|UFDC Home||myUFDC Home | Help|
This item has the following downloads:
1 MEAN VS. VARIANCE: HYDROLOGIC CONTROLS ON WETLAND STRUCTURE AND FUNCTION By JOSEPH M. DELESANTRO A THESIS PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DE GREE OF MASTER OF SCIENCE UNIVERSITY OF FLORIDA 2013
2 2013 Joseph M. Delesantro
3 To my friends and family
4 ACKNOWLEDGMENTS I would like to thank my advisor Matthew Cohen, and committee members, Kath erine Ewel and Mark Brown for all their guidance and support. I would also like to thank Lawrence Korhnak who spent untold hours knee deep in muck with me along the Silver River. Without Larrys experience and patience this project would not have been poss i ble. I would also like to thank Daniel Mclaughlin and David Kaplan who were always there when I needed a sounding bored or advice in my work
5 TABLE OF CONTENTS page ACKNOWLEDGMENTS .................................................................................................. 4 LIST OF TABLES ............................................................................................................ 7 LIST OF FIGURES .......................................................................................................... 8 LIST OF ABBREVIATIONS ............................................................................................. 9 ABSTRACT ................................................................................................................... 10 CHAPTER 1 INTRODUCTION .................................................................................................... 12 2 METHODS .............................................................................................................. 19 Study Area .............................................................................................................. 19 Hydrologic Evaluation ............................................................................................. 20 Productivity ............................................................................................................. 23 Forest Community Structure ................................................................................... 24 Organic Matter Accumulation .................................................................................. 24 Microtopography ..................................................................................................... 26 Statistical Analysis .................................................................................................. 2 6 3 RESULTS ............................................................................................................... 31 Sit Hydrology and Orthogonal Gradients ................................................................ 31 Forest Community Structure ................................................................................... 31 Productivity ............................................................................................................. 32 Morphological Adaptations ...................................................................................... 34 Organic Matter Accumulation .................................................................................. 34 Microtopogr aphy ..................................................................................................... 3 5 4 DISCUSSION ......................................................................................................... 4 7 Comparing Silver River Floodplain to Other Floodplains ........................................ 4 8 Dual Control of Ecosystem Metrics ......................................................................... 4 9 Cypres s Knees ....................................................................................................... 52 Contingent Effects of Hydrology ............................................................................. 53 Soil Organic Matter Recalcitrance Peak ................................................................. 54 Management Implications ...................................................................................... 55
6 APPENDIX A ADDITIONAL SITE C HARACTERISTICS ............................................................... 57 B BASAL AREA AND COVER BY SPECIES ............................................................. 5 8 C ANNUAL LITTERFALL ........................................................................................... 6 0 D SOIL CORE DATA .................................................................................................. 6 1 LIST OF REFERENCES ............................................................................................... 6 4 BIOGRAPHICAL SKETCH ............................................................................................ 7 1
7 LIST OF TABLES Table page 3 1 Hydrologic characteristics of study sites along the Silver River .......................... 37 3 2 Characteristics of study sites along the Silver River ........................................... 38 3 3 Litterfall of study sites along the Silver River ...................................................... 39 A 1 Leaf litterfall indexed to species basal area ........................................................ 57 A 2 Additional Site Characteristics ............................................................................ 5 7 B 1 Long hydroperiod site percent cover by species ................................................ 5 8 B 2 Short hydroperiod site percent cover by species ................................................ 58 B 3 Basal area by species ........................................................................................ 59 D 1 Soil core data ...................................................................................................... 61
8 LIST OF FIGURES Figure page 2 1 Land surface elevation mapping of the Silver River, showing study sites for the comparison of ecological responses to hydrologic mean and variation. .... 28 2 2 Gradient in water level variation along the length of the Silver River ................. 29 2 3 Schematic of a 30 m by 30 m study site and the sampling frames for forest community structure (8 m by 8 m), understory vegetation (0.25 m2), bathymetry (on N S, E W 4 m axis) and litterfall collection (0.5 m2) .................. 30 3 1 Independence of hydrologic metrics .................................................................. 3 6 3 2 Multivariate regression of the response of non Taxodium leaf litterfall to water level STD and hydroperiod by class .................................................................. 40 3 3 Leaf area index .................................................................................................. 41 3 4 Cypress knee structure compared to hydrologic metrics ................................... 42 3 5 Soil organic matter content depth profile from three sites which demonstrate trends in soil organic matter quantity and quality with depth ............................ 43 3 6 Soil organic matter (SOM) content response to hydrologic metrics ................. 44 3 7 Histograms of relative elevation measurements at sites with low, moderate, and high microtopographic relief ....................................................................... 45 3 8 The STD of soil elevation measurements modeled by multivariate regression to hydroperiod and water level STD. ................................................................ 46 C 1 Annual litterfall for four study sites in the Silver River floodplain. ...................... 6 0 D 1 Soil core profiles for long hydroperiod sites. ..................................................... 62 D 2 Soil core profiles for short hydroperiod sites. .................................................... 63
9 LIST OF ABBREVIATIONS HP H ydroperiod MWL Mean water level OM Organic matter SOM Soil organic matter STD Standard deviation
10 Abstract of Thesis Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Master of Science MEAN VS. VARIANCE: HYDROLOGIC CONTROLS ON WETLAND STRUCTURE AND FUNCTION By Joseph M. Delesantro May 2013 Chair: Matthew J. Coh en Major: Forest Resources and Conservation The relative importance of mean vs. variance of exogenous dr ivers to the organization of ecosystem structure and function remains an important open question in ecosystem science Wetlands provide a useful set ting for examining this question, because hydrology acts as the dominant ecosystem driver. However, previous studies have considered variation in hydrology in terms of flood intensity or frequency, which are not independent of the mean. Floodplain wetlands along the Silver River in Florida, USA, provided two natural and orthogonal gradients in the mean and variance, allowing the ir relative controls on ecological structure and function to be evaluated independently. All ecosystem attributes evaluated ( fores t community structure, aboveground net primary productivity, organic matter accumulation, cypress knee morphology, and microtopography ) were controlled by both the mean and variance of hydrologic forcing. Hydroperiod (a measure of hydrologic mean) and water level variation were significantly correlated to soil organic matter content, species specific leaf litterfall (indexed to basal area), and microtopographic variation in multivariate regressions. S tructural attributes of
11 cypress knees and the prevalence of wetland taxa were also correlated to hydroperiod and water level variation. However, the magnitude of response to the mean was consistently larger, suggesting that widespread use of hydroperiod as a predictor of wetland organization is tenable. However, hydrologic variation was an important control, with effects often contingent upon the mean, such that water level variation had a greater ecological impact under short hydroperiod conditions This study underscores the need to consider variation in exogen ous drivers when evaluating, managing, and restoring ecosystems .
12 CHAPTER 1 INT RODUCTION Ecosystems organize in response to exogenous drivers, such as sunlight, hydrology and nutrient availability. Natural gradients in these drivers have provided an opport unity to understand how ecosystems work and thus predict how they change as drivers are intentionally and unintentionally altered. While these exogenous drivers vary in magnitude, duration, frequency, predictability, and timing, most studies are designed t o evaluate gradients in the mean (e.g., mean annual precipitation or hydroperiod), implicitly overlooking effects of variation around the mean (Benedetti Cecchi 2003). This is an important shortcoming because where variation has been studied, i ts effects o n ecosystem structure and function is significant ( Poff and Ward 1989 Odum et al. 1995 Landres et al. 1999, Po r porato et al. 2002 and 2004, Heffernan 2008 ). Moreover, most studies that consider variation quantify variance in terms of event frequency or i ntensity which is generally not independent from the mean ( Benedetti Cecchi 2003) This confounds inferences about the role of variation as a predictor of ecosystem function. To adequately disentangle the independent effec ts of mean and variance, studies must be conducted along orthogonal gradients in the mean and variance. The floodplain wetlands adjacent to spring fed rivers in North Florida provide model systems wherein mean and variance can be disentangled. Steady discharge from the spring vent resul ts in extremely low water level variation at the upstream boundary ( Jawitz and Mitchell 2011). Event driven variation in downstream receiving water bodies extends up the river channel resulting in bottom up flooding dynamics, and establishing a natural gr adient in water level variance along the river from low at upstream locations to high downstream. Variation in soil elevation across the floodplain
13 creates a second natural gradient in the water level mean. These gradients are relevant because hydrology i s the dominant exogenous driver of wetland ecosystems ( Gosselink et al. 1978). Hydrology exerts control over any system by providing water to organisms, but in wetland ecosystems standing water commonly controls the transport of gasses to and from the soi l column and rooting zone. Prolonged inundation results in anoxia and the buildup of toxic chemicals which in turn results in inundation stress. Because hydrology is the dominant exogenous driver of wetland ecosystem organization, these two orthogonal hydrologic gradients provide an appropriate setting for measuri ng the relative effects of mean and variance on ecosystem structure and function. To evaluate the role of these orthogonal hydrologic gradients on ecosystem organization, I chose a suite of ecological metrics responsive to inundation and integrative over time so as not to be confounded by event scale variation. Gradient studies often focus on variation in ecosystem productivity, and vegetation structure and composition ( Connell 1976, Brown 1981, M egonigal et al. 1997). These metrics, general to ecosystems evaluate the stress imposed along a gradient and the adaptations of the system in response to the stress. In wetlands specifically, soil organic matter (SOM) processes are also affected by inundation and the formation of elevated microsites that result from the interaction of productivity and decomposition are an important morphological adaptation of wetland ecosystems ( Beatty 1984, Titus 1990, Scarano et al. 1997, VivianSmith 1997, Pollock et al 1998, Simmons et al. 2011 ). As such, I evaluated the impacts of mean and variance in hydrology on surface morphology and SOM accumulation in addition to ecosystem productivity, and vegetation structure and
14 composition. These ecosystem attributes were expected to be controlled by both the mean and variance. Primary production is among the mos t important ecosystem functions and responds strongly to environmental gradients ( Schuur and Matson 2001). Wetlan d primary productivity is sensitive to gradients in hydrology ( Conner and Day 1976, Mitsch and Ewel 1979, Brinson et al. 1981, Brown 1981, Conner and Day 1992, Megonigal et al. 1997). Flooding frequency has been used to define hydrologic gradients of many studies on wetland productivity ( Johnson and Bell 1976, Keeley 1979, Gosselink et al. 1981, Taylor et al. 1990, Conner and Day 1992, O dum et al. 1995). While it has been demonstrated that flood frequency strongly influences wetland ecosystems, it is not a priori independent of mean water level, making it unc lear whether mean or variation exerts greater control The mean water level may positively ( via moisture or oxygen availability ) or negatively ( via prolonged inundation or desiccation stress) influence productivity Similarly, variation in water level may regulate wetland function via multiple pathways, including promoting productivity when high water events deliver moisture to elevated microsites or when low water events oxygenate sites that are otherwise anoxic. Variation may also limit productivity when changes in water level occur too rapidly or frequently for plants to adjust, or where extreme water levels cause hypoxia or desiccation. In light of the multiple pathways by which the water level regime can affect primary production, and reasoning that water availability would not be a limiting factor in this floodplain ecosystem, I hypothesized th at higher mean water level and greater water level variation would lower productivity.
