Temporal dynamics of leaf phenology in a vegetation mosaic of the campo rupestre, Serra do Cipó, Minas Gerais, Brazil

Link To Share This Poster: https://cdmcd.co/XGgwQj
Live Discussion Link: https://cdmcd.co/LdDg7v

Arthur Bruck, Department of Biodiversity, São Paulo State University (UNESP), Rio Claro, Brazil, Patricia Morellato, Department of Biodiversity, São Paulo State University (Unesp), Rio Claro, Brazil and Bruna Alberton, Department of Biodiversity, UNESP Universidade Estadual Paulista, Rio Claro, Brazil

Background/Question/Methods
The Brazilian campo rupestre (rupestrian grassland) represent a high complex and diverse ecosystem, formed by a grassy-shrub vegetation and composed by a high diversity of species and endemisms. Due the scarcity of phenological records in the tropics for grassy environments, few is known about vegetation responses, rates of productivity, and carbon uptake from these ecosystems. Leaf phenology allows the monitoring of leafing exchange strategies and leads to the understanding of carbon-water relations in plants. Digital cameras are reliable tools to monitor open ecosystems, given their easy setup and acquisition of high quality and temporal resolution data. Here, we used a 7-year data collection of imagery records to investigate the vegetative phenology of the campo rupestre, aiming (i) to describe the seasonal patterns of leaf phenology; (ii) to compare these patterns among different vegetation physiognomies; and (iii) to evaluate shifts in the temporal patterns of the vegetation after a fire event. The study site was monitored by a time-lapse camera, and Regions of Interest (ROIs) were selected corresponding to different vegetation physiognomies within the images. The vegetation index Gcc was calculated from the image pixels and it was used to track leafing transitions along the 7 years (2013-2019).
Results/Conclusions
Approximately 38.000 images were processed, and Gcc time series were extracted from 6 ROIs corresponding to a stony grassland, a wet grassland, a sandy grassland, two rocky outcrops and to the entire landscape. The partial results showed, in general, similar seasonal patterns of leafing among all ROIs, as well as in the amplitude range of the Gcc index values. Our results suggest that common environmental constraint is driving the leaf flushing and senescence for the whole site. In the other hand, regarding the fire event, our results showed different patterns of vegetation post-fire responses. Despite the rapid regeneration observed for all ROIs, Gcc curves presented distinct behaviors among vegetation physiognomies. The wet and sandy grasslands showed the fastest post-fire vegetation recovery, reaching a marked Gcc peak 49 days after fire. The rocky outcrops showed a slower and gradual recovery of all vegetations, without a sharp peak in the Gcc index. Efforts in the development of further analysis are now in progress to integrate environmental factors and extract ecological information to better understand the spatial-temporal dynamics of this highly diverse and endangered vegetation mosaic.