Over several decades, widespread encroachment of Carolina willow (Salix caroliniana), a native shrub, has occurred in areas historically dominated by herbaceous marsh in the upper St. Johns River basin in Florida, USA. Where fire can no longer be used to control mature willow, use of herbicides may be a viable option for management. We evaluated the response of willow communities to application of imazapyr at a low rate and a mixture of imazapyr and glyphosate at high rates at two locations (C40 and Sweetwater) that varied with respect to understory plant composition, willow stand maturity, and hydrologic conditions. Both low and high application rates of herbicides resulted in complete elimination of willow for four years at both sites. By six years post-treatment, willow began to recover in a few treatment quadrats but still had significantly lower willow cover than in control quadrats at both sites, with the exception of the low imazapyr treatment quadrats at the Sweetwater site. Prior to treatment, understory plant communities were essentially uniform across the treatments within a site. Two to three years after treatment, and the concomitant reduction of willow cover, herbaceous ground cover increased substantially, especially that of forbs and graminoids. However, changes in understory plant assemblages were short-lived and by the last sampling year, there were no dissimilarities in the understory between the control and treatments. At the Sweetwater site, two quadrats had sufficient herbaceous cover to carry a prescribed fire in the sixth-year post-treatment. Considering these results, we recommend using low application rates of imazapyr, rather than high application rates of imazapyr/glyphosate, for willow control due to its nearly equal ability to reduce willow canopy cover and to potentially lower off-target effects. However, an integrated approach using varying herbicides and other control techniques (e.g., mechanical control) may be needed if willow develops a resistance to imazapyr. Furthermore, restoration of appropriate hydrologic and fire regimes should be implemented to prolong the treatment effect and to promote long-term sustainability of herbaceous wetlands.

Never before has the resiliency of wetland ecosystems to climatic and anthropogenic stressors been more important or more recognized by those who study these unique ecosystems. The goal of this chapter is to discuss a variety of management and restoration approaches to building resiliency in wetlands that are subjected to changing conditions. We examine wetland responses to changing climatic and hydrologic conditions at multiple spatial (global to microscopic level) and temporal (100-million-year to 1-year) scales which informs our perspective on predicting future wetland responses to both anthropogenic and natural perturbations. Additionally, we introduce the utility of having advanced tools for monitoring changes at the biogeochemical scale, which is likely to be one of the first indicators of change to be detected. The case studies that we present enable us to learn techniques and approaches to address current and future stressors (natural and anthropogenic) on both coastal and inland wetland ecosystems and contain the common thread of carbon sequestration and biogeochemical cycling. We focus on the functional roles of wetlands in providing ecosystem services and how those ecosystem services are best protected, managed, and restored in light of a variety of stressors, such as global climate change, increased water use and demand, and land use changes. Wise-use approaches that enhance wetland biodiversity and resiliency to these changes and impacts are discussed, as are wetland-specific ecosystem services that provide enhanced water quality, water supply, flood protection, storm damage protection, pollution attenuation, and climate change resiliency for adjacent human communities.

State-and-transition models (STMs) have been successfully combined with Dynamic Bayesian Networks (DBNs) to model temporal changes in managed ecosystems. Such models are useful for exploring when and how to intervene to achieve the desired management outcomes. However, knowing where to intervene is often equally critical. We describe an approach to extend state-and-transition dynamic Bayesian networks (ST-DBNs) — incorporating spatial context via GIS data and explicitly modelling spatial processes using spatial Bayesian networks (SBNs). Our approach uses object-oriented (OO) concepts and exploits the fact that ecological systems are hierarchically structured. This allows key phenomena and ecological processes to be represented by hierarchies of components that include similar, repetitive structures. We demonstrate the generality and power of our approach using two models — one developed for adaptive management of eucalypt woodland restoration in south-eastern Australia, and another developed to manage the encroachment of invasive willows into marsh ecosystems in east-central Florida.

In recent decades, invasive shrubs have replaced herbaceous wetlands in many parts of the world. In Florida, the native shrub Salix caroliniana Michx. (Carolina willow) expanded its distribution throughout the upper St. Johns Riv- er, replacing herbaceous marshes with willow swamps. To identify ways to prevent its expansion, we experimentally tested the effects of watering regime, temperature, substrate, and seed source on willow germination and seedling survival. In growth chamber experiments, germination and survival were most affected by watering regime and were greatest in saturated, organic soils. Survival decreased with soil inunda- tion and on drier, sandy soils. Variable texture and nutrient content in native soils had no differential effect on germina- tion or survivability of willow. Time of seed production, seed source, and delay in watering significantly affected germina- tion. Seed germination occurred quickly after being sown. However, seed viability declined just as quickly. Whenever a soil held sufficient water, especially through capillarity, seeds of Carolina willow germinated and survived well. Seasonal manipulation of water levels to flood marshes during seed-fall and to inundate willow seedlings provides managers with an effective strategy for reducing establishment of Carolina willow.