15 Wetland forest communities respond to multiple scales of hydrologic forci ng by adjusting community composition ( Cronk and Fennessey 2001). Wetland species have developed adaptations that allow them to survive under stressful hydrologic regimes. Some (e.g. Taxodium distuchum ) can tolerate a range of hydroperiods, while others are confined to areas of long (e.g. Nyssa aquatica) or short hydroperiod (e.g. Acer rubrum ). I hypothesized that hydrology exerts strong contr ol on forest composition, with greater hydroperiod and water level variation exerting control via selection for in undation tolerant species. In accordance with the intermediate disturbance hypothesis, which predict s diversity patterns along gradients in tidal magnitude ( Sousa 1979, Hacker 1999), I also predicted higher diversity at intermediate levels of hydrologic variation. Notably, however, tidal wetlands are subject to frequent and highly predictable flooding, not observed in non tidal floodplain ecosystems where flooding is stochastic. As such, the applicability of the intermediate disturbance hypothesis to non t idal wetlands is unknown. Wetlands accumulate large quantities of organic matter (OM), making them important settings for carbon storage and associated biogeochemical functions. OM accumulation in an ecosystem represents the balance between production and decomposition and is under strong hydrologic control ( Sahrawat 2003). Moreover, decomposition is responsible for the release and transformation of nutrients from detritus which may greatly affect ecosystem productivity ( Lockaby et al. 1996 ). Hydrology e xerts control on the rate of decomposition in multiple ways. Standing water impedes oxygen transport into the soil, which limits decomposition. In contrast, the rate of decomposition may be enhanced by water availability when water is limiting, or when
16 f looding distributes exoenzymes (Reddy and DeLaune 2008). Evidence that water level variation, and specifically more frequent wetting and drying cycles, increases the rate of decomposition comes from both laboratory and field studies ( Reddy and Patrick 1975 Baker III et al. 2001, Battle and G olladay 2001). Spatial variation in decomposition rates due to microtopography, uneven distributions of organic matter from roots and litterfall and strong impacts of temporal variation in temperature and water level make direct measurements of decomposition complicated and often poorly representative of long term conditions. However, the quantity and quality of accumulated organic matter can be used to measure integrated hydrologic controls on organic matter dynamic s. Organic matter quality refers to how easily organic matter can be mineralized. Generally, high quality organic matter is mineralized be fore low quality organic matter. T herefore, the quali ty of soil organic matter may respond to hydrology when inundatio n affects the rate of decomposition. I hypothesized that the quantity and quality of SOM would increase with longer hydroperiod and that greater variation in water level would decrease SOM quantity and quality Wetland taxa can withstand varying levels of inund ation stress, with a variety of morphological adaptations. Despite a rich literature on these adaptations ( Coutts and Armstrong 1976 Kozlow ski 1984, 1997, Jackson and Colmer 2005, ) it remains unclear which aspects of hydrology control their express ion. Among these morphological adaptation, cypress knees are particularly charismatic, and a ubiquitous feature of these floodplain forests. Cypress knee variation has been attributed to hydrology ( Brown and Montz 1986 Kernell and Levy 1990, Kummer et al 1991, Briand 2000), but the role of knees in tree metabolism remain unclear, as do the controls on knee morphology and
17 frequency. Because it has been suggested that cypress knees play an important r ole in gas exchange ( Kernell and Levy 1990) I predicted that cypress knees will increase in frequency with longer hydroperiod (i.e., conditions requiring greater gas exchange), and increase in height with both greater mean water level and water level variation (i.e., to ensure atmospheric exchange under all inundation conditions). Furthermore, I predict ed that variation in cypress knee height will correspond to variation in water level (i.e., to provide a range of knees that balance the cost of taller knees with the benefit of gas exchange during frequent high water events). Microtopographic variation is a critically important feature of many wetland ecosystems, contributing to productivity and biodiversity ( Beatty 1984, Titus 1990, Scarano et al. 1997, VivianSmith 1997, Pollock et al. 1998, Simmons et al. 201 1, Washuta 2011 ). The mechanisms controlling formation of high and low areas, often referred to as hummocks and hollows respectively, are thought to be self reinforcing feedbacks between soil elevation and below ground root production. In brief, plant gro wth is enhanced on hummocks because of reduced inundation stress. This enhanced productivity on hummocks yields more detritus and below ground biomass. The rate of OM decomposition is also enhanced on hummocks constraining hummock growth ( Hilbert et al. 2000, Eppinga et al. 2009, Belyea and Clymo 2001). S oil elevations are stable where the rate of litter deposits and belowground biomass growth equal the rate of decomposition, which can happen in deeper water (hollows) where production and respiration are low, and at higher elevation (hummocks) where both OM fluxes are high. I hypothesized that the prevalence of hummocks will increase with longer hydroperiod. To that end, I predicted that soil elevations will be bimodal and that
18 the elevation between modes will be positively correlated to hydroperiod. I also predicted that variation in microtopographic relief around a mode would increase with greater water level variation. These predictions span a wide array of ecosystem attributes, from organismal resp onses to community composition to ecosystem processes. Together, they help isolate effects due to water level mean and variation, and provide a system specific test of the hypothesis that both mean conditions and the variation around the mean exert control on ecosystem organization.
19 CHAPTER 2 METHODS Study Area The Silver River is a minimally disturbed spring fed river that flows into the Ocklawaha River, a major tributary of the St. Johns River (Figure 2 1 ). Because the sou rce of the spring flow is the Floridan Aquifer underlying several southeastern U.S. states water level variation near the spring head is extremel y low, both for rivers broadly and also relative to water levels in the Ocklawaha River, which is subject to significant event driven variation. This bottom up flooding regime creates a marked gradient in water level variation along the river (Figure 22 ). Orthogonal to river flow, variation in soil elevation creates a second axis of hydrologic variation in me an depth. The deeper areas of the floodplain are dominated by bald cypress ( Taxodium distichum ) and water tupelo ( Nyssa aquatica ) with an understory comprised of deep water habitat emergent macropytes like pickerel weed ( Pontedaria cordata). Higher eleva tion areas are populated with green ash ( Fraxinas caroliniana), red maple ( Acer rubrum ), sabal palmetto ( Sabal palmetto) and numerous understory plants including maindencane ( Panicum hemitomon) lizards tail ( Saururus cernuus ), and marigold ( Bidens laevi s) Because river and floodplain water levels are regulated by the downst ream Oklawaha River (i.e., backwater flooding), flow velocities throughout the floodplain are very low regardless of flooding depth. Five transects were established perpendicular to the river (Figure 2 1 ), spaced evenly along the river starting 1.5 km from the spring head and extending 6.7 km downstream to a location 0.65 km upstream of the confluence with the Ocklawaha River. Stage data from Jan1 1970 to March1 2011 were obtained from two gauging
20 stations, one 1200 m from the spring head (USGS ID #02239500) and the other 300 meters downstream of the Silver River on the Ocklawaha (USGS ID #02240000). Continuous stage records between these end points were interpolated to estimate d aily stage at each of the five transects. Four years of overlapping river stage data at each transect location allowed us to refine the end point interpolation (R2=0.96, p<0.001), yielding modeled daily stage data since 1970 for any point along the river between the two longterm gauges. Water elevation exceedance probabilities at each location were developed from this 41 year interpolated record, and two target elevations were obtained corresponding to long (4060% exceedance probability) and short hydroperiods (1020% exceedance probability). Using a high resolution Light Detection and Ranging (LiDAR) derived elevation map with a resolution of 5 ft by 5 ft, we identified long and short hydroperiod locations in the floodplain along each transect. Two stu dy sites, each 30 m by 30 m, were selected along each transect for a total of 10 study sites, 5 of which were long hydroperiod, and 5 short hydroperiod spanning a longitudinal gradient in water level variation. The corners and centers of each study site w ere marked, and a benchmark was placed in each plot, to which hydrologic, bathymetric, and soil core elevation measurements could be compared. At each site forest community structure, aboveground net primary productivity (ANNP), soil bathymetry and soil c ores for organic matter accumulation measurements were taken (Figure 2 3). Hydrologic Evaluation A broad array of hydrologic metrics might be used to represent the variance and mean of site hydrology. Critically, however, any metrics for the mean must be orthogonal to metrics for variation to be useful in decomposing their relative causal
21 effects. Many metrics examined were highly correlated to other metrics providing little additional value. After evaluating over 30 possible metrics, mean water level (M WL) and hydroperiod (HP) were chosen to represent the mean hydrologic conditions, and the standard deviation (STD) of water level was chosen to represent water level variance. While these met rics are intuitive and common, they may be quantified at many tem poral scales (e.g., period of record, annual, monthly, weekly, 5 day moving average). Evaluation at many scales yielded highly correlated results and little evidence to support one timewindow over another. As such, all hydrologic metrics were calculated f or the entire 41 year period of record as it is both simple and most integrative. Mean water level and hydroperiod both descri be the mean hydrologic condition but in different ways. H ydroperiod describes how often flooding occurs regardless of depth, wher eas mean water level can be strongly influenced by extreme water levels Because many ecosystem processes may exhibit no additional response to water s level above or below certain elevations mean water level may be less informative than hydroperiod. There fore, hydroperiod is the most appropriate metric of mean. I retained both hydrologic metrics, however, because mean water level is important when an ecological r esponse is driven by a dist inct level (e.g. cypress knee height depth to soil horizon) rather than just the fraction of time a site is inundated To test the assumption that river and floodplain stage are directly connected, we installed high precision pressure transducers (Solinst Level Loggers) in 5 cm diameter shallow wells at least 1 m below ground elevation at each of the 10 sites. The water depths reported by the transducers were linked to surveyed site benchmarks using standing water elevation as the datum, yielding water surface elevations with the same
