Litter Production and Decomposition in Three Conservation Area Marshes

Material Information

Litter Production and Decomposition in Three Conservation Area Marshes
Duffy, Shannon L
Place of Publication:
[Gainesville, Fla.]
University of Florida
Publication Date:
Physical Description:
1 online resource (88 p.)

Thesis/Dissertation Information

Master's ( M.S.)
Degree Grantor:
University of Florida
Degree Disciplines:
Soil and Water Science
Committee Chair:
Committee Co-Chair:
Committee Members:
Graduation Date:


Subjects / Keywords:
Biomass ( jstor )
Everglades ( jstor )
Leaves ( jstor )
Lignin ( jstor )
Marshes ( jstor )
Nutrients ( jstor )
Species ( jstor )
Surface water ( jstor )
Vegetation ( jstor )
Wetlands ( jstor )
Soil and Water Science -- Dissertations, Academic -- UF
accretion -- carbon -- nitrogen -- phosphorus -- wetland
The Everglades ( local )
bibliography ( marcgt )
theses ( marcgt )
government publication (state, provincial, terriorial, dependent) ( marcgt )
born-digital ( sobekcm )
Electronic Thesis or Dissertation
Soil and Water Science thesis, M.S.


The Upper St. Johns River Basin (USJRB) is one of ten major watersheds within the St. Johns River Water Management District. Appropriate understanding and management of wetlands in the USJRB has direct implications on soil subsidence and is vital to achieving the core missions of flood protection, water supply, water quality, and protection of natural systems. Net litter production and decomposition was estimated in three USJRB conservation area marshes using clip plots (n=3). The roles of species community and site characteristics in determining decomposition of litter were examined. Sites represented varied hydrologic disturbance. St. Johns Marsh Conservation Area (SJMCA) is highly altered, Blue Cypress Marsh Conservation Area (BCMCA) is intermediately altered, and Fort Drum Marsh Conservation Area (FDMCA) represents historical conditions. Litter material was characterized by fiber quality and nutrient content over the course of one year of decomposition in the field. Results show quality of litter was correlated with decomposition rate for low quality material (%lignin >40), while hydrology may by more influential for higher quality material. Using annual litter production estimates in conjunction with observed decomposition rates, annual contribution to soil elevation was found to be 0.2 +/- 0.06 cm in BCMCA, 0.18 +/- 0.07 cm in FDMCA, and 0.06 +/- 0.08 cm in SJMCA. Subsidence has resulted in the loss of organic soils and is not quickly reversible due to the slow rate of accretion. ( en )
General Note:
In the series University of Florida Digital Collections.
General Note:
Includes vita.
Includes bibliographical references.
Source of Description:
Description based on online resource; title from PDF title page.
Source of Description:
This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Thesis (M.S.)--University of Florida, 2014.
Co-adviser: CLARK,MARK W.
Statement of Responsibility:
by Shannon L Duffy.

Record Information

Source Institution:
Rights Management:
Applicable rights reserved.
LD1780 2014 ( lcc )


This item has the following downloads:

Full Text


Aquatic Botany, 40 ( 1991 ) 203-224 203 Elsevier Science Publishers B.V., Amsterdam Growth, decomposition, and nutrient retention of Cladium jamaicense Crantz and Typha domingensis Pers. in the Florida Everglades Steven M. Davis South Florida Water Management District, P.O. Box 24680, West Palm Beach, FL 33416-4680, USA (Accepted for publication 7 January 1991 ) ABSTRACT Davis, S.M., 1991. Growth, decomposition, and nutrient retention of Cladium jamaicense Crantz and Typha domingensis Pers. in the Florida Everglades. Aquat. Bot.,40: 203-224. Estimates of phosphorus (P) and nitrogen (N) gains and losses during annual macrophyte growth, death and 2 years decomposition were made along a gradient of surface water nutrient concentrations in the Florida Everglades. Annual rates of P and N allocation to growing leaves, translocation or leaching from dying leaves, and retention in dead leaves of Cladiumjamaicense Crantz and Typha domingensis Pers. were correlated to soluble reactive P and nitrate concentrations in surface water. Rates of each of these processes were higher in T. domingensis than in C. jamaicense, Cladium jamaicense rates increased linearly along the nutrient gradients, but did not fluctuate with yearly variations in soluble reactive P or nitrate concentrations. For T. domingensis, annual rates were strongly correlated with mean annual soluble reactive P and nitrate concentrations during specific sampling years. Responses of C. jamaicense to the nutrient gradient were characteristic of species competitive in an infertile habitat, while responses of 72 domingensis were characteristic of species competitive in a fertile habitat. The main effect of P and N enrichment on leaf nutrient flux was to accelerate translocation or leaching from dying tissue, rather than to increase retention in standing dead leaves. Freshly dead leaves retained only slightly greater quantities of P and N under enriched conditions in comparison to background conditions. After 2 years of decomposition, approximately half of the leaf litter mass remained intact. Increasing P and N concentrations in decomposing leaf litter resulted in net uptake or retention of these elements after 2 years despite decreasing litter mass. The total amounts of P and N that were sequestered annually by T. domingensis after processes of leaf production, mortality and 2 years decomposition were lowest under non-enriched conditions and reached a maximum under a moderate level of enrichment. Wetland ecosystems such as the Everglades, which developed under conditions of low nutrient supply, may offer a finite potential for accelerated nutrient retention when the exogenous nutrient supply increases as a result of human activities. However, a plant species such as C. jamaicense, that is adapted to a low-nutrient environment, may have a low nutrient threshold before it loses its competitive capability and its habitat is invaded by a species such as T. domingensis that is better adapted to a highnutrient environment. 0304-3770/91/$03.50 © 1991 -Elsevier Science Publishers B.V.


204 S.M. DAVIS INTRODUCTION The use of wetlands to remove anthropogenic nutrient inputs from surface water has gained popularity since Boyd (1970) and Steward (1970) noted the nutrient uptake potentials of various macrophyte species. The role of plants in wetland nutrient uptake, however, has remained poorly understood until recently. Organic matter accumulation, plus soil adsorption, appear to be two major processes controlling long-term nutrient immobilization in wetlands (Richardson and Marshall, 1986). Wetland nitrogen (N) budgets are also influenced by denitrification and nitrogen fixation. The role of plants in longterm nutrient retention appears to be largely related to detritus production (Davis and van der Valk, 1978) and the resulting accumulation of organic matter. At an ecosystem level, wetland plants appear to affect nutrient budgets to the extent that they ( 1 ) accumulate nutrients in biomass as they grow, (2) retain nutrients in biomass as they die and (3) retain or accumulate nutrients as they decompose in conjunction with the accretion of organic sediment. In this context, the effectiveness of wetland plants in retaining nutrients appears to be limited and to vary from one ecosystem to another. Senescing plant tissues release nutrients which they accumulated during growth (Davis and van der Valk, 1983; Hopkinson and Schubauer, 1984). Leaching of soluble organic and inorganic compounds during the first few weeks of decomposition results in further loss of nutrients from dead plant tissue (Howard-Williams and Howard-Williams, 1978; Webster and Benfield, 1986). Detritus may continue to lose phosphorus (P) and N after the initial leaching period (Latter and Cragg, 1967; Davis and van der Valk, 1978). In other cases, macrophyte detritus either retains its initial nutrient content or accumulates nutrients during 1-2 years of decomposition (Puriveth, 1980; Day, 1982; Davis and van der Valk, 1983), through the accumulation of microbial protein (Webster and Benfield, 1986 ) or humic N accumulation (Rice, 1982). Inconsistent findings concerning the role of macrophyte detritus in nutrient retention may result from variables such as nutrient supply (Howarth and Fisher, 1976; Saunders, 1976; Almazon and Boyd, 1978; Elwood et al., 1981 ), seasonality in temperature and flooding (Brinson, 1977; Puriveth, 1980; Day, 1982), and plant species (Day, 1982). There is little information concerning the capacity of macrophyte detritus to retain larger quantities of P and N with nutrient enrichment. This study presents a budget for nutrient retention through macrophyte growth, death and 2 years decomposition at eutrophic, transitional and oligotrophic sites along a gradient of surface water nutrient concentrations in the Florida Everglades. Net nutrient gains and losses are quantified in the process whereby Cladium jamaicense Crantz and Typha dorningensis Pets. leaf production leads to detritus accumulation and nutrient retention. Leaf production values of C. jamaicense and T. domingensis (Davis, 1990) are


CLADIUMJAMAICENSE AND TYPHA DOMINGENSIS IN FLORIDA EVERGLADES 205 combined with tissue P and N concentrations to estimate annual allocation to growing leaves and annual retention in standing dead leaves. The P and N content of C. jamaicense and T. domingensis leaf detritus is followed from the time the leaves die through 2 years of decomposition to estimate the detritus decomposition rate and the release or accumulation of P and N during decomposition. The effectiveness of Everglades C. jamaicense and T. domingensis in intercepting P and N is evaluated. Both above-ground and below-ground plant production contribute to nutrient immobilization in organic detritus in wetlands (Richardson and Marshall, 1986 ). In the Florida Everglades, Toth ( 1987, 1988 ) demonstrated that individual C. jamaicense and T. domingensis plants accumulate and hold P and N primarily through leaf production and above-ground detrital accumulation. Long-term nutrient retention by these species, through growth and decomposition of below-ground organs, amounts to less than 20% of the retention associated with leaf production and decomposition. These species thus appear to sequester nutrients primarily through above-ground production and accumulation of leaf litter in the Everglades. Soil water, rather than surface water, probably provides the major source of P and N nutrition for C. jamaicense and T. domingensis, but surface water is the source of anthropogenic nutrient enrichment to both soil and plants in the Everglades. Both surface water and soil concentrations are elevated in the Everglades near inflows of nutrient-enriched agricultural water (Davis, 1990). Anoxic conditions at nutrient-enriched Everglades sites (Reeder and Davis, 1983 ) may also influence nutrient solubility and availability in the substratum (Reddy, 1983; Reddy and Rao, 1983). Consequently, surface water P and N concentrations are used as correlates, rather than direct measures of nutrient enrichment in this study. STUDY AREA The Everglades (Fig. 1 ) represent an oligotrophic wetland ecosystem which has received increased nutrient supplies for nearly 30 years as a result of water management practices. This subtropical, freshwater peatland historically occupied an approximately 1 000 000 ha basin which received water and nutrients mostly from direct rainfall (Davis, 1943; Parker, 1974). As a result, nutrients probably were in limited supply (Steward and Ornes, 1975 ). Most of the remaining Everglades are contained within approximately 500 000 ha of Water Conservation Areas and Everglades National Park. The nutrient supply into the Water Conservation Areas has increased due to pumped inflows of run-off water from drained lands in the Everglades agricultural area. Vegetation change has accompanied the increased nutrient inputs in the Everglades. Vast, nearly monospecific stands of C. jamaicense cover 65-70% of the Everglades marsh (Loveless, 1959 ). The dominance of this large sedge in


LAKE t '~"~ ~ PALM f I ~~BEACH EVERGLADES "~ / ~ "~}! AGRICULTURAL (WATER\ ~ ICULTURAL ~. -" t/ AREA ~ CONSERVATION ~, \ ~ ~ AREA / \ \ /"~ ~ , \ ~, ,WAT E R ~---/J I' "~ r~_~CONSERVATION ~l CJ I \ • " "~AREA 7~ ,,' CONSERVATION~J ~'~J-~ LAUDERDALE AREA " -~ \ \ I il / O EVERGLADES ~ ,,' '~ NATIONAL q > ,~ ~=~00 PARK l ~ ~\ ~__ -L .... ~ ~r ~\ ~;' Florida 0 2 4 6 KILOMETERS E S,TE f 31TE B 0 ~ITE C 0 SITE D BIG RUBBER ~TR EE ISLAND ~ OLD GLORY ISLAND 5.~'t I WATER CONSERVATION AREA 2A Zig. 1. Location of the Florida Everglades, Water Conservation Area 2A and vegetation sample sites.


TABLE 1 Gradients of surface water P and N concentrations (mg l- ’ ) during 19761980. Values represent means and ranges; N= 65 Total P Soluble reactive P Inflows S1 OD 0.097 (0.018-0.564) 0.062 (0.002-0.468) Site A’ 0.109 (0.016-0.490) 0.065 (0.002-0.391) Site B 0.076 (0.014-0.229) 0.03 1 (0.002-O. 167) Site C 0.024 (0.008-0.067) 0.005 (0.002-0.033) Site D 0.008 (0.002-0.038) 0.003 (0.002-0.014) ‘Site A values represent only 1979 and 1980 (N=26). Total N NOx-N NH,-N 4.35 (1.40-12.78) 1.038 (0.004-8.059) 0.34 (0.01-1.02) 4.25 (1.24-7.42) 0.034 (0.004-0.408) 0.23 (0.01-1.49) 3.58 (1.06-9.62) 0.014 (0.004-0.028) 0.26 (0.01-6.66) 3.37 (1.44-7.49) 0.012 (0.004-0.154) 0.03 (0.01-0.17) 2.89 (1.13-5.19) 0.009 (0.004-0.158) 0.05 (0.01-1.14)


208 S.M. DAVIS the Everglades has been attributed to its low nutrient requirements (Steward and Ornes, 1975). Since nutrient supply has increased, T. domingensis has invaded where agricultural inflows enter the marsh (Davis, 1990). The middle of the Water Conservation Areas, designated Water Conservation Area 2A (Fig. 1 ), receives particularly large inflows of agricultural water and nutrients because of a convergence of canal systems on the inflow gates at the north end of the area. A nearly monospecific T. domingensis stand covers 2400 ha below the inflows in Water Conservation Area 2A. Scattered T. domingensis have permeated the sawgrass marsh to the south of this stand during the past decade. Surface water P and N concentrations decline as inflow water flows southward across the marsh, creating a gradient from high concentrations near the inflow structures to values approaching detection limits in the interior marsh (South Florida Water Management District, 1991 ) (Table 1 ). Declining soluble reactive phosphorus (SRP) and nitrate (NO3N) concentrations account for most of the falls in total P and total N along the gradient. The gradient is characterized by high temporal variability in surface water P and N concentrations at eutrophic sites in comparison to lower variability at transitional and oligotrophic sites (Table 1 ). The lower ends of the ranges of surface water concentrations are similar for all sites. Wide fluctuations in concentrations above lower levels, in comparison to consistency in concentrations near lower levels, differentiate eutrophic sites from oligotrophic sites along the gradient. The P and N gradient is further characterized by year-to-year fluctuations in surface water nutrient concentrations, depending on annual variations in water and nutrient inputs into the marsh. The marsh remained flooded year round throughout this study owing to a prescribed water regulation schedule for the area (Davis, 1990). METHODS Leaf biomass turnover and nutrient flux Leaf production and nutrient flux were estimated in C. jamaicense and T. domingensis stands at four sites along the nutrient gradient in Water Conservation Area 2A (Fig. 1 ). These sites were located on a line extending approximately south-southwest into the marsh, at distances of 0.8, 1.6, 3.2 and 6.4 km from the north levee. Estimates were made for two sampling years. Sites B, C and D were sampled during April 1975-April 1976, while Sites A, B and D were sampled during April 1979-April 1980. This combination of three sites per year during two sampling years yielded six estimates of leaf production and nutrient accumulation for each plant species. Methods for estimating annual leaf production were detailed by Davis (1990). Two sampling techniques were employed. The first technique involved quadrat sampling to determine mean annual live leaf biomass. Five


CLAD1UM JAMAICENSE AND TYPHA DOMINGENSIS IN FLORIDA EVERGLADES 209 replicate 0.5 m 2 quadrats were collected monthly from C. jamaicense and T. domingensis stands during the first sampling year. Sampling frequency was reduced to bimonthly during the second year, except at Site A where samples were collected monthly. Living leaves within quadrats were weighed after oven-drying for 72 h at 90°C. The second sampling technique estimated annual leaf turnover rate. At least five newly emerged plants of each species were tagged at each site during each vegetation sampling year. Leaf lengths of each plant were measured monthly throughout the life of the plants. Cumulative life-time leaf growth of each tagged plant was divided by years longevity to estimate annual growth. Annual growth of each plant was divided by mean leaf biomass during its life span to calculate an annual turnover rate of leaf biomass. The mean leaf turnover rate for each stand was multiplied by mean annual leaf biomass, as determined by quadrat sampling, to estimate annual leaf production. Nutrient allocation to growing leaf biomass and release from dying leaf biomass were estimated by combining annual production estimates (Davis, 1990 ) with tissue P and N concentrations. Living leaves collected in biomass samples were analyzed for tissue P and N concentration. Intact standing dead leaves attached to living plants were also collected from quadrats for nutrient analysis, although biomass was not measured. Annual nutrient allocation to growing leaves was estimated by multiplying annual leaf production by mean annual tissue P and N concentrations in living leaves. Annual nutrient retention in dead leaves was estimated by multiplying annual leaf production by mean annual tissue P and N concentrations in freshly dead leaf material. Annual nutrient loss from dying leaves (amounts translocated or leached) was calculated as accumulation during growth minus retention in dead leaves. Leaf decomposition and nutrient flux Litterbag experiments were initiated in July 1977, October 1977 and February 1979 at the locations of previous production studies. Experiments that began in July and October 1977 compared sites B, C and D, while 1979 experiments compared Sites A, B and D. This yielded nine combinations of sites and sampling periods. Intact standing dead leaves were collected from living C. jamaicense and T. domingensis plants at each site the month before litterbags were placed in the marsh. Leaves were cut into 10 cm lengths and oven-dried at 45 °C for 96 h. Litterbags were constructed from 30 X 30 cm squares of fiberglass window screening (six meshes per centimeter) which were loosely folded to contain the leaf material while allowing the entry of macroinvertebrates into the bags. Each bag contained 5 g dry mass of dead leaf material. Litterbags were placed in the water in the C. jamaicense and T. domingensis stands where the dead leaf material had been collected. The bags floated for the first few days until


