Assessing Hydrology, Carbon Flux, and Soil Spatial Variability Within Vernal Pool Wetlands

Vernal pools are small isolated wetlands that are covered by shallow water for variable periods in the winter and spring but may be completely dry for most of the summer and fall. Despite their small size, vernal pools are a dominant wetland type throughout New England. These wetlands are hotspots of floral and faunal biodiversity, as their unique hydrology and landscape characteristics allow them to serve as a home and breeding ground for many distinct plant and animal species. Because of their abundance in New England, vernal pools may also be an important long-term regional storehouse for organic carbon. Despite the functional and ecological values of vernal pools, few studies have investigated how variations in hydrology, pool size, geomorphic setting, and surrounding landscape attributes affect soil carbon in these systems and the associated hydric soils that sequester the carbon. Therefore, the primary objectives of this thesis were to i) assess the effect of landscape characteristics on hydrologic and edaphic conditions; ii) investigate the need for additional hydric soil indicators for vernal pool soils; iii) quantify the relationship between vernal pool hydrology and greenhouse gas fluxes; and iv) evaluate processes of carbon cycling throughout vernal pools. Twenty-one vernal pools in southern Rhode Island were selected and their landscape attributes were characterized using spatial tools in GIS. Sixteen of the 21 pools formed in glaciofluvial deposits on outwash plains, kame terraces, and moraines. The rest formed in till or alluvial landscapes. Vernal pool basin areas ranged from 6 to 381 m and mean slopes of the adjacent landscape ranged from 3 to 20%. Slope class was not significantly correlated to basin area (R = 0.03). Four vernal pools were selected for detailed studies of hydrology, soils, and vegetation in the basin, transition, and upland zones. Water table levels were monitored in each hydrologic zone from June 2015 to October 2016. Median water table levels increased in depth from the soil surface with increased distance from the vernal pool basin. Basin zones were consecutively inundated for the longest period of time, followed by transitional zones; upland zones were never inundated. Water table gradients indicated discharge into pool basins for the majority of the year. Vernal pools with steep slopes showed recharge gradients during periods of significant inundation suggesting a relationship between slope class and hydrology. Vernal pool soils classified as Spodosols, Inceptisols, and Histosols. Although all of the basin and transitional zones met the saturation requirements for hydric soils, 25% of the soils did not meet a hydric soil indicator. Both of these soils were Spodosols, suggesting the need for continued evaluation of hydric soils with spodic morphologies. Twenty-one plant species were identified across all study sites. Specifically, basin zones were dominated by obligate wetland plants, while transitional and upland zones consisted primarily of facultative and facultative wetland plants. Analysis allowed for the identification of plant species that accurately reflected the hydrologic nature of each zone, which affirmed the relationship between vegetation and hydrology. Carbon pools and the contributions to the wetland soil carbon cycle, including leaf litter additions, decomposition of coarse woody debris and leaves, and CO2 from respiration, were monitored during the majority of two growing seasons. On average, basin and transitional zone soils possessed the largest soil organic carbon (SOC) pools (11 kg m), while SOC pools in upland zones were substantially less (8 kg m). Leaf litter additions ranged from 40 to 149 g C m depending on the site and hydrologic zone. Leaf litter bags and wooden dowel rods (representing coarse woody debris) were placed at the surface of each zone in order to investigate above-ground decomposition. Dowel rods were also inserted vertically into the soil to 25 cm in 2015 to investigate below-ground decomposition. On average, basin zones experienced the highest leaf litter loss (12 g C m; 18%), while upland zones experienced the lowest loss (6 g C m; 10%). Basin zones also exhibited the highest mean surface dowel loss in 2015 (4%). Losses from 2016 followed similar trends, but lower precipitation experienced by the pools resulted in significantly lower losses. On average, the highest below-ground decomposition occurred in the upland zone (16%). Analysis comparing vernal pool hydrology to organic matter decomposition yielded results that trended toward significance in 2015 for leaf litter (p = 0.06) and ground dowels losses (p = 0.07). Above-ground decomposition decreased as the water table receded further from the soil surface; conversely, below-ground decomposition increased as the water table receded further from the surface. Carbon dioxide (CO2), methane (CH4), and nitrous oxide (N2O) fluxes were measured between the months of August and November in 2015, and between May and August in 2016. CO2 flux was significantly correlated with soil temperature (p < 0.001) and hydrologic zone across all study sites (p = 0.03). Flux increased as soil temperature increased and as the depth between the soil surface and the water table increased. Although the partitioning of heterotrophic and autotrophic respiration yielded weak correlations (R < 0.5), analysis suggested that root respiration accounted for < 40% of the total CO2 flux. Mean monthly CO2 fluxes were highest in the transitional zones, ranging from 5 to 22 μmole m min across all zones. CH4 flux was significantly correlated with vernal pool hydrology in all study sites (p < 0.003). Positive emissions ranged from 0.02 to 0.03 μmole m min and only occurred in the basin zones during the months of May and June, when the basins were inundated with surface water. Transitional and upland zones exhibited net CH4 absorption, as did the basin zones during months other than May and June. Nitrous oxide fluxes ranged from -0.009 to 0.008 μmole m min and did not vary significantly with hydrologic zone. Despite their small size and ephemeral nature, their significant contribution to greenhouse gas efflux and removal from the atmosphere warrant future investigations and conservation of vernal pools.

