EVALUATING THE USE OF LIGHT DETECTION AND RANGING (LIDAR) INFORMATION TO IMPROVE FOREST MANAGEMENT DECISIONS

• Sponsoring Institution: National Institute of Food and Agriculture
• Project Status: COMPLETE
• Funding Source: MCINTIRE-STENNIS
• Reporting Frequency: Annual
• Accession No.: 1001477
• Project Start Date: Nov 7, 2013
• Project End Date: Sep 30, 2018
• Program Code: [(N/A)]- (N/A)

Recipient Organization
UNIVERSITY OF KENTUCKY
500 S LIMESTONE 109 KINKEAD HALL
LEXINGTON,KY 40526-0001

Performing Department
Forestry & Natural Resources

Non Technical Summary
Forest managers and decision makers are promoting a more active approach to manage forests in Kentucky to increase economic, social, and environmental benefits from these forests in a more sustainable manner. Active management requires rigorous forest inventory and monitoring practices to evaluate how vegetation treatments can better achieve desired management goals. Typically, vegetation treatments are derived from average stand attributes and often applied at the landscape-scale containing several stands with varying conditions. Treating stands with different site potential, species distribution, and vegetation structures under similar treatment guidelines is often inefficient for achieving management goals for the entire landscape. Forest managers and decision makers need efficient ways to accurately estimate stand-level forest attributes to enable them to develop site-specific vegetation treatments for individual stands based on current vegetation structures. This project will evaluate the potential of LiDAR-derived vegetation and terrain information to accurately quantify forest resources and represent surface terrain characteristics for their subsequent use in inventory practices and planning of forestry practices. Automated procedures to obtain tree-level attributes that, when aggregated, can provide accurate stand-level forest information will be developed. Similarly, based on detailed terrain information derived from LiDAR data, computerized procedures will be developed to help design economically efficient forest treatment operations. For demonstration purposes, the developed procedures will be applied to a part of the University of Kentucky's Robinson Forest. However, developed procedures and research findings will be applicable to federal and state agencies involved with research and management of forest and natural resources as well as timber companies, contractors, and private landowners conducting forest treatments.

Animal Health Component

50%

Research Effort Categories

Basic

50%

Applied

50%

Developmental

(N/A)

Classification

Knowledge Area (KA)	Subject of Investigation (SOI)	Field of Science (FOS)	Percent
123	0621	3100	30%
123	0621	2090	35%
123	0621	2080	35%

Knowledge Area
123 - Management and Sustainability of Forest Resources;

Subject Of Investigation
0621 - Broadleaf forests of the South;

Field Of Science
2090 - Statistics, econometrics, and biometrics; 2080 - Mathematics and computer sciences; 3100 - Management;

Keywords

tree-level forest management

forest operations planning

individual tree growth model

Goals / Objectives
There are two main Goals Areas: Goal Area 1. Evaluate the potential of LiDAR-derived vegetation data to improve the quantification of forest resources for use in management, monitoring, and assessment activities. In particular, I will evaluate the ability of LiDAR data to predict tree-level attributes and characterize forest stand structures. · Objective 1. Develop a tree-detection algorithm to identify individual tree locations customized for oak-hickory dominated forest stands commonly found in Kentucky and the Appalachian region. · Objective 2. Develop allometric relationships to estimate derivative tree attributes (i.e. DBH and volume). · Objective 3. Develop individual-tree, distance-dependent growth models for common commercial tree species in Kentucky and the Appalachian region. · Objective 4. Aggregate LiDAR-derived tree-level data to develop auxiliary information for improving the estimates of stand-level attributes. Goal Area 2. Evaluate the potential of LiDAR-derived terrain data to accurately describe surface terrain characteristics for use in the planning of ground-based forest treatment operations. · Objective 5. Evaluate the quality of high-resolution LiDAR-derived DEM delivered by a vendor to represent terrain conditions commonly found in Kentucky. · Objective 6. Develop computerized procedures to delineate areas accessible to ground-based timber harvesting systems. · Objective 7. Develop computerized models to determine the optimal layout design of ground-based harvesting operations including landings, spur-roads, and skid-trails.

