Abstract Suspended sediment is an important regulator of stream habitat quality but notoriously difficult to predict and regulate. This difficulty arises because of high natural variability in suspended sediment yield in space and time. Here we quantified associations between suspended sediment yields and discharge, watershed setting (i.e., physiography and lithology), and disturbance history for 10 temperate mountain watersheds (8.5–6,242 ha) in the U.S. Pacific Northwest (H.J. Andrews Long-Term Ecological Research, LTER) over an ~60-year period. Annual suspended sediment yields varied almost 4 orders of magnitude across space and time. A linear mixed effects model indicated that much of the variation in yields could be explained by the random effect of site (conditional R2 = 0.74) with additional variation explained by the fixed effects (marginal R2 = 0.67) of cumulative annual discharge (p < 0.001) and the variability (standard deviation) of watershed slope (p < 0.001). Two annual sediment yield data points were model outliers, that each occurred within a decade after forest management activities and a large-magnitude storm event at sites with high variability of catchment slope. Other sites had low sediment yields for a range of conditions, including management or flood disturbance. Taken together, our study shows that watersheds with high slope variability have higher suspended sediment yields and may be more vulnerable to increases in sediment yields following disturbances.

Keywords: H.J. Andrews, suspended sediment yield, Oregon, LTER, long-term data, forest management