PERSPECTIVE

PERSPECTIVE

Do geographically isolated wetlands influence

landscape functions?
Matthew J. Cohena,1, Irena F. Creedb, Laurie Alexanderc, Nandita B. Basud, Aram J. K. Calhoune,
Christopher Craftf, Ellen D’Amicog, Edward DeKeyserh, Laurie Fowleri, Heather E. Goldenj, James W. Jawitzk,
Peter Kallal, L. Katherine Kirkmanm, Charles R. Lanej, Megan Langn, Scott G. Leibowitzo, David Bruce Lewisp,
John Martonq, Daniel L. McLaughlinr, David M. Mushets, Hadas Raanan-Kiperwast, Mark C. Rainsu, Lora Smithm,
and Susan C. Wallsv

Edited by Dennis F. Whigham, Smithsonian Environmental Research Center, Edgewater, MD, and accepted by the Editorial Board
December 28, 2015 (received for review June 29, 2015)

Geographically isolated wetlands (GIWs), those surrounded by uplands, exchange materials, energy, and
organisms with other elements in hydrological and habitat networks, contributing to landscape functions,
such as flow generation, nutrient and sediment retention, and biodiversity support. GIWs constitute most
of the wetlands in many North American landscapes, provide a disproportionately large fraction of
wetland edges where many functions are enhanced, and form complexes with other water bodies
to create spatial and temporal heterogeneity in the timing, flow paths, and magnitude of network
connectivity. These attributes signal a critical role for GIWs in sustaining a portfolio of landscape functions,
but legal protections remain weak despite preferential loss from many landscapes. GIWs lack persistent
surface water connections, but this condition does not imply the absence of hydrological, biogeochemical,
and biological exchanges with nearby and downstream waters. Although hydrological and biogeochemical
connectivity is often episodic or slow (e.g., via groundwater), hydrologic continuity and limited evaporative
solute enrichment suggest both flow generation and solute and sediment retention. Similarly, whereas
biological connectivity usually requires overland dispersal, numerous organisms, including many rare or
threatened species, use both GIWs and downstream waters at different times or life stages, suggesting
that GIWs are critical elements of landscape habitat mosaics. Indeed, weaker hydrologic connectivity with
downstream waters and constrained biological connectivity with other landscape elements are precisely
what enhances some GIW functions and enables others. Based on analysis of wetland geography and
synthesis of wetland functions, we argue that sustaining landscape functions requires conserving the
entire continuum of wetland connectivity, including GIWs.
| |
connectivity navigable waters significant nexus

Understanding connectivity—patterns of matter, energy, of ecology and hydrology (1). Connectivity enables dis-
and organism exchanges among landscape elements persal of organisms and flows of water between land-
and across scales—is a challenge that unites the fields scape elements at multiple spatial and temporal scales

a
School of Forest Resources and Conservation, University of Florida, Gainesville, FL 32611; bDepartment of Biology, Western University, London,
ON, Canada N6A 5B7; cNational Center for Environmental Assessment, United States Environmental Protection Agency, Washington, DC 20460;
d
Department of Earth and Environmental Sciences, University of Waterloo, Waterloo, ON, Canada N2L 3G1; eDepartment of Wildlife, Fisheries, and
Conservation Biology, University of Maine, Orono, ME 04469; fSchool of Public Health and Environmental Affairs, Indiana University, Bloomington,
IN 47405; gCSS-Dynamac Corporation, Cincinnati, OH 45268; hSchool of Natural Resource Sciences, North Dakota State University, Fargo, ND
58108-6050; iOdum School of Ecology, The University of Georgia, Athens, GA 30602; jNational Exposure Research Laboratory, United States
Environmental Protection Agency, Cincinnati, OH 45268; kSoil and Water Science Department, University of Florida, Gainesville, FL 32611;
l
Region 4, United States Environmental Protection Agency, Athens, GA 30605; mJoseph W. Jones Ecological Research Center, Newton, GA 39870;
n
Department of Geographical Sciences, University of Maryland, College Park, MD 20742; oWestern Ecology Division, National Health and
Environmental Effects Research Laboratory, United States Environmental Protection Agency, Corvallis, OR 97333; pDepartment of Integrative
Biology, University of South Florida, Tampa, FL 33620; qCDM Smith, Inc., Indianapolis, IN 46204; rDepartment of Forest Resources and
Environmental Conservation, Virginia Polytechnic Institute and State University, Blacksburg, VA 24061; sNorthern Prairie Wildlife Research Center,
United States Geological Survey, Jamestown, ND 58401; tOffice of Wetlands, Oceans, and Watersheds, United States Environmental Protection
Agency, Washington, DC 20460; uSchool of Geosciences, University of South Florida, Tampa, FL 3362; and vWetland and Aquatic Research Center,
United States Geological Survey, Gainesville, FL 32653
Author contributions: M.J.C. and I.F.C. designed research; M.J.C. performed research; M.J.C., I.F.C., E. D’Amico, J.W.J., and C.R.L. analyzed data;
and M.J.C., I.F.C., L.A., N.B.B., A.J.K.C., C.C., E. D’Amico, E. DeKeyser, L.F., H.E.G., J.W.J., P.K., L.K.K., C.R.L., M.L., S.G.L., D.B.L., J.M., D.L.M., D.M.M.,
H.R.-K., M.C.R., L.S., and S.C.W. wrote the paper.
The authors declare no conflict of interest.
This article is a PNAS Direct Submission. D.F.W. is a guest editor invited by the Editorial Board.
Freely available online through the PNAS open access option.
1
To whom correspondence should be addressed. Email: mjc@ufl.edu.

1978–1986 | PNAS | February 23, 2016 | vol. 113 | no. 8 www.pnas.org/cgi/doi/10.1073/pnas.1512650113

(2), with implications for myriad processes, including gene trans- Regulatory Background
port (3, 4), solute and particle movement (5–7), biodiversity sup- Wetland protections in the United States derive from the Federal
port (8, 9), and pollutant retention (10). Multiple flow paths (11) Water Pollution Control Act (FWPCA) (33 USC §1251 et. seq.), also
among landscape elements create a spatially and temporally dy- known as the Clean Water Act (CWA), which establishes the
namic (12–15) portfolio of connections that vary in mode, strength, structure for federal agencies to regulate pollutant discharges into
timing, and direction (16). Landscape functions (e.g., water stor- “waters of the United States” and sets standards for surface water
age, base flow, sediment retention, nutrient cycling, biodiversity (though not groundwater) quality. Before 2001, virtually all wet-
support) emerge from a portfolio of connections (17, 18), which lands (i.e., those that met the regulatory definition) were consid-
can be enhanced (19, 20) and diminished (21, 22) by human activ- ered waters of the United States and federally regulated under the
ities. Quantifying these landscape functions (23, 24) and evalu- CWA. Challenges to federal authority over waters that did not
ating policies adopted to mitigate human impacts to them meet legal standards of “navigability” (45) culminated in two Su-
are imperative. preme Court decisions [Solid Waste Agency of Northern Cook
Wetlands are important landscape elements, providing hydro- County v. United States Army Corps of Engineers, 531 US 159
logical (e.g., floodwater storage and release) (25), biogeochem- (2001); and Rapanos v. United States, 547 US 715 (2006)] that more
ical (e.g., material transformation, carbon sequestration) (26–28), narrowly interpreted the CWA scope (46, 47). In the former, the
and biological (e.g., habitat, refugia) (29–32) functions. Variation Court held that the US Army Corps of Engineers (USACE) excee-
in wetland connectivity to other landscape elements impacts ded its statutory authority in asserting jurisdiction over non-
wetland functions, and thus which services are provided. Wetlands navigable, isolated, intrastate waters (later called “geographically
connected to other landscape elements via persistent surface isolated”) (40) based solely on use by migratory birds. In the latter,
flow paths provide a different blend of storage and water quality a concurrence opinion by Justice Anthony Kennedy limited federal
services than wetlands connected episodically or only via the sub- jurisdiction to wetlands only with a “significant nexus” to naviga-
surface (33, 34). Likewise, varying hydrologic connectivity impacts ble waters. The “significant nexus” term established a legal stan-
organism dispersal and thus community composition (35), biodi- dard for scientific evidence to assert CWA jurisdiction, intending
versity (36, 37), and metapopulation viability (38, 39). Human that protected wetlands possess a more than “speculative or in-
activities that decrease (e.g., dams/levees, land cover change) substantial” connection. Under that standard, a wetland is juris-
or increase (e.g., canals/ditches, human-aided dispersal) wetland dictional where it, either alone or in combination with “similarly
connectivity affect functions in individual wetlands and, by exten- situated” waters in the region, significantly affects the physical,
sion and combination, functions at the landscape scale. chemical, or biological integrity of downstream navigable waters.
Predicting connectivity of geographically isolated wetlands In 2015, in light of these decisions, the USACE and US Envi-
(GIWs) to other landscape elements is crucial to understanding ronmental Protection Agency (EPA) issued regulations defining
their landscape function (40). GIWs are those wetlands surrounded “waters of the United States” (the Clean Water Rule) (48). Based
by uplands. They span many wetland types and hydrogeomorphic
settings (e.g., vernal pools, playas, Delmarva and Carolina bays,
cypress domes, prairie potholes), grouped together by a shared
absence of persistent surface-water connections. Geographic iso-
B D
lation is often interpreted as functional isolation, especially out- C
40°N