22 datum as river stage measurements. S tage data from site wells were compared to river stage data at each transect to test the concordance of floodplain and river stage, and to identify conditions during which the records diverge. Long term stage data from the two permanent gages indicate that average river stage for the one year period of site data collection was fairly representative at the 54 percentiles and 64 percentile for the long term stage gauges ( USGS ID #02239500 and USGS ID #02240000 respectively) During the period of site stage me asurements, there was a period (February to May) during which river and site stage diverged; this corresponded to the end of the dry season and the beginning of the growing season, and indicates that there are regular periods during which rainfall and evapotranspiration control site water level. This period of disconnection between the river and floodplain was removed when assessing the relationship used for predicting site stage with river stage. When the floodplain operates independently of the river th e water level under the floodplain varies more than the river and is generally lower. Therefore predicting site stage with river stage will underestimate water level variation and overestimate stage for the period of disconnection. The stage dynamics during this period are expected to have lower impact on ecosystem structure and function relative to the rest of the year because the water level is general ly below the soil and not impeding gas transfer. To estimate site mean water level and hydroperiod, I use d the short term relationship between site water level and the interpolated river water level to construct a relationship that was then used to back cast conditions over the last 40 years. Note t hat both hydrologic metrics (mean water level and hydroperiod) vary with surface elevation within sites, so all hydrologic metrics are reported for the median elevation, obtained
23 from site bathymetry surveys. Measured site hydroperiods were all within the intended hydroperiod ranges except for the two T 1 sites (upper ri ver) which both had a shorter hydroperiod than the targets, though both show characteristics consistent with their original long and short hydroperiod classifications, principally in SOM content Productivity To compare productivity between study sites, I used litterfall traps to estimate minimum aboveground net annual production. Litterfall measurements are a common and proven method for estimating minimum aboveground net primary production (ANPP) ( Bray and Gorham 1964). Six litter traps (0.5 m2) elevated 1 m off the wetland surface were sampled monthly J uly 2011 to June 2012. Litter trap locations within sites were random (Figure 2 3) but were preferentially placed away from Sabal palmetto because the large palm fronds may intercept and collec t litter Litterfall mass was dried at 80 degrees C for 72 hours before being separated into leafy, woody, reproductive, Taxodium reproductive and miscellaneous components. Leaf litter was f urther subdivided into Taxodium and other (consisting mainly of Nyssa aquatic a Acer rubrum and Fraxinas caroliniana). Each category of litter was weighed and reported as grams per square meter per year Litter and leaf litterfall per unit basal area was also calculated to provide a measure of productivity that accounts for site forest community structure and age variation. The productivity of Sabal palmetto was assessed separately from canopy litterfall At the onset of the project a randomly placed 8 m by 8 m plot was marked off and cleared of all existing Sabal pa lmetto litter at each short hydroperiod site. T he Sabal palmetto litter was collected monthly from the ground of plot. Litter was not collected for two months when substantial flooding was known to have occurred on the sites. As a result the annual total S abal palmetto litterfall was calculated as the 10
24 month average multiplied by 12 months. Sabal palmetto litterfall was not assessed at long hydroperiod sites because of the species very low occurrence at these sites. Projected leaf area index (LAI) was also measured as a proxy for site productivity ( Brown 1981, G holz 1982) Leaf area index was measured by ceptometer (AccuPar LP 80), which evaluates the light passing through the canopy. Measurements, each of which integrates 80 point readings of incident light, were taken every 5 m along seven transects spanning the study site. Forest Community Structure Forest community structure was surveyed in 8 m by 8 m subplots randomly placed within each of four study site quadrants at each site. Diameter at breas t height and species of each tree within the subplots was recorded. Ground cover was surveyed and recorded by percent cover within 3 randomly selected 0.5 m2 areas within each subplot (Figure 2 3 ). The dominance of facultative and obligate wetland species (as defined by the 2012 National Wetlands Plant List; Lichvar et al. 2012 ) within each study site was evaluated by the proportion of basal area, plant density and ground cover, contributed by wetland taxa. Cypress knee density was measured in each forest structure subplot, and height was measured using a laser level elevation relative to the site bench mark. Cypress knees were characterized by their density site average height, and the variance in heights at each site. Organic Matter Accumulation Soil o rganic matter accumulation between sites was compared by evaluating the quantity and quality of organic matter in soil cores. Three 5 cm diameter soil cores, 50 cm long, were taken from random locations within each study site; I avoided taking cores from h ummocks. Because cores were collected when sites were dry in winter,
25 the elevation of the top of each soil core was measured by laser level relative to the benchmark. Sharpened 5 cm diameter PVC pipe was driven into t he ground to a depth of 50 cm. An air tight seal was placed on the top of the pipe and the core was carefully removed. Cores that compacted more than 10 cm or that lost material from the bottom during ex traction were rejected and the sites were resampled. Each core was separated in 2 cm segmen ts. Live root mass was removed and samples were dried at 105 degrees C for 72 hours. Samples were then ground and passed through a # 10 sieve. A two step combustion process was utilized to provide an index of recalcitrance ( Schnitzer and Hoffman 1966, Kris tensesn and Andersen 1987, AngehrnBettinazzi et al. 1988 Sharma 1989, Kristensen 1990 Lopez Capel et al. 2005). Pilot testing indicated that a 1 g sample of soil was sufficient to represent each core segment and would undergo complete combustion of org anic matter in 4 h. The exact weight of each sample plus the weight of the aluminum trays were recorded. The samples were placed into a preheated muffle furnace at 350 degrees C for 2 hours, then removed and placed in a desiccation chamber to cool for 15 m in before being weighed again. Samples were then placed into the preheated muffle furnace at 550 degrees C for an additional 2 hours, after which samples were once again removed and placed in a desiccation chamber to cool for 15 min before being weighed a final time. The total mass of organic matter was measured by the total fractional loss from the combined two step combustion and the recalcitrance index was reported as the mass lost in the second combustion step over the total mass lost The average % SO M was well correlated ( R2 = 0.84, p < 0.001) to the % SOM of the top 10 cm. T herefore the whole core average % SOM was used to analyze OM accumulation in response to hydroperiod and water level STD. Likewise, the mean
26 recalcitrance index for the whole cor e was used to anal yze OM quality in response to hydroperiod and water level STD. Hydroperiod and mean water level were calculated relative to the ground elevation at which each core was collected. H owever, because water level STD is a site level attribute, % SOM and recalcitrance index was averaged at the site level to prevent pseudoreplication. In order to systematically identify peak s in SOM recalcitrance, a piecewise linear model was used. The peak in SOM recalcitrance is identified as the depth at which the model predicting SOM recalcitrance switches from one linear regression to the next. The depth to peaks were also averaged by site for comparison to water level STD Microtopography Site microtopographic relief was obtained using bathymetric survey s in two randomly selected, nonoverlapping, sampling stations in each quadrant of the study sites (n = 8 stations per site Figure 2 3 ). Soil elevation measurements were taken relative to the site benchmark by measuring water depth in areas with standing water and using a laser level everywhere else. Within each sampling station a total of 17 elevation measurements were taken, starting at the sampling station center with four more at random distances along each cardinal direction to a maximum distance of 4 meters from the center (Fig. 2 3). This resulted in a total of 136 soil elevation measurements at each study site. Soil elevation bimodality, variance, and skewness were evaluated for each site. The number of modes in the elevation distribution was analyzed using a BIC analy sis in the MCLUST package in R Statistical Analysis The statistical significance of all results was analyzed by t tests and regression (F test) with a primary and secondary significance level (p <0.05, p<0.10). The
27 secondary significance level was selected to reduce the probability of type II errors because of the studys low statistical power (n=10), and to allow us to investigate trends among the long and short hydroperiod sites independently (n=5). The following methods were utilized in the analysis of each hypothesis on ecosystem metric response. Method 1: The influence of hydroperiod or mean water level and water level STD on each metric of structure and function was analyzed in a multivar iate regression. Method 2: Hydroperiod is discretized into short and long hydroperiod classes and hypotheses are tested again by multivariate regression. Method 3: In order to identify countervailing effects in each hydroperiod class water level STD is compared to metrics of ecosystem structure and function in a single variable regression for each hydroperiod class. Method 4: All hypotheses are tested for long and short hydroperiod sites separately in order to identify sensitivity of ecological metric t o hydrologic metric dependent upon hydroperiod class.
28 Figure 21 Land surface elevation mapping of the Silver River, showing study sites for the comparison of ecological responses to hydrologic mean and variation. Transect along which sites where selected are numbered 15 from upriver to downriver. The L and S denote long and short hydroperiod respectively. The callout demonstrates the selection of study sites based on elevation denoted by grayscale. Elevation and modeled mean water level are used to identify study sites with hydroperiod in the target ranges (40% 60 % inundation for long hydroperiod sites and 10% 20% for short hydroperiod sites).Study sites are chosen based on modeled mean water level Stage is recorded ho urly where transects cross the r iver and at each study site.
29 Figure 22 G radient in water level variation along the length of the Silver River. A) Hydrograph for 2010 showing river stage for the upper (T1) middle (T3) and lower (T5) portions of the Si lver River B) Box and whisker plots of 40 yrs. of back casted stage data for 5 l ocations along the Silver River C) Autocorrelation function (ACF) data for 40 yrs of back casted data stage data at three locations along the upper, middle, and lower portions of the Silver River.
