Spacial dynamics of fires & forest pathogen outbreaks in Yellowstone National Park

Pathogen and pest outbreaks are recognized as key processes in the dynamics of Western forest ecosystems, yet the spatial patterns of stress and mortality are often complex and difficult to describe in an explicit spatial context, especially when considering the concurrent effects of multiple agents. Blister rust, a fungal pathogen, and mountain pine beetle, an insect pest, are two dominant sources of stress and mortality to high-altitude whitebark pine within the Greater Yellowstone Ecosystem (GYE). In whitebark pine populations infested with blister rust or mountain pine beetle, the shift from green to red needles at the outer-most branches is an early sign of stress and infestation. In this analysis, we investigated a method that combines field surveys with a remote sensing classification and spatial analysis to differentiate the effects of these two agents of stress and mortality within whitebark pine. Hyperspectral remotely sensed images from the airborne HyMap sensor were classified to determine the locations of stress and mortality in whitebark pine crowns through sub-pixel mixture-tuned matched-filter analysis in three areas of the GYE in September 2000 and July 2006. Differences in the spatial pattern of blister rust and mountain pine beetle infestation allowed us to separate areas dominated by mountain pine beetle versus blister rust by examining changes in the spatial scale of significant stress and mortality clusters computed by the Ripley's K algorithm. At two field sites the distance between clusters of whitebark pine stress and mortality decreased from 2000 to 2006, indicating domination by the patchy spatial pattern of blister rust infestation. At another site, the distance between significant stress and mortality clusters increased from 2000 to 2006, indicating that the contiguous pattern of mountain pine beetle infestation was the primary source of disturbance. Analysis of these spatial stress and mortality patterns derived from remote sensing yields insight to the relative importance of blister rust and mountain pine beetle dynamics in the landscape.
You can read "Landscape-scale patterns of forest pest and pathogen damage in the Greater Yellowstone Ecosystem" by Jaclyn Hatala et al and published in Remote Sensing of Environment, 2010, here.
Heather Lynch's graduate research broadly examined some of the large scale spatiotemporal patterns of forest ecosystems.  The three aspects of this resarch were:
1.  Spatiotemporal dynamics of insect-fire interactions, i.e. do insect outbreaks promote forest fires and do forest fires promote insect outbreaks? Heather focused on two study sites- British Columbia and Yellowstone National Park.
British Columbia:One of the major forest pests in British Columbia is the western spruce budworm (C. occidentalis).  Heather used spatial statistics to try to uncover any spatiotemporal clustering of fire events and insect damage using data drawn from the British Columbia Natural Disturbance Database.
Yellowstone National Park:There are four major insect pests in Yellowstone National Park (western spruce budworm, Douglas-fir beetle, mountain pine beetle, pine engraver).  One of Heather's projects was to answer the question Did the previous 20 year history of mountain pine beetle damage have a statistically significant effect on the course of the 1988 Yellowstone Fires?  This project was the recipient of the Howard T. Fisher Prize in GIS for 2005.
2.  Remote sensing of insect damage.
One of the major bottlenecks for understanding the large-scale patterns of insect damage is a lack of spatially explicit damage data.  It is particularly difficult to find datasets which cover large areas or multiple time periods.  Satellite imagery has the potential to cheaply and efficiently map large areas anywhere in the world, and to do so repreatedly and consistently.  Heather was particularly interrested in using the LANDSAT satellite, which takes an image of Yellowstone National Park every 16 days with 30m resolution.   However, the LANDSAT satellite may not have the resolution to detect the relatively small outbreaks in Yellowstone National Park, therefore Heather worked with an airborne hyperspectral sensor, known as Hymap, as well as a Japanese satellite system called ASTER.
3.  Mountain pine beetle population dynamics.
Mountain pine beetles represent a major forest disturbance phenomenon in the Rocky Mountain regions, and more recently, in British Columbia and Alberta, Canada.  Mountain pine beetles are interesting mathematically as they disperse by a combination of random movement and pheremone mediated aggregation.  Heather worked to develop a relatively simple model that would capture some of the major spatial characteristics evident in maps of historical damage she compiled for Yellowstone National Park.
You can read more about Heather's conclusions in her paper "The Influence of Previous Mountain Pine Beetle (Dendroctonus ponderosae) Activity on the 1988 Yellowstone Fires" published in the journal Ecosystems, 2006, here.