side the scientific literature, but this interpretation is incorrect

(41, 42). GIWs are not hydrologically, biogeochemically, or bi-
A
ologically isolated (43) from other landscape elements or down-
stream waters. Rather, they span the entire continuum of
landscape connectivity, varying in mode, timing, duration, and H
30°N

magnitude (44), with antecedent moisture, geology, topogra- E

G
phy, land cover, and the specific material or organism. However, Ratio of Wetland Area
lack of persistent surface water connectivity to other elements to Total Area F
1
means they occupy the lower end of the connectivity contin-
0 110°W 100°W 90°W 80°W
uum, with functions controlled by episodic or slow transport
of water and solutes, or constrained dispersal of organisms. A B C D
Unambiguous generalizations about GIW connectivity with
downstream waters are untenable, leading some (44) to argue
the term is misleading. We retain it here as the default term
since 2002 because proximity to drainage features informs our
analyses and because it broadly represents wetlands imperiled
by recent judicial interpretations of US Clean Water Act E F G H
jurisdiction.
To address GIW impacts on landscape functions, we explore
hydrological, biogeochemical, and biological connectivity to nearby
and downstream waters. To support inferences about their functions,
we present an analysis of GIW geometry, size, and arrangement.
Using this analysis and literature synthesis, we evaluate the Fig. 1. Total wetland density (A–H, Upper) varies across the United
States. We selected eight 1,000-km2 blocks (Lower, showing spatial
premise that GIWs contribute to multiple landscape functions, subsets in A–H) to evaluate wetland resources (green) and other
suggesting important consequences of altering their prevalence aquatic habitats (blue) to analyze landscape patterns of geographically
or connectivity. isolated wetlands using data from the National Wetlands Inventory.

Cohen et al. PNAS | February 23, 2016 | vol. 113 | no. 8 | 1979
on extensive scientific literature review (49), the Clean Water Rule 1). These wetlands include California vernal pools (Fig. 1A), prairie
established eight categories of waters for CWA jurisdiction, six of potholes (Fig. 1B), basin wetlands (Fig. 1C), Maine vernal pools (Fig.
which are jurisdictional in all cases, without need for further 1D), playa lakes (Fig. 1E), cypress domes (Fig. 1F), coastal plain
analysis (traditional navigable waters, interstate waters, territorial wetlands (Fig. 1G), and pocosins (Fig. 1H). Other iconic GIW
seas, impoundments of waters of the United States, tributaries, landscapes (e.g., Nebraska Sandhills, Delmarva bays, Appalachian
adjacent waters), and two that can be jurisdictional if case-specific bogs) also merit attention. In each block, we analyzed only lacus-
analysis demonstrates a “significant nexus” with “navigable wa- trine and palustrine wetlands and merged contiguous wetlands.
ters.” Most GIWs fall in the latter. Although NWI limitations are well-documented (50, 51), with
A significant nexus criterion implies that some wetlands, alone omission errors seriously underrepresenting small wetland prev-
or in combination with other similar wetlands, do not significantly alence and extent (51), it provides the seamless, semantically
influence the physical, chemical, or biological integrity of down- consistent, national-scale dataset required to assess geographic
stream waters. One challenge in evaluating this contention is that patterns. Geographic isolation was defined where wetland Euclid-
few studies empirically document GIW connectivity, and time ean distance exceeded 10 m from line or polygon elements in the
variation thereof, across hydrological, biogeochemical, and bio- National Hydrography Dataset (NHD) (1:24,000 scale; nhd.usgs.
logical pathways (49). Further, aggregate functions from complexes gov/). This distance is based on reported NHD positional accuracy
of similarly situated wetlands, a crucial facet in the significant nexus (12 m) and is consistent with previous studies (52, 53). Although
test, remain poorly understood. Finally, the role of weak connec-
NHD neglects drainage network temporal variation (50), it is the
tivity (i.e., where material, energy, or organism exchange is slow
only viable national database for assessing stream proximity.
or episodic) is insufficiently considered or quantified. Justice
We enumerated the total number of wetlands, total wetland area,
Kennedy expressly included functions derived from weak con-
and total wetland perimeter and computed GIW contributions for
nectivity (e.g., slowing water and pollutant transport, limiting bi-
each (Table 1). In all blocks, most wetlands were GIWs, consistent
ological dispersal), stating that “it may be the absence of an
with local analyses (50). Although GIWs were a smaller portion of total
interchange of waters . . . that makes protection of the wetlands
wetland area, they contributed a greater proportion of total wetland
critical to the statutory scheme.” This reasoning implies a signifi-
perimeter in all blocks. Functions enhanced at wetland–upland
cant nexus test based on consideration of functions, not connec-
edges (54) are thus likely to be disproportionately delivered by GIWs.
tivity strength. Under such a test, arbitrary connectivity thresholds
To test the hypothesis that GIWs are small, we used logistic re-
for surface (i.e., based on frequency and duration) or subsurface
gression between wetland size (log-transformed to meet normality
pathways (i.e., based on travel times from distance and flow ve-
assumptions) and geographic isolation. Although fitted slopes dif-
locity) would be problematic. However, if the significant nexus test
fered across blocks (Table 1), the probability of geographic isolation
ultimately focuses on exceeding minimum thresholds in surface
always declined dramatically with increasing size (black lines in Fig. 2).
connectivity or falling below some subsurface travel time thresh-
old, many GIWs will lose legal protections, and landscape func- The odds ratio inferred from fitted slopes (Table 1) suggests that a
tions where weak connectivity increases functional value, or is log-unit increase in wetland size lowers the odds of geographic iso-
even a prerequisite, will be impacted. lation three- to eightfold. GIWs as a class are unambiguously small.
Wetland size controls edge density. Cumulative distributions
Wetland Geography of wetland area and perimeter consistently diverged (dashed and
Evaluating GIW landscape functions requires an inventory of stippled gray lines in Fig. 2), with perimeter rising faster than area,
their size, landscape position (vs. the drainage network and as expected. This divergence increases with wetland area variance
other wetlands), and geometry. Although GIWs exist in many and decreases where shape complexity is size-dependent. The
settings (40) across multiple wetland types, useful generaliza- relationship between area and perimeter-to-area ratio (P:A) (green
tions emerge from geographic analysis of wetland resources dots in Fig. 2) shows how ecotone length varies with size. Small
across representative landscapes. wetlands are more circular (in Fig. 2, dashed black line shows P:A
We obtained National Wetlands Inventory (NWI) (www.fws.gov/ for a circle) than larger wetlands, likely because of mapping res-
wetlands/) data for eight 1,000-km2 landscape blocks across the olution, because larger wetlands coalesce multiple depressions,
conterminous United States where well-known GIW types remain yielding complex shapes, and because human activities impact
prevalent (spatial subsets in Fig. 1 A–H, center coordinates in Table wetland shape (55). Despite nonlinear scaling of edges (Fig. 2),

