GROUNDWATER-DRIVEN INTERMITTENCY REGIMES IN THE ARKANSAS RIVER, KANSAS

Non-perennial rivers and streams provide many important ecosystem services, including supporting wildlife and providing water to society. However, most non-perennial stream research has focused on small, headwater catchments with minimal human activity. As a result, much less is known about groundwater-surface water interactions and their relation to flow and intermittency in larger, human-impacted rivers. Here, we investigate the historical patterns and drivers of stream intermittency in a large (7th order) reach of the Arkansas River near Larned, KS, USA using multi-decadal streamflow, groundwater level, and stream stage data.

We found that the river has two distinct hydrologic regimes: a ‘wet regime’ characterized by near-continuous flow, and a ‘dry regime’ characterized by near-continuous no-flow conditions, with these regimes lasting months to years. Transitions between wet and dry regimes occur when alluvial aquifer water levels fall below (or rise above) the streambed elevation for wet to dry (or dry to wet) transitions. These transitions are associated with the long-term (annual and longer) climatic water balance, as prolonged precipitation deficits cause alluvial aquifer levels to fall and the stream to transition from wet to dry. Agricultural pumping in the underlying High Plains aquifer also contributes to long-term declines in alluvial aquifer levels and stream intermittency due to slow leakage through a semi-confining layer, but the net impact of downward leakage on alluvial aquifer levels and flow is less clear. In contrast, shorter-term (monthly to seasonal) climatic variability most strongly influences flow conditions when water levels in the alluvial aquifer are at or near the streambed elevation. Combined, this indicates that there is a complex interplay between top-down (climatic) and bottom-up (pumping) controls on intermittency in this large river, both of which can have long time lags before their impacts are felt on streamflow.