Progress 04/01/03 to 03/31/09
Outputs
OUTPUTS: Experiments were conducted to evaluate the sensitivity of forest productivity estimates to true versus masked plot locations from the USFS Forest Inventory and Analysis (FIA) program. Computer programs were written to compute metrics of interest from digital spatial data such as soil polygon boundaries. These metrics include measures such as trigonometric function of vertex angles that are expected to be correlated with uncertainty measures for those polygons. Research results were published and presented at conferences, including the Soil Science Society of America (SSSA) conference and Forest Inventory and Analysis (FIA) Science Symposium. Algorithms were developed to aid in the evaluation of spatial data uncertainty. These include: (1) a program to determine the median of a cluster of points (for example, boundary corners from a collection of spatial features), (2) a program to determine the median of a series of polygon boundaries, (3) a program to simulate random errors added to vertices in lines or polygons. Spatial analyses of forestry ordinances were conducted for four counties in Virginia, to demonstrate the ability of spatial analysis to predict potential impacts of ordinances on forest practices. An experiment was conducted to evaluate the effect of boundary error on area estimates using both photo-interpreted and GPS-obtained boundaries around a forest clearing. An evaluation of low-cost GIS tools for forestry applications was completed and published. A literature review was published to document approaches for validating and verifying the performance of models used in policy-relevant forest analyses. A spatial analysis of terrestrial carbon sequestration opportunities in Virginia has been completed. GIS analysis of soils, forest inventory, topography, hydrography and other layers was conducted to identify land areas where carbon sequestration possibilities such as afforestation and conversion of tillage practices would be possible; potential carbon storage impacts were estimated. Three approaches for modeling landslide risk were evaluated using two data sources for study areas in Virginia for which landslide inventories are available. Experiments were conducted to test the appropriateness of the commonly-applied Shi (2002) error band model for polygon boundary uncertainty. Spatial models of forest productivity were evaluated to determine levels of uncertainty based on spatial data inputs and algorithms. The impact of spatially fuzzed plot locations has been documented in a conference paper. Algorithms for determining the spatial median of a cluster of points and the spatial mean of a series of polygon boundaries has been developed and coded in GIS software. Our understanding of the interactions between broad forest structure changes and the resulting impacts on forest breeding bird populations has improved for several species for which habitat models were developed. New models of spatial uncertainty have changed our understanding of the way uncertain spatial data may perform in a variety of analyses; our ability to assess the reliability of spatial analysis results has therefore improved. PARTICIPANTS: Collaborating faculty: Philip Radtke, Carolyn Copenheaver, Tom Fox, Mike Mortimer, Randy Wynne, Valerie Thomas. Graduate Students: Claudia Cotton, Huei-Jin Wang, Brandon Martin, Todd Fearer, Jeffrey Galang, Lauren Stull, Mansi Grover, Aaron Bernard, Laura Johnson Undergraduate students: David Daversa. TARGET AUDIENCES: Target audiences for this effort include GIS practitioners involved in spatial analysis of natural resources data, forest scientists, forest managers, policymakers at local, state, regional, and national levels. PROJECT MODIFICATIONS: Not relevant to this project.
Impacts
Results from this research will aid forest managers in understanding the degree to which common spatial analyses are reliable. Unlike conventional portrayal of uncertainty of estimates (e.g., using confidence intervals), spatial analysts have rarely been able to make reliability statements about their products. Tools developed from this research will be employed to make statements about the likelihood of possible outcomes from spatial assessments. Improved awareness of low-cost, easy-to-use GIS packages should enable more smaller-scale operations (forestry consultants, landowners, small companies) make effective use of these technologies. Better access to these technologies should lead to more informed decisionmaking and more effective resource management. Improved understanding of the magnitude and characteristics of errors in GPS-derived data can help managers and planners assess the utility of results of spatial analyses. Because maps (and estimates derived from them, such as area) are essentially single samples of spatial data, conventional statistical methods for deriving confidence intervals cannot apply. The results described here are a step toward using other approaches to describe the confidence that users may have in map-derived data.
