Abstract
Dissolved oxygen (DO) concentrations reflect hydrologic, climatic, and biological processes that regulate stream ecosystem function, while DO regimes are emergent ecosystem states arising from these interacting hydrologic and biological processes. Thus, DO regimes offer a framework for understanding DO dynamics and drivers across diverse stream networks. In non-perennial streams, recurrent transitions among flowing, pooled, and dry phases alter the balance between physical and biological processes that regulate DO, making these ecosystems particularly vulnerable to oxygen stress and sensitive indicators of how hydrology mediates ecosystem function. We analyzed high-frequency DO, hydrologic, and climatic data from eight non-perennial watersheds across the intermountain west, Great Plains, and southeastern forest regions of the United States to characterize DO regimes and their environmental drivers. Using multivariate clustering, we identified three diel DO states: oxygenated (concentrations near saturation with low diel variability), transitional (large diel fluctuations), and hypoxic (low saturation with little diel fluctuation). Random forest and SHapley Additive exPlanations (SHAP) analyses revealed that active surface drainage network (ASDN) is the key driver of DO regimes: increased ASDN and discharge promoted oxygenated states, while reduced network extent, higher temperatures, and greater light availability intensified biological oxygen demand and favored transitional and hypoxic states. This regime-based framework provides a transferable, mechanistic context for understanding and anticipating changes in DO dynamics across diverse stream ecosystems.
Comments
There are currently no comments
New Comment