30 Figure 2 3 Schematic of a 30 m by 30 m study site and the sampling frames for forest community structure (8 m by 8 m), understory vegetation (0.25 m2), bathymetry (on N S, E W 4 m axis) and litterfall collection (0.5 m2).
31 CHAPTER 3 RESULTS Site Hydrology and Orthogonal Gradients of Mean and Variance The observed gradients in mean and variance varied independently alon g the length of the river (Fig 3 1 ). Flooding frequency, a metric used commonly t o describe hydrologic variation, was much better correlated to mean water level than was water level STD (R2=0.22 vs. R2=0.03). When flooding frequency was defined by the number of floods of X duration, the correlation between mean water level and flood fr equency increased with X ( e.g. When X=60 days, R2=0.9 6 ). Original estimates of site elevation a nd hydroperiod were made from LIDAR data. Later estimates of hydroperiod were obtained by comparing the recorded water level in the floodplain and river at high stage. The original hydroperiod predictions were generally accurate and within the defined hydroperiod classes (Ta ble 3 1), however the post hoc estimate of hydroperiod was mu ch lower than expected at both T 1 sites. Despite this the T 1 sites share many at tributes with their original classification of hydroperiod, especially %SOM, and as such are kept in the original classification. While mean water level and hydroperiod are very well correlated (R2=0.96), both metrics were still used to accommodate the e valuation of processes where actual depths were relevant (mean water level) as well as processes where the wet/dry fractions were more relevant (hydroperiod). Forest Community Structure Basal areas ranged from 31 to 132 m2/ha and were dominated by Taxodium distichum at all sites except T3L and T5S where a handful of large Fraxinas carolinan and Acer rubrum dominated the basal area (Table 32) Basal area was
32 negatively c orrelated to the variation in hydroperiod among sites with in the short hydroperiod cl ass (R2=0.59, p=0.08). Understory percent cover was generally larger at long hydroperiod sites, but was a function of LAI (R2=0.54) and not hydrologic metrics. The prevalence of obligate wetland (OBL) and facultative wetlands (FACW) plants, as defined by the national wetland plant list ( Lichvar 2012), was generally higher (Table 32) and was moderately significantly greater at long hydroperiod sites than at short hydroperiod sites (p= 0.07). However, variation in hydroperiod among sites was only a significant predictor of the prevalence of OBL and FACW plant species at long hydroperiod sites (R2=0.68, p=0.05). The prevalence of OBL and FACW plant species was negatively correlated to water level STD at short hydroperiod sites (R2=0.92, p<0.001) but that stron g c orrelation was absent at long hydroperiod sites. I observed no significant correlation between species richness and any hydrologic metrics. Productivity Litterfall (Table 33) was strongly conce ntrated in the autumn, with 79 % of total annual litterfall occurring between September and December and 91% of leaf litter fall occurring during this period. Litterfall at all sites was dominated by leaf litter; nonleaf litter (branches, reproductive material) comprised a small fraction ( 23 % 7%; mean 1 STD ) of the total mass, and was pool ed for all subsequent analyses. Taxodium accounted for the majority of leaf litterfall throughout the floodplain (56%), but across sites there was a large range in the proportion of T axodium leaf litterfall (28% 88%). Sabal palmetto accounted for a significant proportion of the ANPP estimate of the T2S, T3 S, T4S, and T5S sites (9 % 20 % ). Total l itterfall ( including Sabal palmetto litter ) was moderately correlated to hydroperiod (R2=0.26, p=0.076) and decreased with
33 longer hydroperiod. Litterfall excluding Sabal palmetto was not correlated to metrics of hydrology. Leaf litterfall including Sabal palmetto was also well correlated to hydroperiod (R2=0.59, p=0.006) exhibiting the same trend, however, canopy leaf litterfall was only moderately correlated to hydroperiod (R2=0.28, p=0.067) There was no significant response to water level STD in either case. Taxodium leaf litterfall was positively correlated to water level STD, but only among long hydroperiod sites (R2=0.78, p=0.03) Litterfall per unit basal area as a whole did not correlate to any hydrologic metric. Similarly, Taxodium distchum leaf litterfall per unit basal area alone was not correlated to any hydrologic metrics (Fig ure 3 2. a ) However, litterfall per unit basal area excluding Taxodium dist i chum was significa ntly negatively correlated to water level STD and hydroperiod class in a multivariate model (Figure 3 2. b ). The slopes of independ ent fits of the data for each hydroperiod class are not significantly differen t. The non Taxodium leaf litterfall data for the short hydroperiod sites had a better fit to water level STD than did the long hydroperiod sites (R2=0.53, p=0.09 vs. R2=0.18, p=0.26) Leaf area index varied significantly across sites, but was generally hi gh (range 3.31 to 7.54 m2 m2; Table 1). Replicate measurement of LAI in subsequent growing seasons w as highly correlated (R2 = 0.57) suggesting that site variation observed is fairly robust to inter annual variation. Leaf area index was significantly negatively correlated to hydroperiod (Figure 5). W ater level STD was not a significant predictor of LAI and there was no correlation between LAI and water level STD within hydroperiod classes. While both LAI and leaf li tterfall declined with longer hydroper iod they were
34 only significantly correlated when Sabal palmetto litter was included in leaf litterfall ( R2 = 0.57, p< 0. 001 ). Morphological Adaptations Cypress knees were present at all sites, and varied dramatically in their density mean height, and heig ht variance (Table 3 2 ). C ypress knee density was moderately well correlated to hydroperiod (R2=0.25, p=0.08) decreasing with longer hydroperiod. Cypress knee height was strongly correlated to mean water level (Figure 3 4.a ). Variation in knee height was also positively correlated to water level STD, but only at the long hydroperiod sites (Figure 3 4.b ). The mean elevation of cypress knees was inundated only 5.5% of the year at any site, suggesting that knees respond to high water conditions. The mean elevation of cypress knees was between 50 and 80 cm above the mean water level at all but one site, T1 L, w hich had the shortest modeled hydroperiod of any site (ca. 4%). Organic Matter Accumulation Soil organic matter content was high at all sites (Table 3 2 ), but histic epipedons (i.e., > 20% OM) occurred only at the long hydroperiod sites. Soil organic matter content was strongly depth dependent, but the pattern varied between short and long hydroperiod sites. At short hydroperiod sites there was a consiste nt pattern of declining SOM content with depth. In contrast, at long hydroperiod sites, there was an initial declin e in SOM with depth, but a consistent inflection below which SOM increased again ( Fig ure 35 .a,b ). Soil organic matter recalcitrance also v aried with depth, increasing to a threshold, below which it declined ( Fig ure 3 5 .a,b ). Unlike SOM content,
35 this peak was observed at both long and short hydroperiod sites. The depth to this peak was negatively correlated to mean water level (R2=0.16, p=0.018) Depth averaged % SOM and recalcitrance index were strongly correlated to hydroperiod (Figure 3 6.a ). Average site %SOM was negatively correlated to water level STD and positively correlated to hydroperiod in a multivariate regression (Figure 3 6.b ) While hydroperiod explains most of the variat ion in SOM content, the effects of water level STD are significant an d exert strong effects within hydroperiod classes, especially for short hydroperiod sites (Figure 3 6.c ). Soil organic matter recalc itrance was not correlated to water level STD. Microtopography Although the distributions of soil elevation data were not bimodal (Fig 3 7 ) elevation variation was positively correlated to hydroperiod class and water level STD in a multivariate model ( Fig 3 8 ). The slopes of independent fits of soil elevation data for each hydroperiod class were not significantly different. Soil elevation distributions for all sites were skewed left (Table 3 2 ). However the degree of skewness did not correlate to hydrologic metrics.
36 Figure 31 Independence of hydrologic metrics. A ) Hydroperiod and mean water level are st rongly correlated and covary. B ) Mean water level is not correlated to water level STD demonstrating that hydrologic metrics for mean (hydroperiod and mean water level) are orthogonal of metric of variance (water level STD).
37 Table 3 1. Hydrologic c haracteristics of study sites along the Silver River. Study Site Metric T1L T2L T3L T4L T5L T1S T2S T3S T4S T5S Hydroperiod (% inundati on) 21.7 60.8 46.1 46.2 46.2 4.5 13.3 16.8 12.3 21.0 Mean Water Level (m) 0.22 0.06 0.01 0.00 0.01 0.47 0.34 0.31 0.40 0.31 Water level STD (m) 0.27 0.28 0.31 0.33 0.38 0.27 0.28 0.31 0.33 0.38
38 Table 32 Characteristics of study sites along the Silver River. Study Site Met ric T1L T2L T3L T4L T5L T1S T2S T3S T4S T5S Basal Area (mha ) 63.6 89.4 31.1 57.0 132.5 70.4 55.7 51.7 77.0 41.1 Taxodium 25.7 57.0 19.8 14.6 74.3 54.0 42.1 48.2 46.8 13.3 Non Taxodium 37.8 32.5 11.3 42.4 58.2 16.4 13.6 3.5 30.3 27.8 Project ed Leaf Area Index 4.8 0.08 3.3 0.43 4.2 0.09 4.2 0.55 3.5 0.05 4.9 0.01 7.54 0.15 5.9 0.56 6.5 1.11 5.1 0.19 Stem Density (stems per ha) 703 625 625 508 664 508 2852 586 781 2344 Understory Cover (%) 79 87 116 14 108 43 4 18 0 0 Fra ction Wetland Taxa 0.60 0.94 0.84 0.89 0.99 0.86 0.73 0.70 0.58 0.19 Cypress Knee Density (knees per 100 m) 31.3 16.0 14.5 31.3 45.3 27.0 173.8 87.9 146.5 74.2 Cypress Knee Height (cm) 49 15 70 18 52 23 71 23 78 33 41 17 29 15 26 17 17 17 20 10 Soil Elevation Variance (cm) 6.38 8.70 9.00 8.43 10.22 5.03 4.95 3.73 6.45 6.62 Soil Organic Matter Content (%) 27.7 1.7 56.8 3.5 33.1 1.7 43.9 2.3 34.5 6.6 13.4 2.0 11.8 2.3 10.7 2.4 5.2 0.8 6.3 0.5 Soil Organic Matter R ecalcitrance Index 0.39 0.03 0.26 0.02 0.26 0.01 0.31 0.00 0.31 0.02 0.34 0.01 0.38 0.01 0.32 0.03 0.47 0.02 0.41 0.01
39 Table 33. Litterfall of study sites along the Silver River. Study Site Litterfall (gm 2 yr 1 ) T1L T2L T3L T4L T5L T1S T2S T3S T4S T5S Total Leaf 569 436 404 614 545 642 646 491 626 487 Taxodium 158 252 239 290 343 367 436 433 254 238 Other 411 184 166 324 201 275 210 58 372 249 Wood y 137 85 41 161 91 124 87 85 109 24 Totoal Reproductive 102 99 23 94 81 92 73 62 100 19 Taxodium 72 74 19 53 66 83 67 54 28 2 Other 30 24 4 41 15 8 6 8 72 17 Sabal Palmetto 24 73 132 71 80 Total 811 624 469 874 721 885 882 774 911 61 0
40 Figure 32. Multivariate regression of the response of nonTaxodium leaf litterfall to water level STD and hydroperiod by class. The regression shows that hydroperiod and water level STD influence the leaf litterfall of nonT axodium species
41 Figure 33. Leaf Area Index. The leaf area index of all study sites decreases with hydroperiod even within classes of long and short hydroperiod.