selected for detailed studies of hydrology, soils, and vegetation in the basin, transition, and upland zones. Water table levels were monitored in each hydrologic zone from June 2015 to October 2016. Median water table levels increased in depth from the soil surface with increased distance from the vernal pool basin. Basin zones were consecutively inundated for the longest period of time, followed by transitional zones; upland zones were never inundated. Water table gradients indicated discharge into pool basins for the majority of the year. Vernal pools with steep slopes showed recharge gradients during periods of significant inundation suggesting a relationship between slope class and hydrology. Vernal pool soils classified as Spodosols, Inceptisols, and Histosols. Although all of the basin and transitional zones met the saturation requirements for hydric soils, 25% of the soils did not meet a hydric soil indicator. Both of these soils were Spodosols, suggesting the need for continued evaluation of hydric soils with spodic morphologies. Twenty-one plant species were identified across all study sites. Specifically, basin zones were dominated by obligate wetland plants, while transitional and upland zones consisted primarily of facultative and facultative wetland plants. Analysis allowed for the identification of plant species that accurately reflected the hydrologic nature of each zone, which affirmed the relationship between vegetation and hydrology.
Carbon pools and the contributions to the wetland soil carbon cycle, including leaf litter additions, decomposition of coarse woody debris and leaves, and CO2 from respiration, were monitored during the majority of two growing seasons. On average, basin and transitional zone soils possessed the largest soil organic carbon (SOC) pools (11 kg m -2 ), while SOC pools in upland zones were substantially less (8 kg m -2 ). Leaf litter additions ranged from 40 to 149 g C m -2 depending on the site and hydrologic zone. Leaf litter bags and wooden dowel rods (representing coarse woody debris) were placed at the surface of each zone in order to investigate above-ground decomposition. Dowel rods were also inserted vertically into the soil to 25 cm in 2015 to investigate below-ground decomposition. On average, basin zones experienced the highest leaf litter loss (12 g C m -2 ; 18%), while upland zones experienced the lowest loss (6 g C m -2 ; 10%). Basin zones also exhibited the highest mean surface dowel loss in 2015 (4%). Losses from 2016 followed similar trends, but lower precipitation experienced by the pools resulted in significantly lower losses. On average, the highest below-ground decomposition occurred in the upland zone (16%). Analysis comparing vernal pool hydrology to organic matter decomposition yielded results that trended toward significance in 2015 for leaf litter (p = 0.06) and ground dowels losses (p = 0.07). Above-ground decomposition decreased as the water table receded further from the soil surface; conversely, below-ground decomposition increased as the water table receded further from the surface. Carbon dioxide (CO2), methane (CH4), and nitrous oxide (N2O) fluxes were measured between the months of August and November in 2015, and between May and August in 2016. CO2 flux was significantly correlated with soil temperature (p < 0.001) and hydrologic zone across all study sites (p = 0.03). Flux increased as soil temperature increased and as the depth between the soil surface and the water table increased. Although the partitioning of heterotrophic and autotrophic respiration yielded weak correlations (R 2 < 0.5), analysis suggested that root respiration accounted for < 40% of the total CO2 flux. Mean monthly CO2 fluxes were highest in the transitional zones, ranging from 5 to 22 µmole m -2 min -1 across all zones. CH4 flux was significantly correlated with vernal pool hydrology in all study sites (p < 0.003). Positive emissions ranged from 0.02 to 0.03 µmole m -2 min -1 and only occurred in the basin zones during the months of May and June, when the basins were inundated with surface water. Transitional and upland zones exhibited net CH4 absorption, as did the basin zones during months other than May and June.
Nitrous oxide fluxes ranged from -0.009 to 0.008 µmole m -2 min -1 and did not vary significantly with hydrologic zone. Despite their small size and ephemeral nature, their significant contribution to greenhouse gas efflux and removal from the atmosphere warrant future investigations and conservation of vernal pools. x

INTRODUCTION
Vernal pools are small isolated wetlands covered by shallow water for variable periods in the winter and spring, but may be completely dry for most of the summer and fall (Calhoun and deMaynadier, 2004;Skidds and Golet, 2005). Vernal pools form in many different geomorphic settings, including kettle holes, oxbow scars, and Carolina and Delmarva Bays (Brooks, 2005;Leibowitz, 2003). In New England, these wetlands primarily form in till and outwash on drumlins, outwash plains, and moraines as depressions are created during glaciation and subsequent deglaciation (Tiner, 2003;Skidds and Golet, 2005). They can also form on alluvial landscapes, such as floodplains, as stream and river waters overflow and cut into the surrounding landscape, creating pockets in the land where water may accumulate (Tiner, 2003).
Although some have been found to be upwards of 0.3 ha in size, most New England vernal pools are less than 0.1 ha in size (Brooks and Hayashi, 2002;Skidds and Golet, 2005;Rheinhart et al., 2007;Capps et al., 2014). At the scale that current Soil Survey and National Wetland Inventory maps were created, vernal pools are often missed because of their small size (Witham et al., 1998).
Vernal pools are hotspots of floral and faunal biodiversity, serving as a home and breeding ground for not only unique animal species, but also unique plant species.
For example, one study identified more than 400 different plant species that inhabit vernal pools in the northeastern United States, 20 of which are at risk of extinction in New England (Cutko and Rawinski, 2008). The fact that a large number of the floral and faunal species inhabiting vernal pools are endemic and/or endangered suggests that even small amounts of cumulative vernal pool destruction could lead to local and regional extirpations (Leibowitz, 2003 (Smith et al, 1995;Leibowitz, 2003). Studies have observed a positive correlation between basin area and the span of time vernal pool basins are inundated with water, often referred to as hydroperiod (Brooks and Hayashi, 2002;Skidds and Golet, 2005). Other studies have observed that hydroperiod length can be linked to morphological characteristics of basin soils, such as soil texture and permeability (Calhoun et al., 2013). This suggests the possibility of a relationship between basin size and soil morphology, while also raising the question of whether one characteristic has a greater impact on vernal pool Few detailed studies have simultaneously compared variations in hydrology and related soil morphologies and properties across the region. Studies on a regional level will allow for the creation of a catalogue of the different morphologies, vegetation, and geomorphic surfaces and landforms that are found throughout the northeastern United States. Documenting the differences in vernal pool ecosystems will enable researchers to test hypotheses across varying climates, soil parent materials, and landscape settings (Smith et al., 1995;Lebowitz, 2003;Skidds and Golet, 2005).
Specifically, I investigated the relationship between variations in the landscape attributes of vernal pools and their morphological and hydrological features in the Pawcatuck River Watershed. By initially conducting a broad study of vernal pools within the watershed, I was able to select several pools to study that are representative of common Southern Rhode Island vernal pool characteristics. Focusing on a small subsample of vernal pools allowed for a detailed analysis of specific traits, including landscape characteristics, vegetation distribution, soil morphology, and hydrology.

Site Selection
In initial studies, 65 vernal pools that Skidds and Golet (2005) investigated in the Pawcatuck River Watershed, Rhode Island were evaluated for possible detailed studies. These sites aligned with the breadth of landscape settings and soil parent materials that occur in southern Rhode Island including those adjacent to streams and rivers (alluvial), on outwash plains and kame terraces (glaciofluvial outwash or icecontact materials), and on drumlins and upland settings (lodgement and ablation till) (Skidds and Golet, 2005). Only those vernal pools with hydroperiods that lie within two weeks of the average hydroperiod of these 65 pools were considered for investigation. In addition, field investigations were paired with satellite imagery analysis in order to identify five other potential vernal pools that Skidds and Golet (2005)  Three transects were established at each of the four vernal pools. Each transect extended radially outwards from the center of the basin to the upland zone. Transect locations were determined randomly based on compass orientation. On each transect, three research plots were established, one in each hydrologic zone, resulting in a total of nine research plots per site. Zones and transect locations follow the guidelines of the regional multistate project (NE-1438) this study supports.

Soil Morphological Descriptions
At each site, a soil description to the depth of at least 1 m was made in each zone (Schoeneberger et al., 2012). Samples collected with a bucket auger were utilized in order to describe soil horizons, texture, color, and presence of redoximorphic features in the field. Soil descriptions were analyzed and compared to the field indicators of hydric soils that are currently in use in order to consider whether new indicators should be developed for vernal pools (USDA-NRCS, 2016).

Vegetation Analysis
Vegetation present by zone was identified and analyzed in accordance with the 1987 USACOE Wetland Delineation Manual's regional supplement for the northcentral and northeastern region (U.S. Army Corps of Engineers, 2012).
Documented vegetation was divided into four strata: tree, sapling/shrub, herb, and woody vines. One 30-ft circular radius plot was established at each site in the basin and upland zones, and absolute percent cover of vegetation versus bare soil in each of the four strata was documented. Due to the irregular shape of the transitional zone, absolute percent cover was estimated across the entire zone. Measurements were obtained in July 2016.