210 S.M. DAVIS the litter became waterlogged, after which they gradually sank to the bottom and became incorporated into the litter layer. Three to four replicate bags from each setout were retrieved after 1 month, 1 year, 2 years and varying periods in between. Litter remaining in retrieved bags was dried at 90°C for 48 h plus 45 °C for 24 h, weighed, and analyzed for tissue P and N concentration. Nutrient contents of retrieved litter were calculated by multiplying dry mass by P and N tissue concentrations. Leaf tissue nutrient analyses Leaf material from quadrat samples and retrieved litterbags was ground in a Wiley mill after weighing. Analyses for P and N were made using a Technicon Autoanalyzer II, after solubilization of P by lithium metaborate fusion (Medlin et al., 1969) and Kjeldahl digestion of N using a block digester. For quality control, National Bureau of Standards NBS 1571 (National Bureau of Standards, 1979) was used as an external reference standard for each set of tissue and soil nutrient analyses. Analyses were accepted if values for the standard were _+ 10% of NBS values. Water sampling and nutrient analyses Water samples were collected and water depths were measured at C. jamaicense and T. domingensis sites monthly throughout the study. Samples were collected from all sites on the same date each month. Samples for the analysis of dissolved nutrient fractions were filtered through 0.45/tm Nucleopore filters. Samples for total P analysis were digested by autoclaving at 121 °C ( 15 PSI) using the persulfate procedure. Samples for total N analysis were digested by the Kjeldahl procedure. Analyses were made using a Technicon Auto Analyzer II according to procedures SM424G for total PO4 and SRP (American Public Health Association (A.P.H.A.), 1980), SM418F for NO3-N (A.P.H.A., 1980), SM417G for NH4-N (A.P.H.A., 1980) and EPA351.2 for total Kjeldahl nitrogen (Environmental Protection Agency, 1979 ). Total PO4 and SRP are reported as P. Analyses for SRP and NO3-N were conducted throughout the study, while analyses for total P, total N and NH4-N began in mid1976. RESULTS Surface water SRP and NO3-N concentrations were used to examine correlations of plant nutrient retention to enrichment because these parameters were measured during both vegetation sampling years of 1975-76 and 1979-80. Mean annual SRP and NOa-N concentrations declined southward along the nutrient gradient and were higher during the second sampling year


CLADIUM JAMAICENSE AND TYPHA DOMINGENSIS IN FLORIDA EVERGLADES 211 compared with the first (Davis, 1990). Nutrient gradients in combination with yearly variations yielded mean annual SRP concentrations of 0.0020.036 mg 1-1 and NO3-N concentrations of 0.005-0.050 mg 1-1 for the six combinations of vegetation sites and sampling years. The full range of measured surface water P and N parameters was used to examine correlations of litter nutrient retention to enrichment because all parameters were measured during the 1977-79 and 1979-81 litterbag experiments. Surface water nutrient concentrations during litterbag experiments showed differences between sites and sampling periods similar to those described for SRP and NO3N during vegetation sampling. Nutrient gradients, in combination with yearly variations, yielded mean surface water concentrations of 0.006-0.1 l 6 mg 1-l total P, 0.002-0.066 mg 1-1 SRP, 2.57-4.72 mg 1 -l total N, 0.006-0.031 mg 1l NO3-N and 0.02-0.58 mg 1l NH4-N during the nine litterbag incubations. Leaf biomass turnover and nutrient flux Leaf tissue concentrations of P, but not N, reflected surface water concentrations. Phosphorus concentrations in leaf tissue of both C. jamaicense and T. domingensis differed significantly (P< 0.01 ) among the six site/sampling year treatments. Mean annual leaf tissue P concentrations increased logarithmically with mean annual SRP concentrations in surface water (r= 0.84 and 0.93 for C. jamaicense living and dead leaves, respectively; r= 0.95 and 0.86 for T. domingensis living and dead leaves, respectively) (Fig. 2 ). Tissue N concentrations did not differ significantly along the gradient. A lack of significant differences for N may have resulted in part from the larger variability of tissue N concentrations compared with those of P (Fig. 2). Despite the lack of significant differences, a positive correlation of tissue N concentration to NO3-N was apparent for T. domingensis live leaves (r= 0.95 ). Tissue nutrient concentrations were significantly higher (P< 0.05 ) in live leaves compared with dead leaves. Concentrations in live C. jamaicense leaves exceeded those in dead leaves on average by factors of 2.4 for P and 1.6 for N (Fig. 2 ). Concentrations in live T. domingensis leaves exceeded those in dead leaves by factors 4.4 for P and 2.0 for N. The decline in tissue nutrient concentrations during leaf senescence was proportionately greater for P compared with N and for T. domingensis compared with C. jamaicense. Higher nutrient concentrations in live leaves, compared with dead leaves, indicated substantial translocation or leaching from leaves during mortality. Typha domingensis accumulated higher tissue nutrient concentrations than C. jamaicense during growth, although leaf concentrations after mortality were similar for both species. Concentrations of P and N in live leaf tissue were significantly higher (P<0.05) in T. domingensis compared with C. jamaicense. Phosphorus concentrations in live T. domingensis leaves averaged twice those in C. jamaicense, while T. domingensis N concentrations averaged 1.5


212 S.M. DAVIS C. jamaicense -~ °'6 t !! 0.12q ~ o.o4 t ~,,~ ," a f-~,b I I [ [ [ I ] ] 0 0.01 0.02 0 03 SRP (mg. L -1 ) T_. domingensis @d, & q I I I I ~ I I 0 0.01 0 02 003 SRP (mg.L -~) 1.2 1,0 T t ~ ~ m 0.44 r i ] [ ] i i ~ i i 001 002 003 0.04 0.05 NO3 (mg. L -~) T T a i r t ] i i r [ i r 0.01 0.02 003 004 Q.05 NO3 (mg • L-1) Fig. 2. Tissue P and N concentrations in living (©) and dead ( A ) leaves in relation to mean annual SRP and NO3-N concentrations. Values represent the annual mean + standard error. times those in C. jamaicense (Fig. 2 ). In contrast, dead leaf nutrient concentrations did not differ significantly between species. Although growing T. domingensis leaves were more effective than those of C. jamaicense in concentrating nutrients, dying T. domingensis leaves held a smaller proportion of assimilated nutrients. After growth and senescence, the two species were approximately equally effective in retaining nutrient concentrations in dead leaf tissue. Cladium jamaicense allocated 0.22-1.51 g m -2 year -1 P and 4.7-16.6 g m -2 year-l N to growing leaves, as shown by the upper sets of points in Fig. 3. These points represent estimates for the six combinations of sites and sampling years. Nutrient allocation to growing leaves corresponded to biomass and production estimates (Davis, 1990), and to tissue nutrient concentrations (Fig. 2). Dead leaves retained 0.07-0.74 g m -2 year -~ P and 2.9-10.8 g m -2 year-l N, as indicated by the lower sets of points. Differences between upper and lower points represent nutrient leaching or translocation from leaves during mortality. Because of falls in tissue P and N concentrations during leaf death, dying leaves lost 44-68% of the P and 31-46% of the N which they accumulated during growth. Nutrient allocation to growing C. jamaicense leaves and retention in dead leaves both decreased linearly with distance from the north levee where inflow structures were located (r= 0.97 and 0.99 for P allocation and retention, respectively; r=-0.98 and -0.97 for N allocation and retention, re


CLADIUMJAMAICENSE AND TYPHA DOMINGENSIS IN FLORIDA EVERGLADES 213 1.o: ,~ 05 ~ o a_ 8 0 0 8 ~" 4.0E -~ 30~ 20© ~o ~d 1.0-O o_ 0~)1 0.~)2 0.()3 0~)4 0.~)5 SRP (rag L -1) 15 E lO c g o o o o 0 && & & 8 Distance (kin) of Sites From N Levee 40 3020O0 10O~ 0i 001 0 0 @ ~% 0~2 0'o3 0~4 0~ NO3 (rag • L 1) Fig. 3. Sawgrass nutrient allocation to growing leaves (O) and retention in dead leaves ( A ), in relation to distance from the north levee in Water Conservation Area 2A. Fig. 4. Cattail nutrient allocation to growing leaves (O) and retention in dead leaves ( ), in relation to mean annual surface water SRP and NO3-N concentrations. spectively) (Fig. 3 ). Thus P and N allocation and retention were highest where concentrations were greatest in surface water and soil, toward the upper end of the nutrient gradient. Dying leaves also lost larger amounts of P and N where water and soil concentrations were higher, as indicated by the widening gap between uptake and retention toward the upper end of the nutrient gradient. Allocation of P and N to growing C. jamaicense leaves, and retention in dead leaves, did not reflect differences in surface water NOa-N and SRP concentrations between the two sampling years; correlations to mean SRP and NO3-N concentrations during specific sampling years were not found. Thus, nutrient allocation and retention resulting from C. jamaicense leaf turnover corresponded to general site characteristics of nutrient enrichment, as indicated by long-term surface water nutrient gradients and by distances from inflows. Typha domingensis allocated larger quantities of P and N to growing leaves, translocated or leached larger quantities from dying leaves, and retained larger quantities in dead leaves (Fig. 4) compared with C. jamaicense. Phosphorus allocation to T. dom ingensis leaves of 0.64-4.16 g m-e year-1 averaged 2.7 times that of C. jamaicense, while N allocation to T. domingensis leaves of


214 S.M. DAVIS 9.6-33.4 g m -2 year~ averaged twice that of C. jamaicense. Greater nutrient allocation to T. domingensis leaves resulted from higher biomass and production rates (Davis, 1990), as well as from higher tissue P and N concentrations (Fig. 2 ). Upon mortality, T. domingensis leaves translocated or leached proportionately larger amounts of nutrients than C. jamaicense. Dying T. domingensis leaves lost 71-83% of the P and 33-63% of the N which they accumulated during growth. After allocation during growth and translocation or leaching during mortality, P retention in dead T. domingensis leaves of0.111.00 g m -2 year~ averaged 1.6 times that of C. jamaicense, while T. domingensis N retention of 3.6-15.9 g m -2 year -~ averaged 1.8 times that of C. jamaicense. Typha domingensis differed from C. jamaicense in that nutrient allocation to leaves showed little correlation to the distance of sites from inflows, but increased logarithmically with mean annual nutrient concentrations in surface water during the particular sampling years (r= 0.87 for both P allocation and retention in relation to SRP; r= 0.99 and 0.92 for N allocation and retention in relation to NO3-N) (Fig. 4 ). The gap between growing leaf allocation and dead leaf retention widened as SRP and NO3-N concentrations increased, indicating that dying T. domingensis leaves leached or translocated larger quantities of P and N when surface water concentrations were higher. Typha domingensis resembled C. jamaicense in that both species responded to higher surface nutrient concentrations by increasing P and N allocation to growing leaves, release from dying leaves and retention in dead leaves. However, only in T. domingensis were these processes correlated to yearly variations in surface water nutrient concentrations. 4.0 1 I~ 3.0 2.0 -C. jamaicense .:,:,.z, \~.1 =_ T_. domingensis \ "'2'"'"" '"' "" '=' '."' "'2" '='" 'J" """ '=' 'J" 1977 1978 1979 1977 1978 1979 --nutrient enriched site B,-----transitional site C,-----background site D Fig. 5. Dry mass of sawgrass and cattail leaves retrieved from litterbags during 2 years of decomposition. Values represent the mean + standard error. N= 3.


CLADIUMJAMAICENSE AND TYPHA DOMINGENSIS IN FLORIDA EVERGLADES 215 Leaf decomposition and nutrient flux Patterns of litter decomposition and nutrient flux, as illustrated for July 1977 litterbag experiments, were similar for all experiments. Approximately half of the litter mass remained intact in the bags after 2 years for sites and species combined (Fig. 5). Nutrient enrichment resulted in accelerated decomposition rates of both C. jamaicense and T. domingensis, as evidenced by the smaller litter mass remaining in bags at eutrophic sites relative to transitional and oligotrophic sites. Typha domingensis decomposed more rapidly than C. jamaicense at each site. Phosphorus and N concentrations in decomposing litter of both species increased over 2 years at eutrophic and transitional sites (Figs. 6 and 7). In0.18-o~ 0.14c" o i 0.I00 0.08o 0.02el. C_. jamaicense 0.18 -0.14 -T/z~': 0.10 ~ / T //';" 0.06 / // ~, =~." . ..~. ..... 0.02 .jI /1 I>l /III I>.l /l Io, l>l Izl I¢l I I ,l 1077 1978 1979 T_. domingensis l iii//i ~"[ T/ ,..-" . .=, .~;.-'..'-. ..................... ", ~,,,,,,,~,,,,,,,,,,,,,,,,, 1977 1978 1979 oi 4.0-E ~ ~. 3.0-c ~ ~ 2.0-.~ _ i 1.0-g. 1" z/J. / ..... ~,5~ ..... _, I I loll I I I I lii i I I I i l I>1 Izi I1 I>_1 11 1977 1978 1979 40 t 3.0 ¢_ 20 1 //i /, 1.o \/ ....T./ Ii i1 iiiiii lilt i ii iiiiflll i ~~.~z~|~ ~ 1977 1978 1979 --nutrient enriched site B,----transitional site C, .... background site D Fig. 6. Tissue P concentration and content in sawgrass and cattail leaves during 2 years of decomposition. Values represent the mean_+ standard error. N= 3.


216 S.M. DAVIS == 2.0 o~ 1.5 -ff _ 1.0 _ 0.5 _ Z _O. jamaicense 2.0 ~II 1.s y 1.0 ;> ............. .:~ .....~ 0,5 T. domingensis T-____T .~..//1 1 / I* I I T/ i ~]. -~" T ,.._~-:,;7-" I i ." j • ~ ~.,,~-'~ " "x~''~I I IIIIIIlll IIIIt Ittll IIIII IIII Illllll IIIIIIIIII IIIII 1977 1978 1979 1977 1978 1979 4O 35. O0 e, 30oJ ¢~ 25o 20Z 15,:jl -IT j. -~ /, i S___I i J. ....... ........ { t4_i, .,.t -----= 40 -35 -30 -25 -i 20 -15 -W T 'I~ T "><~/i .. ..... K I ""'~ II -" --[" ..... T ~,! i i i IIIItt lit illlllilll II III Ililllllllltlllllllll[llll 1977 1978 1979 1977 1978 1979 -nutrient enriched site B~-----transitional site C,.... background site D Fig. 7. Tissue N concentration and content in sawgrass and cattail leaves during 2 years of decomposition. Values represent the mean_+ standard error. N= 3. creases in litter P and N concentrations were higher at enriched sites compared to transitional ones, and in T. domingensis compared to C. jamaicense. Phosphorus concentrations at the oligotrophic site changed little, while N concentrations at the oligotrophic site increased over 2 years. Increasing P and N concentrations in decomposing litter resulted in a net increase in content per bag and immobilization of these elements over 2 years, despite decreasing litter mass, at eutrophic and transitional sites. Litter at the oligotrophic site neither gained nor lost significant amounts of P, but accumulated N, during this period. Litter accumulated more P and N toward the upper end of the nutrient gradient. Typha domingensis litter accumulated more P


CLADIUMJAMAICENSE AND TYPHA DOMINGENSIS IN FLORIDA EVERGLADES 217 0.60. 0.50 • 0.40 g "~ 0.30 g 0 ,,,, 0.20 <3 010 C. jamaicense A z D ,~. P Content= -.074274-14.27565 (Total P) z D -104.66287 (Total p)2 r : 0.952 T. domingensis B A x',[ O zk p Content: =00683+17.58563 (Total P) -146.72790 (Total p)2 r = 0,800 r r i i i i 0.02 0.04 0.06 0.08 010 0(~2 0.014 0.06 008 0.10 Total P (mg.L ~) Fig. 8. Changes in the P content of sawgrass and cattail leaves during 2 years of decomposition in relation to mean surface water total P concentration. Values represent changes in mg P per g freshly dead leaf material_+ standard error. N= 3 for 1977 setouts and 4 for 1979 setouts. TABLE 2 Net P and N immobilization ( g m -2 year~ ) by C. jamaicense and T. domingensis resulting from annual leaf turnover and 2 years decomposition. Values represent site means with standard errors in parentheses Site A ~ Site B Site C Site D Phosphorus C. jamaicense 1.11 1.34(0.03) 0.89(0.10) 0.07(0.01 ) T. domingensis 1.42 1.25 (0.12 ) 1.68 (0.24) 0.30 (0.13 ) Nitrogen C.jamaicense 17.4 15.2(0.5) 10.7(1.6) 4.2(0.3) T. domingensis 18.3 15.1 (2.1) 21.3(0.0) 10.8(3.1 ) ~N=I. than C. jamaicense at each site. The amount of P that leaf litter accumulated during 2 years of decomposition was related to the mean surface water total P concentration during that period (Fig. 8 ). Litter of both species accumulated increasing amounts of P as surface water total P concentrations increased from background levels of less than 0.01 mg 1-~ to intermediate levels of enrichment of about 0.06 mg 1-~. However, once total P concentrations exceeded about 0.06 mg 1-1, litter P uptake leveled off, then declined with


