Recreational Trails and Bird Communities

Transcription

1 - 1 - Recreational Trails and Bird Communities INTRODUCTION One of the most insidious challenges facing scientific researchers is their tendency to find what they are looking for: it can be incredibly difficult to disengage the design of experiments and the interpretation of results from one's prior expectations. This tendency is particularly acute in studies of complex systems with many independent, uncontrollable variables where data analysis is difficult and the results are often ambiguous. When others who seek to support their own positions cite the studies, these problems are further compounded. The case in point is a study for the City of Boulder Open Space Department, "Recreational Trails and Bird Communities," performed by Scott G. Miller and Richard L. Knight during the summers of 1994 and As a public document, this report should be available from the City of Boulder Open Space and Mountain Parks Department. Although this study appears not to have been peer reviewed or formally published, selected results drawn from the study have been cited in subsequent publications (see for example, Attachment A). The report is frequently referenced in Open Space management plans. The more extreme local environmentalists assert that the report proves conclusively that recreational trails are a significant cause of habitat fragmentation and reduction in biodiversity. The authors quite simply never made such claims, but a thoughtful review of the report reveals that even the authors' much less dramatic conclusions are subject to misinterpretation. Subsequent widespread erroneous references to the report's conclusions in the public dialog dramatically illustrate the human ability to find what we want to find. The Boulder Outdoor Coalition obtained and reviewed a copy of the original report. We acknowledge our susceptibility to the 'finding what we are looking for' phenomenon, but still believe that careful examination of the report and its data leads to dramatically less extreme conclusions than are usually asserted by those who cite the study. BACKGROUND The researchers undertook this study because they observed that that little was known about trail effects. They noted that: "Our knowledge of the influence of recreational trails on nest predation, nest parasitism, and wildlife community composition is rudimentary." and, that "Most habitat edge studies have been conducted in deciduous forests of the eastern United States with the majority investigating habitat edge effects of small isolated forest fragments surrounded by agricultural fields." The study addressed the following null hypotheses: "(1) species diversity, evenness, and richness did not differ between areas with trails and areas without trails, (2) species diversity, evenness, and richness did not differ with increasing distance from trail edge, (3) bird species composition and abundance did not differ between areas with trails and areas without trails, (4) bird species composition and abundance did not differ with increasing distance from trail edge, and (5) rates of nest predation and brown-headed cowbird parasitism did not differ with increasing distance from trail edge."

2 - 2 - We have included the following definitions of null hypothesis to help put the study s results in context. Philip B. Stark of the University of California, Berkeley s Department of Statistics offers the following in his Statistics Tools for Internet & Classroom Instruction: Null hypothesis: In hypothesis testing, the hypothesis we wish to falsify on the basis of the data. The null hypothesis is typically that something is not present, that there is no effect, or that there is no difference between treatment and control. Bruce L. Bowerman and Richard T. O Connell of Miami University of Ohio offer the following definition in the Glossary for Hypothesis Testing contained in their Business Statistics in Practice text: The statement being tested in a hypothesis test. It usually represents the status quo and it is not rejected unless there is convincing sample evidence that it is false. METHODS The researchers examined breeding bird communities along five recreational trails and seven control transects in the grassland ecosystem, and twelve trails and seven control transects in the forest ecosystem. They surveyed bird communities using line transects established parallel to trails at 0, 50, and 100 m away from trail centers, and on the control transects. During the 1994 and 1995 breeding season, the authors surveyed each of the transects in June and again in July. RESULTS The researchers found no significant differences in species diversity, evenness, or richness between the trail corridors and the control areas. They also found no significant differences in these attributes with increasing distance from the trail edge. The study results indicate that diversity and richness were actually greater along trails than in the control areas, while evenness decreased along grassland trails and increased along forest trails. Readers should note that the overall differences were insignificant, and the authors data did not disprove the first two null hypotheses, as they expected it would. In other words, in the areas the authors studied, their data demonstrates that the presence of trails had no effect on species diversity, evenness, or richness! Refer to Table 1 for details, and note that our interpretations and annotations are shown in red italics in the table and the following figures. The researchers looked for differences in species composition and abundance between the trail corridors and the control areas. Because the density data gathered was insufficient to use standard methods to calculate results, the researchers estimated relative abundance, sometimes based on just a few sightings. Here the abundance of ten of the 13 grassland bird species and 19 of the 29 forest species did not vary with distance from the trails. Comparisons of the 0-meter trail transects with the control transects show that three of the grassland bird species and eight of the forest species were less abundant near the trails. Distance effects were observed for two of the grassland bird species and five of the forest species. One forest bird species seemed more abundant near trails. We worked backwards from the reported data to discern the actual sample sizes. The data is just too limited to justify the authors conclusions for one of the three grassland bird species and perhaps three of the forest species reported to be less abundant near trails (Figure 1). We are especially troubled by the Great-horned Owl data, which includes only a few sightings in a control area. We have similar questions about the Grasshopper Sparrow, the Western Bluebird, and the White-breasted Nuthatch data. The study did find differences in the species composition and abundance between the trail corridors and the control areas (the third and fourth hypotheses). Somewhat less than one-third of the bird species observed showed some response to the trail's presence. The study s final hypothesis dealt with rates of nest predation and brown-headed cowbird parasitism. The researchers found no cowbird parasitism in the grassland areas and

