Realizing ecosystem services: wetland ... - Wiley Online Library

Ecological Applications, 23(7), 2013, pp. 1619–1631 Ó 2013 by the Ecological Society of America

Realizing ecosystem services: wetland hydrologic function along a gradient of ecosystem condition DANIEL L. MCLAUGHLIN1

AND

MATTHEW J. COHEN

School of Forest Resources and Conservation, University of Florida, 319 Newins-Ziegler Hall, P.O. Box 110410, Gainesville, Florida 32611-0410 USA

Abstract. Wetlands provide numerous ecosystem services, from habitat provision to pollutant removal, floodwater storage, and microclimate regulation. Delivery of particular services relies on specific ecological functions, and thus to varying degree on wetland ecological condition, commonly quantified as departure from minimally impacted reference sites. Condition assessments are widely adopted as regulatory indicators of ecosystem function, and for some services (e.g., habitat) links between condition and function are often direct. For others, however, links are more tenuous, and using condition alone to enumerate ecosystem value (e.g., for compensatory mitigation) may underestimate important services. Hydrologic function affects many services cited in support of wetland protection both directly (floodwater retention, microclimate regulation) and indirectly (biogeochemical cycling, pollutant removal). We investigated links between condition and hydrologic function to test the hypothesis, embedded in regulatory assessment of wetland value, that condition predicts function. Condition was assessed using rapid and intensive approaches, including Florida’s official wetland assessment tool, in 11 isolated forested wetlands in north Florida (USA) spanning a land use intensity gradient. Hydrologic function was assessed using hydrologic regime (mean, variance, and rates of change of water depth), and measurements of groundwater exchange and evapotranspiration (ET). Despite a wide range in condition, no systematic variation in hydrologic regime was observed; indeed reference sites spanned the full range of variation. In contrast, ET was affected by land use, with higher rates in intensive (agriculture and urban) landscapes in response to higher leaf area. ET determines latent heat exchange, which regulates microclimate, a valuable service in urban heat islands. Higher ET also indicates higher productivity and thus carbon cycling. Groundwater exchange regularly reversed flow direction at all sites in response to rainfall. This buffering effect on regional aquifer levels, an underappreciated service of isolated wetlands, was provided regardless of condition. Intensive landscapes may benefit most from the hydrologic services that wetlands provide because that is where certain services (floodwater storage, microclimate regulation) are realized. While the portfolio of wetland services clearly changes with disturbance, our results support a revised approach to wetland valuation that recognizes the services that accrue from sustained or enhanced functions in these ‘‘working wetlands.’’ Key words: disturbance; diurnal water-level fluctuation; ecosystem function; evapotranspiration; groundwater exchange; isolated wetlands; land use; urban wetlands; wetland condition assessments; wetland hydrology; wetland mitigation; wetland value.

INTRODUCTION Ecosystem services derive from ecological functions, which are frequently impaired by anthropogenic disturbance. One common effect of human disturbance is to impair the ability of an ecosystem to provide habitat that sustains local and regional biodiversity. Because of the central prominence of the loss of biodiversity functions, ecosystem condition is often evaluated by comparing a site’s habitat integrity vis-a`-vis minimally impacted reference settings (Karr and Chu 1997, Que´tier and Lavorel 2011). While other functions such as carbon Manuscript received 29 August 2012; revised 7 March 2013; accepted 18 March 2013. Corresponding Editor: C. B. Craft. 1 E-mail: mclaugd@ufl.edu

sequestration, adjacency effects on pollination (Ricketts et al. 2008), hydrologic buffering (i.e., infiltration, flood water storage and slow release), microclimate regulation, and contaminant removal are acknowledged as important, they are more challenging to enumerate at appropriate temporal and spatial scales. This raises the question of whether specific ecosystem functions are predicted by rapid assessments of overall ecosystem condition (quantified as departure from reference conditions) (Stander and Ehrenfeld 2009a). If not, a potentially critical disconnect may exist between the provision of ecosystem services (derived from functions) and enumeration of ecosystem value (based on condition). Wetlands are a model setting in which to examine the general question of the association between ecosystem

1619

1620

DANIEL L. MCLAUGHLIN AND MATTHEW J. COHEN

condition and function. Wetlands provide a well-known array of services (MEA 2005), and Federal Clean Water Act mandates (33 U.S. Code section 1251 et seq.; 1972 and as amended) that target no net regional loss have resulted in the need for assessment tools that support Federal and state protection programs and compensatory mitigation efforts. As in other systems, wetland functions are difficult to quantify, especially at their characteristic temporal scale (Hruby 1999). In an effort to provide rapid and repeatable measures of ecosystem function that can be used for compensatory mitigation and to prioritize conservation and restoration actions, a wide variety of assessments of ecological condition have been developed (Fennessy et al. 2004). Condition assessments have been customized to evaluate a variety of wetland attributes (e.g., habitat quality, community structure, hydrologic alteration, water quality, surrounding land use; Fennessey et al. 2007). They are explicitly developed to provide rapid reconnaissance, consistent with their regulatory application to issues like compensatory mitigation, where loss of wetland functions in one area is restored elsewhere to achieve a goal of ‘‘no net loss’’ at the regional scale. For example, the Uniform Mitigation Assessment Method (UMAM), Florida’s legally mandated tool for determining both mitigation requirements (for wetland impacts) and success (for wetland creation or restoration), assesses condition based on surrounding land use and indicators of community structure, hydrology, and water quality (Bardi et al. 2004). Condition assessment methods, such as UMAM, are designed to measure departure from reference conditions (i.e., minimally impacted sites), with lower scores indicating greater departure (Fennessy et al. 2007). Not surprisingly, condition assessments yield low scores for wetlands within high intensity land uses (e.g., urban and agriculture) (Lopez and Fennessy 2002, Reiss 2006, Mack 2007), resulting in reduced compensatory mitigation requirements for their loss. However, many functions that wetlands provide (e.g., microclimate regulation, carbon sequestration) may persist or even be enhanced in these settings despite impaired condition. This highlights an important implicit assumption of using rapid condition assessment methods for enumerating compensatory mitigation, namely that assessed condition is correlated with function, and thus ecosystem service provisioning. Evaluating this implicit assumption is our central goal. We note that not all assessment methods draw inference about function from measured condition. Some, like the federally preferred hydrogeomorphic (HGM) functional assessment approach (Hauer and Smith 1998), target direct assessment of function (Brinson 1993). Like other assessment methods, however, HGM relies on reference sites for calibrating the various structural indicators of function. Reference conditions are presumed to provide an optimal suite of functions characteristic of each wetland class. Crucially,