Publications
- Moan, J. E. 2008. Development and assessment of remotely derived variables in current southern pine beetle (Dendroctonus frontalis Zimm.) hazard mapping in North Carolina, USA. M.S. thesis. Department of Forestry, Virginia Polytechnic Institute and State University, Blacksburg, VA.
- Radtke, P.J., R.L. Amateis, S.P. Prisley, C.A. Copenheaver, D.C. Chojnacky, J.R. Pittman*, and H.E. Burkhart. 2009. Modeling production and decay of coarse woody debris in loblolly pine plantations. Forest Ecology and Management 257: 790-799.
- Cotton, C., S. P. Prisley, and T. R. Fox. 2008. Mapping upland hardwood site quality and productivity with GIS and FIA in the Blue Ridge of North Carolina. In: McWilliams, W., Moisen, G., Czaplewski, R., comps. 2008. Forest Inventory and Analysis (FIA) Symposium 2008; October 21-23, 2008; Park City, UT. Proc. RMRS-P-56CD. Fort Collins, CO: U.S. Department of Agriculture, Forest Service, Rocky Mountain Research Station. 1 CD.
- Prisley, S.P., H-J. Wang, P.J. Radtke, and J. Coulston. 2008. Combining FIA plot data with topographic variables: are precise locations needed In: McWilliams, W., Moisen, G., Czaplewski, R., comps. 2008. Forest Inventory and Analysis (FIA) Symposium 2008; October 21-23, 2008; Park City, UT. Proc. RMRS-P-56CD. Fort Collins, CO: U.S. Department of Agriculture, Forest Service, Rocky Mountain Research Station. 1 CD.
- Thomas, V., J. H. McCaughey, P. Treitz, D. A. Finch, T. Noland and L. Rich. 2009. Spatial modelling of photosynthesis for a boreal mixedwood forest by integrating micrometeorological, lidar and hyperspectral remote sensing data. Agriculture and Forest Meteorology 149:639-654.
- Thomas, V., T. Noland, J. H. McCaughey and P. Treitz. 2009. Lidar-hyperspectral analysis to examine leaf area index, clumping, and canopy biochemistry in a boreal mixedwood environment. Presented at the ASPRS 2009 Annual Conference; March 9-13, 2009; Baltimore, MD.
- Gokkaya, K., V. Thomas, T. Noland, R. Wynne, J. H. McCaughey, I. Morrison and P. M. Treitz. 2009. Prediction of foliar biochemistry using canopy reflectance and structure information. Presented at the Canadian Remote Sensing Symposium; June 22-25, 2009; Lethbridge, Alberta, Canada.
|
Progress 10/01/07 to 09/30/08
Outputs
OUTPUTS: As part of an effort to derive boundary uncertainty estimates, several algorithms were developed to aid in the evaluation of spatial data uncertainty. These include: (1) a program to determine the median of a cluster of points (for example, boundary corners from a collection of spatial features), (2) a program to determine the median of a series of polygon boundaries, (3) a program to simulate random errors added to vertices in lines or polygons. These algorithms have been coded in GIS software and have been used in several research applications. Manuscripts describing their operation are in preparation. The simulation model was used to test the appropriateness of the commonly applied Shi (2002) error band model for polygon boundary uncertainty. Monte Carlo simulation demonstrated that the variability of polygon boundaries when compared to a more accurate original source is not lower in the region between polygon vertices, as suggested by the Shi error model. A manuscript documenting this experiment is in preparation. Spatial models of forest productivity are being evaluated to determine levels of uncertainty based on spatial data inputs and algorithms. The Forest Site Quality Index (FSQI) model has been applied to a six-county area in western North Carolina and correlated with Forest Inventory and Analysis (FIA) data. The impact of spatially fuzzed plot locations has been documented in a conference paper. PARTICIPANTS: Prisley, Stephen P. Principal Investigator Collaborators: P. J. Radtke, H-J Wang, C. Cotton, E. P. Smith, K. Love, L. Carstensen, A. Foy, R. Folledo TARGET AUDIENCES: Target audiences for this research include GIS practitioners involved in spatial analysis of natural resource data. PROJECT MODIFICATIONS: Nothing significant to report during this reporting period.
