Computers and the Crisis in Our Forests

COMPUTERS PROVE MORE RELIABLE THAN SQUIRRELS IN UNEARTHING WOODLAND SECRETS Designing a Forest

David Patton secured the leash on his squirrel, then gently set the furry-tailed, big-eyed, keen-nosed rodent on the fragrant forest floor just outside Flagstaff. He stepped back and waited expectantly, hoping that months of painstaking laboratory training would pay off, and his squirrel would take to his assigned task as expert truffle-finder.

Patton, a Northern Arizona University forestry professor who had studied the role of such squirrels in the ponderosa pine ecosystem for years, wanted to use his trained squirrel to probe links between the rodents, the fungi that colonize the roots of the trees, and the health of the lordly pines themselves. He spent nearly a decade gathering every scrap of information he could find about these brash tree squirrels, using the data to devise a computer modeling program that could be used to predict future squirrel populations.

The squirrel at the end of the leash gazed around the forest with big brown eyes, his alert little face a cartoonist's delight. He trembled. He squeaked. He ran to the end of the leash, halting with a small jerk. Then he ran back and forth across the forest floor, a dance of tremulous excitement circumscribed by the thin tether that connected him to laboratory computers.

Patton sighed and shook his head ruefully. He stooped, unsnapped the leash, and watched months of quixotic effort scurry frantically across the forest floor for the huge corrugated, vanilla-scented bole of the nearest ponderosa pine.

Recalling the moment, Patton shrugs. "The little fellow was so excited, he wasn't going to be looking for mushrooms," he explains. "Besides, I couldn't take it." Some things, after all, don't fit easily into a computer program.

Patton had hoped to use his trained squirrel to refine the computer models designed to help biologists manage the complex interactions of plants and animals in the forest.

Abert's squirrels consume the bulk of each year's pinecone crop, nibble the growing ends of pine branches through the cold winters, and hunt out the nutritious truffles and mushrooms growing on the roots or bark of the giant pines. This fungus sucks nutrients out of the tree roots, but more than compensates by increasing the roots' uptake of essential nutrients and water.

The squirrels survive on the mushrooms during key times of the year but return the favor by spreading the spores of the fungus from tree to tree during their foraging. Therefore the squirrels' search for truffles helps the ponderosa pines survive in a forest that offers barely enough moisture to sustain them.

It's that kind of complex interaction that prompted Patton to train a truffle-sniffing squirrel and has spurred an interdisciplinary team of scientists at NAU to develop and refine computerized forest growth models, augmenting the vast effort by state and federal agencies charged with managing Arizona's forests and wildlife.

The growing computer data bases and the flood of data gathered by ceaselessly orbiting satellites have at last provided scientists with a way to predict the future of forestlands.

"The computer is the only way we have to look at the long-term stuff," said Rick Miller, a forest habitat specialist for the Arizona Game and Fish Department. "It allows us to look at a time frame that we otherwise can't comprehend. But it's got to be used with a lot of caution since even a fivepercent error rate compounded for a 200-year computer simulation adds up to a very large difference in the final result." Flagstaff-based Roger Zanotto, timber staff officer for the Coconino National Forest, noted, "Computers have become the essential tool in forest management." The accuracy and sophistication of those computer models have grown dramatically in the past decade, although computer modeling remains hindered by the lack of detailed data on widely dispersed test plots needed to check computer predictions against reality. In addition, the lack of information about many wildlife species limits the use of computers in predicting the interlocking effects of road building, logging, and cattle grazing on individual wildlife species. Nonetheless, the researchers at NAU and elsewhere hope that the burgeoning use of computers to tackle ever more complex and delicate ecological questions will eventually change the whole emphasis in the management of public lands. Land managers for decades have mostly focused on questions like how much timber can they cut in a given area, and how many cattle can graze without wiping out elk herds. Getting answers to those questions required a lot of guess work, time-consuming calculations, and straightforward economic projections. But the growing mass of information in

THE SQUIRRELS SURVIVE ON THE MUSHROOMS BUT RETURN THE FAVOR BY SPREADING THE SPORES OF THE FUNGUS FROM TREE TO TREE.