Ecological Applications Vol. 23, No. 7

this does not necessarily mean that reference wetlands provide all functions at maximum rates (Brinson and Rheinhardt 1996). Indeed, some functions may be elevated in degraded systems (e.g., denitrification; Stander and Ehrenfeld 2009a), often at the expense of others, particularly biological condition. Departure from reference conditions is treated as degradation even if some functions are enhanced, meaning that a priori determination of disturbance defines what represents optimal functionality for a particular wetland class (Stander and Ehrenfeld 2009a). Wetlands provide a portfolio of services, and although certain functions (e.g., water quality improvement, flood water storage) may be enhanced while others are impaired (e.g., habitat), services like water quality improvement are only realized in contexts where water quality and quantity have been altered, and thus also where habitat services are degraded (Reiss 2006). This implies that there is a trade-off between wetland services, and thus a critical distinction between measuring disturbance (i.e., departure from reference conditions) and enumerating wetland value (Jordan et al. 2007). Hydrology is a core determinant of wetland type, habitat, and processes (Mitsch and Gosselink 2007), and thus the core control on the differential realization of wetland functions across disturbance gradients. Land use influences to hydrologic regime (e.g., changes in depth, duration, and/or spatial extent of flooding) can reduce habitat quality via controls on ecosystem structure and species composition at multiple trophic levels (Richter et al. 1996). However, wetland hydrologic functions extend beyond habitat effects to regulation of downstream flooding, controls on important biogeochemical cycles responsible for improving water quality, local groundwater recharge, and latent heat exchange via evapotranspiration (ET). These hydrologic functions are not, a priori, reduced when hydrologic regime is disturbed and/or biological condition declines. Indeed, nutrient loading, reduced fire regime, and introduction of invasive species can degrade biological condition but increase ecosystem productivity (Keddy 2010) and with it regulation of ET. Moreover, alteration of catchment physical conditions (e.g., soil compaction, impervious surfaces, enhanced drainage networks) can influence wetland and watershed hydrology (Winter 1988), and the role of wetlands in mitigating these disturbances is most important in landscapes where those catchment alterations have occurred. Despite recent work highlighting the importance of direct measurement of site-specific hydrology to assess function, rather than solely using integrative indicators (e.g., redoximorphic features, seasonal high-water lines, buttressed tree trunks) (Cole 2006, Stander and Ehrenfeld 2009a, b, Gebo and Brooks 2012), information about wetland flooding dynamics and exchanges (e.g., ET and groundwater exchange) remains scarce. Exchanges of water between wetlands, adjacent aquifers, and the atmosphere are among the most

October 2013

WETLAND CONDITION VS. FUNCTION

1621

TABLE 1. The wetland site locations (northern Florida, USA) with land use, area, soil series, and condition scores using the Uniform Mitigation Assessment Method (UMAM) and Florida Wetland Condition Index (FWCI). The scores for the Water Environment (WE) category within UMAM are also listed.

Site

Land use

Location (lat., long.)

Area (ha)

R1

Protected park

30.47229, 81.49923

1.06

R2 R3 R4 S

Conserved forest Conserved forest Protected park Silviculture

30.17871, 30.16982, 30.30300, 29.47003,

81.93942 81.93639 82.41403 83.09615

1.05 1.24 0.33 0.67

A1

Agriculture row crops Agriculture row crops Urban Urban Urban Agriculture pasture

29.80023, 82.41417

A2 U1 U2 U3 A3

Soil series

UMAMà

UMAMWEà

FWCIà

0.93

9.0

49.0

0.83 0.83 0.83 0.80

8.0 8.0 8.0 9.0

43.0 44.0 46.2 37.4

0.48

Evergreen-Wesconnett Complex Plummer Fine Sand Plummer Fine Sand Mascotte Fine Sand Clara, Oldtown, and Meadowbrook Pomona Sand

0.77

9.0

17.4

29.79495, 82.41924

0.36

Pomona Sand

0.63

6.0

36.5

82.64868 81.62379 81.76350 82.26923

0.65 0.52 0.81 0.20

Surrency Fine Sand Adamsville Fine Sand Surrency Loamy Fine Sand Pomona Sand

0.60 0.57 0.43 0.30

7.0 7.0 7.0 4.0

30.1 37.9 19.9 9.7

30.20991, 30.11208, 30.20305, 29.74165,

Source: Natural Resources Conservation Service Soil Survey (NRCS 2012). à Source: Deimeke (2009).

important wetland functions. Wetlands play an important role in regulating aquifers through groundwater– surface water exchange (Winter 1999, Sophocleous 2002), and also in regulating local and regional temperatures via their effects on ET and latent heat exchange. Urban vegetation generally alleviates heat island effects (Taha 1997, Peters et al. 2011), and persistent conditions of water availability (the oasis effect; Drexler et al. 2004), dense vegetation structure (the ‘‘clothesline’’ effect; Drexler et al. 2004), and lateral advection of sensible heat (Spronken-Smith et al. 2000) make wetlands in urban landscapes prime locations for realizing this micro-climate regulation service. Moreover, because transpiration dominates the ET flux in forested wetlands (Law et al. 2002), ET and primary productivity are tightly coupled, suggesting that wetland functions related to primary production (e.g., carbon sequestration, biogeochemical cycling) may be inferred from measured ET rates. Despite the strong rationale for direct measurements of hydrologic fluxes to assess wetland function, spatial and temporal variation makes such measurements difficult to obtain (Winter 1999, Drexler et al. 2004). The White (1932) method, a tool developed and now widely used to estimate phreatophyte transpiration from diurnal variation in groundwater levels (Loheide et al. 2005), provides a low-cost, spatially integrative, empirical approach that yields estimates of both ET and groundwater fluxes in surface-water systems. Applications in surface-water systems, however, have been limited due to operational constraints, the recent alleviation of which (McLaughlin and Cohen 2011, McLaughlin and Cohen 2013) has broadened applicability to aquatic ecosystems, providing an opportunity to assess wetland hydrologic services at spatial and temporal scales commensurate with their delivery.

In this work we explored the potential disconnect between hydrologic function and assessed ecosystem condition by measuring both physical and biotically mediated hydrologic responses (including high-resolution White-method estimates of ET and groundwater fluxes) in wetland systems spanning a human-disturbance gradient. Specifically, we investigated how well assessments of condition (via two condition-assessment methods used in Florida) predicted hydrologic services (both with respect to the hydrologic regime and to important hydrologic fluxes) with the hypotheses that (1) assessed ecosystem condition covaries with alteration of hydrologic regime, but that (2) certain hydrologic functions (ET and groundwater exchange) continue to be realized and are even amplified as measured condition declines. METHODS Site descriptions Eleven cypress domes (small circular depressions dominated by Taxodium distichum var. nutans; Ewel 1990) were selected across north Florida (USA); these sites were selected from a larger set of domes previously used to calibrate the Florida Wetland Condition Index (FWCI; Reiss 2006). All 11 sites were hydrologically isolated with no observable surface water connections, and ranged in size from to 0.2 to 1.24 ha (Table 1). While the dominant soil series varied among sites (Table 1), all were characterized by ;1 m sandy soil profiles overlain with an organic layer and confined below by a thick clay layer. Sites were selected from an a priori human disturbance gradient; land uses ranged from minimal human disturbance (i.e., reference) to silviculture, agriculture (improved pasture and row crops), to urban (Table 1).