218 S.M. DAVIS higher levels of enrichment. Unlike P, litter N uptake was not clearly related to surface water N concentrations. Both C. jamaicense and T. domingensis sequestered P and N as the cumulative result of leaf growth, death and 2 years of decomposition (Table 2). Amounts of P and N that were sequestered were estimated by summing nutrient retention by dead leaves plus detritus nutrient immobilization. Nutrient immobilization by leaf detritus as it decomposed over 2 years was calculated by multiplying annual crops of dead leaf material (Davis, 1990) by changes in litter P and N content per unit mass of freshly dead leaf material (Fig. 8 ). Sequestration of P and N by C. jamaicense increased along the nutrient gradient from background Site D to enriched Sites A and B (Table 2 ). Typha domingensis P and N sequestration was also lowest at Site D, but peaked at transitional Site C, and then leveled off or declined further up the gradient at Sites A and B. The total amount of P that was sequestered as the result of annual leaf turnover and 2 years of decomposition ranged from 0.07 to 1.33 g m -2 for C. jamaicense and from 0.30 to 1.42 g m -2 for T. domingensis. Corresponding rates of N sequestration ranged from 4 to 17 g m -2 for C. jamaicenseand from 11 to 21 g m -2 for T. domingensis. DISCUSSION Leaf biomass turnover and nutrient flux Nutrient allocation to growing leaves was greater for T. domingensis than for C. jamaicense. Leaves of T. domingensis accumulated more P and N than C. jamaicense under equivalent conditions, and T. domingensis was capable of allocating more nutrients to leaves during years of elevated nutrient enrichment. T. domingensis leaf turnover rates, being two to three times those of C. jamaicense (Davis, 1990), allowed nutrient allocation to leaves to better reflect temporal changes in surface water concentrations. The ability of T. domingensis to assimilate larger quantities of P and N in growing leaves corresponded to its observed spread into C. jamaicense stands in Everglades areas receiving high-nutrient inflows. A similar intrusion of Typha latifolia L. into macrophyte stands, with an eventual shift to a T. latifolia dominant community, was noted by Kadlec (1987) in the vicinity of effluent discharge in a Michigan wetland. These findings are compatible with models of Shaver and Melillo (1984) and Richardson and Marshall (1986), which predict a vegetation shift toward macrophyte species that are more efficient in nutrient uptake as nutrient supply increases. With nutrient enrichment in the Everglades, a plant species that was more effective in accumulating nutrients in leaf tissue invaded areas occupied by a species that was less effective. Contrasting responses of C. jamaicense and T. domingensis to the nutrient gradient are characteristic of low-nutrient-status plant species from infertile


CLADIUM JA MAICENSE AND TYPHA DOMINGENSIS IN FLORIDA EVERGLADES 219 habitats (C. jamaicense) versus high-nutrient-status species from fertile habitats (T. domingensis) (Grime, 1977; Chapin, 1980). The relatively small leaf growth response of C. jamaicense to temporal variations in surface water nutrient inputs (Davis, 1990) is typical of low-nutrient-status plants. The relatively low nutrient allocation to C. jamaicense leaves, approximately half that of T. dorningensis, under high-nutrient conditions is another trait of lownutrient-status species. A lower rate of nutrient loss from senescing C. jamaicense leaves (via a combination oftranslocation and leaching) also indicates a low-nutrient-status species. The longer leaf longevity, lower leaf growth rate and slower leaf turnover rate of C. jamaicense (Davis, 1990), plus its welldeveloped leaf cuticle, are adaptations that reduce leaching loss in low-nutrient-status plants. In comparison, the larger growth response to changing nutrient availability, higher rates of nutrient allocation to growing leaves under high-nutrient conditions, higher rate of nutrient loss from senescing leaves, shorter leaf longevity, faster leaf growth rate and more rapid leaf turnover of T. domingensis are all characteristics of high-nutrient-status plants which tend to be competitive when the nutrient supply increases. In the fertile habitat near the upper end of the nutrient gradient in Water Conservation Area 2A, the high-nutrient-status traits of T. domingensis appear to have given that species competitive advantage over C. jarnaicense. The effectiveness of growing and dying leaves in retaining P and N remained constant as surface water concentrations of these elements increased. Nutrient retention in freshly dead leaves increased with enrichment at a slower rate than did uptake by growing leaves. Elevated quantities of nutrients which were allocated to growing leaves at enriched sites were mostly translocated or leached from the leaves by the time they died. As a result, dead leaves retained only slightly greater quantities of nutrients under enriched conditions in comparison to background conditions. The main effect of enrichment on leaf nutrient allocation was to increase translocation or leaching from dying tissue, rather than to increase retention in dead leaf material. Since leaf biomass turnover alone was an inefficient mechanism for trapping elevated surface water nutrient inputs, long-term nutrient retention appeared to depend largely on the nutrient dynamics of litter decomposition. The main role of leaf turnover in nutrient retention appeared to be the production of detritus as a substrate for subsequent physical-chemical and microbiological processes. Davis and van der Valk ( 1978 ) drew similar conclusions from temperate stands of Typha glauca Godr. Leaf decomposition and nutrient flux The net effect of production and decomposition was an accumulation of leaf detritus. The finding that approximately half of the leaf litter remained intact after 2 years indicated that decomposition proceeded less rapidly than


220 S.M, DAVIS production, as would be expected in a peatland. Litter decomposition rates in the Everglades were comparable with the 2 year litter half-life reported by Kadlec (1989) in a Michigan peatland. Nutrient enrichment appeared to influence both the annual rate of detritus accumulation and the physical structure of deposited sediments. Detritus accumulation was estimated by combining decomposition rates from this study with production estimates of Davis (1990). The interior C. jamaicense marsh at Site D produced about 894 g m -2 of dead leaf material annually, of which about 45% was lost during 2 years decomposition and about 55% remained as relatively compact, fibrous organic sediment (S.M. Davis, personal observation, 1975-1980) on the marsh floor. In contrast, nutrient-enriched areas that had converted to T. domingensis (Sites A and B ) produced about 2417 g m2 of dead leaf material annually, of which about 52% was lost during 2 years decomposition and about 48% remained as relatively fine, flocculent sediment (S.M. Davis, personal observation, 1975-1980 ). As a net result of production and 2 years decomposition, background C. jamaicense deposited about 492 g m -2 year-l of relatively intact and compact leaf detritus, while nutrient-enriched T. domingensis deposited about 1160 g m -2 yearL of finer, more flocculent sediment. The results of this study indicate that Everglades C. jamaicense and T. domingensis communities deposited P and N in the detritus that resulted from leaf growth, death and 2 years decomposition. Cladium jamaicense and T. domingensis stands accumulated detritus after 2 years of decomposition, contributed to the accretion of organic sediment and deposited nutrients within these sediments. Rates of nutrient storage in organic matter found in this study represented values under conditions of continuous flooding; these values might have differed under more dynamic hydrologic conditions. Rates of nutrient storage reported here did not include contributions by below-ground biomass, which may have increased rates by about 12% for C. jamaicense and by about 16% for T. domingensis (Toth, 1987, 1988). The relevance of nutrient retention after leaf production and 2 years decomposition to longer term nutrient sequestration in accreting organic sediments was not evaluated in this study, but the two processes may have been comparable, as evidenced by the similarity of 2 year nutrient deposition estimates in the Everglades to nutrient accretion measurements in other wetlands. Nutrient deposition estimates for the oligotrophic Everglades site are approximately equivalent to accretion estimates for oligotrophic peatlands and backwater marshes. Mean retention rates of 0.18 g m -2 year-~ P and 7.5 g m-2 year-~ N at the Everglades oligotrophic site (Site D, species combined ) are comparable with calculated values of less than 0.1-0.2 g m -2 year~ P retention and 0.1-4.7 g m -2 year -~ N retention through peat accretion in northern oligotrophic wetlands (Nichols, 1983). Nutrient retention estimates at the Everglades oligotrophic site are also similar to measurements of 0.5 g m -2 year-~ P and 9 g m -2 yearl N retention in accreting sediments in


CLADIUM JAMAICENSE AND TYPHA DOMINGENSIS 1N FLORIDA EVERGLADES 221 a Louisiana backwater marsh (Hatton et al., 1982), although Louisiana sediments contained mineral matter in addition to organics. Litter breakdown has been reported to be largely complete after 2 years in some wetlands (Latter and Cragg, 1967; Chamie, 1976; Day, 1982). Nutrient release from detritus slowed before 2 years in the Everglades, as evidenced by little change in the litter nutrient content during the last 6-12 months of the 2 year decomposition period. Subsequent oxidation and mineralization of detritus may have been impeded by continuous flooding and reducing conditions in the detrital layer (Reeder and Davis, 1983 ). Perhaps rates of long-term nutrient sequestration in accreting organic sediment in the Everglades C. jamaicense community are comparable with rates in oligotrophic peatlands under natural lownutrient conditions. Similar litter half-lives in the Everglades and in a Michigan peatland (Kadlec, 1989) support this interpretation. Where surface water nutrient concentrations were higher in the Everglades, nutrient retention rates in 2-year-old detritus were higher, but apparently only up to a finite capacity. Rates of 1.1-1.4 g m -2 year -I P and 17-18 g m -2 year1 N retention by C. jamaicense and T. domingensis at Site A may represent upper limits of nutrient retention resulting from leaf production and 2 years decomposition in eutrophic Everglades habitat; these rates were equaled or exceeded downstream along the surface water nutrient gradient at transitional sites between Sites A and D. Increased rates of nutrient retention at eutrophic Everglades sites might be expected, based on comparison with other studies. The finding of higher detrital nutrient retention at eutrophic Everglades sites, in combination with the decline in surface water P and N concentrations along the nutrient gradient below inflow structures, agree with the conclusion of Kadlec (1989) that the litter zone in a Michigan peatland receiving secondary effluent contained a large fraction of the nutrients added over a 10 year period. Greater detrital nutrient retention at enriched Everglades sites also agrees with the review of Richardson and Nichols ( 1985 ) of loading-retention relationships in northern wetlands receiving ~vastewater, where P removal increased up to 4.5 g m -2 year1 as loading increased. However, a corresponding decrease in nutrient uptake efficiency (percentage of inputs) as loading increased (Richardson and Nichols, 1985 ) also suggests that the nutrient retention capacity of wetland may be limited, as appears to be the case for Everglades macrophyte stands. Thus, rates of long-term nutrient sequestration may increase with loading, to a limited extent, in a wide variety of wetland systems. The Everglades results suggest that the capacity of this system for long-term nutrient sequestration in accumulating detritus can increase, but only up to a finite level, as anthropogenic nutrient inputs increase. CONCLUSIONS Everglades C. jamaicense and T. domingensis sequester P and N in the accumulated organic sediment that results from annual production, mortality


222 S.M. DAVIS and 2 years decomposition. However, the capacity of vegetation for nutrient retention as a result of these processes is limited. As surface water concentrations of P and N increase along the gradient and during high-discharge years, retention of these elements reaches a maximum under moderate levels of enrichment, between Sites B and C in this study. The Everglades C. jamaicense community is adapted to low nutrient inputs primarily from direct rainfall, has limited capacity to retain higher nutrient inputs and is subject to displacement by a more competitive species ( T. domingensis) under higher-nutrient conditions. The observed spread of T. domingensis into C. jamaicense sites in nutrient-enriched areas of the Everglades, and contrasting competitive strategies of nutrient allocation by the two species, support this concept. Apparently, when the vegetation uptake capacity of stands already invaded by T. domingensis is exceeded during periods of high nutrient inputs, nutrients pass further downstream and the competitive strategies of T. domingensis allow it to invade C. jamaicense sites in the downstream area. Wetland ecosystems which developed under conditions of low nutrient supply, such as the Florida Everglades, may offer a finite potential for accelerated vegetation nutrient retention when the exogenous nutrient supply increases as a result of human activities. However, a plant species such as C. jamaicense, that is adapted to a low-nutrient Everglades environment, may have a low nutrient threshold before it loses its competitive capability and is replaced by a species such as T. domingensis that is better adapted to a highnutrient environment. ACKNOWLEDGEMENTS I gratefully acknowledge the South Florida Water Management District for supporting this research. J.W. Dineen provided encouragement and support throughout the study. Water and soil nutrient analyses were performed by the South Florida Water Management District Chemistry Laboratory. I particularly wish to thank L. Haunert, M. Zaffke, M. Rosen, A.M. Superchi, N.H. Urban, F.E. Worth, D. Cook, S. Green and P.B. Reeder for major contributions to field data collection and laboratory analyses. J.B. Grace, A. Herndon, R.H. Kadlec, M.S. Koch, C.R. Richardson and L.A. Toth provided reviews of the manuscript. REFERENCES Almazon, G. and Boyd, C.E., 1978. Effects of nitrogen levels on rates of oxygen consumption during decay of aquatic plants. Aquat. Bot., 5:119-126. American Public Health Association, 1980. Standard Methods for the Examination of Water and Wastewater. 15th Edition, 1134 pp.


CLADIUMJAMAICENSE AND TYPHA DOMINGENSIS IN FLORIDA EVERGLADES 223 Boyd, C.E., 1970. Vacular plants for mineral nutrient removal from polluted waters. Econ. Bot., 24: 95-103. Brinson, M.M., 1977. Decomposition and nutrient exchange of litter in an alluvial swamp forest. Ecology, 58: 601-609. Chamie, J.P.M., 1976. The effects of simulated sewage effluent upon decomposition, nutrient status and litter fall in a central Michigan peatland. Dissertation, University of Michigan, Ann Arbor, MI, 110 pp. Chapin, F.S., 1980. The mineral nutrition of wild plants. Annu. Rev. Ecol. System., 11: 233260. Davis, C.B. and van der VaIL A.G., 1978. The decomposition of standing and fallen litter of Typha glauca and Scirpusfluviatilis. Can. J. Bot., 56: 662-675. Davis, C.B. and van tier Valk, A.G., 1983. Uptake and release of nutrients by living and decomposing Typha glauca Oodr. tissues at Eagle Lake, Iowa. Aquat. Bot., 16: 75-89. Davis, J.H., 1943. The natural features of southern Florida. Fla. Geol. Soc. Geol. Bull. No. 25, 311 pp. Davis, S.M., 1990. Sawgrass and cattail production in relation to nutrient supply in the Everglades. In: R.R. Sharitz and J.W. Gibbons (Editors), Freshwater Wetlands and Wildlife. Office of Scientific and Technical Information, U.S, Department of Energy, Oak Ridge, TN, pp. 325-341. Day, F.P., Jr., 1982. Litter decomposition rates in the seasonal flooded Great Dismal Swamp. Ecology, 63: 670-678. Elwood, J.W., Newbold, J.D., Trimble, A.F. and Stark, R.W., 1981. The limiting role of phosphorus in a woodland stream ecosystem: effects of P enrichment on leaf decomposition and primary producers. Ecology, 62:146-158. Environmental Protection Agency, 1979. Methods for chemical analysis of water and wastes. EPA-600/4-79-020, 430 pp. Grime, J.P., 1977. Evidence for the existence of three primary strategies in plants and its relevance to ecological and evolutionary theory. Am. Nat., 111: 1169-1194. Hatton, R.S., Patrick, W.H., Jr. and DeLaune, R.D., 1982. Sedimentation, nutrient accumulation, and early diagenesis in Louisiana Barataria Basin coastal marshes. In: V.S. Kennedy (Editor), Estuarine Comparisons. Academic Press, New York, pp. 255-267. Hopkinson, C.S. and Schubauer, J.P., 1984. Static and dynamic aspects of nitrogen cycling in the salt marsh graminoid Spartina alterniflora. Ecology, 65:961-969. Howard-Williams, C. and Howard-Williams, W., 1978. Nutrient leaching from the swamp vegetation of Lake Chilwa, a shallow African lake. Aquat. Bot., 4: 257-267. Howarth, R.W. and Fisher, S.G., 1976. Carbon, nitrogen, and phosphorus dynamics during leaf decay in nutrient-enriched ecosystems. Freshwater Biol., 6:221-228. Kadlec, R.H,, 1987. Northern natural wetland water treatment systems. In: K.R. Reddy and W.H. Smith (Editors), Aquatic Plants for Water Treatment and Resource Recovery. Magnolia, Orlando, FL, pp. 83-98. Kadlec, R.H., 1989. Decomposition in wastewater wetlands. In: D.A. Hammer (Editor), Constructed Wetlands for Wastewater Treatment. Lewis Press, Chelsea, MI, pp. 459-468. Latter, P.M. and Cragg, J.B., 1967. The decomposition ofduncus squarrosus leaves and microbiological changes in the profile of Juncus moor. J. Ecol., 55: 465-482. Loveless, C.M., 1959. A study of the vegetation in the Florida Everglades. Ecology, 40: 1-9. Medlin, J.H., Suhr, N.H. and Bodkin, J.B., 1969. Atomic absorption analysis of silicates employing LiBO2 fusion. At. Absorpt. Newsl., 8: 25-29. Nichols, D.S., 1983. Capacity of natural wetlands to remove nutrients from wastewater. J. Water Pollut. Control Fed., 55: 495-505. Parker, G.G., 1974. Hydrology of the pre-drainage system of the Everglades in southern Florida.


224 S.M. DAVIS In: P.J. Gleason (Editor), Environments of South Florida: Present and Past. Miami Geological Society, Memoir 2. Miami Geological Society, Miami, FL, pp. 18-27. Puriveth, P., 1980. Decomposition of emergent macrophytes in a Wisconsin marsh. Hydrobiologia, 72: 231-242. Reddy, K.R., 1983. Soluble phosphorus release form organic soils. Agriculture Ecosystems Environ., 9: 373-382. Reddy, K.R. and Rao, P.S.C., 1983. Nitrogen and phosphorus fluxes from a flooded organic soil. Soil Sci., 136: 300-307. Reeder, P.B. and Davis, S.M., 1983. Decomposition, nutrient uptake and microbial colonization of sawgrass and cattail leaves in Water Conservation Area 2A. S. Fla. Water Manage. Dist. Tech. Publ. 83-4, 24 pp. Rice, D.L., 1982. The detritus nitrogen problem: new observations and perspectives from organic geochemistry. Mar. Ecol. Prog. Ser., 9: 153-162. Richardson, C.J. and Marshall, P.E., 1986. Processes controlling movement, storage, and export of phosphorus in a fen peatland. Ecol. Monogr., 56: 279-302. Richardson, C.J. and Nichols, D.S., 1985. Ecological analysis ofwastewater management criteria in wetland ecosystems. In: P.J. Godfrey, E.R. Kaynor and S. Pelczarski (Editor), Ecological Considerations in Wetlands Treatment of Municipal Wastewaters. Van Nostrand Reinhold, New York, pp. 351-391. Saunders, G.W., 1976. Decomposition in freshwater. In: J.M. Anderson and A. MacFadyen (Editors), The Role of Terrestrial and Aquatic Organisms in Decomposition Processes. Blackwell Scientific, Oxford, pp. 341-373. Shaver, G.R. and Melillo, J.M., 1984. Nutrient budgets of marsh plants: efficiency concepts and relation to availability. Ecology, 65:1491-1510. South Florida Water Management District, 1991. Surface Water Improvement and Management Plan for the Everglades. Volume III. Tech. Rep., S. Fla. Water Manage. Dist., West Palm Beach, FL, 441 pp. Steward, K.K., 1970. Nutrient removal potentials of various aquatic plants. Hyacinth Control J., 8: 34-35. Steward, K.K. and Ornes, H., 1975. The autecology of sawgrass in the Florida Everglades. Ecology, 56: 162-171. Toth, L.A., 1987. Effects of hydrologic regimes on lifetime production and nutrient dynamics of sawgrass. S. Fla. Water Manage. Dist. Tech. Publ. 87-6, 32 pp. Toth, L.A., 1988. Effects of hydrologic regimes on lifetime production and nutrient dynamics of cattail. S. Fla. Water Manage. Dist. Tech. Publ. 88-6, 25 pp. Webster, J.R. and Benfield, E.F., 1986. Vascular plant breakdown in freshwater ecosystems. In: R.F. Johnston, P.W. Frank and C.D. Michener (Editors), Annual Review of Ecology and Systematics, 17. Annual Reviews, Palo Alto, CA, pp. 567-594.