Project Methods
The main input for this project are two set of LiDAR information acquired over Robinson Forest (RF) during the summer of 2013; a low-density (1 pt/m2) dataset collected during leaf-off season for the purpose of obtaining accurate terrain representations and a high-density (25 pt/m2) dataset collected during leaf-on season to obtain high-quality vegetation information. Both leaf-on and leaf-off LiDAR datasets consist of raw point clouds and four raster layers: intensity layer, digital surface model (DSM), digital elevation model (DEM), and crown height model (CHM). To complete Goal Area 1, a grid of 271 permanent plots established throughout the main tract of RF between 1998 and 2000 will be used to ground-truth LiDAR-derived vegetation data. Inside each permanent plot, tree attributes including species, DBH, merchantable height to 9" top diameter outside bark (DOB) and to 5" top DOB, tree grade, crown class, tree status, and tree class were collected for all trees with DBH ≥ 5 inches. Starting this year, permanent plots will be re-measured and, in addition to the previously collected tree attributes, total height, height to base of live crown, crown ratio, and crown width as well as distance and azimuth from plot center to each tree with DBH ≥ 5 inches will also be collected. After identifying the location of permanent plots on the LiDAR-derived data, a plot radius of 11.35 meters will be used to select only LiDAR-derived data within the permanent plot areas. These selected LiDAR-derived data within each permanent plot will serve as the basis for ground truthing to evaluate the potential of LiDAR-derived vegetation data in order to retrieve tree-level attributes and quantify forest resources. To complete Goal Area 2, 45 terrain plots will be established to determine the ability of LiDAR-derived DEM to accurately represent surface terrain in different slope and ruggedness conditions. DEMs obtained from the USGS archive will be used to calculate slope and terrain ruggedness indices and thus capture terrain heterogeneity at a courser scale than 1-meter LiDAR-derived DEM. Slope values for each 30-meter grid cell will be used to group grid cells into three categories: low (≤30%), moderately (30-60%), and steeply sloped areas (≥60%). Similarly, terrain raggedness index values for each grid cell, calculated as described by Riley et al. (1999), will be used to group grid cells into: level (≤100), moderately (100-400), and highly (≥400) rugged areas. Using a random sampling procedure, 5 terrain plots will be established in each slope/ruggedness category. In each terrain plot, terrain elevation relative to the plot center will be measured every 3 meters along eight transects at 45? intervals spanning 10 meters outwards from the plot center. The mean error, mean absolute error, and RMSE of the elevation calculated for each terrain plot will serve as the initial means of evaluating the quality of the LiDAR-derived elevation data. For more details on the methods used to accomplish each individual objective under each goal area, please refer to the proposal attached.

Progress 11/07/13 to 09/30/18

Outputs
Target Audience:The target audience reached throughout the duration of the project included researchers, forestry practitioners, and decision makers from the private sector as well as from state and federal agencies. In particular, professionals directly involved with forest measurements and forest operations, the two main goal areas of this project. In 2016, research findings from this project were presented at the 38th annual meeting of the Council on Forest Engineering hosted by the University of Kentucky Department of Forestry. The meeting gathered over 80 scientists, practitioners and state and federal agencies from the US as well as researchers from other five countries.Stemming from this project, we also developed a web-based application to calculate and display transportation costs from anywhere in Kentucky to the closes mill destination. This application, named Kentucky Forest Transportation Cost Estimator is available at http://www2.ca.uky.edu/forestry/KY-FTCE/ and is currently being incorporated into the programmatic educational and training activities conducted the Forestry Extension in our department, which reaches professionals, stakeholders and decision makers in all 120 counties in the state of Kentucky. Lastly, the research findings from this project have been made available to the forest management research community via seven peer-reviewed journals, two book chapters, four conference papers, nine presentations, three software programs, two MS thesis, one PhD dissertation, and three peer-reviewed manuscripts still under review. ? Changes/Problems: Nothing Reported What opportunities for training and professional development has the project provided?In terms of training, two MS students and one PhD students graduate with the support of this project. They acquired problem solving skills and received specialized training in forest mensuration and forest operations. Also a web-based application, the Kentucky Forest Transportation Cost Estimator (http://www2.ca.uky.edu/forestry/KY-FTCE/), developed with the support of this project is currently being incorporated into the programmatic educational and training activities conducted the Forestry Extension in our department, which reaches professionals, stakeholders and decision makers in all 120 counties in the state of Kentucky. The project director was also able to attend three national conferences to present research findings from the project.? How have the results been disseminated to communities of interest?Research findings from this project have been made available to the forest management research community via seven peer-reviewed journals, two book chapters, four conference papers, nine presentations, three software programs, two MS thesis, one PhD dissertation, and three peer-reviewed manuscripts still under review. ? What do you plan to do during the next reporting period to accomplish the goals? Nothing Reported