42 Figure 34. Cypress knee structure compared to hydrologic metrics. A ) The height of cypress knees is positively correlated to mean water level. B ) The STD in cypress knee height is positively correlated to the water level STD at long HP sites. However, there is no correlation between the knee height STD and water level STD at short HP sites.
43 Figure 35 Soil organic matter content depth profil e from three sites which demonstrate trends in soil organic matter quantity and quality with depth A) Soil organic matter content depth profile of core T2L1 d emonstrating the trend amoung long hydroperiod sites. Notice the pronouced inflection at intermediate depth. B ) Soil organic matter content depth profile of core T2S 1 that demonstrates the trend amoung short hydroperiod sites. C) Recalcitrance index depth profile of core T4L 3 that demonstrates the trend observed at long and short hydroper iod sites Notice the inflectio n in recalcitrance at depth.
44 Figure 36 Soil organic matter (SOM) content response to hydrologic metrics. A) Perce nt soil organic matter is positively correlated to hydroperiod while the recalcitrance index is negatively correlated to hydroperiod. B ) 3D mesh of the multi variate model describing the response of %SOM to hydroperiod and water level STD. The %SOM increases with hydroperiod and d ecreases with water level STD. C ) The % SOM is positively correlated to water level STD for short hydroperiod sites, but not for long hydroperiod sites.
45 Figure 37. H istograms of relative elevation measurements at sites with low, moderate, a nd high microtopographic relief
46 Water Level STD (m) 0.26 0.28 0.30 0.32 0.34 0.36 0.38 0.40 Soil Elevation STD (cm) 3 4 5 6 7 8 9 10 11 Long HP Short HP Multivariate regression R2=0.80, p=0.002 Figure 38. The STD of soil elevation measurements modeled by multivariate regression to hydroperiod and water level STD.
47 CHAPTER 4 DISCUSSION The floodplain of the Silver River provides a model system for investigating effects of mean environmental conditions independent from those of variation around the mean. Several features are of particular note. First, the back water flooding and length of the river create a strong gradient in hydrologic variation between stable conditions at the spring vent and marked event driven variation at the downstream confluence. This longitudinal gradient is orthogonal to the typical lateral gradients in hydroperiod present in riparian swamps. Second, the absence of scouring floods (floodwaters are typically the slowest velocity flows during these backwater flooding rivers ) means that flooding events are not characterized by variable flow throughout the floodplain and measured conditions are integrative over long time periods. Finally, because the gradients occur over relatively short distances (ca. 8 km), concerns about dispersal and climate variation confounding compari sons among more dis tant sites are alleviated. Despite site aspects that are well suited to addressing the core question of this research, several important aspects of historical land use and floodplain morphology merit consideration. First and foremost, this floodplain was subject to timber extraction in the last 100 years, but the timing and spatial extent of this is largely unknown. Second, water quality in the river has declined over the last 50 years, principally from nitrate enrichment, which may have altered floodplain wetland function (though presumably relatively uniformly across sites). Finally, there are several notable relict channels and lateral inputs of unknown variation across t he floodplain. However, sites w ere located so as to minimize the possible influenc e of these inputs. While these
48 issue s merit consideration in the interpretation of results, none seems to significantly occlude the findings of the study There were a couple possible issue s with the data analysis of the study. First, the shorter than expected hydroperiod of the most upstream long hydroperiod site (T1L) may merit its inclusion in the short hydroperiod class; however, it was kept in the long hydroperiod class because it exhibited many of the characteristics of long hydroperiod sites, i n particular, the high SOM content. Second, Sabal Palmetto may have biased LAI and total litterfall measurements as it was only prevalent at short hydroperiod sites. Comparing Silver River Floodplain to Other Floodplains Although the Silver River floodplain has features that distinguish it from other floodplains in the region, I observed important similarities with other riparian and nonriparian wetlands of the southeastern United States While average leaf litterfall (545 83 g /m2/yr ) fell within the range reported for southeastern floodplain forests ( 405 837 g/m2/yr ) ( Brown 1981, Megonigal et al. 1997, Shure et al. 1985, Cuffney 1988 Conner and Day 1992, Clawson et al. 1996 ) T3L was just below the range. Stem density is not widely reported in the literature and available figures suggest a large range (530 2000 stems/ha; Dabel and Day 1977, Brown 1981, Megonigal et al. 1997) but mean stem density in the Silver River was within this range (1010 stems/ha >2.5 cm DBH) despite t wo s tudy sites above (T2S and T5S) and two below (T4L and T1S) the range. Mean basal area across sites (67 27 m2/ha 60 17 m2/ha excluding T5L: 132 m2/ha ) was higher than values reported for other floodplains (2247 m2/ha; Brown 1981, Megonigal et. al. 1997, Jones 1981 Clawson et al. 1996) potentially as a result of the timing and extent of historical logging, which may have been greater in the large southeastern
49 floodplain forests that have been the focus of most ecological research. This may also be evident in observed Taxodium distchum dominance along the Silver River. Taxodium distchum is a commercially attractive wetland tree species and is very slow to reestablish after harvest ( Conner et al. 1986, Dunn and Shartiz 1987 ). Taxodium distchum accounts for 59% of basal area and 55% of leaf litterfall in the Silver River floodplain. The large basal area and dominance of Taxodium compare more favorably to alluvial swamps such as the Okefenokee Swamp and the Dismal Swamp (52 87 m2/ha of which 50 73 % is from Taxodium ) which are less a ccessible to harvest equipment and have a legacy of protection ( Dabel and Day 1977 Schlesinger 19 78). The large range in SOM content throughout the Silver River floodplain was consistent with other southeastern wetlands. The SOM content of the short hydroperiod sites was within the range of reported SOM content for floodplains (Bruland and Richardson 2006, Craft 2000 ). However, SOM content for the long hydroperiod sites was outside this range, but well within the reported range for nonriverine swamps (B ruland and Richardson 2006) While this is likely partially due to variation in hydroperiod, t his may also result from the lack of high flow velocitie s during l arge floods which limits flooding scour that may arise in other rivers In either case, SOM i n the Silver River floodplain is typical of southeastern wetlands and spans the range of reported values for both floodplain and nonriverine wetlands. Dual Control of Ecosystem Metrics Across all ecosystem attributes measured, my results lend strong sup port to the primary hypothesis of this study, and suggest that both hydroperiod and water level variation regulat e the structure and function of floodplain wetlands. Perhaps more
50 importantly, the results also indicate that hydroperiod and water level vari ation exert interdependent control on ecosystem structure and function. H ydroperiod consistently exerted a stronger effect than water level STD, supporting the use of hydroperiod as the primary metric of hydrology. However, water level STD explained a significant portion of the variation in ecosystem metrics especially within hydroperiod classes and clearly merits consideration. Both mean and variance appear ed to be important regulators of primary pr oduction, though the impact of water level variation wa s species dependent Aboveground net primary productivity estimates (from tree and palm litterfall) and LAI both decreased wit h increased hydroperiod, but were not significantly infl uenced by water level variation. Taxodium distichum leaf litterfall was po sitively correlated to water level variation at long hydroperiod sites; however, Taxodium distichum leaf litterfall per unit basal area was independent of hydrologic metrics. In contrast, leaf litterfall per unit basal area of all other tree species decreased with both hydroperiod and water level variation, as I hypothesized. This suggests that other variables contribute to the regulation of Taxodium distichum productivity (e.g., reduced compet itive stress from other taxa) and highlights the species adapta bility to a wide range of hydrologic conditions. Both mean and variance in water levels also regulated wetland organic matter dynamics. While my results support the hypothesis that SOM quantity increases with longer hydroperiod and decreases with greater water level variation, SOM quality only responded to hydroperiod. Numerous studies have investigated the effects of flooding and inundation on soil processes with varied findings ( Brinson et. al. 1981, Day 1983,
51 Shure and Gottschalk 1986, Lockaby 1996) However few studies have demonstrated a role of water level variation ( Reddy and Patrick 1975, Baker III et al. 2001, Battle and Golladay 2001) increases in which led to decreased SOM content While hydroperiod effect s were generally stronger water level STD explained most of the remaining SOM content variation. Water level variation may enhance decomposition, reducing SOM content either by promoting frequent switching from anaerobic to aerobic decomposition or by allowing more oxygen to enter the soil w hile still widely dispersing exoenzymes (Reddy and Patrick 1975). Due to very low flow velocities within the floodplain, scour of OM is unlikely to be significant. Hydrology appear s to exert less control on SOM recalcitrance than SOM content. Inferences about SOM recalcitrance are likely limited by low method resolution ; however, because labile OM is preferentially consumed, it seems unlikely that water level variation would impact SOM quantity without also influencing SOM quality. Soil elevation bimodality is a hallmark feature of patterned landscapes such as boreal bogs ( Foster et al. 1983, Charman 2002) and the ridgeslough mosaic of the Everglades ( Wu et al. 2006, Watts et al. 2010 ). However the absence of the predicted bimodal elevations suggests t hat feedbacks between hummock elevation and productivity are not as regular. Floodplain wetlands produce more woody debris than the marsh and moss dominated wetlands where bimodality is most often observed. It is possible that woody debris such as large br anches and fallen trees may be creating elevated microsites stochastically at a rate greater than self reinforcing processes (i.e., balance of SOM production and respiration) can regulate their abundance or height. However, the increase in soil elevation v ariation in response to increasing hydroperiod