Climate Data
Weather data from the University of Rhode Island's long-term monitoring (at least 30 years) weather station (Kingston, RI) were obtained in order to learn of local air temperatures maximums and minimums and daily precipitation (Diamond et al., 2013). One HOBO temperature logger was buried 30 cm under the soil surface at Great Swamp and measured soil temperature every 4 hours. These data were used in order to identify the growing season, which is defined as the period of time during which soil microorganisms are active. The time period during which soil temperature was consecutively higher than the benchmark of 41 o F is considered to be the growing season for the vernal pool (Ford, 2014). Due to the close proximity of the study sites, I applied growing season established at Great Swamp to all other study sites.

Hydrological Measurements
At each site, the depth of ponded water or the depth to the water table was monitored and recorded, depending on the hydrologic zone and corresponding inundation/saturation patterns. The depth of ponded water within each vernal pool was measured monthly with a staff gauge. Utilizing staff gauges allowed for extrapolation about hydroperiod length of each pool. Staff gauges supplemented data provided by the Odyssey loggers in the basin zones when surface inundation surpassed the extent of the logger (Witham et al., 1998). A PVC monitoring port comprised of a well screen was installed at each plot in the transitional and upland zones to a depth of approximately 100 cm, allowing water tables to be measured periodically. Wells remained in the vernal pools for the duration of this field study, between June 2015 and August 2015. Only one well was installed in the basin zone and was positioned along the central transect. Odyssey loggers were calibrated and installed in wells along one transect at each site in order to record capacitance water level every 6 hours each day. Due to an inability to assess the exact depths of water tables when they surpassed the depth of the loggers, an analysis of the median depth values were used to assess hydrologic trends (Larson and Runyan, 2009;Rabenhorst, 2014). These measurements were extended to other transects depending on the trends exhibited by their monitoring ports. The maximum logger measurement was recorded when the water table depth exceeded the length of the logger. In addition, manual water table measurements were taken monthly at each well on all three transects. In order to obtain these manual measurements, a digital voltmeter was connected to 18-gauge speaker wire and the wire was then dropped into the wells. Once the wire touched the water, the electrical circuit was completed and a measurement appeared on the voltmeter. Measuring the depth at which the speaker wire reached the water allowed for a determination of water table depth throughout the various hydrological zones (Marti, 2016).
Certain saturation and precipitation conditions need to be met in order for a soil to meet hydric conditions. Precipitation needs to fall within a certain range of normal field conditions (namely, values that fall between the 30 th and 70 th percentiles).
Thirty-year average precipitation totals and 30 th /70 th percentile values were obtained from the nearest WETS station (Kingston, RI4266) and compared to monthly precipitation totals obtained for this study. Saturation conditions require that the water table reside within 25 cm of the soil surface for at least 14 consecutive days during the growing season (NTCHS, 2007). Odyssey logger measurements allowed for these conditions to be evaluated at each zone along the main transect.

Elevation Measurements
Elevation measurements were made along each transect using a laser level.
Initial benchmark measurements were made at the lowest point of each of the three transects, and additional measurements were made at each well. Measurements were also made every 10 feet, extending from the benchmark measurement (0 feet) to 50 feet (Rains et al., 2006).

Site Characteristics and Analysis
Vernal pools occur on a range of landscape depending on the physiographic region. In the northeastern US, the majority of vernal pools are located on glaciated materials (Brooks, 2003;Grant, 2005;Skidds and Golet, 2010 These wetlands can be partitioned into two components; the flat basin in the center of the pool, and the adjacent sloping upland. The transitional zone (zone 2) occurs at the intersection of these two areas. Average percent slopes of the area adjacent to the basins (transitional zone) varied from 3 to 20%, with a mean of 8% (Table 1.1). These values represent the slope of the land extending between the basin and transitional wells (Figure 1.1). While glaciofluvial vernal pools were found to have both the maximum and minimum slopes reported, only 31% of the glaciofluvial sites had an average slope greater than the mean (Table 1.1). This wide range of morphological characteristics in the basins and their adjacent slopes is largely due to the variety of glaciated landforms associated with glaciofluvial parent material (Tiner, 2003;Skidds and Golet, 2005).
Vernal pools are small depressional wetlands which typically have basin areas < 0.1 ha (1000 m 2 ). Skidds and Golet (2005)  Distinguishing a transitional zone from the flat basin zone allowed for a more in-depth analysis of portions of the wetland exhibiting differences in morphology and hydrology. All of the sites I investigated had basins far less than 1000 m 2 (Table 1.1). Brooks et al. (1998) reported that of the 430 vernal pools they investigated in Massachusetts, no pool basin was larger than 500 m 2 . Skidds (2003) reported much larger maximum basin sizes, some ranging upwards of 6000 m 2 . These basin area measurements were made by calculating the areas in ArcGIS and by making field measurements. Skidds (2003) found that of pools < 500 m 2 , ArcGIS measurements were fairly inaccurate. Calculating basin areas based on field measurements may allow for a more accurate assessment of vernal pool sizes on different soil parent materials.
On average, vernal pools on dense till had a greater basin size than any of the other parent materials. In contrast, Skidds (2003) reported that glaciofluvial vernal pools had the largest average basin area. One possible reason for this difference is that I only investigated two vernal pools with dense till while Skidds (2003) (Figure 1.2). In vernal pools, water is lost to the atmosphere primarily through the process of evapotranspiration (Hanes, 2003;Brooks et al., 2004), which follows increases in temperature in the spring when the plants leaf out and drops in the fall when the plants lose their leaves. Thus, water table levels are lowest in the summer months as evapotranspiration increases and soils lose their moisture. In the winter, evapotranspiration is limited and the water table levels respond to precipitation by rising and staying at the highest levels ( Figure 1.1).
Water Comparison of vernal pool elevations with hydrological measurements allowed for an assessment of the water table gradients of each pool (Figure 1.3). When the water table resided deep below the soil surface, the gradient appeared to follow the slope between the transitional and upland zones. While two of the gradients stayed consistent regardless of the month (CAR 2 and GS), the other two appeared to reverse toward the upland when the basin experienced surface inundation (CAR 3 and EP). Hanes (2003) analyzed the hydraulic gradient of vernal pools in an attempt to quantify the impact that various landscape characteristics and water sources have on vernal pool hydrology. In that study, Hanes (2003) found that the relative elevation of a vernal pool must be considered when evaluating groundwater hydrology patterns. In vernal pools with steeper slope gradients surrounding the basin, surface runoff may result in reversed groundwater patterns (Hanes, 2003). The two vernal pools that I studied in which gradient reversal occurred also had the highest percent slopes (Table   1.1).
Understanding hydrologic patterns during the growing season is critical when evaluating hydric soil conditions and identifying hydric soils (Megonigal et al., 1996;Ford, 2014;. Two growing seasons were identified based on soil would have met the saturation criteria necessary to be considered hydric soils. None of the upland soils met the required hydric soil saturation conditions.

Soil Morphology
Soils in the vernal pools were classified as Spodosols, Inceptisols, and Histosols (Table 1.5; Table 1.6; see Appendix 6 for descriptions of the soils in each zone). All of the soils mapped in the upland zones were Inceptisols. These are young, underdeveloped soils that are the most common soil type found in Rhode Island (Rector, 1981 for long periods of time (Fanning and Fanning, 1989). Inundation maintains anaerobic conditions and typically slows organic matter decomposition (Mausbach and Richardson, 1990 (Hanes and Stromberg, 1998;Wang et al, 2005).
Although I only sampled one Histosol in my research, further investigations of the landscape characteristics of vernal pools with Histosol soils may provide insight into a relationship between these characteristics and the formation of Histosols.