Impacts
New models of spatial uncertainty have changed our understanding of the way uncertain spatial data may perform in a variety of analyses; our ability to assess the reliability of spatial analysis results has therefore improved. Our understanding of the interactions between broad forest structure changes and the resulting impacts on forest breeding bird populations has improved for several species for which habitat models were developed.
Publications
- No publications reported this period
|
Progress 10/01/06 to 09/30/07
Outputs
OUTPUTS: Activities- Data collection on forest boundary uncertainty was conducted, using 25 students who delineated a forest boundary from two types of aerial photographs and in the field using GPS equipment. This data collection was conducted as part of a graduate class exercise, thereby providing educational value for the effort. The locations of saw and grist mills in Giles County, Virginia, were analyzed with respect to topography, hydrography, and soils, and related to historical species distributions as recorded in property deeds. This effort stemmed from a forest ecology class and therefore involved an educational component as well. A manuscript has been published. Products- A new spatial uncertainty model based on Bayesian statistics was developed, with a manuscript submitted for publication. Habitat models for forest breeding birds were developed from spatial analysis of Breeding Bird Survey (BBS) and Forest Inventory and Analysis (FIA) data. Spatial combination of these
datasets has enabled the modeling of breeding bird abundance as a function of variables related to forest structure across a broad geographic region. A manuscript has been published and a monograph is in review summarizing this work.
PARTICIPANTS: Prisley, Stephen P. Principal investigator. Collaborators: Philip Radtke, Dean Stauffer, Patrick Keyser, Carolyn Copenheaver, Jeff Galang, Carl Zipper, John Galbraith, Eric Smith, Kimberly Love, Bill Carstensen. Training or professional development: 8 graduate students and 25 undergraduate students were involved in the project.
Impacts
Change in knowledge- New models of spatial uncertainty have changed our understanding of the way uncertain spatial data may perform in a variety of analyses; our ability to assess the reliability of spatial analysis results has therefore improved. Knowledge of the variability of forest boundary locations has been refined through analysis of empirical data. Boundaries delineated from color-infrared photography are more precise than from natural color photography, and both of these approaches yield more precise boundaries than those recorded by consumer-grade GPS units. Variability of mapped forest boundary lines (from 1m resolution photography) ranged from 10 to 16m. Our understanding of the interactions between broad forest structure changes and the resulting impacts on forest breeding bird populations has improved for several species for which habitat models were developed.
Publications
- Fearer, T., S.P. Prisley, D.F. Stauffer, and P.D. Keyser. 2007. A method for integrating the Breeding Bird Survey and Forest Inventory and Analysis databases to evaluate forest bird habitat relationships at multiple spatial scales. Forest Ecology and Management 243(1): 128-143.
- Copenheaver, C.A., S.P. Prisley, J.R. Pittman, M.E. Yonce, C.M.S. Issem, and K.A. Jensen. 2007. The geography of grist, flour, and saw mills: indicators of land-use history in Virginia. Southeastern Geographer 47(1): 138-154.
- Galang, J.S., C.E. Zipper, S.P. Prisley, J.M. Galbraith, and P.F. Donovan. 2007. Evaluating terrestrial carbon sequestration options for Virginia. Environmental Management 39(2): 139-150.
|
Progress 10/01/05 to 09/30/06
Outputs
The Virginia forestry ordinance analysis begun in the last year was completed this year and resulted in the production of two journal articles, a conference proceedings, a trade article, and a award-winning student poster. Effort was also directed at development of tools for the consideration of spatial uncertainty in spatial decision-making. We have collected data (through simulation and student testing) that is being used to refine a widely-used model for spatial uncertainty. Specifically, the Shi error-band model of spatial uncertainty defines zones of less uncertainty at increasing distance from digitized vertices. Preliminary analyses of the data collected suggests that this is unjustified, and we are collecting data to demonstrate that in the case of lines subjected to generalization or simplification, the Shi error-band model is not appropriate. This research has been aided by the completion of a software tool to evaluate a digitized line in a GIS database for
evidence and measurement of simplification.
Impacts
Results from this research will aid forest managers in understanding the degree to which common spatial analyses are reliable. Unlike conventional portrayal of uncertainty of estimates (e.g., using confidence intervals), spatial analysts have rarely been able to make reliability statements about their products. Tools developed from this research will be employed to make statements about the likelihood of possible outcomes from spatial assessments.