1622


Ecosystem condition Condition assessment scores (Table 1) were previously determined (Deimeke 2009) using Florida’s Uniform Mitigation Assessment Method (UMAM) and FWCI. The resulting scores from both methods principally reflect the systems’ surrounding land uses, with high values reported for reference sites, and decreasing values associated with increasing disturbance. FWCI is an intensive assessment method that provides a multi-metric index of biological integrity relative to reference conditions (Reiss 2006). When applied in forested depressional wetlands, FWCI uses detailed monitoring of the presence and density of herbaceous macrophytes and trees and six floristic metrics related to relative abundance, including tolerant and sensitive indicator species, the floristic quality assessment index (FQAI), exotic species, native perennial species, and wetland status species (Reiss 2006). Each metric is scored from 0 to 10, with 10 representing reference biological condition; scores are summed, resulting in an FWCI score ranging from 0 to 60. Details of the method and its implementation can be found in Reiss (2006). UMAM is a rapid assessment method mandated by the State of Florida for determining mitigation requirements (i.e., area multipliers for compensatory mitigation of wetland impacts) and success (Florida Statute 373.414, subsection 18 [2000]; Florida Administrative Code Chapter 62-345). UMAM has three scoring categories: Location and Landscape Support, Water Environment, and Community Structure (Bardi et al. 2004). Similar to FWCI’s metrics, each category is scored on a scale from 0 to 10, with 10 indicating minimally impaired. Scores for the three categories are summed and divided by 30 to yield a final UMAM score ranging between 0 and 1.0. Although these categories, and their associated indicators, are primarily intended to assess habitat suitability (Bardi et al. 2004), the Water Environment category includes indicators for hydrologic alteration. As such, we expected, at minimum, a correlation between this category’s score and metrics of hydrologic regime. Hydrologic regime To evaluate how well assessments of condition predicted hydrologic regime, we compared mean and variance of daily stage and flooding duration across sites of varying condition. We also calculated the frequency and magnitude of daily positive and negative stage changes as a metric of site hydrologic regime (Richter et al. 1996). Wetland stage (relative to ground surface) was continuously measured for two years (2009–2011) with total pressure transducers (Solinst Gold Leveloggers, accuracy ¼ 0.3 cm, resolution ¼ 0.005 cm; Solinst Canada, Georgetown, Ontario, Canada) deployed in monitoring wells in the deepest area of each wetland. Wells were installed to a depth of 90 cm below ground; well casings were screened above and below ground


surface, and constructed of 5.08-cm-diameter, schedule 40 PVC. Total pressure (m H2O) was measured at 15min intervals and corrected for barometric pressure variation collected at equal intervals with barometric pressure transducers (Solinst Barologgers, accuracy ¼ 0.1 cm, resolution ¼ 0.003 cm). Barometric pressure transducers were installed in a dry well below the ground surface, but open to atmospheric pressure variation, to buffer thermal conditions and avoid known temperature sensitivity (McLaughlin and Cohen 2011). Barometrically corrected stage data were verified with direct stage measurements made during site visits. ET and groundwater exchange We compared measured ET and groundwater fluxes across sites to test the hypothesis that certain hydrologic functions remain despite decline in ecosystem condition. The White (1932) method was used to calculate daily ET and net groundwater flow rates (cm/day) from highresolution stage data for days with no rain. The method assumes ET fluxes are negligible at night, allowing net groundwater flows to be inferred from nighttime stage changes; groundwater flow rates are assumed diurnally constant. ET and exfiltration (i.e., groundwater inflow) were calculated with ET ¼ Sy ð24h þ sÞ

ð1Þ

Exfiltration ¼ Sy ð24hÞ

ð2Þ

where Sy is specific yield (dimensionless), s (cm) is the 24-hour stage change (positive values indicate net stage decline), and h (cm/h) is the net groundwater inflow (uncorrected for Sy). The linear slopes of nighttime stage vs. time (between 0:00 and 5:00 hours) for the night before and after each day were averaged to determine h. Sy is defined as the water volume released or taken into storage by a matrix divided by the matrix volume (Healy and Cook 2002); on a unit area basis, Sy represents the input (rain) or output (ET) depth divided by the induced change in water level. Sy applies to aquifer and soil media (soil Sy 1) or surface water systems (water Sy ; 1). High Sy values under flooded conditions, which mute the ET signal vis-a`-vis what is observed in groundwater applications, necessitate the high resolution pressure transducers used here (McLaughlin and Cohen 2011). Sy can also vary with wetland stage, necessitating sitespecific functions that describe stage-dependent variation in Sy (ecosystem-specific yield, ESY; McLaughlin and Cohen 2013). We empirically determined the relationship between stage and Sy (ESY) for each site from the observed ratio of rainfall inputs to induced stage rise (rain : rise) as a function of stage (McLaughlin and Cohen 2013). Onset RG2 data-logging tippingbucket gauges (Onset Computer Corporation, Bourne, Massachusetts, USA) were installed at each site in canopy-free areas to collect continuous rain data from which we extracted accumulated rainfall in 15-minute increments since storm initiation. Rain data were

October 2013


corrected for interception storage since water-level rise occurs due to gross rainfall depth (i.e., measured in canopy-free areas) minus storage capacity in the canopy and sub-canopy. The x-intercept of rise vs. rain plots yielded interception storage (Heimburg 1976), which was subtracted from rain accumulations; we estimated interception storage separately for growing and nongrowing seasons. To isolate the effects of Sy from those of catchment subsidies, we calculated rain : rise values (i.e., interception-corrected rainfall depths divided by measured stage increases) only for storm events that met the following criteria: (1) .3 mm cumulative depth; (2) three hours or less in duration; and (3) not within 24 hours of a preceding storm event. Of the multiple functional forms (e.g., Monod equation, saturating exponential, and quadratic) evaluated, quadratic equations provided the best fit between Sy and stage for all sites; the resulting site-specific ESY functions were used to set Sy in Eqs. 1 and 2. Methodological details for construction of ESY functions are in McLaughlin and Cohen (2013). Calculated daily ET rates were indexed to estimated daily PET (potential evapotranspiration) to control for climatic variability across sites and time. Climate data from weather stations (Real-time Observation Monitor and Analysis Network; available online)2 within ;8 km of each site were used to estimate PET using the Hargreaves method (Hargreaves and Samani 1985). Mean indexed ET rates (ET/PET) during flooded and shallow water table (stage . 30 cm) conditions were compared between sites. Rates determined for drier conditions diverge from ecosystem ET and only indicate the proportion of ET met by groundwater uptake (Shah et al., 2007), introducing error when comparing indexed rates between sites; 30 cm was a cautious threshold to avoid such errors. Ecosystem ET can be strongly regulated by vegetation structure, which is often affected by land use and disturbance. To test the hypothesis that vegetation structure controls ET, we compared site mean ET/PET with measured mean leaf-area index (LAI) for each site. A Sunfleck Ceptometer (Decagon Devices, Pullman, Washington, USA) was used to measure projected LAI based on radiation transmittance (Chen et al. 1997) visa`-vis open-canopy conditions. Discrete LAI measurements were made during early summer and at each node of a 5 3 2 m grid spanning the entire wetland area at 75 cm above ground surface, capturing leaf structure of canopy and subcanopy strata. The instrument automatically averaged discrete LAI measurements and yielded mean LAI for each site. Groundwater exchange from the White (1932) method was available only for days without rain since the method assumes constant groundwater flow and does not account for rain inputs. This constraint limited the 2

