Member and Bayesian mixing models consistently indicate near‐surface flowpath dominance in a pristine humid tropical rainforest

The impacts of forest conversion on runoff generation in the tropics have received much interest, but scientific progress is still hampered by challenging fieldwork conditions and limited knowledge about runoff mechanisms. Here, we assessed the runoff generation, flow paths and water source dynamics of a pristine rainforest catchment in Costa Rica using end member mixing analysis (EMMA) and a Bayesian mixing model (MixSIAR). Geochemical tracer data collected over a 4-week field campaign were combined with tritium data used to assess potential deeper groundwater flow pathways to the perennial stream. The streamflow composition was best captured using three end-members, namely throughfall, shallow (5–15 cm) and deeper (15–50 cm) soil water. We estimated the end-member contributions to the main stream and two tributaries using the two mixing approaches and found good agreement between results obtained from EMMA and MixSIAR. The system was overwhelmingly dominated by near-surface sources, with little evidence for deeper and older groundwater as tritium-derived baseflow mean transit time was between 2.0 and 4.4 years. The shallow soil flow pathway dominated streamflow contributions in the main stream (median 39% and 49% based on EMMA and MixSIAR, respectively), followed by the deeper soil (32% and 31%) and throughfall (25% and 19%). The two tributaries had even greater shallow soil water contributions relative to the main stream (83% and 74% for tributary A and 42% and 63% for tributary B). Tributary B had no detectable deep soil water contribution, reflecting the morphology of the hillslope (steeper slopes, shallower soils and lower vegetation density compared to hillslope A). Despite the short sampling campaign and associated uncertainties, this study allowed to thoroughly assess runoff generation mechanisms in a humid tropical catchment. Our results also provide a first comparison of two increasingly used mixing models and suggest that EMMA and MixSIAR yield comparable estimates of water source partitioning in this tropical, volcanic rainforest environment.Universidad de Costa Rica/[217-B8-276]/UCR/Costa RicaUniversidad de Costa Rica/[217-B8-709]/UCR/Costa RicaUniversidad de Costa Rica/[ED-3319]/UCR/Costa RicaInternational Atomic Energy Agency/[CRP22009]/IAEA/Estados UnidosUCR::Vicerrectoría de Investigación::Unidades de Investigación::Ciencias Básicas::Centro de Investigaciones Geofísicas (CIGEFI)UCR::Vicerrectoría de Docencia::Ciencias Sociales::Facultad de Ciencias Sociales::Escuela de Geografí