Hydric Soils
Soils were classified as hydric or non-hydric (Table 1. When designing this study, I utilized the hydrologic characteristics in order to distinguish the three zones of each vernal pool. I expected that the basin and transitional zones would possess hydric soils and this held true for all of the vernal pools. Two of the wettest soils, however, did not meet any hydric soil indicators: one transitional soil in CAR 2, and one basin soil in CAR 3. Both soils met the saturation requirements for a hydric soil (  (NHSTC, 2008). One of the 18 profiles that met wetland hydrology criteria, however, had a spodic horizon that began at a depth similar to that of the aforementioned CAR 2 and CAR 3 soils (~40 cm). Because this is a test indicator, the possibility exists that the criteria will need to be adjusted in order to account for the broad range of morphological characteristics that Spodosols possess.

Vegetation Analysis
Throughout the three hydrologic zones I encountered 21 different plant species (Table 1.7). While I found many of the same plant species throughout the study sites, these commonly observed species differed in frequency across all pools and zones.
Herbs were the dominant strata in all four basins ( Figure 1.5). Dominant herb species in the basin included Bidens sp., Leersia oryzoies, and Dulichium arundinaceum. These three species were only found in basins that experienced extended periods of inundation: CAR 2, CAR 3, and EP (  (Stewart and Kantrud, 1971) and dominant species may vary at different points throughout the growing season; however, I only sampled once during this period.
Since vegetation in basin zones is primarily herbaceous, relating species richness and dominance to vernal pool hydrology may be both difficult and unreliable.
The transitional and upland zones were dominated by shrubs. Vaccinium corymbosum and Clethra alnifolia were the most common shrub species found in the transitional and upland zones. The distinct differences in strata between the different hydrologic zones supports the connection between vegetation and vernal pool hydrology, and is in line with the system currently used to categorize wetland plants by their indicator status: obligate (> 99% occurrence in wetlands), facultative wetland (66-99% occurrence in wetlands), facultative (33-66% occurrence in wetlands), and facultative upland (1-33% occurrence in wetlands; Table 1

