Daniel B. Botkin

Copyright September 18, 2020 Daniel B. Botkin

This is a curious story, because a major environmental problem was solved not by scientists who said they specialized in the study of salmon, and not by the scientists employed by the state’s large bureaucracy who said they were biologists specializing in the study and maintenance of salmon. The problem was solved by a single old-time fisherman named Jim Welter who said he had no formal training in science.

Here is how the story began. Early in the 1990s, Jim Brown, the head of

Oregon’s Department of Forestry, contacted me about a problem the state was having with its salmon.

Oregon was famous for two natural products: six species of salmon that spawned in its rivers and streams; and magnificent forests of large trees of various species valued widely in the United States and other countries. Jim had heard about various projects I had done, funded by various state and federal agencies, to save specific species and solve other environmental problems.

Jim explained to me that there was a serious problem in his state. There was a consensus that the salmon of Oregon’s rivers were declining. In 1991 the Oregon Legislature charged the Oregon Board of Forestry to commission a study to “assign the relative importance of forest practices” to the decline in salmon and to “make recommendations as to how forest practices can assist in recovery of anadromous fish populations”. The state would provide $1 million for that project. Those salmon, famous as an Oregon food supply, were loved within the United States and worldwide. They were also a major source of jobs, income, and state taxes.

This situation became a famous environmental issue across the United States. For example, when he was president starting in 1993, Bill Clinton went to the Pacific Northwest and pronounced a forestry edict stating that in the Pacific Northwest no logging of trees could take place along salmon streams within twice the length of the tallest tree that could grow right along a stream. He was motivated to make this ruling due to pressures and agreements with a small group of environmental scientists very active in the conservation of nature for many years.

Jim asked me to lead this study, and I agree to do it. There had been a set of actions that were believed to be the causes of the salmon decline. These actions were:

  • Too many sea lions in the Pacific Oceans relatively near shore, who were eating the salmon. Overfishing of salmon since they were both a commercial viable product and a famous sportfishery.
  • Destruction of the rivers as habitat for salmon, especially the building of many dams which the salmon could not cross, and removal of downed trees from the rivers, claimed to block the salmon passages, Large scale massive forest logging, because Oregon’s forests had been a major source of some of the best timber in the world for a wide range of products, including building construction and elegant housewares — dishes, furniture, and bowls and other objects considered beautiful.
  • The amount of gravel of just the right size in their natal stream for depositing eggs was being disrupted. Variations in water flow in streams and in ocean currents as affected by storms, floods, landslides.
  • Every salmon fisherman, every naturalist, in fact almost everybody who has ever walked along a salmon stream knows that salmon have ne of the most complex life histories of any species:
  • Born in a freshwater stream, feeding as youngsters on insect larvae and avoiding larger fish that try to eat them, depending on logs and other natural structures in the streams for places to hide from, hiding between gravel pieces and in small channels and on pools and riffles produced by logs that provide good habitat along main waterways until they are large enough to descend the river and enter the Pacific Ocean. They feed on insects and their larvae which are supported by leaves that fall into the stream from nearby trees and shrubs, and by algae that grow in the stream.
  • Once large enough, they swim downstream to the ocean where they live for quite a number of years. Their life in the ocean was relatively little known — what eats them, what they themselves eat. Three, four, or five years later they return up their natal river and reproduce.

Making even a simple inventory of the six species of salmon found in the study area was a complicated effort. Each species has its own preferred habitats and time during which these habitats are used.

Because they can become prey to birds and other predators, migration downstream is dangerous for young salmon. Crossing the zone between fresh and salt water alone is a feat of an osmotic gymnastics as the fish go through an adjustment so that their bodies can maintain cellular integrity in the dramatically different chemical environment of the ocean. Those that survive this transition can live for quite a number of years in the ocean. Their food, habitats, distribution, competitors, and predators are poorly understood.

Those that survive to adulthood then navigate, often thousands of miles, back to the Northwest coast to find their way past numerous river channels to their natal streams. Swimming and leaping upstream, they reach their birthplace, excavate nursery pits in gravel and rubble, lay and fertilize their eggs. Then most them die (or are caught by fishermen and many kinds of wildlife.

Wild Salmon swimming.

Thus salmon have one of the most complex life cycles of any vertebrate animal and depend on more kinds of habitats than most other organisms. Adding to this complexity is that each species of salmon uses different parts of river systems for spawning and rearing and they spend different amounts of time in fresh waters and in the ocean.

