Paul R. Wetzel

Maintaining tree islands in the Florida Everglades: nutrient redistribution is the key

The Florida Everglades is an oligotrophic wetland system with tree islands as one of its most prominent landscape features. Total soil phosphorus concentrations on tree islands can be 6 to 100 times greater than phosphorus levels in the surrounding marshes and sloughs, making tree islands nutrient hotspots. Several mechanisms are believed to redistribute phosphorus to tree islands: subsurface water flows generated by evapotranspiration of trees, higher deposition rates of dry fallout, deposition of guano by birds and other animals, groundwater upwelling, and bedrock mineralization by tree exudates. A conceptual model is proposed, in which the focused redistribution of limiting nutrients, especially phosphorus, onto tree islands controls their maintenance and expansion. Because of increased primary production and peat accretion rates, the redistribution of phosphorus can result in an increase in both tree island elevation and size. Human changes to hydrology have greatly decreased the number and size of tr...

Effects of nutrient and soil moisture on competition between shape Carex stricta, shape Phalaris arundinacea, and shape Typha latifolia

We investigated the importance of nutrients, soil moisture, arbuscular mycorrhizal fungi (AMF), and interspecific competition levels on the biomass allocation patterns of three wetland perennial plant species, Carex stricta Lam., Phalaris arundinacea L., and Typha latifolia L. A factorial experiment was conducted with high-low nutrient levels, high-low soil moisture levels, and with and without AMF inoculation. Under the experimental conditions, plant inoculation by AMF was too low to create a treatment and the AMF treatment was dropped from the total analysis. P. arundinacea and T. latifolia biomass were 73% and 77% higher, respectively, in the high nutrient treatment compared to the low nutrient treatment. Biomass allocation between shoots and roots remained relatively constant between environmental treatments, although shoot:root ratios of P. arundinacea declined in the low nutrient treatment. For C. stricta, the high nutrient and soil moisture treatments resulted in an increase in biomass of 50% and 15%, respectively. Shoot:root ratios were nearly constant among all environmental conditions. Biomass of T. latifolia and C. stricta was greatly decreased when grown with P. arundinacea. The rapid, initial height growth of P. arundinacea produced a spreading, horizontal canopy that overshadowed the vertical leaves of T. latifolia and C. stricta throughout the study. This pattern was repeated in both high and low nutrient and soil moisture treatments. When grown with P. arundinacea, C. stricta and T. latifolia significantly increased their mean shoot height, regardless of the nutrient or soil moisture level. The results of this experiment suggest that C. stricta and T. latifolia were light limited when growing with P. arundinacea and that canopy architecture is more important for biomass allocation than the other environmental conditions tested. The results also suggest that Phalaris arundinacea is an inherently better competitor (sensu Grime 1979) than C. stricta or T. latifolia.

RESTORATION OF WETLAND VEGETATION ON THE KISSIMMEE RIVER FLOODPLAIN: POTENTIAL ROLE OF SEED BANKS

The composition of seed banks of areas on the drained Kissimmee River floodplain (Florida, USA) that are currently pasture and formerly had been wet prairie, broadleaf marsh, and wetland shrub communities was compared to that of seed banks of areas that have extant stands of these communities. The species composition of the seed banks of existing wet prairie and former wet prairie sites were the most similar, with a Jaccard index of similarity of 55. Existing and former broadleaf marsh and wetland shrub communities had Jaccard indices of 38 and 19, respectively. Although existing and former wet prairie seed banks had nearly the same species richness, species richness at former broadleaf marsh and wetland shrub sites was higher than at existing sites. Mean total seed densities were similar in existing and former wet prairies (700 to 800 seeds m2). However, seed densities in former broadleaf marsh and wetland shrub sites were significantly greater than in comparable existing communities (>4,900 seeds m2 at former sites versus 200 to 300 in existing communities). The higher seed densities in former broadleaf marsh and wetland shrub sites was due to over 4,000 seeds m2 of Juncus effusus in their seed banks. Half of the species that characterize wet prairies were found in the seed banks at former and existing wet prairie sites. At existing broadleaf marsh and wetland shrub sites, most of the characteristic species were found in their seed banks. However, only one characteristic broadleaf species was found in the seed banks of the former broadleaf marsh sites, and no characteristic wetland shrub species were found in the seed banks of the former wetland shrub sites. The seeds of only two non-indigenous species were found in the seed banks of former wetland communities at very low densities. For all three vegetation types, but particularly for the broadleaf marsh and wetland shrub sites, re-establishment of the former vegetation on the restored floodplain will require propagule dispersal from off-site sources.

