Seasonal wetlands in the tropics are important habitats for local and migratory bird species. In the northwestern Pacific of Costa Rica, Palo Verde National Park has one of the most important seasonal wetlands of Central America. The management history of this wetland has shown the impact of invasive plant species such as Parkinsonia aculeata L. whose cover extension and canopy structure impact not only the ecological niches of bird species, but also the wetland hydrology. A 300 m² plot was established in a P. aculeata stand to evaluate the role of P. aculeata on the partitioning and redistribution of precipitation. Gross precipitation (P_Gr ), throughfall (P_TF) and stemflow (P_SF) were measured on a daily basis to determine the interception of precipitation (P_I ) and net precipitation (P_Net ). A total of 43 precipitation events were sampled during the wet season of 2003. We measured 530.5 mm of P_Gr and 458 mm of P_TF, with an average sampling error of 0.7 mm or 6.1%. Canopy storage capacity was estimated at 1.47 mm, throughfall 88.73%, stem flow 2.63% and a total interception of 8.64%, with a P_Net coefficient of 0.9475. The relationships between gross precipitation (P_Gr) with throughfall (P_TF), stemflow (P_SF) and net precipitation (P_Net ) were evaluated using linear regression models. P. aculeata showed to have one of the highest net precipitation and lowest precipitation interception among small trees.

@en

Tropical dry forests (TDF) are endangered ecosystems characterized by a matrix of successional forest patches with structural differences across the Neotropics. Until now, there have been few studies that analyze the partitioning of rainfall by forest interception in TDF. To contribute to the understanding of the TDF impact on the hydrological dynamic at the ecosystem and landscape levels, a rainfall interception study was conducted in Santa Rosa National Park in Costa Rica (SRNP) and in Mata Seca State Park in Brazil (MSSP). In each site, three plots per successional stage were studied. The successional stages were early, intermediate, and late. In each plot the rainfall, throughfall, and stemflow were monitored during one rainy season. The relationship between gross rainfall and water fluxes was evaluated using linear regression models. In general, net rainfall oscillated from 79.3% to 85.4% of gross rainfall in all the plots in MSSP without any trend related to forest succession, due to the effect of a high density of lianas in the intermediate and late stage plots. In SRNP, there was a clear trend of net rainfall among successional stages: 87.5% (early), 73.0% (intermediate), and 63.4% (late). Net rainfall correlated negatively only with plant area index in SRNP (r = −0.755, p < 0.05). This study highlights the need to study rainfall interception in successional stages to estimate net rainfall that reaches the soil. This would provide better hydrological information to understand water balance and water fluxes at the level of forest ecosystems and landscapes.@en