SUMMARY AND CONCLUSIONS
The objectives of this part of my study were to characterize the landscape characteristics of vernal pools and evaluate their impact on edaphic and hydrologic conditions. Vernal pools were found on kame terraces, outwash plains, moraines, till uplands, and floodplains. All basin areas were < 1000 m 2 , which are the areas reported for most New England Vernal pools. Slopes ranged from 3 to 20%, and were Furthermore, the plant species identified accurately reflected the hydrologic nature of each zone.         within the scientific community (Smith et al., 1993;Huntington, 1995;Bridgham et al., 2006;Mitsch et al., 2012;Ricker et al., 2014). Wetlands, which are estimated to store 20-30% of the earth's terrestrial carbon pool (Mitsch et al., 2012), are but one of many landscape types under scrutiny in an attempt to mitigate the impacts of global climate change (IPCC, 2007).
The hydrologic characteristics and landscape setting of a vernal pool largely influence the processes that occur in wetland soils. The decomposition of organic matter is one important process that is largely influenced by these factors. Organic matter decomposition is dependent upon a number of factors, such as the source of the organic matter, soil temperature, soil moisture, and whether aerobic or anaerobic conditions are present. The length of time during which the soils exhibit anaerobic conditions also influences decomposition. Typically, wetland soils accumulate more soil organic matter (SOM) over time than upland soils because of the long periods of saturation that result in anaerobic conditions in these soils, ultimately slowing decomposition rates (Gorham, 1991;Groffman et al, 1996;. Studies on the variation in organic matter decomposition rates across hydrologic gradients of vernal pools have produced conflicting results, largely due to the unpredictable wet-dry cycles associated with the ephemeral nature of vernal pools (McClain et al., 2003;Capps et al., 2014).
Wetlands are hotspots of biogeochemical activity. Whether wetlands serve as net sources or sinks of carbon is dependent upon the biogeochemistry of the systems (Kuzyakov, 2005;Bridgham et al., 2006;Altor and Mitsch, 2008;Davis et al., 2010;Holgerson, 2015). Due to high plant productivity, wetlands are able to easily sequester carbon via plant uptake of atmospheric CO2 through photosynthesis. Plantderived carbon is added to the soil when the roots die or the plants drop leaves (Xiong and Nilsson, 1997) and deadfall (Brinson et al., 1981); contributing to the soil carbon pool. Over time SOM builds up in wetlands because the anaerobic conditions brought on by the saturated soils slow the rates of organic matter decomposition (Whiting and Chanton, 2001;Altor and Mitsch, 2008). These carbon stocks can be quantified and provide insight into the effectiveness of wetlands to serve as a carbon sink.
Organic matter is added to wetland environments by several key sources. For instance, trees contribute organic matter through deadfall and leaf litter deposition.
Furthermore, plant roots provide a source of organic matter throughout soil profiles (Capps et al., 2014). Organic matter serves as an energy source that fuels microbial processes in the soil, creating products emitted not only into the soil, itself, but also into the atmosphere. These atmospheric emissions are of particular concern because they can take the form of several potent GHG: CO2, methane (CH4), and nitrous oxide (N2O) (Moseman-Valiterra et al., 2011;Mitsch et al., 2012). Because of the important role that wetland landscapes such as vernal pools are considered to play in greenhouse gas models, understanding how differences in organic matter source and degree of soil saturation control the quantity of decomposition processes and the production of greenhouse gases is of the utmost importance.
In carbon sequestration studies, sources of carbon are typically associated with Studies have also simulated deadfall and root decomposition through the examination of above-ground and below-ground dowel decomposition in a variety of different environments. Precipitation, temperature, and landscape disturbance are the variables primarily evaluated with dowel decomposition. After a span of approximately three months, several studies found that dowel decomposition did not exceed 10% (O'Lear et al., 1995;Austin and Vitousek, 2000;Bontti et al., 2009).
Little to no research has investigated the influence that the degree and duration of soil saturation has on dowel decomposition.
While wetland soils sequester carbon, they also serve as sources of CO2 through two primary biogeochemical methods: heterotrophic and autotrophic respiration. Heterotrophic respiration primarily results from microbial mineralization of organic matter. Soil microbes decompose SOM in order to utilize it as an energy source. Through respiration, the carbon is released into the atmosphere as CO2.
Depending on environmental conditions, microorganisms will respire either aerobically (via the tricarboxylic acid cycle) or anaerobically (via fermentation). On the other hand, autotrophic respiration primarily results from the respiration of live plant roots (Raich and Schlesinger, 1992;Altor and Mitsch, 2008). Partitioning these two sources allows for a better understanding of the mechanisms behind CO2 flux (Hanson et al., 2000;Ricker et al., 2014). While CO2 flux has been studied extensively in many natural and simulated wetland environments (Raich and Potter, 1995;Bridgham et al., 2006;Altor and Mitsch, 2008;Mitsch et al., 2012;Ricker et al., 2014), research focused on measuring CO2 fluxes and partitionment to the fluxes in vernal pool wetlands is limited to non-existent.
Studies have found that estimates of root contribution to total CO2 flux greatly vary, ranging from 5% to 90% (Andrews et al., 1999). Raich and Schlesinger (1992) examined the mechanisms of soil respiration in a forested soil and found that 26% of total soil respiration resulted from root respiration (autotrophic), while the remaining 74% was attributed to microbial decomposition of organic matter (heterotrophic respiration). In a study quantifying root respiration in forested riparian wetland and upland soils, Ricker et al. (2014) found that mean root respiration ranged from 32-63% of total soil CO2 flux. While efforts have been made to partition the sources of CO2 flux in uplands and wetlands, separately, performing such a study in vernal pools allows for study across a unique hydrologic gradient.
Hydrology also plays a major role in other biological and chemical interactions that take place in the soil (Ricker et al., 2014). The unique characteristics of vernal pools allow for a diagnostic investigation of how the hydrology of a natural environment influences the magnitude of GHG emitted into the atmosphere. GHG such as CO2, CH4, and N2O all contribute the global warming of Earth (Altor and Mitsch, 2008). Due to the differences in the magnitude of their impact on global warming, GHG emissions should be evaluated individually. For instance, the potency of CH4 is 25 times that of CO2, while N2O has a potency of 298 times that of CO2.
Global climate models predict that global surface temperatures will increase 1.5 to 4 o C by the year 2100 (IPCC, 2007). Long-term monitoring of vernal pool climate conditions, such as temperature and rainfall inputs, allows for an improved understanding of vernal pools' vulnerability to global warming.
Wetland soils play a key role in the global carbon cycle not only by contributing CO2, but also by producing CH4 through the process of methanogenesis.
Methanogenesis is an anaerobic process in which microorganisms first degrade organic matter present in the soil. Methanogenic microorganisms utilize the acetate or hydrogen and CO2 produced by this decomposition in order respire, thus producing CH4 which can then be emitted into the atmosphere (Segers, 1997;Altor and Mitsch, 2008). There is a positive correlation between the amount of carbon fixed in wetlands to the amount of CH4 emitted into the atmosphere (Whiting and Chanton, 2001).
Despite their small size, CH4 flux tends to be high in vernal pool wetlands as aerobic soil becomes inundated, reducing the soil's ability to oxidize CH4. When CH4 oxidation exceeds CH4 production through methanogensis, the area is considered to be a sink of CH4 rather than a source (Kuhn, 2015;Holgerson, 2015). Thus, this absorption of CH4 is typical in aerobic soil environments (Kagotani et al., 2001).
In addition to playing a key role in the global carbon cycle, wetland soils serve as an ideal environment for denitrification. Denitrification is the final step in the nitrogen cycle. In this anaerobic process, microorganisms reduce nitrogen oxides into nitrogen gases. In order for this process of reduction to occur, microorganisms need a source of organic carbon. The ultimate product of denitrification is di-nitrogen gas Many studies have investigated CH4 flux (Whiting and Chanton, 2001;Altor and Mitsch, 2008;Segers, 1997;Mitsch et al., 2012) and N2O  Another study divided several vernal pool sites into three similar zones and measured denitrification rates, or N2O production, in each zone after nutrient application. These study sites were located in Maine, and N2O fluxes were measured during the months of May and July 2013. Researchers also found a significant difference between upland denitrification rates and those of the basin and edge zones, as the upland zones consistently produced lower rates (Capps et al., 2014).
These results suggest that vernal pool position does influence CH4 and N2O emissions, and that there is a need for additional assessments of these relationships.
However, these studies were limited by the time period during which sampling occurred. Neither study was able to encompass all four seasons in their research, which is critical in understanding vernal pool dynamics due to the various changes that vernal pools undergo in each season (Zedler, 2003). Furthermore, both studies measured gas flux in a laboratory study rather than in the field. Field measurements allow researchers to examine the relationship between climate and GHG flux (Davis et al., 2010;Chuersuwan et al., 2014;Ricker et al., 2014). The purpose of this study was to investigate the relationships between hydrologic regime and key wetland processes. I examined the balance between inputs and outputs of carbon in vernal pool wetlands. I quantified leaf litter deposition rates throughout the vernal pools (Davis, 2001), as well as the bulk density and loss on ignition (LOI) in order to investigate soil as a sink for carbon (Nelson and Sommers, 1996;Hobson, 1998). However, in order to quantify the ability of the soil to serve as a carbon sink, it is important to also examine losses of carbon from the soil. I examined the rate of organic matter decomposition in the different hydrologic zones of the vernal pools. In order to do this, I replicated three common sources of organic matter; leaf litter, deadfall, and roots, and measured the decomposition that took place over time across a hydrologic gradient. These studies will provide insight into differences in carbon inputs and outputs between upland and wetland environments.
Furthermore, I was interested in determining whether the unique hydrologic and edaphic characteristics of vernal pools cause their CO2, CH4, and N2O fluxes to differ from those emitted by other wetland environments. I investigated the flux of three greenhouse gases, CO2, CH4 and N2O in the different vernal pool zones and examined the role that climate, hydrology, and landscape position plays in gas flux processes. In addition, I compared CO2 flux to root density to estimate the ratio of heterotrophic to autotrophic respiration occurring in vernal pool soils. This comparison allowed for an appraisal of the magnitude by which each process influences CO2 flux (Ricker et al., 2014). These studies will provide a better understanding of carbon dynamics in vernal pool wetlands.

Quantifying Carbon Stocks
Carbon stocks were quantified at each research plot along the three transects at each site (see Chapter 1 for descriptions of the study sites). Soil samples were collected by driving a sharpened aluminum tube (60 cm long and 7.5 cm in diameter irrigation pipe) 50 cm into the soil and excavating the core from the soil. The tubes were capped in the field, moved to the lab, and frozen until they were described.
Electric metal shears were used to open the cores once thawed, and morphological descriptions were made for each core (Schoeneberger et al., 2002), Samples were collected based on the master horizons present (i.e. O horizons, A horizons). Horizon thicknesses was recorded and all of the soil materials for each master horizon was sampled and weighed. Bulk density for each horizon was calculated by dividing the dry weight of the horizon by the horizon volume (cross-sectional area of the core times the horizon thickness) (Soil Survey Laboratory Staff, 2004). Total soil carbon stocks (g/m 2 ) to 50 cm in the soil profile were calculated using estimated SOC content and bulk density and horizon thickness data (Homann and Grigal, 1996).
Estimates for the proportion of SOC in total SOM contents were obtained from a study by Davis (2001) Any day with a temperature higher than this calculated value is considered a growing degree day (Douglas and Rickman, 1992).
The leaf litter bags were retrieved, dried at 60 o C and weighed in order to determine the loss of organic matter over time to determine decomposition rates (Capps et al., 2014). Rods were removed, washed, dried, and weighed in order to determine overall weight loss. Along all three transects, five pre-weighed replicate dowels were inserted vertically into the soil to a depth of 25 cm at each research plot in June of 2015, collected from the field three months later, and prepared in the same manner as the dowel rods placed on the soil surface.
Greenhouse Gas Flux CO2, CH4, and N2O flux was measured using a closed chamber approach. Flux rates were measured at each research plot on the three transects at each site, providing data for each of the three hydrologic zones. Two cylindrical plastic chambers (16 cm in height, 20 cm in diameter) were placed at each site and pushed approximately 2.5 cm into the soil. Using a 20 ml gas-tight syringe, an initial gas sample was taken after securing the chamber's lid, which contained a rubber septum to allow for sampling, followed by samples taken 15 and 30 minutes after the initial sample. In order to mix the gases in the headspace of the chamber prior to sampling, the syringe was pumped three times. Measurements were taken monthly between August and November of 2015 and May and August of 2016. Between each month, the chambers were moved laterally throughout each zone. After sample collection, the syringe contents were immediately transferred into a 15 ml evacuated tube (Amador and Azivinis, 2013). In the field, internal chamber temperatures were also measured when each gas sample was collected and averaged in order to obtain the average chamber temperature during the sampling period. Soil temperature at a depth of 5 cm, and specific chamber volume (m 3 ) were recorded at each time interval of sample collection (Ricker et al., 2014;Waggoner, 2016).
CO2, CH4, and N2O concentrations were measured with a Shimadzu gas chromatograph and recorded in units of ppm (Altor and Mitsch, 2008).
Concentrations were plotted against time and fitted with a linear regression in order to calculate the CO2 flux rates. The mass of each gas present in the sampling chamber, or n (mol), was calculated using the Ideal Gas Law, n=RT/PV, where n=mol CO2 per mol air, R=universal gas constant (0.0821 L atm/mol K), T= chamber internal temperature (K), P=atmospheric pressure (atm), and V= chamber volume (L). The rate of GHG production per unit area was calculated using the slope of the best-fit line, cross-sectional area of the chamber, and volume of air in the chamber (Waggoner, 2016).

