The influence of land cover and storm magnitude on hydrologic flowpath activation and runoff generation in steep tropical catchments of central Panama

Despite abundant research documenting that land use/land cover (LULC) have substantial impacts on the hydrology of humid tropical systems, field-based evidence for the physical mechanisms behind these impacts are still lacking. In particular, our understanding of the hydrologic flowpaths that generate runoff in these systems, and how they vary with respect to LULC is insufficient to inform both physically-based hydrologic modeling and land-use decision-making. In this study, we use end-member mixing analysis (EMMA) of stream chemistry, and hydrometric characterizations of hillslope soil moisture to identify hydrologic flowpaths in humid tropical steep-land catchments of varying LULC: mature tropical forest, young secondary tropical forest, cattle pasture. EMMA was applied to data from 14 storm events (six at the mature forest, five at the young secondary forest, and three at the cattle pasture) that were intensively sampled during the 2017 wet season representing a wide range of rainfall magnitudes and intensities. Additionally, volumetric-soil-moisture responses at multiple depths were characterized during and after 74 storm events occurring from 2015 to 2017. EMMA results indicated that lateral preferential flow within the top 30 cm of the soil profile was a dominant source of runoff generation at the two forested catchments, with the contribution of this flow path increasing with rainfall magnitude and intensity. This was corroborated by volumetric-soil-moisture data, that showed that a perched zone of saturation developed at 30 cm at the time of peak storm runoff during the largest events and lasted for the remaining duration of the event. EMMA indicated that runoff was a combination of infiltration-excess overland flow and lateral subsurface flow in the actively grazed pastoral catchment. There, overland flow contributed 62 % of runoff during the highest runoff rate sampled (35.3 mm/hr) and this contribution increased substantially with storm magnitude. This flowpath identification was also supported by volumetric-soil-moisture data at the pasture, with peak saturation at all depths during the largest storm events occurring up to 30 min after peak runoff. These results provide a mechanistic explanation for previously observed hydrologic differences among tropical LULCs. Additionally, the wide range of hydrologic conditions during these storm events provide a basis for understanding how future changes to this, and similar humid tropical regions will impact hydrological processes and water availability.