Increasingly, foresters are using alter-native silviculture techniques to en-hance habitat, emulate natural dis-

turbances, and address social values.

Considering the scenic beauty of a for-est is one important facet of contem-porary management. A baseline studyin 1990 on the Oregon State Univer-

sity research forest assessed scenic rat-ings for six silviculture treatmentswithin two years of treatment on theharvested stands (Brunson and Shelby1992). Comparisons were made forold-growth, traditional treatments(clearcutting and thinning), and non-traditional treatments (patch cut, snagretention, and two-story). That studyfound scenic ratings were significantly

30 Journal of Forestry • March 2003

Bo Shelby, Jonathan Thompson, Mark Brunson, and Rebecca Johnson

Scenic beauty is one attribute that foresters consider when planning a silviculture treatment.To better understand how scenic quality changes in the years following a harvest, six sites onOregon State University’s research forest were evaluated annually for 11 years. Regressionanalysis was used to describe the change in average ratings over time. Initially there were dif-ferences between sites, but as stand development continued those differences diminishedsubstantially.

Keywords: aesthetics; nontraditional forestry; public perceptions

AB

ST

RA

CT

Above: This old-growth stand in the OregonState University research forest has changed little in the years since this photo was taken in 1990.

Mar

k B

runs

on

Changes in Scenic Quality after HarvestA Decade of Ratings for Six Silviculture Treatments

31March 2003 • Journal of Forestry

higher for nontraditional treatmentsthan for traditional treatments (Brun-son and Shelby 1992). The present ar-ticle describes a followup study ofthose same sites from 1990 to 2000.Two questions were examined:

1. More than a decade after harvest,how do evaluations of the scenic qualityof the silviculture treatments compare?

2. How do scenic evaluationschange through the first 11 years of re-generation?

There is no clear agreement withinthe forestry community about themost appropriate terminology to de-scribe these nontraditional silviculturetreatments. Brunson and Shelby(1992) settled on “new forestry” but

expressed reservations about the term.For the purposes of this article, “non-traditional” will be used as a more gen-eral term to describe the two-story,snag retention, and patch cut sites.These treatments are intended to moreclosely emulate natural disturbanceand emphasize the retention of livetrees, snags, and logging slash forwildlife habitat. The sites are alsoplaced in the landscape in a mannerthat reduces habitat fragmentation andproduces a multi-aged forest. This con-trasts with even-aged forestry, whichhistorically has been the dominantmethod of harvesting in the PacificNorthwest. The clearcut and thinningtreatments in this study will be referredto as “traditional” treatments.

MethodsFollowing the same protocols devel-

oped for the 1990 study (Brunson andShelby 1992), scenic evaluations wereobtained at six sites in the McDonaldResearch Forest near Corvallis, Ore-gon. The sites consisted of one old-growth Douglas-fir (Pseudotsuga men-

ziesii) stand (age 250 to 300) and fiveother stands that had been harvested in1989 and 1990. The traditional treat-ments included a 45-acre clearcut anda thinned stand with 30- to 40-year-old residual trees at a density of ap-proximately 100 trees per acre. Thenontraditional treatments included ahalf-acre patch cut in a 20-acre matrixof mature Douglas-fir (age ~100 years)with approximately one-third of thevolume removed, a snag retention cutwith 1.5 large saw-topped snags peracre, and a two-story stand with eightto 10 overstory trees left per acre.

Logging debris was left exceptwhere replanting required its removal.Hiking trails or skid trails crossed allsites except the snag-retention cut. Allsites were replanted within 18 monthsof harvest. Table 1 describes the sites inmore detail.

Each October from 1990 to 2000, agroup of students enrolled in a junior-level wildland recreation class at Ore-gon State University were taken to theresearch forest. Enrollment rangedfrom 37 to 67 students. Other studies

Table 1. Comparison of silviculture treatments in an Oregon State University research forest.

Harvest type Stand type Stand description

None Old-growth Dominated by Douglas-fir approximately 250–300years old, with diverse maple, grand fir, and Pacificyew understory.

Nontraditional Patch cut Half-acre cuts with all trees removed, scatteredthroughout a 20-acre unit of mature Douglas-fir forest (age 100) with about one-third of the total volume removed.

Two-story 21 acres, Douglas-fir dominated, about two-thirds ofvolume removed, 8–10 scattered overstory trees peracre retained. Similar to shelterwood in appearance,but prescription does not call for removing overstoryafter regeneration establishment.

Snag retention 17-acre clearcut with 1.5 large (> 30″dbh) Douglas-fir snags (saw-topped at ~70 feet) retained per acreas wildlife trees. Tree tops were left where they fellfor habitat.

