The recent 27 April 2011 tornado outbreak spawned three long-track tornadoes that passed across heavily forested landscapes and mountainous regions. One tornado rated EF-4 passed over the western portion of the Great Smoky Mountains National Park, and two other tornadoes, rated EF-2 and EF-3, crossed the mountains of northern Georgia in the Chattahoochee National Forest. Given that this tornado outbreak caused severe economic and human losses to the nation, a prompt and thorough scientific investigation is imperative.
This project combines observed forest damage in complex terrain with simulated vortices in an effort to reconstruct the near-surface wind field during the passage of tornadoes. The objectives are to 1) retrospectively characterize surface wind fields for three strong, long-track tornadoes in heavily-forested landscapes, and 2) determine the aspects of rugged topography that may influence the low-level wind field of strong and relatively rare tornadoes in the mountains. These objectives will allow a test of the hypothesis that significant topographic variations influence the strength and behavior of tornadoes.
Intellectual merit: Because of the difficulties inherent in obtaining direct or remotely-sensed observations of tornadoes in the mountains, alternative approaches that retrospectively characterize the near-surface wind fields must be sought. This rare and notable event provides a unique and valuable opportunity to assess a tornadic wind field along the entire path of the tornado in a heavily forested and mountainous area. Each tornado track through a national forest creates a research opportunity consisting of tens of thousands of trees, each of which provides information on the ambient wind speed and direction when it fell. Combined with information on the physical surroundings, such as slope, topographic position, soil types, and surface elevation, these trees in a tremendous variety of surroundings provide a perfect scenario with which to implement a wind field reconstruction.
To date, no studies incorporate terrain into numerical models of low-level tornado dynamics. Thus, the importance of observational studies that characterize the near-surface tornadic wind field in complex topography remains paramount. Damage surveys relying on aerial photography, with confirmation by ground surveys in a portion of the study area, will reveal tree characteristics, location, and orientation data for use as input to a simulation of the interaction between an idealized vortex and trees.
Broader impacts: The broader benefits to society are plentiful. Results of the research may aid risk assessments by the insurance industry, emergency managers, and forest authorities. The analyses may highlight areas that are more vulnerable to tornadoes in the mountains and will thus shape future structural engineering and geographical placement guidelines for schools, storm shelters, or trailer parks. The tree damage model may reveal characteristics of trees that could eventually lead to improvements in vegetation damage indicators for the Enhanced Fujita scale. Lastly, the PIs expect to educate the general public through online and local news releases regarding the study and its results, with emphasis on the fact that strong tornadoes can and do occur in the mountains. This effort will enhance general awareness of the threat of tornadoes in areas where the public may not recognize the dangers.
In addition, the research will train both graduate and undergraduate students. The PIs and students will participate in various national conferences and workshops. The project will also promote interdisciplinary collaborations from areas such as meteorology, forestry, ecology, environmental engineering, geography, and social sciences.
This grant supported an investigation into forest damage patterns produced by tornadoes passing through heavily forested and rugged terrain. Work began 64 days after the 27 April 2011 tornado outbreak by collecting high-resolution aerial photographs along the entire length of two separate tornado tracks in the Great Smoky Mountains National Park and the Chattahoochee National Forest. The airplane made two passes along each track, giving a total composite image width of about 5000 feet over 56 miles of tornado damage with a pixel resolution of approximately eight inches. These high-resolution images show individual tree trunks, canopies, and root balls. An examination of this imagery allowed us to tag the geographic coordinates of nearly half a million standing and fallen trees. Of these, treefall orientation was labeled for more than 130,000 of the fallen trees. This unique dataset alone will likely provide a valuable source of observations for future studies of forest damage in complex terrain. At the same time, ground surveys of the damage provided ground-truth for the analysis of the aerial photographs, as well as samples of the species, trunk diameter, damage severity, and other characteristics of more than 2600 trees within the damage tracks. Overlaying the tree locations and orientations on detailed terrain maps allowed an examination of the influence of the terrain on the near-surface tornadic wind field. In one instance, for example, the tornado closely followed a winding streambed as it descended a mountain. In other locations, winds were constricted and forced to accelerate through valleys and gaps in the terrain. Observations show that damage increases on steep slopes facing an oncoming tornado. Conversely, the tornadoes also passed straight over steep mountains, into valleys, and over streams. Following a previously-established approach, we developed a model that simulates a tornado and creates virtual blowdowns in forested landscapes and then compares the simulated damage patterns to observed damage patterns determined through the aerial and ground surveys. Iterative refinement of the shape and strength of the simulated tornado continues until the observed and simulated damage patterns match sufficiently well. While this is a feasible approach in more modest terrain, the simple model is unable to capture the complex flow field driven by the topography in these two tornado tracks. An alternative statistical approach to estimating Enhanced Fujita (EF) Scale levels within the forest has shown promising results. The approach uses estimates of tree strength to determine the most likely wind speed required to blow down certain species of trees. Then an assessment of the percentage of fallen trees in a small grid cell, as labeled in the aerial imagery, yields a map of EF-Scale intensity along the entire tornado track. Work continues on this novel approach. Several subprojects grew out of the present work. These efforts contribute to the ecological understanding of tornado disturbance patch size and spatial patterns, as well as to the role of major wind disturbances within forests on regional carbon cycling. An effort to revise and refine the EF Scale will also benefit from the analyses and data from this project. Overall, the grant promoted learning by ten undergraduate students, allowing them to develop field expertise and familiarity with remote sensing and GIS software. The grant also supported two graduate students for one semester and one summer, allowing them to participate in many aspects of the project. Local news releases have shared information about the study and the work has been presented at several national conferences.
