The gist of GIS

Dr. Ian McHarg, a landscape architect who published a book in 1969 titled Design with Nature, is recognized as being the father of Geographic Information System (GIS). This is due to the fact that he pioneered the concept of ecological planning and, in addition, presented ideas dealing with the spatial representation of features in two and three dimensions that were later developed into GIS. In the 1970s, Dana Tomlin, a Ph.D. student at the Harvard Graphics Lab, developed the first raster GIS program called MAP. This software was in the public domain so others developed it further. These individuals were Dr. Duane F. Marble, Dept. of Geography, The Ohio State University, Columbus, OH, who developed OSU MAP-for-the-PC in the 1980s; and Dr. Joseph Berry, Berry & Associates, Ft. Collins, CO, who developed PMAP. The 1980s saw a great proliferation of start-up GIS companies, but in recent times many have disappeared such as PAMAP from Victoria, B.C. and SPANS from Ottawa, Canada. One of these companies, Environmental Systems Research Institute (ESRI), Redlands, CA, under the leadership of Dr. Jack Dangermond, developed the concept of linking attributes stored in a database to map features in 1984. This simple idea revolutionized the GIS industry. At present ESRI has captured about 90 percent of the GIS user community and has clients all over the world. However there are still other GIS software vendors that are alive and doing well such as MAPINFO, MGA from Intergraph, and IDRISI from Clark University to name a few.

Defining Terms

What is GIS? GIS stands for Geographic Information System. It is a system of spatially referenced information, including computer programs that acquire, store, manipulate, analyze, and display spatial data. So the implication is that a computer is needed on which the GIS software is installed, a knowledgeable operator (people-ware) to perform map operations, and digital data in a useable structure and format. GIS was born out of a marriage of several disciplines, namely cartography, earth-bound surveying, database management, and remote sensing. Both the input and output of a GIS are digital maps.

What is a map? A map is a graphic representation of features defined both by position with reference to a coordinate system and by their non-spatial attributes. Maps are all about place and objects in relationship to each other. Maps are composed of points, lines, polygons, and symbols and, if complete, a map includes a legend, scale bar, north arrow, and coordinate system tic marks. Maps are often thematic and produced for a single purpose.

Paper maps of various kinds have been around for centuries and were drawn by hand and initially used for navigation, both on land and sea. Symbology was created by cartographers to represent both natural and man-made features located on a specific area of interest on the earth's surface such as roads, trails, towns, property boundaries, streams, fences, swamps, mountains, lakes, bridges, forts, etc. Rich geographical data have been collected about things such as soils, people, vegetation, climate, etc. Maps allow these data to be geographically referenced and interpreted using both spatial and non-spatial attributes.

Paper maps are drawn to different scales, possess varying resolutions, and utilize different map projections. Map scale is the relationship between distances on a map and the corresponding distances on the earth's surface. Map scale is reported as a ratio of map distance to actual ground distance, e.g. 1:12,000. This means that 1 distance unit on the map equals 12,000 distance units on the ground. Further, if the number of distance units to the right of the colon is small, the map scale is large. On the other hand if the second number is large (say 100,000), then the map scale is said to be small. Resolution deals with how accurately the location and shape of map features can be depicted at a given map scale. A map projection is a mathematical transformation that relates points on a sphere (three dimensions) to those on a plane (two dimensions). Hundreds of mathematical transformations have been developed to convert points on a sphere to those on a planar surface. However, in all cases, no matter which map projection system is chosen, errors either in shape, size, distance, or direction, singly or in combination, are introduced into the final map.

The content of paper maps can be structured in such a way that it can be stored, retrieved and analyzed on a computer through the use of geodatabases. Such digital maps have several advantages over paper maps; namely, that they are made more quickly and cheaply, are made for specific user needs, don't require skilled staff, are easier to make and update, and are dynamic, not static.

