Plant responses to increased inundation and salt exposure: interactive effects on tidal marsh productivity
- First Online:
- Cite this article as:
- Janousek, C.N. & Mayo, C. Plant Ecol (2013) 214: 917. doi:10.1007/s11258-013-0218-6
- 580 Downloads
Flooding and high salinity generally induce physiological stress in wetland vascular plants which may increase in intensity with sea-level rise (SLR). We tested the effects of these factors on seedling growth in a transplant experiment in a macrotidal estuary in the Pacific Northwest. Seven common wetland species were grown at mean higher high water (MHHW, a typical mid-marsh elevation), and at 25 and 50 cm below MHHW in oligohaline, mesohaline, and polyhaline marshes. Increased flooding reduced shoot and root growth in all species, including those typically found at middle or lower tidal elevations. It also generally disproportionately reduced root biomass. For more sensitive species, biomass declined by >50 % at only 25 cm below MHHW at the oligohaline site. Plant growth was also strongly reduced under polyhaline conditions relative to the less saline sites. By combining inundation and salinity time-series measurements we estimated a salt exposure index for each site by elevation treatment. Higher values of the index were associated with lower root and shoot biomass for all species and a relatively greater loss of below-ground than above-ground production in most species. Our results suggest that inundation and salinity stress individually and (often) interactively reduce productivity across a suite of common marsh species. As relative SLR increases the intensity of stress on coastal marsh plants, negative effects on biomass may occur across a range of species and especially on below-ground production.
KeywordsBiomass Ecosystem function Physiological stress Salinity Sea-level rise Submergence
Salt marsh plants occupy the upper intertidal zone of estuaries where steep abiotic gradients of inundation and salinity can strongly impact species distributions and productivity (Pennings et al. 2005; Engels et al. 2011). Longer submergence times lead to greater physiological stress for vascular plants by reducing diffusion of oxygen and CO2, increasing concentrations of toxic substances such as reduced metal ions, and decreasing light availability (Blom and Voesenek 1996; Colmer and Voesenek 2009). Field and mesocosm studies have shown that growth rates in a number of common salt marsh species (e.g., Spartina spp.) decline with increased flooding, although responses are not always linear and can vary by species (Cherry et al. 2009; Kirwan and Guntenspergen 2012; Voss et al. 2013).
Salinity also generally negatively affects estuarine plant growth, particularly for tidal fresh or oligohaline species (Parrondo et al. 1978; Hester et al. 2001; Spalding and Hester 2007; Guo and Pennings 2012). High salinity can impact water potential in plant tissues, impose ion toxicity, and lower nutrient uptake (Gorham et al. 1985). Changes in salinity and flooding also substantially alter the biogeochemical environment of wetland soils. For instance, increased flooding lowers redox potentials (McKee and Mendelssohn 1989) and elevated sulfide concentrations accompany more saline conditions (Howard and Mendelssohn 2000). High sulfide concentrations are toxic for vascular plants and lower biomass, possibly via suppression of nutrient uptake (Koch et al. 1990).
Variation in the distribution of species along salinity and tidal elevation gradients in estuaries suggests that physiological tolerances to these abiotic factors vary among taxa (Wang et al. 2010). High marsh species may be more sensitive to prolonged flooding or higher salt exposure found at lower tidal elevations (Bertness and Ellison 1987). However, relationships between abiotic gradients and plant growth are less well known for species outside of the US Atlantic and Gulf coasts (but see Pearcy and Ustin 1984, for example). In the Pacific Northwest, lower or mid-marsh species such as Triglochin maritima, Jaumea carnosa and Sarcocornia perennis are likely to have greater tolerance to flooding or salinity exposure, while the productivity of high marsh species such as Juncus balticus may decline as these stressors increase. Similarly, species tending to grow in less saline areas of Pacific Northwest estuaries (e.g., Achillea millefolium and Carex lyngbyei) may be negatively affected by higher salinity.
Changes in the intensity of flooding or salinity exposure may also differentially affect above-ground versus below-ground production. For instance, studies with several marsh grasses suggest that stress due to these factors has a stronger negative impact on root production than on above-ground biomass (Rozema and Blom 1977; Langley et al. 2013), but the generality of this finding is uncertain. Theory suggests that plants allocate growth to above or below-ground structures in response to limiting resources such as light and nutrients (Tilman 1988), yet abiotic stressors may also affect this fundamental tradeoff.
