B44B-08
Continuous Ship-borne Methane Measurements on the Upper Mississippi River and Selected Tributaries

Thursday, 17 December 2015: 17:45
2010 (Moscone West)
John Crawford, USGS, Baltimore, MD, United States, Luke C Loken, University of Wisconsin Madison, Center for Limnology, Madison, WI, United States, Mark Dornblaser, US Geological Survey, Boulder, CO, United States, Emily H Stanley, Univ Wisconsin, Madison, WI, United States and Robert G Striegl, USGS Colorado Water Science Center Boulder, Boulder, CO, United States
Abstract:
Despite evidence that streams and rivers contribute immensely to the atmospheric methane budget (~26 Tg CH4 yr-1), very little is known regarding the spatial patterns and controls of methane concentrations in river networks. We present a dataset of high-resolution methane concentrations along a nearly complete river flowpath starting with a small headwater stream (8 km), two larger tributaries (50 and 80 km reaches), as well as the complete length of the upper Mississippi River (1300 km). These systems span from 1st to 9th order and range in discharge from 5 cfs to > 400,000 cfs. Continuous measurements were collected from a moving boat using a flow-through sampling system with cavity ring-down spectroscopy of gas equilibrated water. River methane concentrations ranged from near saturation to > 5 uM with all samples being above atmospheric equilibrium. The extent of methane spatial autocorrelation generally increased with increasing river size (semivariance range = 800, 4000 and 12,000 m), although the largest tributary reach did not exhibit clear spatial autocorrelation structure. Further, all river sections exhibited significant spatial clustering of methane concentrations (Global Moran’s I) and significant hotspots and coldspots of methane (local Moran’s I) associated with changes in benthic geomorphology. Hotspot examples included high methane clusters in organic-rich stream sediments and productive backwaters in the mainstem of the Mississippi River. Incubated anoxic stream sediments illustrated similar patterns, where organic-rich sediments produced substantially greater methane over 24 hours relative to organic-poor substrates. Quantitative PCR analysis of the methanogen gene mcrA also supports the contention that methane is produced at greatest rates in organic-rich stream sediments. Together, our high resolution spatial data and ancillary ecosystem data suggest that river methane is mostly controlled by local sediment processes which become more spatially homogeneous with increasing stream order.