OR/MS Today - February 1999

February 1999

Forestry Problems Chopped Down to Size

By Jenny Summerour

This is another in a series of articles based on interviews with recent Franz Edelman Award finalists. Geared toward practitioners, the articles strive to provide lessons the Edelman authors learned and some pitfalls they encountered during the course of their work. The full Edelman papers are published in Interfaces.

Forest conservation has been a hot topic in U.S. logging areas for many years. While environmental activists have struggled to preserve forests, lumber companies have worked to maintain environmentally and economically sound practices. During the last decade, lumber companies have found themselves under close scrutiny by state and federal regulations designed to protect wildlife, forests and streams. Coinciding with this increase in regulations has been a substantial decrease in the allowable levels of timber harvest on public lands.

The Pacific Lumber Company (PALCO), a large timber-holding company with headquarters in Scotia, Cal., has worked to meet state and federal laws, and in many cases, has maintained company policy management practices even more protective than the regulations require. In 1995, PALCO contracted with VESTRA Resources, a Geographic Information System (GIS) consulting firm that helps clients solve land-use problems, to meet new California Board of Forestry wildlife, fisheries and timber resource requirements. A team of consultants at VESTRA worked with PALCO to develop a 120-year, 12-period forest-ecosystem management plan for its properties. The model they designed is an ecosystem-planning express model called Ep(x).

Initial challenges

Prior to contracting with VESTRA in 1995, PALCO managed its 200,000 acres at the individual timber-sale level using classical forest management practices, along with expertise from local foresters. The classical approach they used involves setting annual harvest volume and harvest-acreage objectives. Unfortunately, wildlife and watershed requirements and policies, at both federal and state levels, were often ambiguous and difficult to comply with.

In 1994, the California Board of Forestry approved a measure to allow large landowners, those with holdings over 50,000 acres, to submit a single sustained yield plan (SYP) for the entire landholding. The SYP was designed to address the long-term health and management of the entire forest ecosystem. Individual timber harvesting plans (THPs) filed with the state would have to comply with the overall strategy outlined in the SYP.

Because of the new, more complex wildlife habitat requirements, VESTRA was challenged to develop a workable methodology to accommodate the new constraints.

"We were trying to apply a real-world situation to something that had been modeled pretty much as small little academic problems up until that point," says Scott Holmen, senior developer at VESTRA Resources. "The biggest challenge was simply that what we were trying to do was technologically totally on the cutting edge. The computer software used to do forest growth modeling had unheard-of flexibility in terms of how it worked.

"Had we tried to do this even five years earlier, simply the computational time it would have taken and the disk space required would have basically meant that we could not have solved the problem with as much detail and complexity as the one we ended up working with," Holmen says. "In addition, we had no real precedence for how to link this kind of planning effort to the GIS side of the problem. That was something that had been discussed and done in a theoretical way, but it had never been attempted on a problem of this scope, to try to model in detail small, one to 10 acre, subunits of some 200,000 acres. We twice ended up blowing out the size of the matrix that was capable of being solved by the LP solver at that time."

Russell Fletcher, who was then VESTRA's senior operations research analyst, says that although the company had experience in GIS and knew what had been developed in terms of models, there was not an existing model that integrated the raw data. "We needed to convert data into a relational database, then take this information from the database and turn it into a linear program input," he says. "So we needed to develop a matrix generator and put it into linear database format. We then had to turn the output into a useable format — a series of maps, reports, etc."

PALCO needed a model that not only addressed the constraints imposed by the SYP and other federal and state laws and regulations, but also optimized the company's overall timberland operations and profitability after it had met those constraints. The model also had to be workable so that local foresters and logging personnel could physically implement it.

Social challenges

In addition to technological challenges, VESTRA had numerous social challenges to contend with. There were ongoing public hearings and protests involving this issue, and the public would have access to model results. The model results also had to be approved by the California Department of Forestry. Therefore, the development of the model had to be done in great detail.

VESTRA had an advantage in working with Henry Alden, who was timberlands manager at PALCO at the time. "We had worked with Henry in his previous capacity, when he was head forester for Michigan California Lumber Company," Holmen says. "He had done some initial long-term planning work for that owner. So he was familiar with the process. He basically was already doing work with us on the GIS side.

"Henry is an educated client — he's involved in the process," Holmen says. "Part of the planning process is scientific, and part is political. He wanted to make sure we did the things that would meet his and his company's goals. This dealt with everything from the range of management operations we should consider to reviewing preliminary model information to see how tree growth measured.