Table 1. The number, area, and perimeter of wetlands, and associated proportions from geographically isolated wetlands (GIWs) in
eight domains (letters under GIW type are from Fig. 1)
Domain Total count Total area, Total perimeter, Odds
GIW type coordinates (% GIW) ha (% GIW) km (% GIW) Pr[GIW] ∼ area ratio

California vernal pools (A) 38.5N, 121.2W 2,163 (82.0) 1,337 (58.5) 584 (67.5) 5.25–1.15 × log(A) 0.317
Prairie potholes (B) 47.1N, 98.2W 6,417 (97.7) 9,509 (64.1) 2,068 (86.5) 7.58–1.09 × log(A) 0.336
Basin wetlands (C) 34.6N, 78.4W 6,507 (94.8) 36,097 (13.3) 5,638 (29.8) 10.9–1.99 × log(A) 0.136
Maine vernal pools (D) 43.5N, 70.6W 5,734 (68.2) 14,093 (16.5) 3,553 (32.0) 6.29–1.53 × log(A) 0.216
Playa lakes (E) 34.1N, 101.7W 420 (82.1) 2,841 (62.8) 386 (72.8) 38.8–7.59 × log(A) <0.001
Cypress domes (F) 29.9N, 82.2W 3,801 (77.6) 27,980 (16.9) 3,490 (40.0) 7.72–1.58 × log(A) 0.205
Coastal plain (G) 31.6N, 82.0W 1,966 (87.4) 63,360 (10.2) 6,606 (20.4) 7.89–1.36 × log(A) 0.258
Pocosins (H) 34.6N, 78.4W 1,387 (73.5) 46424 (7.1) 4,194 (17.1) 6.74–1.33 × log(A) 0.265

Logistic regression results describe the probability of geographic isolation (Pr[GIW]) as a function of wetland area (A) (m2); the odds ratio quantifies how the odds of
geographic isolation change given a unit increase in log(A). All regression results were statistically significant (P < 0.0001).

1980 | www.pnas.org/cgi/doi/10.1073/pnas.1512650113 Cohen et al.

 1.0 1.0 distances. Although GIWs span the entire continuum, their size
and spatial arrangement likely enable functions that wetlands that
0.1
0.5 are larger or nearer streams do not support. Further, GIWs exist
0.01 near other wetlands, likely forming complexes, and contributing
Pr(Geographic Isolated) and Cumulative Proportion

A E to aggregate functionality (57) even when individual wetlands are

0.0 0.001
far from streams.
1.0 1.0

0.1 Landscape Connectivity

0.5 Wetland are focal points for local catchments, and nodes within re-

Perimeter:Area (m-1)
0.01 gional networks. Some functions impact the physical, chemical, and
0.0
B F
0.001
biological integrity of downstream waters, falling under CWA juris-
diction. Other functions, such as carbon storage, microclimate regu-
1.0 1.0
lation, and endangered species habitat, fall outside that legal purview.
0.1 Although protecting GIWs based on their full array of functions may
0.5 be compelling, we consider here only hydrological, biogeochemical,
0.01 and biological connectivity impacts on “traditional navigable” waters.
C G
0.0 0.001
Hydrological Connectivity. A hydrologically isolated system
1.0 1.0
would be both endorheic (i.e., no outflow) and ombrotrophic (i.e.,
0.1 entirely precipitation-fed), conditions that do not represent GIW
0.5 hydrology. Although many natural GIWs (e.g., playa lakes, peat
0.01
bogs) are ombrotrophic, to also be endorheic requires precipitation
D H
0.0 0.001 and evapotranspiration to balance over the long term. Persistent
102 103 104 105 106 107 102 103 104 105 106 107 108 imbalances imply water flux across the system boundary, and thus
2
Wetland Area (m )
hydrologic connectivity. Whether that connectivity extends to
Fig. 2. Across blocks (A–H, maps in Fig. 1), the probability a wetland navigable waters is uncertain (e.g., outflows could evaporate before
is geographically isolated declines with increasing wetland area (solid reaching such waters). However, a cautious assumption given the
black lines). Small wetlands have higher perimeter-to-area ratios than
larger wetlands (green dots) although large wetlands depart more prevalence of wetland complexes (Fig. 3) is that flows eventually
from circular geometry (thick dashed black lines). The cumulative reach navigable waters despite indirect flow paths. A strictly
distribution of wetland perimeter (stippled gray lines) suggests that endorheic system also creates hypersaline conditions from incre-
small wetlands provide a greater fraction of landscape total mental evaporative solute enrichment. Although some GIWs,
perimeter (ecotone) than total area (dashed gray lines), with
especially some prairie potholes, do have salinity indicative of
implications for hydrologic, biogeochemical, and biological functions.
prolonged hydrologic isolation (58), most exhibit only modest ion
enrichment over rainwater (12, 58–63). It follows that GIWs connect
the inference that GIWs provide disproportionate edge density to the hydrologic network over space and timescales sufficient to
per area is clear. maintain low salinity, challenging assertions of hydrologic isolation.
Geographic separation (Euclidean distances) of wetlands vs. Wetlands are storage nodes in flow path networks (64), with
neighbors (nearest wetland) and vs. the drainage network (nearest the mode and strength of hydrological connectivity varying with
stream) yields insights on the prevalence of wetland complexes. time and across wetland settings (Fig. 4A). For some settings, such
Although flow path distances may be more informative for hy- as floodplain swamps, stream network connectivity is obvious and
drologic connectivity, measuring this distance requires high ac- rapid, occurring bidirectionally as water flows into floodplains at
curacy terrain and groundwater level data unavailable for all high river stage, and reverses at lower stage (Fig. 4A) (65). For
blocks. Nearest wetland (green dots in Fig. 3; all wetlands) and others, particularly GIWs that typically lack a persistent surface
stream distances (yellow dots in Fig. 3; GIWs only because non- connection, hydrological connectivity may be less obvious. It
GIW distances are <10 m by definition) were similar across blocks. occurs via unidirectional, episodic, and transient surface connec-
tions when depression storage is seasonally filled (e.g., vernal
Fitted nearest wetland distances followed exponential scaling
pools) (42, 65), or via slower moving subsurface flow paths (58, 66,
(except Playas), consistent with spatially random locations (gray
67). Despite uncertainty in quantifying timescales of hydrologic
lines in Fig. 3). Nearest wetland distances were also shorter than,
connectivity, GIWs have recently been shown to regulate (68, 69)
and uncorrelated with, nearest stream distances, suggesting
and stabilize (70) potentiometric gradients that generate base
stronger interactions with nearby wetlands, forming complexes
flow in streams. These subsurface flow paths may be hard to see
that impact landscape functions in aggregate. Nearest stream
(71), but they are not speculative or insubstantial connections.
distances depended on wetland and drainage density (black lines Indeed, they are often large and quantifiable at both field and
in Fig. 3). In some landscapes (cypress domes, coastal plain, po- landscape scales (41, 66, 72–74). Crucially, the timescales of such
cosins), stream distances were uniformly distributed up to ∼350 m connections are longer than for surface flow paths (75), manifest in
whereas a mode was evident in others (e.g., ∼170 m in basin base flow generation and water chemistry, implying potentially
wetlands and ∼810 m in prairie potholes) (Fig. 3). We infer that decadal delays in observing downstream effects of both wetland
GIW distances from the streams are larger than expected at ran- degradation and restoration activities.
dom, especially for playa lakes (Fig. 1), with implications for or- Hydrological connectivity is temporally dynamic (Fig. 4A) (14,
ganism dispersal and habitat functions (56). Although links 15, 76, 77). Rainfall activates flow paths absent under drier con-
between connectivity strength and distance are uncertain, wet- ditions (16), and generally accelerates flow velocities. This effect is
lands are clearly arrayed in a continuum of sizes and separation true across wetland types, with decreasing catchment travel times

