The world's largest inland delta, located on the Okavango River in Botswana, faces major changes in the annual flooding size and duration due to climatic shifts and increased water use. We examined several parameters of a seasonal floodplain in the Okavango Delta during two years of contrasting flooding size. The small flood (2003) was characterized by high concentrations of total nutrients (2.5 mg N and 1 mg P L⁻¹), high primary production (0.8 mg C m⁻³ d⁻¹), and zooplankton biomass (30 mg DW L⁻¹). Methane production and consumption was considerable and stable isotope analysis suggested that methane oxidation provided a significant input of C to the aquatic food web. The large flood (2004) was characterized by lower volume-specific productivity, lower concentrations of nutrients (1 mg N and 10 µg P L⁻¹), lower primary production (45 mg C m⁻³ d⁻¹), reduced zooplankton biomass (10 µg DW L⁻¹), and low methane production. The density of fish (CPUE) was significantly higher for the large flood compared to the small one. The findings point to the overall importance of flooding size on primary and secondary production, as well as basic food web properties in the delta. Low floods mean higher volume-specific production at the base of the food web. Seasons of large and long lasting floods cause improved circulation and enhanced reproductive success for fish.

Population characteristics of Spartina argentinensis after fire were analyzed. Field experiments were done in temporary flooded tall grassland, dominated by S. argentinensis at the Reserva Federico Wildermuth (Argentina), on burnt and non-burnt plots. The following variables were analyzed: soil seed bank, potential and real establishment of seedlings and the effect of fire on them, percentage of tillers that continue growing after fire, emergence and survival of tillers, number of tillers that differentiated panicles, and production and predation of propagules. Plants resprouting after fire produced a larger number of tillers with greater tiller emergence, recruitment, and survival than tillers of those of the non-burnt plants. The proportion of tillers with panicles and number of full spikelets per plant was greater in burnt plants than in non-burnt plants, but individuals caryopses unitary weight as well as germination percentage and viability were not affected by fire. Caryopses were severely predated and they did not build up a permanent seed bank. In spite of the high output of caryopses, seedling establishment was almost negligible under most circumstances. Fire and environmental conditions provided very limited “windows of opportunity” for seedlings recruitment. Asexual reproduction was the main process of population maintenance and growth, and fire triggers off the output and eventual establishment of sexual propagules.

Created wetlands and water bodies that have wetland characteristics (old farm ponds) appear to provide many of the habitat attributes of natural systems. To compare the biological and physical characteristics of natural and created wetlands, we evaluated water chemistry and a suite of metrics associated with the plant, macroinvertebrate, and avian assemblages at 12 natural and six created systems in central Oklahoma. The natural wetlands had significantly shallower depths and higher turbidity levels than the created wetlands. Of 43 metrics across the three biotic assemblages, seven were significantly different between the two wetland types. The proportions of hemipteran insects from the family Corixidae and insectivorous bird species were both greater in natural than created wetlands. The proportion of perennial plant species, the proportion of invertebrates in the shredder feeding guild, the number of Ephemeroptera, Trichoptera, Sphaeridae, and dragonfly (ETSD) taxa, the proportion of individuals in the dominant bird taxa, and the proportion of avian edge species were all greater in created wetlands than in natural wetlands. The community similarity (based on Jaccard's similarity index) in the two wetland types was 38% for plants, 56% for macroinvertebrates, and 65% for birds. For some individual metrics, assemblage members had similar attributes (e.g., proportion of omnivorous taxa) although the specific taxa often differed between natural and created wetlands. These differences may influence the performance of certain assessment methods when they are applied to created wetlands.

In the northern prairie region of North America, there are millions of small seasonal wetlands. The aquatic ecology of these wetlands is partly controlled by the salinity of the wetland pond water, which affects the vegetation and invertebrate communities. The objective of this study was to identify the key geochemical processes affecting water chemistry in prairie wetlands. We used the combined water and solute mass balance approach to quantify the rates of geochemical reactions in a typical prairie recharge wetland in Saskatchewan, Canada. Sulfate reduction, carbonate mineral dissolution, and processes adding carbon dioxide to the pond were identified as the key geochemical reactions. Sulfate reduction removed more sulfate from the pond than infiltration in each of the four years examined. The average rate of sulfate reduction, 0.07 g m⁻² d⁻¹, was greatest in spring and decreased during the year. Reduced sulfate remains in the sediments but is re-oxidized when the pond dries out and is dissolved into the pond water and sediment pore water when the pond re-floods. X-ray diffraction analyses of wetland soil and mass balance calculations indicate magnesium-calcite is dissolved into the pond water in the spring and precipitates out of solution later in the year, and dissolves into the pond the following year.

I characterized 33 small water bodies from the southern Spain provinces of Seville, Cadiz, and Malaga using hydrochemistry (geochemical analyses), hydrology (water budgets), geological surveys, and local knowledge. Based on hydrogeological criteria (association with permeable materials), water bodies were grouped into three categories: 1) wetlands associated with permeable materials (playa-lakes), 2) wetlands associated with impermeable materials (ponds and pools), and 3) artificial wetlands. The three wetland typologies differed in morpho-structural indices, hydrogeological functioning, water chemistry, flora and fauna, and vulnerability to a range of human impacts, and these attributes need consideration in habitat management and conservation.

Mediterranean fluvial ecosystems are subjected to ever-increasing water demands and to a wide variety of other human impacts with potential negative effects on riparian vegetation. Nevertheless, few studies have compared the importance of human versus natural factors in structuring riparian vegetation in this region, particularly in semi-arid areas. We examined basin-scale responses of riparian vegetation to the main environmental gradient extracted by canonical correspondence analysis in two semi-arid Mediterranean rivers in southern Spain, considering different community attributes (woody, herbaceous, functional groups, and exotics species) and using different metrics (species composition, richness, and cover). The results show two main environmental gradients for riparian vegetation: a prime altitudinal gradient of increasing salinity towards the lower sites, and a second gradient of increasing water pollution and human physical impact parallel to the hydroperiod shortening. Species richness of woody and herbaceous life forms, and cover of woody species, reacted negatively to the increasing salinity downstream, a pattern that was driven by most functional groups. This is consistent with the key role of salinity for vegetation. Furthermore, as for other semi-arid riparian ecosystems, the decline in rainfall coupled with higher temperatures and evaporative stress at the lower elevation open-canopy sites may have constrained herbaceous species. Woody richness and cover declined with shorter hydroperiod and increasing human impacts, a response that was driven by mesic and hydric species. These results agree with the idea that surface flow permanence is an outstanding limiting factor for woody vegetation in semi-arid rivers, possibly exerting more influence on species richness than flood disturbance. However, herbs showed very little response to drought or human impacts compared to woody species, which may be related to the higher recovery rate from disturbance of herbaceous life forms, due to their shorter life-span and higher colonization rates. Generally, exotic species were positively affected by dissolved nutrients (N and P), especially exotic herbaceous richness and cover by nitrogen. As in other Mediterranean rivers, percentage richness of exotics in our rivers was substantially lower compared to fluvial ecosystems in other regions.