http://raws.wrh.noaa.gov/roman

1623

continuity of the data set and systematically underestimated the frequency of groundwater inflow events, which are more likely to occur during and following rain events. To correct for this bias, groundwater flows on days with recorded rainfall were estimated from the residuals of the daily water balance using measured 24hour stage change (corrected for Sy), estimated ET rates, and measured precipitation. However, since ET also cannot be measured on rainy days using the White method, this flux was estimated as the product of PET on each rainy day and the site-specific mean ET/PET measured on days without rain. Daily exfiltration (inflow) on rainy days was then calculated by subtracting rain depths and adding estimated ET to 24-hour stage change (corrected for Sy). We note that exfiltration rates calculated in this way include any overland flows resulting from storm events. These exfiltration data were merged with White method-estimated groundwater flows on non-rainy days, resulting in a complete data set of catchment–wetland exchange. These data were used to evaluate whether a site was a net sink or source of water within the landscape, and the frequency, direction, and magnitude of the implied exchanges. To test our hypotheses, we evaluated correlations between condition scores (FWCI, overall UMAM scores, and also the UMAM Water Environment submetric) and metrics of hydrologic regime (mean and variance in stage, flooding duration, and daily stage change), ET rates, and groundwater fluxes. When comparing regime metrics and fluxes across sites, UMAM scores were used as the abscissa since UMAM is the legally mandated tool in Florida for wetland mitigation. Inferences of the effects of land use intensity were facilitated along this condition gradient because of the apparent influence of land use on assessed condition (Table 1). However, correlation analysis was performed only for assessed condition (a continuous variable) and not for land use category, comparisons of which would require greater replication than was possible for this study. RESULTS Ecosystem condition The 11 wetland sites (north Florida, USA) spanned nearly the entire range of possible values for both wetland condition assessment methods (Table 1), with scores between 9.7 to 49 for the Florida Wetland Condition Index (FWCI), which spans from 0 to 60, and from 0.3 to 0.93 for the Uniform Mitigation Assessment Method (UMAM), which has a maximum value of 1 and where values lower than 0.3 are reserved for sites that have been effectively lost (e.g., paved, filled, ploughed). Moreover, there was strong concordance between the FWCI and UMAM condition scores (linear R 2 ¼ 0.63, P , 0.005), suggesting that UMAM is a reasonable proxy for the more intensive condition assessment provided by FWCI. Condition scores from both assessment methods declined with increasing land

1624



stage variation in urban (U) sites was generally low, even compared with reference (R) sites (Fig. 1b). Similarly, there was no systematic variation in flooding duration (hydroperiod) with measured condition (Fig. 1c, Table 2), though we note that the period of stage record (2009–2011) reported was anomalously dry (cumulative rainfall ;30 cm below normal). Mean positive and negative daily stage changes, a more robust measure of water-level ‘‘flashiness’’ than stage variation alone, similarly revealed no clear trend in wetland storm responsiveness (mean positive stage change) or drying rate (mean negative stage change) with assessed condition (Fig. 2, Table 2). These results do, however, illustrate a fundamental feature of isolated wetland hydrology, with large differences between the magnitude and frequency of stage increases vs. decreases. In all cases, mean positive increases were higher in magnitude (by a factor ranging from 1.74 to 2.58) but occurred less frequently (,30% of the study period). This asymmetry derives from discrete rain events that induce runoff and subsurface flows and create large, but low frequency, stage increases compared to smaller magnitude, but extended drying via evapotranspiration (ET) and infiltration; this general behavior was consistent across all sites. ET and groundwater exchange

FIG. 1. No systematic variation for the 11 wetland sites in (a) wetland stage (mean 6 SD), (b) variance of wetland stage, or (c) percentage of time flooded is evident across a condition gradient (i.e., decreasing Uniform Mitigation Assessment Method [UMAM] scores). Site type codes: R, reference; S, silviculture; A, agriculture; U, urban. For additional wetland site information see Table 1.

use intensity, but UMAM scores more clearly grouped sites by land use category. Site A3 was the only site not to be grouped by land use (agriculture) and had the lowest measured condition due to active grazing in the wetland (Table 1). Hydrologic regime Despite substantial differences across sites in mean water depth (mean stage ranging from 1.1 to 0.1 m) and variation (stage standard deviation ranging from 0.1 to 0.5 m), there was no clear trend with disturbance (Fig. 1a, b) or significant correlation with UMAM or FWCI condition scores (Table 2). Surprisingly, the UMAM Water Environment sub-metric, which includes indicators of water depth and flows, was not significantly correlated with measured mean stage or stage variation (Table 2). Of particular note is the observation that

Site-specific relationships between specific yield (Sy) and stage (ecosystem-specific yield, ESY; Table 3) allowed the White (1932) method to be applied at all stages. However, we limited our analysis to a minimum stage of 0.3 m since the White method may not measure the total evapotranspiration (ET) flux for deeper water tables (i.e., soil moisture in the developing unsaturated zone begins to contribute significantly to total ET; Shah et al. 2007). Mean indexed ET rates (the ratios of measured ET to estimated potential evapotranspiration [PET], ET/PET) suggest a strong amplifying effect of disturbance on wetland ET fluxes (Fig. 3a); higher ET rates (near or above PET; ET/PET 1) occurred in sites with lower condition scores, with the exception of A3. There were strong and significant negative correlations between ET/PET rates and both condition methods when excluding A3 (Table 2). Site A3 is in a heavily grazed cattle pasture, which clearly affected vegetation structure via reduced leaf-area index (LAI) (Fig. 3b); field observations suggest the near total absence of sub-canopy vegetation. In contrast, the other agricultural (including silviculture) and urban sites had increased LAI relative to the reference sites, which led to the elevated ET rates observed (Fig. 3). Linear regression of ET/PET vs. LAI yielded a strongly predictive (linear R 2 ¼ 0.78) and significant (P , 0.001) relationship; the resulting slope (0.09; data not shown) suggests a 9% increase in ET per unit increase in LAI. The frequency and mean rates of exfiltration (inflow) and infiltration (outflow) events (Fig. 4a, b) were similar

October 2013


1625

TABLE 2. Pearson correlation coefficients (r) between hydrologic metrics and ecosystem condition scores (Uniform Mitigation Assessment Method [UMAM], UMAM’s water environment [WE] category, and Florida Wetland Condition Index [FWCI]). r values for ecosystem condition scores Hydrologic metric

UMAM

UMAM-WE

FWCI

Mean stage Stage variance Flooding duration Daily stage change (positive) Daily stage change (negative) ET/PET Cumulative exfiltration/infiltration

0.32 0.08 0.11 0.20 0.24 0.36 (0.76) 0.15 (0.44)

0.29 0.08 0.18 0.26 0.16 0.08 0.02

0.26 0.23 0.16 0.01 0.21 0.38 (0.75) 0.36 (0.70)

Note: Correlation coefficients in parentheses represent values markedly improved by excluding site A3, and boldface values denote significant correlation (P , 0.05). Ratio of measured evapotranspiration (ET) to estimated potential evapotranspiration (PET).

to mean daily stage changes (Fig. 2): exfiltration events were less frequent (with the exception of site A1; Fig. 4a) but with higher mean rates (by a factor of from 1.06 to 2.61; Fig. 4b) compared to infiltration events. All sites, however, experienced episodic switching between net recharge and discharge, regardless of variation in condition (Fig. 4a). In contrast to metrics of flooding dynamics, ratios of cumulative exfiltration to infiltration showed some systematic variation across the disturbance gradient (Fig. 4c). Similar to ET/PET rates, excluding A3 improved the correlations between condition scores and ratios of exfiltration to infiltration, though correlation was only significant for FWCI (Table 2). Reference sites, along with the silviculture site (S), had higher infiltration totals than exfiltration and acted as net recharge systems over the period of study. In contrast, sites with lower condition, with the exception of A3, were closer to being balanced (i.e., exfiltration/infiltration ¼ 1; dashed line in Fig. 4c) or served as net discharge systems (i.e., net sinks). Low LAI and ET in site A3 likely explain that site’s role as a net source, since less of the stored water was lost to the atmosphere. DISCUSSION Ecosystem services derive from ecological functions, and hydrology is a key regulator of those functions in

