Controls on patterns of coarse organic particle retention in headwater streams

Organic matter retention is an integral ecosystem process affecting C and nutrient dynamics and biota in streams. Influences of discharge (Q), reach-scale channel form, and riparian vegetation on coarse particulate organic matter (CPOM) retention were analyzed in 2 headwater streams in northeastern Oregon. Ginkgo biloba leaves were released in coniferous forest reaches and downstream floodplain meadow reaches during spring high flow and summer baseflow. Transitional readies were also analyzed during summer baseflow. Paper strips, simulating sedge blade retention, were released in meadow reaches during high flow. Mean transport distances (S-P) were calculated as the inverse of the longitudinal loss rate (k) of leaves in transport. The metrics S-p, width-specific discharge (Q(w) = Q/stream width), and the mass transfer coefficient (v(dep) = Q(w)/S-p) were used to investigate retention. Values of S-p (0.9-97 m) were 2 to 11 times longer during high flow than baseflow. Mean S-p in forest reaches (29.3 m) was significantly shorter than in meadow reaches (68.9 m) during high flow but not during baseflow. Standardizing k for the scaling effects of Q by analyzing the relationship between Q(w) and S-p, in which the slope equaled the inverse of mean v(dep) of all Ginkgo releases, indicated times when v(dep) was higher or lower than predicted by Q. Values of S-p were driven largely by Q, yet most experiments in which values of v(dep) exceeded those predicted by Q(w) occurred during high flow. Values of v(dep) (0.3-32 mm/s) across experiments were generally inversely related to S-p but did not differ between forest and meadow readies during high flow. Unlike meadow reaches, mean v(dep) in forest readies was higher during high flow (5.2 mm/s) than baseflow (1.1 mm/s). Values of v(dep) were positively related to large wood volume and negatively related to the extent of floodplain inundation during high flow. Yet, in the meadow reach that had lower relative channel constraint, paper strips were transported farther onto the floodplain as Q rose, resulting in long-term (similar to1.5 mo) retention. Despite downstream increases in Q, there were no differences in mean baseflow S-p or v(dep) among readies in either stream, indicating some longitudinal compensation in retention. Alternating associations between retention metrics, and structural elements of the stream channels between flow periods suggests dynamic reach-scale hydrologic-retention thresholds in response to changes in Q. Analysis of v(dep) across experiments indicated that channel morphology, stream wood, and riparian vegetation are major controls on CPOM retention.