Thursday, August 7, 2008 - 4:20 PM

COS 104-9: Linking trace gas emissions and hydrologic variability in coastal plain wetlands under contrasting land uses

Jennifer L. Morse, Marcelo Ardon, and Emily S. Bernhardt. Duke University

Background/Question/Methods

Emissions of the biogenic greenhouse gases N2O and CH4 have been well documented in many natural and human-dominated wetlands, and hydrologic factors are known to be major determinants of these fluxes.  Linking the rates and patterns of trace gas fluxes with environmental variability is a critical step toward estimating landscape-level annual fluxes and building predictive models.  Understanding these relationships is particularly important in hydrologically dynamic ecosystems that are vulnerable to climate change, sea-level rise, and land-use change.  Our research objective is to develop statistical relationships to link patch-scale trace gas flux rates to site hydrology and substrate supply for a range of coastal wetland ecosystems in North Carolina.  We examined four sites with contrasting land-uses and hydrologic regimes:  a large-scale restored riverine wetland (400+ ha), two forested wetlands, and an actively drained agricultural field in Tyrrell County, NC.  We used static chamber incubations and gas chromatography to measure CH4 and N2O fluxes.  We also collected soil solution water from piezometers for chemical analysis (NO3-, NH4+, SO4-2, TDN, TOC, and PO4-3).  To complement the patch-scale sampling, we used dataloggers to record continuous water level, soil redox potential, soil moisture, and soil temperature, at multiple sampling locations in each site.We found high temporal and spatial variability for both N2O and CH4 fluxes, both within and across our study sites. 

Results/Conclusions

In general, we found high CH4 emissions in sites with fully saturated conditions, while N2O fluxes were usually highest in sites that are less than fully saturated (with measurable N2O consumption occurring in some fully saturated sites). Separate stepwise multiple regression models for CH4 and N2O were developed for each site, with variable explanatory power.  Indices of hydrologic variability were developed to characterize antecedent hydrologic conditions for inclusion in models.  Strong predictor variables for N2O included water table elevation, soil temperature, and soil solution [NO3-] (e.g., r2 = 0.70 for July 2007 in restored wetland). Relationships between gas fluxes and environmental parameters allow extrapolation from our measurements to the landscape and annual scales, and permit sensitivity analysis of gas fluxes with respect to changes in temperature and hydrologic variability in these vulnerable ecosystems.