A Satellite Based Remote Sensing Method for Monitoring Insect Disturbance in Forests
Olsson, Per-Ola; Jönsson, Anna Maria; Eklundh, Lars
Lund University, SWEDEN

The northern forests act as a carbon sink and it is estimated that global warming and elevated CO2 concentrations will enhance this capacity to store carbon. However, a changing climate is also likely to increases the frequency and severity of forest disturbance events resulting in reduced capacity of the forests to take up carbon; if disturbances increase more than forest growth there is a risk that forests will turn into carbon sources rather than sinks. Hence, it is important to develop efficient methods that enable early detection of forest damage, and to estimate the extent of these forest disturbances and the impact they have on the carbon cycle. One cause of forest disturbance is insect outbreaks. Insects' response to a changing climate is an area of research with insufficient knowledge. This knowledge gap makes future prediction of insect attacks difficult and the importance of efficient monitoring systems crucial.
Several studies have concluded that satellite based remote sensing can be used to detect insect damage in forests with high accuracy and various change detection techniques have been tested. This study suggests methodology based on time-series analyses to monitor insect induced forest disturbance. It is based on defining a "normal" seasonal trajectory of vegetation indices derived from MODIS satellite imagery. Based on this normal curve a monitoring method is established to identify seasonal deviations and classify them as disturbances depending on their size in relation to natural inter-annual variability. Important factors determining the reliability of the method are that the damaged area is large enough to enable detection and that there are data available from years with no disturbance to establish the baseline conditions. The method has been implemented for detection of damage on spruce forests in southern Sweden, pine forest in Finland and birch forest in northern Sweden. The method detects defoliation and maps the extent of the damaged areas, and as a further development the effect of insect damage on the carbon balance will be estimated.