, M.F.; Bode, M.; Possingham, H.P. Prioritizing Global Conservation Efforts. Nature 2006, 440, 337–340. [Google Scholar] [CrossRef]
  8. Dudgeon, D.; Arthington, A.H.; Gessner, M.O.; Kawabata, Z.-I.; Knowler, D.J.; Lévêque, C.; Naiman, R.J.; Prieur-Richard, A.-H.; Soto, D.; Stiassny, M.L. Freshwater Biodiversity: Importance, Threats, Status and Conservation Challenges. Biol. Rev. 2006, 81, 163–182. [Google Scholar] [CrossRef]
  9. Dodds, W.K.; Perkin, J.S.; Gerken, J.E. Human Impact on Freshwater Ecosystem Services: A Global Perspective. Environ. Sci. Technol. 2013, 47, 9061–9068. [Google Scholar] [CrossRef]
  10. Reid, A.J.; Carlson, A.K.; Creed, I.F.; Eliason, E.J.; Gell, P.A.; Johnson, P.T.J.; Kidd, K.A.; MacCormack, T.J.; Olden, J.D.; Ormerod, S.J.; et al. Emerging Threats and Persistent Conservation Challenges for Freshwater Biodiversity. Biol. Rev. 2019, 94, 849–873. [Google Scholar] [CrossRef]
  11. Craig, L.S.; Olden, J.D.; Arthington, A.H.; Entrekin, S.; Hawkins, C.P.; Kelly, J.J.; Kennedy, T.A.; Maitland, B.M.; Rosi, E.J.; Roy, A.H.; et al. Meeting the Challenge of Interacting Threats in Freshwater Ecosystems: A Call to Scientists and Managers. Elem. Sci. Anthr. 2017, 5, 72. [Google Scholar] [CrossRef]
  12. Brown, J.; Wyers, A.; Bach, L.; Aldous, A. Groundwater-Dependent Biodiversity and Associated Threats: A Statewide Screening Methodology and Spatial Assessment of Oregon. In Groundwater-Dependent Biodiversity and Associated Threats: A Statewide Screening Methodology and Spatial Assessment of Oregon; The Nature Conservancy: Arlington, VA, USA, 2009. [Google Scholar]
  13. Blevins, E.; Aldous, A. Biodiversity Value of Groundwater-Dependent Ecosystems. Nat. Conserv. WSP 2011, 7, 18–24. [Google Scholar]
  14. Rohde, M.M.; Froend, R.; Howard, J. A Global Synthesis of Managing Groundwater Dependent Ecosystems under Sustainable Groundwater Policy. Groundwater 2017, 55, 293–301. [Google Scholar] [CrossRef] [PubMed]
  15. Glasser, S.P. USDA Forest Service Policy on Managing Groundwater Resources. Adv. Fundam. Sci. 2007, 1, 166. [Google Scholar]
  16. Eamus, D. Identifying Groundwater Dependent Ecosystems: A Guide for Land and Water Managers; Land & Water Australia: Canberra, Australia, 2009. [Google Scholar]
  17. Hoyos, I.P.; Krakauer, N.; Khanbilvardi, R. Random Forest for Identification and Characterization of Groundwater Dependent Ecosystems. WIT Trans. Ecol. Environ. 2015, 196, 89–100. [Google Scholar]
  18. Fauvet, G.; Claret, C.; Marmonier, P. Influence of Benthic and Interstitial Processes on Nutrient Changes along a Regulated Reach of a Large River (Rhône River, France). Hydrobiologia 2001, 445, 121–131. [Google Scholar] [CrossRef]
  19. Kløve, B.; Ala-aho, P.; Bertrand, G.; Boukalova, Z.; Ertürk, A.; Goldscheider, N.; Ilmonen, J.; Karakaya, N.; Kupfersberger, H.; Kvœrner, J.; et al. Groundwater Dependent Ecosystems. Part I: Hydroecological Status and Trends. Environ. Sci. Policy 2011, 14, 770–781. [Google Scholar] [CrossRef]
  20. Humphreys, W.F. Hydrogeology and Groundwater Ecology: Does Each Inform the Other? Hydrogeol. J. 2009, 17, 5–21. [Google Scholar] [CrossRef]