Common Beliefs About the History of Salmon and Their Habitats

There had been, in the folklore of Oregon, a set of common beliefs that formed the background for the debate about salmon. These included that, prior to European settlement: there were a superabundance of salmon; forests of western Oregon and northern California were composed of large, ancient trees; this extensive old-growth was essential to the abundance of salmon; the forests and salmon were pretty much at a steady state — constant in abundance and distribution. And finally, it had been thought that Native Americans had little if any effect on the extent and composition of the forests and the abundance of salmon.

LARGE CLEARCUT FORESTED AREA. This photo was taken in Canada, not in Oregon, but represents some of the large logged area also found in Oregon,

(Photo by Daniel B. Botkin)

It was also commonly believed that modern human actions accounted for much of the decline of salmon. These included overfishing, poor forest practices, especially large areas cut of all trees, and especially near streams, and construction of dams. Other commonly mentioned human causes of declines in salmon include channelization of streams, gravel mining of streambeds, hatcheries, predation by birds and marine mammals, road building, and reduction in river flows because of human activities.

How Our Salmon Project Began

Our work seemed straightforward. First establish a small committee of scientists with expertise in the major aspects of what controlled salmon; and then collect data on annual counts of salmon on the major salmon streams and rivers, and then find detailed records of the location, size, and methods of timber harvest that were considered one of the major causes.

I was able to get five top scientists with excellent reputations in the topics essential for us to study. I hired a graduate student working under me at University of California, Santa Barbara, to be my primary assistant, Mark Melancon, who had had lots of experience studying and working as a forest timber harvester I set up a place for him to live in Portland, OR, and rented a very nice apartment in downtown Portland, with an elegant view of the river and the mountains to the east for my wife, Erene and I. We three moved to Portland for the length of our study. I was able to obtain an office and the blessings of the Oregon Forest Preserve for our project, adding to the stature of what we would do.


Once settled in Portland, I began by asking my assistant Mark to contact the primary offices of Oregon’s fishery’s department and ask them to send us the counts by year of the salmon harvest and salmon numbers moving up each of the major salmon rivers; And similarly, I asked Mark to contact the Oregon Forestry Department and ask about the area, harvest methods, and dates of the major logging cuts in the watersheds used heavily by salmon.

He asked for the salmon data. But he was oddly always being put off. After numerous inquiries we came to understand there were very few records of salmon counts ever gathered nor kept.

Problem # 1. The first very thing I was asked to use in my analysis did not exist.

Mark and I went down to the fisheries office to find out if this was true and if so, what had been measured. Throughout our project, we talked many times with the fisheries experts at the Oregon Department of Fisheries. They often took us out to where they were doing a new study, asking our opinions.               None of them knew much about statistical analysis of data nor did they ask many questions about that. Also, we discovered that they did not really understand much about the scientific method. In general, throughout the history of the department, if they did an experiment about the structure of a river, they never used a treated and an untreated stream for comparisons.

Oregon’s Rogue River, one of the nation’s most famous salmon fishing rivers, showing the forests coming down to the river. 

(Photo by Daniel B. Botkin)


Step number 2: Look for a map of Oregon’s forests. Find out where and when large

forest areas were cut that overlapped salmon streams.
I contacted Jim Brown who as I have mentioned was the head of the Oregon Forestry Agency. I told Jim we also needed a map of the state’s forests showing where and when large areas had been cut. He did not know of any such map. He explained that logging records were kept by counties in Oregon. But these records did not include the geographic location, size of the cut, method of harvest. Plus, those records were discarded after several years. As a result, it was not possible to determine the specific history of forest practices in any major watershed.
Records of forest operations notifications were, however, kept by the Oregon Department of Fisheries.  

 Data related to operations beginning in 1979 to the present were available. They included geographic location, size of operation, and method of harvest. However, geographic location information was not adequate to relate these operations to the movement of salmon Douglas Fir old Growth in Oregon (photo by Daniel B. Botkin)

The Oregon Department of Fisheries thought that it should be possible, with considerable effort, to determine the history of timber harvest in many watersheds since 1950, through the use of aerial photo  analyses or  other. Indeed, eventually, A map of all land use conditions was later produced by the Oregon Department of Forestry using remote sensing imagery. In fact, we found a 1988 map based on Landsat remote sensing data, each pixel represents a square area of 320 acres. These data were rather coarse, but they are all that was known to be available.

We also hoped to get records of the numbers of fish released by hatcheries, but these too  were difficult  to obtain, had gaps in  them,  and  were  maintained  for different periods of time. Thus, the sparce data allowed us to provide only broad options for public comparison. However, we also found that little attention had been paid to the regional data that were available, and that therefore new, useful findings could be obtained by objective analyses of these data.