Wetlands have an important role in ecosystem function and biodiversity. Effective management of wetlands requires accurate and comprehensive spatial information on location, size, classification, and connectivity in the landscape. Using a GIS, two provincial wetland maps were compared with regard to their areal correspondence across different ecoregions of New Brunswick. The first consisted of discrete wetland units (vector data) derived from aerial photo interpretation. The second consisted of wet areas modeled by a newly developed depth-to-water index with continuous coverage across the landscape (raster data). This index was derived from a digital elevation model and hydrographic data. The relative advantages and disadvantages of the two approaches were assessed. The two maps were generally consistent with most discrete wetland areas (51%–67 %) embedded in the 0–10 cm depth-to-water class, verifying the continuous modeling approach. The continuous model identified a larger wetland area. Much of this additional area consisted of riparian zones and numerous small wetlands (< 1 ha) that were not captured by aerial photo interpretation. Unlike the discrete map, the continuous model showed the hydrological connectivity of wetlands across the landscape. Both approaches revealed that topography was a major control on wetland distribution between ecoregions, with more wetland in ecoregions with flatter topography.

Reasons for apparent declines in populations of white-winged scoters (Melanitta fusca) in the northern boreal forest are not well understood, but some evidence suggests factors associated with the breeding grounds may be responsible. Climate warming or increased forest fire frequency could adversely affect upland or wetland breeding habitats or key food sources for breeding females or ducklings, which in turn may lower productivity. However, very little is known about wetland habitat preferences of scoters. Determining what habitat features scoters need to breed successfully and whether changes in boreal forest breeding habitat affect scoter productivity are important steps towards understanding their ecology and developing appropriate conservation initiatives. Thus, our overall goal was to characterize features of wetlands used by scoter pairs and broods. Additionally, we compared features of wetlands surrounded by recently burned versus unburned forest to investigate whether fire-induced changes in wetland productivity, water chemistry, or amphipod abundance could affect patterns of scoter habitat use. Scoter pairs and broods used wetlands with more abundant amphipods, a finding that is consistent with other waterfowl studies. However, unlike some previous waterfowl studies, we did not find consistent positive correlations between total phosphorus levels and amphipod abundance or wetland use by scoters. We did not detect differences in our measured water chemistry variables, indices of wetland productivity, or amphipod abundance between wetlands surrounded by burned versus unburned forest.

Wetlands in the floodplains of flood-control impoundments are exposed to different hydrologic conditions than those in natural systems. To gain insight into the impact of hydrology on wetland communities, we examined associations between crustacean zooplankton community structure and hydrologic and local environmental characteristics in 29 ponds surrounding a flood-control reservoir. Fifty-eight species of zooplankton were detected among all ponds, and local species richness ranged from 3–16. Depth and distance from the lake were the most important predictors of species richness and provided 31% of the explanation for variation among ponds. Non-metric multidimensional scaling (NMDS) provided visualization of factors important to community structure. Analysis of similarities revealed hydroperiod and flood frequency were significant factors structuring communities. These results suggest that hydrology has a significant effect on both species richness and community structure of zooplankton communities in floodplain ponds. Elucidating the importance of hydrology to wetlands residing in managed floodplains should assist reservoir managers in understanding the impacts of different management regimes on regional ecology.

Community and environmental gradients within the ecological boundaries of Carolina bay wetlands may provide important information on the interaction between Carolina bays and associated uplands, and may also provide guidance for improved management. We established twelve 30-m transects on the sloping rims of each of six Carolina bays in northeastern South Carolina to characterize the community gradient, as well as important environmental factors producing this gradient. Mid-points of the transects were placed on jurisdictional wetland boundaries. Hydrology, soil properties, and plant species composition were measured within these transects. On average, transects included an elevation change of 0.6 m that corresponded with gradients of hydrology, soil properties, and community characteristics. Decreasing surface soil moisture (i.e., fewer flood events) and decreasing soil nutrients were associated with a shift from shrub-bog vegetation with relatively low alpha diversity and prominence of evergreens to a relatively diverse and heterogeneous community characterized by grasses, herbs, low shrubs, and vines. Ecotones, identified by abrupt changes in community composition, were more frequently found outside jurisdictional wetland boundaries. Likewise, five near-endemic and endemic plant species were found outside the wetland boundaries. Our data reinforce the need for better understanding of how Carolina bays interact with adjacent landscape elements, and specifically how ecological boundaries are influenced by this interaction.

We report the results of a detailed 12-month study of 23 freshwater wetlands and one larger synoptic characterization of 55 freshwater wetlands to test whether a hydrogeomorphic (HGM) classification of the wetlands into lotic (attached to streams) and terrene (groundwater fed) classes meaningfully discriminated wetland surface water chemical composition in the mountainous Catskill-Delaware watersheds of southeastern New York State. Most of these hillslope wetlands are underlain by thin, largely siliceous mineral soils and have minimal peat cover. Nonparametric one-way ANOVA (Kruskal-Wallis) tests based on measurements of SC, Ca² , Mg² , Na , DOC, TDN, TDS, Si, SO₄²⁻, pH, DO, K , Cl⁻, NH₄ , NO₃⁻, TDP, and HCO₃⁻ failed to reject the null hypothesis that the surface water chemistry of lotic and terrene wetlands was identical. Results of the statistical tests showed that the only significantly different chemical species in surface waters from the two HGM landscape classifications were SC, Na , and Cl⁻, which was clearly related to individual wetland proximity to road salt additions. Isotopic analyses of ²H and ¹⁸O for 30 synoptic wetland surface waters also failed to demonstrate significant differences for any of the HGM wetland classes. Based on the results of these data, we caution that landscape position, landform, water flow path, and water body type may not be accurate in making wetland classifications for HGM assessment in all locations. Underlying geology should be considered before making assumptions that water chemistry will differ by landscape position, and wetland functions dependent on water chemistry should be evaluated accordingly.