Because coastal sea-level rise (SLR) due to climate change is expected to increase flooding (and may concurrently increase salt exposure; Parker et al. 2011), productivity rates for many marsh species may decline in the future. Loss of above- or below-ground biomass is likely to negatively affect many important salt marsh functions such as trophic support for estuarine consumers (Wainright et al. 2000), carbon sequestration (Chmura et al. 2003), and soil accretion (Nyman et al. 2006; Li and Yang 2009). Recent estimates of global SLR by the year 2100 vary widely (0.3–1.8 m; Nicholls and Cazenave 2010), but all scenarios suggest that vegetation in future marshes will experience greater flooding if accretion cannot keep pace with SLR (Stralberg et al. 2011).
Using a field transplant experiment, we tested the effects of tidal inundation and salinity on productivity in seven species of marsh plants common in the Pacific Northwest. We tested the following hypotheses: (i) increased submergence and salinity lower root and shoot biomass, (ii) salinity stress exacerbates the negative effects of flooding on plant growth, (iii) inundation shifts ratios of shoot versus root production, and (iv) species vary in their responses to inundation and salinity stress with stronger negative responses in species typically distributed in high marshes or in low salinity areas of estuaries. Since our experiment spanned a range of salinity values typical of temperate estuaries and was consistent with rates of global SLR that may be realized by the end of the 21st century given poor wetland accretion, our study offers tentative predictions about the direction and magnitude of SLR effects on wetland productivity in the region.
Materials and methods
At each site, 18 seedlings per species were randomly divided among six experimental blocks situated along the edge of tidal creeks (Fig. 2). Each block contained three plants per species randomly assigned to one of three elevation treatments: MHHW (a typical mid-marsh elevation) and 25 and 50 cm below MHHW (more typical of low marsh). Terraces were sculpted as needed into the banks of the channels to accommodate the range of elevations tested. The target elevations in each wetland were located using RTK GPS (Trimble R8 receiver) with previously determined relationships between NAVD88 and local tidal datums obtained from benchmark surveys (Janousek and Folger 2012) and by leveling. 378 seedlings were used in the experiment (7 species × 3 sites × 3 elevations × 6 replicate blocks).
Blocks were generally established in southerly-facing areas to decrease shading during the daily course of the sun’s transit. Incident light was measured at ~5–10 cm above each target elevation by taking duplicate measurements with a horizontally-held LiCOR spherical PAR sensor in all blocks at all sites on two separate sunny days (between about 8:30 and 14:45 PST) during the course of the experiment. Light intensity at 25 and 50 cm below MHHW was slightly lower than intensity at MHHW, but this difference seldom exceeded 8 %. During early morning or late evening hours, particular blocks and/or elevations may have been shaded more substantially.
Seedlings were transplanted to the field on 11, 12 June. Pots were partly buried in sediment with the top of each pot at its target elevation. Nearby natural vegetation that might shade plants was trimmed or pinned back. On 14 June, 19 plants that had been lost during the first few days of the experiment (most perhaps dislodged by curious birds) were replaced. Thereafter, plants were checked at least once weekly to assess seedling survival. Death due to physiological intolerance was manifest as wilting and browning tissue (n = 75 of 378 total plants). Missing or severely disturbed (e.g., heavily grazed) plants were not used in data analyses (n = 41). Heavy grazing was rare but easily distinguished from physiological intolerance because remaining tissues were still green. Many missing plants may have died due to physiological intolerance and washed away before observation, so our estimates of mortality and biomass reductions may be somewhat conservative. Weekly maintenance of the experiment included wrack removal, trimming of surrounding vegetation likely to shade plants, or re-addition of potting soil as needed.
To quantify flooding duration and salinity during the experiment, Odyssey water level loggers (set inside stilling wells) and conductivity loggers (Dataflow Systems, Christchurch, NZ; precisions = 0.1 mS cm−1 and ~0.15 cm respectively) were deployed on the banks of the tidal channels at 50 cm below MHHW at each site (Online Resource 1). Prior to deployment, all loggers were checked for calibration accuracy in the lab and minor adjustments to data were made with logger-specific linear regression equations. Once deployed in the field, logger elevations were determined with 4–6 replicate RTK measurements; all were within five cm vertically of the target elevation. The loggers recorded data every 6 min.