data bases and the increasing sophistication of the computer programs should allow scientists to begin asking much more fundamental questions, like what sort of forests we'll want in 50 years, and how decisions made today will affect that future forest. "Basically what we're doing is simulating the growth of the forest under different kinds of management," said Brent Wood, an NAU professor of forestry. "We can forecast what will happen to a given stand of trees or to thousands of acres and determine what its characteristics will be like in 40 years. "Today people ask how many trees they can cut on a given plot and still have goshawks in the forest. But we want people to ask what sort of forest do we want in 50 years; what should it look like; what uses will predominate. Then we can use the models to find the best path to get to that legacy forest." It's a task of daunting complexity, as the ecological parable of the mushroom-eating tree squirrel demonstrates. For instance, the state's rare low-elevation ponderosa pine forests harbor some 250 different species of wildlife and an amazing array of plants and insects. The ponderosa pines dominate the ecosystem, sprouting in clearings, taking on the distinctive yellow-orange bark of maturity at about 150 years of age, and living for up to 800 years before dying and standing for decades as a snag and another century decaying slowly on the forest floor. These trees evolved in a world where wildfires patrolled the forest floor, burning off dead wood and saplings and leaving a forest dominated by widely spaced mature trees so thick-barked and high-branched they were virtually fireproof. All other parts of the forest ecosystem from the birds that nested in the rotted hearts of the snags to data bases and the increasing sophistication of the computer programs should allow scientists to begin asking much more fundamental questions, like what sort of forests we'll want in 50 years, and how decisions made today will affect that future forest. "Basically what we're doing is simulating the growth of the forest under different kinds of management," said Brent Wood, an NAU professor of forestry. "We can forecast what will happen to a given stand of trees or to thousands of acres and determine what its characteristics will be like in 40 years. "Today people ask how many trees they can cut on a given plot and still have goshawks in the forest. But we want people to ask what sort of forest do we want in 50 years; what should it look like; what uses will predominate. Then we can use the models to find the best path to get to that legacy forest." It's a task of daunting complexity, as the ecological parable of the mushroom-eating tree squirrel demonstrates. For instance, the state's rare low-elevation ponderosa pine forests harbor some 250 different species of wildlife and an amazing array of plants and insects. The ponderosa pines dominate the ecosystem, sprouting in clearings, taking on the distinctive yellow-orange bark of maturity at about 150 years of age, and living for up to 800 years before dying and standing for decades as a snag and another century decaying slowly on the forest floor. These trees evolved in a world where wildfires patrolled the forest floor, burning off dead wood and saplings and leaving a forest dominated by widely spaced mature trees so thick-barked and high-branched they were virtually fireproof. All other parts of the forest ecosystem from the birds that nested in the rotted hearts of the snags to

PONDEROSA PINES MATURE AT ABOUT 150 YEARS OF AGE AND LIVE FOR UP TO 800 YEARS BEFORE DYING. THEN THEY STAND FOR DECADES AS SNAGS AND FINALLY SPEND ANOTHER CENTURY DECAYING ON THE FOREST FLOOR.

the elk that grazed in the grass carpeting the forest floor adjusted to that system.

Dramatic changes followed in the wake of loggers, ranchers, and fire fighters. Forest managers for decades used the forests to provide as much lumber and beef as possible, gradually transforming the forest into grassless thickets of young trees with only a handful of the giants that once dominated the landscape. Some unexpected problems have emerged as a result, underscoring the pitfalls of trying to manage something as complex as an ecosystem.

Years of fire suppression protected saplings and allowed the buildup of dead wood on many acres of the forest floor, creating a tinderbox forest threatened by crown fires that can climb the ladder of saplings to the tops of the once-fireproof mature trees.

Years of cutting down and hauling away snags as potential fire hazards have robbed the forest of what proved to be an essential wildlife habitat because many of the birds and bats that help control insect infestations depend on holes in the snags for nesting.

Expensive efforts to control mistletoe infestations that sap the strength of trees proved a losing proposition when biologists realized that the close-packed stands of trees in the modern logged forest made such infestations inevitable.

Time after time, the human actions proved to be stones plunked into an ecological pond sending ripples outward with unexpected speed and intensity.

