Recreation Monitoring at Acadia National Park

Page created by Dorothy Foster
 
CONTINUE READING
Recreation Monitoring at Acadia National Park
Robert Manning, Charles Jacobi, and Jeffrey L. Marion

Introduction
ACADIA NATIONAL PARK IS ONE OF THE MOST INTENSIVELY USED NATIONAL PARKS in the
United States. While its annual visitation (2.2 million visits in 2004) does not rise to the lev-
els of some of the “crown jewel” western national parks (Yellowstone National Park, for
example, accommodated 2.9 million visits in 2004), visits to Acadia are concentrated on its
comparatively small size of less than 50,000 acres. Yellowstone, by comparison, is spread
across 2.2 million acres. Given the intensive character of visitor use at Acadia, it is vital to
monitor recreational use and its associated impacts to help ensure protection of important
park resources and the quality of the visitor experience.
     Over the past decade, Acadia has            tem. These indicators and standards give
undertaken an expanding program of recre-        focus to the monitoring program, enhanc-
ation-related monitoring and associated          ing its efficiency and effectiveness. Second,
activities. This program has been guided by      computer-based simulation modeling has
the National Park Service (NPS) Visitor Ex-      been used to help conduct the monitoring
perience and Resource Protection (VERP)          program. Simulation models can be used to
framework (NPS 1997; Manning 2001).              estimate park conditions (as defined by
VERP was designed by the NPS to address          indicator variables), reducing the need for
carrying capacity and related recreation         on-the-ground monitoring, and can be used
management issues. The underlying ration-        as a more “proactive” monitoring approach
ale of the VERP framework is to (1) define       by estimating the maximum amount of visi-
desired resource and social conditions in        tor use that can be accommodated without
terms of empirically based indicators and        violating standards. Third, several issues
standards, (2) monitor indicator variables,      associated with monitoring on the carriage
and (3) apply management actions to              roads are described. These issues have aris-
ensure that standards have been main-            en over several years of experience with this
tained. Thus, monitoring resource and            monitoring program. Finally, the recreation
social conditions at Acadia has become an        monitoring program at Acadia is expanding
important part of park planning and man-         to address other areas of the park. A pro-
agement.                                         gram of natural and social science research
     This paper describes four aspects of        has been undertaken to help guide this
the monitoring program applied to Acadia.        monitoring. The final section of the paper
First, a suite of indicators and associated      briefly describes this program of research as
standards was initially formulated to guide      it applies to resource-based impacts of
monitoring of the park’s carriage road sys-      recreation, including societal judgments
Volume 23 • Number 2 (2006)                                                                   59
Visitor Impact Monitoring