FIG. 2. Daily stage change shows no systematic variation in mean magnitude or relative frequency of increases or decreases in stage across a disturbance gradient (i.e., decreasing UMAM scores). The relative frequency of positive stage changes are shown as percentages of days during the study period when stage increased. For site codes see Fig. 1 legend.

wetlands. To that end, we sought to document how well condition assessments, which are frequently used to guide compensatory mitigation, predict hydrologic functions of wetlands along a disturbance gradient. We argue that this requires distinguishing between impacts to the hydrologic regime imposed on a wetland, which influences ecosystem structure and function, and the estimation of other hydrologic functions (e.g., hydrologic fluxes that derive from ecosystem processes). Value to society accrues from both hydrologic regime (e.g., habitat) and hydrologic flows (e.g., water availability) (Brauman et al. 2007). Our results, which demonstrate that both can be obtained in parallel from the same time series of water-level measurements, provide a new means of synthesizing hydrologic regime and function, and explicitly test the hypothesis that condition assessment captures wetland function. Our approach to the problem focused on a single hydrogeomorphic (HGM) class of wetland, those that are hydrologically isolated (i.e., depressional); we further refined the setting by focusing only on forested wetlands in north Florida (i.e., cypress domes, an HGM subclass). Isolated wetlands, which occur in many landscapes (e.g., prairie potholes, Carolina bays, raised bogs, vernal pools; Tiner 2003), are replicated landscape units facilitating whole-ecosystem measurements (e.g.,

1626



TABLE 3. Quality of fit for ecosystem-specific yield (ESY) relationships, along with soil specific yield (Sy,soil), which represents site-specific minimum Sy values, and the Sy value (Sy,max) at maximum observed stage. Site R1 R2 R3 R4 S A1 A2 U1 U2 U3 A3

R2

Sy,soil

Sy,max

0.73 0.72 0.50 0.46 0.61 0.73 0.75 0.74 0.74 0.98 0.53

0.11 0.11 0.16 0.13 0.09 0.25 0.15 0.11 0.17 0.13 0.22

0.72 0.96 0.84 0.78 0.78 0.71 0.81 0.79 0.73 0.99 0.76

The 11 wetland sites in northern Florida, USA. Site type codes: R, reference; S, silviculture; A, agriculture; U, urban; for more site information, see Table 1.

evapotranspiration [ET], groundwater exchange) that can be treated as independent. Hydrologic investigations of isolated wetlands are additionally motivated by recent Supreme Court decisions (i.e., Solid Waste Agency of Northern Cook County v. U.S. Army Corps of Engineers (531 U.S. 159) (2001)—the SWANCC case; Rapanos v. United States (571 U.S. 715) and Carabell v. U.S. Army Corps of Engineers (126 S. Ct. 2208 (2006)) that limit Federal Clean Water Act protections over these wetland types. We fully acknowledge that

FIG. 4. (a) Frequency and (b) mean rates of daily net exfiltration (inflow) and infiltration (outflow) events reveal that all sites switched between net recharge and discharge systems and, in most cases, experienced less-frequent, but highermagnitude, exfiltration events. (c) Ratios of cumulative exfiltration to infiltration demonstrate that systems in lower land use intensities were net recharge systems over the period of study, whereas systems in higher land use intensities tended to be more balanced or serve as net sinks. The horizontal dashed line denotes neutrality between cumulative exfiltration and infiltration. Note that exfiltration was defined here as groundwater inflow plus any overland flow from storm events.

FIG. 3. (a) Measured mean daily evapotranspiration (ET) rates indexed to potential evapotranspiration (PET) (error bars denote 95% confidence intervals) and (b) leaf-area index (LAI) were higher in systems with decreased condition (i.e., lower UMAM scores) with the exception of Site A3, which was actively grazed pasture resulting in decreased LAI and ET/PET. Regressing ET/PET against LAI results in a strong and significant relationship (R 2 ¼ 0.78; P , 0.001).

selecting a single class intrinsically limits the generality of the findings. Indeed, our results, which support a more nuanced view of land cover effects on wetland hydrology than is broadly adopted, underscore the critical need for comprehensive hydrologic monitoring in wetlands spanning HGM classes (Cole 2006, Stander and Ehrenfeld 2009a, b, Gebo and Brooks 2012). The literature is replete with regional tools for assessing ecological condition; the variety and continuing emergence of such metrics underscores the magni-

October 2013


tude of the challenge of doing this well. The strong concordance between the rapid (i.e., UMAM [Uniform Mitigation Assessment Method; Bardi et al. 2004]) and intensive (i.e., FWCI [Florida Wetland Condition Index; Reiss 2006]) condition methods used in this study confirms that the basis of assessment is repeatable and generalizable. This is important and to be expected given the intensive investment in the development of these tools by Florida’s State agencies, and their widespread use for wetland assessment and mitigation. We note that UMAM was designed to primarily assess habitat integrity, which is the sole basis of assessment via FWCI, but also was designed to consider wetland attributes other than biological condition (e.g., soil erosion, topographic features, and water quality); the strong correlation between the two methods suggests that these other attributes are either of modest influence or strongly co-vary with biological condition. Does condition predict hydrologic regime? The most surprising finding of this study was that we observed no association between assessed condition and wetland hydrologic regime (Table 2). This was unexpected since UMAM and FWCI assess disturbance (i.e., departure from reference conditions), and changes in hydrologic regime have been observed in response to human disturbance across a variety of aquatic ecosystems (e.g., rivers, Richter et al. 1996; lakes, Coops et al. 2003; and floodplain wetlands, Cole and Brooks 2000). Perhaps more importantly, the UMAM water environment (WE) category, which was designed to specifically assess hydrologic alteration, did not predict variation in hydrologic metrics between sites (Table 2), underscoring the difference between inferences from rapid indicators and those from hydrologic monitoring (Stander and Ehrenfeld 2009a). Calibrating indicators of hydrologic regime with reference conditions imbeds a priori predictions about the impacts of disturbance on hydrologic regime. While land use disturbances can alter wetland hydrology (Winter 1988, Ehrenfeld 2000), Stander and Ehrenfeld (2009a) found little support for such predictions about the impacts of land use effects on hydrologic regime within HGM classes. Our results support the latter body of evidence, with no clear effect of disturbance intensity. Reference sites were indistinguishable from impacted sites in any of the regime metrics we computed, including hydroperiod, mean water depth, variation in depth (Fig. 1), and average rate of depth change (i.e., ‘‘flashiness’’; Fig. 2). In fact, reference sites occupied both extremes in mean stage (deep and shallow), and included the site with the highest stage variance (i.e., an attribute commonly associated with urbanization; Ehrenfeld et al. 2003). Even the rate-of-change metric (Fig. 2), which documents how quickly wetland stage responded to rainfall (positive rate of change) and how quickly water was lost via infiltration and ET (negative rate of change), showed no clear variation with