Impacts
What was accomplished under these goals? Active management requires rigorous forest inventory and monitoring practices to evaluate how vegetation treatments can better achieve desired management goals. Quantifying forest attributes (i.e., trees, basal area, and volume) is traditionally conducted though field sampling, which is time consuming and expensive. Often resulting attributes have large estimation errors due to reduced sampling intensity and large vegetation variation. An important contribution of this project was a significant step forward towards the remote quantification of forest attributes. We developed cutting-edge, automated tree detection procedures applicable to forests with complex vegetation structures, such as the eastern deciduous forests of the Cumberland plateau. The potential benefit of using automated procedures, such as the one developed here, can impact all professionals involved in forest management as quantifying forest attributes is the first essential tasks before managing, monitoring, and conserving forests and natural resources. The accuracies of these procedures were not only evaluated using hundreds of permanent plots, but also applied to identify tree locations over large area covering tens of thousands of acres. Another important contribution of this project is the development of computerized procedures to optimize the efficiency of ground-based timber harvesting operations. Traditionally, the location of log landings, access rounds, and skid-trails is determined manually prior harvesting or by equipment operators at the time while harvesting. The developed procedure determines the optimal location of log-landings, access roads, and skid-trails that minimize harvesting costs (skidding, road and landing construction, and transportation) while also reducing soil disturbances caused by the traffic of heavy harvesting equipment. These optimization procedures have the potential to provide efficacy gains to all those involved in timber harvesting ranging from private and public landowners, logging firms and state and federal agencies involved in timber harvesting operations.