about appropriate standards for such                than 50 mi of unpaved roads constructed at
impacts.                                            the direction of John D. Rockefeller, Jr., in
                                                    the early 1900s, represent one of the most
Indicators and standards to guide                   significant resources in the park. Originally
monitoring                                          built for horse-drawn carriages, the carriage
      Parks are complex resource and social         roads are now used primarily for hiking and
systems that can be characterized by a mul-         biking and have become extremely popular.
titude of ecological and experiential vari-         However, increased use has created con-
ables. However, monitoring such variables           cern for the quality of the recreation experi-
can be time-consuming and expensive.                ence. In response to this concern, a pro-
Therefore, monitoring programs must be              gram of research was initiated to help for-
designed carefully to ensure that they focus        mulate indicators and standards for the car-
on a relatively small number of indicator           riage-road experience.
variables that best meet selected criteria or            A first phase of research focused on
characteristics (Schomaker 1984; Stankey            identifying potential indicators. A survey of
et al. 1985; Marion 1991; Merigliano 1990;          a representative sample of carriage road vis-
Whittaker and Shelby 1992; National Park            itors was conducted. Using both open- and
Service 1997; Manning 1999).                        closed-ended questions, visitors were asked
      Moreover, within the context of the           to report on what added to or detracted
VERP framework, standards must be for-              from the quality of their experience on the
mulated for indicator variables. Standards          carriage roads. Two types of indicators were
are generally defined as the minimum                identified: one was crowding-related and
acceptable condition of indicator variables         concerned the number of visitors seen on
(National Park Service 1997; Manning                the carriage roads; the other was conflict-
1999). Standards provide vital reference            related and addressed several “problem
points for monitoring programs. Within the          behaviors” experienced on the carriage
VERP framework, standards are thresholds            roads, including bicycles passing from
that define when and where management               behind without warning, excessive bicycle
action is needed. Without such standards,           speed, people obstructing the carriage
findings from monitoring programs are dif-          roads by walking abreast or stopping in
ficult to interpret and provide little direct       groups, and dogs being off leash.
guidance to managers.                                    The first phase of research also docu-
      For these reasons, the recreation-relat-      mented existing patterns of use on the car-
ed monitoring program at Acadia began               riage roads and visitor attitudes toward a
with an effort to formulate a suite of indica-      variety of management alternatives. The
tors and standards. This work was initiated         carriage roads support a diversity of recre-
in the mid-1990s on the park’s system of            ation opportunities defined both spatially
carriage roads (Manning et al. 1998; Jacobi         and temporally. Some areas and times are
and Manning 1999) and has expanded to               relatively heavily used, while others accom-
include all major visitor use areas within the      modate relatively light levels of use. Despite
park. The initial application to the carriage       the problem behaviors noted above, most
roads will be used to illustrate this work.         visitors supported maintaining the current
      The carriage roads, a system of more          mix of carriage road users—hikers, bikers,

60                                                                          The George Wright Forum
Visitor Impact Monitoring

and equestrians. Based on these findings,                Visitors were shown the photographs
park management decided to maintain a               in random order and asked to rate their
diversity of carriage road experiences by           acceptability on a scale from –4 (“very un-
establishing two types of recreation oppor-         acceptable”) to +4 (“very acceptable”).
tunity “zones” for the carriage roads as            Study findings are shown in Figure 2. This
defined by location, time of day, and time of       figure is called a social norm curve (or
year. However, both of these zones would            impact acceptability curve) and represents
continue to accommodate all types of visi-          the aggregate acceptability ratings for the
tors. The two carriage road zones would be          sample of visitors. The norm curve docu-
defined by the same indicators, but different       ments the relationship between increasing
standards would be set.                             use levels and the quality of the visitor expe-
     A second phase of research focused on          rience.
formulating standards for indicator vari-                Two other sources of information were
ables. This research also used a survey of a        developed to help formulate standards for
representative sample of carriage road visi-        crowding on the carriage roads. The survey
tors and adopted normative theory and               using the photographs described above was
related empirical techniques (Shelby and            administered to a representative sample of
Heberlein 1986; Vaske et al. 1986; Man-             residents of communities surrounding the
ning 1999). As applied in outdoor recre-            park, and a computer-based simulation
ation, norms are generally defined as stan-         model of visitor use on the carriage roads
dards that individuals and groups use to            (described in the next section) was devel-
evaluate social and environmental condi-            oped to provide more detailed information
tions in parks and related areas (Shelby and        on the relationship between total daily use
Vaske 1991). If visitors have normative stan-       level of the carriage roads and PPV condi-
dards concerning relevant aspects of recre-         tions. Based on all of this information, stan-
ation experiences, then such norms can be           dards were formulated for the indicator
studied and used as a basis for formulating         variable of PPV conditions.
standards.                                               Standards were also formulated for the
     Because of the relatively large number         four problem behaviors described above.
of visitors on the carriage roads, crowding         Visitors were asked to report the maximum
was measured in terms of persons-per-               number of times it would be acceptable to
viewscape (PPV), incorporating a visually           experience each of these behaviors during a
based measurement approach (Manning et              trip on the carriage roads. The resulting
al. 1996; Manning and Freimund 2005).               norms were used as a basis of formulating
The viewscape for the carriage roads (the           standards.
length of carriage road that can be seen at              After indicators and standards for the
any one time) averages approximately 100            carriage road experience were formulated, a
m. A series of photographs was prepared             similar program of work has been conduct-
that showed a range of zero to 30 visitors on       ed in all other major visitor use areas of the
a typical 100-m section of the carriage             park; indicators and standards are now
roads. The photographs were prepared                being formulated for these areas. Examples
using digital photo-editing software.               of the range of resource and social indica-
Sample photographs are shown in Figure 1.           tors being formulated at Acadia include trail