52 and water level variation as well as the skewness of the distributions towards higher elevations may be diagnostic of low prevalence hummocks which form as a result of feedbacks between organic matter accumula tion, primary production an d reduced hydroperiod. While hydroperiod explains most of the total variation in soil elevation variance, water level variation explained a significant portion of t he variation in topography within hydroperiod classes Although I found no evidence of increased species richness with water level variation or soil elevation variation, other studies have observed that microtopography is an important factor in maintaining and enhancing biodiversity ( Beatty 1984, Titus 1990, Scarano et al. 1997, VivianSmith 1997, Pollock et al. 1998, Simmons et al. 2011, Washuta 2011 ) Cypress Knees Despite insufficient evidence to demonstrate that cypress knees provide direct benefits to the tree, my data clearly support the hypothesis that mean and va riation in cypress knee height is regulated by hydrology This in turn, strongly suggests a dynamic feedback between tree investments in cypress knee development and exogenous drivers I n ote in particular that across sites knees consistently grew to a h eight not inundated more than 5% of the time. The ongoing debate about the role of knees remains unresolved, but these data provide evidence that may help reject some prevailing hypotheses For example, Kummer et al. (1991) suggested that the primary func tion of cypress knees was to extract nutrients from stumps Brown and Montz (1986) suggested that knees may store starches Lamborn (1890) proposed that knees helped to stabilize cypress trees by penetrating the subsurface root mat. Finally Kramer et a l. (1952) suggested that cypress knees growth is triggered by inundation
53 stress. In this work, longer hydroperiod did not result in greater knee prevalence, but did affect knee height, which would suggest a feedback between environmental controls on inundation stress and tree investment in knee biomass. I posit that even modest gas exchange is sufficient to alleviate inundation stress creating a feedback that regulates cypress knee growth. Within the long hydroperiod sites, water level variation is strongly correlated with variation in cypress knee heights, suggesting that cypress invests in a range of knee heights that bal ance the increasing metabolic cost s of knees as they get taller with the relative frequency of high stage events; in short, cypress trees invest in a portfolio of knees that is commensurate with their hydrologic regime. These results support the contention in Kernell and Levy (1990) that cypress knees respond to hydrology and that the height of the tallest knees correlates to high water events Contingent Effects of Hydrology Many measured attributes of ecosystem structure and function exhibited greater sensitivity to water level variation w ithin a single hydroperiod class, suggesting strong interaction effects wherein the influence of hyd roperiod and water level variation are interdependent These interactions are difficult to identify due to the low power of the study, however, several significant results support this supposition. I observed strong effects of water level variation on leaf litterfall per unit basal area (decrease), wetland taxa prevalence ( decrease ), and SOM content ( decrease) at short hydroperiod sites. Similarly I observed strong effects of hydroperiod variation on basal area, but principally at short hydroperiod sites. T his likely arises because all of these functions are controlled by impacts of soil saturation on gas exchange w ith the atmosphere. At short hydroperiod sites mean water level s are typically below the soil surface. As a result,
54 variation in water level may result in greater change to the status of gas transfer. Likewise, even small differences in mean water level or hydroperiod among short hydroperiod sites may result in significantly greater gas exchange. In contrast water levels during the growing season in the long hydroperiod sites are generally above the soil surface Consequently, differences in mean water level are less likely to significantly expose the soil column to allow gas exchange, muting the impacts of both variation in mean water level acro ss long hydroperiod sites and the effects of water level variation within sites. The mean and variation in the height of cypress knees was more responsive to hydrology at the long hydroperiod sites. Cypress knees at the short hydroperiod site may not hav e been particularly sensitive to hydrology because soil saturation, not inundation depth, controls knee formation at short hydroperiod sites Therefore knees need only to br eak the soil surface at short hydroperiod sites to be useful to the tree Soil Org anic Matter Recalcitrance Peak I expected that SOM recalcitrance would increase with depth as decomposition incrementally consumes labile SOM deposited at the soil surface. This prediction assumes that the age of organic matter increases with depth in the soil column. However, in every soil core, SOM recalcitrance increased to a peak and then decreased indicating that labile OM is being introduced into the soil column throughout the soil profile This OM may be introduced by roots and root exudates. Further more, this suggests that below the SOM recalcitrance peak ( usually 2030 cm below the soil surface) the OM lability is conserved due to increased anoxia and thus reduced decomposition. This provides evidence that a periodically aerobic soil layer (acrotel m)
55 overlays a permanently anoxic soil layer (catotelm) throughout the floodplain and that the thickness and depth of the acrotelm is a function of mean water level Management Implications Wetlands are conferred special protections because of their signi ficant role in delivery of ecosystem services. Among the most important services are biodiversity carbon storage, and nutrient retention and transformations ( Gren et al. 1994), the loss of which has implications at multiple scales. This study show s that carbon storage is partially controlled by water level variation. The accumulation of organic matter also stores and retains nutrients. This study has also shown that microtopography is also influenced by water level variation and other studies have shown microtopography to enhancing biodiversity ( Beatty 1984, Titus 1990 Scarano et al. 1997, Pollock et al. 1998, V ivian Smith 1997, Simmons et al. 2011, Washuta 2011) While managing for water level variation is clearly important in maintaining ecosystem serv ices current regulatory approaches to managing environmental flows often fail to adequately consider hydrologic variation (though see Richter et al. 1997) The majority of regulatory methods for determining the maximum allowable withdraw al from a water bo dy are based on maintaining flows and level s which prevent significant harm to dependent water bodies and wetlands ; such is the case with the Minimum Flows and Levels (MFLs) mandate placed on Floridas environmental regulators by the 1972 Florida Water Res ources Act Minimum flows and level s are often quantified and represented largely by flow duration curves which relate flow /stage to the percentage of the time it is exceeded ( Beecher 1990, Gillilan and Brown 199 7, Tharme 2003, Neubauer et al. 2008) While this method acknowledges that there is no meaningful
56 minimum but rather a suitable regime, it does not allow for management of the duration and return intervals of flooding events, which are crucial components of the hydrologic regime and characterize w ater level variation ( Gordon et. al. 1991 Neubauer et al. 2008). M ore robust methods like those implemented by the St.Johns River Water Management District (SJRWMD) in north east Florida, define the duration and frequency of a handful of flooding events ( Neubauer et al. 2008). While this practice addresses the importance of variation, it only maintains a few hydrologic events rather than protect ing the water level variation regime As we seek to conserve our valuable ecosystems and the underlying drivers th at regulate their structure and function, we need to evaluate and manage for both aspects of the drivers, defining not only a characteristic mean, but also a characteristic variance.
57 APPENDIX A ADDITIONAL SITE CHARACTERISTICS Table A 1 Leaf litterfall indexed to species basal area. Leaf Litterfall per unit Basal Area (gm 2 yr 1 ) Study Site T1L T2L T3L T4L T5L T1S T2S T3S T4S T5S Total 8.9 4.9 13 11 4.1 9.1 12 9.5 8.1 12 Taxodium 6.1 4.4 12 20 4.6 6.8 10 9 5.4 18 Other 11 5.7 15 7.6 3.5 17 15 16 12 9 Table A 2 Additional site characteristics. Study Site Metric T1L T2L T3L T4L T5L T1S T2S T3S T4S T5S Soil Elevation Skew 0.65 0.35 0.65 0.17 1.57 0.95 0.59 0.11 0.35 0.21 Species Richness 20 10 7 11 10 19 9 18 10 6 Recalcitra nce Peak Depth (cm) 26 5.8 16 6.4 27 5.2 24 0.0 27 4.8 38 5.5 35 1.6 21 1.6 22 2.1 30 1.6
58 APPENDIX B BASAL AREA AND COVER BY SPECIES Table B 1. Long hydroperiod site percent cover by species. A zero indicates species presence without contribution to overall cover. Basal area by species (m2/ha) Table B 2. Short hydroperiod site percent cover by species. A zero indicates species presence without contribution to overall cover. Basal area by species (m2/ha) Site Acer rubrum Bidens sp. Boehmeria cylindrica Carpinus caroliniana Diodia virginiana Hydrocotyle bowlesioides Hymenocallis sp. Mikania scandens Panicum abscissum Pe rsicaria amphibia Phyla fruticosa Taxodium dichtum Toxicodendron radicans Ulmus americana Unknown grass Vitis rotundifolia T5S T4S 0 T3S 17 0 5 2 40 6 5 0 32 30 T2S 5 0 1 0 T1S 26 8 55 0 105 11 35 10 3 4 7 Site Lemna minor Panicum abscissum Ulmus americana Acer rubrum Pontederia cordata Boehmeria cylindrica Saururus cernuus Rhynchospora sp. Lobelia cardinalis Hydrocotyle americana Ludwigia sp. Persicaria amphibia Fraxinus pennsylvanica T5L 100 18 1 1 T4L 5 5 0 4 T3L 68 13 19 0 T2L 13 51 0 15 1 0 T1L 20 1 3 3 28 1 12 1
59 Table B 3 Basal are by species (m2/ha). Site Taxodium distichum Acer rubrum Frasinua pennsylvanica Ulmus americana Nyssa aquatica Sabal palmetto Persea palustris Quercus nigra Fraxinus americana Liquidambar styracuflua T1L 25.74 8.8 6 4.41 16.24 8.32 0.63 T2L 56.97 5.36 27.11 T3L 19.79 4.91 6.39 T4L 14.61 6.02 20.66 13.49 1.11 1.14 T5L 74.29 15.78 1.17 41.25 T1S 53.98 3.26 12.43 0.76 6.67 T2S 42.07 0.06 13.58 20.70 T3S 48.19 1.10 21.72 2.42 T4S 46.75 0.03 0.44 1.70 4.71 11.34 16.75 T5S 13.27 27.58 0.21 31.78
60 APPENDIX C ANNUAL LITTERFALL Figure C 1 Annual litterfall for four study sites in the Silver River floodplain. A) The most upstream long hydroperiod s ite. B) The most upstream short hydroperiod site. C) The most downstream long hydroperiod site. D) The most downstream short hydroperiod site.
