Project
Spatio-temporal approaches for satellite monitoring of global forest change dynamics
Humanity is highly depended on forests for a number of products (e.g. woods) and services (e.g. mitigation of global warming). However human activities such as deforestation and forest degradation lead to emission of greenhouse gases from forest thus contributing to global warming. Evidence exist that the warming of the globe so far leads to frequent occurrence of extreme climate conditions such as heat waves and droughts which weaken the capacity of the forest ecosystems to act as sinks for carbon. Despite these negative effects, tracking the effects of human activities and extreme climate conditions on forest in a systematic and timely manner remains challenging. It also remains highly challenging to better predict how forests are likely to respond to future climate forcing especially under continuous global warming. In this thesis our overall research goal is to develop an analysis framework for tracking in near real-time the forest change events in the pan-tropical regions of the globe using satellite time series data, and to better understand how these events affect vegetation activities in pan-tropics. To achieve this goal, first we will improve and optimise near real-time detection and monitoring of deforestation and forest degradation from satellite image time series by focusing on specific sites in pan-tropical regions.
This will result into a framework for detecting and monitoring deforestation and forest degradation in near real-time from satellite image time series. Next, we will develop an analysis framework for tracking in near real-time the effect of extreme climate conditions on vegetation activities in pan-tropics. Thereafter, we will integrate and calibrate the two frameworks developed above into one analysis framework for tracking forest dynamics resulting from deforestation, forest degradation, fire and extreme climate conditions, and subsequently up-scale this framework to the entire pan-tropical region. The up-scaling of this analysis framework will result into the time series data of disturbance events, thus allowing us to evaluate the impact of deforestation, forest degradation, fire and extreme climate conditions on vegetation activities. In addition, time series data on deforestation and forest degradation from our analysis can also be used as activity data for estimating GHG emissions from pan-tropical forests. We will then use time series data of disturbance events to assess the performance of terrestrial biosphere models in simulating the forest dynamics, focusing on the contribution of human-climate disturbance events to the biasness usually noted between observed- and modelled vegetation activities. In the end, we will use this information to understand how terrestrial biosphere models can be improved. Overall, information resulting from this research will be useful for development of a biosphere-atmosphere index.
This will result into a framework for detecting and monitoring deforestation and forest degradation in near real-time from satellite image time series. Next, we will develop an analysis framework for tracking in near real-time the effect of extreme climate conditions on vegetation activities in pan-tropics. Thereafter, we will integrate and calibrate the two frameworks developed above into one analysis framework for tracking forest dynamics resulting from deforestation, forest degradation, fire and extreme climate conditions, and subsequently up-scale this framework to the entire pan-tropical region. The up-scaling of this analysis framework will result into the time series data of disturbance events, thus allowing us to evaluate the impact of deforestation, forest degradation, fire and extreme climate conditions on vegetation activities. In addition, time series data on deforestation and forest degradation from our analysis can also be used as activity data for estimating GHG emissions from pan-tropical forests. We will then use time series data of disturbance events to assess the performance of terrestrial biosphere models in simulating the forest dynamics, focusing on the contribution of human-climate disturbance events to the biasness usually noted between observed- and modelled vegetation activities. In the end, we will use this information to understand how terrestrial biosphere models can be improved. Overall, information resulting from this research will be useful for development of a biosphere-atmosphere index.