3 - 3 - no relationship between cowbird parasitism and trail proximity in the forest areas. They calculated nest survival rates and the probability of finding a nest close to a trail. Because of sampling problems however, the researchers could calculate the probability of finding a nest near a trail only for the grassland areas. Here they found a significant, positive correlation between nest survival and distance from trails (Figure 2), and between the probability of finding a nest and distance from trails (Figure 3). It s interesting to note that nests along a trail corridor have essentially the same probability of surviving as nests in the control areas. Survival probabilities right next to the trail are lower than in the control areas, but they increase with distance from the trail, and eventually exceed the control area probabilities. It also worth noting that the probability of finding a nest increases with distance from the trail and significantly exceeds the probability of finding a nest in the control transects. Together, these factors indicate that nesting success in grassland areas is actually greater along the trail corridor than in the control areas. A study by the University of Colorado s Professor Bill Merkle observed a similar effect in riparian corridors (Attachment B). The researchers alluded to this possibility in their final discussion without drawing any conclusions: "Whether or not the avoidance of trails by nesting birds has any consequence on population size is unclear. Fewer nests near trails may be compensated by a greater number away from trails..." SUMMARY Here, in plain language, are the study s major findings: 1. No significant differences in species diversity, evenness, and richness between the trail corridors and the control areas. 2. No significant differences in species diversity, evenness, and richness with increasing distance from the trail edge. 3. Some differences in species composition and abundance between the trail corridors and the control areas. Less than one-third of the bird species observed showed some response to the trail's presence. 4. No cowbird parasitism in the grassland areas and no relationship between cowbird parasitism and trail proximity in the forest areas. 5. As distances from the trail increase, both the probability of finding nests and nest survival increase. The average nest survival probability is identical for trails and control areas. In grasslands, overall nesting success is actually greater along the trail corridors. The authors included several cautions regarding interpretation of their results. In particular: "Our results indicate that trails may be viewed by some species as edges, while this is not the case for others. Thus, caution should be used when equating trails as typical habitat edge. "The intensity of recreational activities on our study area was extremely heavy. Therefore, caution should be used when extrapolating results obtained in our study to landscapes lacking either urban development, intense recreational pressure, or both." Neither the study's results nor the authors' discussion justify extreme claims that recreational trails are a significant cause of habitat fragmentation and reduction in biodiversity. Our review demonstrates that trail effects are limited and localized, and bird communities actually compensate, and even take advantage of those effects.

4 - 4 - Table 1 Species diversity, eveness (Simpson), and richness Diversity Evenness Richness Grassland Transects 0 Meter Meter Meter Avg Control Delta (vs. cntl) 22% -22% 60% Diversity Evenness Richness Forest Transects 0 Meter Meter Meter Avg Control Delta (vs. cntl) 7% 15% 22%

5 - 5 - Figure 1 Data from Knight & Miller Figures 1 & 2-Relative abundance Grassland trails Vesper Sparrow Grasshopper Sparrow Western Meadowlark Trails transects Total m. Sightings Per 100 m Sightings Per 100 m Sightings Per 100 m 0 Meter 11, Meter 11, Meter 11, Control transects 9, Forest trails Mountain Chickadee Mourning Dove Western Bluebird Trails transects Total m. Sightings Per 100 m Sightings Per 100 m Sightings Per 100 m 0 Meter 13, Meter 10, Meter 10, Control transects 8, Townsend's Solitaire Great-horned Owl Western Wood-pewee Trails transects Total m. Sightings Per 100 m Sightings Per 100 m Sightings Per 100 m 0 Meter 13, Meter 10, Meter 10, Control transects 8, Chipping Sparrow Pygmy Nuthatch White-breasted Nuthatch Trails transects Total m. Sightings Per 100 m Sightings Per 100 m Sightings Per 100 m 0 Meter 13, Meter 10, Meter 10, Control transects 8, Stellar's Jay American Robin Trails transects Total m. Sightings Per 100 m Sightings Per 100 m 0 Meter 13, Meter 10, Meter 10, Control transects 8,

6 - 6 - Figure 2 Grasslands Trail Control 0 Meter Meter Meter Meter Meter Avg Forest 0 Meter Meter Meter Meter Meter Avg

