Deciphering global patterns of forest canopy rainfall interception (FCRI): A synthesis of geographical, forest species, and methodological influences
Carlos R de Mello,Li Guo,Chuan Yuan,André F Rodrigues,Renato R Lima,Marcela C N S Terra
DOI: https://doi.org/10.1016/j.jenvman.2024.120879
Abstract:Forest canopy rainfall interception (FRCI) is an essential hydrological process that governs water and biogeochemical cycles in forest ecosystems. Identifying patterns and relationships of FCRI using a systematic review is key to improving our knowledge supporting new experiment research, modeling, and application. In this meta-analysis, we aimed to delineate the canopy interception (CI), throughfall (TF), and stemflow (SF) concerning geographical and forest variables and experimental methodologies. We leveraged peer-reviewed 170 articles across 234 sites globally, extracting TF, CI, SF, geographical, forest, and experimental aspects. We applied multivariate statistical procedures to discern the principal influences on TF, CI, and SF and examined their multicollinearity. In addition, we developed Generalized Linear Models (GLM) for CI and TF. Global TF experiments indicate that the predominant rainfall devices, number of sample trees, number of events, and monitoring length are 10-20 devices (81% fixed), 3-6 trees, 30-50 events, and 10-30 months. Predominant global values of TF, CI, and SF are 70-80% (median = 73%), 20%-30% (median = 23.9%), and <1.0% (median = 1.87%), respectively. Global models of CI and TF were responsive to T, LAI, and D (respectively, R2adj of 0.196** and 0.206**). Temperate forests mirrored the global model (R2adj of 0.274** and 0.31**, respectively). The Subtropical CI model was fitted based on P and DBH (R2adj = 0.245*), and the TF model was based on E, D, and LAI (R2adj = 0.532**); the Mediterranean CI model was based on T, Basal, and LAI (R2adj = 0.45*), while TF was based on P, Basal, and LAI (R2adj = 0.671**). The Tropical CI model was based on T and H (R2adj = 0.396*), and the TF model, LAI, and P (R2adj = 0.35*). This meta-analysis underscores the importance of comprehending the hydrological processes in forested areas as they are pivotal in mitigating climate change impacts.