The use of computers to predict the forest's future is intended to head off those kinds of ecological complications. But building a good computer model can be a tough problem in a forest where tree squirrels both consume pinecones and help water roots. Fortunately agencies like the Forest Service have begun constructing massive data bases that provide ready access to the information needed to design a forest. The Forest Service's Geographic Information System (GIS) includes measurements of tree growth and density across millions of acres of public land.

Information flows into the system from a variety of sources. Scientists maintain test plots to provide the real-world measurements that generate the equations simulating tree growth with allowances for rainfall, soil type, tree density, and a host of other factors. Measurements taken by orbiting satellite can also be fed into the system, offering precise data on whether the trees crowd together so closely that they shade the forest floor, where ponderosa pine gives way to thicker stands of spruce and fir at higher elevations, and the types of vegetation growing in open areas within the forest.

Studies by biologists offer estimates regarding the flow of water in streams, the nesting preferences of goshawks, the moisture in the soil, the growth of elk herds, the number of woodpeckers per acre. Tallies by logging companies track tree growth and stand densities.

Ranchers provide estimates of grass growing on the forest floor, water sources, and even activities of mountain lions, coyotes, and bears.

All of this information flows into the whirring brains of number-hungry computers, which can pull together a steady stream of maps, statistics, and projections linked to any given patch of ground. If a timber company wants to cut trees, the computer can rapidly pinpoint potential impacts as diverse as the likely effect on spotted owls and the amount of mud that will wash into streams from logging roads. The computer can spit out maps showing the wildlife species affected, the watersheds impacted, and the likely spin-offs in neighboring parcels.

"Back in the '80s, we couldn't paint a visual picture of how all these things would interact," said the Coconino's Zanotto. "We would lay out a big map, then do an overlay of a soil survey, then another overlay of a hydrologic survey. And by the time you get about four of these overlays on top of each other, nothing makes sense. All of a sudden, you've got a can of worms in front of you."

The computer programs designed by scientists like Patton and his colleagues at NAU take the next step, predicting how tree density, squirrels, rainfall, erosion, tree growth, grass cover, and hundreds of other variables will interact in the course of the 100 or 200 years it would take a forest to return to presettlement conditions.

That makes it possible to glimpse the answers to crucial questions.

If you want 20 percent of the forest to return to stands of virtually fireproof oldgrowth trees by the year 2050, how many trees of what size can you cut each year for the next 50 years? If you want to build in a 90-percent certainty that old-growth dependent species like goshawks and spotted owls will still be with us in 100 years, how much will you have to reduce the timber harvest? If you want to increase the elk

IF YOU WANT 20 PERCENT OF THE FOREST TO RETURN TO STANDS OF VIRTUALLY FIREPROOF OLD-GROWTH TREES BY THE YEAR 2050. HOW MANY TREES OF WHAT SIZE CAN YOU CUT EACH YEAR FOR THE NEXT 50 YEARS?

THE FOLLOWING MAPS ARE CURRENTLY ACTIVE

TOTAL ITEMS 484

herds by 20 percent in the course of the next 30 years, how many cattle can you afford to have chomping on the same grass? If you want to encourage the timber industry to build new sawmills able to use smaller trees that now crowd public lands, how many of those smaller trees can you guarantee them annually for the next 50 years without wiping out the spotted owls? Of course even the best models can still provide only general estimates. Mostly that's because the data to be analyzed is not as comprehensive as it needs to be. Tree growth can vary dramatically from one slope to another, depending on soil, rainfall, sunlight, drainage, and even soil bacteria. The computer projections can be accurate only if scientists have studied enough real pieces of the forest to keep the equations connected to the real world. What's worse, fundamental questions remain unanswered regarding most insects and animals in the forest. How heavily do goshawks depend on pristine old-growth forest? Why do outbreaks of pine bark beetles that can lay waste to thousands of acres at a time come in 20-year cycles? Why do most forest birds spend most of their time on the mature pines and snags when they're available? What role do the little feathered brown creepers that nest in snags play in controlling insect outbreaks? The answers to all of these questions could dramatically affect the future of the forest and the accuracy of computer estimates. That's why scientists like Dave Patton continue to tromp through the forest, extracting cores to measure tree growth and undertaking odd projects like using a squirrel to hunt truffles. With a little luck and a lot of work, perhaps they'll help ensure there's a forest healthy enough to make a squirrel's nose quiver a century hence.