Digital map data in a GIS are represented either in vector (feature) or raster (grid) format depending on the data source. For example, all remotely-sensed data are raster, which consists of rows and columns of equal-sized cells (x, y coordinate) having a specific z-value assigned to each cell. On the other hand, vector data consist of a number of points, identified either singly by x, y location or as a series of points joined together to form a line segment if only connectivity exists or as a polygon if closure exists as well.

Within a GIS a map is generally composed of not a single layer but multiple layers of digital data. Each layer contains a set of features having a common theme, i.e. topography in the form of contours of equal elevation above sea-level, or hydrography (rivers and lakes), roads (paved and unpaved), vegetation, average precipitation zones, solar radiation, soil survey areas, human population density, wildlife habitat, etc. These layers can be examined and displayed separately or combined using map algebra. Any operation that can be performed on numbers (addition, subtraction, multiplication, or division) can be applied to map layers.

Due to rapid advances in computer technology, hardware, and GIS software, as well as the acquisition and availability of digital data from satellite sensors on orbiting platforms above the earth and GPS (Global Positioning System) receivers, complicated maps can be created and displayed in two and three dimensions. Further, advanced tools have been provided by GIS software vendors that permit very sophisticated spatial analyses that have applications in a wide range of disciplines from health care and sociology to recreation and forest management to forest fire suppression and renewable energy to transportation and climate change, to name only a few.

Types of GIS operations

A GIS is designed to allow a user to obtain answers to an almost unlimited number of spatial questions. The main question is: "Where am I (or is an object or are a series of objects) in relationship to everything else?" not just "Where am I (or is it or are they)?"

A GIS can perform a number of 2-dimensional analyses such as overlay operations, querying and theming, corridor and proximity analyses, network analyses and report generation from planning rings.

A GIS can generate a number of products from a 3-dimensional DEM (digital elevation model) such as block diagrams; profiles and horizons; cut-and-fill earthwork calculations; contour maps; viewsheds showing areas that are visible and invisible from a user-specified point, line or boundary; maps of slope (rate of change in elevation), aspect (orientation of non-zero slope), convexity (upward sloping), and concavity (downward sloping); and shaded relief maps.

Other GIS functions such as surface modeling, draping, creation of orthographic and perspective views as well as triangular irregular networks (TIN) can also be executed.

GIS and forest management challenges

Let me now turn my attention to the management sciences, specifically natural resources management. It is generally agreed that natural resources managers cannot continue to do business as usual, in part because of the increased complexity of our global village. We need to 1) have greater integration of different land uses and values, 2) focus on changing public values and uses of our resources, 3) identify different approaches to making management decisions, and finally 4) have better management tools and improved knowledge of management consequences. Although not all of these needs can be met through the use of GIS technology, it would be part of a management toolbox.

Current forest management situations are more complicated than they were forty years ago for two reasons. First, with the advent of the information age in the late sixties, natural resource managers have had to deal with ever increasing amounts of data. These data have inundated professionals due to the time and effort necessary to acquire and process them. Second, our publics/clients have become more knowledgeable, more educated, more concerned about their environment. So forest managers have had to become more sophisticated through the use of analytical tools such as tree growth simulators, harvest schedulers, economic analyses, cruise/inventory report writers, optimizers for road construction, sediment yield models, wildlife population models, and forest-wide/area-specific planning models, to name a few. In turn these tools have required the use of computers to execute them and the acquisition of appropriate data sets for input.

The outputs of these computer programs have usually been exhibited in the form of tabular reports. What has been the problem with these reports? They have provided natural resource managers with only half the picture, namely the WHAT. But what about the WHERE? Where on the ground are actual or planned changes taking place? The only efficient, organized way of addressing this problem is to utilize Geographic Information System (GIS) technology. GIS has become an effective and efficient planning and decision-making tool for natural resource managers.