Publications
- Bernard, A. M. 2006. Geospatial modeling of forest road networks and their effect on stream macroinvertebrate communities. M.S. Thesis. Department of Forestry, Virginia Polytechnic Institute and State University, Blacksburg, VA.
- Prisley, S. P., D. R. Daversa and M. J. Mortimer. 2006. Estimation of forest area affected by local ordinances: a Virginia case study. Southern J. of Applied Forestry 30(4):188-195.
- Mortimer, M. J., L. Stull and S. Prisley. 2006. Forest-related ordinances in Virginia: A case study in regulatory de-evolution. Southern J. of Applied Forestry 30(4):196-205.
- Mortimer, M. J. and S. P. Prisley. 2006. Virginias local forestry-related ordinances: prospects for change? Virginia Forests 62(3):5-7.
- Grover, M., D. Bosch, and S. Prisley. 2005. Incentives for a forest landowner to sequester and trade carbon under uncertainty: impact of hurricanes in the southeastern United States. World Resource Review 18(1):115-141.
- Daversa, D. R., S. P. Prisley and M.J. Mortimer. 2006. Spatial analysis of potential impacts of local forestry ordinances. In: Prisley, S., P. Bettinger, I-K. Hung, and J. Kushla, eds. Proceedings of the 5th Southern Forestry and Natural Resources GIS Conference, June 12-14, 2006, Asheville, NC. Warnell School of Forestry and Natural Resources, University of Georgia, Athens, GA. (In press)
|
Progress 10/01/04 to 09/30/05
Outputs
Efforts during this period included continued data collection for evaluation of boundary error using both photo-interpreted and GPS-obtained boundaries around a forest clearing. The database now consists of 81 independent observations of the boundary; 26 from GPS and 55 from photo delineation. Preliminary analysis shows variation in the confidence band around the photo-derived boundary due to photo interpretation differences among observers. GPS boundaries had wider, but more consistent boundaries. An evaluation of low-cost GIS tools for forestry applications was completed and published. A spatial analysis of forest areas subject to local ordinances was conducted. Four study counties in Virginia were selected and ordinances restricting forestry operations within buffer zones of streams, roads, and parcel boundaries were identified. Using publicly-available spatial data on these features and Virginia forest cover, we found that nearly 30 percent of operable forests (in
patches at least 20 acres [8.09 ha] in size) would be impacted by local ordinances. Furthermore, indirect effects of parcelization would affect another 10 percent of operable forests. An analysis of terrestrial carbon sequestration opportunities in Virginia has been completed. GIS analysis of soils, forest inventory, topography, hydrography and other layers was conducted to identify land areas where carbon sequestration possibilities such as afforestation and conversion of tillage practices would be possible; potential carbon storage impacts were estimated.
Impacts
Results from this research will help to expand the use of geospatial technologies such as GIS and GPS by forest managers and planners. Improved awareness of low-cost, easy-to-use GIS packages should enable more smaller-scale operations (forestry consultants, landowners, small companies) make effective use of these technologies. Better access to these technologies should lead to more informed decisionmaking and more effective resource management. Improved understanding of the magnitude and characteristics of errors in GPS-derived data can help managers and planners assess the utility of results of spatial analyses. Because maps (and estimates derived from them, such as area) are essentially single samples of spatial data, conventional statistical methods for deriving confidence intervals cannot apply. The results described here are a step toward using other approaches to describe the confidence that users may have in map-derived data.
Publications
- Bernard, A. and S. P. Prisley. 2005. Digital mapping alternatives: GIS for the busy forester. Journal of Forestry 103(4):163-168.
- Johnson, L. Y. 2005. Evaluation of Bureau of Land Management protocols for monitoring stream condition. M.S. Thesis, Department of Forestry, Virginia Polytechnic Institute and State University, Blacksburg, VA. 101 p.
- Galang, J. A., C. E. Zipper, S. P. Prisley, J. M. Galbraith and R. H. Wynne. 2005. Evaluating terrestrial carbon sequestration options for Virginia. Report to the Southeastern Regional Carbon Sequestration Partnership. 51 p.
