Title
On the Hydroclimate-Vegetation Relationship in the Southwestern Amazon During the 2000–2019 Period
Date Issued
24 March 2021
Access level
open access
Resource Type
journal article
Author(s)
Sorbonne Université
Institut des Géosciences de l'Environnement
Institut des Géosciences de l'Environnement
Publisher(s)
Frontiers Media S.A.
Abstract
The southern Amazonia is undergoing a major biophysical transition, involving changes in land use and regional climate. This study provides new insights on the relationship between hydroclimatic variables and vegetation conditions in the upper Madeira Basin (~1 × 106 km2). Vegetative dynamics are characterised using the normalized difference vegetation index (NDVI) while hydroclimatic variability is analysed using satellite-based precipitation, observed river discharge, satellite measurements of terrestrial water storage (TWS) and downward shortwave radiation (DSR). We show that the vegetation in this region varies from energy-limited to water-limited throughout the year. During the peak of the wet season (January-February), rainfall, discharge and TWS are negatively correlated with NDVI in February-April (r = −0.48 to −0.65; p < 0.05). In addition, DSR is positively correlated with NDVI (r = 0.47–0.54; p < 0.05), suggesting that the vegetation is mainly energy-limited during this period. Outside this period, these correlations are positive for rainfall, discharge and TWS (r = 0.55–0.88; p < 0.05), and negative for DSR (r = −0.47 to −0.54; p < 0.05), suggesting that vegetation depends mainly on water availability, particularly during the vegetation dry season (VDS; late June to late October). Accordantly, the total rainfall during the dry season explains around 80% of the VDS NDVI interannual variance. Considering the predominant land cover types, differences in the hydroclimate-NDVI relationship are observed. Evergreen forests (531,350 km2) remain energy-limited during the beginning of the dry season, but they become water-limited at the end of the VDS. In savannas and flooded savannas (162,850 km2), water dependence occurs months before the onset of the VDS. These differences are more evident during extreme drought years (2007, 2010, and 2011), where regional impacts on NDVI were stronger in savannas and flooded savannas (55% of the entire surface of savannas) than in evergreen forests (40%). A spatial analysis reveals that two specific areas do not show significant hydroclimatic-NDVI correlations during the dry season: (i) the eastern flank of the Andes, characterised by very wet conditions, therefore the vegetation is not water-limited, and (ii) recent deforested areas (~42,500 km2) that break the natural response in the hydroclimate-vegetation system. These findings are particularly relevant given the increasing rates of deforestation in this region.
Volume
3
Language
English
OCDE Knowledge area
Oceanografía, Hidrología, Recursos hídricos Investigación climática
Scopus EID
2-s2.0-85109367516
Source
Frontiers in Water
ISSN of the container
26249375
Sponsor(s)
We wish to thank the following agencies/organisations for providing access to data: The developers of GPM-IMERG product, the Climate Hazards Group Infrared Precipitation for providing CHIRPS data (information available at http://chg.geog.ucsb.edu/data/chirps/), the GRACE mission (available at http://grace.jpl.nasa.gov), MODIS (available at https://lpdaac.usgs.gov/), The CRU-JRA (https://catalogue.ceda.ac.uk/), the SNO-HYBAM observatory for providing river discharge data (available at: http://www.ore-hybam.org), and the European Space Agency (ESA) Climate Change Initiative (CCI) and the Copernicus Climate Change Service (C3S) for providing the Global Land Cover maps (https://maps.elie.ucl.ac.be/CCI/viewer/). Funding. This research has been supported by the French MOPGA program funded by Sorbonne Université, ANR and IRD (ref. MOPGA scholarship and ref. ANR-18-MPGA-0008, AMANECER Project). PA was supported by Universidad de Antioquia through the grant CODI PRG2017-16264 and Minciencias Program 1115-852-70719 (contract # 490-220).
Sources of information: Directorio de Producción Científica Scopus