Cohen et al. PNAS | February 23, 2016 | vol. 113 | no. 8 | 1981
 1.0 y = 0.23e-0.005x y = -0.04ln(x) + 0.29 declines where wetlands are lost (87, 88). As with hydrologic
R2 = 0.95 2
R = 0.31
y = 0.22e -0.004x functions, water quality functions likely vary with wetland con-
0.1
R2 = 0.93 nectivity and size (Fig. 4), but no systematic synthesis compares
0.01 2
GIWs to other wetland types across biogeochemical functions.
y = -2E-08x + 6E-05x + 0.015
A E 2
R = 0.55 Wetlands are important for sediment retention because low
0.001
flow velocities (83) enhance settling and because plant sediment
1.0 y = 0.31e-0.008x y = 1.28e-0.014x
R2 = 0.95 R2 = 0.96 stabilization limits resuspension (87, 88). Because of their size (Ta-
-0.003x
y = norm(814,937) y = 0.13e
0.1 R2 = 0.98
2
R = 0.97
ble 1) and landscape position (surrounded by uplands, distant from
Proportion of Wetlands

streams) (Fig. 3), GIWs generally receive the first landscape flush of
0.01 solutes and sediments, creating deposition and retention hot spots
B F (41, 89). Low surface connectivity in GIWs also limits subsequent
0.001
entrainment and export, providing long term storage (90).
1.0 y = norm(5,61) y = 0.38e-0.007x
R2 = 0.95 R2 = 0.98 Wetlands effectively retain nutrients, preventing downstream
y = norm(172,521) y = 0.07e-0.001x
0.1 R2 = 0.97 R2 = 0.85 transport. However, nutrient retention efficiency for GIWs vs.
other wetlands remains unknown, necessitating inference by
0.01 analogy to streams and lakes. Stream nutrient retention decreases
C G with increasing size because of variation in the following: (i)
0.001
y = 0.26e -0.007x
y = 0.31e-0.008x channel morphology that controls contact between solutes and
1.0
R2 = 0.94 R2 = 0.95 sediments (91); (ii) chemical gradients (i.e., concentration, redox
y = 0.08e-0.002x
0.1 y = 0.28e-0.006x R2 = 0.87 potential) (92) that controls reaction rates; and (iii) allochthonous
2
R = 0.95
inputs per unit storage (91, 93). Similar size-dependent function-
0.01
ality has been shown in lakes for plant biomass (94), organic and
D H mineral burial (95), species richness (96), and fish yield (97), im-
0.001
50 500 50 500 plying that small lakes contribute disproportionately to landscape
Distance (m) functions (98). Four attributes of GIWs suggest similar size-
Fig. 3. Distributions of wetland distance to nearest wetland neighbor dependent variation in biogeochemical reactivity. First, high primary
(green dots) and stream (yellow dots; GIWs only) for the eight production and anaerobic soils in GIWs enable retention of met-
landscape blocks (A–H, maps in Fig. 1). All fitted models (lines) were
als, nutrients, and pesticides in organic matter (99), and processes
statistically significant (P < 0.001). Results suggest that GIWs exist in
complexes with other wetlands, enabling cumulative landscape such as denitrification that remove nitrogen (100, 101). Second,
functions from interactions of numerous “similarly situated waters” like headwater streams and small lakes, GIWs dominate the total
even where stream distances are long. number of wetlands (Table 1) and generally occupy headwater
positions (Fig. 3). Thus, GIWs interact first with solute and particle
fluxes off the land, leading to dramatically enhanced reactivity
as landscape wetness increases (Fig. 4A) (15), a process long re-
(102). Further, GIWs likely exchange water and solutes with other
ferred to as variable source area runoff (72, 78). Such connectivity
wetlands before discharge to the drainage network; this wetlands-
variation may be most pronounced in wetlands where fill-and-spill
in-series configuration can enhance retention efficiency (41, 89).
flow dynamics dominate because groundwater flow is limited by
Third, GIWs are small (Fig. 2) (38, 64, 103), with high perimeter
low permeability aquifers (41, 72). Vernal pools, for example, ex-
length per unit area (Fig. 2). Size-dependent reactivity in streams
hibit episodic surface connectivity, when rainfall fills depression
(91) and lakes (95, 98) is controlled by edge-to-area geometry. By
storage, but slow groundwater connectivity during dry periods
analogy, because GIWs have high perimeter:area (Fig. 2), we ex-
(79). Such bimodal connectivity highlights the role of depression
pect commensurate increases in reactivity. An inverse correlation
storage in limiting peak flow frequency, magnitude, and duration between wetland size and water quality (104) supports this in-
and illustrates why snapshot assessments likely underestimate ference. Finally, long residence times due to intermittent or slow
connectivity (20, 80). connections (105, 106) facilitate completion of kinetically limited
Watershed responses are partly controlled by wetland number, reactions (e.g., P sorption into minerals, complex organic molecule
area, and distribution, as well as connection paths (i.e., surface vs. mineralization), enhancing sink functions. Although maximum re-
subsurface) and velocities (Fig. 4B) (16, 81). Wetlands connected via tention efficiency occurs when reaction rates and residence times
perennial surface flow paths contribute dynamic storage during align (107), loss of high reactivity and long residence time land-
high flows (green in Fig. 4). In comparison, wetlands connected via scape elements alters overall fluxes, particularly when GIWs are
intermittent fill-and-spill dynamics (yellow in Fig. 4) or via subsurface embedded in solute-generating areas (e.g., where fertilizer is ap-
flow paths only (red in Fig. 4) constrain peak flow volumes, delay plied) (101). Timescales for detecting changes may be long (75),
peak timing, impact recession rates, and control base flow (82). indicating impacts principally to base flow chemistry. Inferences
Indeed, recent modeling suggests that water storage in GIWs im- based on nutrients apply to other contaminants [metals (108) and
pacts downstream flow (69) and enables groundwater exchange, pesticide (109)], where retention is enabled by low redox conditions
ultimately buffering stream flow variation (70). Watershed discharge and organic matter storage, common features of all wetland set-
integrates the entire continuum of hydrological connectivity, not tings, including GIWs (28).
just rapid or surface-connected flow paths. Wetlands spanning the entire connectivity continuum protect
water quality, with GIWs likely playing important roles in sediment
Biogeochemical Connectivity. Wetlands are hotspots for sedi- retention, base flow chemistry, and solute retention where resi-
ment deposition (83), nutrient retention and transformation (84, 85), dence time is a key determinant of retention efficiency. Moreover,
organic matter cycling and storage (27, 28), and metal and pes- it is precisely by performing functions along slow-velocity flow
ticide (86) immobilization. Predictably, downstream water quality paths to the drainage network, an attribute interpreted as weak

1982 | www.pnas.org/cgi/doi/10.1073/pnas.1512650113 Cohen et al.