Lake Michigan drowned river-mouth wetlands have a unique geomorphology and hydrology. Macroinvertebrate communities in these systems respond to multiple biotic and abiotic factors that are not well understood. In June and August 2003, we sampled macroinvertebrate communities at 22 sites in four Lake Michigan drowned river-mouth wetlands. Sites were distributed along gradients of anthropogenic disturbance, vegetation, and sediment types. The relative influences of anthropogenic disturbance, vegetation, and sediment type on macroinvertebrate community composition were determined using non-metric multidimensional scaling (NMDS) and multi-response permutation procedures (MRPP). The depth of organic deposits best explained the gradients revealed with NMDS and MRPP for both sampling dates. The MRPP did not detect differences in community composition among vegetation types and wetlands with different levels of anthropogenic disturbance. These results suggest that 1) macroinvertebrate community structure in Great Lakes drowned river-mouth habitats is influenced substantially by sediment characteristics, and 2) anthropogenic practices that affect the deposition of organic sediment in coastal wetlands (e.g., eutrophication and hydrologic manipulation) will likely affect macroinvertebrate community structure.

Australia's large desert floodplains are among the world's most hydrologically variable wetlands and vegetation in these habitats changes dramatically over time in response to flooding and drought. Annual forb and grass communities in these desert floodplains rely on large, diverse soil seed banks as critical sources of propagules for recruitment. I investigated the effects of flooding on seedling emergence from the soil seed bank of the Cooper Creek floodplain in arid central Australia. My objective was to determine the effects of differences in both short-term flooding characteristics as well as the longer-term flood history of sediments on the composition of plant communities establishing from the soil seed bank. I conducted a greenhouse experiment in which sediments collected from high (inundated at least once every 1–5 yrs), medium (inundated approximately once in every 5–10 yrs), and low (inundated less than once a decade) flood frequency zones were subjected to different durations of submergence, rates of drawdown (i.e., duration of post-submergence waterlogging), and seasonal timing of flooding. The abundance of emerging seedlings more than doubled in response to longer overall durations of wetting, but the relative duration of submergence (i.e., 8 vs. 4 weeks) versus subsequent soil waterlogging (i.e., 4 vs. 8 weeks) within this period did not have a significant effect. The duration of submergence versus waterlogging did however influence the composition of emerging seedlings with almost twice as many annual monocots appearing following longer durations of submergence. All dominant species emerged frequently following both summer and winter flooding, suggesting opportunistic germination strategies. The composition of moderately abundant and rare species, however, differed in response to the seasonal timing of inundation with monocots more common following summer floods and forbs following winter flooding. Common species were well-distributed in the soil seed bank across the flood history gradient, and the species richness of emerging seedlings did not vary between different flood frequency zones. However, seedlings of species emerging from low flood frequency zone samples tended to be less abundant and differed in composition compared with those emerging from samples of high and medium flood frequency zones. My results suggest that changes to both short- and long-term flooding patterns, through flow management or climate change, are likely to affect vegetation responses to inundation in these desert floodplains. Reductions in flood pulse magnitude and frequency, for instance, could result in a decline in the abundance of valuable pasture grasses germinating in response to summer flooding and a loss of hydrophytic species in rarely flooded areas.

Salt marshes along the northeastern coast of the United States are increasingly subject to changes in hydrology and enrichment with nitrogen as a result of human activities. We conducted a greenhouse experiment to determine the response of Phragmites australis, Spartina alterniflora, and their root-associated microbial communities to these environmental perturbations. Two sets of treatments were compared: 1) saturated versus drained hydrology under low N enrichment and 2) low versus high N enrichment under drained hydrologic conditions. Unvegetated sediments were planted with either Phragmites australis or Spartina alterniflora, and after one growing season, sediment characteristics, macrophyte biomass, and sediment microbial community structure, as described by phospholipid fatty acids (PLFAs), were analyzed. Under all conditions tested, Spartina root production was significantly greater than Phragmites. While Spartina invested more biomass in roots, Phragmites invested proportionally more biomass in shoots and rhizomes, and Phragmites response to drained hydrology or to an increase in N also differed from that of Spartina. Under N enrichment, the rate of Phragmites stem production doubled, and under drained conditions the ratio of Phragmites shoot:root biomass increased, while Spartina biomass ratios remained unchanged. Although Spartina root biomass was significantly greater than that of Phragmites, under drained conditions the Spartina sediment PLFA diversity was significantly lower than the PLFA diversity in both Phragmites and unvegetated sediments. Under saturated conditions, vegetated sediments exhibited greater PLFA diversity, while no diversity differences were seen in unvegetated sediments under the two hydrologic conditions. Six PLFAs were responsible for 80% of the separation seen within the Principal Components Analysis ordination space. Significant differences in these PLFAs were due to hydrology when comparing saturated vs. drained sediments, and predominantly due to the plant species when comparing N treatments under drained conditions. Our results suggest that macrophyte root association can influence the structure of estuarine sediment microbial communities, but that saturated hydrological conditions may override the plant influences.

Gunnera perpensa L., a wetland herb, is extensively harvested and used as traditional medicine. In many areas of KwaZulu-Natal, South Africa, populations are under threat of overexploitation from the herb. Successful cultivation would reduce harvesting of natural populations and could help conserve the species. To provide information for cultivation, we assessed the effects of various flooding regimes on plant growth and biomass accumulation, as well as on morphological and physiological adaptations of this species to waterlogging stress. Plant rhizomes collected from the field were cultivated for three weeks, and then subjected to drained, 20%, 40%, and 80% flooded treatments for 12 weeks. Plant height, total dry biomass accumulation, and carbon dioxide exchange increased significantly with increase in level of flooding from the drained to the 40% flooded treatments, and thereafter decreased as level of flooding increased to 80%. Soil E_h and root specific gravity decreased with increases in flooding. In rhizotron experiments, in which plants were subjected to 40%, 80%, and 100% flooding, root growth as well as biomass accumulation decreased as flooding increased. Photosynthetic performance, monitored along a natural soil moisture gradient in the field, indicated that carbon dioxide exchange, quantum yield of photosystem II (PS II), and electron transport rate (ETR) through PS II were higher at a soil moisture of 45% than at 25% or 60%. This study demonstrated that G. perpensa grows best in moderately reduced soils that are moist but not waterlogged. Adaptations to waterlogging include increased aerenchyma in roots and petioles, and decreased below ground biomass accumulation.

