From the Publisher
By Roland Mangold

3D Forest Modeling: A Personal Perspective
By David Balcaen

If one wished to design a building that looked and functioned in a certain way, one would likely hire an architect. When wanting to create a logging project that looks and functions in a specific way, however, the choices are quite a bit narrower. Very few people have addressed these issues in an overall fashion. Heretofore the process was, drag out the topographic map, draw some sight lines, and scratch away at a photo to simulate the impact of logging. This is far from an exact science, to say the least. Other techniques include using the terrain lines to assist in a design of "pleasant appearing clear cuts," a tall order to fill and an oxymoron where most people are concerned. The science of forestry deals with many aspects of growing new plantations of trees, doing a census of the trees already standing, engineering the act of logging, and managing both the plethora of regulations and, to some degree, how logging looks. Integrating all these factors, plus all the while creating a pleasant-looking logging site, is a challenge in visually sensitive areas. With the advent and refinement of digital elevation models and digital land cover information, a new field of forestry consultation has opened up. I like to call it "viewscape generation."
      I live in central British Columbia, Canada. Logging here is a popular pastime, and I will use as my example a fictitious company called Loggall International ('where clear-cut and burn is a way of life'). All kidding aside, this is the view most people have of the British Columbia forest industry. To some degree, this is true. But over the past decade, the climate has changed as to what is acceptable to society regarding the way logging looks. This rising awareness of the physical attributes of logging has opened the door for photo-realistic landscape modeling and the use of 3D modeling tools heretofore absent from forestry planning.
      Here in the province of B.C., one has a choice of which digital elevation models to use. Groupings of files, called TRIM for short, contain contours, DEM points, roads, break lines and other files-all in a UTM projection-with some NAD 27 and some NAD 83. These files are basically a variation on the Microstation format theme. The ones I trust the most are the random DEM points with break lines. Each 1:20,000 map sheet covers 11.1 km north-to-south, and approximately 13 km east-to-west. When combined with break lines, this data accurately describes a model of the land. Break lines are key elements in that they describe locations and elevations along valley bottoms, ridges and other "breaks" in the terrain. This is very important when doing visual simulations! Next we have the land cover information, which is pivotal to putting the correct foliage on the DEM at the correct location. In B.C., it takes the form of complex files that describe both forest and non-forest land cover by using labeled polygons and an associated data file for each respective polygon.