 Geographically Isolated: community composition (116), with historical connectivity imprinted
intermittent surface
connectivity on contemporary diversity patterns (117). Geographic isolation se-
lects plants with long-lived seeds (118) or long-distance dispersal
A Channel Connected: strategies (e.g., via fauna or wind) (113), and animals such as am-
frequent/persistent
surface connectivity phibians that require competitor exclusion for all or part of their life
cumulative
probability

cycles (119), or that rely on dynamic heterogeneity in aquatic re-

sources (120, 121). For example, increased nearest wetland distance
reduces local species richness of both native and nonnative fauna
travel time (log scale) (36) and flora (113) but enhances both landscape biodiversity (122,
123), by creating taxonomically distinct communities, and meta-

cumulative
Geographically Isolated:

probability
infrequent/absent increasing population stability (124), by creating spatial heterogeneity in the
surface connectivity watershed
wetness drivers of subpopulation dynamics. Frequent dry conditions and the
absence of persistent surface connectivity can preclude fish pop-
travel time (log scale) ulation establishment or recruitment after extirpation (35, 125). Fish
preclusion has implications for survival of juvenile amphibian, crus-
cumulative

tacean, and insect fauna (32, 126–128), many of which disperse as

probability

Surface Flowpath
adults to upland habitats or downstream water bodies. Indeed,
Stream or other
Navigable Water
Groundwater Flowpath models predict that loss of GIW habitats would impact a wide array
travel time (log scale)
of fauna, not just permanent residents, and most prominently, tur-
B tles, amphibians, birds, and small mammals (36), many of which are
imperiled (38). That many organisms use both GIWs and down-
cumulative probability

w
flo ; stream waters [turtles (129), birds (130, 131), snakes (132), and alli-
rm ion
Sto erat lain
n gators (30)], in different seasons (121) or life stages (30, 133),
ge oodp logy
Fl rpho ia; tion
;
mo fug en ; illustrates that the entire connectivity continuum, including GIWs,
Re t ret age ion
lat
m
n
e sto r
r e gu ion impacts habitat heterogeneity and redundancy, regional biodiver-
di k le la t
Se Ban tab egu sity, and, thus, the biological integrity of downstream waters.
ter te r
Wa solu

travel time (log scale)

Humans Alter Connectivity
Although losing wetland area poses the most obvious threat,
Fig. 4. (A) Hydrological connectivity between wetlands and downstream
changing wetland connectivity also impacts landscape functions,
waters varies in space and time. Watershed travel time distributions
are shortest for wetlands with persistent surface connectivity (green), and humans have altered the continuum of connectivity (134).
intermediate (and bimodal) for wetlands with episodic surface Modifications to reduce flood risks increase connectivity in head-
connectivity (gold), and longest for wetlands without surface connections water areas (e.g., via agricultural ditches, hardened urban streams,
(red). Travel time distributions decline with increased watershed
and drainage canals) and reduce connectivity downstream (e.g., via
wetness. (B) Convolution of travel times means watershed exports are
impacted by wetlands spanning the connectivity continuum. Differential levees disconnecting river and floodplain). Coupled to systematic
protections across wetlands alter connectivity patterns and impact losses of some network elements (135), especially small wetlands
watershed functions and related ecosystem services. (55), these changes impact the time and space variation of con-
nectivity (66). Wetland losses are strongly biased toward removing
wetlands far from drainage features, changing network topology
connectivity, that high rates of solute and particulate matter re-
(55), and impacting services derived from longer residence times,
tention are enabled.
including sediment storage (87), nutrient and floodwater retention
(25, 87), and controls on flood timing and magnitude (136). Net-
Biological Connectivity. Variation in wetland connectivity affects
work changes also impact aquatic ecosystem structure (137) and
biota and, thus, ecological condition (60, 61) within and across
composition (24, 113) by affecting dispersal (17, 138, 139). Land use
wetlands (110–113). However, biological connectivity, unlike wa-
intensification (i.e., cropping, urbanization) also leads to more
ter and solutes, is not always constrained by flow direction. For
regular wetland shapes (55), lowering perimeter-to-area ratios and
organisms that move over land or through air, connections among
impacting functions associated with wetland edges (140). By short-
nodes are constrained by terrestrial landscape properties and circuiting or removing GIW storage, water, solute, and sediment
species traits, rather than hydrologic flow paths (111). Organism retention functions are lost. By changing distances between wet-
movement among habitat patches occurs in response to eco- lands (e.g., by removing GIWs), preferentially protecting large or
logical processes operating at multiple spatial scales. Travel dis- perennially connected wetlands, and changing land cover between
tances also vary in time, responding strongly and nonlinearly to wetlands, crucial biological functions have also been lost (43).
climate forcing (114), fire and other natural disturbances, and
human impacts. Wetlands contiguous to surface dispersal path- Science and Policy Challenges
ways (e.g., streams, flyways) or near refugia (e.g., lacustrine hab- Conservationist and writer Aldo Leopold once remarked in refer-
itats) differ in community structure from wetlands where dispersal ence to species protection that “to keep every cog and wheel is
is restricted, or desiccation more frequent, either due to shallow the first precaution of intelligent tinkering.” We apply this advice
basin form or exclusively subsurface hydrologic connectivity (113). to protecting healthy landscapes. Wetlands connect to other
Geographic isolation does not imply biological isolation (115) water bodies, including downstream waters along a time-varying
but can constrain aquatic plant and animal movement. Spatial and continuum of connectivity, from permanent to transient, from fast
temporal heterogeneity in the frequency, timing, and duration to slow, and from principally surface to exclusively subsurface flow
of connectivity affects water-mediated movement and thus paths. Landscape functions arise from the cumulative effects of the

Cohen et al. PNAS | February 23, 2016 | vol. 113 | no. 8 | 1983
full complement of connection types and strengths. As such, se- and quantifies the frequency, timing, and duration of wetland
lectively eliminating some connection types and protecting others connectivity across multiple flow paths, and the myriad ways
inadvertently prioritizes some functions over others, without ade- in which weak or slow connectivity is important, a logical and
quate rationale. The significant nexus test explicitly requires reg- precautionary inference is that all wetlands influence landscape
ulators and the regulated community to evaluate the functions that functions.
GIWs provide. Our analysis and synthesis suggests that GIWs,
which generally have less frequent or slower hydrologic connec- Acknowledgments
tions than other wetlands, support a multitude of landscape This paper arose from a “Geographically Isolated Wetlands Re-
functions, enhancing many, and provide some that other wetlands search Workshop” cohosted by the US Environmental Protection
do not. As such, there is no obvious and nonarbitrary connectivity Agency Office of Research and Development and the Joseph W.
threshold (e.g., based on travel distance or time) to designate Jones Ecological Research Center in Newton, GA, November 18–
protections for downstream waters. Even before the Supreme 21, 2013. We thank the organizers. Donna Downing, Barbara
Court decisions to limit federal protections, many GIWs were lost Bedford, and Arnold van der Valk provided valuable comments on
(52). Those that remain are imperiled by alterations to their ge- an early draft. Information in this document has been funded in
ometry, connectivity (141, 142), surrounding land cover, and part by the US Environmental Protection Agency. This manuscript
now legal protections. Although the consequences of these has been subjected to agency review and has been approved for
changes require further research, GIW losses alter the portfolio of publication. The views expressed in this journal article are those of
landscape connectivity with negative effects on downstream the authors and do not necessarily reflect the views or policies of
waters. As the scientific community increasingly recognizes the US Environmental Protection Agency.