Created wetlands are often limited in soil organic matter, a product that usually accumulates with long-term ecosystem succession. Although many studies have tested the effect of adding organic material to these systems, few have quantified the effect of various loadings of organic matter (OM) in created wetlands. The purpose of this study was to determine how vegetation composition, standing crop biomass, and woody vegetation development varied in a created freshwater wetland with respect to different loadings (0, 56, 112, 224, or 336 Mg ha⁻¹) of a soil OM amendment. Soil C, N, and P were positively related to loading rate, as was soil surface elevation. Species richness, evenness, and diversity measurements, along with the Ellenberg Community Coefficient Similarity Index, suggested an overall similarity of plant assemblages regardless of loading rate. Standing crop biomass (580–790 g m⁻²) was not significantly correlated with OM loadings, but showed a significant curvilinear relationship with plot surface elevation. Woody vegetation development was correlated with OM loadings, plot elevation, and soil P, indicating a positive relationship with all three factors. An amendment loading of 112 Mg ha⁻¹ provided the maximum benefit because it provided soil nutrient levels that were within the range of natural wetlands while also minimizing changes in soil surface elevation due to the added bulk material.

Soil processes often exhibit spatial heterogeneity that may influence plant community structure. This study was conducted to determine fine-scale spatial patterns and degrees of variability of soil nutrients and plant communities within different vegetation types in a stream floodplain in southwestern West Virginia. One 5 m × 5 m site was established in each of two vegetation/drainage types: pasture (PA) and seasonal wetland (SW). Sites were located ∼25 m apart on flat bottomland. A 10 m × 1 m transect was also established perpendicular to the visible boundary between PA and SW drainage types, but separate from the PA and SW sites. Each site was divided into 1-m² plots (n  =  25) and the transect was divided into 0.25-m² plots (n  =  40). Mineral soil was taken to a 5 cm depth. Soil organic matter was measured as loss-on-ignition. Extractable NH₄ and NO₃⁻ were determined before and after laboratory incubation (28 days at 27°C) to determine net N mineralization and nitrification. Cations were analyzed using inductively coupled plasma emission spectrometry. Vegetation was assessed using estimated percent cover (sites) and aboveground harvested biomass (transect). Mean organic matter was significantly higher (P < 0.05) in SW than in PA (10.6% and 8.3%, respectively). Nitrification was nearly 100% of mineralization in all soils, and was significantly lower (P < 0.05) in PA than in SW (0.7 and 1.8 µg NO₃⁻-N/g soil/d, respectively). Aluminum was significantly higher (P < 0.05) in SW than in PA (202.1 and 0.5 µg Al/g soil, respectively). Calcium and pH were significantly higher (P < 0.05) in PA than in SW (768.5 µg/g soil and 4.4, respectively). Magnesium was significantly higher (P < 0.05) in SW than in PA (174.2 and 121.9 µg/g soil, respectively). Transect results were similar to PA and SW sites and an abrupt transition was found between PA and SW site types. Vegetation analysis revealed two distinct communities with SW dominated by wetland species and PA dominated by a mixture with slightly more upland species. Transect vegetation also consisted of largely wetland species within wetland and a mixture in pasture, however Arthraxon hispidus (Thunb.) Makino dominated for ∼1 m at the boundary. Spatial variability of organic matter was much lower than spatial variability of nitrification. Thus, availability of organic substrates to N-processing microbes is less variable than N processing itself, underlining the complexity of biotic factors responsible for regulating soil N processes.

Impacts of co-occurring biotic and abiotic environmental variables on terrestrial mollusc communities have rarely been studied. In wetlands, terrestrial molluscs can drown or become stranded by rising water levels and they are often prey for predaceous beetles and small mammals. I used coverboard traps over a three year period to study the biotic and abiotic gradients that may structure a wetland prairie terrestrial mollusc community in western Oregon, USA. A non-metric multidimensional scaling (NMDS) ordination indicated that shrews (Sorex vagrans) and predaceous beetles were the primary gradients associated with the terrestrial mollusc community. Snails were associated with the presence of shrews, slugs were allied with predaceous beetles, and the vectors representing the shrew and predaceous beetle gradients were opposed, suggesting that beetles preyed upon snails while the shrews consumed both beetles and slugs. These hypothesized predator-prey relationships among molluscs, shrews, and beetles coincide with studies of Sorex gut contents and predaceous carabid beetle life history. The mollusc community was not associated with the water cover gradient in the NMDS ordination, but small mammal burrows and predaceous beetles were found in greater abundance in unflooded mound habitats. This suggests that standing water may temporally concentrate both predators and prey in a shared unflooded environment during the rainy season. The high metabolic demands of shrews and an apparent preference for both beetle and molluscan prey suggest that shrews may exert a strong top-down effect on invertebrate communities.

Drained depressional wetlands are typically restored by plugging ditches or breaking drainage tiles to allow recovery of natural ponding regimes, while relying on passive recolonization from seed banks and dispersal to establish emergent vegetation. However, in restored depressions of the southeastern United States Coastal Plain, certain characteristic rhizomatous graminoid species may not recolonize because they are dispersal-limited and uncommon or absent in the seed banks of disturbed sites. We tested whether selectively planting such wetland dominants could facilitate restoration by accelerating vegetative cover development and suppressing non-wetland species. In an operational-scale project in a South Carolina forested landscape, drained depressional wetlands were restored in early 2001 by completely removing woody vegetation and plugging surface ditches. After forest removal, tillers of two rhizomatous wetland grasses (Panicum hemitomon, Leersia hexandra) were transplanted into single-species blocks in 12 restored depressions that otherwise were revegetating passively. Presence and cover of all plant species appearing in planted plots and unplanted control plots were recorded annually. We analyzed vegetation composition after two and four years, during a severe drought (2002) and after hydrologic recovery (2004). Most grass plantings established successfully, attaining 15%–85% cover in two years. Planted plots had fewer total species and fewer wetland species compared to control plots, but differences were small. Planted plots achieved greater total vegetative cover during the drought and greater combined cover of wetland species in both years. By 2004, planted grasses appeared to reduce cover of non-wetland species in some cases, but wetter hydrologic conditions contributed more strongly to suppression of non-wetland species. Because these two grasses typically form a dominant cover matrix in herbaceous depressions, our results indicated that planting selected species could supplement passive restoration by promoting a vegetative structure closer to that of natural wetlands.