Publications

• Type: Conference Papers and Presentations Status: Published Year Published: 2015 Citation: Staats, W., Contreras, M. 2015. Elevation error of LiDAR-derived DEM in the complex terrain and vegetation condition of eastern deciduous forests. Presented at the Council of Forest Engineering (COFE) 38th Annual Meeting. July 19-22, 2015, Lexington, KY, USA.
• Type: Journal Articles Status: Published Year Published: 2016 Citation: Contreras, M., Parrott, D., Chung, W. 2016. Designing skid-trail networks to reduce skidding cost and soil disturbance for ground-based timber harvesting operations. Forest Science 62(1):48-58.
• Type: Journal Articles Status: Published Year Published: 2016 Citation: Hamraz, H., Contreras, M., Zhang, J. 2016. A robust approach for tree segmentation in deciduous forests using small-footprint airborne LiDAR data. International Journal of Applied Earth Observation and Geoinformation 52:532-541.
• Type: Conference Papers and Presentations Status: Published Year Published: 2014 Citation: Parrot, D.L., Contreras, M. 2014. Comparison of operator-designed and computer-generated skid-trail networks. In Proceedings: Global Harvesting Technology of the Council of Forest Engineering (COFE). 37th Annual Meeting. Moline, IL, USA.
• Type: Conference Papers and Presentations Status: Published Year Published: 2014 Citation: Staats, W., Contreras, M. 2014. LiDAR-derived forest canopy metrics and their influence on spatial distribution of plethondotid salamander populations. Presented at the 24th IUFRO World Congress and SAF national convention, October 5  11, 2014, Salt Lake City, USA.
• Type: Journal Articles Status: Published Year Published: 2017 Citation: Contreras, M., Freitas, R., Ribeiro, L., Stringer, J.W., Clark, C. 2017. Multi-camera surveillance systems for time and motion studies of timber harvesting equipment. Computers and Electronics in Agriculture 135:208-215.
• Type: Conference Papers and Presentations Status: Published Year Published: 2014 Citation: Contreras, M., Parrott, D., Stringer, J.W. 2014 Quantifying potential benefits of implementing computer generated skid trail networks. Presented at the 24th IUFRO World Congress and SAF national convention, October 5  11, 2014, Salt Lake City, USA.
• Type: Journal Articles Status: Published Year Published: 2017 Citation: Hamraz, H., Contreras, M., Zhang, J. 2017. A scalable approach for tree segmentation within small-footprint airborne. Computers and Geosciences 102:139-174.
• Type: Journal Articles Status: Published Year Published: 2017 Citation: Contreras, M., Staats, W., Yang, J., Parrott, D. 2017. Quantifying the accuracy of LiDAR-derived DEM in deciduous eastern forests of the Cumberland Plateau. Journal of Geographic Information System 9:339-353.
• Type: Journal Articles Status: Published Year Published: 2017 Citation: Hamraz, H., Contreras, M., Zhang, J. 2017. Vertical stratification of forest canopy for segmentation of under-story trees within small-footprint airborne LiDAR point clouds. ISPRS Journal of Photogrammetry and Remote Sensing 130:385-392.
• Type: Journal Articles Status: Published Year Published: 2017 Citation: Hamraz, H., Contreras, M., Zhang, J. 2017. Forest understory trees revealed within sufficiently dense airborne laser scanning point clouds. Scientific Reports 7:6770.
• Type: Book Chapters Status: Published Year Published: 2018 Citation: Contreras, M., Hamraz, H., Zhang, J. 2018. Uso de tecnolog�a de datos LiDAR en inventarios forestales remotos (Use of LiDAR technology in remote forest inventories) In Geom�tica Aplicada, Mena, C., Ormaz�bal, Y., Barrientos, V. (Eds). Aerophotogrammetry Service  Chilean Air Force & University of Talca  Chile. pp 139-148. Available from: http://editorial.utalca.cl/docs/ebook/geomatica.pdf
• Type: Book Chapters Status: Published Year Published: 2018 Citation: Hamraz, H., Contreras, M.A. 2018. Remote sensing of forests using discrete return airborne LiDAR. Recent Advances and Applications in Remote Sensing. Ming Hung (Ed), IntechOpen, DOI: 10.5772/intechopen.71777. Available from: https://www.intechopen.com/books/recent-advances-and-applications-in-remote-sensing/remote-sensing-of-forests-using-discrete-return-airborne-lidar
• Type: Journal Articles Status: Under Review Year Published: 2019 Citation: Hamraz, H., Jacobs, N., Contreras, M., Clark, C. 2018. Deep learning for conifer/deciduous classification of airborne LiDAR 3D point clouds representing individual trees. Remote Sensing of Environment
• Type: Websites Status: Published Year Published: 2018 Citation: Kentucky Forest Transportation Cost Estimator Available at: http://www2.ca.uky.edu/forestry/KY-FTCE/
• Type: Other Status: Other Year Published: 2015 Citation: LiDAR Data Format Converter - free software to convert LiDAR data from LAS to ASCII format. Avalilable for download at: http://www2.ca.uky.edu/forestry/FMRL_Website/SoftWare.html
• Type: Other Status: Other Year Published: 2015 Citation: Forest Road Viewer - free software to estimate earthwork movement for proposed, low-volume forest roads, and visualize proposed road in an GIS environment
• Type: Journal Articles Status: Under Review Year Published: 2019 Citation: Contreras, M., Stringer, J.W., Parrott, D. 2019. Retroactive comparison of operator-designed and computer-generated skid-trail networks on steep terrain. Croatian Journal of Forest Engineering
• Type: Journal Articles Status: Submitted Year Published: 2019 Citation: Contreras, M. 2019. Improving the accuracy of stand attributes through the use of small footprint airborne LiDAR in deciduous forests of Eastern Kentucky. Cnadian Journal of Forest Research
• Type: Theses/Dissertations Status: Published Year Published: 2015 Citation: Wesley Staats, 2015. Use of LiDAR-derived terrain and vegetation information in a deciduous forest in Kentucky. MS Forestry, University of Kentucky.
• Type: Theses/Dissertations Status: Published Year Published: 2018 Citation: Hamid Hamraz. 2018. Automated tree-level forest quantification using airborne LiDAR. PhD Dissertation, PhD Computer Science, University of Kentucky.
• Type: Theses/Dissertations Status: Under Review Year Published: 2019 Citation: Rafael Freitas. 2019. Use of camera surveillance systems to conduct motion studies and develop terrain dependent cycle time equations. MS Forestry, University of Kentucky.