One of the common problems facing planners has been the issue of resolution. At what level should natural resources planning be done? Forest-wide (regional) or area-specific (local)? With GIS technology this no longer is an issue. The level of resolution can be changed as called for by the particular set of analyses that need to be conducted in order to solve a natural resource problem provided that the necessary digital data are accessible and have been acquired.

For me the use of emerging technologies such as GIS (40 years old now) has provided new ways to analyze and visualize large sets of spatial and non-spatial data, and to play "what if" games on a computer to determine the effects of various alternative forest management actions without actually carrying them out on the ground. Let me give you some examples of natural resource management problems which involve the use of GIS technology to solve them.

Examples of GIS applications

In this section I describe a number of completed research projects that utilized GIS, remotely sensed imagery, and GPS receivers to obtain solutions to applied research project objectives which were impossible or at minimum very difficult to achieve prior to the advent of GIS technology. Truly the use of GIS has caused managers, planners, and scientists to change how they work to solve complex natural resource problems.

Some years ago the Fish and Wildlife Service created a war chest through an act of Congress to purchase private lands having high species diversity and high management risk. Before these lands could be purchased, maps needed to be constructed on a state-by-state basis, at an appropriate scale, depicting their location and size. This meant that four map layers had to be created, namely vegetation (flora) distribution, selected animal/insect/bird/fish (fauna) distribution, management risk, and ownership. Because the vegetation map data did not exist, remotely sensed data from satellites were used to identify and delineate classes of vegetation. The other map data were likewise acquired or generated. Then the four digital map layers were overlaid and a resultant thematic map displayed private lands in various classes of species diversity and management risk categories.

Fifteen years ago the Montana Department of State Lands expressed a strong interest in GIS technology because of increasingly complex management issues dealing with multiple use, conflict resolution, and checkerboard land ownership patterns. The Swan State Forest was designated as a pilot GIS study area and a long-term cooperative research project was started with the School of Forestry at the University of Montana to develop an Integrated Resource Information System (IRIS) on a personal desktop computer. The heart of this process was a database management system that was linked to a mapping and spatial analysis facility (initially PAMAP and later ArcView) as well as various natural resource models. A key component was an area-specific planning model that optimally identified new road segments to be constructed over up to five planning periods (decades) so as to provide access to proposed resource projects (cutting units). These resource projects were constrained by adjacency, corridors, visuals, water quality and wildlife requirements. Both timber and non-timber production were considered as the primary management objective. The road segments handled different traffic types and flows (amount and direction) and were closed either temporarily or permanently during some planning periods. Upon completion of the optimization process, various feasible solutions were displayed in tabular as well as visual form. The resulting maps displayed the boundaries of the proposed cutting units and the road segments to be constructed to them over multiple planning periods (color coded). Through a series of menus and data entry screens the complexity of the system was hidden from the user (trained natural resource manager), who nevertheless had full capability to formulate complex "what-if" management scenarios without having knowledge about the specific necessary commands to execute specific software components of the system.

This planning and decision-making tool was later modified and enhanced jointly by the staff of the USDA Forest Service Forestry Science Laboratory and U of M forestry staff to become MAGIS, Multi-resource Analysis and Geographic Information System (Zuuring et al., 1995). That second-generation modeling system has been used extensively to determine how to reduce fuel loading in the wildland urban interface (WUI) so as to protect homes from the damage of wildland fires and assist in the efficient and timely deployment of fire crews. MAGIS has also been used in determining the spatial distribution and possible sustainable flow of bio-fuels derived from small-diameter trees and the subsequent placement of small bio-generating plants as alternative energy sources. This information has been used by the high school in Darby, Montana, that has realized substantial energy savings during the winter for heating in the amount of $70,000 per year.