- Mortimer, M. J., S. P. Prisley, D. Daversa and L. Stull. 2005. The prevalence and effects of local forest-related ordinances in the Commonwealth of Virginia. Report to Virginia Board of Forestry. 52 p.
- Galang, J. S. 2004. A comparison of GIS approaches to slope instability zonation in the central Blue Ridge Mountains of Virginia. M.S. Thesis. Department of Forestry, Virginia Polytechnic Institute and State University, Blacksburg, VA. 121 pp.
|
Progress 10/01/03 to 09/30/04
Outputs
Efforts during this period included an evaluation of the impact of GPS sampling intervals on polygon area estimates, a review of modeling guidelines for carbon accounting, and an evaluation of spatial data resolution on models of landslide risk. Multiple GPS traverses around a photo-identifiable area on campus were conducted to observe temporal variation in GPS location accuracy and its effects on polygon area. Subsets of GPS locations were used to simulate different sampling frequencies in field digitizing of area features. In a convex shaped area, increasing the spatial interval between GPS sample locations reduces the estimated area. Models are an integral component of forest carbon accounting, used widely to convert forest inventory data collected for management purposes into carbon units for evaluation of sequestration rates. We reviewed literature on generally accepted guidelines for evaluation of models in applications such as forest carbon modeling where
quantitative results have policy impacts; a synthesis paper was published. Landslides have been associated with road construction and some forestry operations in steep terrain. Numerous models have been developed to assess the risk of certain terrain conditions to slope failures resulting in landslides. Three modeling approaches were evaluated using two data sources for study areas in Virginia for which landslide inventories are available. The qualitative map combination approach was most efficient at categorizing high risk areas.
Impacts
Results from this research will help to expand the use of geospatial technologies such as GIS and GPS by forest managers and planners. Improved awareness of low-cost, easy-to-use GIS packages should enable more smaller-scale operations (forestry consultants, landowners, small companies) make effective use of these technologies. Better access to these technologies should lead to more informed decisionmaking and more effective resource management. Improved understanding of the magnitude and characteristics of errors in GPS-derived data can help managers and planners assess the utility of results of spatial analyses. Because maps (and estimates derived from them, such as area) are essentially single samples of spatial data, conventional statistical methods for deriving confidence intervals cannot apply. The results described here are a step toward using other approaches to describe the confidence that users may have in map-derived data.
Publications
- Prisley, S. P. and M. J. Mortimer. 2004. A synthesis of literature on evaluation of models for policy applications, with implications for carbon accounting. For. Ecol. & Mgt. 198(1-3):89-103.
- Johnson, L. Y. and S. P. Prisley. 2004. The effect of spatial accuracy on analysis of range management and stream macroinvertebrates. 2004 Annual Conference, American Soc. Photogrammetry and Remote Sensing, Denver, CO. May 2004. (Poster)
- Radtke, P. J., Westfall, J. A. and H. E. Burkhart. 2003. Conditioning a distance-dependent competition index to indicate the onset of inter-tree competition. Forest Ecology and Management 175:17-30.
- Amateis, R. L., P. J. Radtke, and G. D. Hansen. 2004. The effect of spacing rectangularity on stem quality in loblolly pine plantations. Canadian Journal of Forest Research 34(2):498-501.
- Radtke, P. J., Prisley, S. P., Amateis, R. L., Copenheaver, C. A. and H. E. Burkhart. 2004. A proposed model for deadwood C production and decay in loblolly pine plantations. Environmental Management (In press).