© 2014 Shannon Duffy


3 ACKNOWLEDGMENTS It is with sincere gratitude that I acknowledge the many people who have helped me throughout my education and especia lly with this research effort. I thank m y chair , Dr. Todd Osborne, and my other committee members Dr. Angelique Bocknak, Dr. Mark Clark, and D r. Rex Ellis for guidance throughout this process. I especia lly thank Todd for his patience, encouragement , and advice on keeping things in perspective. The moral support he provided is a privilege to have had as a student. I also thank Angelique for he r mentorship and help with every aspect of this project. Many thanks to the St. Johns River Water Management District staff, especially Kimberly Ponzio and Ken Synder for both facilitating safe access to the research sites and sharing their extensive knowl edge during field days. Field assistance was also provided by Lawrence Keenen and Tim Miller. I thank Brenhan Street, Yu Wang and Katelyn Foster assistance in the laboratory. I also owe thanks to Mike Sisk and Linda Cowart, whose logistical and administ rative support was essential . Colleagues and friends within the Soil and Water Science D epartment have provided more support and guidance than they probably realize, and I owe them many thanks. I must also recognize the support of the Department of Hous ing and Residence Education. My position with Housing is what made graduate school possible. I am very grateful for the professional development and lifelong friendships that I have gained . Lastly, I sincerely thank my family and friends for their tremendous encouragement and support throughout this process.


4 TABLE OF CONTENTS page ACKNOWLEDGMENTS ................................ ................................ ................................ .. 3 LIST OF TABLES ................................ ................................ ................................ ............ 6 LIST OF FIGURES ................................ ................................ ................................ .......... 7 ABSTRACT ................................ ................................ ................................ ................... 10 CHAPTER 1 INTRODUCTION ................................ ................................ ................................ .... 12 Background Information on Biogeochemical Processes ................................ ......... 13 Dec omposition of Detritus ................................ ................................ ....................... 14 Upper St. Johns River Basin ................................ ................................ ................... 16 Objectives ................................ ................................ ................................ ............... 20 2 LITTER PRODUCTION ................................ ................................ .......................... 21 Influential Factors ................................ ................................ ................................ ... 21 Vegetation Community ................................ ................................ ..................... 21 Hydrology ................................ ................................ ................................ ......... 23 Other Factors ................................ ................................ ................................ ... 24 Research Questions ................................ ................................ ............................... 24 Methods ................................ ................................ ................................ .................. 24 Site Desc riptions ................................ ................................ .............................. 24 St Johns Marsh Conservation Area ................................ ........................... 27 Blue Cypress Marsh Conservation Area ................................ .................... 29 Fort Drum Marsh C onservation Area ................................ ......................... 30 Biomass Harvest ................................ ................................ .............................. 31 Results ................................ ................................ ................................ .................... 33 Discussion ................................ ................................ ................................ .............. 36 3 LITTER DECOMPOSITION ................................ ................................ .................... 39 Rate Influencing Factors ................................ ................................ ......................... 39 Nutrients ................................ ................................ ................................ ........... 40 Fiber Quality ................................ ................................ ................................ ..... 41 Hydrology ................................ ................................ ................................ ......... 42 Research Questions ................................ ................................ ............................... 43 Field Methods ................................ ................................ ................................ ......... 43 Lab Methods ................................ ................................ ................................ ........... 49 Results ................................ ................................ ................................ .................... 51 Decomposition Rates ................................ ................................ ....................... 51


5 F iber Analysis ................................ ................................ ................................ ... 54 Nutrients ................................ ................................ ................................ ........... 57 Percent totals ................................ ................................ ............................. 57 Mass difference ................................ ................................ .......................... 60 Ratios ................................ ................................ ................................ ......... 64 Multivariate Analysis ................................ ................................ ......................... 64 Discussion ................................ ................................ ................................ .............. 66 Field Decomposition ................................ ................................ ......................... 66 Influential Site Conditions ................................ ................................ ................. 67 Biotic influence ................................ ................................ ........................... 67 Abiotic influence ................................ ................................ ......................... 69 Fiber Analysis ................................ ................................ ................................ ... 70 Nutrients ................................ ................................ ................................ ........... 71 Multiple Variables ................................ ................................ ............................. 72 Summary ................................ ................................ ................................ .......... 73 4 CONCLUSIONS ................................ ................................ ................................ ..... 75 APPENDIX A ADDITIONAL FIGURES ................................ ................................ ......................... 78 LIST OF REFERENCES ................................ ................................ ............................... 85 BIOGRAPHICAL SKETCH ................................ ................................ ............................ 88


6 LIST OF TABLES Table page 1 1 Soil characteristics for sites in St Johns Marsh Cons ervation Area .................... 20 1 2 Species of interest for this study. ................................ ................................ ........ 20 2 1 Above ground biomass (g) collected from clip plots in July, 2012 and July, 2013. . ................................ ................................ ................................ ................. 35 2 2 Annual litter production contribution to soil . . ................................ ....................... 36 3 1 Litter bag contents ................................ ................................ .............................. 46 3 2 Sampling timeline. . ................................ ................................ ............................. 48 3 3 Mean annual d ecomposition rate constants. ................................ ...................... 52 3 3 Continued: Mean annual d ecomposition rate constants ................................ ..... 53 3 4 Evaluation of difference between composite and individual samples of each species. ................................ ................................ ................................ .............. 55 3 5 Coefficients of correlation between % nutrients and time ................................ ... 58 3 6 Coef ficients of correlation between decomposition rate constants (k) for each species and different variables. ................................ ................................ .......... 65 4 1 Calculated values of vegetation input to soil. ................................ ...................... 76 4 2 Annual litter contribution to soil elevation after decomposition. ......................... 76


7 LIST OF FIGURES Figure page 2 1 Wetland habitat types in St. Johns Marsh Conservation Area ............................ 22 2 2 Wetland habitat types in Blue Cypress Marsh Conservation Area ...................... 22 2 3 Wetland habitat types in Fort Drum Marsh Conservation Area ........................... 23 2 4 Location of Upper St. Johns River Basin Conservation Area Marshes ............... 26 2 5 Stu dy site locations. . ................................ ................................ ........................... 27 2 6 Location of the SJM WIL stu dy site. . ................................ ................................ .. 28 2 7 Location of SJM SAW and SJM TY P sites. . ................................ ....................... 29 2 8 Loca tion of BCMCA and FDMCA sites. . ................................ ............................. 30 2 9 Standing live, standing dead, and litter b iomass at each site in 2012 . ................ 34 2 10 Relative amount of different species collected at each site in 2012. ................... 34 2 11 Standing live, standing dead, and litter biomass at eac h site in 2013. . ............... 35 2 12 Annual litter contribution of C to soil. BCMCA and FDMCA n=3, SJM n=2. ....... 36 3 1 Heat sealing final edge of litter bag after filling with sample and label using an impulse sealer.. ................................ ................................ ................................ .. 44 3 2 Schematic of the litter bag deployment at each study site. ................................ . 47 3 3 Top and bottom view of a litter bag after being incorporate d in the litter layer . ... 48 3 4 Sc hematic of the fiber analysis process ................................ ............................. 51 3 5 Mean annual decomposition rate constants by species. . ................................ ... 54 3 6 Mean annual decomposition rate constan ts by site ................................ ............ 54 3 7 Comparison of mean % lignin content in different species . ................................ 57 3 8 Percent C in litter over time ................................ ................................ ................ 58 3 9 Percent N in litter over time ................................ ................................ ................ 59 3 10 Percent P in litter over time ................................ ................................ ................ 59


8 3 11 Change in mass of C over time . ................................ ................................ ......... 61 3 12 Change in mass of N over t ime from each litter bag type. ................................ .. 62 3 13 Change in mass of P over time from each litter bag type.. ................................ . 63 3 14 Carbon to N ratio over time for each species ................................ ..................... 64 3 15 Relationship between litter quality and decomposition rates. ............................. 66 3 16 Roots growing into litter material. Photo courtesy of author. ............................. 69 A 1 Mass loss over t ime of litter bag sample at each site, with corresponding exponential decay equations.. ................................ ................................ ............ 79 A 2 Percent mass loss over time by species . ................................ ............................ 80 A 3 Fiber fractions of each sample. ................................ ................................ ........... 81 A 4 Change in quality over time, comparing compo site bags and individual bags. . .. 82 A 5 Change in quality over time for in each species at each site. ............................. 83 A 6 C:N Ratios ................................ ................................ ................................ .......... 84


9 LIST OF ABBREVIATIONS BCM CA Blue Cypress Marsh Conservation Area C Carbon FDMCA Fort Drum Marsh Conservation Area N Nitrogen OM Organic matter P Phosphorus SJMCA St. Johns Marsh Conservation Area SJM SAW St. Johns Marsh sawgrass site SJM TYP St. Johns Marsh Typha site (cattail) SJM WIL St. Johns Marsh willow site SJR St. Johns River SJRWMD St. Johns River Water Management District SOM Soil organic matter TFMCA Three Forks Marsh Conservation Area TC Total carbon TN Total nitrogen TP Total phosphorus USJRB Upper St. Johns River Basin


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 LITTER PRODUCTION AND DECOMPOSITION IN THREE CONSERVATION AREA MARSHES By Shannon Duffy August 2014 Chair: Todd Osborne Major: Soil and Water Science The Upper St. Johns River Basin (USJRB) is one of ten major watersheds within the St. Johns River Water Management District. Appropriate understanding and management of w etlands in the USJRB has direct implications on soil subsidence and quality, and protection of natural systems. Net litter production and decomposition was estimat ed in three USJRB conservation area marshes using clip plots (n=3) . The roles of species community and site characteristics in determining decomp osition of litter were examined. Sites represented varied hydrologic disturbance. St. Johns Marsh Conservatio n Area (SJMCA) is highly altered , Blue Cypress Marsh Conservation Area (BCMCA) is intermediately altered, and Fort Drum Marsh Conservation Area (FDMCA) represents historical conditions . Litter material was characterized by fiber quality and nutrient conte nt over the course of one year of decomposition in the field. Results show quality of litter was correlated with decomposition rate for low quality material (%lignin >40) , while hydrology may by more influential for higher quality material. Using annual litter production estimates in conjunction with observed decomposition rates, annual


11 contribution to soi l elevation was found to be 0.20 0.06 cm in BCMCA, 0.18 0.07 cm in FDMCA, and 0.06 0.0 9 cm in SJMCA. Subsidence has resulted in the loss of organic soi ls and is not quickly reversible due to the slow rate of accretion.


12 CHAPTER 1 INTRODUCTION Soil organic matter (SOM) in wetlands is an important reservoir of global c arbon (C) ( Kayranli et al . , 2010; Marín Muñiz et al . , 2014 ). Drainage of wetlands results in significant soil subsidence and loss of C to the atmosphere ( Morris et al., 2004; Ewing and Vepraskas, 2006 ; Van Dam, 2012 ) . Subsidence occurs via two mechanisms. Primary subsidence is the compaction or settling of soil , and secondary subsidence is the oxidati o n of SOM. Oxidation of SOM results in the transfer of carbon from soil to the atmosphere , and may be cause d by microbial decomposition or fire (Ewing and Vepraskas, 2006). Together, primary and secondary subsidence lower the elevation of the soil following drainage ( Ewing and Vepraskas, 2006 ; Vam Dam, 2012). Accumulation of SOM is the opposing process that builds soil and raises elevation. Input of SOM is often derived from decomposing vegetative materia l , except in cases such as treatment wetl ands that receive external OM in puts (Kayranli et al., 2010) . Accumulation of SOM is therefore a function of litter production from plants and subsequent decomposition (Mitch and Gosselink, 1986 ; Kayranli et al., 2 010 ) . Organic soils are threatened and C stores are released when drainage ca uses subsidence to occur at a faster rate than can be counterbalanced by accretion (Vam Dam, 2012). Wetlands are well known for accumulating organic matter over time (Reddy and DeLuane, 2008) . The reasoning for this can be summarized in two points: 1) wetlands are highly productive systems and 2) saturated conditions reduce the amount of oxygen available for microbially mediated decomposition, resulting in slower decomposition of detritus as compared to uplands. When soils are saturated, the water displaces air from pore spaces resulting in a reduced environment after dissolved oxygen in the water is


13 used up . Additiona lly, standing water on the surface of the soil significantly reduces the rate of oxygen diffusion (approximately 10,000 times slower through water than through air) (Reddy and DeLaune, 2008). Any dissolved oxygen in the water is quickly used up, necessita ting alternative, less efficient electron acceptors instead of oxygen to be utilized by microbes in catabolism of OM , which slows down the decomposition process (McLatchey and Reddy, 1998; Reddy and DeLaune, 2008). This results in the buildup of OM over t ime in wetlands, and explains the rapid loss of OM due to oxidation in cases where wetlands dry and oxygen is abundant . Background Information on Biogeochemical Processes One of the many functions of wetlands is to store, release, and transform nutrients. The elements of interest for this research are carbon (C), nitrogen (N), and phosphorus (P). Carbon is a foundational element, creating complex chains and rings and bonding with other key elements to facilitate the processes and form structures that are required for life (Wade, 2006). Not only is it essential for life, C also plays a significant role in affecting global climate. The soil carbon pool represents a substantial fraction of global C, approximately 1500 1600 Pg, which is over twice the amoun t found in the atmosphere (approximately 700 Pg) (Raich and Potter, 1995; Bolin et al., 1979 ; Kayranli et al., 2010 ). Wetland soils account for approximately 68% of the soil C pool, despite representing a much smaller surface area of the earth compared to terrestrial systems ( Schlesinger , 1991). On the global scale, living and dead plant biomass includes approximately 700 Pg and 150 Pg of C, respectively (Bolin et al., 1979). The flux of C between pools is important to understand because the various form s and amounts of C in each reservoir have drastic impacts on global climate. The focus of this research is on the transfer of C from the atmosphere to the biosphere (in production of


14 plant biomass ) and from the biosphere to the soil (in decomposition of d etrital material). As decomposition occurs, respiration also contributes C back to the atmospheric pool. Nitrogen and P are also required for life and are essential parts of DNA and other nucleic acids, proteins, amino sugars, and other compounds. Becaus e of the direct linkages of N and P to C in complex compounds, the fates of these elements are directly tied to C. In addition to those organic forms of N and P, both elements have inorganic components to their cycles. In order for plants to utilize N or P, they must be in their inorganic forms. These elements are cycled in wetlands by microbes which decompose organic matter (OM), transforming organic N and P into bioavailable forms which can then be utilized in primary production by plants. Seasonal cy cles of dead plant material inputs to the soil continue to provide the microbial community with OM to sustain the cycle. Depending on the amount of N and P available, wetlands systems can be described as either N limited or P limited. Excess of either or both of these elements can cause eutrophication of a wetland (Westlake et al., 1998). Decomposition of Detritus It is well recognized that the breakdown of detrital material in ecosystems is an important pathway of energy flow and transformation of materi al to other forms (Webster and Benfield, 1986; Davis, 1991). Dead plant material which accumulates on the surface of the soil contributes C to microbial communities, which catabolize the C compounds for energy (McLatchey and Reddy, 1998; Reddy and DeLaune , 2008). The catabolism of C by microbes also effects the N and P cycles by breaking down organic molecules containing these elements. Under anaerobic conditions, organic N is mineralized to ammonium in an oxidative process regulated by microbes (Reddy a nd DeLaune, 2008). The ammonium may then be either immobilized an d taken up into


15 living biomass converting back to organic N , or leave the sys tem by entering the atmosphere either by volatilization, forming NH 3 , or by undergoing subsequent nitrification/de nitrification, forming N 2 and NO 2 (Reddy and DeLaune, 2008). Phosphorus, on the other hand, does not have a significant gaseous form . It remains in the system, either accumulating in the soil, taken up as new biomass, or leached into the water column. A significant factor in determining the fate of N during decomposition is the amount of C available and the microbial requirements for assimilation of C and N. Microbes maintain a C:N ratio of 10; since they obtain C and N for cell synthesis from detrital material, the C:N ratio of the detritus will determine if ammonification outweighs immobilization or vice versa in the C cycle (Reddy and DeLaune, 2008) . If the C:N ratio of detritus is greater than the critical C:N ratio needed by microbes to maintain t heir ratio of 10, then net immobilization occurs (inorganic N will be assimilated by microbes). If the C:N ratio is lower than the critical ratio, net ammonification occurs and organic N is released. The critical C:N ratio is dependent on the conditions; aerobic decomposition by microbes requires a C:N ratio of 25, while anaerobic conditions require a C:N ratio of 100 because the efficiency of C assimilation is lower (Reddy and DeLaune, 2008). Phosphorus storage or release similarly depends on the C:P r atio of detrital material and microbial requirements. Critical C:P ratios under aerobic conditions are as follows: C:P <200 will result in net mineralization (organic P transformed to inorganic and released), C:P between 200 and 300 yields no gain or loss of inorganic P, and C:P>300 results in net immobilization (inorganic P transformed to organic P and incorporated in biomass) (Reddy and DeLuane, 2008). Organic P from detrital material