1 Tetzlaff D, et al. (2007) Connectivity between landscapes and riverscapes: A unifying theme in integrating hydrology and ecology in catchment science. Hydrol
Processes 21:1385–1389.
2 Peters DPC, et al. (2008) Living in an increasingly connected world: A framework for continental-scale environmental science. Front Ecol Environ 6:229–237.
3 Hughes JM, Schmidt D, Finn DS (2009) Genes in streams: Using DNA to understand the movements of freshwater fauna and their riverine habitat. Bioscience
59:573–583.
4 Bertuzzo E, et al. (2011) Prediction of the spatial evolution and effects of control measures for the unfolding Haiti cholera outbreak. Geophys Res Lett 38:L06403.
5 Creed IF, Sanford SE, Beall FD, Molot LA, Dillon PJ (2003) Cryptic wetlands: Integrating hidden wetlands in regression models of the export of dissolved organic
carbon from forested landscapes. Hydrol Processes 17:3629–3648.
6 Creed IF, Beall FD (2009) Distributed topographic indicators for predicting nitrogen export from headwater catchments. Water Resour Res 45:W10407.
7 Gall HE, Park J, Harman CJ, Jawitz JW, Rao PSC (2013) Landscape filtering of hydrologic and biogeochemical responses in managed catchments. Landscape Ecol
28:651–664.
8 Labonne J, Ravigne V, Parisi B, Gaucherel C (2008) Linking dendritic network structures to population demogenetics: The downside of connectivity. Oikos
117:1479–1490.
9 Carrara F, Altermatt F, Rodriguez-Iturbe I, Rinaldo A (2012) Dendritic connectivity controls biodiversity patterns in experimental metacommunities. Proc Natl Acad
Sci USA 109(15):5761–5766.
10 Heathwaite AL, Quinn PF, Hewitt CJM (2005) Modelling and managing critical source areas of diffuse pollution from agricultural land using flow connectivity
simulation. J Hydrol (Amst) 304:446–461.
11 Cvetkovic V, Carstens C, Selroos JO, Destouni G (2012) Water and solute transport along hydrological pathways. Water Resour Res 48:W06537.
12 Leibowitz SG, Vining KC (2003) Temporal connectivity in a prairie pothole complex. Wetlands 23:13–25.
13 Jensco KG, et al. (2009) Hydrologic connectivity between landscapes and streams: Transferring reach- and plot-scale understanding to the catchment scale. Water
Resour Res 45:W04428.
14 Botter G, Bertuzzo E, Rinaldo A (2010) Transport in the hydrologic response: Travel time distributions, soil moisture dynamics and the old water paradox. Water
Resour Res 46:W03514.
15 Harman C (2015) Time-variable transit time distributions and transport: Theory and application to storage-dependent transport of chloride in a watershed. Water
Resour Res 51:1–30.
16 Bracken LJ, Crocke J (2007) The concept of hydrological connectivity and its contribution to understanding runoff-dominated geomorphic systems. Hydrol
Processes 21:1749–1763.
17 Pringle CM (2003) What is hydrologic connectivity and why is it ecologically important? Hydrol Processes 17:2685–2689.
18 Fisher SG, Sponseller RA, Heffernan JB (2004) Horizons in stream biogeochemistry: Flowpaths to progress. Ecology 85:2369–2379.
19 Mehnert E, et al. (2007) Denitrification in the shallow ground water of a tile-drained, agricultural watershed. J Environ Qual 36(1):80–90.
20 Lang M, McCarty G, McDonough O, Oesterling R, Wilen W (2012) Enhanced detection of wetland-stream connectivity using LiDAR. Wetlands 32:461–473.
21 Galat DL, et al. (1998) Flooding to restore connectivity of regulated large-river wetlands: Natural and controlled flooding as complementary processes along the
lower Missouri River. Bioscience 48:721–733.
22 Elmore AJ, Kaushal SS (2008) Disappearing headwaters: Patterns of stream burial due to urbanization. Front Ecol Environ 6:308–312.
23 Leibowitz SG, Loehle C, Li B, Peterson EM (2000) Modeling landscape functions and effects: A network approach. Ecol Modell 132:77–94.
24 Jackson CR, Pringle CM (2010) Ecological benefits of reduced hydrologic connectivity in intensively developed landscapes. Bioscience 60:37–46.
25 Acreman M, Holden J (2013) How wetlands affect floods. Wetlands 33:773–786.
26 Whigham D, Jordan T (2003) Isolated wetlands and water quality. Wetlands 23:541–549.
27 Bridgham SD, Megonigal JP, Keller JK, Bliss NB, Trettin C (2006) The carbon balance of North American wetlands. Wetlands 26:889–916.
28 Capps KA, et al. (2014) Biogeochemical hotspots in forested landscapes: Quantifying the functional role of vernal pools in denitrification and organic matter
processing. Ecosystems (N Y) 17:1455–1468.
29 Tiner RW, Jr. (1984) Wetlands of the United States: Current Status and Recent Trends (US Department of the Interior, Washington, DC).
30 Subalusky AL, Fitzgerald LA, Smith LL (2009) Ontogenetic niche shifts in the American Alligator establish functional connectivity between aquatic systems. Biol
Conserv 42:1502–1514.
31 Semlitsch RD, Bodie JR (2003) Biological criteria for buffer zones around wetlands and riparian habitats for amphibians and reptiles. Conserv Biol 17:1219–1228.
32 Mitchell JC, Paton PWC, Raithel CJ (2008) The importance of vernal pools to reptiles, birds, and mammals. Science and Conservation of Vernal Pools in
Mortheastern North America, eds Calhoun AJK, de Maynadier PG (CRC, Boca Raton, FL), pp 169–193.
33 Brinson M (1993) Changes in the functioning of wetlands along environmental gradients. Wetlands 13:65–74.
34 Bullock A, Acreman M (2003) The role of wetlands in the hydrologic cycle. Hydrol Earth Syst Sci 7:358–389.

1984 | www.pnas.org/cgi/doi/10.1073/pnas.1512650113 Cohen et al.