In the early 1990s the Montana Legislature passed HB340, "Forest Land Tax Act," which changed the then method of taxing privately-owned, commercial forested land from a yield to a potential forest productivity basis, effective January 1, 1994. To implement such a system of taxation required the use of GIS technology linked with a biogeoclimatic process model. Through the use of remotely sensed data acquired from satellites, all land in the state of Montana was classified as forested or non-forested at a resolution of five acres and larger. Then forested lands were further classified into commercial and noncommercial. Finally, commercial forested lands were grouped into five forest land productivity classes. The potential forest productivity was estimated through the use of two particular computer models called FOREST-BGC and MT-CLIM. Six main inputs drove the process model; namely, elevation, slope, aspect, leaf area, soil water-holding capacity, and climate (temperature and precipitation). Discrete classes were devised for each model input parameter. Six map layers were created to depict the spatial distribution of each characteristic on a pixel-by-pixel basis. Through the use of map algebra, estimated potential forest productivity values were assigned to geographic locations within commercial forested lands and grouped into four to six classes on a seventh map layer. Finally, an eighth map layer showing the property boundaries of privately-owned commercial forested lands was constructed at a scale of 1:15,840 and overlaid with the map containing the estimated forest productivity classes. Acreage reports were also produced.

In the province of Alberta, fire suppression personnel utilized a GIS to determine the optimal placement of new fire lookout towers and the upgrading (raising) of existing towers whose field of view had been reduced due to the increased height of trees in surrounding stands. New lookout towers were placed over the landscape in such a fashion that each tower had a 25-mile radius field-of-view. A DEM (digital elevation model) was constructed and the curvature of the Earth's surface was taken into consideration as well as the location of timber stands and average tree height. Elevation values were adjusted for tree height and maps depicting visible and invisible zones were created around each existing and proposed new tower location. The invisible zones were further subdivided into hidden (not more than 100' below the horizon) and dead (more than 100' below the horizon) zones. It was assumed that fires starting in the hidden zones could still be identified and located from the lookout towers but those starting in the dead zones would have to be identified from aircraft. Existing towers were raised to varying heights depending on trial adjustments to the height of the observer's eye above the landscape and the pre-defined minimum acceptable percentage of the field-of-view (viewshed) visible and hidden.

From March 2006 until December 2008 University of Montana College of Forestry & Conservation faculty (2) and research staff (3) in cooperation with the Montana Department of Revenue redeveloped the production of forest productivity maps for the purpose of taxing owners of private forested land (Zuuring et al., 2008). This time forest productivity was based on a soil-site model which had an error term (a measure of the difference between an actual observed value and an estimated model value) not possible with a deterministic model like FOREST-BGC. This model was driven by nine inputs: temperature, solar radiation, precipitation, slope, elevation, and four soil variables measured in the top 24 inches of a soil profile. The four soil variables were available water content (AWC), cation exchange capacity (CEC), PH, and percent bare ground. The model output was site index (SI), a measure of site quality that captures the growth potential of the site. Specifically, site index is the height of a free-to-grow vigorous tree at a reference age of 50 years. Through the use of FPS, Forest Projection System software package, a plantation of Douglas fir was grown at 10ft by 10ft spacing to a rotation age of 150 years under different combinations of site index and clumpiness, a measure of site occupancy. From these simulations the culmination of mean annual increment (CMAI) in board feet/acre/year (maximum growth) was determined and that was used as a measure of potential forest productivity. Adjustments were made to productivity to account for reduced stand density due to decreasing rainfall. Through map algebra, using a raster calculator, forest productivity maps were generated for 48 out of 56 Montana counties containing private forested lands. Those forest productivity values were separated into non-commercial and commercial forest and the latter had its productivity grouped into five grades, labeled fair to excellent. Land owners' property boundaries were then draped over these maps to determine the acres in each productivity class and in turn a graduated tax was applied to those productivity classes. The beauty of this approach was that complete digital data were available for all nine model inputs across the whole state of Montana so that predicted forest productivity was driven by topographic, climate and soil factors—all the agents that act on vegetation growth, both above and below the ground. Finally, the model is sensitive to climate change so future forest productivity values can be adjusted accordingly.

