“Knowing where things are, and why, is essential to rational decision making”
— Jack Dangermond

The spatial pattern of disturbance events is commonly analyzed to estimate ecological effects.  However, we can go a step further by analyzing the spatial pattern of disturbance drivers to get landscape and regional estimates of the effects of disturbance regimes, meaning the cumulative effect of a disturbance-ecological system.  This is far more powerful, but requires considerable number crunching.  One example is our analysis of multiple disturbances on water resources across the lower 48.  A second would be the 20 degree of latitude analysis of an emergent, climate driven disturbance process not seen at broad scales before. 

A third is the role of wind, landslides, and cedar decline on regional carbon balance in southeast Alaska - an area the size of Florida that contains carbon stocks equivalent to ~8% of the lower 48 forests put together.

Landslide susceptibility if you consider wind disturbances and cedar decline as interacting factors (left), if you don't (middle), and the difference (right).  Scale goes from 0 (no chance) to 1 (100% probability).  A scaled up version of this analysis was used to map cumulative, regional exposure to landslides and wind and tie those results to regional carbon balance (below).

GIS Datasets

A variety of GIS and remote sensing datasets inform the work.  The USFS has provided excellent landslide maps which have enabled us to explore how better to predict their distribution on the landscape; new remote sensing resources have allowed us to analyze forest mortality and expansion across all of southeast Alaska (currently expanding down to northern California). 

Distribution of forest disturbances across southeast Alaska (anthropogenic disturbances removed) and a comparative distribution of their spatio-topographic contexts at a latitudinal scale.  The analysis is conducted at the 30m scale. From Buma and Barrett 2015.


Utilizing a variety of statistical analyses, we look at how disturbances are distributed across the landscape, how they interact, and how that can improve natural hazard forecasting, increase our knowledge of ecological processes, and help us monitor the effects of climate change.

Several variables interact in complex ways to shape the distribution of landslides on the Alaskan landscape.  From Buma and Johnson 2015.

Several variables interact in complex ways to shape the distribution of landslides on the Alaskan landscape.  From Buma and Johnson 2015.

We can go to a true "Big Data" approach by employing new technologies like Google Earth Engine, which allows us to look at broad scales and fine grains quickly and efficiently.  For example, together with colleagues at NC State and the USFS, we are exploring how disturbance processes change with scale across the continent and what that means for conservation and ecosystem monitoring in different ecoregions.  By combining information about the biotic communities and the pattern of disturbances, we can improve decision making processes regarding reserve design (how big is big enough?), management planning (what is the optimal size and variance in silvicultural treatments?) and climate change modeling (what's a baseline disturbance distribution?).

Differences in spatial pattern are apparent at multiple scales, from 30m pixel arrangements in a small area to big, landscape spanning shapes (>100 km2).  By analyzing them simultaneously, we can extract significant amounts of new information regarding ecosystem functioning.  By doing this simultaneously across the continent, a new understanding of the similarities and differences between different ecoregions emerges.