When it rains, water remains in the ground for variable amounts of time before it is taken up by plants or becomes streamflow. Understanding how long water stays in the ground before it is taken up by plants or becomes streamflow helps predict what will happen to the water cycle in future climates. Some studies suggest that plants take up water that has been in the ground for a long time; in contrast, we find that plants take up a significant amount of recent rain.

Mountains store and release water, serving as water towers for downstream regions and affecting global sediment and carbon fluxes. We use stream and rain chemistry to calculate how much streamflow comes from recent rainfall across seven sites in the Andes mountains and the nearby Amazon lowlands. We find that the type of rock and the intensity of rainfall control water retention and release, challenging assumptions that mountain topography exerts the primary effect on watershed hydrology.

Himalayan rivers are full of giant boulders that rarely move except during glacial lake outburst floods (GLOFs), which therefore must be important drivers of erosion in the Himalayas. GLOFs are rare, so little is known about their long-term erosional impact. We found that rivers in Nepal have channel geometry that, compared with markers of upstream glaciation, confirm GLOFs as a major control on erosion. This previously unrecognized control should be accounted for in landscape evolution studies.

Arctic rivers erode into permafrost and mobilize organic carbon, which can react to form greenhouse gasses or be re-buried in floodplain deposits. We collected samples on a permafrost floodplain in Alaska to determine if more carbon is eroded or deposited by river meandering. The floodplain contained a mixture of young carbon fixed by the biosphere and old, re-deposited carbon. Thus, sediment storage may allow Arctic river floodplains to retain aged organic carbon even when permafrost thaws.

We use microCT scanning to improve the quality of ³He exposure ages measured in detrital magnetite. We show that the presence of inclusions can significantly increase the measured amount of ³He and thereby the exposure age. By prescreening magnetite with microCT and analyzing only inclusion-free grains, this problem can be avoided. We also calibrate the cosmogenic ³He production rate in magnetite relative to ¹⁰Be in quartz, which can be used for similar studies in the future.