Volume 23 • Number 2 (2006)                                                                     61
Visitor Impact Monitoring

     1a                                                                1b

     1c                                                                1d

     1e                                                                1f

     Figure 1. Sample study photographs showing a range of visitor use on the carriage roads.

Figure 2. Social norm curve for PPV on the carriage roads.
62                                                                                              The George Wright Forum
Visitor Impact Monitoring

width and erosion, visitor encounters along            can be accommodated on the carriage roads
trails, number of visitors at one time at              without violating crowding-related (PPV)
attraction areas, and number of cars on park           standards. Finally, the model can also be
roads.                                                 used to facilitate a program of adaptive
                                                       management (Lee 1993; Stankey et al.
Computer simulation modeling as a                      2005) by estimating (“monitoring”) the
monitoring tool                                        effects on PPV levels of alternative manage-
     As noted earlier, monitoring indicator            ment practices, such as redistributing use
variables can be time consuming and costly.            over time and space and altering the mix of
Moreover, some indicators, like PPV condi-             carriage roads users (Lawson et al. 2003).
tions along the carriage roads, can be inher-
ently difficult to observe. Simulation mod-            Monitoring recreation on Acadia’s
els offer a potentially attractive alternative to      carriage roads
on-the-ground monitoring. Once a simula-                    Based on the research described above,
tion model is developed, it can be used to             several indicators were established for the
estimate the condition of indicator vari-              quality of the carriage road experience,
ables.                                                 including PPV levels and problem behav-
     A simulation model of carriage road               iors. Standards were then formulated for
use was developed in the 1990s (Manning                these indicator variables. For PPV levels,
et al. 1998; Wang and Manning 1999; Ja-                the simulation model of carriage road use
cobi and Manning 1999; Manning and                     estimated that the carriage roads could
Wang 2005). The model was constructed                  accommodate up to 3,000 visitors per day
using diary reports by visitors of their travel        without violating the PPV standards that
routes and times along the carriage roads              were formulated. Because total daily use of
and counts of the number of visitors enter-            the carriage roads was potentially easier to
ing each of the 11 major access points to the          monitor than PPV levels, this became a
carriage roads. These and related data were            focus of the recreation monitoring program.
processed using the general purpose simu-              Total daily use of the carriage roads is mon-
lation software, Extend. The model was                 itored by means of an electronic trail count-
designed to estimate PPV levels along the              er. Over a period of several years, a series of
carriage roads and can be run at any daily             one-day censuses of carriage road use were
use level of the carriage road system. The             conducted by stationing volunteer ob-
park’s monitoring program measures daily               servers at each of the 11 major carriage road
use of the carriage roads through an elec-             entrances. Resulting counts of total daily
tronic trail counter and uses the simulation           carriage road use were correlated with
model to estimate PPV levels (the crowd-               simultaneous trail counter readings, and a
ing-related indicator variable) to ensure that         strong correlation was found. Based on this
crowding-related standards are not violat-             analysis, PPV levels are monitored on an
ed. Moreover, the model has been used in a             annual basis using trail counter readings.
more “proactive monitoring” approach                   Trail counter readings from 1997 through
(Lawson et al. 2003). The model was run at             2004 suggest that carriage road use has
increasing levels of daily carriage road use           remained relatively stable and that PPV-
to estimate the maximum daily use level that           related standards have not been violated.