1627

land use intensity. The former result was particularly surprising since urban drainage effects are assumed to produce much higher inflow rates during rainfall events. There are several explanations for why we observed poor association between assessed condition and hydrologic regime. First, local soil and topographic variation may exert primary controls on flooding dynamics compared to anthropogenic influences that disturb biological condition. While these sites are all classified within the same HGM subclass, they exist within different edaphic and geologic settings (Table 1), and the influence of those factors is difficult to control for. Second, isolated wetlands are, to a large extent, fed principally by direct rainfall and a highly localized catchment. Wetland systems with surface water connections (e.g., floodplains, strands, and fringe marshes) are likely more influenced by disturbances to catchment physical conditions over larger spatial scales. Finally, it may be that the expected effects of intensive land use are only observed in more intensive settings than those sampled here. All of the urban sites are, strictly speaking, peri-urban, including single and multi-family residential settings along with golf courses and small parking lots; this may miss the most urbanized and more intensive agricultural sites. While there was no evidence of a threshold intensity above which hydrologic function ceased, that threshold plausibly exists, and requires additional sampling to identify. Does condition predict hydrologic fluxes? Hydrologic functions include groundwater and ET flows, which are difficult to measure directly. The approach presented here, using the White (1932) method, offers an empirical, low-cost quantification of both fluxes under inundated conditions and facilitates cross-site and cross-class comparisons that have, to date, been impractical. It also facilitates empirical assessments of ecohydrologic controls on these functions (e.g., the reciprocal effects of community composition, ecological succession, and nutrient status) that are beyond the scope of this study. What we can infer, however, from these direct measurements is the provision of ecosystem services closely linked with ET, such as local climate regulation, ecosystem productivity, carbon sequestration, and biogeochemical functions. ET directly indicates the rate of latent heat exchange that regulates local temperatures (Taha 1997). In addition to this direct service, measured ET provides a reasonable proxy for primary productivity (Law et al. 2002), particularly in forested systems where transpiration dominates the ET flux. The cascade of services derived from high primary production, such as carbon sequestration and biogeochemical processes that improve water quality (e.g., denitrification, Hernandez and Mitsch 2007; dehalogenation of toxins, Mohn and Tiedje 1992) were not directly measured. However, strong coupling between ET and productivity makes extrapolating differences in ET to differences in

1628


services derived from primary production plausible, particularly since hydroperiod and water depth differences that also regulate those services were not systematically affected by disturbance. In contrast to metrics of hydrologic regime, there were strong and significant correlations between assessed condition and ET, albeit negative ones (Table 2). This relationship between disturbance and ET was clearly mediated by variation in ecosystem structure (specifically, leaf-area index [LAI]). Urbanization and agriculture land uses reduce, if not completely remove fire as part of the natural regime, and also increase nutrient loads (Ehrenfeld 2000, Houlahan and Findlay 2004); both can impair biological condition but also increase vegetation structure and biomass (Keddy 2010). This general expectation is strongly supported by our observations of increased LAI in sites with lower assessed condition, with the exception of an agricultural site (A3) that was heavily grazed (Fig. 3). LAI is likely the most informative metric of vegetation structure when estimating productivity and water-use rates since it is a direct measure of the surface exchange area for carbon and water (Vose et al. 1994). The relationship between indexed ET rates (ET/PET) and LAI was particularly strong (R 2 ¼ 0.78, P , 0.001), which provides support for both the inference of links between ET and productivity, and the utility of the methods used to estimate ET. Groundwater fluxes, the other major wetland hydrologic function studied here, are arguably more difficult to measure, but are integral to regional hydrologic system function. Wetlands are commonly, but often erroneously, thought of as locations for preferential groundwater recharge (Winter 1999). Our data suggest a more nuanced role for isolated wetlands, namely their ability to buffer aquifer dynamics, which may ultimately prove even more important. Isolated depressional wetlands have been observed to regularly switch in groundwater flow direction (Winter 1999, Riekerk and Korhnak 2000) in response to differential fluctuations in groundwater and surface water levels induced by ET and rainfall. Following rainfall events, water levels in upland areas increase more rapidly than in wetlands due to Sy (specific yield) differences; this induces net exfiltration, where wetlands receive groundwater from their surroundings. The same Sy differences have the reverse effect during inter-event periods, as the groundwater is drawn down more quickly in response to ET than water levels in inundated wetlands. As such, during these periods, groundwater fluxes are out of the wetland. These sink–source reversals ultimately buffer surficialaquifer dynamics, limiting the range and rate of aquifer variation with stabilizing effects on the regional drainage network and upland ecosystems. In short, these reversals suggest that isolated wetlands provide water storage capacitance in the landscape that attenuates water-table responses to rainfall variation.


All of our sites exhibited regular sign reversals in groundwater exchange and thus, by extension, the ability to buffer aquifer dynamics, regardless of variation in assessed condition (Fig. 4). Correlations between ratios of cumulative exfiltration to infiltration and assessed condition (Table 2), however, suggest an influence of vegetation structure (and thus condition) on net groundwater fluxes. Wetlands surrounded by forested uplands (i.e., reference and silviculture land uses), as well as the low LAI agricultural site, act as net recharge systems, whereas wetlands in areas with higher land use intensity and with lower assessed condition tend to act as net discharge systems (Fig. 4c). The explanation is likely twofold, one in response to wetland vegetation properties, and the other in response to properties of the surrounding uplands. The first is that wetlands with lower LAI and ET will tend to maintain greater head differences vis-a`-vis their surroundings and create more persistent infiltrating conditions. The statistical association between LAI and the ratio of cumulative exfiltration to infiltration was significant (r ¼ 0.55, P , 0.01). The second derives from the fact that the adjacent uplands also control the hydraulic gradient. In reference and silviculture sites the adjacent uplands tend to be forests, where transpiration during periods with no rain maintains a strong gradient for wetland infiltration. Urban sites, in contrast, may lose less water to ET in the uplands and yield more water to the wetland via exfiltration following rain events. We note that both groundwater recharge to sustain surficialaquifer levels and water storage of excess flows from uplands represent important hydrologic functions that all wetlands in this study appear to provide, regardless of condition. Condition vs. function This study reinforces the contention (Hruby 2001, Stander and Ehrenfeld 2009a) that function and condition are not necessarily positively correlated, at least in settings similar to our studied systems (i.e., isolated wetlands with varying land use intensity) and in regard to the hydrologic functions that are identified as among the core services wetlands provide. This conceptual discord between what we value (i.e., a diverse set of functions that create ecosystem services) and how we commonly assess that value (i.e., departure from reference conditions) is a particular challenge for wetland management and mitigation, which are implemented to sustain and protect the many ecosystem services wetlands provide. Because of the diversity of services and the stated goal of ‘‘no net loss’’ of wetland function at landscape and regional scales, validating the often implicit hypothesis that impaired wetland condition means diminution in functionality overall, and not just functions directly linked to condition, is critically important. Our study provides direct evidence that measured condition is a poor proxy for some functions. There was

