Julia Hackländer (OpenGeoHub) presents "Mapping land potential / tracking land degradation using EO data".
Abstract:
While land is increasingly degrading, robust monitoring approaches are required to identify land degradation processes and to ultimately tackle those. Primary production is one of the key indicators for determining impacts of land degradation, which can be approximated by different parameters such as fPAR, the fraction of absorbed radiation, or more directly, gross primary productivity (GPP). Previous work has mapped primary productivity globally at coarse resolution, however, drivers behind land degradation, such as agriculture or grazing, might be missed unless mapped at fine-scale resolution. Hence, spatially continuous high resolution maps are needed for a more accurate monitoring of land degradation. Land degradation can be monitored by comparing the state not only to the immediate past and the surrounding, but also to its potential. This potential can be derived from data on pre-industrial times or from ecologically similar regions under nature protection. To identify degradation dynamics, in this research the gap between actual and potential primary productivity will be evaluated globally. Different approaches to model GPP from earth observation data will be evaluated by comparing empirical light-use efficiency models with machine learning and hybrid approaches. Based on primary productivity time series analyses, the current and past resilience of ecosystems to abrupt and gradual disturbances will be evaluated to identify the most fragile ones and the impact on biodiversity.