35 Snodgrass JW, Bryan AL, Jr., Lide RF, Smith GM (2001) Factors affecting the occurrence and structure of fish assemblages in isolated wetlands of the upper coastal
plain, USA. Can J Fish Aquat Sci 53:443–454.
36 Gibbs JP (1993) Importance of small wetlands for the persistence of local populations of wetland-associated animals. Wetlands 13:25–31.
37 Scheffer M, et al. (2006) Small habitat size and isolation can promote species richness: Second-order effects on biodiversity in shallow lakes and ponds. Oikos
112:227–231.
38 Semlitsch RD, Bodie JR (1998) Are small, isolated wetlands expendable? Conserv Biol 12:1129–1133.
39 Pope SE, Fahrig L, Merriam HG (2000) Landscape complementation and metapopulation effects on leopard frog populations. Ecology 81:2498–2508.
40 Tiner RW, Jr. (2003) Geographically isolated wetlands of the United States. Wetlands 23:494–516.
41 Winter TC, LaBaugh JW (2003) Hydrologic considerations in defining isolated wetlands. Wetlands 23:532–540.
42 Zedler PH (2003) Vernal pools and the concept of “isolated wetlands”. Wetlands 23:597–607.
43 Leibowitz SG (2003) Isolated wetlands and their functions: An ecological perspective. Wetlands 23:517–531.
44 Mushet DM, et al. (2015) Geographically isolated wetlands: Rethinking a misnomer. Wetlands 10.1007/s13157-015-0631-9.
45 Adler RW (2015) US Environmental Protection Agency’s new Waters of the United States Rule: Connecting law and science. Freshw Sci 34(4):1595–1600.
46 Downing DM, Winer C, Wood LD (2003) Navigating through Clean Water Act jurisdiction: A legal review. Wetlands 23:475–493.
47 Leibowitz SG, Wigington PJ, Jr, Rains MC, Downing DM (2008) Non-navigable streams and adjacent wetlands: Addressing science needs following the Supreme
Court’s Rapanos decision. Front Ecol Environ 6:364–371.
48 US EPA and US ACE (US Environmental Protection Agency and US Army Corps of Engineers) (2015) Clean Water Rule: Definition of “Waters of the United States”.
80. Federal Register, pp 37054–37127.
49 US EPA (2015) Connectivity of Streams and Wetlands to Downstream Waters: A Review and Synthesis of the Scientific Evidence (US Environmental Protection
Agency, Washington, DC), EPA/600/R-14/475F.
50 Lang M, McCarty G, Oesterling R, Yeo I-Y (2012) Topographic metrics for improved mapping of forested wetlands. Wetlands 33:141–155.
51 Tiner RW, Jr (1997) NWI Maps: What They Tell Us. National Wetlands Newsletter 19(2):7–12.
52 Tiner RW, Jr (2003) Estimated extent of geographically isolated wetlands in selected areas of the United States. Wetlands 23:635–652.
53 Lane CR, D’Amico E, Autrey B (2012) Isolated wetlands of the Southeastern United States: Abundance and expected condition. Wetlands 32:753–767.
54 Holland MM, Whigham DF, Gopal B (1990) The characteristics of wetland ecotones. The Ecology and Management of Aquatic-Terrestrial Ecotones, Man and the
Biosphere Series, eds Naiman RJ, Décamps H (CRC, Boca Raton, FL), Vol 4, pp 171–198.
55 Van Meter KJ, Basu NB (2015) Signatures of human impact: Size distributions and spatial organization of wetlands in the Prairie Pothole landscape. Ecol Appl 25(2):
451–465.
56 Urban D, Keitt T (2001) Landscape connectivity: A graph-theoretic perspective. Ecology 82:1205–1218.
57 Uden DR, Hellman ML, Angeler DG, Allen CR (2014) The role of reserves and anthropogenic habitats for functional connectivity and resilience of ephemeral
wetlands. Ecol Appl 24:1569–1582.
58 LaBaugh JW, Winter TC, Rosenberry DO (1998) Hydrologic functions of prairie wetlands. Great Plains Res 8:17–37.
59 Newman MC, Schalles JF (1990) The water chemistry of Carolina Bays: A regional survey. Arch Hydrobiol 118:147–168.
60 Reiss KC (2004) Developing biological indicators for isolated forested wetlands in Florida. PhD Dissertation (University of Florida, Gainesville FL).
61 Lane CR, Brown MT (2007) Diatoms as indicators of isolated herbaceous wetland condition in Florida, USA. Ecol Indic 7:521–540.
62 Rains MC, Dahlgren RA, Williamson RJ, Fogg GE, Harter T (2008) Geological control of physical and chemical hydrology in vernal pools, Central Valley, California.
Wetlands 28:347–362.
63 Korfel CA, Mitsch WJ, Hetherington TE, Mack JJ (2010) Hydrology, physiochemistry, and amphibians in natural and created vernal pool wetlands. Restor Ecol
18:843–854.
64 Lane CR, D’Amico E (2010) Calculating the ecosystem service of water storage in isolated wetlands using LiDAR in north central Florida, USA. Wetlands
30:967–977.
65 Junk WJ, Bayley PB, Sparks RE (1989) The flood-pulse concept in river-floodplain systems. Proceedings of the International Large River Symposium, Canadian
Special Publication of Fish and Aquatic Sciences, ed Dodge DP (Fisheries and Oceans Comunications Directorate, Ottawa, ON, Canada), vol 106, pp 110–127.
66 Pyzoha JE, Callahan TJ, Sun G, Trettin CC, Miwa M (2007) A conceptual model for a forested Carolina bay depressional wetland on the Coastal Plain of South
Carolina, USA. Hydrol Processes 22:2689–2698.
67 McLaughlin DL, Cohen MJ (2013) Realizing ecosystem services: Wetland hydrologic function along a gradient of ecosystem condition. Ecol Appl 23(7):
1619–1631.
68 Min JH, Perkins DB, Jawitz JW (2010) Wetland-groundwater interactions in subtropical depressional wetlands. Wetlands 30:997–1006.
69 Golden HE, et al. (2015) Relative effects of geographically isolated wetlands on streamflow: A watershed-scale analysis. Ecohydrology, 10.1002/eco.1608.
70 McLaughlin DL, Kaplan DR, Cohen MJ (2014) A significant nexus: Isolated wetlands influence landscape hydrology. Water Resour Res 50:7153–7166.
71 Lang M, Bourgeau-Chavez L, Tiner R, Klemas V (2015) Advances in remotely sensed data and techniques for wetland mapping and monitoring. Remote Sensing of
Wetlands: Applications and Advances, eds Tiner R, Lang M, Klemas V (CRC, Boca Raton, FL), pp 74–112.
72 Rains MC, Fogg GE, Harter T, Dahlgran RA, Williamson RJ (2006) The role of perched aquifers in hydrological connectivity and biogeochemical processes in vernal
pool landscapes, Central Valley, California. Hydrol Processes 20:1157–1175.
73 Ferone JE, Devito KJ (2004) Shallow groundwater-surface water interactions in pond-peatland complexes along a Boreal Plains topographic gradient. J Hydrol
(Amst) 292:75–95.
74 McLaughlin DL, Cohen MJ (2014) Ecosystem specific yield for estimating evapotranspiration and groundwater exchange from diel surface water variation. Hydrol
Processes 28:1495–1506.
75 Sanford WE, Pope JP (2013) Quantifying groundwater’s role in delaying improvements to Chesapeake Bay water quality. Environ Sci Technol 47(23):
13330–13338.
76 Vanderhoof MK, Alexander LC, Todd JM (2015) Temporal and spatial patterns of wetland extent influence variability of surface water connectivity in the Prairie
Pothole Region, United States. Landscape Ecol, 10.1007/s10980-015-0290-5.
77 McGuire KJ, McDonnell JJ (2010) Hydrologic connectivity of hillslopes and streams: Characteristic time scales and nonlinearities. Water Resour Res 46:W10543.
78 Dunne T, Black RD (1970) Partial area contributions to storm runoff in a small New England watershed. Water Resour Res 6:1296–1311.
79 Devito KJ, Hill AR, Roulet N (1996) Groundwater-surface water interactions in headwater forested wetlands of the Canadian Shield. J Hydrol (Amst) 181:127–147.
80 Ali GA, Roy AG (2009) Revisiting hydrologic sampling strategies for an accurate assessment of hydrologic connectivity in humid temperate systems. Geogr
Compass 3:350–374.
81 Creed IF, Sass GZ (2011) Digital terrain analysis approaches for tracking hydrological and biogeochemical pathways and processes in forested landscapes. Forest
Hydrology and Biogeochemistry 216:69–100.
82 Winter TC, Harvey JW, Lehn Franke O, Alley WM (1998) Ground Water and Surface Water: A Single Resource (US Geological Survey, Denver).
83 Craft CB, Casey WP (2000) Sediment and nutrient accumulation in floodplain and depressional freshwater wetlands of Georgia, USA. Wetlands 20:323–332.
84 Reddy KR, Kadlec RH, Flaig E, Gale PM (1999) Phosphorus retention in streams and wetlands: A review. Crit Rev Environ Sci Technol 29:83–146.
85 Whigham DF, Chitterling C, Palmer B (1988) Impacts of freshwater wetlands on water quality: A landscape perspective. Environ Manage 12:663–671.
86 Gambrell RP (1994) Trace and toxic metals in wetlands: A review. J Environ Qual 23:883–891.
87 Johnston CA (1991) Sediment and nutrient retention by freshwater wetlands: Effects on surface water quality. CRC Crit Rev Environ Control 21:491–565.

