Technical ReportPDF Available

Using computer visualizations to help understand how forests change and develop

Authors:

Abstract and Figures

INTRODUCTION Trying to understand the effects of alternative management practices over long periods puts an enormous burden on decision-makers or the public when making decisions. One means of easing the burden is to develop a range of precise, carefully developed illustrations of the important characteristics of the anticipated changes. Computer visualizations can be a powerful tool for such illustrations. The significant advantage of the computer image is that it is readily manipulated to represent the impact of management activities on forest growth. The work described in this paper offers some guidance toward making the power of the visual world a valid and reliable surrogate for the real world that we manage. As part of a larger study (see Daniel and Vining chapters for other aspects), computer visualizations were developed—using the kind of data a forest manager will typically be able to access—to clearly show the development of the forest over time and in response to different management actions. This paper will discuss four aspects of the process managers need to consider in using visualizations and then provide a specific example of how these issues were addressed in our study.
Content may be subject to copyright.
2006, Orland, B. and C. Ursavas. Using computer visualizations to help understand how
forests change and develop. In, S. McCaffrey (ed.) The Public and Wildland Fire
Management: Social Science Findings for Managers. Gen. Tech. Rep. NRS-1. Newtown
Square, PA: U.S. Department of Agriculture, Forest Service, Northern Research Station. 187-
196.
Contact:
Brian Orland.
Penn State University, University Park, PA 16802. (boo1@psu.edu)
INTRODUCTION
Trying to understand the effects of alternative management practices over long periods puts an
enormous burden on decision-makers or the public when making decisions. One means of easing the
burden is to develop a range of precise, carefully developed illustrations of the important characteristics
of the anticipated changes. Computer visualizations can be a powerful tool for such illustrations. The
significant advantage of the computer image is that it is readily manipulated to represent the impact of
management activities on forest growth.
The work described in this paper offers some guidance toward making the power of the visual world a
valid and reliable surrogate for the real world that we manage. As part of a larger study (see Daniel and
Vining chapters for other aspects), computer visualizations were developed—using the kind of data a
forest manager will typically be able to access—to clearly show the development of the forest over time
and in response to different management actions. This paper will discuss four aspects of the process
managers need to consider in using visualizations and then provide a specific example of how these
issues were addressed in our study.
KEY FINDINGS: Making Visualizations Useful to Forest Managers
Managers need to consider four issues when using visualizations:
Visual realism: To what degree does the image match the real world?
Information-driven visualization: Can it be shown that the image is an accurate representation
of forest conditions?
Change over time: Can it show an essential component of forest landscape management,
change over time?
Challenges in using visualizations:
o Using sampled data
o Data availability and quality
Visual Realism
Using digital or scanned photographs, it is feasible, and quite easy, to use image editing tools such as
Adobe Photoshop® to create highly realistic images of landscapes modified by management and
changing over time. Figure 1 provides an example from a study of forest harvest practices (Orland et al.
1994). The visualizations were created by referencing a large library of photographs of known ground
conditions that provided visual templates, but most projects do not have the luxury of such libraries.
While the images themselves are convincing representations, the visual changes are not necessarily
connected to any underlying information such as forest density, species mix, diameter, or terrain.
Orland and Ursavas, Visualization as a tool for understanding how forests change and develop 2
Figure 1.—Images representing attribute levels for variable “residuals in cut area”.
Information-driven Visualization
Generally, investigation of such complex issues as forest management requires that images accurately
represent measured or predicted ground conditions. Some software developers have created tools that
provide visual representations of forest inventory data. The Stand Visualization System (SVS) and
Envision, a landscape-scale visualization tool, were developed by Robert McGaughey and his
colleagues (McGaughey 2003), and SmartForest was developed by the author and his collaborators
(Orland 2003). Both use USDA Forest Service forest inventory data to create visual representations of
forest stands. In the former case, this is done at the scale of a 1- to 4-acre plot with no reference to the
landscape context; in the latter case, this is done at landscape scale in the context of other stands and
including the representation of topography.
Another necessity is a dataset with sufficient detail to show noncommercial or less important species
beside the dominant forest types and including details of the shrub and herbaceous components of the
forest. Each of these has significant visual impact, but requires that sufficiently detailed data are
available. Such information is particularly important when discussing treatments to decrease fuel loads;
studies have found that understory vegetation is an important component of scenic beauty ratings
(Ryan 2005).