Traditional Thinning 8 acres, Douglas-fir dominated plantation, thinnedto approximately 100 trees per acre. Residual treesare 30–40 years old, with understory of sparseherbs and shrubs.

Clearcut 45 acres, all trees removed.

NOTES: All harvested sites except thinning received herbicide site preparation and were replantedto approximately 200 trees per acre; the clearcut, snag retention, and two-story sites have denserstocking due to natural regeneration. All harvested sites were cut between winter 1989 and winter1990.

have found that college students’ scenicassessments are similar to those of thepublic, both generally (Daniel andBoster 1976; Anderson and Schroeder1983) and at these specific sites (Brun-son 1991; Brunson and Reiter 1996).Instructions were given to the respon-dents explaining the purpose of thestudy and directing them to respondonly for the stand of interest. If theyhad additional questions about the sitesthey were asked to wait until after thesurveys for all sites were completed. Onthe self-administered questionnaire,students were asked to judge the “scenicquality of this location.” Responseswere on a nine-point acceptability scale,with –4 being the most unacceptable,zero being “neutral,” and +4 being themost acceptable. Silviculture treatmentsand past evaluations were discussed in afollowup class session.

FindingsTable 2 shows the mean scenic rat-

ings for each site from 1990 to 2000.Positive ratings indicate that, on aver-age, the site was rated “acceptable,” andnegative scores indicate an “unaccept-able” rating. To allow us to explore thechanges in evaluations from 1990 to2000, a regression line was fitted to theaverage ratings for each site over time(fig. 1). A linear equation reasonablyrepresents the relationship betweentime and average scenic ratings for all

the harvested sites except the patch cut(r 2 = .42–.82).

For the patch cut, average ratingsimproved for the first six years and de-clined for the last five, so the relation-ship is best described by a quadraticequation (r 2= .76). There was no sig-nificant difference between the 1990and 2000 average ratings (one-sided pvalue = 0.39).

For all years, the old-growth site re-ceived the highest average rating, andthere was no significant change in aver-age ratings between 1990 and 2000(one-sided p value = 0.158). The re-gression lines for the two-story (slope =0.12, r 2 = 0.56), thinned (slope = 0.18,r 2 = 0.84), and snag retention (slope =0.19, r 2 = 0.43) stands were fairly sim-ilar to each other; all showed signifi-cant improvement between 1990 and2000 (one-sided p values all below0.03). The clearcut had the lowest eval-uation in 1990, showed the greatestimprovement (slope = 0.32, r 2 = 0.82),and improved significantly from 1990to 2000 (one-sided p value <0.001).

Figure 2 shows the percentage of re-spondents who rated each stand “accept-able” (+1 or higher), which may be of in-terest from a policy point of view. A har-vest method might be defined as meetingthe public’s standards if judged accept-able by some proportion of the public;for the purposes of this discussion, a sim-ple majority (>50 percent) is used.

During the initial years after har-vest, the old-growth site and the threenontraditional sites all met this stan-dard, whereas the traditional sites didnot. In 2000, when the sites had 11 to13 years to recover, all treatments metthe standard and the distinction be-tween the sites decreased. The mostnotable improvement was in theclearcut site; only 21 percent of re-spondents rated it acceptable in 1990,compared to 68 percent in 2000.

DiscussionPrior studies have shown that scenic

quality is at its lowest in the time pe-riod directly after a clearcut or thinning(Hull and Buhyoff 1986). But little re-search has been done to investigate therate of recovery on the types of nontra-ditional harvests discussed here. How-ever, low scenic quality has been asso-ciated with small stems, large amountsof downed wood, and evidence of me-chanical disturbance (Benson and Ull-rich 1981). Because these features arepresent on all recently harvested sites,it was not surprising to see that ratingshad generally improved more than adecade after harvest. This was enoughtime to allow the more obvious evi-dence of logging operations to fade ascoarse woody debris began to decay,understory plants obscured stumps andslash, and replanted Douglas-firs beganto thrive. This may also explain why

32 Journal of Forestry • March 2003

Table 2. Mean scenic-quality ratings for different stands.