An Early 20th Century Map of Oregon’s Forest is discovered

Several years later, when we were still heavily involved in this study, we got a call from the Oregon Department of Forestry telling us that a very old map that had been produced about 1910 and had been discovered still in existence. It turned out that the

then State Forester believed a map of the state’s forests and their recent harvest was important. He was said to have spent something like two-thirds of the state’s forestry budget for that year on having the map made. But after he retired, nobody knew what to do with the map, it was so large. So it got rolled up and put in a trash receptacle in one of the men’s rooms. A few years later, one of the employees decided the map had no use and stuck it outside with the day’s trash.

However, by good luck, the forestry employee on night duty saw the map in the garbage and decided it was too important to throw out. He withdrew it from the trash and brought it the next day to the office of the then state forester. It had been stored but forgotten about again until we turned up, and someone remembered about this map.

Miraculously, he was able to find it and, to our good fortune, he gave it to us.

World War II Tank Batteries Use Oregon Trees, and many logs lying in Salmon Rivers are taken away

We took many trips to different rivers and forests in western Oregon. On one, we came to a stream that had no dead tree logs in it. We had thought these dead trees were a common feature of Oregon’s streams, when a tree died and fell into a stream, nobody bothered to remove it. The Fisheries scientist taking us on the tour said that two things led to the removal of dead trees from of the salmon streams. First, during World War II, the American tanks had large electric batteries and some species of Oregon’s western forests made the best containers for these batteries — the wood did not conduct electricity and the toughness of the wood and bark keep the batteries in working order for a long time.

During the same time and later, those in charge of the salmon streams decided that fallen logs created pools, where the water moved slowly downstream, and ripples — fast water areas. They decided that these ripple areas slowed the salmon, both those swimming upstream to reproduce and the young salmon swimming downstream. With the military demand for western Oregon logs, it seemed reasonable to remove most of these dead logs from the stream, under the belief that doing so would improve salmon travel and reproduction.

However, instead of treating this as a standard scientific experiment, which would have begun by taking pairs of streams, measuring salmon migrating in each, and then removing the logs from one. Then a comparison could be made to see which stream was most productive of salmon. Instead, they just cleared fallen trees from one river and did not count the salmon in another river not so cleared.

Oregon Salmon River with the Forest lands (Photo by Daniel B. Botkin)

Instead, the Oregon Department of Fisheries just cleared many streams without doing any research tests. They were unfortunately unaware of scientific methods which could have been employed had the project been scientifically driven

However useful the wood was for American World War II tanks; the result was a great reduction in salmon production in the cleared streams. It turned out that the dead logs created pools and rapids. This was well known to all salmon sport and commercial fishermen Those rapids carried insects rapidly down into the pools, and the quiet waters of the pools allowed adult salmon to stay near the downstream edge of a rapid and catch insects with relatively little energy use on their part. Thus, the fallen logs created a much better habitat for the salmon. But the removal was already done, and that was that.

We suggested that the current salmon scientists could do a double-blind test right then, to see how effective the logs were: take a stream and put a lot of logs into in and compare it to another stream where that was not done. Months later, these salmon scientists proudly took us to see one of the streams in which they had done exactly what we had suggested. I asked where was the partner stream into which they had not put dead logs and then counted the salmon produced there. They admitted that they had not bothered to pick a parallel stream, counted the salmon produced in that one as well

in order to compare with the treated one. Here we were half a century later and these “scientists” had still not come to understand one of the basics of empirical tests about nature. They seemed surprised at our commentary about that aspect of scientific method. Once again, another discouraging note had to be made in our research diary.

Public Meetings and an old time Fisherman Figures out the correct analysis to be done.

At the beginning of my work on this project, I told the heads of all the various state agencies that were involved with salmon that my plan included holding public meetings. A strong believer in democracy, I thought that the people should have their say as well as technical experts. The government bureaucrats tried to talk me out of this, saying that this would be crazy, that some nutcases would get up and never stop talking and say useless things.  Clearly, they were afraid of what the public might bring up and did not like having the public interfering with their approach to their work. I replied that I had a solution to such possibilities: any citizen could apply to speak at any of the public meetings, but each would have to submit a written statement, that would be read by members of my committee to decide if the suggested statement was valuable enough to be heard.

These public meetings became very popular, and I quickly learned that one of the reasons state department heads did not want to let the public speak was because many who did speak brought up topics which the state employees had not mentioned. One person, for example, said that while all the focus was on the effect of forests on salmon the salmon only reached forested areas by swimming first past large agricultural tracts which were next to the ocean shore. Therefore, the effect of those farming methods should be considered.

