Disentangling litter decomposition trajectories in streams: Ecological insights from a global multimodel analysis

Stream litter decomposition is a key ecosystem function commonly modeled using a single exponential decay function. However, the flow-driven nature of stream environments may lead to temporally heterogeneous decay dynamics that are poorly captured by this approach. Here, we evaluated the decomposition process of leaf litter in streams with four decomposition models (single exponential, double exponential, asymptotic, and Weibull) using 523 time series observations across the globe. Results showed that (1) the single exponential model provides a simple and reliable description of litter decomposition dynamics and captures a broad temporal pattern of mass loss; (2) double exponential and asymptotic models were applicable to only a subset of observations, whereas the Weibull model showed relatively consistent applicability across datasets; and (3) climate zone, leaf type, litterbag mesh size, and stream water temperature emerged as the key factors associated with decomposition parameters from different models. Overall, our results suggest that the single exponential model provides a robust description of stream litter decomposition, whereas alternative formulations such as the Weibull model could provide complementary views of decomposition dynamics in stream ecosystems.