Reported changes in precipitation may impact processes linked to growth and survival, with potential consequences to carbon balance in the tropics. Furthermore, precipitation regimes partially modulate the occurrence of forests and savannas in these regions. However, the local mechanisms responsible for the resilience of those ecosystems remain uncertain. We aim at understanding how forests and savannas under the same precipitation regime respond in terms of GPP to seasonal rainfall variability. We hypothesize that savannas respond faster to precipitation by changing their greenness (EVI2) because savanna trees and grasses usually occur in areas far away from the water table, while riparian forests occur in valleys with shallower water tables. We sampled a total of 20 (200m2) field plots, with 10 in cerrado stricto sensu (savanna) and 10 in riparian evergreen forests, at the Chapada dos Veadeiros National Park (PNCV), located within the Cerrado biome, in Brazil. We measured tree cover using a concave densiometer, and calculated the coupling, i.e. Kendall correlation, between EVI2 from Landsat8 and the monthly precipitation calculated from CHIRPS dataset, using time-series from 2013 through 2018. In savannas where trees and grasses co-exist, we standardized coupling values by tree cover to account only for the effect savanna trees have on the EVI2. Forests and savannas respond differently to precipitation. Maximum coupling in savannas is greater than in forests but when considering only trees in both ecosystems, we found similar responses. Furthermore, savannas respond faster than forests. Our results indicate that the GPP may depend on the proximity of the water table. However, further studies are necessary to understand which mechanisms are driving the patterns found here. Therefore, forests and savannas at the PNCV are functionally distinct, in agreement with structural differences they present, particularly taking into account the differences in the response time of forest and savanna communities.
- 127 Reads