Historically, wetlands along the St. Johns River, Florida, were dominated by herbaceous marshes. However, in the last 50 years many areas transformed to shrub‐dominated wetlands, at the same time a system of levees and canals was constructed to control flooding. We tested the role of water management in controlling Carolina willow (Salix caroliniana), a native shrub that accounts for most of this shift. We assessed survival and growth of seedlings and cuttings on four artificial islands. We planted willow seedlings and cuttings at the spring waterline and at three higher levels (+17.5, +35, and +50 cm) and evaluated their responses to natural hydrologic fluctuations. Overall, seedlings had lower survival than cuttings. Highest mortality occurred during summer floods and willows greater than 50 cm above marsh surface had the highest survivorship. Surviving seedlings attained similar height and biomass among elevations, but the cuttings had greater stem diameter, stem height, and biomass at higher elevations. In the second experiment, we planted seedlings and short (25 cm) and tall (50 cm) cuttings at the waterline and at three higher levels (+25, +35, and +50 cm) in artificial ponds with controlled water levels. Before flooding, seedlings at the highest elevation suffered some mortality due to desiccation, but after flooding, they had the highest survival. Elevation did not affect cutting survival, but those at the lowest elevation had the greatest height and biomass. Hydrologic manipulation can be a powerful tool to control willow establishment. However, its success depends on timely and prolonged inundation or water drawdown.

We identify and discuss the potential risks associated with implementing a prescribed fire program in Florida's Upper St. Johns River basin and evaluate how these risks are being addressed. Specific risks are the threat of cattail (Typha spp.) expansion in recently burned nutrient-rich areas, and the potential threat prescribed burning poses to endangered species, particularly nesting snail kites (Rostrahamus sociablis plumbeus) in a fragmented habitat matrix. We discuss the ecological basis for the use of prescribed fire in the upper basin, including the risks associated with fire exclusion. Altered water budgets, increased nutrient levels, and perhaps a decrease in fire frequency during the past half century have been accompanied by substantial increases in the acreage dominated by willow (Salix caroliniana) and cattail in the basin. To determine the efficacy of fire at reducing willow cover and its effect on mixed sawgrass (Cladium jamaicense) and cattail communities, an experimental bum of 316 hectares was conducted in June 1994. Burning caused a significant (P<0.05) reduction in both cover and density of willows taller than 1.5 meters. Average willow cover was 65% less than preburn levels after one year. Willows not killed by the fire responded with increased sprout production. Average density of willow sprouts increased from 0.11 sprouts per square meter preburn to 3.65 sprouts per square meter postburn (P<0.05). In burned sites dominated by a mixed saw grass-cattail community, saw grass density was unchanged; however, cattail densities more than doubled (P<O.OOOl). The greatest increases in cattail densities occurred at sites which had low cattail densities prior to the bum. Soil phosphorous levels at unburned sites declined significantly by mid-June, whereas soil phosphorous at burned sites remained elevated above levels reported to favor cattail expansion through mid-July. Fire is frequently prescribed by managers to reduce woody shrubs and enhance the vegetation mosaic in wetlands. Managers must consider the potential for negative impacts from fires (such as accelerated cattail expansion), and the potential negative impacts to endangered species. Managers must develop methods for weighing these risks against expected benefits. A better understanding of biotic responses to interactions of fire, hydrology, and nutrients is needed. Citation: Miller, Steven J., Kimberli J. Ponzio, Mary Ann Lee, Lawrence W Keenan, and Steven R. Miller. 1998. The use of fire in wetland preservation and restoration: are there risks? Pages 127-139 in Teresa L. Pruden and Leonard A. Brennan (eds.). Fire in ecosystem management: shifting the paradigm from suppression to prescription. Tall Timbers Fire Ecology Conference Proceedings, No. 20. Tall Timbers Research Station, Tallahassee, FL.

Approximately eight hectares of shrub swamp wetland, primarily consisting of willow (Salix caroliniana Michx.), were roller-chopped to test the efficacy of this management technique in returning the plant community to one dominated by herbaceous species. Roller-chopping caused severe disturbance to the treatment area. Few willows were left standing and a majority of trees were uprooted entirely. Three months following chopping, limited re-sprouting of willow was observed. However, by the following March, no live willows were found in the treatment area. Willow remained absent for three years post-treatment. Hydrologic conditions may have influenced willow mortality through drought-induced stress and subsequent inundation in the treatment area. In contrast, these hydrologic conditions did not appear to affect willows in the control area. Following the elimination of willow, herbaceous plant species re-colonized the treatment areas, increasing in richness and abundance. Free-floating plants became the dominant species in the newly opened areas. Gradually, free-floaters became less important while rooted emergents, such as graminoids and forbs, became more abundant. This study indicates that, under dry conditions followed by flooding, roller-chopping can be an effective method of willow control. However, utilization of heavy equipment may alter topography, compact soil, and affect drainage patterns. Therefore, managers should exercise caution when applying this technique.