16 has a number of possible fates depending on a variety of factors. It may leach from the plant tissue and enter the water column as dissolved organic P, become buried in the soil ove r time stabilize as soil OM , adsorb to mineral and clay surfaces, breakdown to inorganic forms through mineralization, hydrolysis, or photolysi s (Reddy and DeLaune, 2008). It can be expected that during decomposition, some organic P will be released and some will be stored long term. As with any environmental research topic, there are many factors that influence the process of decomposition and nutrient cycles, and therefore make it impossible to develop one universal model that holds true for all circumstances (Webster and Benfield, 1986). Different models of vascular plant breakdown have been described, though there are trade offs between the m. Selecting a model which gives good precision may not be generally applicable across different locations, and a model that may have the advantage of generality is less precise or realistic (Webster and Benfield, 1986). In order to obtain site specific data for the USJRB, the decomposition rates of various wetland vegetation litter were recorded and the relationship between variables that may influence those rates were analyzed. Upper St. Johns River Basin The north flowing St. Johns River (SJR) headwa ters begin with expansive floodplain marshes at the southern end of the system. The low laying freshwater marshes, settled between the Atlantic Coastal Ridge to the east and the Osc eola Ridge to the west, supply water that gradually flow north and beco me the headwaters of the longest river in the state. Today, the marshes h ave been drastically reduced in area following conversion of the marshes to agricultural land uses and other development (Brenner et al., 2001, SJRWMD, 2007; SJRWMD, 2008; SJRWMD, 2009) . Like much


17 of Florida throughout the first half of the 20 th century, the region was drained with the construction of dikes and ditches in order to provide flood protection and create usable land for agriculture and development. The drainage resulted in 62% of the 100 year floodplain and 42% of the annual floodplain being converted to agricultural land by the 1970s (SJRWMD, 2009). These modifications to the natural system significantly changed the hydrology and ultimately caused a shift in the vegetative species composition in the mosaic of headwater wetlands that makeup the upper basin of the SJR . After decades of drainage, the importance of floodplain marshes for the ecosystems services they provide began to be recognized. Beginning in the late 1970s , land in the Upper St. Johns River Basin (USJRB) marshes was acquired in parcels by the St . Johns Water Management D istrict (SJRWMD) to be protected. The SJRWMD manages these conservation area marshes with multiple purposes in mind: flood control, restor ing hydrology, improving water quality, protecting habitat for native flora and fauna, protecting cultural resources, and providing educational and recreation opportunities for the public (SJRWMD, 2007; SJRWMD, 2008; SJRWMD, 2009). Hydroperiod and fire regime are the two factors that most influence the characteristics of the marshes. The land management plans prepared by the SJRWMD use these factors as tools to restore or maintain the flood plain marshes; by controlling the water and fire regi mes . In addition to regular monitoring and chemical treatments of exotics , they can encourage the health of certain species and habitats and help restrict the spread of invasive species (SJRWMD, 2007; SJRWMD, 2008; SJRWMD, 2009) . However, past alteration s of the hydroperiod have already affected the species


18 composition, which has in turn resulted in changing the fuel type available for fires (SJRWMD , 2007) . Salix caroliniana has established communities in the USJRB where historical drainage and compartme ntalization of marshes has favored the shrub over the native herbaceous species (Quintana Ascencio et al., 2013). Fire alone is not a sufficient management tool in controlling Salix because it resprouts prolifically and within two years recovers stem dens ity (Lee et al., 2005). Fuel type is a factor the effectiveness of fire as a management tool in maintaining the floodplain marsh, and mechanical treatment of shrub species is necessary to encourage herbaceous growth and preferred fuel before fire can be effective (SJRWMD, 2007) . Addition ally, prescribed burns are challenging to schedule and carry out due to the proximately of highways and developments nearby, and the danger imposed by smoke. In general, changes to hydrology and lack of fire have facilita ted the invasion of shrub species in these areas (SJRWMD, 2007; SJRWMD, 2008; SJRWMD, 2009). Therefore, the management of the hydroperiod of these marshes as well as invasive species control plays an important role in the restoration and conservation of n ative habitats in this area . Quintana Ascencio et al. (2013) found that hydrologic manipulation with the correct timing can help control the establishment and growth of Salix in early stages of its life cycle. They recommend targeting seedlings and small saplings with extended periods of complete inundation to ensure mortality followed by drawdowns lasting for weeks to months to limit seed germination and establishment. Research by Brenner et al. (2001) has been done in the Upper St. Johns Basin (USJRB) t o quantify historical trends of sediment and nutrient accumulation using paleoecological methods. They found that sediment accretion rates in the USJRB were


19 on average 0.33 ± 0.05 cm yr 1 since around the year 1900, and since around 1963, the mean peat ac cumulation rate was 0.53 ± 0.11 cm yr 1 (Brenner et al., 2001). The higher rate in recent years was attributed to two factors: an increase in mass accumulation in recent years and compaction of older deposits (Brenner et al., 2001). They also found that since the 1970s, P accumulation rates are approximately 2 17 times higher and C and N accumulation rates were 1 3 times higher compared to 1920 rates. They concluded that increased P loading to the system drove increased rates in primary productivity, res ulting in higher C and N burial since the 1970s. The differences were more pronounced for P rates than C or N since both C and N are lost from the system by means of soil respiration and denitrification while P remains in the sediment (Brenner et al., 200 1). Soil subsidence and respiration in the USJRB were studied by Van Dam in 2012. Using soil cores retrieved from the sites in a laboratory study, he found that subsidence occurred at a rate of 4.7 cm yr 1 after drainage. Results from an in situ study of C flux showed total C respiration, including both CO 2 and CH 4, equal to 447 ± 79 g C m 2 yr 1 under flooded conditions and 2603 ± 338 g C m 2 yr 1 under drained conditions (Van Dam, 2012), with CO 2 being the main form of gaseous loss. Soil chara cteristics were analyzed in 11 sites, three of which correspond to the same sites sampled in SJMCA for this research and are listed below in Table 1 1. Nutrient flux from soil cores was measured in the laboratory and showed a significant increase in flux for N and P after cores were drained for 215 days, and no significant difference in dissolved organic C flux after drainage (Van Dam, 2012).


20 Table 1 1. Soil characteristics for sites in St Johns Marsh Conservation Area (Mean ± Std Error) . Site Mean Bulk Density (g cm 3) TC (%) TN (%) TP (g kg 1 ) S JM WIL 0.13 ± 0.004 50.1 ± 1.3 3.19 ± 0.06 0.63 ± 0.13 SJM SAW 0.17 ± 0.08 47.8 ± 1.0 3.13 ± 0.08 0.76 ± 0.14 SJM TYP 0.15 ± 0.004 45.5 ± 1.1 3.2 ± 0.04 1.26 ± 0.10 Source: (Van Dam, 2012) Objectives The objectives of this research are to determine the litter production of different vegetative communities in the USJRB and to determine to what extent they contribute to SOM. Both litter production and decomposition were measured to quantify the input of org anic matter from vegetation to the soil . Communities from site s with varying levels of disturbance were studied to determine to what extent different vegetation types contribute to the SOM pool and nutrient turnover. Species sampled are listed in Table 1 2. Table 1 2. Species of interest for this study. Scientific Name Common name Type Cladium jamaicense Sawgrass Sedge, emergent Panicum hemitomon Panic grass Grass Polygonum species Knotweed Typha domingensis Cattail Emergent Salix caroliniana Carolina willow Shrub or small tree Cephalanthus occidentalis Buttonbush Shrub or small tree


21 CHAPTER 2 LITTER PRODUCT ION Wetlands are among the most productive ecosystems on the planet (Mitsch and Gosselink, 1986; Keddy, 2000; Rocha and Goulden , 2009). Primary producers, which are able to build biomass using sunlight and atmospheric C , are the basis on which entire ecosystems depend for energy. The main factors which limit productivity in plants are often water and available nutrients; however in wetlands both of these resources are often abundant, which explains why productivity in wetland systems is so high (Keddy, 2000; Rocha and Goulden, 2009). Influential Factors Vegetation Community Vegetation communities are the defining feature of we tland types and determine many of the characteristics of a wetland . For example, a shift from a sawgrass dominated herbaceous marsh to woody shrubs such as willow, which has already occurred in the northern part of SJMCA, will affect the wildlife niches pr esent, affect biodiversity, increase the level of evapotranspiration, change the way fire is carried through the marsh and its effects, and potentially change the soil characteristics over time (Quintana Ascencio et al., 2011). Primary productivity is als o expected to differ amongst communities. Due to inherent differences between the photos ynthetic parts of different species, such as surface area and positioning of leaves, primary production will naturally depend on the vegetation communities present (Ke ddy, 2000, Rocha and Goulden, 2009). With shifting communities in the USJRB , it is important to understand the potential effects and feedbacks that may result from a chan ge in community structure . Figures 2 1, 2 2, and 2 3 show the distribution of wetland habitat types


22 across each conservation area marsh. These maps are useful to gain a general understanding of the greater region surrounding the individual study sites. Of the five sites sampled, four were herbaceous marsh (light green on the maps) and one was shrub dominated (yellow on the maps). Figure 2 1: Wetland habitat types in St. Johns Marsh Conservation Area . Figure 2 2: Wetland habitat types in Blue Cypress Marsh Conservation Area .


23 Figure 2 3: Wetland habitat types in Fort Drum Marsh C onservation Area . Hydrology The hydrologic regime of a wetland includes the flow pattern, duration, depth, and frequency of flooding, and seasonality . T he flow of water through a wetland generally has a positive relationshi p with the productivity. G reate r flows yield higher productivity while stagnant wetlands generally have lower productivity (Brinson et al. , 1981 ; Mitsch and Gosselink, 1986). This may be due to flowing water contributing nutrients to the system, while nonflowing wetlands may rely more on rainwater and therefore have lower nutrient inputs. In addition to flow, the hydroperiod offers insight to potential effects on productivity. A study by Mitsch and Ewel (1979) on cypress productivity found that sites with the most difference in water level between wet and dry season yielded the highest productivity, likely due in part to aeration of t he roots during the dry season. The hydrology also may indirectly affect net primary productivity by controlling the species composition, with certain sp ecies adapted for particular hydrologic regimes . In the Everglades, Cladium jamaicense productivity was found to


24 have a strong negative correlation with water levels (Childers et al., 2006). One year following a decline in Cladium productivity, Eleocharis sp. i ncreased i n stem density, indicating that a hydrologically driven decline in productivity for one species may result in a shift to another species more adapted to the hydroperiod (Childer et al., 2006). Other Factors Ecosystems such as wetl ands and marshes are inherently complex. In addition to species diversity/richness and hydrology, primary productivity is also a function of climate, nutrient input and av ailability, light availability , herbivory, interactions with micorrhizae, biogeochem istry in the soil (affecting respiratory requirements for plants), wind (involved in seed dispersal and po llination), fire frequency, among others (Mitsch and Gosselink, 1986; Wes tlake et al. , 1998; Rocha and Goulden, 2009). As a product of the interactio ns of all these factors, wetland productiv ity i s inherently site specific and t heref ore should be determined for studied marsh systems. Research Questions This study aims to answer the following: 1) What are the annual contributions of litter material to the soil elevation? 2) How much C is contributed to the soil by litter? Methods Site Descriptions The research was conducted in Fort Drum Marsh Conservation Area (FDMCA) , Blue Cypress Marsh Conservation Area (BCMCA) , and St Johns Marsh Conservation Area (SJ MCA) (Figure 2 4 ). The three areas comb ined encompass a total of 294 km 2 of mixed habitat types, with the majority consisting of floodplain marsh (SJRWMD 2007; SJRWMD, 2008; SJRWMD, 2009) . Three locations in SJMCA, and one each in


25 BCMCA and FDMCA were used in this study. The three sites in SJMCA have differing vegetation communities and are denoted as SJM WIL (a Salix Caroliniana dominated site), SJM SAW (a Cladium jamaicense site), and SJM TYP (a Typha domingensis site). The sites in BCMCA and FDMCA were mostly comprised of sawgrass, shrubs including willow and buttonbush, and a variety of other emergent wetland plants. Sites were selected to represent historically natural vegetation communities ( Cladium ) and com munities resultant from hydrologic disturbance ( Salix and Typha ), with SJMCA representing the most disturbed sites , BCMCA representing an intermediate hydrologic disturbance, and FD MCA representing the most pristine condition.


26 Figure 2 4 : Location of Upper St. Johns River Basin Conservation Area Marshes .


27 Figu r e 2 5 : Study site locations. Map data © 2013 Google , SIO, NOAA, U.S. Navy, NGA, GEBCO. St Johns Marsh Conservation Area Each of the three sites sampled in SJMCA are relatively close to the C 4 0 canal , the primary drainage canal for the US Army Corp of Engineers Upper Basin Flood Control Project, which runs north south adjacent to the eastern berm containing the area. The SJMCA is bordered by agricultural lands to the west, and Three Forks Marsh Conservation Area (TFMCA) to the east. East of TFMCA is more agricultural land which quickly transitions to suburban an d urban development, including I 95 approximately 16 km east of SJ MCA . The northernmost and lowest elevation site sampled was dominated by Salix caroliniana (SJM WIL) . The tree line of Salix began approximately 20 m from a side canal going west from the main C 40 canal, and the sampling location was several meters past the beginning of the tree line (F igure 2 6 ). The remaining two sites within SJMCA were dominated by Cladium jamaicense (SJM -


28 SAW) and Typha domingensis (SJM TYP), located approximately 5.6 km and 7.6 km south of SJM WIL, respectively. Both sites were simi larly west of the C 40 canal, but without proximity to any smaller side canals (figure 2 7 ). Cladium at SJM SAW was present in thick pockets with clearings of other herbaceous plants in between. Typha at SJM TYP was sparse with a dense cover of P olygonum and would be best described as a mixed herbaceous community. Of the three conservation area marshes studied, SJM is the most disturbed from its natural state due to hydrologic alterations. Figure 2 6 : Location of the SJM WIL study site. Map data © 2013 Google. C 40 canal


29 Figure 2 7 : Location of SJM SAW and SJM TYP sites . Map data © 2013 Google. Blue Cypress Marsh Conservation Area Blue Cypress Marsh Conservation Area (BCMCA) is located approximately 31.5 km south of the SJM WIL site, and is situated at a slightly higher elevation (Figure 2 8) . The sampling location within BCMCA is adjacent to an infrequently used airb oat trail, but no major canals . The vegetation community is very dense sawgrass ( Cladium ) punctuated with mixed shrubs including willow ( Salix ) and buttonbush ( Cephalanthus ) next to a smaller open area of floating and submerged aquatic vegetation. Sampling was conducted several meters into the stand of sawgrass . Around February/March of 2013, there appeared to be a new airboat trail that was not previously there running through the BCMCA site, and possibly running over one of the clip plots. The PVC poles were still in place, but it should be noted that there may have been disturbance to the plot. C 40 canal


30 Fort Drum Marsh Conservation Area Fort Drum Marsh Conservation Area (FDMCA) is the southernmost and highest elevation site in this study (Figure 2 8) . The sampling location at this site is the most remote compared to the other four, located near an infrequently used airboat trail and not close to any canals or boat ramps for visitor access. The vegetation community at the sampling site is very diverse and includes mixed woody shrub species (willow, buttonbush, wax myrtle), emergent vegetation (arrowhead, pickerel weed, false nettle, sawgrass), and submer ged vegetation (bladderworts). The Fort Drum site represent s the most natural hydrologic regime and vegetative community. Figure 2 8: Location of BCMCA and FDMCA sites. Map data © 2013 Google.


31 Biomass Harvest Above ground plant biomass was collected from the five study sites using direct harvesting methods . At all sites except for SJM WIL, three replicate 1 m 2 plots were randomly located . Above ground vegetati on and the litter layer within the plot were collected and brought back to the lab for analysis. Below ground biomass was not accounted for in this study. Plants were clipped down to near the soil surface, leaving roots and rhizomes intact. Biomass from each plot was sorted by species and separated by standing live, standing dead, or litter. Vegetation was dried, and weighed. Three replicates from each site were averaged to find the mean production per m 2 . Due to the density of trees at the willow si te, three 0.25 m 2 plots were used to collect biomass. The concentration of tree trunks at the site prohibited the use of 1 m 2 plots. When litter material was collected, there was stand ing water over the soil surface. W ith each scoop of leaf litter, wate r was disturbed and more flowed into plot. Litter material was abundant but may be slightly over estimated in plots because of the uneven distribution of litter due to movement by water. Clipped material included overhanging branches above the plot, as i f a vertical column extended above the plot . Branches within reach were included, higher branches out of reach were estimated by clipping lower branches to compensate. It is important to note that entire trees were not harvested for this study, and a lar ge proportion of the biomass at this site is in trees. Therefore it is difficult to compare this site with the four others using the same methods. Additional biomass at the willow site (standing live and standing dead willow) was collected in addition to the three plots, to ensure enough material for use in the litter decomposition study (C hapter 3) . The additional biomass (not included in plot weights) came from several nearby broken branches with an abundance of dead leaves that had


32 not yet dropped. En tire branches (no greater than 3 cm diameter, some much smaller) were collected to efficiently gather the hanging dead leaves, though the woody material was not used in the li tter decomposition study. After one year, plots were revisited and vegetation was collected again from each plot. Plots from BCM CA , FDM CA , and SJM SAW were reharvested. Plots from SJM TYP were lost, and perma nent plots were not installed at SJM WIL due to the drastically different vegetation community. The biomass collected included regrowth of the clipped plants over one year, as well as any othe r plants growing into the plot including new plants that may not have been rooted in the plot t he first year. Biomass was again sorted by standing live or standing dead and by s pecies, dried, and weighed to estimate the annual net primary productivity. Since the collected vegetation was bagged in the field by plot and sorted upon return to the lab, some of the vegetation collected in both 2012 and 2013 was difficult to smaller pieces of material that were of unknown species as well as species that contributed a very sm all proportion of the total biomass in the plot. To reduce the risk of losing plots in future studies, a more permanent marker should be installed indicating the area which was clipped, such as rebar which can be inserted into the soil in each of the four corners of the plot, and a more visible flagging method (perhaps something taller that airboaters would avoid). Regrowth after that level of cut back may not represent the same amount of new growth an intact plant has on a given year. Also, the destr uctive technique used in clipping and collecting vegetation created 1m 2 gaps in the wetlands, which may have a


33 different amount of growth and regrowth within a year than an undisturbed 1m 2 plot would have due to light and space availability opening up oppo rtunities for different species to move in. In future studies, a comparison of methods in addition to direct harvest including leaf area index, gas exchange, and/or mean shoot density might be helpful in providing a better estimate of the true net product ion (Westlake et al., 1998; Brinson et al., 1981). Contribution from vegetation to soil elevation was estimated by dividing the production of litter biomass (g cm 2 ) by the mean bulk density of the soil (g cm 3 ). The resultant volume (cm) is the maximum increase in soil elevation that can be expected if the total net production of litter became soil. Decomposition of litter will be accounted for in chapter 3. Results Figure 2 9 shows the mean biomass (g m 2 ) collected at each site, including standing live, standing dead, and litter material . Blue Cypress Marsh Conservation Area had the most biomass, with approximately 2433 g m 2 of total combined weight, most of which was comprised of Cladium . Figure 2 10 s hows the relative amounts of the different species of above ground biomass collected at each site. The standing dead material collected is an estimate of litter input for the following year.