Big deal.
We have all this detailed and wonderful data, but how do we turn it into a photo-realistic model of the land? World Construction Set is the landscape architect's ultimate tool. In essence, this tool prevents costly mistakes. It builds the models with the data just described, and gives one the flexibility to "model" a new plan in minutes, saving many hours of time and creating a product that just wasn't possible beforehand. The land cover is equally well handled, from broad sweeps of entire continents to the details along a roadside stopping area, complete with shrubs and a picnic table. There is other modeling software available as well. Terragen, Maps 3D, World Builder, and several other titles are among them. To their merit, they all contribute to the world of viewscape generation, each one "upping the ante" in the competitive arena of software development. I have chosen World Construction Set, as it most closely matches the needs of a forester doing viewscapes of potential logging projects. For example, it has the feature of standing 2D trees up on a 3D-terrain model, accurately portraying the forest. The other packages each have their own strong points, and I will leave it up to the reader to discover what they are.
      So what does all this computing power mean to a forester? Let's say that you or your clients are concerned about proposing and executing a logging project that needs to meet certain visibility goals. Photo-realistic modeling allows you to alter the location and shape of the proposed project, as well as of the logging method itself. One gains instant feedback to any such changes, an unprecedented perspective in the visual design process. It is the WYSIWYG (what you see is what you get) of the landscape design world. This is also a dream-come-true for planners. In the past I have been involved with many "trial and error" techniques for visually designing logging; photo-realistic modeling turns this into "trial and success."
      To give you an example of photo-realistic modeling in the forestry visualization field, I will present a typical project. My client, Loggall International, has the authority to log a specific volume of timber from an area south of a large lake. Aside from the fish, many year-round residents call the lake home, and in the summertime the population swells considerably. Despite its name, Loggall International has learned that, if a company plans to alter the land, it must involve all of the people who have a vested interest in the land. The viewshed in question contains many low hills against a backdrop of mountains, and the area of concern is visible from most of the opposite shore of the lake. Loggall wants its logging plan designed to meet all of the environmental, silvi-cultural and viewshed goals, plus also wants it to be acceptable to local residents. The company's engineers have drawn up a preliminary plan that includes cut blocks, cutting method to be used, roads, and initial viewpoints. The plan is e-mailed to us, along with the digital elevation model (DEM) and the land cover information.
      Our crew goes to work, importing the DEM, converting the land cover information to a raster data set, and importing the vector data that describes the proposed logging project. The land-cover-information raster file, or bitmap image, depicts tree species, species mix, plus the height and density of the trees. We had already written our own utility to create this raster image. And a word about the "crew." All of us have a forestry background first, and subsequently added the skills of photography, 3D modeling, and methods of public presentation.
      Ecosystem data can be applied using vector data, raster data and "rules of nature." Vector data is the most accurate and best suited to foreground detail, and the actual development plan itself. Raster data is as accurate as one wishes it to be, generally better suited to broad sweeps of land up to 50 km from one's point of view. The "rules of nature" concept is a very easy and accurate way to cover with foliage all areas outside the scope of interest, yet within the field of view. "Rules of nature" places ecosystems on the terrain, following rules that specify elevation limits, azimuth skew of this limit, concavity or convexity of the land, and the preponderance of an ecosystem to either creep up a ridge, or else fill in lowlands. This is powerful stuff.
      We are placing 2D trees on a 3D-terrain model. This is fast, much faster than placing 3D trees on a 3D terrain, and infinitely more accurate than draping a photo over the terrain. With the trees standing up on the terrain, the screening effect of trees is quite evident. Partial cuts, small patch cuts, thinning, multiple age levels- any forest composition can be modeled. At this point we require site photos to ensure that our digital model really does match reality. A reconnaissance trip finds us standing at the viewpoints where photographs are taken-both normal and telephoto-to be used in matching the texture, color and lighting of our model. If the logging plan involves leaving some or most of the trees in place, then a walk through the area ensures that the trees we are planning to leave are truly there. Back at the office these site photos, and the other information that has been gathered, are imperative in order to verify the model. Our watchwords remain, "Verify, verify, verify!"
      While we were on this reconnaissance trip, photographs of trees similar to the ones to be left in the selection cutting area were shot. These were photographs of both mature and immature Douglas fir, and some mature pines. These images are used in our model to ensure that the trees depicted in the model match the trees that are to be left standing on the land. The trees in this instance happen to be sparsely limbed on their lower half, so it is important for model accuracy that the correct images are used. We have a large library of tree images, sorted by bio-region and elevation, and find this to be a convenient arrangement. Using local-area tree images is important, so that the model will take on the appearance of the area in which one is working.
      Lakes, roads, clouds, streams, and detailed foreground information as described by the vector data are all imported, and these aspects add a tremendous amount of validity and believability to the verified model. Another powerful tool that is foreign to most foresters is the use of procedural textures, especially useful in portraying a forest's floor, water surfaces, and vector objects such as road surfaces and side cuts. Textures are powerful communicators, and our subconscious willingly accepts an image that includes these seemingly non-forestry related aspects of that image. Having credibility with the interest groups that are going to make choices from these images is vital.
      Now that we have built the model, with development plans applied and previews sent to the client, the process of design can proceed. At this point, the client may want to involve others who will review the plan and suggest changes to it. We can model any form of logging there is, from clear-cutting to a 10 percent thinning of a stand. After viewing the preview images, Loggall decides to go with a combination of small clear-cuts in the pine stand, and selection logging in the mixed fir and pine stands.
      This plan fits with how they wish to regenerate the forest after logging to provide the desired look. Remember, the goal is not to hide the logging, but rather to design it to fit within the constraints present. At this stage of the process, another set of preview images is generated and e-mailed to Loggall. The company is happy, the images are then sent to several of the others involved in the process, and a preliminary agreement is obtained.
      Final high-resolution, print-quality images are rendered and printed on photo-quality paper. These images and some of the site photos are then labeled, placed in a covering document that explains what is going on, and then both are printed as an Adobe PDF (portable document file). The PDF file is great for posting on a web site for all to see, and allows the proposal to be e-mailed to whomever is interested. Printed documents and PDFs are a flexible pair of communication tools. Loggall is happy with the product, the public interest groups are happy to know what the logging will look like, regulators are pleased not to be in the middle of a controversy, and I am happy to be making a living doing this sort of work.
      I find that World Construction Set is the best tool that I have, by a long shot, for predicting what logging will ultimately look like. Even Loggall International (where clear-cut and burn was a way of life) has come of age, taking its social responsibilities in hand by making use of this very accurate tool. To complete the story, Loggall won the approval of each of the interest groups due to the company's openness, flexibility, and by presenting the logging plan in a clear and understandable manner.
      Before my forestry career began, I worked as a television news cameraman. I was always drumming up stories, and the news director started me off on the right foot when he kept asking me, "Who cares and why?" about each of my stories. I grumbled at this a bit, but soon came to realize that one may as well do something that others care about. I am still asking these questions today, as I model the forest and seek to design logging projects that address each one of them.

About the Author:
David Balcaen operates his own forestry consulting business in the interior of British Columbia, specializing in forestry visuals. He has been using World Construction Set ever since its first version was issued in 1994. David has taught 3D landscape modeling to other foresters with considerable success.

Back