Volume 23 • Number 2 (2006)                                                                        63
Visitor Impact Monitoring

     In addition to the monitoring de-              undesignated (visitor-created) trails on
scribed above, actual PPV counts are con-           Little Moose Island (LMI) and designated
ducted every three years at selected loca-          trails on Isle au Haut (IAH). Research has
tions and times. An observer records the            been directed at selecting appropriate indi-
number of visitors on 100-m viewscapes at           cators of resource impacts and developing
15-second intervals. These counts help              protocols for monitoring these indicators.
“ground truth” and validate the PPV esti-           This work has integrated natural and social
mates derived from the trail counter and            science studies to yield indicators and asso-
computer simulation model.                          ciated standards that are both ecologically
     Monitoring of problem behaviors on             and experientially meaningful. This work
the carriage roads is conducted by means of         will be extended to undesignated and desig-
a short visitor survey conducted every three        nated trails on Mount Desert Island in
years. Respondents are asked to report how          2007.
many of the four problem behaviors they                   The creation and proliferation of
have experienced. Resulting data suggest            undesignated trails is a common problem in
that standards are being violated for two           parks that can directly affect sensitive plant
problem behaviors despite enactment of              communities, rare flora and fauna, and
several management practices (promulga-             wildlife habitats (Leung and Marion 2000).
tion of “rules-of-the-road” for the carriage        Visitors seeking to access scenic overlooks,
roads, “courtesy” patrols of the carriage           water resources, or merely to explore, often
roads, and production of an educational             trample vegetation sufficiently to create ex-
video shown at local bicycle rental shops)          tensive informal trail systems. Resource de-
designed to address this issue. If data from        gradation on these trails is often severe due
the next monitoring cycle in 2006 suggest           to lack of professional trail design, con-
that standards are still being violated, a          struction, and maintenance. Such un-
decision will be made to either take addi-          planned trail networks generally receive no
tional management actions to reduce the             environmental reviews, yet they cause
incidences of problem behaviors (e.g., edu-         direct trampling and resource degradation,
cate visitors, adopt behavioral regulations,        habitat fragmentation, and can further
limit carriage road use) or formulate               spread invasive species. While some degree
revised, more realistic standards.                  of visitor impact is unavoidable, excessive
                                                    trail impacts threaten natural resource val-
Monitoring resource impacts                         ues, visitor safety, and the quality of recre-
     Acadia National Park visitors partici-         ation experiences.
pate in a variety of recreation activities,               On LMI four types of monitoring pro-
including driving, boating, hiking, horse           cedures have been experimentally applied:
riding, and biking, and these activities have       (1) global positioning system (GPS) surveys
an array of effects or “impacts” on natural         of the location and extent of visitor-created
resources, including vegetation, soils, water,      trails, (2) trail condition class assessments
and wildlife. To date, resource-related mon-        based on vegetation and soil loss, (3) photo-
itoring and associated research has focused         graphic monitoring from permanent GPS-
on hiking and camping impacts to vegeta-            mapped photopoints, and (4) a point sam-
tion and soils, including assessments of            pling method with a systematic sampling

64                                                                          The George Wright Forum
Visitor Impact Monitoring

interval of 300 ft following a randomized                         (54.8% of the total) were classified as hav-
start. LMI is a very small, pristine island                       ing impacts that were barely distinguishable
located off the Schoodic Peninsula and is                         or being in a lightly impacted condition
only accessible to hikers during low tide.                        (Class 0 or 1), 2,864 ft (18.3%) were
     The GPS and condition class surveys                          assessed as being heavily impacted with
revealed a surprisingly large network of vis-                     eroded treads (Class 5) (Table 1). These
itor-created trails, totaling 15,618 ft (2.96                     procedures and data illustrate the most effi-
mi) (Figure 3 & Table 1). While 8,552 ft                          cient methods for tracking the proliferation

        Figure 3. Trails on Little Moose Island.