61 APPENDIX D SOIL CORE DATA Table D 1. Soil core data. Mean water level and hydroperiod are calculated relat ive to the elevation of the core top. Soil Core Organic Matter Fraction Recalcitrance Index Depth to Recalcitrance Peak (cm) Mean Water Level Hydroperiod T1L 1 0.30 0.36 17 22 0.21 T1L 2 0.26 0.42 30 21 0.22 T1L 3 0.27 0.39 29 20 0.24 T2L 1 0.53 0. 29 25 7 0.59 T2L 2 0.56 0.25 11 10 0.64 T2L 3 0.61 0.25 11 13 0.67 T3L 1 0.36 0.27 24 4 0.51 T3L 2 0.32 0.26 34 4 0.41 T3L 3 0.32 0.27 22 5 0.40 T4L 1 0.42 0.32 24 0 0.45 T4L 2 0.46 0.31 24 16 0.66 T5L 1 0.40 0.31 21 8 0.53 T5L 2 0.38 0.29 30 11 0.57 T5L 3 0.25 0.33 32 10 0.56 T1S 1 0.11 0.35 36 50 0.03 T1S 2 0.13 0.33 33 47 0.04 T1S 3 0.16 0.36 45 46 0.04 T2S 1 0.09 0.37 33 33 0.13 T2S 2 0.15 0.37 37 29 0.16 T2S 3 0.12 0.39 35 30 0.15 T3S 1 0.07 0.35 21 33 0.15 T3S 2 0.13 0.28 23 24 0.21 T3S 3 0.12 0.31 19 25 0.21 T4S 1 0.06 0.46 24 38 0.14 T4S 2 0.04 0.49 20 52 0.08 T5S 1 0.06 0.41 27 28 0.22 T5S 2 0.07 0.40 31 40 0.16 T5S 3 0.07 0.42 31 36 0.18
62 Organic Matter Fraction 0.0 0.1 0.2 0.3 0.4 0.5 0.6Depth (cm) 0 10 20 30 40 Recalcitrance Index 0.0 0.1 0.2 0.3 0.4 0.5 0.6 Labile OM Recalcitrant OM Recalcitrance Index T2L-1Organic Matter Fraction 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0 10 20 30 40 Recalcitrance Index 0.0 0.1 0.2 0.3 0.4 0.5 T3L-1Organic Matter Fraction 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0 10 20 30 40 Recalcitrance Index 0.0 0.1 0.1 0.2 0.2 0.3 0.3 0.4 T4L-3Organic Matter Fraction 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7Depth (cm) 0 10 20 30 40 Recalcitrance Index 0.0 0.1 0.2 0.3 0.4 0.5 T5L-1Organic Matter Fraction 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0 10 20 30 40 Recalcitrance Index 0.0 0.1 0.2 0.3 0.4 0.5 T1L-1 Figure D 1. Soil core profiles for long hydroperiod sites. One soil core profile is shown for each site. Profiles show the organic matter fraction split between the labile and recalcitrant fractions as well as the recalcitrance index.
6 3 T1S-2Organic Matter Fraction 0.00 0.05 0.10 0.15 0.20 0.25Depth (cm) 0 10 20 30 40 Recalcitrance Index 0.0 0.1 0.2 0.3 0.4 0.5 Labile OM Recalcitrant OM Recalcitrance Index T2S-1Organic Matter Fraction 0.00 0.02 0.04 0.06 0.08 0.10 0.12 0.14 0.16 0.18 0 10 20 30 40 Recalcitrance Index 0.0 0.1 0.2 0.3 0.4 0.5 T3S-2Organic Matter Fraction 0.00 0.05 0.10 0.15 0.20 0.25 0.30 0 10 20 30 40 Recalcitrance Index 0.0 0.1 0.2 0.3 0.4 T4S-2Organic Matter Fraction 0.00 0.02 0.04 0.06 0.08Depth (cm) 0 10 20 30 40 Recalcitrance Index 0.0 0.1 0.2 0.3 0.4 0.5 0.6 T5S-1Organic Matter Fraction 0.00 0.02 0.04 0.06 0.08 0.10 0.12 0 10 20 30 40 Recalcitrance Index 0.0 0.1 0.2 0.3 0.4 0.5 0.6 Figu re D 2. Soil core profiles for short hydroperiod sites. One soil core profile is shown for each site. Profiles show the organic matter fraction split between the labile and recalcitrant fractions as well as the recalcitrance index.
64 LIST OF REFERENCE AngehrnBettinazzi, C., P. Lscher, et al. (1988). "Thermogravimetry as a Method for Distinguishing Various Degrees of Mineralisation in Macromorphologically defined Humus horizons." Zeitschrift fr Pflanzenernhrung und Bodenkunde 151 (3): 177183. Baker III T. T., B. G. Lockaby, et al. (2001). "Leaf Litter Decomposition and Nutrient Dynamics in Four Southern Forested Floodplain Communities." Soil Sci. Soc. Am. J. 65(4): 13341347. Battle, J. M. and S. W. Golladay (2001). "Hydroperiod Influence on Breakdown of Leaf Litter in Cypress Gum Wetlands." American Midland Naturalist 146(1): 128145. Beecher H A (1990) Standards for instream flows. Rivers 1 (2):97 109 Beatty, S. W. (1984). "Influence of Microtopography and Canopy Species on Spatial Patterns o f Forest Understory Plants." Ecology 65(5): 14061419. Belyea, L. R. and R. S. Clymo (2001). "Feedback control of the rate of peat formation." Proceedings of the Royal Society of London. Series B: Biological Sciences 268(1473): 13151321. Benedetti Cecch i, L. (2003). The Importance of the Variance Around the Mean Effect Size o f Ecological Processes ." Ecology 84(9): 23352346. BondLamberty, B. and A. Thomson (2010). "Temperatureassociated increases in the global soil respiration record." Nature 464 (728 8): 579582. Bray, J. R. and E. Gorham (1964). Litter Production in Forests of the World. Advances in Ecological Research J. B. Cragg, Academic Press. Volume 2: 101 157. Briand, C. H. (2001). "Cypress knees: an enduring enigma." Arnoldia (Jamaica Plain) 60(4): 19 25. Brinson, M. M, A. E. Lugo, and S. Brown. ( 1981) Primary productivity, decomposition and consumer activity in freshwater wetlands. Annual Review of Ecology and Systematics 12:123161. Brown, S. ( 1981) A comparison of the structure, pri mary productivity, and transpiration of cypress ecosystems in Florida. Ecological Monographs 51:403427. Brown, C. A. and G. N. Montz (1986). Baldcypress : the tree unique, the wood eternal Baton Rouge, Claitor's Pub. Division.
65 Bruland, G. and C. Richardson (2006). "Comparison of Soil Organic Matter in Created, Restored and Paired Natural Wetlands in North Carolina." Wetlands Ecology and Management 14(3): 245 251. Charman, D. J. (2002). Peatlands and environmental change. Chichester, West Sussex, Engla nd; New York, J. Wiley. Clawson, R. G., B. G. Lockaby, et al. (2001). "Changes in Production and Nutrient Cycling across a Wetness Gradient within a Floodplain Forest." Ecosystems 4 (2): 126 138. Connell, J.H. ( 1976) Diversity in tropical rain forest an d coral reefs. Science 199 : 13021310. Conner, W. H., and J. W. Day, Jr. ( 1976) Productivity and composition of a baldcypress water tupelo site and a bottomland hardwood site in a Louisiana swamp. American Journal of Botany 63 :13541364. Conner, W. H., et al. (1986). "Natural regeneration of baldcypress (Taxodium distichum (L.) Rich.) in a Louisiana swamp." Forest Ecology and Management 14(4): 305 317. Conner, W. H ., and J. W. Day, Jr. ( 1992) Water level variability and litterfall productivity of forested freshwater wetlands in Louisiana. American Midland Naturalist 128:237245. Coutts, M. and W. Armstrong (1976). "Role of oxygen transport in the tolerance of tr ees to waterlogging." Tree physiology and yield improvement : 361 385. Craft, C. B. and W. P. Casey (2000). Sediment and nutrient accumulation in floodplain and depressional freshwater wetlands of Georgia, USA ." Wetlands 20 (2): 323 332. Cronk, J. K. and S. Fennessy (2001). Wetland Plants: Biology and Ecology Taylor & Francis. Cuffney, T. F. (1988). "Input, movement and exchange of organic matter within a subtropical coastal black water river flood plain system." Freshwater Biology 19(3): 305 320. Dabel C. V. and F. P. Day, Jr. (1977). "Structural Comparisons of Four Plant Communities in the Great Dismal Swamp, Virginia." Bulletin of the Torrey Botanical Club 104 (4): 352 360. Day, F. P., Jr. (1983). "Effects of Flooding on Leaf Litter Decomposition in Microcosms." Oecologia 56 (2/3): 180 184.
