Summary of results
Is there an answer to mapping old growth?
An examination of two projects conducted with remote sensing and GIS
Robert A. Norheim
A number of different factors influenced the way in which the
projects were conducted, including the time and budgets allocated
for the projects, the nature of the data used, and how the projects
applied the definition of old growth.
Timeline
Perhaps the most important constraint that affected both projects was the timeline. Both projects were performed quite rapidly. Jerry Franklin called the Pacific Meridian Resources project a "kind of a crash contract" (U.S. House of Representatives 1989: 109). Similarly, Morrison was under intense pressure to perform his work rapidly, as we have seen.
However, the frenzied schedule of both projects was a result of the Forest Service's refusal for many years to perform a comprehensive, reliable, and uniform inventory. Had the agency not simply tried to do the minimum it could get away with to manage for the spotted owl, and had instead taken on the problem enthusiastically from the time that Forsman first identified it (or even somewhat later), it might have avoided the scrutiny it received from both Congress and environmental groups. The institutional nature of the agency was also to let each forest use its own definition for old growth, resulting in the hodgepodge of definitions shown in Table 3-4. As mentioned, due to the agency's focus on providing sawtimber, the agency did not bother to inventory forests in wilderness areas, which could not be cut. Astoundingly, inventories were not even updated to account for cutting that had taken place, never mind fires or other disturbances. Thus, each forest did its own inventory by its own definition and on its own schedule, and not even the whole forest was inventoried. When Congress asked the agency for a single figure as to the amount of old growth, the agency could not supply one that was reliable. Thus, Congress mandated the agency to come up with an answer within a year. The Wilderness Society, tired of the delaying tactics of the agency, decided to do their own inventory.
The short timeline given by Congress forced the Forest Service to go outside the agency and to the technology of satellite remote sensing. Previous inventory work had been done by the agency itself using photo interpretation. Pacific Meridian Resources tried to accommodate what foresters in the agency were used to, by adapting its remote sensing techniques to mimic those of photo interpretation. Pacific Meridian also had to rely on the staff of the agency to assist them with verification of their work. These techniques had varying rates of acceptance by different ranger districts and different forests.
While the delaying tactics of the agency also were a cause for
The Wilderness Society to do its mapping project, it was a different
set of circumstances that forced Morrison to turn to remote sensing.
Clearly he favored photo interpretation, and wished to use satellite
imagery only for what was most easily discernible (recent clearcuts).
However, the events outlined in Chapter One made it clear that
at least decisions were going to finally be made in 1991, by the
agencies, courts, and Congress. The Society wanted to have its
own dataset completed and available to decision makers to inform
them of the Society's position in the conflict. Perhaps another
implicit goal of the Society was to not allow the dataset that
the Forest Service came up with to exist in a vacuum and be accepted
as the only truth. By providing an alternative dataset, questions
would be raised as to the accuracy of the Forest Services' maps.
While there would also be doubts about The Wilderness Society's
maps because of the discrepancy, the Society nevertheless wins
with a draw, by casting doubt on the Forest Service's maps.
Remote sensing
As already mentioned, the choice of using satellite imagery was relatively new to the Forest Service. While the technology had existed since the late 1960s, the first Forest Service conference on remote sensing occurred only in 1986. It is worth examining the fundamental nature of remotely sensed imagery - both photo interpretation and satellite remote sensing have their roots in military technology (Artis 1994; Taylor 1995). Landsat was designed for use in agricultural inventories; it is worth asking whether it is appropriate technology for studying natural forest growth.
On the other hand, when The Wilderness Society turned to using
satellite imagery for the second part of its project, it could
not afford Thematic Mapper (TM) imagery, and was forced to use
the earlier multi-spectral scanner (MSS) technology. TM has several
advantages over MSS, including more bands, greater spatial resolution,
and finer spectral resolution. This is not much of a disadvantage
for Morrison for the four forests studied in this thesis, where
Morrison used the imagery only to determine recent clearcuts.
However, the accuracy of Morrison's mapping of the other nine
Forests is certainly likely to be less than the Pacific Meridian
Resources TM-based dataset. It is interesting to note that both
projects experimented with SPOT panchromatic imagery. Pacific
Meridian did not feel that SPOT data was a cost-effective supplement
to TM imagery. On the other hand, Morrison (who received the data
inexpensively under a cooperative agreement) was quite pleased
with the improvement of the results over MSS.
Geographic scope
The differing geographic scopes might also be attributed to the cultures of the institutions involved. In particular, Morrison did not map two areas of the Mt. Baker-Snoqualmie National Forest that are administered by other forests. A large percentage of both of these areas is in wilderness. Because the Society was interested in Morrison's data for the purpose of advocating for the preservation of more old growth, old growth that was already preserved was not as much of an issue. Similarly, Morrison did not study the National Parks in the region as Congress ordered the Forest Service to do. This makes clear that The Wilderness Society's motivation for the project was to understand where the unprotected old growth was, not how much old growth was left in total, or how much there had been in the past.
Another question that might be raised is why the projects focused on the westside National Forests exclusively. While the change in biogeographic regime certainly is noticeable as one crosses the Cascades, and there is some debate as to whether lodgepole pine forests should ever be classified as old growth, this is not to say that there are no spotted owls and no old growth Douglas-fir forests east of the crest. In the Oregon and southern Washington Cascades, the climatic change is more gradual than in northern Washington due to the lower elevations. This may help explain why Morrison mapped areas of the Gifford Pinchot and Mt. Hood that are east of the defined crest. Yet it is odd that neither project attempted to quantify old growth by any definition on eastside forests.