  21. Brown, J.; Bach, L.; Aldous, A.; Wyers, A.; DeGagné, J. Groundwater-Dependent Ecosystems in Oregon: An Assessment of Their Distribution and Associated Threats. Front. Ecol. Environ. 2011, 9, 97–102. [Google Scholar] [CrossRef]
  22. Kløve, B.; Ala-Aho, P.; Bertrand, G.; Gurdak, J.J.; Kupfersberger, H.; Kværner, J.; Muotka, T.; Mykrä, H.; Preda, E.; Rossi, P.; et al. Climate Change Impacts on Groundwater and Dependent Ecosystems. J. Hydrol. 2014, 518, 250–266. [Google Scholar] [CrossRef]
  23. Pérez Hoyos, I.C.; Krakauer, N.Y.; Khanbilvardi, R.; Armstrong, R.A. A Review of Advances in the Identification and Characterization of Groundwater Dependent Ecosystems Using Geospatial Technologies. Geosciences 2016, 6, 17. [Google Scholar] [CrossRef]
  24. Condon, L.E.; Atchley, A.L.; Maxwell, R.M. Evapotranspiration Depletes Groundwater under Warming over the Contiguous United States. Nat. Commun. 2020, 11, 873. [Google Scholar] [CrossRef] [PubMed]
  25. Winter, T.C.; Harvey, J.W.; Franke, O.L.; Alley, W.M. US Geological Survey Circular 1139. Ground Water Surf. Water A Single Resour. 1998, 50, 2–50. [Google Scholar]
  26. Haynes, A.B.; Briggs, M.A.; Moore, E.; Jackson, K.; Knighton, J.; Rey, D.M.; Helton, A.M. Shallow and Local or Deep and Regional? Inferring Source Groundwater Characteristics across Mainstem Riverbank Discharge Faces. Hydrol. Process. 2023, 37, e14939. [Google Scholar] [CrossRef]
  27. Briggs, M.A.; Harvey, J.W.; Hurley, S.T.; Rosenberry, D.O.; McCobb, T.; Werkema, D.; Lane Jr, J.W. Hydrogeochemical Controls on Brook Trout Spawning Habitats in a Coastal Stream. Hydrol. Earth Syst. Sci. 2018, 22, 6383. [Google Scholar] [CrossRef] [PubMed]
  28. Ferguson, G.; Gleeson, T. Vulnerability of Coastal Aquifers to Groundwater Use and Climate Change. Nat. Clim. Change 2012, 2, 342. [Google Scholar] [CrossRef]
  29. Noss, R.F.; LaRoe, E.T.; Scott, J.M. Endangered Ecosystems of the United States: A Preliminary Assessment of Loss and Degradation; US Department of the Interior, National Biological Service: Washington, DC, USA, 1995; Volume 28. [Google Scholar]
  30. Lamptey, B.L.; Barron, E.J.; Pollard, D. Impacts of Agriculture and Urbanization on the Climate of the Northeastern United States. Glob. Planet. Chang. 2005, 49, 203–221. [Google Scholar] [CrossRef]
  31. Eggleston, J.; Mccoy, K. Assessing the Magnitude and Timing of Anthropogenic Warming of a Shallow Aquifer: Example from Virginia Beach, USA. Hydrogeol. J. 2014, 23, 105–120. [Google Scholar] [CrossRef]
  32. Kaushal, S.S.; Groffman, P.M.; Likens, G.E.; Belt, K.T.; Stack, W.P.; Kelly, V.R.; Band, L.E.; Fisher, G.T. Increased Salinization of Fresh Water in the Northeastern United States. Proc. Natl. Acad. Sci. USA 2005, 102, 13517–13520. [Google Scholar] [CrossRef]
  33. Climate Change 2007: Impacts, Adaptation and Vulnerability. Contribution of Working Group II to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change—ProQuest. Available online: https://www.proquest.com/openview/10ac1638830c779bb315884da3366533/1?pq-origsite=gscholar&cbl=32284 (accessed on 21 April 2023).