Change Over Time
As Figure 2 demonstrates, the necessity to consider change over time is a central aspect of forest
management. To project the changes in biophysical components of the forest, the Forest Vegetation
Simulator (FVS) (Dixon 2003) is one of a family of tools developed to enable forest managers to project
future forest conditions. In wide use by forest management agencies, FVS and its derivatives are
capable of modeling very complex growth processes. In the context of silvicultural and harvest
operations, users can specify parameters for a range of forest operations including thinning and
planting as well as major treatments anticipated for the study sites. The growth model takes into
account over-performing and underperforming trees, as well as mortality among out-competed or
senescent trees, and includes natural regeneration of both commercial and noncommercial species.
Output data from FVS can be used by each of the visualization tools identified earlier to create images
of the forest under a range of management scenarios and at time-steps into the future as specified in
creating the FVS projections. The resultant images are a powerful tool for communicating the
implications of management programs, especially to groups from multidisciplinary backgrounds where
the visualization serves as a common meeting ground for their different understandings.
Challenges in Using Visualizations
Visualization is especially challenging where trees and shrubs are recognizable as individuals of
different species and are clumped or dispersed with respect to one another, yet the information
gathered and projected about their growth and change is based on sampled data and statistically
summarized. In foresters’ terms the “stand” or “block” is a fundamental unit of forest management, each
being defined as an area of relatively homogeneous forest of consistent topographical characteristics
Orland and Ursavas, Visualization as a tool for understanding how forests change and develop 3
such as slope and aspect. The same conditions apply to urban and recreational forest, where
management actions are taken on individual and recognizable trees, yet information about the forest is
maintained as numbers of trees per acre, with little or no spatial information maintained at the tree-by-
tree level.
One critical necessity at the heart of using such visualizations is to accept that the image is from
sampled data and does not represent a real location—even though the visualization is sufficiently
realistic to create a plausible sense of place. Users must accept that scenes are no more than
surrogates for “the real world” and that all that is necessary in the context of a project such as this one
is to ensure that the landscape behaves plausibly. Managers will need to constantly remind people to
separate themselves from considering each location as “real.”
The other closely related fundamental considerations are the availability and quality of the data used to
develop the visualizations. FVS is a highly developed tool, but its capabilities are dictated by the
completeness and accuracy of the input data. One thing that visualization is especially useful for is to
show the errors and omissions in available data; in other scientific fields a principal use of visualization
is in data verification. FVS is also a statistically derived numerical modeling tool—thus its projections
have uncertainties associated with them. That uncertainty, while challenging to visualize, is
nevertheless an important issue to consider when using visualization in decision-making. The realism of
the visual imagery may create a false sense of confidence in what are, in fact, best approximations of
what the future will bring (Orland et al. 2001).
VISUAL CASE STUDY
On July 4, 1999, a powerful windstorm affecting the Boundary Waters Canoe Area Wilderness resulted
in widespread forest blowdowns—areas of completely uprooted or snapped-off conifer and deciduous
trees (USDA Forest Service 2000). The blowdown area is in northern Minnesota and across the
Canadian border in western Ontario. The opportunity to monitor the recovery of this important area has
resulted in forest inventory data of high quality, collected as part of an intensive inventory for ecological
modeling (Gilmore et al. 2003). The data included the species and size of each stem more than 6 mm
in diameter.
Using that data alone, it was possible to create very convincing images using SmartForest
(http://www.imlab.psu.edu/smartforest
)—a tool developed as a landscape-scale visualization tool
capable of distributing forest stand data according to stand boundaries defined via the ArcGIS®
Geographic Information System and over terrain derived from a USGS Digital Elevation Model. We will
use SmartForest images as the basis for our discussion here, but the principles apply to images
created using other visualization tools.
Figure 2.—Actual site conditions photographed in 2001 and 2003.
Orland and Ursavas, Visualization as a tool for understanding how forests change and develop 4
Figure 3.—In-stand photos and visualization: Salvage area residuals.
Figures 3 and 4 represent different forest conditions. It is evident that the computer-generated images
lack the realism of the photographs. However, it is also evident that the distribution of large woody
material within the immediate forest stand is similar in numbers, species, and sizes of trees—the
component of forest management most likely to be impacted by policy changes.