Interestingly, Congress later asked for another report that gave
the total amount of old growth across all ownerships. This report
includes old growth in all of California, Oregon, and Washington,
including both sides of the Cascades (Bolsinger and Waddell 1993).
Application of the ecological definition of old growth
Both studies professed to be using the same "PNW-447" ecological definition of old growth (Old-Growth Definition Task Group 1986). However, had they interpreted and applied the definition identically, they should have come up with the same acreage figures for and spatial layout of old growth. Clearly, they each interpreted the definition differently. This is not unreasonable considering the different technologies used in the two projects. As mentioned before, some of the characteristics in the definition could not be seen from either the air or space, and both projects had to interpret the definition.
Morrison distinguished what he considered to be true old growth from "other ancient forest", which had many but not all of the attributes in the old growth definition. Pacific Meridian Resources was mapping what it considered to be "potential old growth" (until required to say "old growth" by the Forest Service), a broader category. Thus, one might suppose that all the acreage that Morrison found to be old growth would be similarly classified by Pacific Meridian. However, the agreement rate was only 70 to 76 percent. While this is a majority, it is not the overwhelming majority one would expect.
Similarly, one might expect that Pacific Meridian's "potential" category would have included the stands in Morrison's other ancient forest category. For two forests, the agreement rates are 68 to 72 percent; however, for the other two, the agreement rates are only 35 and 50 percent. Thus, a detailed examination of the data unfortunately provides no insight into the ways in which the definitions used by the two teams compared. It would be interesting to compare the rest of the Oregon forests, to see whether the applications of the definition were more uniform on forests that were classified by remote sensing by both projects.
That the Forest Service required Pacific Meridian to drop the
word "potential" from the map titles is worth bringing
up again here. This is a clear imposition of institutional biases
on what was ostensibly an objective study. It was in the interests
of the Forest Service to identify as much old growth as possible,
and not to muddy the issue with words like "potential".
The more old growth found, the more reliable the agency's earlier
estimates would appear to be and the less creditable the environmental
organization's earlier criticism would be. Similarly, the more
old growth found, the more that would be available for allocation
between ecosystem protection and the agency's traditional timber
interest.
Budget
The resources available to each project certainly had some differing impacts on their methodologies and results. The differences between the TM and MSS imagery that the two projects used have already been discussed. The staffs of the projects were another variable related to the budgets of the project. Pacific Meridian Resources had people lined up to hire the instant the contract was awarded. Morrison had to rely on the efforts of many volunteers and had higher turnover because of the nature of volunteer work. It is impossible to evaluate the actual skill levels of the respective staffs, but the volunteer nature of Morrison's group does raise some questions.
Hardware and software was also an area of difference. Pacific
Meridian Resources used state-of-the-art workstations and software;
for example, they were a beta site for the ERDAS-Arc/INFO "live
link" (Green 1991). Morrison had a donated 80386 PC and cobbled
together various software systems including borrowed time on Washington
State Department of Wildlife computers. On the other hand, the
economics of the hardware and GIS software changed significantly
for everyone in the late 1980s. Such a project could not have
been taken on at virtually any price just a few years earlier.
Institutional agenda
The respective agendas of the institutions that sponsored the two old growth projects affected the way they were conducted.
While there are clearly many differences between The Wilderness Society and the Forest Service, the Society has grown significantly in recent years and has gained more bureaucracy. Thus, one can also draw some similarities between the non-profit and agency; for example, the increasing removal of the Society's leadership from its field staff, its inclination to take fewer risks, and its confusion of the goals of self-perpetuation with its goal of advocacy (Boerner and Kallery 1994).
Morrison wanted to ask many more questions than how much old growth there is. The questions "How much old growth had there been" and "how do natural processes affect the longevity of old growth" are certainly as important in trying to decide how much old growth ecosystem to preserve. Yet the nature of the institutions - not only the Society but also Congress, the agencies, the courts - pressed him to deliver just the simple answer. It is ironic that it is the images from his Olympic report, that show the change in the extent of old growth over time (Morrison 1990: fig. 13, page 19), that are the most powerful and are the ones that are reproduced in the media. Or perhaps it is not ironic, but instead an indication that the question is really not "how much is there?" (the question both studies were trying to ask) but instead is "how much was there, and how much is left?"
The focus of the Forest Service was to get a simple answer to supply to Congress. While they had initiated a pilot project with Pacific Meridian Resources to do vegetation mapping of eastside National Forests, the sudden urgency of the old growth issue made them break up the westside contract into phase one, old growth only, and phase two, all vegetation. Pacific Meridian adapted their procedures to look for only old growth in phase one. Undoubtedly, they would have come up with somewhat different results had they examined the westside forests for all vegetation classes from the outset.
The culture of the Forest Service was also to discredit Morrison. Perhaps he was considered a "traitor" by some of his ex-colleagues. While both Pacific Meridian and Morrison had expressed a wish that the two projects had worked cooperatively, to try to come up with an answer that both could agree on, the Regional Office apparently vetoed any such cooperation. The footnote in the EPA report mentioned above, asking for any interest groups with old growth maps to share them with the Forest Service is certainly amusing given the treatment that Morrison reported he received.