  34. Magness, D.R.; Morton, J.M.; Huettmann, F.; Chapin III, F.S.; McGuire, A.D. A Climate-Change Adaptation Framework to Reduce Continental-Scale Vulnerability across Conservation Reserves. Ecosphere 2011, 2, 1–23. [Google Scholar] [CrossRef]
  35. Smit, B.; Wandel, J. Adaptation, Adaptive Capacity and Vulnerability. Glob. Environ. Change 2006, 16, 282–292. [Google Scholar] [CrossRef]
  36. Snyder, S.D.; Loftin, C.S.; Reeve, A.S. Predicting the Presence of Groundwater-Influenced Ecosystems in the Northeastern United States with Ensembled Models. Water 2023, 15, 4035. [Google Scholar] [CrossRef]
  37. Omernik, J.M.; Griffith, G.E. Ecoregions of the Conterminous United States: Evolution of a Hierarchical Spatial Framework. Environ. Manag. 2014, 54, 1249–1266. [Google Scholar] [CrossRef] [PubMed]
  38. McGarigal, K.; Compton, B.W.; Plunkett, E.B.; Grand, J. Designing Sustainable Landscapes: Development and Hard Development Settings Variables. 2020. Available online: https://umassdsl.org/data/ecological-integrity-metrics/ (accessed on 12 September 2022).
  39. Plunkett, E.B.; McGarigal, K.; Compton, B.W.; Jackson, S.D.; DeLuca, W.V.; Grand, J. Designing Sustainable Landscapes Products, Including Technical Documentation and Data Products. 2022. Available online: https://umassdsl.org/Data/ (accessed on 12 September 2022).
  40. Raney, P.A.; Leopold, D.J. Fantastic Wetlands and Where to Find Them: Modeling Rich Fen Distribution in New York State with Maxent. Wetlands 2018, 38, 81–93. [Google Scholar] [CrossRef]
  41. Theobald, D.M.; Harrison-Atlas, D.; Monahan, W.B.; Albano, C.M. Ecologically-Relevant Maps of Landforms and Physiographic Diversity for Climate Adaptation Planning. PLoS ONE 2015, 10, e0143619. [Google Scholar] [CrossRef] [PubMed]
  42. Soil Survey Staff, Natural Resources Conservation Service, United States Department of Agriculture. Soil Survey Geographic (SSURGO) Database for the Northeastern United States; United States Department of Agriculture: Washington DC, USA, 2020. [Google Scholar]
  43. Hijmans, R.J.; Graham, C.H. The Ability of Climate Envelope Models to Predict the Effect of Climate Change on Species Distributions. Glob. Chang. Biol. 2006, 12, 2272–2281. [Google Scholar] [CrossRef]
  44. Bradley, B.A.; Wilcove, D.S.; Oppenheimer, M. Climate Change Increases Risk of Plant Invasion in the Eastern United States. Biol. Invasions 2010, 12, 1855–1872. [Google Scholar] [CrossRef]
  45. Watling, J.I.; Romañach, S.S.; Bucklin, D.N.; Speroterra, C.; Brandt, L.A.; Pearlstine, L.G.; Mazzotti, F.J. Do Bioclimate Variables Improve Performance of Climate Envelope Models? Ecol. Model. 2012, 246, 79–85. [Google Scholar] [CrossRef]
  46. Fewster, R.E.; Morris, P.J.; Ivanovic, R.F.; Swindles, G.T.; Peregon, A.M.; Smith, C.J. Imminent Loss of Climate Space for Permafrost Peatlands in Europe and Western Siberia. Nat. Clim. Change 2022, 12, 373–379. [Google Scholar] [CrossRef]
  47. Kløve, B.; Allan, A.; Bertrand, G.; Druzynska, E.; Ertürk, A.; Goldscheider, N.; Henry, S.; Karakaya, N.; Karjalainen, T.P.; Koundouri, P.; et al. Groundwater Dependent Ecosystems. Part II. Ecosystem Services and Management in Europe under Risk of Climate Change and Land Use Intensification. Environ. Sci. Policy 2011, 14, 782–793. [Google Scholar] [CrossRef]
  48. Griebler, C.; Avramov, M.; Hose, G. Groundwater Ecosystems and Their Services: Current Status and Potential Risks. In Atlas of Ecosystem Services: Drivers, Risks, and Societal Responses; Schröter, M., Bonn, A., Klotz, S., Seppelt, R., Baessler, C., Eds.; Springer International Publishing: Cham, Switzerland, 2019; pp. 197–203. ISBN 978-3-319-96229-0. [Google Scholar]
  49. Thornton, M.M.; Shrestha, R.; Wei, Y.; Thornton, P.E.; Kao, S.-C.; Wilson, B.E. Daymet: Daily Surface Weather Data on a 1-Km Grid for North America, Version 4 R1 2020. Available online: https://daac.ornl.gov/DAYMET/guides/Daymet_Daily_V4.html (accessed on 20 January 2023).