Current projects utilizing GIS

New decision-making and planning tools as related to natural resource management have been developed that can be used by individuals having little or no knowledge about GIS. An example of such a tool is CAMAS (Computer Assisted Management and Analysis System) a Spatial Decision Support System (SDSS) developed by the staff of the School of Forestry GIS Laboratory to provide the manager of Lubrecht Experimental Forest and forestry students with a set of management tools (Riekena, 1991; Zuuring, 1991). Through a unified, easy-to-use interface, a collection of software packages can be accessed and executed on desktop computers. Little knowledge is required, on the part of the user, with respect to the input requirements associated with a specific computer program or how it executes its tasks. Necessary inputs and resultant outputs are stored in a series of linked master and GIS databases. Records from these databases can be viewed and modified/edited and reports can be generated through a data management system (DMS). Some examples of tools integrated into the spatial realm of a GIS are tree growth simulators, harvest schedulers, road network optimizers, inventory/cruise compilers, investment analyzers, mapping facilities/spatial analysis functions (GIS), and database management system (DBMS). Within the GIS environment a set of commonly used spatial analysis functions (such as querying/theming and overlaying) can be invoked through keyboard function keys.

Another software package that has a lot of potential utility for resource managers is INFORMS (INtegrated Forest Resource Management System) developed by John Heasley (1991), a private contractor with ASRC Management Services (USGS), for the USDA Forest Service. USFS foresters from the Butte Ranger District of the Deerlodge National Forest demonstrated the three-dimensional data visualization capabilities of INFORMS as applied to a mountain pine beetle (MPB) infestation in lodgepole pine stands. The visual effects of various silvicultural practices ranging from no action to clear-cutting were clearly demonstrated over time. The user could either display the visual effects of four silvicultural practices simultaneously on a computer color monitor holding time fixed, or else show the effects of a single silvicultural practice over four time periods. This software package utilizes a DEM (digital elevation model) imported from a GIS which is shown as a solid-colored three-dimensional perspective view (blue sky and light brown [summer] or white [winter] ground). Variable-sized stylized tree stem forms are placed over the landscape in accordance with stand polygon boundaries and attribute information about stand density, species composition, and tree size. Tree stems with live foliage are colored green while dead trees with and without foliage are colored brown and gray respectively. In addition a tree growth simulator and a MPB infestation model are linked in a spatial framework to display changes in tree growth and rate/severity of the MPB infestation over time.

Already in the Pacific Northwest, GIS technology has been utilized to identify and delineate old growth forests that are prime habitat for spotted owls and pileated woodpeckers. Maps of elk habitat suitability have been produced that utilize a combination of distance from hiding cover, forage, and distance from roads to develop a habitat suitability index that is classified into three grades: poor, medium, and good. Now, in the North Fork of the Flathead River watershed, wildlife biologists have been studying the movements of wolves in relationship to topography and elk and deer populations. Through the use of GIS hydrological features (streams, rivers, lakes), wolf tracks and known ungulate kill sites have been draped over a number of edge-matched DEMs (digital elevation models) and displayed as a three-dimensional image (perspective view). Various other spatial investigations such as proximity and network analysis may reveal information about what conditions are favorable and unfavorable to wolf population growth and stability.

Likewise on the west slopes of the Mission Mountains, wildlife biologists are studying the cumulative effects of encroaching residential development and associated road construction on critical grizzly bear habitat using the spatial analysis capabilities of a GIS. In the vicinity of Hungry Horse Reservoir, avalanche chutes are being delineated and the unique vegetation within them is being classified since these areas are critical to grizzly bears in the spring. These features are then digitized and spatially analyzed in a GIS together with remotely sensed data from satellites to determine if these chutes and the vegetation in them can be identified and monitored from space.