On raised bogs, the distribution of Sphagnum species is determined by their distance to the water table, but occasionally species are able to survive outside their niche. Hollow species that persist in hummock vegetation are assumed to profit from the higher water content of the surrounding hummock species, although the mechanism responsible is unclear. In this study, we elucidated the role of lateral hummock water transport (LHWT) and precipitation on the water content of hollow species occurring in hummocks. This was tested using a full factorial field transplantation experiment with cores of Sphagnum cuspidatum in a high and a low hummock. Treatments included direct precipitation (present or absent) and LHWT (present or absent). Fresh weights of the cores were measured at regular time intervals. Our results show a relatively large effect of precipitation on the water content in both the high and low hummock, whereas LHWT only seemed to be an important source of water in the high hummock, which was relatively dry. Furthermore, LHWT played an important role only after large precipitation events, suggesting that lateral water transport is indirectly affected by rain. This study shows that precipitation alone can explain the persistence of hollow species in high hummocks, whereas it was less important for hollow species in low hummocks. Our data suggest that the survival and potential expansion of hollow species in higher hummocks strongly depends on the intensity and frequency of rain events. Changes in precipitation patterns may result in a loss of Sphagnum diversity in hummocks.

California black rails (Laterallus jamaicensis coturniculus) occur in two disjunct regions: the southwestern USA (western Arizona and southern California) and northern California (Sacramento Valley and the San Francisco Bay area). We examined current status of black rails in the southwestern USA by repeating survey efforts first conducted in 1973–1974 and again in 1989, and also examined wetland plant species associated with black rail distribution and abundance. We detected 136 black rails in Arizona and southern California. Black rail numbers detected during past survey efforts were much higher than the numbers detected during our more intensive survey effort, and hence, populations have obviously declined. Plants that were more common at points with black rails included common threesquare (Schoenoplectus pungens), arrowweed (Pluchea sericea), Fremont cottonwood (Populus fremontii), seepwillow (Baccharis salicifolia), and mixed shrubs, with common threesquare showing the strongest association with black rail presence. Plant species and non-vegetative communities that were less common at points with black rails included California bulrush (Schoenoplectus californicus), southern cattail (Typha domingensis), upland vegetation, and open water. Black rails were often present at sites that had some saltcedar (Tamarix ramosissima), but were rarely detected in areas dominated by saltcedar. We recommend that a standardized black rail survey effort be repeated annually to obtain estimates of black rail population trends. Management of existing emergent marshes with black rails is needed to maintain stands of common threesquare in early successional stages. Moreover, wetland restoration efforts that produce diverse wetland vegetation including common threesquare should be implemented to ensure that black rail populations persist in the southwestern USA.

Since the late 1980s, the United States Fish and Wildlife Service (USFWS) has helped restore hundreds of wetlands in Manitowoc County, Wisconsin, in an effort to enhance the production of waterfowl and other wetland associated species. During summer 2004, we re-examined 11 restorations to determine how their plant communities had changed since 1992. In addition, we re-evaluated waterfowl and anuran communities at eight restorations. Because 1992 data were compiled from separate studies, all three groups of organisms were simultaneously examined at only two wetlands. Significant decreases in plant diversity and plant species richness were detected between 1992 and 2004, while significant increases were detected for species classified as obligate or facultative wet. These changes suggest that the plant communities at the restorations have matured since 1992. Use of the wetlands by waterfowl and anurans, on the other hand, did not exhibit significant change over this period. Regular monitoring of restorations over even longer periods will provide new insights into the way in which restored communities develop and whether current restoration methods have succeeded in establishing stable, species-rich wetland communities.

Coastal restoration strategies in the Mississippi River Delta Plain include use of direct sediment additions as well as river diversions to reintroduce fresh water and sediments into coastal marshes. Often, these diversions are located in the upper parts of the coastal basins, where fresh water floating marshes predominate. Floating marshes are wetlands of emergent vascular vegetation that have a significant mat of live and dead roots, dead organic material, and mineral sediments. Mats move vertically as ambient water levels rise and fall. As such, extensive water exchange is maintained below the mat, while overland sheet flow is reduced or eliminated, limiting inorganic sediment input. Thus, the effect of sediment introduction into floating marsh wetlands is unknown. In this study, we documented marsh mat response to Mississippi River sediment addition and measured change in soil properties, species composition change, and growth of vegetation. The study included two sites located in seasonally floating thin-mat marshes dominated by spikerush (Eleocharis baldwinii). At each site, 16 plots were established. Each 1 m² plot was randomly assigned one of four treatments: none, low, medium, and high sediment additions, with the highest sediment addition representing approximately 15 years of deposition from an existing river diversion at Caernarvon, Louisiana. With increasing sediment addition, significant increases in bulk density, and significant decreases in percent organic matter in the top 25 cm of the marsh mat were observed. Most of the sediment added remained in the top 10 cm of the marsh mat. No significant differences in water level over the mat were found among treatments, indicating that buoyancy was not affected by sediment addition. Aboveground biomass tended to increase with sediment addition, although the differences were not statistically significant. Belowground biomass was not significantly affected by the addition of sediment. These results indicate that sediment additions at the level supplied by existing diversion structures should not sink floating marsh mats and may have slight positive effects on vegetation biomass.

Many wetland restoration projects are initiated with phosphorus (P) retention as a primary objective. While undisturbed wetlands often are net sinks for P and other nutrients, there is evidence that newly flooded restoration wetlands on former agricultural land initially release P to surface waters. The objectives of this study were to: 1) measure P release from soils to overlying surface waters that would occur when re-flooding agricultural fields to restore a lake fringe wetland connected to Upper Klamath Lake, Oregon; and 2) identify management strategies to abate nutrient release from soils during restoration to minimize P loading to Upper Klamath Lake. We simulated the process of re-flooding soils using mesocosms in a laboratory experiment. The soils were flooded with lake water, and the water was replenished on a weekly basis. The net P flux from soils to surface water was estimated by measuring differences in P concentrations between water that had been in the mesocosms and the lake water used for replenishment. After the flooding experiment, we measured the concentrations of four forms of soil P using a modification of the Hedley procedure, to examine relationships between soil P chemistry and P release. The majority of P was released in the first two days of the experiment, and all detectable P was released by the end of the second month. We estimated that 1–9 g P/m² were released from the soils to the water column over the course of the experiment, which amounted to 1%–16% of total soil P. Scaling up to the entire wetland, this totals approximately 64 tons P released over 3,000 ha. We did not find any statistically significant relationships between any of the four forms of soil P and the amount of P released in the flooding experiment. Even though we demonstrate here that P is released while undertaking wetland restoration projects on former agricultural land, it is likely to be a temporary process, and once the wetland begins to resume more natural hydrological and biogeochemical functions and vegetation structure, it will re-start the process of soil accretion and P sequestration.

