Remote sensing of vegetation cover dynamics and resilience across southern Africa

A. Harris, A. S. Carr, J. Dash

Research output: Contribution to journalArticlepeer-review


Southern Africa supports a significant portion of the world's floral biodiversity but predicted changes in climate are likely to cause adverse impacts on the region's ecosystems and biodiversity. Knowledge regarding the resilience of vegetation cover is important for understanding the potential impact of anthropic or climatic change. The length of time vegetation cover takes to recover from disturbances can provide an indication of ecosystem resilience. We investigated spatial and temporal patterns in the persistence of vegetation cover across southern Africa (1982-2006) and used persistence probability plots to estimate decay times of NDVI trends as a means to characterise the potential resilience of key southern African biomes. Patterns of positive and negative NDVI trend persistence were spatially coherent, indicating collective dynamic behaviour of vegetation cover. Persistence probability plots indicated differences in resilience between biomes. Mean recovery times from negative NDVI trends were shorter than for positive trends in the Savanna and Nama Karoo, whereas the Succulent Karoo exhibited the shortest mean lifetime for positive NDVI trends and one of the longest mean lifetimes for negative trend survival, implying potentially slow recovery from environmental disturbance. The results show the potential of satellite-time series data for monitoring vegetation cover resilience in semi-arid regions. © 2013 Elsevier B.V.
Original languageEnglish
Pages (from-to)131-139
Number of pages8
JournalInternational Journal of Applied Earth Observation and Geoinformation
Issue number1
Publication statusPublished - 2014


  • Africa
  • Land degradation
  • NDVI
  • Remote sensing
  • Trend analysis


Dive into the research topics of 'Remote sensing of vegetation cover dynamics and resilience across southern Africa'. Together they form a unique fingerprint.

Cite this