34 Figure 2 9 : Standing live, standing dead, and litter b iomass at each site in 2012 . n=3. Figure 2 10 : Relative amount of different species collected at each site in 2012 .


35 In July 2013, plots were harvested again in BCM CA , FDM CA , and SJM SAW. Vegetation harvested was brought back to the lab, sorted by species, dried and weighed. The totals for each site are shown in Figure 2 11 . In 2013, BCMCA had the highest total combined biomass weight at approximately 635 g m 2 . Figure 2 11 : Standing live, standing dead, and litter biomass at each site in 2013. n =3 f or BCM and FDM, n =2 for SJM. Table 2 1: Above ground biomass (g) collected from clip plots in July, 2012 and July, 2013 . n =3. 2012 2013 Total g collected g m 2 Total g collected g m 2 BCM CA 4896 1632 669 1541 514 151 FDM CA 2166 722 8.29 1535 511 194 SJM SAW 2227 742 345 1214* 404 87.7 SJM TYP 1551 517 299 No data *Value adjusted based on two of the three clip plots (one was lost in the field). The total above ground biomass in the two clip plots re harvested was 809 g.


36 Potential input from the vegetation communities at each site based on annual production and a mean soil bulk density of 0.173 (Van Dam, 2012) are listed in Table 2 2. Litter contributions of C to soil were calculated using annual litter production and the %C in the standing dead material. This is an estimate of C input before decomposition. Table 2 2: Annual litter production contribution to soil. Note: decomposition not included. BCMCA (n =3 ) FD M CA (n =3 ) SJMCA SAW (n =2 ) E levation (cm) 0.297 0.0 88 0. 296 0.0 67 0. 234 0.0 51 C (g m 2 ) 94.4 53.0 44.8 2.0 20.5 4.6 Figure 2 12 : Annual litter contribution of C to soil. BCM CA and FDM CA n=3, SJM n=2. Discussion Litter production estimates based on the annual production in clip plots give insight to the rate at which OM in soil is accreted in the USJRB. Though the standing dead material represents the next input to the litter later, for this study the total annua l production was used to calculate input because it is assumed that all of the biomass


37 produced in a year will eventually transition to standing dead and ultimately litter material. The most hydrologically disturbed site (SJMCA) had the slowest soil build ing potential (Table 2 2). These rates do not take into acc ount subsequent decomposition (C hapter 3). Surprisingly, the amount of biomass collected in 2013 was similar across sites despite the sites having very different initial harvests in 2012. It was expected BCMCA , which had the highest biomass density in 2012, would have similarly high biomass production in 2013 when compared wi th the other sites. However, BCMCA was similar to FDM CA in 2013, with 514 151 g m 2 and 511 194 g m 2 respectively . A pos sible reason for this observation is light availability. Plots at the FDM CA site were scattered throughout the mixed marsh, and most had ample light reaching them. Plants in this site were of mixed heights and types with varied spacing and in general see med to allow more light to reach the water/soil surface. The site at BCM CA had clip plots located in a very dense stand of sawgrass. Once clipped, the plants within those plots were shaded on all sides by the tall sawgrass, which may account for the lowe r amount of biomass in 2013 than expected. Additionally, biomass collected at FDMCA was more diverse which m ay account for why regrowth after clipping was m ore substantial than at BCMCA , where Cladium dominated. The data between 2012 and 2013 indicate th at BCMCA has greater overall storage of C in biomass than FDMCA or SJMCA, but that storage does not necessarily translate to greater inputs of litter to the soil. It is also important to note that three randomly selected 1 m 2 plots may give a good representation of the immediate area, but a wider radius at each site would include a more diverse selection of biomass. Wetlands are known for being difficult to sample due to the clumped or gradient nature of the populations prese nt (Westlake et


38 al., 1998). There fore, the data presented here are noted to be representative of the immediate area for each site, and not the entire conservation area marshes which are comprised of a variety of wetland types and may have differing levels of litter product ion.


39 CHAPTER 3 LITTER DECOMPOSITION The process of decomposition of litter material can be divided into three main components: the initial leeching of labile C to the water column, the microbial processing of the material, and the physic al breaking of the material into smaller and smaller pieces (Webster and Benfield, 1986). Rate I nfluencing F actors Characteristics of the litter material that affect the rate of decomposition include the nutrient content in the plant material , the fiber quality , and the presence of chemical inhibitors (Webster and Benfield, 1986). This research effort will address the former two by analyzing nutrient content and fiber fractionation over time during decomposition. Environmental factors including temperat ure and flooding regime are further sources of variation in rates (Mitsch and Gosselink, 1986). Conant et al. (2008) found that more recalcitrant OM has greater temperature sensitivity than labile OM in litter and soil. Differences between sites in this study should give insight to some of the effects of hydrology. The relative importance and interactions between these factors is not a constant, but rather depends on the unique set of variables for a given site and litter type (Meentemeyer, 1978). S yner gistic effects on decomposition in a mixed species sample may also affect the rate (Schmidt, 2002). In a recent study in an upland environment, mixed types of litter material were found to have different microbial communities and chemical characteristics (specifically, phospholipid fatty acid concentrations) in the first 10 months of decomposition compared with single litter types (Chapman et al., 2013). The researchers concluded that a mixture of litter type contributes to synergistic decomposition by in fluencing the microbial community


40 development (Chapman et al., 2013). Accordingly, this study incorporates both individual and composite litter bags. Nutrients Of the litter characteristics, N in particular may have a strong effect on breakdown rates, and species with high initial N concentrations in the leaves typically break down faster than others ( Melillo et al., 1982; Webster and Benfield, 1986 ; Serna et al., 2013 ). When comparing Typha a nd Cladium , both of which are present at the sites in this study, Davis (1991) found that although Typha was better able to concentrate nutrients in biomass during growth than Cladium, both species had similar nutrient content in dead tissue because Typha did not retain those accumulated nutrients as well as Cladium during senescence. Along a nutrient gradient in the Everglades, Typha averaged 2.7 times the allocation of P to leaves and twice the N allocation compared to Cladium (Davis, 1991) . ad aptations to living in low nutrient environments include longer leaf longevity , smaller growth , and lower turnover rates of leaves. B eing adap ted to low nutrient conditions make Cladium likely to leech less nutrients in comparison with Typha (Davis, 1991. ) Serna et al. (2013) also found strong linkages between nutrient content and decomposition rates. They found that between three Everglades plants, Cladium jamaicense, Eleocharis cellulosa, and Nymphaea ordorata, comparatively high initial TN and TP leve ls in N. ordorata correlated to a rapid rate of decomposition, while the other two species had lower nutrient contents and slowe r rates. They suggest that the higher nutrient species better supported microbial decomposers, and that P content in litter may be a good indicator of decomposition rate.


41 Fiber Quality The quality of the litter material, which describes the relative amount of lignin and other carbon fractions, is well recognized as a significant indicator of the decomposition rate (Meentemeyer, 19 78; Roberts and Rowland, 1998; Osborne, 2005). Quality of the recalcitrance of the material, w ith highly recalcitrant components such as lignin being the most resistant to decomposers. In this study, the groups of interest in reference to fiber quality are: soluble fiber, hemicellulose, cellulose, and lignin. The soluble fiber group contains a var iety of compounds which are easily leeched and include cellular components such as sugars, fatty acids, and amino acids, while the hemicellulose and cellulose fractions are important for structural components of plant cell walls (Reddy and DeLaune, 2008). Cellulose, which is not found in algal cannot be catabolized by mammals (Wade, 2006). Microbes do the work of hydrolyzing cellulose for energy, and the resulting carbo hydrate products are able to be digested by consumers (Wade, 2006). Lignin is also an integral part of the structural material in vascular plants. Its structure is comprised of phenol groups linked in random branches with other C groups, and it does not have regular intervals of linkages (Reddy and DeLaune, 2008). Because of its low degradability, lignin content of litter material is a main factor in predicting decomposition rates (Meentemeyer, 1978; Roberts and Rowland, 1998).


42 Hydrology The environment in which litter material decomposes can fluctuate between aerobic and anaerobic conditions depending on the changing water regime. The inhibition of decomposition under anaerobic conditions has been established (McLatchey and Reddy, 1998; Wright and Reddy , 2001). However, due to the variability in litter type and other factors, the relationship between decomposition and hydrology is not always clear or as expected. For example, in a study on the effects on variable hydrology in a re created Everglades, r esearchers predicted that a decrease in water table depth would yield more rapid decomposition due to the more favorable aerobic environment, but did not see this effect in their experimental results (Serna et al., 201 3 ). They found that under dry conditi ons, the percentage of dry mass remaining over time was greater than or equal to the amount remaining under wet conditions, and concluded that the chemical characteristics of the species was a better predictor of decomp osition rate. Wright et al . ( 2013 ) f ound that for some litter types, the lowering of the water table may result in 10% more mass loss over timescales greater than 15 months. Their data also supported the idea that litter type is a more important factor in predicting decomposition rates than changes to the water table . Freeman et al. (1996) found that lowering of the water table in a peatland did not increase microbial respiration, but did result in increased enzyme stimulation resulting in increased mineralization. They conclude that the l owering of the water table does not stimulate synthesis of new enzymes by microbial activity but instead reduces the inhibition of existing enzymes, stimulating mineralization.


43 Research Questions This study focuses on the following questions for the USJ RB: 1) How does litter quality vary between species ? 2) How d o decomposition rates vary across sites with different hydrologic regimes ? 3) What factors are most strongly related to decomposition rate? Field Methods Litter bag and leaf pack studies have been used for decades as a method to monitor and quantify the mass loss (and sometimes leaf area loss, depending on the study) as litter decomposes (Schmidt, 2002). Litter bags can be made of a variety of materials, o ften sewn or stapled together, which ultimately encase the plant material in some kind of net or mesh so that they can be left in the field and incorporated in the litter layer and exposed to the environmental conditions at the site (Schmidt, 2002). This technique is well established and provides a relatively good way to retrieve material left in the field for monitoring over time, because the mesh bag or net prevents the sample from being lost (or outside material from being added) while still allowing fo r exposure to the elements. T here are several caveats to consider when collecting decomposition data from litter bags. The bag can protect the plant material inside from processing by detritivores (which contribute to decomposition) by excluding any gra zers which are larger than the size of the mesh (Schmidt, 2002). Invertebrates contribute to decomposition by direct consumption of the material as well as promoting more active microbial decomposition by breaking it down (which increases the surface area ) and inoculating the litter (Schmidt, 2002; Webster and Benfield, 1986). This leads to underestimating the true


44 decomposition rate. However, for inland marsh systems such as those in this study, herbivory is considered to be a minor factor in decomposi tion (Mitsch and Gosselink, 1986). If the mesh size is increased to allow access for invertebrates it would result in more particulate loss from the bag , possibly overestimating decomposition (Schmidt, 2002). Litter bags used in this study were made using vinyl coated window screening . Screen was cut into 25 cm by 36 cm rectangles, folded in half and heat sealed on the Samples were weighed, placed in the bag along with a label, and the final open end was heat sealed using an impulse sealer resulting in a closed packet of litter material (Figure 3 1). Two different species of vegetation were used at each site, each species bagged separately in individual bags and also toge ther in a composite bag. The composite bag is meant to represent the natural litter layer . Figure 3 1. Heat sealing final edge of litter bag after filling with sample and label using an impulse sealer. Photo courtesy of author. Vegetation material used in the litter bags came from the vegetation previously collected in July 2012 in the productivity clip plots ( C hapter 2 ). Standing dead material was used for most bags, except for some leaf material from the woody species. The


45 Salix and Cephalanthu s leaves were a mixture of standing live and standing dead leaves. This was due to the limited amount of standing dead leaves available at the time of collection for those species (BCM CA site) . All vegetation was thoroughly dried before weighing . The li tter material used in each bag at each site varied depending on which were mos The Salix/Cephalanthus bag (BCM B) was a mixture of leaf litter from both of these shrub species which were about equally present and scattered at the BCMCA purposes of this study; the composite bag at BCM CA was these two shrubs plus Cladium . This was done in order to have enough leaf litter to use in the bags, and because it was represen tative of the site. For material that was very abundant, 10 g samples were used for each bag. For everything else, 5 g samples were used. Table 3 1 describes the detailed contents of each bag at each site.


46 Table 3 1 Litter bag co ntents . Site Label Contents of bag SJM SAW A 10 g Cladium B 10 g Panicum C 5 g composite (2.5 g Cladium+ 2.5 g Panicum) SJM WIL A 5 g Salix B 5 g Polygonum C 5 g composite (2.5 g Salix + 2.5 Polygonum) SJM TYP A 10 g Typha B 5 g Polygonum C 5 g composite (2.5 g Typha + 2.5 g Polygonum) FDM CA A 10 g Cladium B 5 g Polygonum C 5 g composite (2.5 g Cladium + 2.5 g Polygonum) BCM CA A 10 g Cladium B 5 g Salix/ Cephalanthus C 5 g composite (2.5 g Cladium + 2.5 g Salix and Cephalanthus ) The litter bag decomposition study began in January 2013 at each of the five study sites. Four strands of litter bags were deployed at each site, one for each subsequent sampling time over the course of the following year. Each strand consisted of thre e types of litter material (A, B, and C bags), in triplicate, for a total of nine litter bags per strand. Bags were wired together and attached to a 1 m pvc pole that was inserted into the soil to mark the location of the bags. When deployed, bags were s paced out radially from the center PVC pole and placed so that they were lying flat at the soil surface next to each other, not overlapping (Figure 3 2).


47 Figure 3 2. Schematic of the litter bag deployment at each study site. Litter bags were collected four times per site over the course of a year. Table 3 2 summarizes the sampling dates for each site. Some variation on collection dates between sites is due to inclement weather preventing access to every site on each given trip. All bags were recove red except for TYP A1, TYP A2, and TYP B1 from the 4th sampling time, which were lost in the field. Mass remaining over time was defined by the material left in the bag after gentle rinsing with DI water upon retrieval from the field. Over time , the bags became well incorporated with the litter layer at the soil surface, and some bags contained soil that had passed through the mesh (Figure 3 3). Several bags had living vegetation growing through them, and a few were torn and appeared chewed on. Any mate rial that was obviously not part of the original sample was removed (such as living vegetation and small invertebrates) and the litter was gently rinsed of soil . The litter materia l was completely dried and weighed to determine the mass lost.


48 Figure 3 3. Top and bottom view of a litter bag after being incorporated in the litter layer (SJM WIL site). Photos cou rtesy of author. Table 3 2: Sampling timeline. The total number of days left in the field is in parentheses below the sampling date. SJM WIL SJM SAW SJM TYP BCM FDM All samples deployed 1 16 13 (0) 1 16 13 (0) 1 16 13 (0) 1 17 13 (0) 1 17 13 (0) 1st strand collected 4 11 13 (85) 4 11 13 (85) 4 11 13 (85) 4 11 13 (84) 4 11 13 (84) 2nd strand collected 7 11 13 (176) 7 11 13 (176) 7 11 13 (176) 7 24 13 (188) 7 11 13 (175) 3rd strand collected 11 20 13 (308) 12 17 13 (335) 11 20 13 (308) 11 20 13 (307) 11 20 13 (307) 4th strand collected 1 31 14 (380) 1 27 14 (376) 1 27 14 (376) 1 31 14 (379) 1 31 14 (379)


49 To describe the decomposition rate using the measurement of the mass lost over time, the following exponential decay formula was used (Brinson et al., 1982; Westlake et al., 1998 ; Bärlocher, 2005). or, where: x t = dry weight of remaining litter after period of time x 0 =dry weight of initial sample of litter t= duration of experiment and k = the rate coefficient constant The rate constant, k, w as calculated using an exponential trend line. This incorporated data points from each collection time over the course of a year, rather than just the final weight and initial weight. The reported k is the mean of three replicates for each species at eac h site. Lab Methods Retrieved samples were analyzed for C, N, and P content as well as C quality. Carbon quality was determined using methods from Ankom Technology, which have been established for approximating the various fiber fractions of the sample (Roberts and Rowland, 1998; Osborne, 2005). The method is borrowed from the feed and forage industry where it is used to determine the quality of forage f or digestion in livestock. The fractionation of neutral detergent fiber, acid detergent fiber, acid detergent lignin, and crude fiber are approximations for soluble fiber, hemicellulose, cellulose, lignin, and ash components of the sample (Osborne, 2005). Therefore, the percentage of each for the samples in this study should be considered an estimate rather than a precise number. A potential source of error could be that particles did not fully rinse out of the


50 bag after each digestion, but because the m ethod includes standard rinsing times and procedures, any error is expected to be consistent across samples. Dried samples were ground with a Wiley mill and passed through as size 20 mesh. For each sample, 0.5 ±0.05 g was weighed and placed in an F 57 F ilter bag from Ankom Technology and heat sealed using an impulse sealer. Samples were rinsed, dried, and weighed after each step in a series of digestions to estimate the relative proportions of soluble fiber, hemicellulose, cellulose, lignin, and mineral content in each. Figure 3 4 represents a schematic outline of the process. Digestions were performed using the Ankom 200 Fiber Analyzer and required chemicals from Ankom Technology. In the final step, samples were ashed in a muffle furnace at 550°C for 4 hours . Ankom Technology protocol indicated to ash the entire filter bag with the sample inside, using a blank bag for a correction factor. After one round of ashing, it was determined by laboratory staff that the filter bags should not be placed in the muffle furnace due to the filter bag material. For all samples presented here, the ashing step was modified. Instead of placing the entire bag in the furnace, bags were cut open and the sample was transferred to a pre weighed crucible for ashing. Howev er, because of the porous nature of the bags, there was a very small amount of sample that was lost during this transfer. Therefore, mineral content is slightly underrepresented and lignin content slightly overrepresented in this study.