Figure 4.—In-stand photos and visualization: Undisturbed, mature forest area.
Much realism can be achieved with full representation of shrubs and grasses. Figure 5 illustrates the
effects of adding those components to the scene. Just as in photographs of the real landscape, the
addition of a vigorous shrub component to the visualization can be as visually devastating as it is in the
real world.
Orland and Ursavas, Visualization as a tool for understanding how forests change and develop 5
Figure 5.—Tree component; with grass and half-density shrubs; and with full shrubs.
Figure 6.—Recovery from salvage: Natural regeneration (top) vs. planting with pines.
Images show 2002 conditions.
Figure 7.—Growth: 2022 (top) and 2052 visualizations of the planting scenarios in figure 6.
The images in figure 6 show the early stages of recovery from a salvage operation in an area heavily
impacted by the blowdown. The top images show the regeneration from natural seed sources; the
lower images show the effect of deliberately planting a mix of red and white pines. Figure 7 shows
representations (in years 2022 and 2052) of the planted forest conditions in figure 6. When we look at
Orland and Ursavas, Visualization as a tool for understanding how forests change and develop 6
the first set of images, the impact of the pine planting is very clear and seen in the context of the
residual hardwoods. In the “growth” images, individual trees can be tracked between the time-steps,
although the major visual change is the density of the youthful growth.
In the instances visualized in figures 6 and 7, base data were available for those shrubs and forbs
present in 2001, but growth and development data for groundcover and shrub species are generally not
a component of the growth models. Although the resulting images of those components were thus not
accurate to the anticipated conditions, if such detailed information had been available the improved
validity of that aspect of the visualization might well mask changes in the major vegetative
component—the trees—just as in the photographs of the real location. Figure 4 indicated the technical
feasibility of creating accurate images of groundcover, given adequate data.
The resultant images were used in a survey (fig. 8) (Daniel, this volume) that was used to solicit public
input on desirable management scenarios for a much-impacted forest in more than 200 face-to-face
interviews. To express the passage of time represented in the image sets, they were shown as
animations stepping viewers through five time-steps. They were also used by Merrick and Vining (this
volume) to investigate what forest elements people pay attention to when determining visual
preferences.
Figure 8.—Typical survey page.
REFERENCES
Dixon, Gary S., comp. 2003. Essential FVS: a user’s guide to the Forest Vegetation Simulator. Internal
Report. Fort Collins, CO: U.S. Department of Agriculture, Forest Service, Forest Management
Center. 193 p
Gilmore, Daniel W.; Kastendick, Douglas N.; Zasada, John C.; Anderson, Paula J. 2003. Alternative
fuel reduction treatments in the Gunflint Corridor of the Superior National Forest: second year
results and sampling recommendations. Res. Note NC-381. St. Paul, MN: U.S. Department of
Agriculture, Forest Service, North Central Research Station. 8 p.
McGaughey, R. 2003. SVS—The Stand Visualization System. Portland, OR: U.S. Department of
Agriculture, Pacific Northwest Research Station. [Available online:
http://forsys.cfr.washington.edu/svs.html.
Orland and Ursavas, Visualization as a tool for understanding how forests change and develop 7
Orland, B., 2003. SmartForest—Interactive Forest Visualization. State College, PA: Penn State
University, Department of Landscape Architecture. http://www.imlab.psu.edu/smartforest
Orland, B.; Budthimedhee, K.; Uusitalo, J. 2001. Considering virtual worlds as representations of
landscape realities. Landscape and Urban Planning. 54: 139-148.
Orland, B.; Daniel, T.C.; Haider, W. 1994. Calibrated images: landscape visualizations to meet rigorous
experimental design specification. In: Proceedings, Decision support 2001—resource technology
94; 1994 September 12-16; Toronto, ON. Bethesda, MD: American Society for Photogrammetry
and Remote Sensing: 919-926.
Ryan, R.L. 2005. Social science to improve fuels management: a synthesis of research on aesthetics
and fuels management. Gen. Tech. Rep. NC-261. St Paul, MN: U.S. Department of Agriculture,
Forest Service, North Central Research Station. 58 p.
U.S. Department of Agriculture, Forest Service. 2000. Gunflint Corridor fuel reduction: final
environmental impact statement. Gunflint Ranger District. Grand Marais, MN: U.S. Department of
Agriculture, Forest Service, Superior National Forest. 294 p.
