Characterization of the inflow and outflow components explaining water balance dynamics in the wetlands of Palo Verde National Park, Costa Rica

Details

Supervisors

Javaux, Mathieu ; Alonso, Alice

Faculty

Faculté des bioingénieurs

Degree label

Master [120] : bioingénieur en sciences et technologies de l'environnement, à finalité spécialisée

Abstract

Covering 5 to 8 percent of the Earth's surface, wetlands are vital ecosystems at the intersection of land and water, providing essential ecological functions such as pollutant filtration, shoreline protection and climate regulation. Palo Verde National Park, located in northwestern Costa Rica, is among the most ecologically significant tropical wetlands in Central America. Due to human activities, since the 1970s, this Ramsar Wetland has faced significant ecological disruptions affecting its hydrology and biodiversity. Indeed, these disturbances have led to the invasion of native species like Typha domingensis and a significant decline in bird populations. The exact causes and consequences of these disturbances are likely multifaceted and are still poorly understood. To understand these impacts, this study identifies the hydrological processes shaping the water balance of Palo Verde sub-wetland. An exploratory analysis of hydrological data available over a 15-year period (2003-2018) is conducted. The components analysed are precipitation, evapotranspiration and river overflow. The research began by correcting, extrapolating or comparing raw data for each water balance component, followed by intra-annual and inter-annual analyses. The simplified water balance calculated from these components is then compared to actual water levels in the lagoon to identify discrepancies and potential errors. This study identifies inaccuracies in precipitation data affecting the water balance of the Palo Verde sub-wetland, primarily due to underestimation by the Palo Verde rain gauge. Remote sensing products TRMM 3B42 and GSMaP V8 are employed for validation, though they are not entirely performant. The research also aims to refine the characterization of evapotranspiration, estimated through ET remote sensing products MOD16A2, PML-V2, and GLEAM v4.1a. The analysis reveals notable discrepancies in evapotranspiration estimates among these products. The PML-V2 product provides a detailed characterization of evapotranspiration components, highlighting the predominance of plant transpiration. MOD16A2 and GLEAM v4.1a provide consistent trends in hydric stress, despite some unexplained biases. Additionally, the study assesses a model for extrapolating missing river water levels. This model, based on astronomical phenomena, performs well in capturing when overflow events occur due to tides but is less accurate in quantifying the intensity of these overflows. The calculated water balance shows significant discrepancies compared to the actual water levels in the lagoon, indicating a substantial missing flux. This is hypothesized to be soil infiltration. Estimations based on the water balance, suggest that infiltration rates are 4 to 5 times higher during the dry season, likely due to water seeping into the cracks of dry vertisols.