51 Figure 3 4: Sch ematic of the f iber analysis p rocess . Samples were further ball milled and passed through a size 40 mesh before analyzing for nutrients. Phosphorus content was analyzed using the ascorbic acid method and spectrophotometer at the Wetland Biogeochemistry La b, University of Florida . Carbon and N were analyzed by Waters Agricultural Laboratories , Inc. in Camilla, Ga. Results Decomposition Rates Cladium jamaicense at the BCMCA and FDMCA sites had the slowest annual decomposition rates (k=0.21 0.04 and k=0.22 0.02 respectiviely). Cladium jamaicense at the SJM SAW site decomposed faster (k = 0.57 0.04) than at BCMCA and FDMCA. Individually bagged Salix litter also had a slow decomposition rate with k= 0.26 0.03. The fastest decomposition rates occurred at th e SJM TYP site with Polygonum with k=1.2 0.29. Polygonum also had fast decomposition rates at other sites (k= 1.09 0.09 and 0.83 0.09 at SJM WIL and FDMCA sites, respectively). Individually bagged Panicum as well as composite bags of Cladium/Panicum,


52 Cla dium/Polygonum, and Salix/Polygonum had mid range decomposition rates (k= 0.65 0.05, 0.56 0.02, 0.50 0.03, and 0.58 0.03, respectively). Table 3 3 summaries the mean k for each set of samples. Table 3 3: Mean a nnual decomposition rate constants. n=3 Samp le ID Species k WIL A Salix caroliniana 0.26 0.03 WIL B Polygonum species 1.09 0.09 WIL C Salix caroliniana and Polygonum species mix 0.58 0.03 SAW A Cladium jamaicense 0.57 0.04 SAW B Panicum hemitomon 0.65 0.05 SAW C Cladium jamaicense and Panicum hemitomon mix 0.56 0.02 TYP A Typha species 0.85 0.09 TYP B Polygonum species 1. 2 0.29 TYP C Typha species and Polygonum species mix 1.01 0.24


53 Table 3 3 Continued: Mean annual decomposition rate constants. n=3 . Sample ID Species k BCM A Cladium jamaicense 0.21 0.04 BCM B Salix caroliniana , Cephalanthus occidentalis 0.59 0.06 BCM C Cladium jamaicense , Salix caroliniana , and Cephalanthus occidentalis mix 0.38 0.05 FDM A Cladium jamaicense 0.22 0.02 FDM B Polygonum species 0.83 0.09 FDM C Cladium jamaicense and Polygonum species mix 0.50 0.03 Figure 3 5 summarizes the mean decomposition rate constants of the individually bagged species. S amples from composite bags are not included in this figure . Typha and Polygonum decomposed the fastest, while Cladium and Salix were the slowest. Surprisingly, the rate of the Salix/Cephalanthus decomposition was faster than Salix alone. Recall that the Salix/Cephalanthus bag was a mixture of leaf litter from both of these shrub spe study. Figure 3 6 summarizes the average decomposition rate constant for each site.


54 Figure 3 5: Mean annual decomposition rate constants by species. Means were calculated using individually bagged samples. There was not a statistical difference between the rates of Cladium and Salix (p =0.7695). Rates between other species were statistically different. Figure 3 6 : Mean annual decomposition rate constants by sit e . n = 9 . F iber Analysis The results of the fiber analysis for each site and each sam pling period are summarized in F igure A 3 in the appendix. The mineral component of the litter was 0.63 0.58 1.02 0.39 0.52 0 0.2 0.4 0.6 0.8 1 1.2 1.4 SJM-WIL SJM-SAW SJM-TYP BCMCA FDMCA Mean k Site


55 near negligible for all samples (<1%). For almost all sites and species, the placement of litter in an individual bag or a composite bag did not influence the proportion of soluble fiber, hemicellulose, cellulose, and lignin over time. The only exception was at the BCMCA Salix/Cephalanthus . When mixed with Cladi um in the composite bag (labeled C b ), the Salix/Cephalanthus had a 7 15% higher proportion of lignin, and 9 13% less soluble fiber than the Salix/Cephalanthus in the individual bag (labeled B) across the first three samplin g times (Figure A 4). Table 3 4 shows the p values for fiber quality comparisons between individual and composite bags. Table 3 4 : Evaluation of difference between composite and individual samples of each species. Listed p values (non parametric, Steel The only sam ple that had significantly different results between composite and individua l bags was Salix/Cephalanthus. soluble fiber hemicellulose cellulose L ignin Cladium (n= 71) 0.184 0.0641 0.0855 0.0562 Panicum (n=24) 0.1749 0.4705 0.5834 0.4025 Polygonum (n=68) 0.4766 0.8753 0.7824 0.6334 Salix (n=24) 0.4705 0.9310 0.7075 0.7075 Salix/Cephalanthus (n=24) <0.0001 0.0003 0.0004 <0.0001 Typha (n=22) 0.1213 0.1985 0.3068 0.4098 Of all samples, litter material containing the highest proportion of soluble fiber was Typha material in both the individual and composite bag from the last sampling period at the SJM TYP site, with approximately 48% soluble fiber. The least amount of so luble fiber was in Cladium at the BCM CA and FDM CA sites at the third sampling time,


56 with approximately 15%. Cladium at the BCM CA site across the first three sampling periods had the highest proportion of hemicellulose ( 38%) while Salix from both the composite and individual bags from the first two sampling times had the least with less than 3%. Polygonum (from individual bag, 1 st collection time and composite bag, 2 nd collection time at FDM CA ) and Panicum (from individual bag, 2 nd colle ction time at SJM SAW) had the highest proportion of cellulose at 43%. The lowest cellulose was found in Typha from the last collection at the SJM TYP site, in both the individual and composite bag with between 11% and 12% cellulose. Lignin, a component of most interest in relation to decomposition rates, was highest in Salix at the SJM WIL site at the 3 rd collection time, with 61% lignin. The smallest proportion of lignin was found in Polygonum ( 9.5%) from the composite bag at SJM TYP from the last co llection time, Cladium ( 10%) from the individual bags at BCM CA and SJM SAW, and the composite bag at SJM SAW all from the first collection time, and Polygonum from the FDM site composite bag from the 2 nd collection time ( 10.5%). For detailed graphs showi ng the change in soluble fiber, hemicellulose, cellulose, and lignin over time for each species at each site, refer to F igures A 4 and A 5 in the appendix . Because lignin content is of particular interest in predicting decomposition rates, the percent age lignin of the different species sampled was compared using one way ANOVA. S pecies were significantly different in their lignin content (p<0.0001) except for the following: Typha and Cladium (p=0.5562), Salix/Cephalanthus and Salix (p=0.1381) and Typha an d Polygonum (p=0.0619).


57 Figure 3 7 : Comparison of mean % lignin content in different species . Species with different letters have significantly different lignin content (Steel Dwass, p<0.05). Cladium n=71, Polygonum n=67, Panicum, Salix, and Salux/Cephalanthus n=24, Typha n=22. Nutrients Percent totals Total C content of all samples ranged from approximately 38% to 53%, with the lowest % C in Polygonum and Typha samples and the highest %C in the Salix and Salix/Cep halanthus samples. Total N ranged from approximately 0.5% to 3.4% with Cladium at the low end with the least %N and Salix, Cephalanthus, Polygonum, and Typha all with higher N content. Total P ranged from <0.01% in Cladium to 0.22 0.23% in Polygonum and Typha . A AB B C C D


58 Percent C increased over time for all species of litter material except Typha and had a moderately significant correlation with time for Cladium, Panicum, Polygonum, and Salix but no significant correlation for the Salix/Cephalanthus sample or Typha (Figure 3 8 , T able 3 8). Over time, %N increased in the litter material for all species (Figure 3 9 ). Correlation coefficients were significant for all species, but most significant for Typha with r= 0.96 (T able 3 5 ). Percent P increased over time for all samples except Salix and was most strongly correlated with Typha (r=0.93) (F igure 3 10 , T able 3 8) . Table 3 5 : Coefficients of correlation between %nutrients and time . Cladium Panicum Polygonum Salix Salix/Cephalanthus Typha %C vs time 0.59 0.5 3 0.62 0.42 0.36 0.31 %N vs time 0.55 0.79 0..59 0.67 0.71 0.96 %P vs time 0.45 0.07 0.06 0.26 0.73 0.93 F igure 3 8 : Percent C in litter over time .


59 Figure 3 9 : Percent N in litter over time . Figure 3 10 : Percent P in litter over time .


60 Mass difference The change in mass over time of C , N , and P in the litter material was calculated using the measured weight of the samples after each retrieval and the corresponding percent nutrient content for that sample. Although for many samples the nutrie nt values became more concentrated (higher percentage over time), the mass difference indicates that overall, nutrients were released from some of those samples. Figures 3 10 , 3 11 , and 3 1 2 show the overall change in mass of C , B , and P over time. As ex pected during decomposition, C steadily declined over time for all samples. Nitrogen had mixed results, with about half of the samples losing nitrogen over time ( Polygonum, Panicum, and Salix) while the others generally gained nitrogen over time ( Cladium and Typha ). Phosphorus was generally lost from the litter material as decomposition occurred, with the exception of Typha and some of the Cladium samples, which gained P . Overall, treatment of litter in an individual bag or composite bag did not have a si there was no significant difference between the treatments for any species. For N, the only sample that showed a significant difference was the Salix/Cephalanthus bag, which also differed in fiber quality between the individual and composite treatments. Percent N was higher for Salix/Cephalanthus in the composite bag than when bagged individually. For C, placement in mixed or individual bag made a difference for Panicum, Polygonum, and Typha but not for Cladium, Salix, or Salix/Cephalanthus.


61 Figure 3 11 : Change in mass of C over time; A) SJM WIL site, B) SJM SAW site, C) SJM TYP site, D) BCM CA site and E) FDM CA site . Data used was from individually bagged samples only (composite samples excluded). A B C D E


62 Figure 3 12 : Change in mass of N over time from each litter bag type. Data used were from individually bagged samples only (composite samples excluded) . Positive change in mass (gain in N) indicates net immobilization and negative change (loss of N) indicates net mineralization. Sets represent different collection periods over the course of 1 year.


63 Figure 3 13 : Change in mass of P over time from each litter bag type . Data used were from individually bagged samples only (composite samples excluded) Positive change in mass (gain in P) indicates net immobilization and negative change (loss of P) indicates net mineralization. Sets represent different coll ection periods over the course of 1 year.


64 Ratios The change in C:N ratio over time for each s pe cies is depicted in F igure 3 13 Panicum and Typha showed the strongest correlation with r values of 0.93 and 0.84. Figure 3 14 : Carbon to N ratio over time for each species . Multivariate Analysis For all comparisons related to the decomposition rate, data from individually bagged samples were used. Composite bags were not included in these statistics because the rate calculated for those bags does not necessarily reflect the rate of each species within the bag, but rather the overall rate of the two. Percent lignin content, percent N content, and the C:N ratio were chosen as factors to analyze based on a literature review of their potential i nfluence on decomposition rates. For Cladium and for Salix/Cephalanthus , there was moderate correlation between the rate constant and


65 each of the three variables. Table 3 6 lists these coefficients of correlation highlighted in green; all other samples s howed weak or no correlations. Table 3 6 : Coefficients of c orrelation between decomposition rate constants (k) for each species and different variables. Figure 3 1 5 shows the general relationship between litter quality and decomposition rate. In low quality samples (%lignin >40), the lignin content was moderately correlated with the decomposition rate (r= 0.374). For higher quality samples, the %lignin was not correlated with the decomposition rate. % lignin %N C:N Cladium k 0.4 043 0.6 300 0.6 248 Panicum k 0. 1377 0. 0868 0.1 047 Polygonum k 0.0287 0. 1514 0.04 35 Salix k 0.0942 0. 2605 0.3 242 Salix/ Cephalanthus k 0. 5866 0.5 496 0. 6333 Typha k 0.1 008 0. 2116 0. 2353


66 Figure 3 15 : Relationship between litter quality and decomposition rates . Discussion Field Decomposition In comparing the many factors that influence decomposition processes, it is important to note that the species used in litter bags at each site was determined by what was present at each site. Fortunately Cladium and Polygo num were common at multiple sites, so there was some opportunity for comparison across sites . However, Typha was only used at the SJM TYP site, which makes it difficult to discern whether to attribute differences in decomposition rate to the species quali ties or the conditions at the site. In reality , it is expected to be a combination of these factors. For the purposes of characterizing each site, this study is useful because the data collected represents the natural condition and species present at eac h site. However, for better analysis on


67 the relative influence of species type and environmental conditions on the rates of decomposition, it would be helpful to use all the same species of litter across all sites. As a substitute for species type, litte r types were grouped by % lignin for comparison across sites. A future lab study utilizing the consistent litter material, soil cores from each site, and controlled temperature and various hydrologic conditions would be useful in understanding the role of water level fluctuation on decomposition rate, such as in Wright et al. ( 2013 ) . There are tradeoffs to the focusing on fewer variables by doing a lab study; the advantage is more control on the particular variable of interest, but a major disadvantage is removing so many of the factors and complex interactions that occur in situ including wind, sun, detritivores, additional inputs on litter and subsequent burial, and more. Some of these may be simulated as best as possible in a lab study, but results will still likely differ from an in situ study. Influential Site Conditions Biotic influence Invertebrates are a natural part of the decomposition process and contribute by shredding litter material into smaller pieces, which increases the overall surface are a available for microbial activity. It was observed from the first collection trip to the TYP site that an ant mound had built up around the center PVC pipe that the litter bags were attached to. Ants were very active in and around the bags, and had chew ed through some of the screen material. Though this case was the most obvious, invertebrate activity and damage to the screening material was not unique to this site; many samples experienced some small tears or holes in the bag and virtually all bags had evidence of invertebrate presence. Because these animals help to break up the detrital material into


68 smaller pieces, this influence is not considered a problem with the experiment but rather a part of the natural processes being studied. However, in ord er to better represent these factors amongst the samples, the design of the litter bag layout should have been randomized rather than in orderly strands. For example, the ant mound situation was mainly experienced by the composite bags which were closest to the PVC pipe, and not by the individual species bags which were further out radially. Litter bags were also impacted by vegetation growing directly through them, gradually creating a hole in the screen which it could push through. This was most common at the SJM WIL site, where the bags became incorporated into the litter layer the fastest and most thoroughly of all the sites, but also occurred occasionally at other sites. Not only did plants grow through the screening, they also grew directly into the litter mater ial inside the bag. Figure 3 16 shows a piece of litter material, peeled open longwise, which became substrate for roots of other vegetation to grow in and on. As much as possible, any plants which grew into the bags were removed before meas uring the final weight, including the small roots growing through the litter. Therefore, any impact new plant material had on the final weight of the bags should be negligible. However, it is possible that the live plants in the bags play a role in nutri ent cycling occurring within the litter bag. This could provide a direction for further research at these sites in regards to nutrient pools and bioavailability.