Plants were retrieved from the field on 18 July after 5 weeks of exposure to the treatments. In the lab, plants were separated from the pots and rinsed over 1 mm sieves to remove soil. Each plant was cut into shoot and root sections (roots were not processed for Agrostis) and then dried (≥3 days at 60 °C) for biomass determination. Any dead but attached leaves were included in above-ground biomass estimates if the plant itself was alive (they were generally uncommon). Some fine roots were inevitably lost in processing the samples. A subset of samples was re-dried and reweighed to ensure that drying was relatively complete.
Analyses were conducted in R 2.14.1 (R Development Core Team 2012). To test inundation, site (salinity), and their interactive effects on growth, dry weights were analyzed with linear mixed models (package ‘lme4’) followed by Type III analysis of deviance (package ‘car’) where site and elevation were fixed factors and block nested within site was the random factor. Each species was analyzed separately. Plants determined to have died due to physiological intolerance were treated as having zero biomass. Root and shoot biomass values were square-root transformed prior to analyses because this generally improved homoscedasticity and/or normality.
Elevation effects on shoot:root biomass ratios [ln(shoot mass/root mass)] were also compared for all species (except Agrostis) within the oligohaline site with mixed models (block was the random factor), Type III analysis of deviance, and pair-wise Tukey’s HSD means comparisons with the package ‘lsmeans’ at α = 0.05. Zero biomass values for dead plants could not yield ratios, and since all Achillea, Deschampsia, Grindelia and Juncus plants died in at least one site by inundation treatment (e.g., 50 cm below MHHW at the polyhaline site), shoot:root ratios were only analyzed for the oligohaline site for those species. Most Triglochin and Plantago plants survived in all treatments, so site (salinity) and site by elevation effects on ratios were also examined for those species using two factor mixed models as per analyses of shoot dry mass data.
Using the time series of salinity and water level obtained at each site, a salinity exposure index was determined for each elevation by site combination (nine treatments) by summing all salinity values recorded during submergence at each elevation and then dividing sums by the total number of values in the time series (~8,700 points over 5 weeks). The resulting index was unitless and theoretically would range from 0 to ~33 in non-hypersaline estuaries. The index does not account for salts left in soil porewater during emersion. Review of the salinity time series data suggested that there was some salinity stratification in the water column during high tides, so our method of computation slightly overestimated cumulative salinity exposure at MHHW and 25 cm below MHHW. Shoot and root biomass values (square root-transformed) and ln(shoot mass/root mass) values for each species were compared linearly with the index by type I regression to determine the slope of the relationship (species sensitivity) and amount of variance accounted for by the index (R2).
Plants at MHHW were inundated for 3–5 % of the experimental period while plants transplanted to 25 and 50 cm below MHHW were flooded for 9–13 and 17–20 % of the experimental period respectively. Plants at the oligohaline site tended to experience somewhat longer inundation times than seedlings at the other sites due perhaps to ~5 cm measurement error in the initial determination of the vertical position of MHHW by GPS. During high tide events, daily mean and maximum salinities varied significantly among the sites (repeated measures ANOVAs, both P < 0.0001), with maximum salinities reaching ~30 at the polyhaline site (Fig. 3). Salinity increased modestly over the course of the experiment as spring precipitation gave way to drier summer conditions on the Oregon coast. Plant survival was high (≥80 %) for all species at all elevations at the oligohaline site, but was reduced at one or more elevations at the other sites for most species (Online resource 2).