Volume 23 • Number 2 (2006)                                                                                 65
Visitor Impact Monitoring

     Table 1. Condition class and length of visitor-created trails on Little Moose Island in 1996.

and degradation of undesignated trails.                                cover of vegetation, organic litter, exposed
Park staff analyzed and employed these data                            soil, rock, and roots). This type of data is
to define a limited subset of 1.09 mi of des-                          presented for illustration purposes in the
ignated trails, a 63% reduction, that retain                           next section on IAH.
access to the island’s principal destination                                On IAH two trail condition assessment
sites. These trails received subtle mainte-                            methodologies, the point sampling and
nance work to facilitate their use, while 1.86                         problem census methods, were integrated
mi of trail were closed to use with brush and                          to provide quantitative data describing con-
small trail closure/revegetation signs. A                              ditions for several impact indicators. A
trailhead sign with a Leave No Trace mes-                              point sampling method with a systematic
sage was installed asking visitors to protect                          sampling interval of 500 ft, following a ran-
sensitive vegetation by staying on the main                            domized start, was the primary method
trails or rock ledges.                                                 (Leung and Marion 1999b; Marion and
      Photopoint monitoring can be easily                              Leung 2001). Data assessed for a subset of
added to GPS surveys to provide a visual                               the indicators are reported here, including
documentation of changing trail conditions.                            tread width, maximum tread incision, and a
Park staff purposely identified and took                               new, more efficient variable interval method
photos at 24 photopoints in 2001 and                                   for determining cross-sectional area (CSA)
2003, including points along the newly des-                            of soil loss (see Figure 4). While more time-
ignated trail system and along the trails                              consuming than maximum incision, CSA
identified for closure. Few changes in vege-                           provides an accurate measure of trail soil
tation cover were detectable along the des-                            erosion that can be extrapolated to provide
ignated trails but photos revealed substan-                            an estimate of total soil loss from each trail.
tial recovery beginning along the closed                                    A problem census method integrated
trails. Finally, the point sampling survey                             into the monitoring procedures provided
provided additional quantitative transect                              census information on three specific trail
data on indicators such as tread width, soil                           impact problems: excessive erosion (>5 in),
loss, and tread substrates (e.g., percent                              excessive muddiness, and number of infor-

66                                                                                                   The George Wright Forum
Visitor Impact Monitoring

      Figure 4. Diagram of the cross-sectional area method (variable intervals) for assessing soil erosion on trails.

mal trails branching from the formal trail                       which equates to a loss of about 55 truck-
since the last sample point (Leung and                           loads of soil. Problem assessment data
Marion 1999a). As they hiked, field staff                        (Table 3) also reveal very few occurrences of
looked for and recorded the beginning and                        excessive soil erosion and muddiness, with
ending distances from the starting point for                     a lineal extent of only 368 and 292 ft/mi,
all occurrences of these problems. In con-                       respectively. These data illustrate the capa-
trast to point sampling, this method pro-                        bility of more measurement-based trail
vides census data on the extent and location                     assessment protocols that provide quantita-
of specific pre-defined problems, facilitat-                     tive data for tracking degradation.
ing management efforts to rectify such                           Preliminary standards of quality have been
impacts. Both methods can be applied con-                        selected for tread width and maximum inci-
currently at an assessment rate of approxi-                      sion, and monitoring will be replicated to
mately one mile/hour by two surveyors.                           evaluate these approximately every five
      All 12 NPS trails on IAH were                              years.
assessed; trail length ranged from 0.16 to                            An associated program of social sci-
3.99 mi, with a mean of 1.56 mi and a total                      ence research was also conducted on trail-
of 20.1 mi. Use data were developed for                          related impacts at IAH. Trail widening and
each trail based on trail use monitoring and                     erosion, visitor-created trails, and vegeta-
development of a simulation model of visi-                       tion loss affect park resources, but these
tor use (Manning et al. 2004). Based on nat-                     impacts can also degrade the quality of the
ural breaks in the use data, IAH trails were                     visitor experience. At what point do such
classified as low-use (0.3–1.0 encoun-                           impacts become unacceptable from an
ters/mi), moderate-use (4.0–6.0 encoun-                          experiential perspective? To help answer
ters/mi), and high use (10.0–11.0 encoun-                        this question, two series of photographs
ters/mi).                                                        were prepared illustrating a range of recre-
      Table 2 presents point sampling data                       ation-related trail impacts on IAH, includ-
for each trail with summaries by level of trail                  ing trail erosion and proliferation of visitor-
use and for the entire island. Results reveal                    created trails (Manning et al. 2004).
increasing degradation with use level,                           Preparation of these photographs was guid-
though conditions on even the high-use                           ed by the assessment of trail conditions
trails are good—which we attribute to rela-                      described above. The photographs for visi-
tively low use, good design, and durable                         tor-created trails are shown in Figure 5. As
substrates. However, even with a mean max-                       with the photographs of visitor use on the
imum incision value of 1.4 in, estimated soil                    carriage roads described earlier, these pho-
loss for the entire trail system totals 553 yd3,                 tographs of resource impacts were incorpo-
Volume 23 • Number 2 (2006)                                                                                             67
Visitor Impact Monitoring