  50. Hijmans, R.J.; Phillips, S.; Leathwick, J.; Elith, J. Dismo: Species Distribution Modeling; R Package Version 1.0-12; The R Foundation for Statistical Computing: Vienna, Austria, 2015. [Google Scholar]
  51. Gorelick, N.; Hancher, M.; Dixon, M.; Ilyushchenko, S.; Thau, D.; Moore, R. Google Earth Engine: Planetary-Scale Geospatial Analysis for Everyone. Remote Sens. Environ. 2017, 202, 18–27. [Google Scholar] [CrossRef]
  52. Jeschke, J.M.; Strayer, D.L. Usefulness of Bioclimatic Models for Studying Climate Change and Invasive Species. Ann. N. Y. Acad. Sci. 2008, 1134, 1–24. [Google Scholar] [CrossRef] [PubMed]
  53. Zhang, K.; Zhang, Y.; Tao, J. Predicting the Potential Distribution of Paeonia Veitchii (Paeoniaceae) in China by Incorporating Climate Change into a Maxent Model. Forests 2019, 10, 190. [Google Scholar] [CrossRef]
  54. Kiraç, A. Potential Distribution of Two Lynx Species in Europe under Paleoclimatological Scenarios and Anthropogenic Climate Change Scenarios. Cerne 2021, 27, e102517. [Google Scholar] [CrossRef]
  55. Qiao, H.; Soberón, J.; Peterson, A.T. No Silver Bullets in Correlative Ecological Niche Modelling: Insights from Testing among Many Potential Algorithms for Niche Estimation. Methods Ecol. Evol. 2015, 6, 1126–1136. [Google Scholar] [CrossRef]
  56. Phillips, S.J.; Anderson, R.P.; Schapire, R.E. Maximum Entropy Modeling of Species Geographic Distributions. Ecol. Model. 2006, 190, 231–259. [Google Scholar] [CrossRef]
  57. Wood, S.N. Fast Stable Restricted Maximum Likelihood and Marginal Likelihood Estimation of Semiparametric Generalized Linear Models: Estimation of Semiparametric Generalized Linear Models. J. R. Stat. Soc. Ser. B (Stat. Methodol.) 2011, 73, 3–36. [Google Scholar] [CrossRef]
  58. Baldwin, R.A. Use of Maximum Entropy Modeling in Wildlife Research. Entropy 2009, 11, 854–866. [Google Scholar] [CrossRef]
  59. Viera, A.J.; Garrett, J.M. Understanding Interobserver Agreement: The Kappa Statistic. Fam. Med. 2015, 37, 360–363. [Google Scholar]
  60. Allouche, O.; Tsoar, A.; Kadmon, R. Assessing the Accuracy of Species Distribution Models: Prevalence, Kappa and the True Skill Statistic (TSS). J. Appl. Ecol. 2006, 43, 1223–1232. [Google Scholar] [CrossRef]
  61. McElfish, J.M.; Kihslinger, R.L.; Nichols, S. Setting Buffer Sizes for Wetlands. Natl. Wetl. Newsl. 2008, 30, 6–17. [Google Scholar]
  62. Marczak, L.B.; Sakamaki, T.; Turvey, S.L.; Deguise, I.; Wood, S.L.R.; Richardson, J.S. Are Forested Buffers an Effective Conservation Strategy for Riparian Fauna? An Assessment Using Meta-Analysis. Ecol. Appl. 2010, 20, 126–134. [Google Scholar] [CrossRef] [PubMed]
  63. Allan, D.; Erickson, D.; Fay, J. The Influence of Catchment Land Use on Stream Integrity across Multiple Spatial Scales. Freshw. Biol. 1997, 37, 149–161. [Google Scholar] [CrossRef]
  64. Schiff, R.; Benoit, G. Effects of Impervious Cover at Multiple Spatial Scales on Coastal Watershed Streams1. JAWRA J. Am. Water Resour. Assoc. 2007, 43, 712–730. [Google Scholar] [CrossRef]