Carbon Dioxide Respiration Partitioning
The ratio of heterotrophic (microbial) to autotrophic (root) respiration was evaluated using root biomass/CO2 flux relationship (Ricker et al, 2014;ArchMiller and Samuelson, 2015). This method is based on the linear relationship between root biomass (independent variable) and total CO2 concentrations emitted from the soil (dependent variable). Based on this relationship, microbial respiration is assumed to be represented by the y-intercept of the linear regression model, and at this point, there is no root biomass present in the soil. Without root biomass, it is assumed that all of the CO2 respiration can be attributed solely to microbial activity (Kuzyakov, 2006).
At the time of each gas sampling session, a root corer (6.5 cm in diameter) was used to collect soil samples of a known volume to a depth of 10 cm. Samples were taken from the center of each cylindrical plastic chamber. The samples were transferred to the laboratory and air dried. Once dry, tweezers were used to remove all root biomass materials from the soil samples. Roots were washed with a calgon solution in order to remove any residual soil, air dried, and weighed. Total root biomass for each sample was plotted against the CO2 respiration rates in order to evaluate how much respiration can be attributed to roots versus microbial activity (Ricker et al., 2014).

Hydrological Measurements
At each site, the depth to the water  (Capps et al., 2014;Kuhn, 2015).

Climate Data
Weather data from the University of Rhode Island's long-term monitoring (at least 30 years) weather station (Kingston, RI) were obtained in order to learn of local air temperatures maximums and minimums and daily precipitation (Diamond et al., 2013). Surface soil temperature was measured by inserting a digital thermometer into the top 5 cm of the soil in each zone on the main transect at the time of gas sampling.
These data were compared to the monthly variations between GHG fluxes in the vernal pools in order to investigate the relationship between climate and GHG flux (Mitsch et al., 2012;Chuersuwan et al., 2014).

Carbon Stocks
Soil carbon pools are a function of plant additions, such as roots, leaf litter, and deadfall, and the loss of that carbon through the process of decomposition (Bontti et al., 2000;Capps et al., 2014;Ricker et al., 2014). Average SOC pools in the upper 50 cm of the soils ranged from 8 to 11 kg m -2 , with the upland having the average lowest value (Table 2.1). The mean SOC pool estimated in the upland zone (8 kg m -2 ) was very similar to the 9 kg m -2 that Davis et al. (2014) reported for Typic Udipsamments in southern New England. Increased soil wetness slows rates of organic matter decomposition such that wetter soils have greater SOC pools than drier soils (Zdruli et al., 1995;Davis et al., 2004). Thus, the basin and transitional zones had the largest pools. The basin soils had the greatest variations, ranging from 5.5 to 19 kg m -2 . The EP basin soil was dominated by organic soil materials (Table 1.