Although I have already mentioned the use of GIS in previous projects focused on fire control, I want to mention the ongoing efforts of the National Center for Landscape Fire Analysis (NCLFA) located in the U of M College of Forestry & Conservation. It was established in 2001 and has been applying innovative science and technology to solving on-the-ground natural resources management problems (http://firecenter.umt.edu). NCLFA has made numerous innovative uses of GIS technology dealing with various aspects of wildland fire management.

A seamless set of all-ownership GIS maps was created for the Bob Marshall Wilderness Complex (BMWC) which covers over 1.5 million acres. These maps represented three fire management themes: vicinity, flight hazard, and suppression/fire perimeter. In addition, a toolbar named WFSA GIS Tools 1.0 was developed and integrated into the Rapid Assessment of Values At Risk (RAVAR) analysis and reporting process which is a component of the national prototype Wildland Fire Decision Support System (WFDSS). This toolbar was used by the Ninemile Ranger District of the Lolo National Forest and has the ability to perform spatial analyses on and generate reports from map layers such as locations of homes and structures, agency-delineated management areas, land ownership, threatened and endangered species habitat, recreational resources, and wilderness areas.

Likewise, the Montana Wildland Fire Base Map application has been developed by NCLFA staff to allow a GIS analyst to generate a base map in minutes for an area where a wildland fire is occurring. It takes care of data mining, data generalization and presentation, graphic refinement, and map compilation, thus freeing the wildland fire GIS specialist to focus on the creation, display, and analysis of fire incident data. Such useful products as vicinity, briefing, and Incident Action Plan (IAP) maps can be produced to aid deployment of fire crews and equipment in a timely and efficient manner previously not possible. An interactive web-based map viewer has been developed to allow anyone having an internet connection to view the Montana Wildland Fire Base Map.

As can be seen natural resource managers are faced with solving problems of an increasingly complex nature requiring the analysis of more data, and the only way that this can be done effectively and efficiently is by the utilization of the spatial analysis capabilities of a GIS and other high technology modeling and visualization tools.

Recent GIS developments and concluding remarks

In the last six years substantial GIS advances have been made in the delivery of customized web pages that contain GIS functionality. Now it is possible for individuals and agency staff to access digital data from data warehouses like the NRCS Soil DataMart or NRIS (Montana State Library) and utilize that digital data to perform specific spatial analyses without having to purchase GIS software. This capability not only allows GIS neophytes to "play" with digital maps but also can allow editing of corporate spatial database records to be updated in the field. This reduces the costs that corporations and agencies pay to store and retrieve spatial data in a timely and accurate manner.

On a quite different front, data visualization software that was previously used exclusively by Hollywood, game designers, and special effects experts has now included GIS functionality such that the images have spatial integrity and a series of such images can be stitched together into a movie clip. Examples of this are Visual F/X and WCS (World Construction Set). The latter is the leading software package for professional photorealistic terrain modeling, visualization, rendering, and animation.

Just as word-processing revolutionized writing, so GIS technology has radically changed the way spatial data are analyzed and displayed. Not only are non-geographers generating more and more maps but natural resource managers can now process large data sets and solve multi-resource problems using the expertise of a group of subject matter specialists together with the input from special interest groups and the general public. With this powerful tool, then, comes the responsibility of its users to make sure that misinformation is not produced because, unlike statistical analyses, misused spatial analyses are very hard to duplicate.

The increased analysis capabilities and visual displays offered by GIS have allowed natural resource (and other) professionals to do their jobs better. A challenge is posed nonetheless, due to the increasingly broad accessibility of GIS to "neophytes." This can be a blessing or a curse. Collectively we have a responsibility to see that this technology is not abused and that users are appropriately educated. Through the acquisition of reliable digital data combined with the proper use of GIS including other allied technologies like GPS and remote sensing, we can produce reliable maps and make this world a better place.

Ahl, R. & Zuuring, H. (2002, April 9-11). "Extrapolation of spatially explicit forest processes across biophysically similar watersheds." A paper presented at Intermountain GIS Users Conference, Big Sky, MT.