  65. Shi, P.; Zhang, Y.; Li, Z.; Li, P.; Xu, G. Influence of Land Use and Land Cover Patterns on Seasonal Water Quality at Multi-Spatial Scales. Catena 2017, 151, 182–190. [Google Scholar] [CrossRef]
  66. Shu, X.; Wang, W.; Zhu, M.; Xu, J.; Tan, X.; Zhang, Q. Impacts of Land Use and Landscape Pattern on Water Quality at Multiple Spatial Scales in a Subtropical Large River. Ecohydrology 2022, 15, e2398. [Google Scholar] [CrossRef]
  67. Karmalkar, A.V.; Bradley, R.S. Consequences of Global Warming of 1.5 °C and 2 °C for Regional Temperature and Precipitation Changes in the Contiguous United States. PLoS ONE 2017, 12, e0168697. [Google Scholar] [CrossRef]
  68. Fernandez, I.J.; Birkel, S.; Simonson, J.; Lyon, B.; Pershing, A.; Stancioff, E.; Jacobson, G.L.; Mayewski, P.A. Maine’s Climate Future: 2020 Update. 2020. Available online: https://digitalcommons.library.umaine.edu/cgi/viewcontent.cgi?article=1005&context=climate_facpub (accessed on 10 March 2023).
  69. Notaro, M.; Lorenz, D.; Hoving, C.; Schummer, M. Twenty-First-Century Projections of Snowfall and Winter Severity across Central-Eastern North America. J. Clim. 2014, 27, 6526–6550. [Google Scholar] [CrossRef]
  70. Melillo, J.M.; Richmond, T.; Yohe, G.W. Climate Change Impacts in the United States: The Third National Climate Assessment; U.S. Global Change Research Program: Washington, DC, USA, 2014. [Google Scholar]
  71. Karl, T.R.; Meehl, G.A.; Miller, C.D.; Hassol, S.J.; Waple, A.M.; Murray, W.L. Weather and Climate Extremes in a Changing Climate; US Climate Change Science Program: Washington, DC, USA, 2008. [Google Scholar]
  72. Brookfield, A.E.; Macpherson, G.L.; Covington, M.D. Effects of Changing Meteoric Precipitation Patterns on Groundwater Temperature in Karst Environments. Groundwater 2017, 55, 227–236. [Google Scholar] [CrossRef]
  73. Thomas, B.F.; Behrangi, A.; Famiglietti, J.S. Precipitation Intensity Effects on Groundwater Recharge in the Southwestern United States. Water 2016, 8, 90. [Google Scholar] [CrossRef]
  74. Wang, H.; Stephenson, S.R. Quantifying the Impacts of Climate Change and Land Use/Cover Change on Runoff in the Lower Connecticut River Basin. Hydrol. Process. 2018, 32, 1301–1312. [Google Scholar] [CrossRef]
  75. Santos, M.J.; Smith, A.B.; Dekker, S.C.; Eppinga, M.B.; Leitão, P.J.; Moreno-Mateos, D.; Morueta-Holme, N.; Ruggeri, M. The Role of Land Use and Land Cover Change in Climate Change Vulnerability Assessments of Biodiversity: A Systematic Review. Landsc. Ecol 2021, 36, 3367–3382. [Google Scholar] [CrossRef]
  76. Potter, K.; Douglas, J.; Bricj, E.; DeFries, R.S.; Asner, G.P.; Houghton, R.A. Impacts of Agriculture on Aquatic Ecosystems in the Humid United States. Ecosyst. Land Use Change Am. Geophys. Union 2004, 153, 31–40. [Google Scholar]
  77. Bai, Y.; Ochuodho, T.O.; Yang, J. Impact of Land Use and Climate Change on Water-Related Ecosystem Services in Kentucky, USA. Ecol. Indic. 2019, 102, 51–64. [Google Scholar] [CrossRef]
