Continental patterns of submarine groundwater discharge reveal coastal vulnerabilities

See allHide authors and affiliations

Science  12 Aug 2016:
Vol. 353, Issue 6300, pp. 705-707
DOI: 10.1126/science.aag1058

Water dissolving and water removing

Not all groundwater ends up flowing into rivers. Some is discharged directly into the ocean along the coast. Although much lower in volume than water transported by rivers, such submarine groundwater discharge can be a hidden source of dissolved ions, nutrients, or contaminants from human activities. Sawyer et al. performed a high-resolution continental-scale analysis of fresh groundwater discharge along the coastline of the United States. In total, more than one-fifth of coastal waters are vulnerable to groundwater-borne contamination.

Science, this issue p. 705


Submarine groundwater discharge (SGD) delivers water and dissolved chemicals from continents to oceans, and its spatial distribution affects coastal water quality. Unlike rivers, SGD is broadly distributed and relatively difficult to measure, especially at continental scales. We present spatially resolved estimates of fresh (land-derived) SGD for the contiguous United States based on historical climate records and high-resolution hydrographic data. Climate controls regional patterns in fresh SGD, while coastal drainage geometry imparts strong local variability. Because the recharge zones that contribute fresh SGD are densely populated, the quality and quantity of fresh SGD are both vulnerable to anthropogenic disturbance. Our analysis unveils hot spots for contaminant discharge to marine waters and saltwater intrusion into coastal aquifers.

Submarine groundwater discharge (SGD) influences global geochemical cycles and coastal water quality by delivering chemical compounds and dissolved ions from land to sea (1, 2). SGD includes two primary components: fresh, land-derived groundwater that infiltrates on land, and salty, ocean-derived groundwater that infiltrates offshore and returns to the sea (3). Although small in volume, fresh SGD exports naturally derived elements such as calcium and silicate at rates that rival those of rivers (4, 5). Fresh SGD is sensitive to human disturbance, and mixtures of fresh and saline SGD transport nutrients and other contaminants offshore. Therefore, the spatial distribution of fresh SGD has a direct impact on patterns of coastal water quality. High rates of nutrient-rich SGD, for example, contribute to harmful algal blooms and hypoxia (6, 7). SGD patterns also influence ocean temperature (8) and alkalinity (9), which are key controls on marine ecological and biogeochemical processes.

To understand the influence of fresh SGD on biogeochemical cycles and coastal water quality, rate assessments are needed at global and local scales. However, fresh SGD is difficult and costly to measure, and observations are scarce (8). Fresh SGD is heterogeneous and diffuse, unlike river discharge, which is concentrated at discrete and readily measurable points along the coast. Although many techniques exist to measure fresh SGD, most measurements are focused on a handful of well-studied, easily accessible locations. The majority of these locations are on the Atlantic coast of the United States (8) (Fig. 1). In the absence of measurements, water budgets have been used to map fresh SGD at high resolution over small coastal regions (10) and low resolution across the global oceans (11). These disparate scales of analysis leave critical gaps in our understanding of SGD. High-resolution estimates are needed across large regions to reveal relationships among climate, geology, and SGD and to identify where coastal water resources are vulnerable to degradation.

Fig. 1 SGD estimates from this study and the literature.

Our SGD values correlate well with local estimates (12) but are generally lower. Circles are from seepage meters, squares are from water budget analysis, and triangles are from multiple methods. Solid symbols indicate fresh SGD estimates. Open circles indicate total (fresh and saline) SGD. We calculated total SGD rates from our fresh rates using the relation SGDtotal = 1.1 SGDfresh + 470 m2/year (14, 30).

Here, we present high-resolution continental-scale estimates of fresh SGD across the contiguous United States. Our estimates are based on a simple water budget analysis and state-of-the-art continental-scale hydrography and climate data sets. Recharge zones, or contributing areas, for fresh SGD are defined with high-resolution hydrographic data. We assume that recharge zones are the wedge-shaped land areas outside stream catchments where water flows directly to the coast (Fig. 2, inset). We then assume that recharge across these coastal catchments is the component of precipitation that infiltrates and would become base flow to a stream, if a stream were present, but instead flows to the coast. This recharge rate is derived from three decades of reanalysis of climatic data (12).

Fig. 2 Map of fresh SGD rates along the contiguous United States coast.

On the West Coast, fresh SGD increases from south to north (point A to point B) with net precipitation (Net P) while drainage length (DL) remains consistent. The shape of recharge zones (inset in map) dictates local variability; recharge zone a has shorter DL, whereas zone b has longer DL. Fresh SGD is calculated from the product of infiltrating precipitation (I) and DL (DL=A/L, where A is recharge zone area and L is coastal length). Noninfiltrating runoff (R) does not contribute to fresh SGD. Expanded view of Cape Cod, Massachusetts shows coastal recharge zones colored by rate of fresh SGD.

