A freshwater marsh, created by artificially draining a 105-ha shallow lake (Rice Lake), is a suspected source of phosphorus (P) to the Detroit Lake watershed, a major recreational destination for northwestern Minnesota. P loadings to the main drainage canal increase during the summer months, when the water table typically declines. On the assumption that increased aeration causes P release through mineralization of organic matter, local authorities have proposed that the water level in the marsh be maintained at a higher level. To test this assumption, and to examine the impacts of water table control on greenhouse gas emissions, intact soil cores from 2 sites in the wetland were subjected in the laboratory to three water table levels and incubated for 6 wk. Dissolved reactive P (DRP), reduced Fe, pH, dissolved organic carbon (DOC), and oxidation-reduction potential (Eh) were monitored in the pore water. Carbon dioxide (CO2), methane (CH4), and nitrous oxide (N2O) fluxes were determined from headspace gas samples. Gas fluxes were converted to global warming potentials in CO2 equivalents. DRP dynamics differed between sites. One site displayed higher DRP flux (P<0.01) from the highest water table treatment, while the other site showed only trace levels of porewater DRP, with no significant differences among treatments. Greenhouse gas fluxes were similar at both sites, and significant differences were observed among treatments. At both sites, the highest water table treatment displayed lower cumulative N2O flux (P<0.05) and higher cumulative CH4 flux (P<0.05) than the lowest water table treatment. Cumulative CO2 equivalent data suggests that the highest water table treatment has the greatest global warming potential (significant at P<0.06). Thus, the proposal to increase water levels has the potential to increase both P release and total greenhouse gas emissions, although issues related to process variability across the wetland were difficult to assess.