66 Dunn, C. and R. Sharitz (1987). "Revegetation of a Taxodium Nyssa forested wetland following complete vegetation destruction." Vegetatio 72(3): 151157. Eppinga, M. B., et al. (2009). "Nutrients and Hydrology Indicate the Driving Mechanisms of Peatland Surface Patterning." The American Naturalist 173 (6): 803 818. Foster, D. R., G. A. King, et al. (1983). "Origin o f string patterns in boreal peatlands." Nature 306 (5940): 256 258. Gillilan, D. M. and T. C. Brown (1997). Instream Flow Protection: Seeking A Balance In Western Water Use, Island Press. Gholz, H. L. (1982). "Environmental Limits on Aboveground Net Primary Production, Leaf Area, and Biomass in Vegetation Zones of the Pacific Northwest." Ecology 63(2): 469481. Gosselink, J. et al. ( 1981 ) Ecological factors in the determination of riparian wetland boundaries. Wetlands of bottomland hardwood forests in J. R. Clark and J.Benforado, editors. Elsevier Scientific, New York, New York, p 197219 Gosselink, J. and R. Turner (1978). "The role of hydrology in freshwater wetland ecosystems." Freshwater Wetlands: Ecological Processes and Management Potential 197 8, R. E. Good, D. F. Whigham and R. L. Simpson, editors, Academic Press, New York, p 6378. 6 fig, 3 tab, 36 ref. Gordon, N. D., et al. (1991). Stream hydrology: an introduction for ecologists John Wiley and Sons Ltd. Gren, I.M., et al. (1994). "Primar y and secondary values of wetland ecosystems." Environmental and Resource Economics 4 (1): 55 74. Hacker, S. D. and M. D. Bertness (1999). Experimental Evidence f or Factors Maintaining Plant Species Diversity i n a New England Salt Marsh." Ecology 80(6): 2 064 2073. Heffernan, J. B. (2008). Wetlands a s a n Alternative Stable State i n Desert Streams ." Ecology 89(5): 12611271. Hilbert, D. W., et al. (2000). "Modelling and Analysis of Peatlands as Dynamical Systems." Journal of Ecology 88(2): 230 242. Jackson, M. B. and T. D. Colmer (2005). "Response and Adaptation by Plants to Flooding Stress." Annals of Botany 96(4): 501 505. Jawitz, J. W. and J. Mitchell (2011). "Temporal inequality in catchment discharge and solute export." Water Resources Research 47 (1 0): W00J14.
67 Jones, R. H. (1981). A classification of lowland forests in the northern coastal plain of South Carolina (Doctoral dissertation, Clemson University). Johnson, F. L. and D. T. Bell (1976). "Plant Biomass and Net Primary Production along a Fl oodFrequency Gradient in the Streamside Forest." Castanea 41 (2): 156 165. Keeley, J. E. (1979). "Population Differentiation along a Flood Frequency Gradient: Physiological Adaptations to Flooding in Nyssa sylvatica." Ecological Monographs 49(1): 89 108. Kernell, J. L. and G. F. Levy (1990). "The Relationship of Bald Cypress [Taxodium distichum (L.) Richard] Knee Height to Water Depth." Castanea 55(4): 217 222. Kozlowski, T. T. (1984). "Plant Responses to Flooding of Soil." BioScience 34(3): 162 167. Ko zlowski, T. T. (1997). "Responses of woody plants to flooding and salinity." Tree Physiology 17(7): 490. Kramer, P. J., W. S. Riley, et al. (1952). "Gas Exchange of Cypress Knees." Ecology 33(1): 117 121. Kristensen, E. (1990). "Characterization of bioge nic organic matter by stepwise thermogravimetry (STG)." Biogeochemistry 9 (2): 135 159. Kristensen, E. and F. Andersen (1987). "Determination of organic carbon in marine sediments: a comparison of two CHN analyzer methods." Journal of Experimental Marin e Biology and Ecology 109(1): 15 23. Kummer, H., M. Brandt, et al. (1991). "Nutritional exploitation of dead trunks: another function of cypress knees (Taxodium distichum)?" Trees 5 (2): 122123. L amborn R. H. (1890). The knees of the bald cypress tree: a new theory of their function. Science ns 15(365): 6567. Lichvar, R., E. United States. Army. Corps of, et al. (2012). The National Wetland Plant List Hanover, NH, US Army Engineer Research and Development Center, Cold Regions Research and Engineering Laboratory. Landres, P., et al. (1999). Overview of the use of natural variability concepts in managing ecological systems. 9: 1179 1188. Lockaby, B. G., et al. (1996). "Hydroperiod Influences on Nutrient Dynamics in Decomposing Litter of a Floodplain Forest." Soil Sci. Soc. Am. J. 60(4): 12671272.
68 Lopez Capel, E., S. P. Sohi, et al. (2005). Use o f Thermogravimetry Differential Scanning Calorimetry t o Characterize Modelable Soil Organic Matter Fractions ." Soil Sci. Soc. Am. J. 69(1): 136140. Lugo, A. E. (1978). Stress and ecosystems Energy and Environmental Stress in Aquatic Systems, Selected Papers from a Symposium held at Augusta, Georgia November 24, 1977. CONF 771114. p 62101, 1978, 8 fig, 5 tab, 114 ref. Megonigal, J. P., W. H. Conner, et al. (1997). "Aboveground Production in Southeastern Floodplain Forests: A Test of the Subsidy Stress Hypothesis." Ecology 78(2): 370 384. Mitsch, W. J., and K. C. Ewel. ( 1979) Comparative biomass and growth of cypress in Florida wetlands. American Midla nd Naturalist 101 :417 426. Neubauer, C., G. Hall, et al. (2008). "Minimum Flows and Levels Method of the St. Johns River Water Management District, Florida, USA." Environmental Management 42(6): 11011114. Poff, N. L. and J. V. Ward (1989). "Implications of Streamflow Variability and Predictability for Lotic Community Structure: A Regional Analysis of Streamflow Patterns." Canadian Journal of Fisheries and Aquatic Sciences 46(10): 18051818. Pollock, M. M., R. J. Naiman, et al. (1998). Plant Species Ric hness i n Riparian Wetlands A Test Of Biodiversity Theory ." Ecology 79(1): 94 105. Porporato, A., P. DOdorico, et al. (2002). "Ecohydrology of water controlled ecosystems." Advances in Water Resources 25(8 12): 13351348. Porporato, A., E. Daly, et al. ( 2004). "Soil Water Balance and Ecosystem Response to Climate Change." The American Naturalist 164 (5): 625 632. Odum, W. E., E. P. Odum, et al. (1995). "Nature's Pulsing Paradigm." Estuaries 18 (4): 547555. Reddy, K. R. and R. D. DeLaune (2008). Biogeoche mistry of wetlands: science and applications CRC PressI Llc. Reddy, K. R. and W. H. Patrick (1975). "Effect of alternate aerobic and anaerobic conditions on redox potential, organic matter decomposition and nitrogen loss in a flooded soil." Soil Biology and Biochemistry 7 (2): 87 94. Richter BD, Baumgartner JV, Braun DP (1997) How much water does a river need? Freshwater Biology 37:231 249 Sahrawat, K. L. (2003). Organic matter accumulation in submerged soils. Advances in Agronomy Academic Press. Volume 81: 169 201.
69 Schnitzer, M. and I. Hoffman (1966). "A Thermogravimetric Approach To The Classification Of Organic Soils." Soil Sci. Soc. Am. J. 30(1): 63 66. Schuur, E. and P. Matson (2001). "Net primary productivity and nutrient cycling across a mesic to wet precipitation gradient in Hawaiian montane forest." Oecologia 128(3): 431442. Schlesinger, W. H. (1978). "Community Structure, Dynamics and Nutrient Cycling in the Okefenokee Cypress SwampForest." Ecological Monographs 48 (1): 43 65. Sharma, H. S. S. (1989). "Thermogravimetric analysis of flax shives degraded by pleurotus ostreatus and ceraceomyces sublaevis." Thermochimica Acta 138 (2): 347 357. Shure, D. J., M. R. Gottschalk, et al. (1986). "Litter Decomposition Processes in a Floodplain Forest ." American Midland Naturalist 115 (2): 314 327. Scarano, F. R., et al. (1997). "Plant establishment on flooded and unflooded patches of a freshwater swamp forest in southeastern Brazil." Journal of Tropical Ecology 13(06): 793803. Simmons, M. E., X. Ben Wu, et al. (2011). "Plant and Soil Responses to Created Microtopography and Soil Treatments in Bottomland Hardwood Forest Restoration." Restoration Ecology 19 (1): 136 146. Sousa, W. P. (1979). "Experimental Investigations of Disturbance and Ecological Su ccession in a Rocky Intertidal Algal Community." Ecological Monographs 49 (3): 227 254. Taylor, J. R., et al. (1990). "Bottomland hardwood forests: their functions and values." Ecological processes and cumulative impacts: Illustrated by bottomland hardwood wetland ecosystems : 13 86. Tharme, R E (2003) A global perspective on environmental flow assessment: emerging trends in the development and application of environmental flow methodologies for rivers. River Research and Applications 19:397 441 Titus, J. H. (1990). "Microtopography and Woody Plant Regeneration in a Hardwood Floodplain Swamp in Florida." Bulletin of the Torrey Botanical Club 117(4): 429437. Vivian Smith G. ( 1997) Microtopographic Heterogeneity and Floristic Diversity in Experimental Wetland Communities. Journal of Ecology 85(1): 71 82.
70 Ward, J and J. Stanford (1983) Intermediate Disturbance Hypothesis: An Explanation for Biotic Diversity Patterns in Lotic Ecosystems Dynamics of Lotic Systems 347 356 p, 2 Watts, D., M. Cohen, et al. (2010). "Hydrologic Modification and the Loss of Self organized Patterning in the Ridge Slough Mosaic of the Everglades." Ecosystems 13(6): 813 827. Washuta, A. B. (2011). Microtopographical effects of clearcut ting of Taxodium Nyssa swamps. Gainesv ille, Fla, University of Florida. Wu, Y., N. Wang, et al. (2006). "An analysis of spatial complexity of ridge and slough patterns in the Everglades ecosystem." Ecological Complexity 3 (3): 183192.
71 BIOGRAPHICAL SKETCH Joseph Delesantro received his bachelors degree in environmental engineering from the University of Florida in 2010. He remained at the University of Florida moving over to the School of Forest Resources and Conservation to pursue his masters degree, c ompleted in 2013. He plans to find employment in the field of environmental hydrology and work towards the sustainable use of water resources.