Shoot and root growth
Aboveground mean (±SE) dry mass (g) for seven common species of marsh plants grown at MHHW and 25 and 50 cm below MHHW in oligohaline, mesohaline and polyhaline marshes in the Yaquina Estuary
Species and elevation
Factor effects (χ2)
Site × elevation
0.715 ± 0.167
0.838 ± 0.101
0.028 ± 0.028
0.450 ± 0.105 (−37)
0.037 ± 0.037 (−96)
0.000 ± 0.000 (−100)
0.198 ± 0.040 (−72)
0.000 ± 0.000 (−100)
0.000 ± 0.000 (−100)
0.563 ± 0.137
0.656 ± 0.128
0.116 ± 0.027
0.232 ± 0.084 (−59)
0.183 ± 0.035 (−72)
0.006 ± 0.004 (−95)
0.119 ± 0.020 (−79)
0.049 ± 0.008 (−93)
0.000 ± 0.000 (−100)
0.191 ± 0.037
0.176 ± 0.027
0.037 ± 0.008
0.092 ± 0.010 (−52)
0.062 ± 0.016 (−65)
0.009 ± 0.006 (−77)
0.063 ± 0.008 (−67)
0.028 ± 0.008 (−84)
0.000 ± 0.000 (−100)
0.272 ± 0.051
0.516 ± 0.083
0.228 ± 0.034
0.184 ± 0.027 (−32)
0.223 ± 0.050 (−57)
0.025 ± 0.011 (−89)
0.194 ± 0.060 (−29)
0.153 ± 0.047 (−70)
0.000 ± 0.000 (−100)
Juncus balticus ssp. ater
0.049 ± 0.012
0.047 ± 0.010
0.005 ± 0.005
0.021 ± 0.003 (−58)
0.025 ± 0.007 (−47)
0.000 ± 0.000 (−100)
0.006 ± 0.002 (−88)
0.002 ± 0.002 (−96)
0.000 ± 0.000 (−100)
0.623 ± 0.071
0.716 ± 0.083
0.283 ± 0.065
0.240 ± 0.037 (−62)
0.293 ± 0.037 (−59)
0.130 ± 0.024 (−54)
0.188 ± 0.020 (−70)
0.164 ± 0.027 (−77)
0.047 ± 0.021 (−83)
0.029 ± 0.007
0.032 ± 0.008
0.018 ± 0.001
0.017 ± 0.002 (−41)
0.024 ± 0.010 (−25)
0.005 ± 0.002 (−73)
0.011 ± 0.002 (−63)
0.013 ± 0.002 (−61)
0.001 ± 0.001 (−93)
Salinity exposure index
Linear relationships between square root-transformed shoot or root biomass and the cumulative salt exposure index
Juncus balticus spp. ater
Prolonged flooding and increasing salinity generally negatively affect salt marsh plant growth, with variation in responses among species (Konisky and Burdick 2004). Climate change impacts such as sea-level rise may exacerbate these stressors with consequent declines in coastal wetland productivity. The results of our study contribute to this body of knowledge by (i) documenting negative effects of elevated salinity and inundation on plant production across a suite of common species and functional groups in Pacific Northwest marshes, (ii) showing correlated and interactive effects of salinity and flooding on production in several species, and (iii) demonstrating greater impacts of flooding stress on root productivity than on above-ground biomass. We discuss the implications of these findings for understanding marsh plant responses to abiotic stressors and potential SLR impacts on tidal wetland structure and function.
Flooding and salinity effects on growth
Inundation duration had a negative effect on above-ground and below-ground biomass in all of the species we studied. For example, there was a reduction in short-term shoot and root growth by more than 40 % for most species grown at 25 cm below MHHW, corresponding with about a two-fold increase in flooding times relative to MHHW. In fact, growth was severely reduced by greater flooding for the middle or high marsh species Achillea, Agrostis, Juncus, and Deschampsia. Some of this reduction in growth was due to having a higher frequency of dead plants at these depths (we assigned zero biomass values to dead plants for the statistical analyses), but surviving plants also tended to be smaller with depth. For instance, there were no dead Achillea individuals at the oligohaline site, but plants grown at 25 and 50 cm below MHHW had 37 and 72 % lower shoot mass respectively than the least flooded treatment.
Negative inundation effects on tidal wetland plant growth have been observed for many Atlantic and Gulf coast marsh species (McKee and Mendelssohn 1989; Gough and Grace 1998; Baldwin et al. 2001; Konisky and Burdick 2004). For instance, Gough and Grace (1998) found that increased submergence strongly decreased Spartina patens biomass in the Gulf of Mexico. West coast species have been relatively poorly studied, but in British Columbia, Pidwirny (1990) observed a strong positive relationship between tidal elevation and above-ground biomass in Schoenoplectus americanus, a pattern that may be driven by inundation.
Individual salt marsh species responses to inundation may decline monotonically with greater flooding, or may exhibit some optimal flooding level beyond which increasing or decreasing inundation leads to declines in production. Morris et al. (2002) showed that Spartina alterniflora production was relatively low in higher marsh and hypothesized that after a maximum growth rate at about 40 cm below mean high tide, growth would again decline with further submergence. Such unimodal relationships between biomass and flooding have also been described for S. patens and Schoenoplectus americanus (Kirwan and Guntenspergen 2012). Although the number of elevation treatments we tested wasn’t sufficient to identify the shape of the flooding-biomass response, it is notable that growth declined at elevations below MHHW in all of the species we studied, including a species common in lower marsh habitat (Triglochin) and one presumed to be more tolerant of high flooding conditions based on previous research (Agrostis; Rozema and Blom 1977). Our results suggest that elevations below MHHW are not physiologically optimal for a range of common species in Pacific Northwest marshes.
