Large Area Monitoring with a MODIS-based Disturbance Index (DI) Sensitive to Annual and Seasonal Variations

Disturbance of the vegetated land surface, due to factors such as fire, insect infestation, windthrow and
harvesting, is a fundamental driver of the composition forested landscapes with information on disturbance
providing critical insights into species composition, vegetation condition and structure. Long-term climate
variability is expected to lead to increases in both the magnitude and distribution of disturbances. As a
consequence it is important to develop monitoring systems to better understand these changes in the
terrestrial biosphere as well to inform managers about disturbance agents more typically captured through
specific monitoring programs (such as focused on insect, fire, or agricultural conditions). Changes in the
condition, composition and distribution pattern of vegetation can lead to changes in the spectral and thermal
signature of the land surface. Using a 6-year time series of MODerate-resolution Imaging Spectroradiometer
(MODIS) Land Surface Temperature (LST) and Enhanced Vegetation Index (EVI) data we apply a previously
proposed Disturbance Index (DI) which has been shown to be sensitive to both continuous and
discontinuous change. Using Canada as an example area, we demonstrate the capacity of this Disturbance
Index to monitor land dynamics over time. As expected, our results confirm a significant relationship
between area flagged as disturbed by the index and area burnt as estimated from other satellite sources
(R2=0.78, pb0.0001). The DI also demonstrates a sensitivity to capture and depict changes related to insect
infestations. Further, on a regional basis the DI produces change information matching measured wide-area
moisture conditions (i.e., drought) and corresponding agricultural outputs. These findings indicate that for
monitoring a large area, such as Canada, the time series based DI is a useful tool to aid in change detection
and national monitoring activities.