HARNESSING REMOTE SENSED AND CLOUD-BASED DATA TO IMPROVE THE MANAGEMENT OF CONSERVATION AND BIODIVERSITY IN INDONESIA

Indonesia is one of the most biodiverse countries on Earth. Conservation management in such an environment relies on an integration between technology and local knowledge and wide stakeholder participation. This research presents a novel approach to processing satellite remote sensing scenes of the Indonesian archipelago, to provide the longest ever record of landuse change as baseline data for conservation activities. 144,438 individual Landsat scenes covering 222 spatial tiles comprising the entire country were used to map spatial and temporal patterns in deforestation over the 43 years between 1972 and 2016. The processing stream comprised an innovative machine-learning algorithm utilising matrix completion and wavelet analysis to improve the annual resolution of imagery and to reduce the occurrence of cloud-a ected pixels. The analysis improves upon both the spatial and temporal resolution of space-based deforestation mapping and shows that total rate of Indonesian deforestation surpassed that of Brazil in 2009. Deforestation trends are also linked to a range of land-use changes, for example the conversion of forested land into agricultural areas. More than 150,000 km 2 of the country are now occupied by palm oil plantations, of which 30,000 km 2 are within forested areas designated as protected land. Furthermore, 42% of the remaining forested land in Indonesia are classi ed as degraded land, reducing the quality of the remnant forest cover with important impacts for sustainable land management and biodiversity. The outcomes of this novel deforestation and land-use mapping approach led to the development of an Information Technology framework for use by the Indonesian government to aid data collection in conservation management. This information collection complies with global metadata structures to contribute to conservation and biodiversity management. The framework employed incorporates cloud-based biodiversity data based on standard formats for use in further research. Finally, a mobile application was developed to assist forest rangers to collect data in remote areas. Preliminary usage showed that implementation of the mobile application helped to record fauna movement patterns and forest structure information. The mobile application and web GIS are well used by forest practitioners, with 66% of forest rangers using the approach. These results demonstrate the importance of remotesensing based mapping and of technology integration in biodiversity and conservation management e orts in Indonesia