        Table 2. Point sampling data for tread width and soil loss.
rated into a survey of visitors to IAH. The                     dards. Indicators provide a focus for the
resulting social norm curve for visitor-creat-                  monitoring program and standards provide
ed trails is shown in Figure 6. These and                       reference points that inform interpretation
related data are also being used to help for-                   of resulting monitoring data and ultimately
mulate standards for recreation-caused                          guide management action. Identification of
resource impacts.                                               indicators and formulation of standards is
                                                                informed by a continuing program of natu-
Conclusion                                                      ral and social science research.
     Recreation monitoring is an expand-                             Monitoring of recreation at Acadia
ing, evolving, and increasingly important                       includes both resource and experiential
element of park management at Acadia.                           variables. The monitoring program, and the
This monitoring program is conducted                            research upon which it is based, explicitly
within the context of the NPS VERP frame-                       recognize the interrelationships that can
work, which requires identification of indi-                    characterize the resource and social dimen-
cators and formulation of associated stan-                      sions of park and recreation management.
68                                                                                     The George Wright Forum
Visitor Impact Monitoring

    Table 3. Problem assessment data for number of occurrences and lineal distance of excessive soil erosion and muddi-
    ness.
For example, natural science research helps                     designed to maximize efficiency and pro-
document the ecological impacts of recre-                       vide measures of cross-validation where
ation and how these impacts progress over                       possible.
time, while social science research helps                             Recreation monitoring began at Acadia
understand the degree to which these                            nearly 10 years ago, but was relatively nar-
impacts degrade the quality of the visitor                      rowly focused on the quality of the visitor
experience. An integrated program of natu-                      experience on the carriage roads.
ral and social science research can help                        Monitoring has since been extended to
develop a more comprehensive and coordi-                        other areas of the park and to both experi-
nated suite of recreation-related indicators                    ential and resource-related visitor impacts.
and standards.                                                  Ultimately, a suite of resource and experien-
      Recreation monitoring at Acadia is                        tial indicators and standards will be formu-
conducted through a combination of field-                       lated for all major areas of the park, and
based measurements (e.g., measures of trail                     management of recreation use and related
width and erosion, direct observation of                        impacts will be guided by data derived from
PPV levels, visitor surveys), remote sensing                    the park’s program of recreation monitor-
(trail counters), and computer simulation                       ing.
modeling. This combination of methods is
References
Cole, D. 2005. Computer Simulation Modeling of Recreation Use: Current Status, Case
    Studies, and Future Directions. General Technical Report RMRS-GTR-143. Ogden,
    Utah: U.S. Department of Agriculture–Forest Service, Rocky Mountain Research
    Station.
Volume 23 • Number 2 (2006)                                                                                               69
Visitor Impact Monitoring

      5a                                                                     5b

      5c                                                                     5d

     Figure 5. Study photographs showing increasing levels of visitor-created trails.

