Use of Landsat multi-temporal imagery to assess secondary growth Miombo woodlands in Luanshya, Zambia

Abstract

Miombo woodlands are the most extensive natural forest type in subtropical Africa, supporting the livelihood of over 100 million urban and rural settlers. These forests continue to experience extensive deforestation and land degradation due to land use land cover (LULC) changes. The aim of this study is to use remotely sensed images from the Landsat archive to provide the baseline for the spatial extent of Miombo woodlands in Luanshya district of the Copperbelt Province, Zambia. It also assesses the implications of spatial-temporal changes for the conservation of these woodlands. A hybrid classification method involving ISODATA and Support Vector Machine classifiers was used to generate LULC maps for 1986, 1998, 2001, 2004, 2011 and 2016, with an overall accuracy of 82.0%–99.6% and kappa coefficient of 0.80–0.99. Unlike previous studies that have mainly focused on comparison between two or three time periods, a longitudinal analysis over the 20-year periods of this study shows that there are complex net and gross changes in different LULC types over different scales. Landscape change metrics indicate that woodlands have been declining in extent, but are partly offset by regeneration. Rates of deforestation (regeneration) for the periods 1986–1998, 1998–2001, 2001–2004, 2004–2011 and 2011–2016 are 10% (1%), 38% (30%), 15% (7%), 11% (8%) and 9% (2%) per annum, respectively. These results suggest that over 68% of Miombo woodlands present in 2016 are second-growth forests. Restoration strategies such as assisted natural regeneration are necessary in order to accelerate the recovery of natural forests. Targeting under-used and degraded land for assisted natural regeneration and empowering local communities in sustainable environmental stewardship are crucial in remediating against continued forest degradation.

Keywords: conservation, Landsat, Miombo, resilience, secondary growth, woodlands