Progress 10/01/16 to 09/30/17

Outputs
Target Audience: Nothing Reported 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?During the reporting period, results have been disseminated in the form of three peer-reviewed journals. What do you plan to do during the next reporting period to accomplish the goals?During the next reporting period I plan to finish all ramaining analyses and finish the reprepatation of manuscript describing the methods used to achieve the remaing objective.

Impacts
What was accomplished under these goals? During this reporting period, we have published two journal publications under Goal Area 1 and another journal publication under Goal Area 2. Progress towards completing objectives 2-4 are still underway and two manuscripts describing the details are in preparation. Similarly, Objective 6 has been completed and the manuscript addressing objective 7 is under review. ?

Publications

• Type: Journal Articles Status: Published Year Published: 2017 Citation: Hamraz, H., Contreras, M., Zhang, J. 2017. Forest understory trees revealed within sufficiently dense airborne laser scanning point clouds. Scientific Reports 7:6770.
• Type: Journal Articles Status: Published Year Published: 2017 Citation: Hamraz, H., Contreras, M., Zhang, J. 2017. Vertical stratification of forest canopy for segmentation of under-story trees within small-footprint airborne LiDAR point clouds. ISPRS Journal of Photogrammetry and Remote Sensing 130:385-392.
• Type: Journal Articles Status: Published Year Published: 2017 Citation: Contreras, M., Staats, W., Yang, J., Parrott, D. 2017. Quantifying the accuracy of LiDAR-derived DEM in deciduous eastern forests of the Cumberland Plateau. Journal of Geographic Information System 9:339-353. (IF = 0.98)

Progress 10/01/15 to 09/30/16

Outputs
Target Audience: Nothing Reported 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?During the reporting period, results have been disseminated mainly in the form ofpeer-reviewed journals. What do you plan to do during the next reporting period to accomplish the goals?During the next reporting periodI plan to finishall analyses using LiDAR-derived and field-collected tree data required to complete Objectives 2 and 3. I expect manuscripts describing methods related to Objectives 5 and & will also be published. Finally, four additional publications (2 journal publications and 2 book chapters) related to Objectives2 will be completed.

Impacts
What was accomplished under these goals? During this reporting period, Objective 1 has been completed and two journal publications describe the method and the application of the developed tree-detection algorithm. Preliminary analyses of the LiDAR-derived tree data and field-collected tree data have also been finished to address Objectives 2 and 3. Progress towards completing Objective 4 is still underway. Objective 5 has been completed and the manuscript we prepared summarizing the results is still in the review process. Progress towards completing Objective 6is still underway. Objective 7has also been completed and a manuscript is in preparation which describes the methods.

Publications

• Type: Journal Articles Status: Published Year Published: 2016 Citation: Hamraz, H., Contreras, M., Zhang, J. 2016. A robust approach for tree segmentation in deciduous forests using small-footprint airborne LiDAR data. International Journal of Applied Earth Observation and Geoinformation 52:532-541.
• Type: Journal Articles Status: Published Year Published: 2016 Citation: Hamraz, H., Contreras, M., Zhang, J. 2016. A scalable approach for tree segmentation within small-footprint airborne. Computers and Geosciences 102:139-174.

Progress 10/01/14 to 09/30/15

Outputs
Target Audience:The target audience reached during this reporting period include the 80 perticipant of the 38th annual meeting of the Council on Forest Engineering (COFE) hosted by the University of Kentucky Department of Forestry during July 19-22, 2016. This meeting gathered scientists, practicioners, and state and federal agenciesfromthe US as well asresearcher from other 5 countries. 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?During the reporting period, results have been disseminated in the form of conference papers and presentations as well as in peer-reviewed journals What do you plan to do during the next reporting period to accomplish the goals? Nothing Reported

Impacts
What was accomplished under these goals? During this reporting period, important process has been made to accomplish the objectives stated in this project. We are in the last stages of testing thetree-detection algorithm under Objective 1 and preparing a manuscript describing the methods. Objective 5 has been completed and a manuscript summarizing the results is under review, Objective 7 is also partially completed: a model has been develop to design ski-trail networks that simultaneously minimi skidding costs and soil disturbances. Progress towards the remaining objective is under way.

