Using satellite remote sensing to detect forest degradation in the coastal forests of Tanzania

Abstract

The Coastal Forests of Tanzania are essential for biodiversity, as well as being a vital supply of ecosystem services to local communities. However, their proximity to large cities such as Dar Es Salaam and large ports on the Indian Ocean make them targets for larger scale deforestation for international markets. Much of the activity in these forests is poorly monitored and recorded, and much is illegal, which creates an unsustainable resource for the future. Tanzanian forest services are under-resourced and poor infrastructure impacts their ability to monitor such a large area. The ability to detect forest degradation from space to prioritise monitoring and conservation efforts would be beneficial in the preservation of Tanzanian Coastal forests. This paper uses three satellite remote sensing products: Sentinel 1 GRD, Sentinel 1 SLC and the ALOS Mosaic to identify a suitable method for detecting forest degradation in the region. Consideration was taken to make sure all data and methods were open source to ensure the continuity and accessibility of the project to Tanzanian counterparts. Radar remote sensing is available in all weather conditions and day and night which is suited to the often cloud covered Coastal Forests. All radar scenes were pre-processed, validated with a comprehensive field survey from 2016, ten forests in total were analysed. Overall the ALOS and the GRD methods are recommended due to the balance between accuracy and accessibility. SLC did not produce meaningful results.