Monitoring Vegetation in South America

Trends and ENSO/AAO Driven Variability in NDVI Derived Productivity and Phenology alongside the Andes Mountains

ARSC Group

Increasing water use and droughts, along with climate variability and land use change, have seriously altered vegetation growth patterns and ecosystem response in several regions alongside the Andes Mountains. Thirty years of the new generation biweekly normalized difference vegetation index (NDVI3g) time series data show significant land cover specific trends and variability in annual productivity and land surface phenological response.

Land cover types based on data from Eva et al.[ 33 ]. The black circles are the sample areas representative of the major land cover types that are used to show the phenology and some examples of the seasonal and interannual NDVI time series data within the indicated study area extent.

Productivity is represented by the growing season mean NDVI values (July to June). Arid and semi-arid and sub humid vegetation types (Atacama desert, Chaco and Patagonia) across Argentina, northern Chile, northwest Uruguay and southeast Bolivia show negative trends in productivity, while some temperate forest and agricultural areas in Chile and sub humid and humid areas in Brazil, Bolivia and Peru show positive trends in productivity.

Significant positive and negative trends (p ≤ 0.05) in annual mean NDVI based on 30 years of continuous NDVI time series data (1982–2011) are shown in the right side figure. Land cover is provided as a reference.

The start (SOS) and length (LOS) of the growing season results show large variability and regional hot spots where later SOS often coincides with reduced productivity. A longer growing season is generally found for some locations in the south of Chile (sub-antarctic forest) and Argentina (Patagonia steppe), while central Argentina (Pampa-mixed grasslands and agriculture) has a shorter LOS.

Thirty-year significant trends (p ≤ 0.05) in the SOS and LOS indicating regions where SOS is earlier or later and where the LOS is longer or shorter. Many of these shifts span 30 to 90 days over 30 years, but are often very variable from year to year.

Some of the areas have significant shifts in SOS and LOS of one to several months. The seasonal Multivariate ENSO Indicator (MEI) and the Antarctic Oscillation (AAO) index have a significant impact on vegetation productivity and phenology in southeastern and northeastern Argentina (Patagonia and Pampa), central and southern Chile (mixed shrubland, temperate and sub-antarctic forest), and Paraguay (Chaco).

Annual Mean NDVI values (1982–2011) are responding most strongly to winter MEI and summer AAO data (p ≤ 0.05).

Land cover types based on data from Eva et al.[ 33 ]. The black circles are the sample areas representative of the major land cover types that are used to show the phenology and some examples of the seasonal and interannual NDVI time series data within the indicated study area extent.

Significant positive and negative trends (p ≤ 0.05) in annual mean NDVI based on 30 years of continuous NDVI time series data (1982–2011) are shown in the right side figure. Land cover is provided as a reference.

Thirty-year significant trends (p ≤ 0.05) in the SOS and LOS indicating regions where SOS is earlier or later and where the LOS is longer or shorter. Many of these shifts span 30 to 90 days over 30 years, but are often very variable from year to year.

Annual Mean NDVI values (1982–2011) are responding most strongly to winter MEI and summer AAO data (p ≤ 0.05).