This seemed completely realistic. Afterwards I asked one of the heads of Oregon State Departments why they had never brought this up. The one most closely involved with agricultural practices said in reply “I guess we’ve just been too clever to keep this subject off the table.”

One of our public meetings was held in Gold Beach Oregon, a town doing well because it was at the mouth of the Rogue River, one of the nation’s most famous salmon streams. We scheduled the meeting for daytime. But the fishermen leaders called to say they all worked during the day and could not attend a daytime meeting. At my suggestion two meetings were set up. One for daytime and one in the evening

As it happened the afternoon meeting was filled, and almost all were commercial salmon fishermen and fishing guides. When I opened the meeting, all these fishermen sat with their hands folded or else in postures that meant they were hostile. Then the first question was asked: “Professor Botkin, do you believe that Oregon’s salmon are in trouble and something has to be done about it?” I replied “I’m just a professor of biology at U C Santa Barbara and I don’t know much of anything about salmon. I was just asked to run this study and I have no opinion.”

Immediately, all the audience relaxed. They must have concluded here was somebody in charge of a major study who was totally open minded.

An older man stood up. He said his name was Jim Welter. “I’m eighty years old and been salmon fishing all my life, and don’t know anything about science. But it just makes sense that if these salmon are born and reared in freshwater streams and spend about a year there, and then go to the ocean and return when they’re three or four, that the amount of water flowing in the stream where they were born ought to make a big difference in how many survive and return.”

That made a lot of sense to me, and it was refreshing to hear something constructive, especially when I had only recently learned that the Bonneville Power Administration, which built and ran the big dams on the Columbia and Snake rivers, had spent $2.5 billion on salmon research and restoration and, according to one of their top executives who spoke to me, those dollars hadn’t yielded a single sign of improvement in the salmon. How could a big agency spend that much money and have absolutely nothing to show for it? I wondered. Jim Welter, the 80-year-old salmon fisherman, who figured out how to forecast the size of a salmon run 4 years in advance.

The best science done by anyone living and working in the state of Oregon

Jim Welter explained that on his own he had gone to the state of Oregon’s Department of Fish and Game and got the data for the counts of salmon crossing a dam on the Rogue and the Umpqua rivers, the only two rivers where the state actually counted salmon.

Then Jim went to the U.S. Geological Survey and got annual water flow data on

these same two rivers. If you haven’t had to deal with state level agencies of the kind we were working with in Oregon, you have to understand that what Jim did took a lot of effort and time.

Next, a young man who had come with Jim, came forward with a very large rolled up paper, which they fastened to the top of the blackboard behind me in the meeting room. He had graphed water flow four years before and the number of returning salmon up these two rivers in the present year. Definitely, these two graphs overlapped. On my staff was a professor of forestry at Rutgers University, who taught statistics.  He looked at the graphs and was very impressed. He and Jim had a long talk, and subsequently he was given Jim’s data. The two worked together on it. Indeed, there was a strong statistical correlation between water flow 4 years before and the present year’s salmon abundance. We did elaborate statistical tests that demonstrated that 80% of the

variation in salmon numbers could be accounted for by water flow alone. Just introducing a little of the ambient environmental factors into the forecasting methods could make a tremendous difference. We published this as a scientific paper, but by the time I had finished running my project, I knew of no other salmon scientist who had paid attention to our finding, nor of any scientific publication other than ours that discussed this finding. This was the case even though newspapers published very positive articles about our findings.

Jim Welter’s analysis became our best and reliable forecasting tool. This was the best, in fact the only, statistical analysis anybody ever presented to my committee. The so-called fisheries experts on the state’s Department of Fisheries had never thought in these terms. Jim and I became life-long friends, talking and writing each other not just about fisheries but life in general, and how our common friends were doing. This is the little-known story that demonstrates one of the basic problems that troubles our attempts to conserve our living resources. Those in charge often do not have any training in statistical analysis nor in any methods of analysis wildlife changes in abundance. 

And the salmon are telling us, time after time, that we must change these methods.

(Analyses reported here were spawned by the testimony of Mr. Jim Welter of Brookings, Oregon at CSE’s January 1993 public hearing in Gold Beach, Oregon. Mr. 

Welter conjectured that spring chinook escapement would be significantly explained by water flow three years earlier, and he exhibited supporting graphs. He observed that such relationships, if confirmed, might yield useful forecasts three years in advance of escapements)