Publications

• Type: Journal Articles Status: Accepted Year Published: 2015 Citation: Contreras, M., Parrott, D., Chung, W. 2015. Designing skid-trail networks to reduce skidding cost and soil disturbance for ground-based timber harvesting operations. Forest Science. http://dx.doi.org/10.5849/forsci.14-146
• Type: Conference Papers and Presentations Status: Published Year Published: 2015 Citation: Staats, W., Contreras, M. 2015. Elevation error of LiDAR-derived DEM in the complex terrain and vegetation condition of eastern deciduous forests. Presented at the Council of Forest Engineering (COFE) 38th Annual Meeting. July 19-22, 2015, Lexington, KY, USA.
• Type: Journal Articles Status: Under Review Year Published: 2015 Citation: Staats, W., Contreras, M., Parrott, D. 2015. Elevation error of LiDAR-derived DEM in the deciduous eastern forests on the Cumberland Plateau. In Photogrammetric Engineering and Remote Sensing

Progress 11/07/13 to 09/30/14

Outputs
Target Audience: Nothing Reported 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? Preliminary results have been disseminated in the form of three conference papers and a peer-reviewed publication currently under review in Forest Science. Future dissemination effort will also include non-technical magazines, conference proceedings, and through information posted in the "Forest Management Research Lab" at the Department of Forestry, University of Kentucky (www.ca.uky.edu/forestry/fmrl_website) What do you plan to do during the next reporting period to accomplish the goals? Nothing Reported

Impacts
What was accomplished under these goals? This project is evaluating the potential of LiDAR-derived vegetation and terrain information to accurately quantify forest resources and represent surface terrain characteristics for their subsequent use in inventory practices and planning of forestry practices. We are developing automated procedures to obtain tree-level attributes that, when aggregated, can provide accurate stand-level forest information. Moreover, we are developing computerized procedures to help design economically efficient forest treatment operations, which are based on detailed terrain information derived from LiDAR data. For Goal Area 1 of this project, we have accomplished a major milestone. We completed data collection of 271 permanent plots across the study area (University of Kentucky's Robinson Forest). This data is crucial for ground validating LiDAR-derived tree locations and the completion of objectives 1 through 4. For Goal Area 2 of this project, we also completed field data collection on 45 terrain plots to evaluate the accuracy of LiDAR-derived terrain models. This is a basic step for the completion of Objective 5. We are currently analyzing the collected data and to laterincorporate results into required procedures to complete the remaining objectives of both Goals Areas.

Publications

• Type: Journal Articles Status: Under Review Year Published: 2014 Citation: Contreras, M., Parrott, D., Chung, W. 2014. Designing skid-trail networks to reduce skidding cost and soil disturbance for ground-based timber harvesting operations. In Forest Science.
• Type: Conference Papers and Presentations Status: Accepted Year Published: 2014 Citation: Parrot, D.L., Contreras, M. 2014. Comparison of operator-designed and computer-generated skid-trail networks. In Proceedings: Global Harvesting Technology of the Council of Forest Engineering (COFE). 37th Annual Meeting. Moline, IL, USA.
• Type: Conference Papers and Presentations Status: Accepted Year Published: 2014 Citation: Staats, W., Contreras, M. 2014. LiDAR-derived forest canopy metrics and their influence on spatial distribution of plethondotid salamander populations. Presented at the 24th IUFRO World Congress and SAF national convention, October 5  11, 2014, Salt Lake City, USA.
• Type: Conference Papers and Presentations Status: Accepted Year Published: 2014 Citation: Contreras, M., Parrott, D., Stringer, J.W. 2014 Quantifying potential benefits of implementing computer generated skid trail networks. Presented at the 24th IUFRO World Congress and SAF national convention, October 5  11, 2014, Salt Lake City, USA.