Unmanaged Traditional Nontraditional

Year n Old-growth Clearcut Thinning Patch cut Snag retention Two-story

1990 42 +2.83a –1.69b –0.40c +1.07d +0.48d +0.93d

1991 49 +2.58a –1.24b –0.61bc +1.71d –0.71bce +0.12ce

1992 67 +3.12a –1.84b +0.20c +1.35d –1.20be +0.18c

1993 62 +2.92a –1.62b +1.23c +1.68d –0.11e +0.47e

1994 43 +2.95a –0.67b +0.67c +2.12ad –1.19be +0.23c

1995 53 +2.77a +0.10b +0.75bc +2.15a +1.23cd +0.83cd

1996 48 +3.02a –0.31b +1.23c +2.00cd –0.09b +1.10c

1997 54 +3.02a +0.44b +1.17bc +1.56c +0.77bc +0.79bc

1998 37 +2.86a +1.68b +1.73b N/A +0.95b +1.27b

1999 41 +3.25a +0.80b +1.05b +1.49b +0.93b +1.05b

2000 41 +3.05a +0.80b +1.44bc +0.95bc +1.38bc +1.85c

r 0.54 0.91 0.84 0.87 0.65 0.75r 2 29% 82% 70% 76% 43% 56%Slope* +0.02 +0.32 +0.18 N/A +0.19 +0.12p value 0.086 <0.001 0.001 <0.01 0.029 0.007a,b,c,d,e,f Ratings with different superscripts are significantly different within rows, using the LSD multiple comparison test.* Slope is equal to the annual rate of change in the evaluation.

March 2003 • Journal of Forestry 33

the clearcut had the highest rate of im-provement, given that it showed thegreatest impact immediately after har-vest and had the most rapid regrowthbecause of aggressive site preparationand lack of overstory competition.

The old-growth site was the high-est-rated and had the least variabilitythroughout the study. These resultswere expected, because a decade ofgrowth makes little difference in theappearance of 200- to 300-year-oldtrees.

The clearcut site is now a denseplantation of young Douglas-firs. Al-though it is among the lowest-ratedsites each year, its rating has improvedsteadily throughout the study. Will theratings continue to improve? Theyoung trees are now reaching 12 to 15feet in height, and in future years theview of the stand (from the road at thetop of the slope) will be obstructed bythe new growth. There is evidence that,beyond some moderate point, increas-ing density is associated with a decreasein scenic beauty—in effect, the near-view vistas become less interesting(Tahvanainen et al. 1996). The nextdecade of research will show if thescenic ratings are affected by the ob-struction and the corresponding lack of“visual penetrability.”

The thinned site has also improvedover time and has had positive ratingsfor the past nine years. This findingsupports other research showing thathigh scenic beauty is associated withmature trees and low to moderatestand densities (Ribe 1989). Ratingswill likely continue to improve over thenext decade as the trees grow older andthe visual evidence of harvesting di-minishes.

The patch cut is the only site whosescores over time cannot be character-ized by a straight linear function. Wehypothesize the initial improvement(years 1–6) is due to the reduction inevidence of logging accompanied by anincrease in forb cover. After year 6,however, the brush in the understorygrew up enough to look “messy” andcaused a decrease in ratings. Ribe(1991) has shown that other factors,such as the character of nondominantvegetation (forbs versus shrubs) andwoody material, may complicate a sim-

ple linear relationship. We did not col-lect data on individual characteristicsof the site, so it is not possible to spec-ify any effect these variables may havehad on the average ratings.

The score for the snag retention siteincreased slowly through the study pe-riod. Scores might be raised further ifthe snags were topped in a more “nat-ural-appearing” manner. For example,dynamite-topped trees have a morejagged appearance and look more likewind-snapped trees than do saw-

topped trees. Research suggests that theratings would also improve if respon-dents were informed about the wildlifebenefits of the snags (Brunson and Re-iter 1996).

The most remarkable result of thisstudy may be the tendency of scenicratings for harvested sites to becomesimilar over time. Immediately afterharvest in 1990, the scenic evaluationsfor the harvested sites were quite dif-ferent, with the traditional treatments(particularly the clearcut) scoring

Figure 2. Percentage of respondents who rated a site “acceptable,” 1990 and 2000.

Old-growth

Patch cut

Two-story

Snag retention

Thinning

Clearcut

Percent

100

80

60

40

20

0

1990

2000

Figure 1. Regression lines fit to the average scenic rating for each site from 1990 to 2000.

Old-growth (r 2 = 0.29)

Patch cut (r 2 = 0.76)*‡

Clearcut (r 2 = 0.76)*

Two-story (r 2 = 0.56)*

Thinning (r 2 = 0.82)*

Snag retention (r 2 = 0.42)*

4

3

2

1

0

–1

–2

–31990 1992 1994 1996 1998 2000

* = p value < 0.05 ‡ = includes quadratic term

Rating

lower. Over time, however, the ratingsfor sites with the biggest impacts haveimproved the most, and the evalua-tions in 2000 are remarkably similar.This does not mean they look thesame, and scenic quality is still belowthat of the old-growth stand. But theshort-term advantages of the nontradi-tional treatments have faded, and thepresent similarity of the scenic-qualityratings is remarkable.