Litterfall Additions and Decomposition
In temperate forests such as those in New England leaf litter is typically the largest source of carbon to the soils (Davis et al., 2004;Ricker et al, 2014). Leaf litter SOC inputs varied among zones and sites (Figure 2.2a). Mean SOC additions ranged from 58 to 149 g C m -2 in the basin zones, from 54 g C m -2 to 121 g C m -2 in the transitional zones, and from 40 g C m -2 to 129 g C m -2 in the upland zones. Ricker et al. (2014)  These data suggest that leaves that are newly lain on the soil surface have relatively slow initial decomposition rates relative to additions of carbon.
On average, in three out of the four study sites, basin zones exhibited the most leaf litter decomposition, followed by the transitional zone, with upland zones exhibiting the least decomposition. The average loss of mass from the leaves applied ranged from 10 to 18%. Bags in the basin zone yielded the highest percent loss, ranging from an average of 13.5 to 19.2% mass lost after three months. Average leaf litter losses ranged from 10.2 to 15.8% in the transitional zones and from 10.1 to 18% in the upland zones (Figure 2.3 Leaf litter decomposition rates were much lower in 2016 than those experienced during the previous year (Figure 2.3). Average losses ranged from 2.5 to 6.2% of the leaf litter's initial mass in the transition zone, 2.5 to 3.8% in basin zones, and from 1.5 to 5.3% in upland zones. One possibility for the difference in decomposition rates between years is the period during which the litter bags were deployed. Leaf litter bags were placed in the vernal pools a month later (July) in 2015 than in 2016. Records of growing degree days, however, were very similar (approximately 2500 growing degree days during both experiments) suggesting that a difference in heat energy in the soil system was not a strong contributor to the difference in decomposition between the two years (http://uspest.org/cgibin/ddmodel.us).
Another explanation for the differences in decomposition between years could be moisture content. Both Moore et al. (1999) and Austin and Vitousek (2000) reported increased decomposition of leaf litter and deadfall at the soil surface with Wooden dowels secured at the soil surface in 2015 displayed trends similar to the 2016 studies. In 2015, surface dowels' mean loss ranged from 0.7 to 2.6% in the transitional zones and 1.8 to 2.6% in the upland zones (Figure 2.5). In the basin zone, one mean value was as high as 13.3% in the basin with the remaining three below 2.5% loss. Excluding the one high value, these losses fall in the range of values (<10% loss) reported by Austin and Vitousek (2000) for decomposition of dowels secured at the soil surface.
Wooden dowels placed below the soil surface showed much higher losses due to decomposition than those at the surface (Figure 2.6). Ground dowels placed in the basin zone exhibited the lowest amount of decomposition in (on average, between 9.5 and 14% of mass lost) while the upland zone had the highest (on average, between 10.2 and 20.3% of mass lost). Transitional zones averaged between 9.1 and 18.9% loss (Figure 2.6). Mean ground dowel decomposition in all zones exceeded the losses observed by O' Lear et al. (1999), in which losses after 3 months did not exceed 10%.
A study by Bontti et al. (2009) investigated wooden dowel decomposition and encompassed both above-and below-ground decomposition by positioning the dowels vertically, with half buried and half existing above the soil surface. Even when including the potential for both methods of decomposition, losses did not exceed 10%.
One likely explanation for these differences is the differences in hydrological conditions between studies. Both took place in either forested or grassland environments; neither took place in dedicated wetlands.
Differences in trends revealed by above-and below-ground decomposition are likely attributed to differences in the amount of moisture maintained by organic substrates at the soil surface versus below the soil surface. Materials laid on the soil surface are directly exposed to wind and sunlight, while materials in the subsurface are shielded from these elements. These differences in exposure cause surface substrates to dry faster than subsurface substrates. Furthermore, subsurface materials are surrounded by soil, and thus the moisture maintained by these materials is augmented by the moisture held in the soil pore space. Conversely, surface materials are not entirely surrounded by the soil, meaning that they are not exposed to soil moisture content to the same extent as the subsurface materials. These differences suggest that ideal conditions for microorganisms at the soil surface differ from those for microorganisms below the surface.
Comparison of vernal pool subsurface hydrology with organic matter decomposition trended toward significance in 2015 for leaf litter (p = 0.06) and ground dowel losses (p = 0.07). Leaf litter decomposition in 2015 decreased as the water  (Table 2.2). Soil saturation results in an anaerobic soil environment, which inhibits microbial decomposition, overall slowing organic matter decomposition rates in wetlands (Whiting and Chanton, 2001;Altor and Mitsch, 2008). Thus, in the basin where the water table levels were the highest, decomposition was the lowest. Conversely, ground dowel decomposition was highest in the upland zone in three out of the four study sites (Figure 2.6). As discussed in Chapter 1, the upland zones possessed the only soils that did not meet the saturating conditions necessary to be considered a hydric soil at any point during the study. Aerobic soil conditions provide ample oxygen diffusion for microorganisms, serving as an environment that facilitates quick organic matter decomposition (Kristensen et al., 1995). Soils that are too dry may stress the microorganisms and impede decomposition, as some moisture is necessary to fuel these processes (Davidson and Janssens, 2006). Ground dowel decomposition increased as the soil environment became increasingly aerobic, creating a more favorable environment for the microorganisms. As previously discussed, dowels secured below the soil surface are more inclined to hold onto moisture than organic materials at the surface. Because of this, ground dowels in the basin were likely saturated, resulting in anaerobic conditions and the inhibition of microbial decomposition. On the other hand, organic materials at the surface dried more readily, facilitating moist conditions rather than saturated conditions, and providing the ideal amount of moisture for the microorganisms. Comparison of above-and below-ground decomposition with hydrology affirms the strength of the relationship between field moisture conditions and the efficiency of microorganisms.
Differences in decomposition between leaf litter and surface dowels can be attributed to differences in their carbon to nitrogen ratios (C:N). On average, broadleaf foliage has a C:N ratio of 35:1, while dowels, have a ratio of approximately 100:1 (Wade and Fay, 1989;McGroddy et al., 2004). The C:N ratio determines how easy it is for microorganisms to decompose organic matter (Brinson et al., 1981). A C:N ratio of 30:1 readily facilitates decomposition (USDA-NRCS, 2011). Because leaves have a C:N ratio much closer to this ideal value than wooden dowels, leaves tend to decompose more readily. Typically, wooden dowels also have higher lignin contents than leaf litter. Lignin is a material located in plant cell walls that provide the plant with woodiness and rigidity. Materials with higher lignin contents are often associated with larger C:N ratios. Bontti et al. (2009) investigated the relationship between lignin and organic matter decomposition and discovered that increased lignin contents drive decomposition processes. Substrate surface area also influences decomposition rates (Jackson et al., 1997). The surface area of one wooden dowel is approximately equivalent to that of one leaf. Because each litter bag contained 10 g of leaves, the surface area available for decomposition was much larger than that available on a single wooden dowel. This increased surface area may have contributed to the heightened decomposition exhibited by the leaf litter.

Carbon Dioxide
The most common GHG of concern for global warming are CO2, CH4, and N2O (Altor and Mitsch, 2008;Moseman-Valtierra et al., 2011). I measured each of these gases monthly during the months of August through November in 2015 and May through August in 2016. Simple linear regression analyses between CO2 flux and soil temperature were used to evaluate the influence that the soil environment has on the mechanisms responsible for CO2 flux. There was a significant correlation between soil temperature and CO2 flux in all of the study sites (Table 2.3; p < 0.001). In all vernal pools, as soil temperature increased, CO2 flux also increased (Table 2.3).
Similar increases in respiration were reported by Reth et al. (2005), Davis et al. (2010), and Ricker et al. (2014). Warm soil environments encourage both microbial activity and plant growth, which are the primary mechanisms driving soil respiration.
Thus, increased atmospheric temperatures resulting from climate change may also lead to increased CO2 emissions.
Both soil moisture and soil temperature affect soil respiration (Kuzyakov, 2005;Davis et al., 2010;Ricker et al., 2014). The relationship between soil moisture, soil temperature, and soil respiration is very complex. Heterotrophic and autotrophic respiration is inhibited when the soil is saturated resulting in anaerobic conditions.
Respiration is also inhibited when the soil is too dry, as excessively dry conditions induce stress on the microbial and plant communities. I measured soil saturation, but it should be noted that depth of the water table gives some indication of the moisture status of the soil. Linear regression analysis revealed a significant interaction between water table depth and CO2 flux in one vernal pool (GS; p = 0.026). Results trended towards significance in one other vernal pool site (EP; p < 0.1). At these sites, as the distance from the soil surface to the water table increased, CO2 flux increased, as well ( Figure 2.7). These results suggest that as the water table receded further below the soil surface, the aerobic conditions encouraged soil respiration. Similar trends were apparent when investigating below-ground organic matter decomposition; decomposition increased with increased depth between the soil surface and the water table (highest decomposition was reported in the driest soils). While there are multiple processes that contribute to soil respiration, these similar trends support the current belief that microbial respiration is favored in unsaturated soil environments (Inglett et al., 2012;Kuhn, 2015). All basin zone soils were saturated within the upper 25 cm at least during the months of May and June, while one site extended these saturated conditions through the month of July (CAR 3). The upper saturation of soils in the transitional zone varied as well, ranging from 1 to 2 months (Figure 2.8).
Further investigations of CO2 flux emitted during months when the upper portion of the soil exhibits anaerobic conditions could provide a more accurate assessment of differences between aerobic and anaerobic respiration.
Analysis of CO2 flux across all study sites revealed a significant relationship between zone and flux (p = 0.03; Table 2.4). Within CAR 3, CO2 flux also varied significantly by hydrologic zone (p < 0.001). In both of these comparisons, the transitional zone produced a significantly greater CO2 flux than the basin zone.
Although there was no significant difference between basin flux and upland flux, the increased flux exhibited by the drier transitional zone in comparison with the wetter basin zone supports the relationship suggested by linear regression analysis of hydrology versus subsurface organic matter decomposition; decomposition increased in drier soils. The literature offers contradicting evidence regarding the relationship between hydrology and soil respiration. For example, Davis et al. (2010) investigated soil respiration in seasonally saturated versus better drained upland soil environments and found that the very poorly drained soils emitted the lowest CO2 fluxes overall. In contrast, Ricker et al. (2014) found no significant difference in CO2 flux when comparing poorly-drained wetland and upland environments. Raich and Schlesinger (1992) suggested that the specific type of wetland determines whether wetlands or non-wetlands produce more CO2 to the atmosphere. Such contrasting evidence merits further investigation into the contribution of CO2 emissions from wetlands into the atmosphere.
Although soil temperature is the main factor influencing soil respiration (Kicklighter et al. 1994), soil moisture also influences soil respiration (Davis et al., 2010). When fit with a multiple regression, water table level and soil temperature were significantly correlated to CO2 flux in all vernal pool study sites (p < 0.001;  (1984) and Davis et al. (2010) found a stronger relationship between CO2 flux and soil temperature than with soil moisture. For example, Davis et al. (2010) found that 72-88% of CO2 flux could be attributed to soil temperature, but when analyzing soil moisture effects only 18-43% of the flux could be explained by soil moisture. The significant relationships between CO2 and soil temperature (Table 2.3) suggest that the methods I used for gas sampling and analysis provide at a minimum a relative measure of the controlling processes in regard to CO2 flux in vernal pools.
For understanding SOC sinks and sources, it is critical to identify whether the source of CO2 from the soil is a function of microbial decomposition of SOC or root respiration. Thus, I utilized a linear regression technique relating root biomass to CO2 flux to estimate root respiration (Ricker, 2014). Although the correlations were generally very weak (all R 2 values < 0.5; Table 2.5), the analysis suggested that root respiration accounted for < 40% of the CO2 flux (Table 2.6). The exception was the basin average in October which was twice as much (78%) as the next highest value (39%). If this high value is excluded, average root respiration values are < 30% suggesting that microbial decomposition is the dominant process toward CO2 efflux.
Root respiration varied significantly with hydrologic zone during the months of September, October, and November (p < 0.04; Table 2.7). In September, percent root respiration was significantly higher in the upland zone than the transitional zone, while in October, percent root respiration was significantly higher in the basin zone than in the transitional and upland zones. Percent root respiration was also significantly higher in the basin zone than in the transitional zone during the month of November (Table 2 Likewise, further investigation evaluating root versus microbial respiration in vernal pools by season would be helpful in ascertaining a more accurate relationship between these two pathways.