Our results also suggest a strong and consistently negative effect of salinity on plant growth. Seedlings of all species had the lowest biomass at the most saline site. Achillea, Deschampsia, Agrostis and Juncus grew particularly poorly there. However, we found that many species (Grindelia, Juncus, Triglochin, Agrostis, and Plantago) grew approximately equally well at the mesohaline and oligohaline sites. In fact, Grindelia had maximum growth at the mesohaline site when grown at MHHW. Guo and Pennings (2012) showed that some halophytes (Sarcocornia and Batis) actually grew better under more saline conditions, exceptions to the well entrenched ecological paradigm that competition, not physiological preference, limits some halophytes to more stressful niches. We should note that the three sites in our study may have differed in other environmental characteristics (in addition to salinity) such as cumulative daily irradiance. However, even if salinity was not fully responsible for the geographic differences observed in growth, our data clearly show that plant growth varied spatially within the estuary. These experimental results are consistent with field observations across several Oregon estuaries suggesting that higher above-ground biomass tends to be found in less saline wetlands (Janousek, personal observation).
Overall, our data show inter-specific variation in sensitivity to salinity and flooding stress but also demonstrate that negative effects on biomass occurred across a range of plant functional groups in tidal marshes: succulent and non-succulent forbs (Triglochin, Plantago, Achillea), perennial grasses (Agrostis, Deschampsia), a rush (Juncus) and a sub-shrub (Grindelia). As observed in other work (McKee and Mendelssohn 1989; Gough and Grace 1998), flooding and salinity effects on productivity varied to some degree by species. In our study, the succulent forbs Plantago and Triglochin appeared to be the most tolerant of both submergence and salinity exposure, consistent with predictions from field distributions. Plantago is not particularly common in the Yaquina estuary, it but appears to favor low to middle marsh elevations in more saline marshes. Triglochin grows across a wide range of salinities and elevations in Oregon estuaries, but it is typically most abundant in low marsh.
Root versus shoot production
We found that flooding disproportionately reduced root production in the majority of species we studied. Relative root growth in Spartina patens and Schoenoplectus americanus also decreased with greater inundation (Kirwan and Guntenspergen 2012; Langley et al. 2013). The mechanisms underlying flooding-induced shifts in shoot:root ratios in our study are unknown, but theory about plant growth strategies in response to resource acquisition may be germane (Tilman 1988). Plants that were submerged longer may have had more access to nutrients; phosphate, for instance, may become more available at low soil redox levels (Lamers et al. 1998). Darby and Turner (2008) showed that shoot:root ratios increased in Spartina alterniflora with phosphorus additions, presumably due to decreased root foraging (see also Deegan et al. 2012 for nitrogen effects). Changes in nutrient availability may thus represent an important link between flooding stress and plant growth strategies.
Similar to arguments about below-ground resource availability, changes in shoot:root ratios may have occurred because plants at lower tidal elevations were possibly more light limited (due to longer submergence in turbid waters or because there was greater silt deposition on leaves at depth). Finally, more waterlogged and anoxic soils may have promoted shallow root growth and restricted the total soil volume utilized by plant roots, ultimately affecting root biomass.
Plantago and Triglochin growth responses in our study suggested that elevated salinity may also impact shoot:root ratios, but only in combination with greater flooding (Online resource 4). Rozema and Blom (1977) found a similar interactive effect on Agrostis stolonifera: inundation increased shoot:root ratios only under moderately saline conditions (50 % seawater) but not in freshwater. On the other hand, shoot:root dry mass ratios in Juncus gerardii tended to increase in their study with more flooding at both salinities. Haines and Dunn (1976) found that high salinity lowered shoot:root ratios in Spartina alterniflora but did not affect shoot:rhizome ratios. Salinity effects on the relative allocation of biomass to plant structures appear to vary among wetland species (Parrondo et al. 1978 and references therein, Pearcy and Ustin 1984). More research is needed to understand how abiotic gradients in salt marshes differentially affect allocation of above- versus below-ground biomass.
