Estrutura do dossel, dinâmica florestal e fenologia foliar com uso de aeronave remotamente pilotada na Amazônia Central

Abstract

A number of difficulties in studying the canopy of tropical forests can be resolved with the use of high spatial and temporal resolution optical remote sensing data, which enables mapping of individual tree crowns. To investigate some challenging canopy processes, we used 2 cm resolution Remotely Piloted Aircraft System (RPAS) imagery and forest dynamics data from plots remeasured bi-annually over a period of more than two decades. These data were utilized to quantify light environment and canopy positions in relationship to tree size distributions and long-term average growth rates. Sixteen overflights performed with RGB (Red, Green, Blue) and NGB (Near Infrared, Green, Blue) digital cameras were used to detect leaf phenology and its spectral response over the period of January 2016 to January 2017. Results demonstrated that crown diameter measured from the image was a fairly strong predictor for stem diameter at breast height (DBH), with relationship of 61%. Of all 1244 inventoried trees with DBH > 10 cm, 40% were in the canopy and 60% in the understory. A 25 cm DBH tree was more than 50% likely to be in the canopy. The canopy trees larger than this diameter followed an exponential distribution, typical of tropical forests. The understory individuals followed an exponential distribution and all individuals fitted a Weibull distribution. The mean annual increment (MAI) of all individuals measured in the period from 1996 to 2015 was 1.67 ± 0.09 mm.year-1. The canopy trees grew twice as large as those of understory with an MAI of 2.38 ± 0.19 and 1.17 ± 0.07 mm.year-1, respectively. Phenology results demonstrated an increase in the budding of new leaves and vegetation indexes in the dry period, confirming the green-up on plateau and valley environments. The highest peaks of new leaves occurred during September on the plateau and during October and November on valley. Many trees on the plateau anticipated the leaf change in March-April, probably due to the El Niño 2015-2016 drought, and this anticipation was not visualized on the valley, suggesting that the proximity of the water table influences directly on trees’ response to drought events. The percentage of trees with crowns totally and partially exposed that did leaf flushing was 60% and 30%, respectively. Results improved the canopy studies and revealed crown distributions and leaf phenology of trees from different light environments at this old growth Central Amazon forest site.