October 2013


no association observed between metrics of wetland hydrologic regime (hydroperiod, flashiness) and measures (rapid or intensive) of ecosystem condition in our isolated systems, which means that services that respond to water-level dynamics (e.g., carbon sequestration, denitrification) are not inhibited in landscapes with more intensive use. This inference is even more pronounced for those wetland functions that were directly measured. The ability of wetlands to buffer surficial aquifer dynamics, for example, was realized at all sites despite dramatic variation in ecosystem condition. Wetlands in agricultural and urban areas actually exhibited slightly greater exfiltration than infiltration (i.e., they are relative water sinks), which suggests that the service of storage of excess water is being realized. Similarly, we measured higher ET rates in urban landscapes, which translates into enhanced latent heat exchange, the cooling effects of which are acutely needed in those landscapes (Taha 1997). Increased LAI and ET with decreasing condition scores supports the conclusion that primary production actually increases with declining condition. This, in turn, suggests increased potential for carbon sequestration and denitrification, a process that is prevalent and even amplified in disturbed wetlands (Jordan et al. 2007, Stander and Ehrenfeld 2009a). Location is critically important for delineating which wetland functions are realized as services as opposed to those that remain potential services (Hruby 1999). In other words, the provision of ecosystem services is context dependent, especially for services that are provided at local spatial scales. For example, it is clear that wetlands in minimally altered landscapes sustain higher habitat quality, meaning that biodiversity services are realized there and are degraded in other settings where, among other things, dispersal, nutrient loading, species composition, and fire frequency are changed. However, the contributions of wetlands in minimally impacted settings to mitigating local microclimate warming, improving degraded water quality, and/or storing excess water remain as potential services because such impairments do not occur in those settings. In contrast, while urban and agricultural wetlands clearly have reduced habitat functions (i.e., consistently low condition scores), other services (increased latent heat exchange, water storage, and water-quality improvements) are realized and perhaps even enhanced because these wetlands are in the proper locations to ameliorate stresses typical of human-dominated landscapes. To complicate matters further, the effects of certain functions are themselves context dependent. For example, stimulation of ET may represent a beneficial service in some landscape settings where it contributes to alleviating urban heat-island effects; in another landscape, however, additional water use actually decreases water yield to the regional hydrologic system, which may be unfavorable. Similarly, storage of water may be at once beneficial (e.g., increasing denitrification) and

1629

deleterious (e.g., support vectors of disease) (Ehrenfield 2004). In short, the enumeration of the value of a wetland, in part, is a dialog about prioritizing among the many services it provides. Amplification of certain functions (e.g., latent heat exchange, carbon sequestration, and water-quality improvements) may be at the expense of others (e.g., habitat), representing a trade-off between ecosystem services. It is this trade-off that is not fully addressed with current wetland regulation and mitigation, which heavily weights habitat integrity and underestimates the value of ‘‘working wetlands.’’ While not all assessment methods are restricted to habitat integrity (e.g., HGM), such methods use reference sites to calibrate structural indicators of function (Hauer and Smith 1998), and make the explicit assumption that ‘‘optimal’’ functional capacity occurs in reference conditions. Operationally, this means that disturbance (i.e., departure from reference conditions) reduces wetland value a priori, even where specific functions are enhanced (Jordan et al. 2007). The disconnect between condition and function critically underscores the distinction between assessing disturbance and enumerating wetland value. Recognizing this distinction, Fennessy et al. (2007) suggest an approach where condition assessment scores can be adjusted with ‘‘value-added metrics’’ to account for functions of degraded systems identified as valuable but that are not captured by the method applied. Alternatively, methods like HGM that specifically assess multiple services via functional indicators can be further calibrated with measured functions across a disturbance gradient (e.g., Jordan et al. 2007) rather than calibrated only to reference conditions. The current assumption of optimality in reference conditions creates circumstances where enhanced function (e.g., higher denitrification in degraded wetlands; Jordan et al. 2007) is interpreted as departure from reference conditions and thus leads to lower scores. Calibrating functional indicators with measured functions, however, would provide means to identify specific functions that are enhanced (or maintained) in degraded systems and allow for scores to be adjusted accordingly. Score-adjustment methods (e.g., adjusting HGM scores, ‘‘value-added metrics’’ for condition-based assessments) are tenable for refining ecosystem value provided that efficient methods are identified to quantify wetland function within specific wetland classes. Because of the rapid assessment requirements of most assessment methods, this ultimately necessitates rapid and specific proxies of measured rates (e.g., soil characteristics for denitrification rates [Jordan et al. 2007]; LAI for ET rates [this study]). Finally, this approach requires a formal weighting mechanism to assign relative importance to individual functions and to enable trade-off analyses between functions. We note, however, that a weighting scheme is somewhat context specific in response to the fact that

1630


different landscape locations benefit from different ecosystem services. The goal of ‘‘no net loss’’ of wetlands requires tools for measuring the myriad functions (or their proxies) that wetlands provide and then weighting the contextdependent importance of those functions. We are optimistic that current assessment methods can be adjusted to consider what we argue is undervaluation of the functions and services of disturbed wetlands. As these adjustments accrue, however, it may become apparent that all wetlands are uniformly valuable (for different reasons). Arriving at this conclusion would have the advantage of simplifying wetland assessment dramatically, but clearly requires a level of empirical support not yet available in the literature. Our intent in the provocative suggestion that, after explicitly considering functions and trade-offs, wetlands may be uniformly valuable is not to question the importance or validity of ecosystem condition (e.g., biodiversity and habitat functions), but rather to critically evaluate the implied correlation between condition and function and thus challenge the way in which we value systems that are judged to be degraded. This argument is central to wetland mitigation but is also general to ecosystem management. Conservation and restoration of critical habitats should remain a principal priority, but it is also imperative to recognize the importance of ‘‘working’’ ecosystems proximate to disturbance, where the consequences of human actions are most acute and where ecosystem services that mitigate the negative effects of those stressors are realized. ACKNOWLEDGMENTS This work was supported by the Environmental Protection Agency under grant EPA #L CD-95417909-0. We gratefully acknowledge the field assistance of Joseph Delasantro, Jake Diamond, and David Kaplan. LITERATURE CITED Bardi, E., M. T. Brown, K. C. Reiss, and M. J. Cohen. 2004. UMAM. Uniform mitigation assessment method training manual. Web-based training manual for Chapter 62-345 FAC for wetlands permitting, prepared by the Center for Wetlands, University of Florida. Florida Department of Environmental Protection, Talahassee, Florida, USA. Brauman, K. A., G. C. Daily, T. K. Duarte, and H. A. Mooney. 2007. The nature of value and ecosystem services: an overview highlighting hydrologic services. Annual Review of Environment and Resources 32:67–98. Brinson, M. M. 1993. A hydrogeomorphic classification for wetlands. Wetlands Research Program Technical Report WRP-DE-4. U.S. Army Corps of Engineers Waterways Experimental Station, Vicksburg, Mississippi, USA. el.erdc. usace.army.mil/wetlands/pdfs/wrpde4.pdf Brinson, M. M., and R. Rheinhardt. 1996. The role of reference wetlands in functional assessment and mitigation. Ecological Applications 6:69–76. Chen, J. M., P. M. Rich, S. T. Gower, J. M. Norman, and S. Plummer. 1997. Leaf area index of boreal forests: theory, techniques, and measurements. Journal of Geophysical Research 102:429–443.


