Publications

Venter, ZS; Scott, SL; Desmet, PG; Hoffman, MT (2020). Application of Landsat-derived vegetation trends over South Africa: Potential for monitoring land degradation and restoration. ECOLOGICAL INDICATORS, 113, 106206.

Abstract
Monitoring vegetation change is important because the nature, extent and rate of change in key measures, such as plant biomass, cover and species composition, provides critical insight into broader environmental and land use drivers and leads to the development of appropriate policy. We used Landsat data between 1984 and 2018 to produce a map of Enhanced Vegetation Index (EVI) change over South Africa at 30 m resolution and an interactive web application to make the analysis both globally applicable and locally meaningful. We found an increase in EVI of 0.37 +/- 0.59% yr(-1) (mean +/- standard deviation), confirming global vegetation greening trends observed with lower-resolution satellites. Mesic, productive biomes including the Albany Thicket and Savanna, exhibited the largest greening trends while browning trends were dominant in more arid biomes, such as the Succulent Karoo and Desert. Although overall EVI trends correspond to vegetation index trends derived from the Advanced Very-High-Resolution Radiometer (8 km resolution), the relative scarcity of Landsat data availability during the 1980 s is a potential source of error. Using repeat very-high-resolution satellite (ca. 3 m resolution) imagery and ground-based photography as reference, we found good correspondence with EVI trends, revealing patterns of degradation (e.g. woody plant encroachment, desertification), and restoration (e.g. increased rangeland productivity, alien clearing) over selected landscapes. The utility of the EVI trend layer to government and industry for monitoring ecosystem changes will be enhanced by the ability to distinguish climatic from anthropogenic drivers of change. This may be partially achieved though interactive exploration of the EVI trends using the application found here: http://evitrend.zsv.co.za

DOI:
10.1016/j.ecolind.2020.106206

ISSN:
1470-160X