Search for: All records

Stable isotope variations are extremely useful for flow partitioning within the hydrologic cycle but remain poorly understood throughout the tropics, particularly in watersheds with rapidly infiltrating soils, such as Andisols in Central America. This study examines the fluctuations of stable isotope ratios (δ¹⁸O and δ²H) in the hydrologic components of a tropical coffee agroforestry watershed (~1 km²) with Andisol soils in Costa Rica. Samples were collected in precipitation, groundwater, springs, and stream water over 2 years. The local meteoric water line for the study site was δ²H = 8.5 δ¹⁸O + 18.02 (r² = 0.97,n = 198). The isotope ratios in precipitation exhibited an enriched trend during the dry season and a notable depletion at the beginning of the wet season. The δ¹⁸O compositions in groundwater (average = −6.4‰, σ = 0.7) and stream water (average = −6.7‰, σ = 0.6) were relatively stable over time, and both components exhibited more enriched values in 2013, which was the drier year. No strong correlation was observed between the isotope ratios and the precipitation amount at the event or daily time‐step, but a correlation was observed on a monthly scale. Stream water and base flow hydrograph separations based on isotope end‐member estimations showed that pre‐event water originating from base flow was prevalent. However, isotope data indicate that event water originating from springs appears to have been the primary driver of initial rises in stream flow and peak flows. These results indicate that isotope sampling improves the understanding of water balance components, even in a tropical humid location, where significant variations in rainfall challenge current modelling efforts. Further research using fine‐scale hydrometric and isotopic data would enhance understanding the processes driving spring flow generation in watersheds.

By mediating evapotranspiration processes, plant canopies play an important role in the terrestrial water cycle and regional climate. Substantial uncertainties exist in modeling canopy water interception and related hydrological processes due to rainfall forcing frequency selection and varying canopy traits. Here we design a new time interpolation method “zero” to better represent convective‐type precipitation in tropical regions. We also implement and recalibrate plant functional type‐specific interception parameters for rainforests and oil palm plantations, where oil palms express higher water interception capacity than forests, using the Community Land Model (CLM) versions 4.5 and 5.0 with CLM‐Palm embedded. Reconciling the interception scheme with realistic precipitation forcing produces more accurate canopy evaporation and transpiration for both plant functional types, which in turn improves simulated evapotranspiration and energy partitioning when benchmarked against observations from our study sites in Indonesia and an extensive literature review. Regional simulations for Sumatra and Kalimantan show that industrial oil palm plantations have 18–27% higher transpiration and 15–20% higher evapotranspiration than forests on an annual regional average basis across different ages or successional stages, even though the forests experience higher average precipitation according to reanalysis data. Our land‐only modeling results indicate that current oil palm plantations in Sumatra and Kalimantan use 15–20% more water (mean 220 mm or 20 Gt) per year compared to lowland rainforests of the same extent. The extra water use by oil palm reduces soil moisture and runoff that could affect ecosystem services such as productivity of staple crops and availability of drinking water in rural areas.