Cole, C. A. 2006. HGM and wetland functional assessment: six degrees of separation from the data? Ecological Indicators 6:485–493. Cole, C. A., and R. P. Brooks. 2000. Patterns of wetland hydrology in the Ridge and Valley Province, Pennsylvania, USA. Wetlands 20:438–447. Coops, H., M. Beklioglu, and T. L. Crisman. 2003. The role of water-level fluctuations in shallow lake ecosystems—workshop conclusions. Hydrobiologia 506–509:23–27. Deimeke, E. 2009. Efficacy of condition assessments to predict ecosystem function in isolated wetlands. Thesis. University of Florida, Gainesville, Florida, USA. Drexler, J. Z., R. L. Snyder, D. Spano, and K. T. Paw U. 2004. A review of models and micrometeorological methods used to estimate wetland evapotranspiration. Hydrological Processes 18:2071–2101. Ehrenfeld, J. G. 2000. Evaluating wetlands within an urban context. Ecological Engineering 15:253–265. Ehrenfeld, J. G. 2004. The expression of multiple functions in urban forested wetlands. Wetlands 24:719–733. Ehrenfeld, J. G., H. B. Cutway, R. Hamilton IV, and E. Stander. 2003. Hydrological description of forested wetlands in northeastern New Jersey, USA—an urban/suburban region. Wetlands 23:685–700. Ewel, K. C. 1990. Multiple demands on wetlands. Florida cypress swamps can serve as a case study. BioScience 40:660– 666. Fennessy, M. S., A. D. Jacobs, and M. E. Kentula. 2004. Review of rapid assessment methods for assessing wetland condition. EPA/620/R-04/009. U.S. Environmental Protection Agency, Washington, D.C., USA. Fennessy, M. S., A. D. Jacobs, and M. E. Kentula. 2007. An evaluation of rapid methods for assessing the ecological condition of wetlands. Wetlands 27:543–560. Gebo, N. A., and R. P. Brooks. 2012. Hydrogeomorphic (HGM) assessments of mitigation sites compared to natural reference wetlands in Pennsylvania. Wetlands 32:321–331. Hargreaves, G. H., and Z. A. Samani. 1985. Reference crop evapotranspiration from temperature. Applied Engineering in Agriculture 1:96–99. Hauer, F. R., and R. D. Smith. 1998. The hydrogeomorphic approach to functional assessment of riparian wetlands: evaluating impacts and mitigation on river floodplains in the U.S.A. Freshwater Biology 40:517–530. Healy, R. W., and P. G. Cook. 2002. Using groundwater levels to estimate recharge. Hydrogeology Journal 10:91–109. Heimburg, K. 1976. Hydrology of some northcentral Florida cypress domes. Thesis. University of Florida, Gainesville, Florida, USA. Hernandez, M. E., and W. J. Mitsch. 2007. Denitrification potential and organic matter as affected by vegetation community, wetland age, and plant introduction in created wetlands. Journal of Environmental Quality 36:333–342. Houlahan, J. E., and C. S. Findlay. 2004. Estimating the ‘‘critical’’ distance at which adjacent land-use degrades wetland water and sediment quality. Landscape Ecology 19:677–690. Hruby, T. 1999. Assessments of wetland functions: what they are and what they are not. Environmental Management 23:75–85. Hruby, T. 2001. Testing the basic assumption of the hydrogeomorphic approach to assessing wetland functions. Environmental Management 27:749–761. Jordan, T. E., M. P. Andrews, R. P. Szuch, D. F. Whigham, D. E. Weller, and A. D. Jacobs. 2007. Comparing functional assessments of wetlands to measurements of soil characteristics and nitrogen processing. Wetlands 27:479–497. Karr, J. R., and E. W. Chu. 1997. Biological monitoring: essential foundation for ecological risk assessment. Human and Ecological Risk Assessment: An International Journal 3:993–1004.

October 2013


Keddy, P. A. 2010. Wetland ecology. Principles and conservation. Second edition. Cambridge University Press, New York, New York, USA. Law, B. E., et al. 2002. Environmental controls over carbon dioxide and water vapor exchange of terrestrial vegetation. Agricultural and Forest Meteorology 113:197–120. Loheide, S. P., II, J. J. Butler, Jr., and S. M. Gorelick. 2005. Estimation of groundwater consumption by phreatophytes using diel water table fluctuations: a saturated–unsaturated flow assessment. Water Resources Research 41(7):W07030. http://dx.doi:org/10.1029/2005WR003942 Lopez, R. D., and M. S. Fennessy. 2002. Testing the floristic quality assessment as an indicator of wetland condition. Ecological Applications 12:487–497. Mack, J. J. 2007. Developing a wetland IBI with statewide application after multiple testing iterations. Ecological Indicators 7:864–881. McLaughlin, D. L., and M. J. Cohen. 2011. Thermal artifacts in measurements of fine-scale water level variation, Water Resources Research 47(9):W09601. http://dx.doi.org/10. 1029/2010WR010288 McLaughlin, D. L., and M. J. Cohen. 2013. Ecosystem specific yield for estimating evapotranspiration and groundwater exchange from diel surface water variation. Hydrological Processes. http://dx.doi.org/10.1002/hyp.9672 MEA [Millenium Ecosystem Assessment]. 2005. Ecosystems and human well-being: wetlands and water synthesis. World Resources Institute, Washington, D.C., USA. Mitsch, W. J., and J. G. Gosselink. 2007. Wetlands. John Wiley, Hoboken, New Jersey, USA. Mohn, W. M., and J. M. Tiedje. 1992. Microbial reductive dehalogenation. Microbial Reviews 56:482–507. NRCS. 2012. Natural Resources Conservation Service Soil Survey. http://websoilsurvey.nrcs.usda.gov/app/WebSoilSurvey. aspx Peters, E. B., R. V. Hiller, and J. P. Mcfadden. 2011. Seasonal contributions of vegetation types to suburban evapotranspiration. Journal of Geophysical Research 116:GO1003. http:// dx.doi.org/doi: 10.1029/2010JG001463 Que´tier, F., and S. Lavorel. 2011. Assessing ecological equivalence in biodiversity offset schemes: key issues and solutions. Biological Conservation 144:2991–2999. Reiss, K. C. 2006. Florida wetland condition index for depressional forested wetlands. Ecological Indicators 6:337– 352.

1631

Richter, B. D., J. V. Baumgartner, J. Powell, and D. P. Braun, D. P. 1996. A method for assessing hydrologic alteration within ecosystems. Conservation Biology 10:1163–1174. Ricketts, T. H, et al. 2008. Landscape effects on crop pollination services: are there general patterns. Ecology Letters 11:499–515. Riekerk, H., and L. V. Korhnak. 2000. The hydrology of cypress wetlands in Florida pine flatwoods. Wetlands 20:448–460. Shah, N., M. Nachabe, and M. Ross. Extinction depth and evapotranspiration from ground water under selected land covers. Ground Water 45:329–338. Sophocleous, M. 2002. Interactions between groundwater and surface water: the state of the science. Hydrogeology Journal 10:52–67. Spronken-Smith, R. A., T. R. Oke, and W. P. Lowry. 2000. Advection and the surface energy balance across an irrigated urban park. International Journal of Climatology 20:1033– 1047. Stander, E. K., and J. G. Ehrenfeld. 2009a. Rapid assessment of urban wetlands: Do hydrogeomorphic classification and reference criteria work? Environmental Management 43:725–742. Stander, E. K., and J. G. Ehrenfeld. 2009b. Rapid assessment of urban wetlands: Functional assessment model development and evaluation. Wetlands 29:261–276. Taha, H. 1997. Urban climates and heat islands: albedo, evapotranspiration, and anthropogenic heat. Energy and Buildings 25:99–103. Tiner, R. W. 2003. Geographically isolated wetlands of the United States. Wetlands 23:494–516. Vose, J. M., P. M. Dougherty, J. N. Long, F. W. Smith, H. L. Gholz, and P. J. Curra. 1994. Factors influencing the amount and distribution of leaf area of pines stands. Ecological Bulletins (Copenhagen) 43:102–114. White, W. N. 1932. A method of estimating ground-water supplies based on discharge by plants and evaporation from soil—results of investigation in Escalante Valley, Utah. U.S. Geological Survey Water Supply Paper 659-A. pubs.usgs. gov/wsr/0659a/report.pdf Winter, T. C. 1988. A conceptual framework for assessing cumulative impacts on the hydrology of nontidal wetlands. Environmental Management 12:605–620. Winter, T. C. 1999. Relation of streams, lakes, and wetlands to groundwater flow systems. Hydrogeology Journal 7:28–45.