Western Boreal Plain peatlands can play an important role in the global nitrogen cycle by storing N in peat and potentially releasing large amounts of N to the atmosphere. In this study, biological denitrification rates were measured in marsh and fen vegetation zones in two boreal peatland-pond complexes in northcentral Alberta, Canada. Assuming negligible winter denitrification, we estimated annual denitrification rates of 11 g N·m⁻² in marshes and 24 g N·m⁻² in fens. Two techniques were employed to measure denitrification: 1) measurements of direct N₂-flux were taken from intact cores in gas-tight N-free chambers, and 2) nitrous oxide (N₂O) flux was measured in the two fens using in situ chambers. N₂ fluxes ranged from 2.14–4.19 mg N·m⁻²·h⁻¹ in marshes and 6.19–6.81 mg N·m⁻²·h⁻¹ in fens. N₂O release from fen peat ranged from consumption to 0.025 mg N·m⁻²·h⁻¹. Peat with higher carbon and moisture content was a source of N₂O whereas peat with lower carbon and moisture content was a sink. Surface water did not appear to be a major source of nitrate for denitrification. However, denitrification rates were positively correlated with peat extractable nitrate. Combined with mineralization studies, this indicated that soil nitrification provided most of the substrate for denitrification.

Soil phosphorus (P) enrichment from external inputs can result in considerable changes in wetland ecosystem structure. What is not known is whether many of the microbial physiological measures that are effective at determining ongoing impact are equally sensitive to reductions in the soil P content. The study was conducted over a two year period (1999–2000) in two areas located in Blue Cypress Marsh Conservation Area (BCMCA), an area (Enriched) with historically elevated soil P (1,544 mg kg⁻¹ in 1995, 877 mg kg⁻¹ in this study) and a reference area (Reference) with background soil P contents (698 mg kg⁻¹). Nutrient loading to this wetland was terminated in 1994. Microbial ecophysiology measures were obtained quarterly and consisted of soil microbial biomass carbon (MBC) content, β-glucosidase and acid phosphatase, and end products of anaerobic microbial metabolism (CO₂ and CH₄). All measures exhibited significant temporal variation with higher values for MBC content, enzyme activities, and respiration rates during summer months (June and September) and lower in the winter months (December and March). We found no significant differences between site mean MBC (Enriched: 7.24, Reference: 8.22 mg⁻¹ kg⁻¹), CH₄ production (Enriched: 4.41, Reference: 4.73 µmol CH₄ gr⁻¹ d⁻¹), or β-glucosidase activity (Enriched: 56.14, Reference: 57.70 µg MUF gr⁻¹ h⁻¹). The site mean acid phosphatase (Enriched: 56.92, Reference: 78.74 µg MUF gr⁻¹ h⁻¹) and CO₂ production rates (Enriched: 11.00, Reference: 13.69 µmol CO₂ gr⁻¹ d⁻¹) were found to be significantly different. Microbial communities at the two sites were different in terms of their metabolic activities, but not in terms of C-pathways. We also found that enzyme profiles at the enriched site did not change appreciably over the two year period. The results obtained in this two year study suggest that most microbial community ecophysiology measures were not responsive to decreasing concentrations of P. However, at both sites, β-glucosidase and anaerobic microbial activities were higher in the second year (41.80 vs. 68.27 µg MUF gr⁻¹ h⁻¹ and 10.34 vs.13.85 µmol CO₂ gr⁻¹ d⁻¹) and acid phosphatase activities lower (72.54 vs. 63.11 µg MUF gr⁻¹ h⁻¹). A drawdown that took place in the winter months of late 1999 and early 2000 might have released labile soil components, resulting in increases in overall metabolic activities and repression of the acid phosphatase activities. This has management consequences as P from the enriched areas can be remobilized and move further downstream in surface water.

Only a few studies have investigated foliar air-surface exchange associated with wetland plants and none have investigated this exchange in an experimental setting as a function of different soil and water mercury (Hg) exposure concentrations or monitored foliar Hg concentrations. In this study, foliar total Hg (THg) and methyl Hg (MeHg) concentrations and foliar Hg flux were investigated using Typha latifolia growing within controlled mesocosms. Exposure scenarios included combinations of two soil exposure Hg concentrations (0.03–0.1 and 0.38–0.44 µg g⁻¹), and two water exposure Hg concentrations (4–8 and 40–140 ng L⁻¹). Soil and water Hg concentrations were not correlated with foliar total Hg concentrations or foliar Hg flux. Foliar Hg fluxes measured with a gas exchange chamber were low, and atmospheric deposition to foliage was the dominant flux for all exposures, except for those plants growing in the low Hg in water and soil scenario. Based on data developed, it is suggested that Hg concentrations in foliage of Typha latifolia growing in media contaminated with Hg from historic mine waste were influenced primarily by assimilation of Hg from the atmosphere and not contaminated by water or sediment. In contrast, foliar MeHg concentrations followed a temporal pattern that was similar to observed changes in water MeHg concentrations. This indicated that MeHg in foliage could have been derived from the rooting media and was assimilated by different processes than THg in leaves.

We examined trends in the movement and source of water in a headwater wetland in North Shropshire, UK. Six piezometer nests along two transects were monitored over an 18 month period, and flownets were derived to estimate the rate and direction of water movement through the wetland and the interaction between precipitation and groundwater discharge. Individual water sources are identified using stable isotopes and seasonal differences in the composition of wetland soil-water are described. Variations in dissolved inorganic and organic carbon (DIC and DOC) were measured in water samples collected from discrete points in the wetland and the adjacent river and were interpreted using the hydrological data. The results suggest that end-members for DOC and DIC can be identified across the range of sampling sites: a groundwater spring (GS) had the lowest DOC and high DIC (DOC  =  5.6 ± 4.5 mg/l; DIC  =  36.7 ± 4.6 mg/l); a shallow well (WS) had the highest DOC and DIC (DOC  =  32.5 ± 18.7 mg/l; DIC  =  61.9 ± 18.9 mg/l); while surface-water (WSW) had the lowest DIC (20.6 ± 12.1 mg/l). Water fluxes between the wetland and river are estimated using the Dupuit-Forcheimer approximation to highlight the degree to which some headwater wetlands may act as a carbon source to ecosystems downstream. These wetlands are potentially a significant pool of C but are particularly sensitive to future changes in groundwater levels.