69 Figure 3 16 : Roots growing into litter material. Photo courtesy of author. Abiotic influence Due to the efficiency of oxygen as an electron acceptor in microbial processes, it is expected that the status of an environment as either aerobic or anaerobic will influence decomposition processes. Of the five study sites, only BCMCA had inun dated conditions for the entire duration of the study. Fort Drum Marsh Conservation Area was inundated for most of the study, a total of 357 out of 379 days. Exact hydrologic data is not available for the three sites in SJMCA, however based on observatio ns during sampling trips those sites appeared to be consistently drier than both BCMCA and FDMCA. This may partly explain why Cladium decomposition rates were much faster at the SJM SAW site than at BCMCA and FDMCA. The SJM TYP site, which had the highes t decomposition rate constant on average of all sites (k=0.95), was very dry at the first collection time. Exposure to sunlight may also play a role in influencing decomposition. Photochemical reactions are a known mechanism that influences the breakdown of dissolved organic P in lakes, but less is known about its significance in wetlands (Reddy and DeLaune, 2008). In this study, vegetation provided complete or near complete


70 shade at three of the sites (SJM WIL, BCMCA, and FDMCA) while SJM SAW and SJM TY P had partial shade or more seasonal changes that impacted amount of shade. Direct sunlight might also contribute to accelerated decomposition by raising the temperature and promoting biological activity. Although quantitative data was not collected, obs ervations of each site indicate that sunlight was probably not a major factor at any of the sites except possibly for SJM TYP, which had litter bags in direct sunlight for part of the year. Fiber Analysis In general, the relative proportion of lignin for e ach species increased over time (Figure A 4 D ). This is expected due to the highly recalcitrant nature of lignin; during the course of decomposition, other components are catabolized and the remaining lignin is proportionally greater. The soluble fiber fraction had mixed results depending on species but most surprising was the clear increase over time in Typha litter. As the most labile fraction, it was expected that the soluble fib er would significantly decrease early on, but instead it shows an increase or stable percentage for many of the species sampled (Fig ure A 4 A ). One possible explanation is the addition of new material to t he litter bags, such as biofilm growing on the lit ter and/or algae, which would contribute primarily to the soluble group. If this is the case, it explains why there is an increase proportions of each of the othe r fractions are off set as these were calculated as parts of a whole; if the soluble fraction has outside inputs, it will dilute the calculated relative proportions of each of the other fractions. This is important to keep in mind when comparing data from this study with other studies . Hemicellulose either remained fairly constant or decreased slightly over time, wh ile cellulose had a general tren d of


71 decreasing, indicating that the microbial communities are utilizing cellulose as their main energy source (Fig. A 4 C &D) . Surprisingly, Salix litter had a fairly consistent breakdown of fiber quality throughout the decomposition period. I t was expected that the leafy litter would be the fastest to break down, but instead it had one of the highest lignin c ontents and slowest decomposition rates. Cladium at BCM CA and FDM CA also had little change in the fiber quality over time, which is not as surprising as Cladium is known for having very slow rates of decomposition. However, Cladium at the SJM TYP site sh owed a clear decrease in %cellulose over time ( Figure A 5 C) which corresponds to the higher rate found a t that site as well. A probable explanation for the accelerated rate for Cladium at SJM TYP is the hydrology; this site was generally drier and fluctu ated between wet and dry much more than BCM CA or FDM CA , which were inundated for all or most of the study. Nutrients Based on the C:N ratio and the mass differences in N over time, net immobilization/net mineralization should be predictable. For high C:N ratios (>25) in litter , net immobilization typically occurs . At that level, m icrobes are unable to meet their N requirement from the litter material alone and they must take up nitrate and ammonium ions to build biomass (Brady and Weil, 2008; Reddy and De Laune, 2008). The critical C:N ratio in determining net immobilization/net mineralization is very different for anaerobic conditions. Poor energetic efficiency under anaerobic conditions requires less N and therefore a much higher C:N ratio (up to 100) can yield net mineralization during decomposition (Reddy and DeLaune, 2008). Cladium and Typha had high C:N ratios (F igure A 6), and therefore should have net immobilization occur.


72 Figure 3 12 supports this, indicating a gain in N over time in the litte r of Typha and most Cladium samples . Salix had amongst the lowest C:N ratio, and based on this would likely have net mineralization occur. This is not clearly supported by Figure 2 11, indication that other factors are involved in determining the fate o f N for Salix litter . Figure 3 12 shows the difference in N (g) over time, with negative numbers indicating a loss in N over time (net mineralization) and positive n . Since aerobic/anaerobic conditions may quickly change based on water level, and that si gnificantly changes the critical C:N ratios required to predict immobilization/mineralization, the calculated mass differences in N in the litter is a better indicator of net immobilization or net mineralization. It is interesting that at most sites, one l itter type appears to be undergoing net mineralization while the other experiences net immobilization. Further investigation is needed on what causes this; it does not appear to be related to C:N ratio of the litter types or the fiber quality. Multiple Va riables Figure 3 15 shows that for samples with high lignin content, the quality of the litter is likely an influential factor in determining the decomposition rate, but for those samples with moderate to lower lignin content (higher quality litter), other factors may determine the rate. This is evident in the wide range of decomposition rates for samples with lignin content in the 10% 25% range. A logical explanation for this spread is that for those samples, hydro logy may be the driving factor in determining decomposition rate. Of that group, the far left samples (smallest k) were from the sites with the longest hydroperiod. Moving toward the right, samples are from sites that were generally drier throughout the year.


73 Summary Cladium and Salix de composed slowest of the species sampled, while Typha was the fastest . Though the expectation was that each species would have a unique rate constant, Cladium and Salix were statistically the same. Rates were fastest at the SJM TYP site (k= 1.02 0.23 ), fol lowed by SJM WIL (k=0.63 0.3) and SJM SAW (0.58 0.05 ). The slowest rates occurred at BCM CA (k=0.39 0.14 ) followed by FDM CA (k=0.53 0.22 ) . In general, the most natural sites (FDMCA and BCMCA) had the slowest rates of decomposition and the most hydrologica lly altered sites (SJMCA) had faster rates. This may be the result of both the difference in water levels as well as the shift in species composition , especially at the SJM TYP site where both species sampled decomposed quickly . Under the assumption that hydrologic regime is a factor in determining decomposition rates, results here suggest that if disturbance similar to the SJMCA occurs at other sites in the USJRB, a side effect may be faster decomposition of litter. While the impact to N and P cycles is more difficult to predict, the impact to the C cycle is clear. Fa ster decomposition results in more of the detrital C being release d annually and less input to the SOM pool. Alternatively, hydrologic restoration of SJMCA to a more natural state may decr ease the decomposition rates to reflect rates similar to those f ound for BCMCA and FDMCA, thereby enhancing C storage. Factors that relate to the rate constant were not consistent for all samples. For Salix and Cladium, lignin content and C:N ratio were correlated with decomposition rate , but for other species there was no correlation. Findings suggest that for species with higher lignin content (low quality) , the quality may play a larger role i n determining the


74 decomposition rate . This has implicatio ns for the spread of Salix in the USJRB. With surprisingly poor litter material, Salix litter will accumulate over time and may provide better C storage in detrital material compared with other species.


75 CHAPTER 4 CONCLUSIONS This research aimed to characterize the litter production and decomposition rates in three conservation area marshes in the USJRB . In conjunction with other research efforts in recent years, the data collected in this study will be useful for SJRWMD in th e development of a subsidence prediction model for the USJRB . Accumulation of OM in a wetland, which counteracts subsidence, can result from bot h increased primary production and decreased decomposition in wetlands (Mitsch and Gosselink, 1986). The net r esult depends on whether or not C loss through subsidence is outweighed by accumulation of soil built by litter fall . Utilizing annual production estimates (C hapt er 2) and decomposition rates (C hapter 3) for USJRB marshes, the annual net input of biomass to the soil was determined with the exponential decay rate formula . Where: = the litter production in one year (estimated by biomass collection after one year post harvest) k = the mean k determined for each site t = one year and is the remaining biomass that contributes to the soil


76 Table 4 1: Calculated values of vegetation input to soil. k (g m 2 ) Annual Litter Production (g m 2 ) Annual Soil Input BCMCA (n=3) 0.39 203 114 138 77 FDMCA (n=3) 0.52 98 4.5 58 2.7 SJMCA SAW (n=2) 0.58 46 10 2 6 5.7 Annual addition to soil elevation can now be more accurately estimated using the biomass of detrital material remaining after one year of decomposition . Results are summarized in table 4 2. Rates found for the USJRB are similar compared to rates recorded for comparable systems. Glaser et al. (2012) found that over the past 4000 years, the sediment accretion rate in the Everglades was 0.2 mm yr 1 . Van Dam (2012) found that the potential s ubsidence rate for SJMCA following drainage was 4.7 cm yr 1 . Table 4 2: Annual litter contribution to soil elevation after decomposition. BCMCA FDMCA SJMCA SAW Elevation (cm) 0.201 0.0 59 0.176 0.0 67 0. 05 5 0. 078 One of the current issues in the USJRB is the spread of Salix . With a shift in vegetation community such as this , the litter inputs to the soil will also change. This research effort sheds light on the characteristics of different litter types that will aid in predicting some of the effects of any future changes in species composition . Based on the findings from this research , Salix has a very poor litter quality and relatively slow decomposition rate. This held true for Salix at both the BCM CA site, which was inundated throughout th e study, and at SJM WIL, a more hydrologically variable site .


77 This indicates that expansion of Salix i n these marshes will contribute high lignin litter to the soil persist longer relative to other species . To unde rstand the full potential of the amount of litter Salix contributes, further study is needed du e to the inapplicability of the clip plots at the SJM WIL site in this study. Measuring annual litter contributions using l itter fall traps would be useful for this species. It was observed that the SJM WIL site had the most new litter accumulation occur on top of the bags over the course of the year , by far . The bags became complete ly incorporated into the soil surface due to burial by new litter fall, while a t the other sites the bags were generally still laying on top of the soil surface. Based on this observation and the data on Salix decomposition rate, it can be anticipate d that an invasion of Salix would be highly productive and result in a greate r increase i n marsh elevation relative to herbaceous communities. In terms of water level management, it appears that a shortened hydroperiod m ay increase rates of litter decomposition for material with lignin content in the 10% to 25% range, but may not affect material with higher lignin content. This indicates that for Salix, litter decomposition will be slow regardless of inundation frequency. The slow rate of accretion found for USJRB marshes indicates that the restoration of soils which are lost th rough oxidation will be a very slow process . This (2012) that water tables should not be lowered past 10 cm below the soil surface to protect against soil oxidation.




79 Figure A 1: Mass loss over time of litter bag sample at each site, with corresponding exponential decay equations. A) SJM WIL, B) SJM SAW, C) SAW TYP, D) BCM, E) FDM. A B C D E


80 Figure A2: Percent mass loss over time by species (data from sites combined; composite bag rates excluded).


81 Figure A3: Fiber fractions of each sample.


82 Figure A4: Change in quality over time, comparing composite bags and individual bags. The only obvious difference between a leaves from two shrub species, and the com posite bag was a mix of three species total (Salix, Cephalanthus, and Cladium) This sample was located at the BCM CA site. The added complexity of two and three species may account for the difference, whereas the other bags contained only one and two spe cies.


83 Figure A5: Change in quality over time for in each species at each site.


84 Figure A6: C:N Ratios .


85 LIST OF REFERENCES Bolin,B . , Degens,E.T, Duvigneaud,P and Kempe,S., 1979. The global biogeochemical carbon cycle The Global Carbon Cycle, Edition: SCOPE Report 13, Chapter: 1, pp.1 56 . Publisher: J. Wiley & Sons, Chichester, New York, Brisbane, Toronto Brenner, M., Schelske, C.L. and Keenan, L.W., 2001. Historical rates of sediment and nutrient accumulation in marsh es of the Upper St. Johns River Basin, Florida, USA. Paleolimnology, 26, 241 257. Brinson, M.M., Lugo, A.E. and Brown, S., 1981. Primary Productivity, Decomposition and Consumer Activity in Freshwater Wetlands. Annual Review of Ecolog y and Systematics, Vol. 12, 123 161 . Childers, D.L., Iwaniec, D., Rondeau, D., Rubio, G., Verdon, E. and Madden, C.J. 2006. Responses of sawgrass and spikerush to variation in hydrologic drivers and salinity in Southern Everglades marshes. Hydrobiologia, 569; 273 292. Cona nt, R.T., Drijber, R.A., Haddix, M.L., Parton, W.J., Paul, E.A., Plante, A.F., Six., J. and Steinweg, J.M. (2008). Sensitivity of organic matter decomposition to warming varies with its quality. Global Change Biology. 14, 868 877. Daoust, R.J. and Childer s, D.L., 1998. Quantifying aboveground biomass and estimating net aboveground primary production for wetland macrophytes using a non destructive phenometric t echnique, Aquatic Botany, 62:2, 115 133 . Davis, S.M., 1991. Growth, decomposition, and nutrient r etention of Cladium jamaicense Crantz and Typha domingensis Pers. In the Florida Everglades . Aquatic Botany, 40; 203 224. Ewing, J.M. and Vepraskas, M.J., 2006. Estimating primary and secondary subsidence in an organic soil 15, 20, and 30 years af ter dra inage. Wetlands, 26 (1); 119 130. Freeman C . , Liska G . , Ostle N . J ., Lock, M.A., Reynolds, B. and Hundson, J . 1996. Microbial activity and enzymic decomposition processes following peatland water table drawdown. Plant and Soil 180:121 127 . Glaser, P.H., Vo lin, J.C., Givnish, T.J., Hansen, B.C.S. and Stricker, C.A., 2012. Carbon and sediment accumulation in the Everglades (USA) during the past 4000 years: rates, drivers, and sources of error. Geophysical Research Biogeosciences. 117 , G03026, doi:10.1029/2 011JG001821 Google Earth, 2013. Upp er Saint Johns River Basin sites lat 27.789096 ° , lon 80.567800 ° elev 20 ft, eye alt 55.15 mi. [Viewed September 2013]. Methods to study litter decomposition: a practical guide. Dordrecht, Springer. 37 42.


86 Kayranli, B., Scholz, M., Mustafa A. and Hedmark, A., 2010. Carbon storage and fluxes within freshwater wetlands: a critical review. Wetlands, 30; 111 124. Lee, M.A.B., Ponzio, K.J., Miller, S .J., 2005. Response of willow ( Salix Caroliniana Michx.) in a floodplain marsh to a growing season prescribed fire. Natural Areas Journal. 25(3); 239 245. Marín Muñiz, J.L., Hernández , M.E. and Moreno Casaolam P. , 2014. Comparing soil carbon sequestrati on in coastal freshwater wetlands with various geomorphic features and plant communities in Veracruz, Mexico. Plant Soil, 378; 189 230. McLatchey, G.P. and Reddy, K.R., 1998. Regulation of organic matter decomposition and nutrient release in a wetland s o il. Environmental Quality 27, 1268 1274 . Meentemeyer , V. 1978. Macroclimate the Lignin Control of Litter Decomposition Rates . Ecology,59 ( 3 ) 465 472 . Melillo, J.M., Aber, J.D. and Muratore , J.F. 1982. Nitrogen and Lignin Control of Hardwood Leaf Litter De c omposition Dynamics. Ecology 63, 621 626. Mitsch, W.J. and Ewel, K.C., 1979. Comparative Biomass and Growth of Cypress in Florida Wetlands. American Midland N aturalist, Vol. 101, No. 2, 417 426 . Morris, D.R., Glaz, B. and Daroub, S.H., 2004. Organic so il oxidation potential due to periodic flood and drainage depth under sugarcane. Soil Science, 169 (8); 600 608. Osborne , T.Z. Dissertation , 2005. Characterization, mobility, and fate of dissolved organic c arbon in a wetland ecosystem . Dissertation, University of Florida, Gainesville, Fl. Osborne, T.Z., Inglett, P.W., Reddy, R.K., 2007. The use of senescent plant biomass to investigate relationships between potential particulate and dissolved organic matter in a wetland ecosystem. Aquatic Botany 86, 53 61 . Quintana Ascencio, P., Fauth, J.E., Castro Morales, L.M. and McCauley, L., 2001. Ecological Studies of Willow (Salix caroliniana): Project Final Report. Department of Biology, University of Central Florida, 4000 Central Florida Boulevard, Orlando , Florida 32816. Quintana Ascencio, P.F., Fauth, J.E., Castro Morales, L.M., Ponzio, K.J., Hall, D. and Snyder, K. 2013. Taming the beast: managing hydrology to control Carolina Willow ( Salix caroliniana ) seedlings and cuttings. Restoration Ecology Raic h, J. W. and Potter, C. S., 1995. Global patterns of carbon dioxide emissions from so ils, Global Biogeochem. Cycles, 9 (1), 23 36 .


87 Roberts, J.D. and Rowland, A.P., 1998. Cellulo se fractionation in decomposition studies using detergent fibre pre treatment methods . Communications in Soil Science and Plant Analysis , 29 ; 11 14 . Rocha, A.V. and Goulden, M.L., 2009. Why is marsh productivity so high? New insights from eddy covariance and biomass measurements in a Typha marsh. Agricultural and Forest Meteorolo gy, 149 (1), 159 168 . Serna, A., Richards, J.H. and Scinto, L.J . 2013. Plant decomposition in wetlands: effects of h ydrologic variation in a re created E verglades. J.Environ. Qual. 42; 562 572 . Schmidt, J.M., 2002. Litter Decomposition in Created and Adjacent Forested Wetlands of the Coastal Plain of Virginia. Thesis submitted to the Faculty of the Virginia Polytechn ic I nstitute and State University. St Johns River Water Management District, 2007. Upper St Johns River Basin Surface Water Improvement and Management Plan. St Johns River Water Management District, 2009. Fort Drum Marsh Conservation Area Land Management Plan Summary St Johns River Water Management District, 2007 . Three Forks Conservation Area Land Management Plan Summary St Johns River Water Management District, 2008 . Blue Cypress Marsh Conservation Ar ea Land Management Plan Summary Van Dam, B. 2012. Organic soil subsidence and nutrient leaching in an impacted Florida wetland. Thesis submitted to the University of Florida. Wade, L.G. 2006. Organic Chemistry. Pearson Prentice Hall. Webster, J.R. and Benfield, E.F., 1986. Vascular plant breakdown in freshwater ecos ystems. Ann. Rev. Ecol. Syst., 17, 567 94. Wetlands. Cambridge University Press. Wright, A.L. and Reddy, K.R., 2001. Heterotrophic Microbial Activity in Northern Everglades Wetland Soils. Soil Sci. Soc. Am. J. 65, 1856 1864 . Wright, E.L., Black, C.R., Cheesman, A.W., Turner, B.L. and Sjögersten, S., 2013. Impact of simulated changes in water table depth on ex si tu decomposition of leaf litter from a Neot ropical peatland. Wetlands, 33, 217 226.


88 BIOGRAPHICAL SKETCH Shannon Duffy was born in Hialeah, F lorida in 1987 and grew up in Sunrise, Fl. After completing high s chool in 2005, she earned her Bachelor of S cie nce in Environmental Science from the School of Natural Resources and Environment at the University of Florida in 2009 . Following graduation, she gai ned experience with Everglades r progr am. She then left sunny Florida to learn about different (and much colder) wetland s at the Prairie Wetlands Learning Center in Fergus Fal ls, Minnesota, where she worked with the United S tates Fish and Wildlife Service in environmental education. She retu rned to Florida in 2010 with an expanded set of skills in utilizing the outdoors as a classroom, which she applied to her full time position with the Marshall Foundation as an Education Associate. In 2011, she returned to the University of Florida to purs ue a Ma ster of Science degree in soil and water s cience.