Climatic determinants of photosynthetic activity in humid subtropical forests under different forestry activities

Abstract

Forests are key ecosystems providing a variety of contributions to societies that are based on the ecological processes occurring in them. Understanding the relationship between ecological processes and the climate system is essential to predict how they will respond to possible future climatic conditions and trajectories. In this study, the relationship between climate variables and vegetation dynamics was studied in different plots of preserved native forest, selectively logged forests, and pine plantations in an area of humid subtropical forest in Misiones, Argentina. Time series analysis via gaussian processes and multiple regressions with the moving average of explanatory variables combined with Bayesian model selection using the lowest Bayesian Information Index as selection criteria were used to determine the relationship between the Enhanced Vegetation Index (EVI) and meteorological variables including temperature, rainfall, global radiation, and potential evapotranspiration. Our results showed that EVI's variability is best explained by a combination of the moving average of temperature, global radiation, and the logarithm of potential evapotranspiration. The estimated EVI value is the minimum predicted value of each linear function of these variables, indicating that these ecosystems are conditioned by the most limiting meteorological variable for vegetation growth at each time among potential evapotranspiration, global radiation, and temperature, but not by a linear combination of these. The different ecosystems, in turn, responded differently to climatic variables. Both logged forests and pine plantations exhibited near-zero slopes with both global radiation and temperature, indicating they are close to temperature and radiation saturation points. Moreover, they presented a higher sensitivity to these variables than preserved forests that were more sensitive to atmospheric water demand. The proposed methodology allowed us to separate the external climatic influence in the forest photosynthetic activity from the internal vegetation processes. Moreover, the final model was capable to capture the multi-scale temporal patterns of forest vegetation.