Cohen et al. PNAS | February 23, 2016 | vol. 113 | no. 8 | 1985
88 Zedler JB (2003) Wetlands at your service: Reducing impacts of agriculture at the watershed scale. Front Ecol Environ 1:65–72.
89 Cohen MJ, Brown MT (2007) A model examining hierarchical wetland networks for watershed stormwater management. Ecol Modell 201:179–193.
90 Cheesman AW, Dunne EJ, Turner BL, Reddy KR (2010) Soil phosphorus forms in hydrologically isolated wetlands and surrounding pasture uplands. J Environ Qual
39(4):1517–1525.
91 Peterson BJ, et al. (2001) Control of nitrogen export from watersheds by headwater streams. Science 292(5514):86–90.
92 Hedin LO, et al. (1998) Thermodynamic constraints on nitrogen transformations and other biogeochemical processes at soil-stream interfaces. Ecology
79:684–703.
93 Alexander RB, Smith RA, Schwarz GE (2000) Effect of stream channel size on the delivery of nitrogen to the Gulf of Mexico. Nature 403(6771):758–761.
94 Duarte CM, Kalff J, Peters RH (1986) Patterns in biomass and cover of aquatic macrophytes in lakes. Can J Fish Aquat Sci 43(10):1900–1908.
95 Downing JA, et al. (2008) Sediment organic carbon burial in agriculturally eutrophic impoundments over the last century. Global Biogeochem Cycles 22(1):
GB1018.
96 Dodson SI, Arnott SE, Cotting KL (2000) The relationship in lake communities between primary productivity and species richness. Ecology 81:2662–2679.
97 Youngs WD, Heimbuch DG (1982) Another consideration of the Morphoedaphic Index. Trans Am Fish Soc 111:151–153.
98 Downing JA (2010) Emerging global role of small lakes and ponds: Little things mean a lot. Limnética 9(1):9–24.
99 Marton JM, et al. (2015) Geographically isolated wetlands are important biogeochemical reactors on the landscape. Bioscience 65(4):408–418.
100 Jordan SJ, Stoffer J, Nestlerode JA (2011) Wetlands as sinks for reactive nitrogen at continental and global scales: A meta-analysis. Ecosystems (N Y) 14:144–155.
101 Lane CR, et al. (2015) Denitrification potential in geographically isolated wetlands of North Carolina and Florida, USA. Wetlands 35:459–471.
102 Whitmire SL, Hamilton SK (2005) Rapid removal of nitrate and sulfate in freshwater wetland sediments. J Environ Qual 34(6):2062–2071.
103 Gibbs JP (2000) Wetlands loss and biodiversity conservation. Conserv Biol 14:314–317.
104 Ghermandi A, Van Den Bergh JCJM, Brander LM, de Groot HLF, Nunes PALD (2010) Values of natural and human‐made wetlands: A meta-analysis. Water
Resour Res 46:W12516.
105 Holland JF, et al. (2004) Effects of wetland depth and flow rate on residence time distribution characteristics. Ecol Eng 23:189–203.
106 Werner TM, Kadlec RH (2000) Wetland residence time distribution modeling. Ecol Eng 15:77–90.
107 Powers SM, Johnson RA, Stanley EH (2012) Nutrient retention and the problem of hydrologic disconnection in streams and wetlands. Ecosystems (N Y) 15:435–449.
108 Mays PA, Edwards GS (2001) Comparison of heavy metal accumulation in natural wetlands and constructed wetlands receiving acid mine drainage. Ecol Eng
16:487–500.
109 Stehle S, et al. (2011) Pesticide risk mitigation by vegetated treatment systems: A meta-analysis. J Environ Qual 40(4):1068–1080.
110 Haig SM, Mehlman DW, Oring LW (1998) Avian movements and wetland connectivity in landscape conservation. Conserv Biol 12:749–758.
111 Gibbons JW (2003) Terrestrial habitat: A vital component for herpetofauna of isolated wetlands. Wetlands 23:630–635.
112 Euliss NH, et al. (2004) The wetland continuum: A conceptual framework for interpreting biological studies. Wetlands 24:448–458.
113 Boughton EH, Quintana-Ascencio PF, Bohlen PJ, Jenkins DG, Pickert R (2010) Land-use and isolation interact to affect wetland plant assemblages. Ecography
33:461–470.
114 McIntyre NE, et al. (2014) Climate forcing of wetland landscape connectivity in the Great Plains. Front Ecol Environ 12:59–64.
115 Van der Valk AG, Pederson RL (2003) The SWANCC decision and its implications for prairie potholes. Wetlands 23:590–596.
116 Sheldon F, Boulton AJ, Puckridge JT (2002) Conservation value of variable connectivity: Aquatic invertebrate assemblages of channel and floodplain habitats of
a central Australian arid-zone river, Cooper Creek. Biol Conserv 103:13–31.
117 Dias MS, et al. (2014) Global imprint of historical connectivity on freshwater fish biodiversity. Ecol Lett 17(9):1130–1140.
118 Brose U (2001) Relative importance of isolation, area and habitat heterogeneity for vascular plant species richness of temporary wetlands in east-German
farmland. Ecography 24:722–730.
119 Snodgrass JW, Komoroski MJ, Bryan AL, Jr, Burger J (2000) Relationships among isolated wetland size, hydroperiod, and amphibian species richness:
Implications for wetland regulations. Conserv Biol 14:414–419.
120 Roe JH, Georges A (2008) Maintenance of variable responses for coping with wetland drying in freshwater turtles. Ecology 89(2):485–494.
121 Beaudry F, de Maynadier PG, Hunter ML, Jr (2009) Seasonally dynamic habitat use by Spotted (Clemmys guttata) and Blanding’s Turtles (Emydoidea balndingii)
in Maine. J Herpetol 43:636–645.
122 Holland RF, Jain S (1988) Vernal pools. Terrestrial Vegetation of California, eds Barbour MG, Major J (California Native Plant Society, Sacramento, CA), 2nd Ed, pp
515–533.
123 Bedford BL, Godwin KS (2003) Fens of the United States: Distribution, characteristics and scientific connection versus legal isolation. Wetlands 23:609–629.
124 Bohonak AJ, Jenkins DG (2003) Ecological and evolutionary significance of dispersal by freshwater invertebrates. Ecol Lett 6:783–796.
125 Shulman RS, Chase JM (2007) Increasing isolation reduces predator: Prey species richness ratios in aquatic food webs. Oikos 116:1581–1587.
126 Colburn EA (2004) Vernal Pools: Natural History and Conservation (McDonald and Woodward, Blacksburg, VA).
127 Calhoun AJK, de Maynadier PG (2008) Science and Conservation of Vernal Pools in Northeastern North America (CRC, Boca Raton, FL), 392 pp.
128 Ryan ME, Palen WJ, Adams MJ, Rochefort RM (2014) Amphibians in the climate vise: Loss and restoration of resilience of montane wetland ecosystems in the
western US. Front Ecol Environ 12:232–240.
129 Joyal LA, McCollough M, Hunter ML, Jr (2001) Landscape ecology approaches to wetland species conservation: A case study of two turtle species in southern
Maine. Conserv Biol 15:1755–1762.
130 Silveira JG (1998) Avian uses of vernal pools and implications for conservation practice. Ecology, Conservation, and Management of Vernal Pool Ecosystems: Proceedings
from a 1996 Conference, eds Witham CW, Bauder ET, Belk D, Ferren WR, Jr, Ornduff R (California Native Plant Society, Sacramento, CA), pp 92–106.
131 Naugle DE, Johnson RR, Estey ME, Higgins KF (2001) A landscape approach to conserving wetland bird habitat in the prairie pothole region of eastern South
Dakota. Wetlands 21:1–17.
132 Roe JH, Kingsbury BA, Herbert NR (2004) Comparative water snake ecology: Conservation of mobile animals that use temporally dynamic resources. Biol
Conserv 118:79–89.
133 Bodie JR, Semlitsch RD (2000) Spatial and temporal use of floodplain habitats by lentic and lotic species of aquatic turtles. Oecologia 122:138–146.
134 Pringle CM (2003) The need for a more predictive understanding of hydrologic connectivity. Aquat Conserv 13:467–471.
135 Johnston CA (2013) Wetland losses due to row crop expansion in the Dakota Prairie Pothole region. Wetlands 33:175–182.
136 Poff NL, et al. (1997) The natural flow regime. Bioscience 47:769–784.
137 Bunn SE, Arthington AH (2002) Basic principles and ecological consequences of altered flow regimes for aquatic biodiversity. Environ Manage 30(4):492–507.
138 Fagan WF (2002) Connectivity, fragmentation, and extinction risk in dendritic metapopulations. Ecology 83:3243–3249.
139 Campbell Grant EH, Nichols JD, Lowe WH, Fagan WF (2010) Use of multiple dispersal pathways facilitates amphibian persistence in stream networks. Proc Natl
Acad Sci USA 107(15):6936–6940.
140 Wörman A, Kronnäs V (2005) Effect of pond shape and vegetation heterogeneity on flow and treatment performance of constructed wetlands. J Hydrol (Amst)
301:123–138.
141 Babbitt KJ, Tanner GW (2000) Use of temporary wetlands by anurans in a hydrologically modified landscape. Wetlands 20:313–322.
142 Rains MC, Landry S, Rains KC, Seidel V, Crisman TL (2013) Using net wetland loss, current wetland condition, and planned future watershed condition for wetland
conservation planning and prioritization, Tampa Bay Watershed, Florida. Wetlands 33:949–963.

1986 | www.pnas.org/cgi/doi/10.1073/pnas.1512650113 Cohen et al.