Heterogeneity of phosphorus distribution in a patterned landscape, the Florida Everglades

The biologically mediated transfer of nutrients from one part of a landscape to another may create nutrient gradients or subsidize the productivity at specific locations. If limited, this focused redistribution of the nutrient may create non-random landscape patterns that are unrelated to underlying environmental gradients. The Florida Everglades, USA, is a large freshwater wetland that is patterned with tree islands, elevated areas that support woody vegetation. A survey of 12 tree islands found total soil phosphorus levels 3–114 times greater on the island head than the surrounding marsh, indicating that the Florida Everglades is not a homogeneous oligotrophic system. It was estimated that historically 67% of the phosphorus entering the central Everglades was sequestered on tree islands, which are ~3.8% of the total land area. This internal redistribution of phosphorus onto tree islands due to the establishment of trees may be one reason that marshes have remained oligotrophic and may explain the spatial differentiation of the patterned Everglades landscape.

Biogeochemical Processes on Tree Islands in the Greater Everglades: Initiating a New Paradigm

Scientists’ understanding of the role of tree islands in the Everglades has evolved from a plant community of minor biogeochemical importance to a plant community recognized as the driving force for localized phosphorus accumulation within the landscape. Results from this review suggest that tree transpiration, nutrient infiltration from the soil surface, and groundwater flow create a soil zone of confluence where nutrients and salts accumulate under the head of a tree island during dry periods. Results also suggest accumulated salts and nutrients are flushed downstream by regional water flows during wet periods. That trees modulate their environment to create biogeochemical hot spots and strong nutrient gradients is a significant ecological paradigm shift in the understanding of the biogeochemical processes in the Everglades. In terms of island sustainability, this new paradigm suggests the need for distinct dry-wet cycles as well as a hydrologic regime that supports tree survival. Restoration of historic tree islands needs further investigation but the creation of functional tree islands is promising.

Tree Island Ecosystems of the World

Tree islands are defined as patches of woody vegetation within a freshwater wetland matrix dominated by non-woody species. Ecosystems with tree islands as a prominent landscape feature are found throughout the world, suggesting that they arise from a common mechanism of formation. This chapter considers the ecological processes that foster the development and maintenance of tree islands and compares tree islands found in the Florida Everglades to other tree island ecosystems. From the wide diversity of tree island ecosystems two common characteristics emerged: 1) a general mechanism of island formation and 2) vegetation communities that are a subset of the surrounding lowland forests. All tree islands form through a combination of directional, moving waters and biological activity. Islands are initiated with a physical-chemical point of formation such as a bedrock topographic high or low or a minerotrophic groundwater outflow. Biotic factors, usually plants, respond to that point of formation by raising the surface elevation of the island above the surrounding water level through deposition of plant litter. Plants also bind soil substrata or increase island sedimentation by stabilizing the point of formation. Other biotic factors of tree island formation include termites and seed dispersal by animals, primarily birds. Review of the literature also found that the vegetation on tree islands is a subset of the surrounding regional forest community. No endemics or rare plant species are reported to grow on tree islands. Hydrology is the primary factor affecting tree island vegetation in all systems and controls community composition, species richness, and vegetation zonation. Hydrology also controls succession and ultimately is linked to island development. Secondary succession is relatedto the ecosystem disturbance regime. Fire, flooding, and droughtare disturbances common to all tree island ecosystems. All tree island ecosystems form in extraordinarily flat landscapes. If overlandwatershave a low velocityand climatic conditions support peat formation the result is a peatland tree island ecosystem, of which the Florida Everglades is an example. Highervelocity water flows, such as in or alongrivers, result in non-peatland tree island ecosystems. Non-peatland tree islands are subject to abiotic factors common to riverine systems: high water velocity, rapid changes in hydrology, and alluvial geologic forces. These abiotic factors may have a greater influence on island formation than biotic factors, especiallyin the early stages of island development. Peatland tree island systemsare just as dynamic, but biotic factors may dominate the formation of these islands.Tree islands have long distance ecological links that extend far beyond the apparent boundaries of the island, requiring an expansive wetland complex to support them. Human shave affected nearly all tree island ecosystems through physical restructuring, discharge of wastes, or the introduction of exotic species. Although tree islands appear to be resilient, their alteration and destruction in the Everglades clearly illustrates that they can be destroyed or greatly modified by human activities.

Landscape analysis of tree island head vegetation in Water Conservation Area 3, Florida Everglades