To validate our estimates, we compiled 18 local estimates of SGD across the United States (12). We sought representation from diverse locations along the Pacific, Gulf, and Atlantic coasts and favored studies that used direct near-shore measurements to estimate the fresh component of SGD wherever possible. We excluded sea-based measurements using radon or radium tracer techniques, which can capture a large saline component of SGD (3, 13, 14). Sea-based measurements sometimes predict substantially larger SGD rates than near-shore measurements, simulations, and water budget–based estimates (14, 15). Our predicted SGD rates are correlated with local estimates but are consistently lower (Fig. 1). The magnitude of discrepancy shows no apparent relationship with geology, climate, land use, or population density. Some field measurements may overestimate fresh SGD, because sites are often selected where fresh SGD is likely to be focused (for example, in permeable sands or bay heads where groundwater flow paths converge). Although field measurements may overestimate fresh SGD, our approach likely underestimates them because coastal recharge zones could import groundwater from upland catchments (16). These potential additional groundwater sources are not included in our analysis. Furthermore, fresh SGD estimates from water balance approaches tend to decrease with increasing spatial resolution (17), and our analysis uses high-resolution hydrography data. The approach nevertheless allows for unprecedented mapping of fresh SGD.

At the local scale, our analysis exposes a strong heterogeneity in fresh SGD rates (Fig. 2, expanded map). This heterogeneity can be explained by the variability in land area that contributes groundwater to a given length of coastline. We define the coastal drainage length, which represents the average distance that groundwater travels from its point of recharge to the coast (Fig. 2, inset). The drainage length equals the recharge area for fresh SGD divided by the length of coastline where discharge occurs. It varies strongly with local topography and locations of coastal rivers (Fig. 2, expanded map). As a result, the spatial variability in fresh SGD over a typical 100-km segment of shoreline is almost as large as the variability at the continental scale. For example, the coefficient of variation for SGD within 100 km of San Francisco is 0.76, whereas its coefficient of variation along the entire West Coast is 1.10 (Fig. 2). Because of this strong local variability, SGD measurements at a single site cannot be extrapolated to other nearby sites with high confidence. Moreover, human modifications to coastal drainage networks affect patterns of fresh SGD. For example, fresh SGD rates are low in some areas of Florida with highly altered drainage networks (fig. S2).

At the continental scale, patterns in fresh SGD depend on both drainage geometry and climate. The influence of climate is clear along the West Coast from Southern California to Washington (Fig. 2), where net precipitation and fresh SGD both increase by more than 90%, but coastal drainage length is consistent. Meanwhile, the influence of drainage length is evident across East and West coasts. For example, net precipitation is similar in the Pacific Northwest and the mid-Atlantic, but fresh SGD rates are ~50% greater in the Pacific Northwest because of the abundance of long coastal drainage lengths in steep terrain (Fig. 2 and fig. S1).

At the continental scale, recharge areas for fresh SGD constitute a small portion of the total land area (0.4% of the contiguous United States). However, these areas drain more water than the continental interior on an areal basis because they receive more net precipitation. The total volumetric rate of fresh SGD from the contiguous United States to the oceans is 15 ± 4 cubic km/year, or <1% of total land runoff (18). Because this simple water budget analysis may underestimate fresh SGD by up to 40% (Fig. 1), rates may be as high as 25 ± 7 cubic km/year, or <2% of runoff. These continental-scale estimates are in line with previous estimates for the contiguous United States [1 to 10% of runoff (19)] and the world [6% of runoff (11)]. Note that saline SGD is substantially greater than the fresh component estimated here (3) and may be as large as 300 to 400% of global runoff (20), but the fresh component is most vulnerable to contamination and other anthropogenic disturbances. Although volumetrically small, fresh SGD can carry large contaminant mass loads. For example, in some parts of the world, fresh SGD delivers up to 30 times as much nitrogen to the coast as rivers (21, 22).

The average volumetric flux of fresh SGD per unit length of coastline is 420 m2/year, but rates span orders of magnitude (Fig. 3, inset). Although SGD is ubiquitous, concentrated discharge zones contribute the majority of fresh SGD to the oceans: Half of all fresh SGD is focused along only 14% of the coast. Interestingly, rates of fresh SGD follow a log-normal distribution (Fig. 3, inset), like permeability values for the shallow earth (23). Permeability is difficult to measure because it ranges by orders of magnitude and is scale-dependent. A particular strength of our analysis is that it does not require permeability data but relies instead on standardized topographic and climatic data sets.

