Landscape-scale patterns of wind disturbance
Jeff Cannon has been the lead hands-on person conducting the various steps of this research. We actually are trying to combine a study of landscape-scale disturbance patterns (e.g. patch size and spatial distribution, relationship to topography), with the additional goal of getting estimates of the amount of carbon transferred from living to decomposing. The latter is essential information for the national-scale estimates of carbon impact of tornadoes, described on one of the other Research pages. Our initial analysis concentrated on the north GA April 2011 tornado track, which is where we have three study sites. Subsequent analyses will apply the same steps to the Great Smoky Mountains National Park tornado track from April 2011.
In July 2011, we obtained high-resolution aerial photos of both tornado tracks. Using hundreds of training points in the images (points we manually classified by eye), Jeff conducted a supervised classification of the tornado damage swath. The air photo shown here is the westernmost 5 km of the north GA tornado track; the small colored dots are locations of our Boggs Creek site sampling plots.
Each pixel of the tornado track was classified into one of five severity levels, and then neighboring pixels were grouped into 20x20 m cells; the result of this classification is shown here, where severity increases from green to yellow to red.
Once the classification is completed for the entire tornado track, landscape-scale analyses can reveal the frequency distribution of patch sizes, the relative abundance of levels of severity, how severity of damage changes with topography or other spatial variables, etc. The analysis of the north GA track is now complete, and we are hard at work on the Smokies track.
Even before we have completed analysis of the second tornado track, several broad trends are apparent. First, tornado damage is extremely heterogeneous in terms of track width and severity. Obviously, tornadoes are not on the ground along their entire track length, so some areas have no damage. The width of the damage swath varies five-fold or more. Severity of damage follows a decreasing frequency distribution, with most areas having relatively light damage, and a small fraction experiencing the most severe damage. Damage patches fit a patterns described as a “dissolved bull’s-eye”. Severity was strongly influenced by topography, but the effect was complex and in places lower elevations experienced more damage than higher elevations.
As mentioned above, a simultaneous goal of this work was to get estimates of the amount of biomass and therefore carbon transferred from living to decomposing pools by the tornado. We have completed analyses only for the north GA tornado track, but we can use those findings to extrapolate out to other tornadoes, to the entire late April 2011 tornado outbreak, or even to total of all U.S. tornadoes in a given year. Because we know the average pre-storm standing biomass from our field survey plots (266,300 kg/ha), we can multiply that amount by severity to estimate the amount of biomass knocked down, and then multiply out how many hectares were affected at different levels of severity, to get an estimate of the total biomass knocked down by this tornado. Our findings suggest that 423,814 Mg or 0.424 Tg of biomass was downed by the north GA tornado. Making the simple assumption that roughly 50% of aboveground biomass is carbon, that suggests that 0.212 Tg of carbon was knocked down by the north GA April 2011 tornado.