Created wetlands are increasingly used to mitigate wetland loss. Thus, identifying wetland creation methods that enhance ecosystem development might increase the likelihood of mitigation success. Noting that the microtopographic variation found in natural wetland settings may not commonly be found in created wetlands, this study explores relationships between induced microtopography, hydrology, and plant species richness/diversity in non-tidal freshwater wetlands, comparing results from two created wetland complexes with those from a mature reference wetland complex in northern Virginia. Elevation, steel rod oxidation depth, and species cover were measured along replicate multiscale (0.5 m-, 1 m-, 2 m-, and 4 m-diameter) tangentially conjoined circular transects in each wetland. Microtopography was surveyed using a total station and results used to derive three roughness indices: tortuosity, limiting slope, and limiting elevation difference. Steel rod oxidation depth was used to estimate water table depth, with data collected four times during the growing season for each study site. Plant species cover was estimated visually in 0.2 m² plots surveyed at peak growth and used to assess species richness, diversity, and wetland prevalence index. Differences in each attribute were examined among disked and non-disked created wetlands and compared to a natural wetland as a reference. Disked and non-disked created wetlands differed in microtopography, both in terms of limiting elevation difference and tortuosity. However, both were within the range of microtopography encompassed by natural wetlands. Disked wetlands supported higher plant diversity and species richness than either natural or non-disked wetlands, as well as greater within-site species assemblage variability than non-disked wetlands. Irrespective of creation method, plant diversity in created wetlands was correlated with tortuosity and limiting elevation difference, similar to correlations observed for natural wetlands. Vegetation was more hydrophytic at disked sites than at non-disked sites, and of equivalent wetland indicator status to natural sites, even though all sites appeared comparable in terms of hydrology. Results suggest that disking may enhance vegetation community development, thus better supporting the goals of wetland mitigation.

Phosphorus (P) is frequently the limiting nutrient in aquatic ecosystems, so wetland P attenuation is of particular landscape importance. Providing reliable knowledge about the capacity of wetlands to provide P sequestration is limited by knowledge of soil sorption capacities. We examined P-sorption in wetland soils using samples (n  =  326) collected from 171 wetlands across three southeastern ecoregions, stratifying by land use intensity (reference vs. impacted), vegetation (forested vs. herbaceous), and hydrologic setting (riverine vs. non-riverine). Single-point isotherm values ranged from −73 to 990 mg P kg⁻¹ (mean  =  462.7 ± 295.2 mg P kg⁻¹). Using a mixed-effects ANOVA, no significant P-sorption differences were observed for vegetation or condition, and only a moderate effect of hydrology (p  =  0.01). We observed a strong ecoregion effect (Regions IX ≈ XIV ; XII; p < 0.001) and a strong interaction between ecoregion and condition (p < 0.001). Site-level and within-site random effects were both significant (p < 0.001 and p  =  0.04, respectively), though the latter were small. The overall model explained only 29% of observed variance, suggesting limited generality for prediction. Given increased sample density requirements for assessment of P-sorption, pedotransfer functions (PTF), which estimate hard-to-measure properties from more readily observable properties, may be useful. We developed and validated two PTF models by relating observed sorption with 1) biogeochemical P-sorption covariates (total P/C, water extractable P, oxalate extractable Fe/Al/P/Ca/Mg) and 2) visible/near-infrared (VNIR) diffuse reflectance spectra. Advantages of VNIR for predicting soil properties include low cost, high sample throughput, minimal sample preparation and reagent waste, and high analytical precision. Standard error of prediction (SEP), r², and relative performance determinant (RPD) values were compared between PTF models for hold-out validation data. Models were of comparable utility; with SEP values of 144.1 vs. 157.2 mg P kg⁻¹, r² values of 0.61 vs. 0.69, and RPD values of 1.61 vs. 1.88 for the biogeochemical vs. VNIR models, respectively. Given other advantages of using VNIR spectra, it appears to be a useful tool for mapping and monitoring P sorption in wetland soils.

Wetland restoration projects attempt to recreate the hydrology found in natural wetlands, but little is known of the water budgets associated with wetlands in their natural state. The objective of this study was to compute the water budgets of three natural Carolina bay wetlands in Bladen County, North Carolina, USA. DRAINMOD models of various locations in the bays were calibrated with measured water table depths over a 2-yr period using inputs of rainfall, air temperature, and soil physical properties. The models were successful in simulating water table depths at all well locations during the calibration period with average absolute deviations between simulated and measured water table depths of approximately 4 cm. Measured and simulated data revealed very shallow (< 0.1 m) water table depths at all of the bays. Groundwater inflow was a significant component of the water balance at locations near the perimeters of the bays, ranging from 3%–26% of the total water input for these sites during the study period. A semi-confined aquifer below one of the bays was likely the source of groundwater inflow for that bay. Meanwhile, locations near the centers of the bays did not have groundwater inflow as an input to their water budgets. Groundwater outflow for the centers of the bays ranged from 2%–21% of rainfall. Areas near the perimeters of the bays were recharge, discharge, or flow-through wetlands depending on hydrologic conditions at the sites. Areas near the centers of the bays exhibited characteristics of recharge wetlands only. These results were consistent across the three Carolina bays studied, and can be used to better understand the hydrology of natural Carolina bays, improving the success of restoration projects of similar sites.