Fig. 3 Coastal vulnerability map.

Vulnerability to offshore contamination (dark blue) is identified where the rate of fresh SGD is above average (420 m2/year) and developed or agricultural land use is above average (27.7%). Vulnerability to saltwater intrusion (magenta) is identified where low fresh SGD or high groundwater extraction may cause complete saltwater invasion. Light blue areas of coastlines are vulnerable to both offshore contamination and saltwater intrusion. Inset shows histogram of fresh SGD rates for the contiguous United States.

Most of the global population lives near and depends on coastal water resources and fisheries. Thus, high-resolution data sets are imperative for identifying coastal waters that may be vulnerable to “hidden” contaminant loads from fresh SGD. Within the contiguous United States, 3% of the population inhabits recharge areas for fresh SGD, which represent only 0.4% of the total land area. Although 72% of recharge areas were undeveloped as of 2011, conversion to agricultural and urban land use is ongoing. With coastal land development, nutrient loads to groundwater from septic tanks and fertilizers are increasing. Making matters worse, wetland loss reduces coastal resilience to contaminant loading because wetlands are efficient contaminant filters. Regions with above-average fresh SGD and land use development are particularly vulnerable to groundwater-borne contamination, and these regions represent 12% of the coastline. Vulnerable regions include the northern Gulf Coast from Mississippi to the Florida Panhandle, northern Atlantic Coast, and Pacific Northwest (Fig. 3). These regions have previously been shown to have high potential nitrogen inputs from SGD to coastal waters (24). Vulnerable regions should be monitored for direct nutrient inputs from fresh SGD.

High-resolution maps of SGD are also useful for assessing vulnerability of coastal aquifers to saltwater intrusion (Fig. 3). In populated areas, groundwater extraction subtracts from the recharge available for fresh SGD and can lead to saltwater intrusion. Most coastal aquifers are more sensitive to groundwater extraction than sea level rise (25). We predicted vulnerability to saltwater intrusion using an analytical solution for the position of the freshwater-saltwater interface (26, 27), assuming a population-dependent groundwater extraction rate that directly subtracts from the rate of fresh SGD (12). Coastlines are considered vulnerable where the toe of the freshwater-saltwater interface reaches the groundwater divide of the coastal aquifer, which implies imminent and full saltwater intrusion. These regions represent 9% of the coastline and include confirmed locations of saltwater intrusion such as Long Island, New York (28) and Los Angeles, California (29) (Fig. 3).

Because of the highly heterogeneous nature of fresh SGD, neighboring coastal zones can be vulnerable to saltwater intrusion and discharge of groundwater-borne contaminants to the ocean. Regions of mixed vulnerability include southeastern Florida and New Jersey, among many others. Only a small fraction of coastline (<1%) is dually vulnerable to both saltwater intrusion and offshore contamination, including the heavily developed and populated areas of San Francisco, Los Angeles, and Long Island (Fig. 3). If these urban areas rely primarily on groundwater to meet their resource demands, the resulting deductions to fresh SGD may cause full saltwater intrusion. Conversely, effective groundwater management may sustain fresh SGD rates, but contaminant loads to the coast may be high.

As the resolution of global hydrographic data improves, this same approach can be used to predict global distributions of fresh SGD and vulnerabilities in coastal water quality. Vulnerable regions will shift and likely grow with coastal land use change, population growth, climate change, and sea level rise. In many areas, rates of fresh SGD will decrease as impervious pavement expands and groundwater withdrawals increase. Regions with high rates of fresh SGD that are currently vulnerable to offshore contamination may instead become vulnerable to saltwater intrusion. Because rates of fresh SGD are highly heterogeneous, spatial estimates are imperative for identifying monitoring needs and assessing threats to coastal water quality on both sides of the land-sea boundary, in onshore aquifers and marine surface waters.

Supplementary Materials

Materials and Methods

Supplementary Text

Figs. S1 and S2

References (3153)

References and Notes

  1. See supplementary materials on Science Online.
Acknowledgments: We thank three anonymous reviewers for their suggestions, and M. Durand, H. Michael, C. Russoniello, and J. Heiss for discussions. Supported by the Jet Propulsion Laboratory, California Institute of Technology, under a contract with NASA, and grants from the NASA SWOT and Sea Level Science Teams (C.H.D. and J.S.F.); NSF grant EAR-1446724; and the Ohio State University School of Earth Sciences. The authors declare no competing interests. Fresh SGD rates and associated data are freely available at
View Abstract

Navigate This Article