Monitoring Paraná River wetland dynamics using MODIS NDVI time series

SALVIA, MARÍA MERCEDES; KANDUS, PATRICIA; GRINGS, FRANCISCO MATIAS; KARSZENBAUM HAYDEE

Orlando

Congreso; 9 th INTECOL International Wetlands Conference "WETLANDS IN A COMPLEX WORLD"; 2012

The largest wetlands in South America are associated with the floodplains of the big rivers. Most of them belong to tropical and subtropical humid climate (Amazonas, Orinoco) but some extend to the subtropical-temperate zone, like those associated to the Paraná River. These wetlands have subregional extension mostly covered by herbaceous vegetation and showing a high spatial and temporal variability of the water table that constrains biogeochemical cycles and fluxes, and supports a very rich and particular biota well adapted to a wide range of water availability and hydroperiods. Paraná River wetlands are important habitat sustaining commercial fisheries, cattle ranching and apiculture providing also roughness surfaces for flood regulation. Multitemporal satellite images are now of standard use in land cover classification and monitoring of large areas and NDVI (Normalized Difference Vegetation Index) temporal series have been used for monitoring, by means of constructing seasonal activity profiles, obtaining critical information on phenology and seasonality of vegetation (Paruelo et al, 2001, Orshan, 1989). Even when there is a vast bibliography on the analysis of low or medium resolution NDVI temporal series, most of them were developed for terrestrial environments. Although there are only a few examples for wetland areas, we consider that time series of NDVI could be of great interest to assess processes and functions of the large wetlands of the Paraná floodplain. Therefore, at regional scale, we used a 8-year of MOD13Q1 product in order to analyze TERRA-MODIS NDVI signal over the large wetland ecosystems of the Paraná River Delta. Our objective was to identify spatial and temporal change patterns based on the hydrologic and plant phenology behavior and to gain insight on the impact of extreme hydrological events (EHE) on the river floodplain. For this, to assess the effect of EHE on wetland NDVI patterns, first we begin by studying vegetation patterns as seen by MODIS in years when no EHE took place (undisturbed). Second, we propose simple models to explain NDVI year-long variation due to phenology. Then using robust statistical tools, we investigate anomalies in NDVI patterns for different types of vegetation in different conditions, usually related to EHE. These anomalies were analyzed using its resilient time and temporal behavior to associate them to extreme environmental disturbances. We have found that marsh vegetation is both the more sensitive to floodplain forcing (flooding, drought, fire) and the one with less resilient time. The effect of these disturbances in marsh vegetation in future wetland dynamics is considered. In particular, the effect of the reduction of the vegetation drag coefficient due to the destruction of marsh overall biomass. We are currently working in mapping and classifying the parameters of the model used for the sample analysis, in order to compose a map that considers both the land cover type and its annual and interannual dynamics.