7 - 7 - Figure 3 Trail Control 0 Meter Meter Meter Meter Meter Avg Delta (vs. cntl) 48%

8 - 8 - Attachment A An interpretation of the Knight and Miller study INFLUENCE OF RECREATIONAL TRAILS ON BREEDING BIRD COMMUNITIES By Scott G. Miller, 1,3 Richard L. Knight, 1 and Clinton K. Miller, 2,4 Abstract. This study investigated the influence of recreational trails on breeding bird communities in forest and mixed-grass prairie ecosystems in Boulder County, Colorado during 1994 and Species composition, nest predation, and brood parasitism by Brown-headed Cowbirds were examined near and away from existing recreational trails. Results indicated that composition and abundance of birds were altered adjacent to trails in both grasslands and forest ecosystems. In particular, some species like interior forest species did not occur, or occurred in lower densities, near trails than they occurred at a greater distance from trials; whereas some species, mainly generalists like edge species, were more abundant near trails. For the majority of species found in reduced numbers near trails, the zone of influence of trails appears to be about 75 meters (240 ft.). Certain species, exhibited reduced numbers as far as 100 meters (320 ft.) away from trails. Within the grassland ecosystem, birds were less likely to nest near trails. Within both ecosystems, nest predation was greater near trails. It has been previously documented (Hickman, 1990) that avian nest predators were attracted to narrow, open corridors. In forests, the rate of brood parasitism was not influence by trails. No brood parasitism was found in the grassland ecosystem although Hickman, 1990, had documented cowbirds were attracted to nature trails as narrow as 2.5 meters in deciduous forests. Management of natural areas must entail not only proper trail placement, but also recreational management. Consolidation of trails to certain areas (i.e. edges of forests and grasslands) will reduce the fragmentation of large blocks of habitat, maintaining less-disturbed areas for species sensitive to fragmentation. Nature area personnel can inform recreationists of how their activities affect wildlife, and how they can modify their regime of disturbance by staying on trails and keeping pets leashed, to minimize disturbances along the trails. 1 Department of Fishery and Wildlife Biology, Colorado State University; Fort Collins, Colorado 80523; 2 Department of Open Space, 66 S. Cherryvale Road, Boulder Colorado 80303; 3 U.S. Fish and Wildlife Service, Alamos/Monte Vista National Wildlife Refuge, 9383 El Rancho Lane, Alamos, Colorado 81101; 4 The Nature Conservancy, Northern Tallgrass Prairie Ecoregion, PO Box 816, Clear Lake, South Dakota Literature Cited Hickman, S Evidence of edge species attraction to nature trails within deciduous forest. Natural Areas Journal 10:3-5 For more information, see Miller, Scott G., Richard L. Knight and Clinton K. Miller, "Influence of Recreational Trails on Breeding Bird Communities" in Ecological Applications, 8 (1), 1998, pp

9 - 9 - Attachment B Abstract of Boulder County Ecosystem Symposium Paper, March 22, 2002 THE EFFECTS OF RECREATIONAL TRAIL-USE ON THE BEHAVIOR AND NESTING SUCCESS OF AMERICAN ROBINS AND YELLOW WARBLERS Presenter: Bill Merkle (University of Colorado) This research examined the effects of recreational trail-use on the behavior and nesting success of American robins (Tardus migratorius) and yellow warblers (Dendroica petechia) in willow/cottonwood riparian habitats located on public open space lands in Boulder County, Colorado. Nests were located and monitored in riparian corridors with and without trails to determine nesting success. In addition, behavioral observations were conducted on pairs of birds associated with active nest-sites. Vegetation at each study site, as well as each nest-site, was also quantified. Results indicate that increasing intensity of recreational trail-use was associated with increased incubation behavior and reduced feeding of nestlings by female American robins. Male robins appeared to increase their feeding at higher use sites, countering the reduction in female feeding. Surprisingly, both American robins and yellow warblers achieved higher nesting success with increasing intensities of trail-use possibly due to some potential nest predators being displaced from trail sites by disturbance from recreational users. For robins, nesting success was lowest at low use trails. Some of the refuge effect for yellow warblers was reduced by increased parasitism by brown-headed cowbirds with increasing trail use. 9TH BOULDER COUNTY ECOSYSTEM SYMPOSIUM, March 22, 2002

10 Attachment C Habitat Fragmentation Definition BIO-302: Population Ecology and Conservation Ecology Habitat Fragmentation Definition It is the most visible form of habitat alteration. A large, continuous area of habitat is both reduced in area and divided into two or more fragments. It has two components: 1) reduction of the total amount of habitat 2) habitat isolation Causes It s the result of several human activities and/or natural causes: 1) fires, agriculture, urbanization, geologic events, etc. 2) it can also occur when area is reduced to only a minor degree by roads, railroads, pipelines, etc. (Trombulak and Frissell 2000)