In marshes, fire is considered vital in restricting woody invasion and maintaining herbaceous dominance. However, many shrub species are not killed by fire and respond to this disturbance by resprouting. We assessed the response of coastal plain willow (Salix caroliniana Michx.), a common shrub in southeastern wetlands, to a growing season prescribed fire using three metrics: stem density, stem basal area, and cover. We sampled burned and unburned sites before the fire and annually for four years thereafter. Cover of understory species were sampled before the fire and annually for two years after the fire.The initial response of willow was prolific resprouting. However, stem density returned to pre-fire levels by the second year. Basal area and canopy level (>1.5 m) cover decreased after the fire and remained lower throughout the duration of the study. Basal area and canopy cover declined significantly in the unburned site, but the cause of these declines could not be identified. Despite these declines, basal area and canopy cover in the burned sites were lower than in the unburned sites for each of the four years after the fire, although differences were not always significant at the p<.05. Cover of dominant understory species, sawgrass (Cladium jamaicense Crantz) and cattail (Typha domingensis Pers.) declined after the fire and did not recover to pre-fire levels within two years after the fire. Understory species richness increased after the fire.

Expansion of woody species into herbaceous wetlands is a serious concern in wetland management. Prescribed fire is often used as a tool to manage woody species, although many species resprout after fire making control problematic. In this study, we assessed the usefulness of repeated dormant season fires for controlling Salix caroliniana (Michx.) in a floodplain marsh in Florida. Salix is a common shrub in southeastern marshes that resprouts prolifically after fire. We compared stem basal area, stem density, and cover of Salix in three adjacent sites in a floodplain marsh in east central Florida. One site was burned once in February 1997, another site was burned in February 1997 and then again in March 1999 and one site was left unburned. At the unburned site, Salix stem basal area, stem density, and cover increased over the course of the study. In the two burned sites, the first fire destroyed large diameter stems and stimulated production of sprouts. As a result, stem basal area and cover decreased but stem density remained unchanged. The second fire caused a decline in stem density and a further decline in cover. Changes in understory species composition and cover could not be attributed to the fires. Our results suggest that dormant season fires are effective in reducing Salix cover and basal area, and that repeated fires have greater effects than a single fire.

The importance of fire to the maintenance of herbaceous plant communities in Florida wetland ecosystems is widely acknowledged. However, despite the acceptance of fire as a natural and necessary disturbance, ecosystem responses to fire in these systems are still poorly understood. Of particular concern is the effect of fire on the dynamics of plant communities dominated by Cladium jamaicense Crantz and Typha domingensis Pers. High nutrient levels, primarily phosphorus, and prolonged hydroperiods have been associated with Typha expansion into Cladium dominated communities. Recent studies suggest that fire is a disturbance that may play a facilitative role in this process. The objective of this study was to monitor the long-term effects of a single prescribed fire on Cladium and Typha densities in a freshwater marsh in Florida. Transects located at two burned sites and one unburned site were sampled prior to and annually for four years following a prescribed, lightning-season fire. There was a significant increase (P < 0.01) in Typha at both burn sites for two years after the fire. However, this increase was temporary since Typha density declined to pre-burn levels in the third and fourth years post-burn. Cladium density at the burned sites either increased or remained unchanged throughout the study period. When the control site unexpectedly burned in the fourth year of the study, density changes of Typha were similar to those observed at the original burn sites. Overall, we did not see any lasting changes in Cladium and Typha as a result of the fires, even though soil nutrient levels and hydroperiods were within levels documented to enhance Typha expansion.

The objective of this research was to test the effectiveness of several treatments at raising germination percentages of sawgrass (Cladium jamaicense) seeds. Sawgrass seed lots from two years (1991 and 1995) were tested in two separate germination experiments that were run from 1994 to 1995 and 1995 to 1996. Treatments aimed at breaking dormancy were abrasion with sand paper, steeping in hot water, dry heating, soaking in nitric acid or sodium hypochlorite, cold, moist stratification (for a duration of 3 days or 1 month), and supplements of exogenous chemicals (gibberellic acid or potassium nitrate). A combination treatment of wet heat followed by soaking in gibberellic acid was tested. In the 1994–1995 experiment, no treatment was effective at increasing germination over that of untreated seeds (P>0.05). Treatments with dry heat, abrasion, the combination treatment, and the water control significantly reduced germination. In contrast, in the 1995–1996 experiment, treatment with sodium hypochlorite (bleach) significantly increased sawgrass germination in comparison with untreated seeds. Cold, moist stratification for one month significantly increased germination over that found in the water control. The highest germination of nearly 80% was achieved in seeds that were treated with bleach. The results of this experiment suggest that the disinfectant properties of bleach may be one mechanism through which sawgrass germination is enhanced.

The objective of this study was to determine the effects of three hydrologic treatments (saturated, shallowly inundated, and deeply inundated) on sawgrass (Cladium jamaicense Crantz) germination. In addition, germination of sawgrass was evaluated under two different sowing treatments. There was a significant difference (P < 0.05) in germination of sawgrass seeds between the three hydrologic treatments. However, there was no significant difference in germination between different sowing treatments. Germination was highest in saturated samples (average 30.2%), intermediate in shallowly inundated samples (average 16.3%), and lowest in deeply inundated samples (average 9.3%). Faster growth and higher survival of sawgrass seedlings occurred under saturated conditions than under inundated conditions, There was also a distinct seasonal germination pattern in all samples, showing a surge of germination in mid-October to early-November.