                   Figure 6. Social norm curve for visitor-created trails.
Jacobi, C., and R. Manning. 1999. Crowding and conflict on the carriage roads of Acadia
    National Park: An application of the Visitor Experience and Resource Protection frame-
    work. Park Science 19:2, 22–26.
Lawson, S., R. Manning, W. Valliere, and B. Wang. 2003. Proactive monitoring and adaptive
    management of social carrying capacity in Arches National Park: An application of com-
    puter simulation modeling. Journal of Environmental Management 68, 305–313.
Lee, K. 1993. Compass and Gyroscope: Integrating Science and Politics for the Environment.
    Washington, D.C.: Island Press.
Leung, Y-F., and J.L. Marion. 1999a. Assessing trail conditions in protected areas:

70                                                                                      The George Wright Forum
Visitor Impact Monitoring

    Application of a problem assessment method in Great Smoky Mountains National Park,
    U.S.A. Environmental Conservation 26:4, 270–279.
———. 1999b. The influence of sampling interval on the accuracy of trail impact assessment.
    Landscape and Urban Planning 43:4, 167–179.
———. 2000. Recreation impacts and management in wilderness: A state-of-knowledge
    review. In Wilderness Science in a Time of Change Conference–Volume 5: Wilderness Eco-
    systems, Threats and Management. D.N. Cole, S.F. McCool, W.T. Borrie, and J.
    O’Loughlan, comps. Proceedings RMRS-P-15-Vol-5. Ogden, Utah: U.S. Department
    of Agriculture–Forest Service, Rocky Mountain Research Station, 23–48.
Manning, R. 1999. Studies in Outdoor Recreation: Search and Research for Satisfaction.
    Corvallis: Oregon State University Press.
———. 2001. Visitor experience and resource protection: A framework for managing the
    carrying capacity of national parks. Journal of Park and Recreation Administration 19,
    93–108.
Manning, R., and W. Freimund. 2004. Use of visual research methods to measure standards
    of quality for parks and outdoor recreation. Journal of Leisure Research 36:4, 552–579.
Manning, R., C. Jacobi, W. Valliere, and B. Wang. 1998. Standards of quality in parks and
    recreation. Parks and Recreation 33:7, 88–94.
Manning, R., S. Lawson, P. Newman, M. Budruk, W. Valliere, D. Laven, and J. Bacon. 2004.
    Visitor perceptions of recreation-related resource impacts. In Environmental Impacts of
    Ecotourism. Cambridge, Mass.: CABI International, 259–272.
Manning, R., D. Lime, W. Freimund, and D. Pitt. 1996. Crowding norms at frontcountry
    sites: A visual approach to setting standards of quality. Leisure Sciences 18:1, 39–59.
Manning, R.E., and F.I. Potter. 1984. Computer simulation as a tool in teaching park and
    wilderness management. Journal of Environmental Education, 15:3, 3–9.
Manning, R., and B. Wang. 2005. Acadia National Park carriage roads: Estimating the effect
    of increasing use on crowding-related variables. In Computer Simulation Modeling of
    Recreation Use: Current Status, Case Studies, and Future Directions. General Technical
    Report RMRS-GTR-143. Ogden, Utah: U.S. Department of Agriculture–Forest Ser-
    vice, Rocky Mountain Research Station, 50–54.
Marion, J.L., and Y-F. Leung. 2001. Trail resource impacts and an examination of alternative
    assessment techniques. Journal of Park and Recreation Administration 19:3, 17–37.
McCool, S.F., D.W. Lime, and D.H. Anderson. 1977. Simulation modeling as a tool for
    managing river recreation. In Proceedings: River Recreation, Management and Research
    Symposium. General Technical Report NC-28. St. Paul, Minn.: U.S. Department of
    Agriculture–Forest Service, North Central Forest Experiment Station, 304–311.
Merigliano, L. 1990. Indicators to monitor the wilderness recreation experience. In
    Managing America’s Enduring Wilderness Resource. St. Paul, Minn.: University of
    Minnesota, 156–162.
National Park Service. 1997. VERP: The Visitor Experience and Resource Protection (VERP)
    Framework—A Handbook for Planners and Managers. Denver: NPS.
Schomaker, J.H. 1984. Writing quantifiable river recreation management objectives. In
    National River Recreation Symposium Proceedings. J.S. Popadic, D.I. Butterfield, D.H.