"Some of our other clients have been less involved in the early phases," Holmen says. "Because they are less involved in the whole planning process, they care less about what the answer is."

Although Alden was able to relate VESTRA's work to PALCO decision-makers, he says he encountered a lack of OR knowledge from the federal agencies. "The biggest problem we had was gaining understanding and acceptance from the agency side," Alden says. "We brought in a group of independent experts who participated in the process and could witness that we did the rational thing in each decision point."

Alden maintains that a communicator is essential in OR work. "OR companies need somebody who is technically capable and articulate," he says. "If exper#tise doesn't exist at a company, they need someone who can sense missing information and put it into lay terms."

One of the biggest social challenges the modelers faced was trying to incorporate all the parties involved, from the stakeholders to the actual end users. "The place where we're falling apart now is that some of the groups that don't have a direct stake weren't consulted enough," Holmen says. "They're challenging pieces of it because they think we should have done something different or considered something else.

"We let PALCO do most of the decision-making in terms of whom they would pay attention to and what range of options they would consider," he says. "There's no way they could consider all the groups that wanted to have everything done — it's a pretty politically controversial piece of ground. We basically served as the people who did the actual modeling and OR work, and gave them advice on what we thought was the most sensible thing to do and tried to help them interpret what the answers meant."

Ep(x) — A dramatic difference

Although Holmen says he isn't totally satisfied with how the model works (one problem is size limitations for data storage), there have been some notable achievements. On a fiscal level, the model increased annual earnings at PALCO by $22 million a year in 1996 and 1997. On an environmental level, it has generated a more optimal mix of wildlife habitat acres, including spotted-owl nesting habitat.

Fletcher points out two things the model did: "First, it allowed the company to be profitable and to achieve its objectives of satisfying the stockholders and increasing revenue. Secondly, it meets the federal and state regulations or exceeds them and provides for the environment in the long-term sustanability of their forest land."

More specifically, the model helped PALCO to predict forest growth more accurately. As opposed to just a few runs, the model could do hundreds of runs and therefore produce more accurate predictions for decision-makers.

Also, during this process VESTRA discovered that prohibiting harvesting in certain areas (as in the classical approach) was not beneficial in the long run. They determined that light harvesting in these areas actually improved the habitat for wildlife and made the land more valuable.

Holmen says there are also benefits to the users' education. "One benefit of a planning system like this is that it really forces you to focus on what it is you're interested in accomplishing," he says. "That is part of the whole planning problem."

Alden agrees. "This is dramatic in my mind," he says. "This is not as solid as scheduling UPS trucks, but it sends you down a path you otherwise might not have taken. The model has helped in communicating choices in a credible business language to the ultimate decision-makers."

Lessons learned

As Holmen reflects on VESTRA's consultation with PALCO, he says the biggest change he would make in the process would be to get more involvement with the people at the field level. "You can't ask them for advice all the time, but the field staff needs to know what went on behind the scenes," he says. "Otherwise, they just see a plan presented. I would try to get more feedback from them — find out whether or not this kind of planning is meeting their needs and what their questions are. If they don't accept the plan and implement it as we expect them to, then the plan really can't succeed.

"They're the people who are ultimately responsible for laying out the harvest units, doing the management, and tracking the changes to make sure things are on track," Holmen continues. "That's not a simple task. Without their acceptance and involvement in the planning process, it has a much greater chance for failure in terms of actual implementation."

Holmen says he also wishes VESTRA had been given more time to do a prototype on a small area. "We had planned to (do a prototype), and the political situation changed to the point where suddenly, we went from, 'We'll do a little pilot and test this stuff out and get some feedback' to 'Oh my god, we've got to do the whole thing and we've got a very short time to do it in.'

"In PALCO's case, it wasn't too bad, because Henry had worked with some of this planning before, and he had a pretty good idea of what we needed and the quality of data we would be able to get from him and get back to him," he says.

"It's sort of this iterative process," Holmen summarizes. "The easiest way to get that started is to do a prototype; show them the responses of a more-or-less completed plan which they may or may not like, and get feedback on what it is that's good and bad about it. Then, when you do the whole property, you've got a chance of fixing some of the things they see as flaws."

Jenny Summerour is managing editor of OR/MS Today. She can be reached via e-mail at jsummerour@lionhrtpub.com

Table of Contents

OR/MS Today Home Page

Lionheart Publishing, Inc.
506 Roswell Street, Suite 220, Marietta, GA 30060, USA
Phone: 770-431-0867 | Fax: 770-432-6969
E-mail: lpi@lionhrtpub.com
URL: http://www.lionhrtpub.com