] The San Joaquin Freshwater Marsh (SJFM) is a seasonally flooded Typha wetland in Southern California that is characterized by high rates of Aboveground Net Primary Production (ANPP) and a large accumulation of standing leaf litter. The ANPP, Gross and Net Ecosystem CO2 Exchange (GEE and NEE), and Enhanced Vegetation Index (EVI) at the SJFM fluctuate by ∼40% from year to year, in ways that are not directly attributable to variation in weather or the maximum green Leaf Area Index (LAImax). We tested the hypothesis that this variation is caused by a negative feedback between ANPP, the buildup of leaf litter, shading of green leaves by litter, a reduction in GEE and NEE, and a subsequent reduction in ANPP. Litter manipulations on replicated plots demonstrated that the presence of standing litter decreased plot-level NEE by 17 to 47% and surface EVI by 25 to 48%, even as green Leaf Area Index (LAIgreen) was held constant. Plot level NEE and surface EVI remained tightly correlated, and largely decoupled from LAIgreen, as standing litter was varied. This pattern paralleled that observed for the entire marsh, where NEE and EVI remained tightly correlated, and largely decoupled from LAImax, from year to year. Correcting LAIgreen and LAImax for the amount of shading caused by standing litter improved the correlations between LAI and EVI and NEE, indicating that EVI and NEE are most sensitive to the amount of unshaded LAI. The accumulation of standing litter at the SJFM decouples the relationships between LAI and EVI and NEE, and appears to be important for controlling the interannual variability observed at the site.