Methane
CH4 is considered to be one of the most potent GHG (Altor and Mitsch, 2008;Moseman-Valtierra et al., 2011). Vernal pool hydrology was significantly correlated to CH4 flux in all study sites (p < 0.003). All zones of CAR 2 and EP exhibited net  (Table 2.8).
CH4 flux varied significantly with hydrologic zone in all of the vernal pool study sites when analyzed via a one-way ANOVA both individually (p < 0.04) and across all pools (p < 0.001; Table 2.9). Flux was significantly higher in the basin zone than the upland zone at all sites, and was significantly higher in the basin zone than the transitional zone in three out of the four sites (results trended towards significance in GS). This can be attributed to the nature of methanogenesis as an anaerobic process.
Methanogenesis only occurs in areas of very low redox potential which is facilitated by periods of extended saturation (Kuhn, 2015) such as vernal pool basins.
Methanogenesis occurs when CO2, which is not thermodynamically favored by microorganisms, is the electron acceptor primarily available for redox reactions. were only present in flooded, anaerobic environments of low redox potential (Segers, 1998;Whiting and Chanton, 2001). Previous analyses suggested that subsurface organic matter decomposition rates are slowest in the wettest zones. It is possible that the activity of detritivores, or decomposing microorganisms, is heightened in aerobic environments. Further analyses investigating the microbial communities and activity in vernal pool basin versus upland environments in pools could provide insight into the reasoning behind differences in decomposition and methanogenesis efficiency.
The mean negative flux values indicate that during those months, CH4 was being absorbed by the soil over time, rather than emitted. The amount of CH4 removed from the atmosphere by soils is estimated at 30 Tg per year. Methane absorption is common in soils dominated by an aerobic environment. Such absorption is a process commonly associated with forested soils (Kagotani et al., 2001). Aerobic versus non-hydric soils, Altor and Mitsch found that hydric soils produced significantly more CH4, which was the case in my study, as well. Kuhn (2015) measured CH4 flux based on grams of soil rather than on an area basis, so exact quantities of flux could not be compared. However, the study compared flux in the same three zones as I used and found the highest production in the basin zones. The positive CH4 flux emitted by the basin zones may have implications on climate change, as CH4 as a greenhouse gas has a potency of 25 times that of CO2.

Nitrous Oxide
N2O is produced in wetland soils through the process of denitrification, in which microorganisms reduce nitrogen oxides into nitrogen gases. Specifically, N2O is an intermediate gas produced when denitrification does not proceed to completion (Bernhard, 2010). In this process, a source of nitrate (NO3 -) is necessary as an electron acceptor during the microbial metabolism (Capps et al., 2014). Because my sites are located in protected management areas, immediate sources of NO3 -, such as fertilizer, industrial waste, and urban drainage, may not be readily available. This may be why little N2O appeared to be emitted to the atmosphere from the vernal pools ( Figure 2.11). Other processes, however, also control how much N2O is released, including nitrification, nitrifier denitrification, and coupled nitrificationdenitrification (Moseman-Valiterra et al., 2011). These processes would likely require an oxic-anoxic interface to oxidize ammonium to nitrate, which are typically present at the top of the water

SUMMARY AND CONCLUSIONS
Wetland soils store much of the earth's terrestrial carbon pool, and thus their soil carbon budget has been increasingly investigated within the scientific community in an attempt to mitigate the impacts of global climate change. In this study, I evaluated the relationships between hydrologic regime and key wetland processes by examining the balance between inputs and outputs of carbon in vernal pool wetlands. Comparison of hydrology and decomposition trended toward significance in leaf litter and subsurface wooden dowel decomposition. Above-ground decomposition increased as the water table level became closer to the soil surface, while belowground decomposition increased as soils became more aerobic. Because substrates at the soil surface dry more readily than substrates below the surface, these results suggest that microbial decomposition is favored at moist field conditions, which were achieved at the surface in the basin zone and below the surface in the upland zone.
Depending on environmental conditions, microorganisms will respire either aerobically (via the tricarboxylic acid cycle) or anaerobically (via fermentation).
Aerobic respiration is typically a faster process than anaerobic respiration, supporting evidence obtained in this study that subsurface organic matter decomposition occurs at faster rates in aerobic environments (upland) than in anaerobic environments (basin).
In both years, the greater amount of decomposition experienced by the leaf litter than the wooden dowels affirms the important roles that C:N ratios and surface areas play in decomposition processes.
Because of its nature as a potent GHG, understanding CO2 flux dynamics is critical at a time when climate change threatens the environment. In order to evaluate these dynamics, I monitored GHG flux in the three hydrologic zones of the vernal pool. As the Earth's atmosphere warms, soil temperature will increase as well,