Tree islands, forested islands in an herbaceous freshwater wetland landscape, are a major landscape feature in the Florida Everglades. The vegetation communities on the heads of 31 tree islands, including eight islands with recreational camp structures, were assessed throughout Water Conservation Area 3 to determine their composition, structure, and distribution across the landscape. The islands were a sample of the most elevated islands in the local landscape. Measures of forest canopy (> 3 m) and subcanopy (1–3 m) structure and composition, including cover, species richness, number of exotics, and total canopy basal area were ordinated onto six hydrologic variables estimated from the South Florida Water Management Model (v5.5) simulation from 1984 to 1997, and history of recent fire. Ordination allowed identification of four island groups: Group A, higher islands, most with camp structures, low or no canopy structure, high level of fire history, driest hydrology, and largest number of exotic species in canopy; Group B, variable canopy development including many plots without canopy cover, some fire history, and exotics in sub-canopy; Group C., highest islands with well developed canopy structure and no canopy exotics; and Group D, low elevation islands, wettest hydrology, no exotics, and deep peat soils. Cluster analysis of the vegetation cover data was used to identify sub-canopy and canopy communities of the island groups. Our results indicated that the forest canopy of elevated tree islands is similar throughout the central Everglades and that differences in tree island forest composition and structure were the result of local differences in island topography, hydrology, direct human disturbance, and past fire history. Canopy composition and structure were strongly correlated with extreme wet or dry hydrologic events rather than mean or median annual water levels. Fern species were also found to be a ubiquitous component of the sub-canopy. The results of this study identify potentially successful species and provide some basic guidelines for restoring the forest head communities of degraded tree islands.

Potential propagule sources for reestablishing vegetation on the floodplain of the Kissimmee River, Florida, USA

After channellization of the Kissimmee River, the primary land use of the drained floodplain was cattle pasture but included sod farms. A project to restore the river began in 1999. One of its goals is to reestablish the three dominant, pre-channelization vegetation types (wetland shrub, broadleaf marsh, and wet prairie) in areas where they previously were found. We investigated whether indicator species of these three vegetation types were present in 53 permanent quadrats on the drained floodplain. All seven indicator species were found in the permanent quadrats. We also examined three potential sources of propagules (relict wetlands, seed banks, and several surrogates of hydrochory) for these indicator species. All seven species were found in adjacent relict wetlands; and six were found in the seed banks of permanent quadrats. Based on binomial logistic regressions, the presence of relict wetlands and surrogates for flooding (relative elevation, total days flooded) can predict the presence or absence of most of these indicator species. Sod farming reduced the presence of wet-prairie and broadleaf marsh indicator species in permanent quadrats, in adjacent relict wetlands and in the seed bank. The potential importance of relict wetlands for the re-vegetation of the floodplain was our most important finding.

Use of a reciprocal transplant study to measure the rate of plant community change in a tidal marsh along a salinity gradient

In a tidal marsh on the Savannah River (Georgia, USA), rate of plant community change along a salinity gradient was measured using a reciprocal transplant study. Donor sods were moved in all possible combinations from freshwater to brackish sites and from brackish to freshwater sites at four different locations. The reciprocal aspect of the experiment also allowed the determination of how the rate of plant community change is affected by the direction and level of displacement along the salinity gradient. Stem densities of each species were counted in each transplanted plot in June and October for a 30-month period. Plant community structure and composition changed by a significantly measurable amount within 6 to 18 months of a change in salinity. However, the time required for the transplanted sods to resemble their surrounding communities (at the p≤0.05 level) ranged from 6 to more than 30 months, with some transplanted sods never resembling the surrounding plant communities during the study period. If freshwater or oligohaline sods were moved to more saline environments, environmental conditions appeared to have an overriding effect on the vegetation and community change was rapid, occurring in 6–10 months (mean= 9.3 months, SE=1.9). Shifts from brackish to fresher sites on the salinity gradient delayed community change to about 18 months (mean=15.3 months, SE=1.7) and appeared to be controlled by biotic factors such as vegetative expansion and interspecific competition.

Analysis of Tree Island Vegetation Communities

This chapter analyzes vegetation changes on tree islands surveyed in Water Conservation Area 3A between 1977-1986. Presence/absence was measured annually on 24m x 3m plots located on the head of each island. Based on a 15 year hydrologic record, the islands were divided into two groups: frequently flooded and frequently dry islands. In 1981 the frequently flooded islands experienced a drawdown, while the frequently dry islands experienced a drawdown and fire. After fire disturbance, frequently dry islands had a greater number of herbaceous species than frequently flooded islands. The lower number of herbaceous species on frequently flooded islands was maintained even after the flooded islands experienced drier conditions during drawdown. Ordination analysis of all species revealed that of the 15 environmental factors examined, the hydrological gradient represented by the 10 year mean water level and the time period since the last fire disturbance were most important, explaining about 45%of the variation of the data. A separate analysis of woody species frequency correlated most strongly with time since last bum and the difference between the maximum and minimum water levels over 15 years (extreme drought and flooding events) and the average amount of time that the islands were flooded. Woody species may respond to hydrologic parameters over a longer time frame, but long-term hydrologic data (greater than 15 years) was not available for analysis. Drawdown on the frequently flooded islands enabled rapid recruitment of some woody species from the seed bank. Based on the ordination analysis, species were divided into general groups along low, moderate, or high water and time since last bum parameters. From these life history traits, island vegetation composition was predicted under different environments.