Resistance and resilience of stream metabolism to high flow disturbances

Streams are ecosystems organized by disturbance. One of the most frequent and variable disturbances in running waters is elevated flow. Yet, we still have few estimates of how ecosystem processes, such as stream metabolism (gross primary production and ecosystem respiration; GPP and ER), respond to high flow events. Furthermore, we lack a predictive framework for understanding controls on within-site metabolic responses to flow disturbances. Using 5 years of high-frequency dissolved oxygen data from an urban- and agricultural-influenced stream, we estimated daily GPP and ER and analyzed metabolic changes across 15 isolated high flow events. Metabolism was variable from day to day, even during lower flows; median and ranges for GPP and ER over the full measurement period were 3.7 (minimum, maximum = 0.0, 17.3) and −9.6 (−2.2, −20.5) g O2 m−2 d−1. We calculated metabolic resistance as the magnitude of departure (MGPP, MER) from the mean daily metabolism during antecedent lower flows (lower values of M represent higher resistance) and estimated resilience as the time until GPP and ER returned to the prior range of ambient equilibrium. We evaluated correlations between metabolic resistance and resilience with characteristics of each high flow event, antecedent conditions, and time since last flow disturbance. ER was more resistant and resilient than GPP. Median MGPP and MER were 0.38 and −0.09, respectively. GPP was typically suppressed following flow disturbances, regardless of disturbance intensity. The magnitude of departure from baseflow ER during isolated storms increased with disturbance intensity. Additionally, GPP was less resilient and took longer to recover (0 to >9 d, mean = 2.5) than ER (0 to 6 d, mean = 1.1). Prior flow disturbances set the stage for how metabolism responds to later high flow events: the percent change in discharge during the most recent high flow event was significantly correlated with M of both GPP and ER, as well as the recovery intervals for GPP. Given the flashy nature of streams draining human-altered landscapes and the variable consequences of flow for GPP and ER, testing how ecosystem processes respond to flow disturbances is essential to an integrative understanding of ecosystem function.