Are nontraditional treatmentsworthwhile, given that they improvescenic quality in the short term? It isimportant to consider these resultsfrom several time contexts. Elevenyears is not a long time from the per-spective of a forester who thinks interms of rotation ages, which for theseOregon Coast Range sites are typically40 to 50 years. But to the residentswho live around the research forest andthe recreationists who use the area, 11years may seem much longer. To them,the improved scenic quality associatedwith nontraditional silviculture may bewell worth the costs, especially in situ-ations where harvests can be seen frompeople’s homes or recreation sites. In aseparate study on this research forest,Johnson et al. (1994) used computer-based image-capture technology toshow homeowners adjacent to the for-est how the view from their homesmight change under different silvicul-ture scenarios. Results suggested thatvisual impacts are more important inspecific valued settings, such as one’sown home (Johnson et al. 1994). Thestudy also showed that the averagelength of residence was eight years. Theshort-term effects are particularly sig-nificant given this short “rotation” ofhomeownership; five to 10 years oflooking at a clearcut in one’s back yardmay seem like a long time.

The convergence of scenic beautyscores found in this research is applicableonly to these sites in the highly produc-tive Douglas-fir stands of the OregonCoast Range. In lower productivity areassuch as the Rocky Mountains or evenmuch of the Cascades, where rotationsmay be over 60 years, the length of timeto convergence is likely to be longer, andthe relative benefits of nontraditional sil-viculture approaches for scenic qualityprotection are probably greater.

34 Journal of Forestry • March 2003

More than a decade after harvest, the clearcut is now adense plantation of 12- to 15-foot Douglas-fir. Although itsratings are still comparativelylow, it has improved the mostover the course of the study.

The patch cut showed initial improvement, but its ratings decreased as brush built up inthe understory.

The snag retention site has improved throughout the study,although not as rapidly as theclearcut site.

The thinned site has receivedpositive ratings for the past nine years as the evidence ofharvesting has diminished.

Although the two-story site received positive ratingsthroughout the study, it hasimproved less rapidly thansome of the others.

Pho

tos

by J

onat

han

Tho

mps

on

Finally, these results should not betaken as an argument against usingnontraditional treatments for otherpurposes, such as wildlife habitat, em-ulating natural disturbances, andmaintaining biodiversity. There aremany factors to consider when decid-ing on a silviculture approach, andscenic beauty is only one.

Literature CitedANDERSON, L.M., and H. SCHROEDER. 1983. Applica-

tion of wildland scenic assessment methods to theurban landscape. Landscape Planning 10:219–37.

BENSON, R.E., and J.R. ULLRICH. 1981. Visual impacts offorest management activities: Findings on public prefer-ences. Research Paper INT-262. Ogden, Utah: USDAForest Service, Intermountain Research Station.

BRUNSON, M. 1991. Effects of traditional and “new for-estry” practices on recreational and scenic quality ofmanaged forests. PhD thesis, Oregon State University,Corvallis.

BRUNSON, M., and D.K. REITER. 1996. Effects of eco-logical information on judgments about scenic im-pacts of timber harvest. Journal of EnvironmentalManagement 46:31–41.

BRUNSON, M., and B. SHELBY. 1992. Assessing recre-ational and scenic quality: How does new forestryrate? Journal of Forestry 90(7):37–41.

DANIEL, T.C., and R.S. BOSTER. 1976. Measuring land-scape esthetics: The scenic beauty estimation method. Re-search Paper RM-167. Fort Collins, CO: USDA For-est Service, Rocky Mountain Forest and Range Ex-periment Station.

HULL, R.B., and G.J. BUHYOFF. 1986. The scenic beautytemporal distribution method: An attempt to makescenic beauty assessments compatible with forestplanning efforts. Forest Science 32(2):271–86.

JOHNSON R.L., M.W. BRUNSON, and T. KIMURA. 1994.Using image capture technology to assess scenic valueat the urban/forest interface: A case study. Journal ofEnvironmental Management 40:183–95.

RIBE, R.G. 1989. The aesthetics of forestry: What hasempirical preference research taught us? Environmen-tal Management 13(1):55–74.

———. 1991. A general model for understanding theperception of scenic beauty in northern hardwoodforests. Landscape Journal 9(2):86–101.

TAHVANAINEN L., L. TYRVAINEN, and I. NOUSIAINEN.1996. Effect of afforestation on the scenic value ofrural landscape. Scandinavian Journal of Forestry Re-search 11:397–405.

Bo Shelby (bo.shelby@orst.edu) is profes-sor, Jonathan Thompson is graduate re-search assistant, and Rebecca Johnson isprofessor, Department of Forest Resources,Oregon State University, Corvallis, OR97331; Mark Brunson is associate profes-sor, Department of Environment and So-ciety, Utah State University, Logan.

35March 2003 • Journal of Forestry