  78. Pham, H.V.; Sperotto, A.; Torresan, S.; Acuña, V.; Jorda-Capdevila, D.; Rianna, G.; Marcomini, A.; Critto, A. Coupling Scenarios of Climate and Land-Use Change with Assessments of Potential Ecosystem Services at the River Basin Scale. Ecosyst. Serv. 2019, 40, 101045. [Google Scholar] [CrossRef]
  79. Vaighan, A.A.; Talebbeydokhti, N.; Bavani, A.M. Assessing the Impacts of Climate and Land Use Change on Streamflow, Water Quality and Suspended Sediment in the Kor River Basin, Southwest of Iran. Env. Earth Sci. 2017, 76, 543. [Google Scholar] [CrossRef]
  80. Serov, P.; Kuginis, L.; Williams, J.P. Risk Assessment Guidelines for Groundwater Dependent Ecosystems, Volume 1—The Conceptual Framework; NSW Department of Primary Industries, Office of Water: Sydney, Australia, 2012. [Google Scholar]
  81. Keiter, R.B. Toward a National Conservation Network Act: Transforming Landscape Conservation on the Public Lands into Law. SSRN J. 2018, 42, 61. [Google Scholar] [CrossRef]
  82. Li, P.; Wu, J. Sustainable Living with Risks: Meeting the Challenges. Hum. Ecol. Risk Assess. Int. J. 2019, 25, 1–10. [Google Scholar] [CrossRef]
  83. Li, P. To Make the Water Safer. Expo Health 2020, 12, 337–342. [Google Scholar] [CrossRef]
  84. Beier, P.; Hunter, M.L.; Anderson, M. Conserving Nature’s Stage. Conserv. Biol. J. Soc. Conserv. Biol. 2015, 29, 613–617. [Google Scholar] [CrossRef]
  85. Hao, R.; Yu, D.; Liu, Y.; Liu, Y.; Qiao, J.; Wang, X.; Du, J. Impacts of Changes in Climate and Landscape Pattern on Ecosystem Services. Sci. Total Environ. 2017, 579, 718–728. [Google Scholar] [CrossRef]
  86. Côté, I.M.; Darling, E.S. Rethinking Ecosystem Resilience in the Face of Climate Change. PLoS Biol. 2010, 8, e1000438. [Google Scholar] [CrossRef] [PubMed]
  87. Olds, A.D.; Pitt, K.A.; Maxwell, P.S.; Connolly, R.M. Synergistic Effects of Reserves and Connectivity on Ecological Resilience. J. Appl. Ecol. 2012, 49, 1195–1203. [Google Scholar] [CrossRef]
  88. Mumby, P.J.; Chollett, I.; Bozec, Y.-M.; Wolff, N.H. Ecological Resilience, Robustness and Vulnerability: How Do These Concepts Benefit Ecosystem Management? Curr. Opin. Environ. Sustain. 2014, 7, 22–27. [Google Scholar] [CrossRef]
  89. Snyder, S.D.; Loftin, C.S.; Reeve, A.S. Vulnerability of Groundwater Influenced Ecosystems in the Northeastern United States: U.S. Geological Survey Data Release. 2024. Available online: https://www.sciencebase.gov/catalog/item/660ab004d34e4df16bd5898f (accessed on 10 March 2023).
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Article Metrics

Citations

Article Access Statistics

Journal Statistics
Multiple requests from the same IP address are counted as one view.
Download PDF
Spatiotemporal Variations in Snow Cover on the Tibetan Plateau from 2003 to 2020
Previous
Remote Measurement of Tide and Surge Using a Deep Learning System with Surveillance Camera Images
Next