Recent studies found substantial variability in plant community integrity of wetlands in the Prairie Pothole Region (PPR) of central North Dakota, USA. We speculated that this variability might be connected to the nature of the surrounding landscapes and that a link might exist between landscape spatial metrics and wetland condition. We explored this potential link, using a case study in the PPR. A combination of remote sensing, geographic information systems (GIS), and landscape spatial metrics was used to: 1) examine the condition-landscape pattern relationship of temporary and seasonal wetlands, and 2) develop a landscape-level decision support tool for rapid assessment of wetland condition. We sampled 73 wetlands in the study area. We used the Index of Plant Community Integrity (IPCI) as our measure of wetland condition. A wetland landscape was defined by a 300 m radius circular area (0.283 km²) around each habitat. Quantitative characterization of landscape pattern was conducted using metrics computed from land cover categorization maps processed from multi-temporal Landsat satellite data. Ordination of wetland samples in a multivariate space of landscape metrics using non-metric multidimensional scaling revealed strong associations between wetland condition and 10 landscape metrics, primarily among seasonal wetlands. The Landscape Wetland Condition Analysis Model (LWCAM) was developed and validated for rapid quantitative assessment of wetland condition. The model was based on three landscape metrics considered most important for use in the PPR: 1) grassland percent core area of landscape, 2) grassland largest patch index, and 3) the number of patches per unit area. We concluded that surrounding natural grasslands and landscape fragmentation were the most important influences on the structure and plant community condition of wetland ecosystems.

The objective of this research was to compare and contrast C dynamics within plots occupied by Phragmites australis, Typha spp., and Sagittaria latifolia in a Lake Erie coastal wetland (Ohio, USA). The effect of each species on above- and belowground biomass, soil C pools, soil labile C, litter decomposition rates, and microbial catabolic response profiles were analyzed. Phragmites australis and Typha spp. produced significantly more aboveground biomass (1,522 ± 464 and 1,177 ± 164 g DM m⁻², respectively) than S. latifolia (500 ± 80 g DM m⁻²), although no difference was observed in terms of belowground biomass. After 208 days in the field, litter of S. latifolia had lost 72% of its initial mass while only 47% of the litter of Typha spp. and P. australis had decomposed. This coupled process of high primary production and slow litter decomposition within P. australis and Typha spp. communities did not translate into greater accumulation of C in the soil. In fact, we observed lower rates of C mineralization and greater biomass of methanogens in the S. latifolia plots. Despite similar water level, soil conditions in the S. latifolia community was more saturated, which might have limited availability of C for microbial consumption in these plots. Microbial catabolic responses to 24 substrates demonstrated distinct differences in the respiration responses of the soil microbial communities of the three macrophyte species. The microbial community found in the rhizosphere of P. australis was particularly responsive to phenolic acids. Few differences in C fluxes and pools were observed between plots occupied by P. australis and Typha spp., but the replacement of S. latifolia by one of the two other species could have a significant effect on the C cycle in the Great Lakes coastal wetlands.

Composition of salt marsh vegetation is important to wetland ecosystem health, and monitoring invasive species is critical. The purpose of this study was to examine the utility of airborne hyperspectral imagery in mapping salt marsh vegetation in Humboldt Bay, California, USA. An unmixing algorithm was applied to spatial and spectral image subsets. Overall accuracy among Spartina densiflora, Salicornia virginica, and Distichlis spicata was assessed at 85.1%. Algorithm prediction between observed and predicted percent cover ranged from r²  =  0.32 to r²  =  0.53, an improvement on comparable studies. Percent cover prediction was least accurate for Distichlis spicata, due to initial endmember selection. Use of the Pixel Purity Index in conjunction with field work likely aided in identifying the best candidates for the linear unmixing technique.

We described phytoplankton productivity in a floodplain wetland of the Lower Paraná River, Argentina. Four samplings encompassing periods of high and low water levels were conducted in a highly colored shallow lake. Photosynthesis-irradiance (P-E) curves and the areal photosynthetic rate (P_A) were estimated following the ¹⁴C assimilation technique. Likewise, physical and chemical variables and phytoplankton composition, density, and chlorophyll a concentration were measured. Phytoplankton assemblages shifted from cyanobacteria blooms in summer to cryptophycean dominance in winter, and co-dominance of cryptophytes and chlorophytes in autumn. Assimilation number P_max ranged from 3–7.8 µg C (µg Chl a h)⁻¹ and peaked in early summer (low water level) when water color was highest, suggesting that phytoplankton productivity was not depressed by the high content of colored humic acids. Photosynthetic efficiency (α) ranged from 0.021–1 µg C (µg Chl a h)⁻¹ µmol photons⁻¹ m² s and reached its maximum value during winter even when the assemblage, dominated by cryptophyceans, did not achieve light saturation. In early and late summer and in autumn, optimal irradiance (E_opt) ranged from 544–1,397 µmol photons m⁻² s⁻¹. The highest P_A (207 mg C m⁻² h⁻¹) was registered in late summer (high water level) when the lowest mean irradiance (E_mean) was observed (341 µmol photons m⁻² s⁻¹). The lowest P_A (28 mg C m⁻² h⁻¹) occurred in winter when E_mean was maximum (1,432 µmol photons m⁻² s⁻¹). Our results indicate that productivity was similar to those recorded for other latitudes and appeared not limited by the humic content of the water because phytoplankton was dominated by algae well-adapted to low light conditions.

A recent conflict between Lebanon and Israel began July 12, 2006, and ended August 14, 2006, with one of the hardest hit regions being the Tyre or Sour area. There was concern about environmental damage to one of the most prominent environmental sites in the region, the Tyre Coast Nature Reserve (TCNR) and the artesian springs within it. The recharge zone of these springs lies to the east of the TCNR in an area that was subject to severe bombing. The main objective of this study was to assess the impact of the bombing on the water quality of these springs by comparing water quality data collected before the July war with those collected shortly after the cessation of hostilities. The physicochemical characteristics of the water samples were mostly within normal ranges. However, a spike in the concentrations of nickel and chromium did appear in one post-conflict sample. Therefore, to ascertain that no permanent contamination has occurred, continuous monitoring of the artesian springs and the receiving portion of the Mediterranean should be carried out.

We studied richness and composition of the cladoceran species in 61 temporary peat-pools (pools within high-altitude peatland mires) in the Cordillera del Tunari in Cochabamba (Bolivia) during one wet season. Of the 21 species collected, two were new to science and five were new records to Bolivia. Across all 61 pools, species richness per pool varied from 3–16 (mean  =  8.3 2.8 SD). Rarefaction analysis revealed that more than 80% of all collected species was represented by a subsample of 25 pools. This is the first study to comprehensively present the cladoceran fauna of temporary peat-pools in the high Andes. The cladoceran diversity is comparable with many permanent systems in the region and other temporary pool systems around the world, and underlines the conservation value of these peatland systems (bofedales) for the aquatic biodiversity of South America.