Key points are not available for this paper at this time.
The interaction between the forest canopy and precipitation is a fundamental process for understanding the hydrological cycle in forests. Physical models have been applied to estimate canopy water interception, and their efficiency has been tested based on metrics used to assess hydrological models. For eucalyptus plantations in Brazil, more studies are needed on the canopy rainfall interception model. Thus, we calibrated the Gash model using two complete hydrological years of observation in a eucalyptus plantation in southeastern Brazil. The model’s parametrization was conducted using 17 trees individually in different planting spacings (3 m × 2 m, 3 m × 3 m, and 3 m × 5 m). The average values of the model’s parameters were taken to represent the forest, and the average parameters for each planting spacing were used to assess the model’s performance according to the planting spacings. We used NSE, KGE, and Pbias statistical metrics to assess the model’s performance. For individual trees and rainfall events, the model showed an average NSE and Pbias of 0.59 and 18.2%, respectively, meaning a “satisfactory” performance for eight trees and “poor” performance for nine trees. When the model was averaged for the entire forest and individual rainfall events were considered, the metrics were improved, being 0.643 for NSE and 8.2% for Pbias, indicating a “good” model performance, which was strengthened by an average KGE of 0.746. Regarding the model for the planting spacings, the best results were found for the 3.0 m × 2.0 m spacing (“a good performance”). For the other spacings, Pbias was higher than 15%, leading to inferior performance, but with the NSE and KGE compatible with “good” performance. The practical implications of our findings are significant, as they can be used to enhance the accuracy of models for a better understanding of the hydrological cycle in eucalyptus forests in Brazil, thereby contributing to more effective forest management and conservation.
Neto et al. (Mon,) studied this question.