|
Progress 10/01/02 to 09/30/03
Outputs
Efforts during this six-month period focused on evaluations of the spatial accuracy of Global Positioning System (GPS) data collected for forest plan development and the implications of that accuracy on determination of areas of stands or treatment areas. In addition, a comparative analysis of low-cost Geographic Information System (GIS) software packages was conducted. GPS data were collected at benchmarks on the Virginia Tech campus, at polygonal areas around a parking lot, and at parcel corners and stand boundaries in a forested tract in eastern Texas. Multiple GPS receivers were used in each experiment. Results from benchmark tests indicated that location errors from the receivers followed the Rayleigh distribution, with mean errors of 3m for the Garmin Etrex receiver and 10m for the Magellan 315 receiver. Error in polygon area (compared to polygon delineation from high-resolution aerial photographs) was on the order of 2 percent. Point errors have been found to
be correlated in time as well, and analysis is ongoing to describe the statistical properties of that correlation. These results can help to provide levels of confidence in GPS-derived point locations and area estimates. Knowledge of likely error distributions from low-cost GPS receivers can inform resource managers of the benefits associated with expenditures associated with higher-cost, higher-accuracy GPS receivers. Errors following the observed benchmark error distributions were simulated using a spreadsheet-based simulation software. The standard deviation of point errors is being correlated with the standard deviation of area errors (analysis incomplete). It had been mathematically determined that the influence of errors at an individual point (vertex) on the polygon area is equivalent to one-half of the distance between points nearest neighbors on the polygon boundary. For example, if a point location is in error by 1m, and the neighbors of the point are 100m apart, then the
area of the polygon will be in error by 50 m2. These results will enable resource mappers to evaluate which corners of a parcel, for example, are most influential in determining polygon area and thus represent the most important sites to accurately determine location coordinates. GSY=3
Impacts
Results from this research will help to expand the use of geospatial technologies such as GIS and GPS by forest managers and planners. Improved awareness of low-cost, easy-to-use GIS packages should enable more smaller-scale operations (forestry consultants, landowners, small companies) to make effective use of these technologies. Better access to these technologies should lead to more informed decisionmaking and more effective resource management. Improved understanding of the magnitude and characteristics of errors in GPS-derived data can help managers and planners assess the utility of results of spatial analyses. Because maps (and estimates derived from them, such as area) are essentially single samples of spatial data, conventional statistical methods for deriving confidence intervals cannot apply. The results described here are a step toward using other approaches to describe the confidence that users may have in map-derived data.
Publications
- Radtke, P. J., T. E. Burk and P. V. Bolstad. 2002. Bayesian melding of a forest ecosystem model with correlated inputs. Forest Science 48:701-711.
- Radtke, P. J., J. A. Westfall and H. E. Burkhart. 2003. Conditioning a distance-dependent competition index to indicate the onset of inter-tree competition. Forest Ecology and Management 175:17-30.
- Henning, J. G. and P. J. Radtke. 2003. Non-destructive stem taper measurements: Can you see the trees through the point-cloud? In: Proc. Joint Conference of the Southern Mensurationists and Northeastern Mensurationist Organization; Oct. 6-8, 2003; Roanoke, VA. (In press)
- Amateis, R. L. 2003. Quantitative tools and strategies for modeling forest systems at different scales. In: Amaro, A., D. Reed and P. Soares, eds. Proc. Modelling Forest Systems, CAB International. pp. 87-95.
- Amateis, R. L., M. Sharma and H. E. Burkhart. 2003. Scaling growth relationships from seedling plots using similarity analysis. Forest Science 49:188-195.
- Amateis, R. L., M. Sharma and H. E. Burkhart. 2003. Using miniature-scale plantations as experimental tools for assessing sustainability issues. Canadian Journal of Forest Research 33:450-454.
- Burkhart, H. E. 2003. Suggestions for choosing an appropriate level for modeling forest stands. In: Amaro, A., D. Reed and P. Soares, eds. Proc. Modelling Forest Systems, CAB International. pp. 3-10.
- Burkhart, H. E. 2003. Assessing the sustainability of plantation management. Forest Landowner 62(2):18-21.
- MacFarlane, D. W., E. J. Green, A. Brunner and R. L. Amateis. 2003. Modeling loblolly pine canopy dynamics for a light capture model. Forest Ecology and Management 173:145-168.
- Sharma, M. and H. E. Burkhart. 2003. Selecting a level of conditioning for the segmented polynomial taper equation. Forest Science 49:324-330.
- Sharma, M., R. L. Amateis and H. E. Burkhart. 2003. Forest stand dynamics and similarity theory. Ecological Mod. 167:165-180.
- Westfall, J. A. and R. L. Amateis. 2003. A model to account for potential correlations between growth of loblolly pine and changing ambient carbon dioxide concentrations. Southern Journal of Applied Forestry 27:279-284.
|
|