Monitoring global land cover fractions and change from dense satellite time series

Summary

It is essential for land owners, governments and scientists across the world to know what kind of land cover is in their areas of interest: how much land is occupied by trees, shrubs, grass, crops, bare soil, water and built-up area. Even more useful is knowing how much of it is changing over time, and being able to compare it across countries or even parcels of land. My PhD thesis research focused on improving the ways to create and update global land cover maps. I explored new ways to map land cover fractions, to speed up a change detection algorithm so it can be applied to the whole globe, and to improve the consistency of multi-temporal land cover maps using artificial intelligence. Some of these improvements have already been taken up by leading global land cover map producers.