Progress 10/01/10 to 09/30/13
Outputs
Target Audience: Forest resource managers, particularly working in the Coast Redwood region; remote sensing and fire ecology research communities. Changes/Problems:
Nothing Reported
What opportunities for training and professional development has the project provided?
Nothing Reported
How have the results been disseminated to communities of interest?
Nothing Reported
What do you plan to do during the next reporting period to accomplish the goals?
Nothing Reported
Impacts
What was accomplished under these goals?
Three models of tree mortality were developed, using: 1) imagery alone, 2) fusion of imagery with post-fire LiDAR data, and 3) fusion of imagery with differenced (pre- to post-fire) LiDAR data. Mortality was classified as low (0-25% of trees >10” DBH), moderate (25-50% of trees >10” DBH), or high (50-100% of trees >10” DBH). The accuracy of each model in estimating mortality class was evaluated using data from permanent plots over three years following the fire, to allow comparison of each method.
Publications
- Type:
Conference Papers and Presentations
Status:
Published
Year Published:
2013
Citation:
Bishop, B.D., Dietterick, B.C., and Mastin, T. 2014. Classification of fire-caused tree mortality in a redwood forest using digital orthophotography and LiDAR. Abstract B41F-04 presented at 2013 Fall Meeting, AGU, San Francisco, CA.
|
Progress 10/01/11 to 09/30/12
Outputs
OUTPUTS: A third year of post fire aerial imagery was delivered, but not corresponding GPS points which are needed for basic processing before this imagery can be utilized for modeling. Additional ground control registration is planned for Spring, which will entail surveying in existing control data plots to gain a greater degree of precision in their mapped locations. Imagery from LANDSAT, NAIP, and Quickbird were processed for NDVI/dNDVI/dNBR, and tasselled cap transformation, for the burn area. A variety of metrics including percentile heights (90th, 75th, 50th, 25th, 10th), skewness and kurtosis of height distribution, coefficient of variation of intensity, and Ripley's K (various distances) were generated from LiDAR points, either for the entire burn area (by grid cells) or for the set of permanent inventory plots which make up the response dataset for the project. All of the above have potential to contribute information about burn severity and mortality, and will need to be evaluated alone and in combination. PARTICIPANTS: Dietterick, Brian, Project Director, Cal Poly; Mastin, Tom, co-PD, Cal Poly; Bishop, Brian, graduate student, Cal Poly; Olsen, Chris, collaborator, Executive Director, Naval Postgraduate School, Monterey, CA; Two graduate students from the Naval POstgraduate School have utilized the post-fire LiDAR data to complete theses related to NPS objectives. TARGET AUDIENCES: Forest resource managers, particularly working in the Coast Redwood region; research and remote sensing communities PROJECT MODIFICATIONS: Nothing significant to report during this reporting period.
Impacts
Preliminary modeling of relationships between indices and metrics above and mortality has been done, suggesting areas which are more and less likely to yield significant results. Further refinement is needed, but it seems likely that fusion of LiDAR data with imagery will increase the accuracy of mortality modeling. Digital aerial imagery previously obtained was georeferenced using a variety of DEMs (LiDAR bare earth vs. LiDAR DSM vs. photo-derived DSM) to determine which yielded the clearest, most consistent image from year to year. Through partnership with NPS, we were able to obtain an additional set of postfire imagery, from the Quickbird satellite. This will allow comparison between imagery obtained via different platforms and at different resolutions, to determine which is optimal for this type of analysis.
Publications
- No publications reported this period
|
Progress 01/01/11 to 12/31/11
Outputs
OUTPUTS: A description of the project and its objectives, methods and data collection was presented at the Redwood Science Symposium in June 2011. This presentation was directed to an audience of resource managers, researchers, land owners and interested members of the public, all with a focus on the Coast Redwood forest type. The session included other presentations on the topic of remote sensing as a tool for mapping and monitoring forest resources. A publication derived from this presentation is due to appear in the conference proceedings, to be published in early 2012. As a General Technical Report from the US Forest Service, the conference proceedings will be widely accessible and visible to both the research and forest management communities. PARTICIPANTS: Two graduate students (1 student x .5 time) in the Forestry Sciences M.S. degree program from Cal Poly State University are developing theses to complete the objectives of this project. One project director, co-investigator, and research assistant at Cal Poly are involved in project administration, analysis and oversight of the graduate students. A partnership with the Remote Sensing Group at the Naval Postgraduate School in Monterey, CA has developed through this project. Dr. Chris Olsen, Executive Director of the Remote Sensing Group has provided match funding and in-kind contributions to support this project. In turn, Cal Poly has provided the NPS with LiDAR data, and access to nearby research facilities in Santa Cruz, CA. This has enabled NPS graduate students to work on remote sensing projects with access to a nearby field location. Two graduate students from the Naval Postgraduate School have utilized the post-fire LiDAR data to complete theses related to NPS objectives. Richard C. Olsen, Remote Sensing Group Executive Director, Physics Dept, PH/Os, 833 Dyer Road, Naval Postgraduate School, Monterey, CA 93943, olsen@nps.edu. TARGET AUDIENCES: Forest resource managers, particularly working in the Coast Redwood region are the primary target audience. Findings regarding the mortality observed from this example, and the utility of using various remote sensing approaches is the information that will be most useful to that audience. Additionally the research and remote sensing communities will be interested as well, regarding the methods, data and their utility as determined in this project. PROJECT MODIFICATIONS: Nothing significant to report during this reporting period.
Impacts
Airborne LiDAR data was collected during leaf-off conditions six months following the fire. This data was delivered by the vendor and subsequently processed into a number of derived layers. Post-fire layers derived from the LiDAR included improved bare-earth DEM, updated canopy height model, updated canopy cover layer and a new LiDAR intensity image. These layers are being used to evaluate tree and forest stand conditions that existed following the fire. In addition, aerial color and infrared imagery was acquired at one and two years following the Fire. The imagery complements the structural height information from the LiDAR and provides the spectral characteristics of trees and stands following the fire. Field-based forest inventory data has been collected in years one and two following the fire to provide a ground-truth evaluation of tree mortality at 83 plots within the study area. Plot data has been entered into a forest inventory database to be used in the development of statistical models that will relate remote sensing measurements to the observed tree mortality. Additional field surveying was conducted to evaluate the vertical accuracy of the bare-earth DEM produced following the fire. This field survey data will be used to quantify the improvements in accuracy that can be achieved when near-ground vegetation is removed prior to an airborne LiDAR flight.
Publications
- White, R. and B. Dietterick, 2012. Use of LiDAR and multispectral imagery to determine conifer mortality and burn severity following the Lockheed Fire. Proceedings: 2011 Redwood Science Symposium, USDA Forest Service Pacific Southwest Research Station, Santa Cruz, California. (in press).
|
|