Interactive effects of flooding and salinity
Dual salinity and flooding stresses may have more severe impacts on marsh plant productivity than changes in only one factor. In mesocosm and field studies, Konisky and Burdick (2004) and Spalding and Hester (2007) found interactive effects of inundation duration and salinity exposure on production across a variety of marsh species. We found significant site by inundation interactions for Achillea, Grindelia and Juncus. Interactive effects may be particularly prominent in species with a low tolerance for salinity because increased flooding lengthens exposure periods to osmotic stress and may lead to salt accumulation in marsh soils (Spalding and Hester 2007). One species for which there was a particularly strong interactive effect (Achillea) tends to occur in fresher marshes in the Pacific Northwest. Additionally, the combination of increasing submergence and elevated salinity may increase sulfide toxicity (Webb and Mendelssohn 1996).
Regardless of the specific abiotic factors that were physiologically averse to plant growth (e.g., sulfide toxicity, soil hypoxia) relationships between biomass and the combined stress imposed by salinity and inundation were well captured by our salinity exposure index. Growth of all species was negatively correlated with this index, and it explained a relatively large proportion of variation in growth response (>50 %) in at least half of the species tested. For instance, the calculated index was very similar between MHHW at the polyhaline site (0.9) and 50 cm below MHHW at the oligohaline site (1.0) and a number of species (Agrostis, Grindelia, Juncus) showed very similar above-ground biomass responses (Table 1). Given spatial data on inundation duration and salinity within tidal wetland ecosystems, it may be a useful tool to predict or map variation in plant production.
Sea-level rise implications
The reduced plant growth observed in our study may have several important consequences for marsh structure and function under future SLR scenarios. First, overall declines in plant production may lead to less input of detritus for wetland food webs. Second, loss of productivity is expected to impair accretion processes that are necessary for marshes to keep pace with SLR. Shorter or less dense plant canopies may trap less suspended sediment carried during flooding tides, reducing mineral inputs to wetland soils (Voss et al. 2013). Declines in below-ground production may further slow accretion with loss of surface root growth (Nyman et al. 2006). Moreover, reduced root or rhizome production may weaken marsh banks leading to increased susceptibility to erosion (Deegan et al. 2012). If marshes cannot accrete at rates matching future SLR, then they must migrate laterally to persist (Stralberg et al. 2011). However, steep river valleys at the landward edges of marshes in the coast range of the Pacific Northwest and existing human infrastructure (roads and dikes) may limit this capacity.
Finally, spatial patterns of species biomass within individual estuaries may shift with altered gradients in inundation and salinity stress. SLR is expected to increase the relative abundance of more salt- or flood-tolerant species and may alter inter-specific interactions such as competition. The three species estimated in field surveys (Janousek and Folger, unpublished data) to have the greatest above-ground percent cover in emergent marshes in the Yaquina estuary—Deschampsia cespitosa (16 %), Agrostis stolonifera (13 %) and Juncus balticus (12 %)—each showed marked declines in root and shoot biomass with increased flooding and higher salinity. There was also a disproportionate loss of below-ground biomass in Deschampsia and Juncus (important mid and high marsh species respectively) with increasing stress (Figs. 4, 6). This suggests broad vulnerability of the tidal marsh flora to increased flooding and salt stress.
In this study we found reduced growth with higher submergence and salinity exposure for a suite of marsh species in the Pacific Northwest, including lower marsh species like Triglochin and more apparently salt-adapted species such as Plantago and Grindelia. We also found a disproportionate reduction in root productivity that may have consequences for the accretion potential of tidal marshes and the stability of marsh edges. SLR may increase both salinity exposure and submergence times in future salt marshes, and the additive or multiplicative effects of these stressors on plant growth may reduce rates of productivity across the entire marsh landscape.
We thank C. Folger, D. Fultz, M. Frazier, R. Loiselle, B. Weigel and R. Norton for research assistance and those who reviewed previous versions of the manuscript including C. Weilhoefer, M. Frazier and anonymous reviewers. The information in this publication has been funded by the U.S. Environmental Protection Agency. It has been subjected to review by the National Health and Environmental Effects Research Lab and approved for publication. Approval does not signify that the contents reflect the views of the Agency, nor does mention of trade names or commercial products constitute endorsement or recommendation for use.