Butler, E. (2005). Modeling forest planning trade-offs on the Colorado front range, using MAGIS, an optimization, spatial decision support tool. Unpublished MS thesis, The University of Montana, Missoula, MT. 109 p.

Heasley, J.E. (1991, March 12-14). "INFORMS—The Integrated Forest Resource Management System." Invited paper presented at USDA For. Serv. R-6/PNW GIS Conference, Tacoma, WA.

Hebblewhite, M. (1998). Effects of landscape characteristics on wolf-killed elk kill-site location in Banff National Park, Alberta. [Unpublished document]. The University of Montana, Missoula, MT. 29 p.

Jones, J.G., Chew, J.D. & Zuuring, H.R. (1999, April 5-9). Applying simulation and optimization to plan fuel treatments at landscape scales. Proceedings of Fire Economics, Policy and Planning: Bottom Line Symposium, San Diego, CA.

Riekena, J. (1991, February 12-15). The CAMAS project: Building a computer-assisted management and analysis system for natural resource managers. A paper presented at the 5th Annual GIS Symposium "Applications in a Changing World", Vancouver, BC, Canada.

Zuuring, H.R. (1989, September 24-27). Teaching GIS Concepts: The University of Montana School of Forestry Experience. In Forestry on the Frontier, Proceedings of the SAF National Convention (pp. 49-51). Spokane, Washington.

Zuuring, H.R. (1991 March 12-14). CAMAS: A resource manager's toolbox. Invited paper presented at USDA For. Serv. R-6/PNW GIS Conference, Tacoma, WA.

Zuuring, H.,& Manasi, M. (1990, April 10-12). Estimating forest productivity from remotely sensed data and topographic variables using the spatial analysis capabilities of a GIS. In Management and productivity of western-montana forest soils, symposium proceedings (pp. 183-187). Boise, Idaho. USDA For. Serv. Gen. Tech. Rep. INT-280. Intermtn Res. Sta., Ogden, UT.

Zuuring, H.R. (1992). New tools to meet new perspectives. Western Wildlands 17 (4), 34-38.

Zuuring, H.R. (1994). General GIS principles (Revised). [Unpublished document]. School of Forestry, Univ. of Montana, Missoula, MT. 52 p.

Zuuring, H.R., Wood, W.L., & Jones, J.G. (1995). An overview of MAGIS: A Multi-Resource Analysis and Geographic Information System. USDA Forest Service, Intermtn. Res. Sta., Res. Note INT-RN-427. Ogden, UT. 6 p.

Zuuring, H.R., Chew, J.D., & Jones, J.G. (1999, May 18-20). Sequential use of simulation & optimization in analysis and planning. Proceedings of "The Bitterroot Ecosystem Management Research Project: What We Have Learned" Symposium, Missoula, MT.

Zuuring, H.R., Jones, J.G. & Chew, J.D. (1997, May 28-31). Applying simulation and optimization to address forest health issues at landscape scales. In Proceedings of the 7th symposium on systems analysis in forest resources (p. 470). Traverse City, MI. USDA Forest Service General Technical Report NC-GTR-205, St. Paul, MN.

Zuuring, H.R., Chew, J.D. & Jones, J.G. (2001, June 25-28). Planning Fuel Treatments at Landscape Scales using SIMPPLLE and MAGIS. In Proceedings of EU fire conference, Kalamata, Greece.

Zuuring, H.R. & Troutwine, J. (2005, July 25-29). Decision support models for economically efficient integrated forest management. Paper presented by J. Troutwine (written by Zuuring & Troutwine) at 25th ESRI International User Conference, San Diego, CA.

Zuuring, H.R., Milner, K., Ahl, R. & Hansen, M. (2008). Final report for MT DOR forest productivity project. [Unpublished document]. College of Forestry & Conservation, The University of Montana, Missoula MT, 59812. 83 p.