ResearchGate has not been able to resolve any citations for this publication.
Article
Fuel loadings need to be considered in two ways: 1) the total fuel loadings of various size classes and 2) their distribution across a site. Fuel treatments in this study affected both. We conclude that 1) mechanical treatments of machine piling and salvage logging reduced fine and heavy fuel loadings and 2) prescribed fire was successful in reducing fine fuel loadings (fuels less than 3 inches in diameter) but less successful than salvage logging and mechanical piling in reducing heavy fuel loadings (fuels greater than 3 inches in diameter). On July 4, 1999, unprecedented thunderstorm downbursts, also known as derechos (wind speeds of 75 to 110 mph), caused wind damage to approximately 477,000 acres of sub-boreal forest in the Superior National Forest, including the Boundary Waters Canoe Area Wilderness and adjacent Gunflint Corridor (fig. 1). To reduce the risk of wildfire and protect the public, four management alternatives were considered with varying degrees of management intensity. After extensive public review, the management alternative selected for implementation by the USDA Forest Service includes three fuel reduction treatments on 4,714 acres in the Gunflint Corridor: (a) prescribed burning, (b) salvage harvesting, and (c) piling of down trees with and without burning (USDA Forest Service 2000). The purpose of this paper is to provide second-year results on the efficacy of these fuel reduction treatments and provide recommendations for future sampling in similar situations.
Article
The development of computer tools for creating and representing virtual worlds has dramatically increased our abilities to capture salient aspects of the environment and communicate them to audiences remote from the landscape under study. The speed and quality of generation of visual imagery, as well as the editing power computers offer, has already dramatically extended the capabilities of practitioners and researchers in representing scenic views. The immersion, motion, and sound offered by virtual worlds may greatly extend the ecological validity of environmental representations and allow for deeper and more meaningful study of the effects of the real world on human experience. However, while the tools are being widely and increasingly adopted, there has been little discussion of the nature of the advantages being gained, nor the cautions that may be necessary. This paper examines the suitability of Virtual Reality (VR) technology for supporting environmental decision-making. We discuss and categorize different aspects of human-computer interfaces and then discuss to what extent the attributes of VR correspond to the needs of landscape representation.
SVS—The Stand Visualization System. Portland, OR: U.S. Department of Agriculture, Pacific Northwest Research Station. [Available online
  • R Mcgaughey
McGaughey, R. 2003. SVS—The Stand Visualization System. Portland, OR: U.S. Department of Agriculture, Pacific Northwest Research Station. [Available online: http://forsys.cfr.washington.edu/svs.html.
Calibrated images: landscape visualizations to meet rigorous experimental design specification
  • B Orland
  • T C Daniel
  • W Haider
Orland, B.; Daniel, T.C.; Haider, W. 1994. Calibrated images: landscape visualizations to meet rigorous experimental design specification. In: Proceedings, Decision support 2001-resource technology 94; 1994 September 12-16;
Social science to improve fuels management: a synthesis of research on aesthetics and fuels management
  • R L Ryan
Ryan, R.L. 2005. Social science to improve fuels management: a synthesis of research on aesthetics and fuels management. Gen. Tech. Rep. NC-261. St Paul, MN: U.S. Department of Agriculture, Forest Service, North Central Research Station. 58 p.
Gunflint Corridor fuel reduction: final environmental impact statement
U.S. Department of Agriculture, Forest Service. 2000. Gunflint Corridor fuel reduction: final environmental impact statement. Gunflint Ranger District. Grand Marais, MN: U.S. Department of Agriculture, Forest Service, Superior National Forest. 294 p.
SVS-The Stand Visualization System
  • R Mcgaughey
McGaughey, R. 2003. SVS-The Stand Visualization System. Portland, OR: U.S. Department of Agriculture, Pacific Northwest Research Station. [Available online: http://forsys.cfr.washington.edu/svs.html.
SmartForest-Interactive Forest Visualization
  • B Orland
Orland, B., 2003. SmartForest-Interactive Forest Visualization. State College, PA: Penn State University, Department of Landscape Architecture. http://www.imlab.psu.edu/smartforest
MD: American Society for Photogrammetry and Remote Sensing
  • O N Toronto
  • Bethesda
Toronto, ON. Bethesda, MD: American Society for Photogrammetry and Remote Sensing: 919-926.