Volume 23 • Number 2 (2006)                                                              71
Visitor Impact Monitoring

     Anderson, and M.R. Popadic, eds. Baton Rouge: Louisiana State University, 249–253.
Shechter, M., and R.C. Lucas. 1978. Simulation of Recreational Use for Park and Wilderness
     Management. Baltimore: Johns Hopkins University Press.
Shelby, B., and T.A. Heberlein. 1986. Carrying Capacity in Recreation Settings. Corvallis:
     Oregon State University Press.
Shelby, B., and J.J. Vaske. 1991. Using normative data to develop evaluative standards for
     resource management: A comment on three recent papers. Journal of Leisure Research
     23, 173–187.
Smith, V.K., and R.L. Headly. 1975. The use of computer simulation models in wilderness
     management: A case study of the Adirondack Forest Reserve. In Management Science
     Applications to Leisure Time. S. Ladany, ed. Amsterdam: North-Holland Publications.
Smith, V.K., and J.V. Krutilla. 1976. Structure and Properties of a Wilderness Travel
     Simulator: An Application to the Spanish Peaks Area. Baltimore: Johns Hopkins
     University Press.
Stankey, G., R. Clark, and B. Bormann. 2005. Adaptive Management of Natural Resources:
     Theory, Concepts, and Management Institutions. General Technical Report PNW-654.
     Portland, Ore.: U.S. Department of Agriculture–Forest Service, Pacific Northwest
     Forest and Range Experiment Station.
Stankey, G., D. Cole, R. Lucas, M. Peterson, S. Frissel, and R. Washburne. 1985. In The
     Limits of Acceptable Change (LAC) System for Wilderness Planning. General Technical
     Report INT-176. Ogden, Utah: U.S. Department of Agriculture–Forest Service,
     Intermountain Forest and Range Experiment Station.
Van Wagtendonk, J., and D. Cole. 2005. Historical development of simulation models of
     recreation use. In Computer Simulation Modeling of Recreation Use: Current Status,
     Case Studies, and Future Directions. General Technical Report RMRS-GTR-143.
     Ogden, Utah: U.S. Department of Agriculture–Forest Service, Rocky Mountain
     Research Station, 3–9.
Wang, B., and R.E. Manning. 1999. Computer simulation modeling for recreation manage-
     ment: A study on carriage road use in Acadia National Park, Maine, USA.
     Environmental Management 23, 193–203.
Whittaker, D., and B. Shelby. 1991. Developing good standards: Criteria, characteristics,
     and sources. In Defining Wilderness Quality: The Role of Standards in Wilderness
     Management—A Workshop Proceedings. General Technical Report PNW-305. Portland,
     Ore.: U.S. Department of Agriculture–Forest Service, Pacific Northwest Forest and
     Range Experiment Station, 6–12.
Vaske, J., A. Graefe, B. Shelby, and T. Heberlein. 1986. Backcountry encounter norms:
     Theory, method, and empirical evidence. Journal of Leisure Research 18, 137–153.
Robert Manning, University of Vermont, 356 Aiken Center, Burlington, Vermont 05405;
     robert.manning@uvm.edu
Charles Jacobi, Acadia National Park, P.O. Box 177, Bar Harbor, Maine 04609-0177; char-
     lie_jacobi@nps.gov
Jeffrey L. Marion, U.S. Geological Survey, Patuxent Wildlife Research Center, Virginia
     Tech, Forestry (0324), Blacksburg, Virginia 24061; jmarion@vt.edu
72                                                                    The George Wright Forum
You can also read