adaptive management of prairie grouse:how do we get … management of prairie grouse • aldridge et...

9922 SSPPEECCIIAALL CCOOVVEERRAAGGEE

WWiillddlliiffee SSoocciieettyy BBuulllleettiinn 22000044,, 3322((11))::9922––110033 PPeeeerr rreeffeerreeeedd

Prairie grouse have been declining throughout NorthAmerica over the last century (Braun et al. 1994; Braun1998; Connelly et al. 1998; Applegate et al. 2000a, b).Some of the most marked declines have been recordedfor sage-grouse (Centrocercus spp.; Braun 1998) andgreater prairie-chickens (Tympanuchus cupido; Applegate

et al. 2000a) over the last few decades. Prairie grousepopulation declines are associated with direct loss ofhabitat and continued fragmentation and degradation ofexisting habitat (Braun et al. 1994). Management effortshave been relatively unsuccessful, and populations con-tinue to decline. As a result, several species or distinct

Adaptive management ofprairie grouse: how dowe get there?by Cameron L.Aldridge, Mark S. Boyce,and Richard K. Baydack

Managing prairie grouse has been largely a reactive process without any “true” manage-ment experiments being implemented, thereby limiting our ability to learn from man-agement and enhance conservation efforts for declining prairie grouse populations. Ina few cases where the potential existed for a passive or active adaptive approach, mon-itoring was insufficient to detect effects of changes in management practices. Similarproblems appear to occur at planning stages in attempts to implement adaptive man-agement for prairie grouse populations, preventing proper consideration of soundadaptive experiments that advance learning. Successful adaptive management beginswith stakeholder gatherings following a policy planning process, which includes manysteps, beginning with goal identification and understanding of uncertainties and culmi-nating in model simulations to understand potential management policies. By follow-ing this process, the opportunity to implement successful management experimentscan be enhanced. We discuss the successes and failures of prairie grouse managementusing 2 case studies, 1 for prairie sharp-tailed grouse (Tympanuchus phasianellus) inManitoba and 1 for greater sage-grouse (Centrocercus urophasianus) in southernAlberta. We describe ways in which active adaptive management could improve ourunderstanding of prairie grouse population declines and outline a policy planningprocess that, if followed, will allow adaptive management to be successfully imple-mented, enhancing prairie grouse management and conservation.

AAbbssttrraacctt

adaptive management,Alberta, Centrocercus urophasianus, conservation plans,grazing, greater sage-grouse, habitat, Manitoba, policy planning, prairie sharp-tailedgrouse, Tympanuchus phasianellus

KKeeyy WWoorrddss

Address for Cameron L. Aldridge and Mark S. Boyce: Department of Biological Sciences, University of Alberta, Edmonton, Alberta, Canada T6G2E9; e-mail for Aldridge: [email protected]. Address for Richard K. Baydack: Faculty of Environment, University of Manitoba, Winnipeg,Manitoba, Canada R3T 2N2.

Adaptive management of prairie grouse • Aldridge et al. 9933

populations in Canada and the United States have beenpetitioned or listed as threatened or endangered(Anonymous 2000, Aldridge and Brigham 2003). Toooften, prairie grouse managers have taken a wait-and-seeor trial-and-error approach (Hilborn 1992, Halbert 1993).This has limited our ability to learn about population reg-ulation and habitat limitations. We believe that adaptivemanagement (Holling 1978, Walters 1986, Walters1997b), appropriately implemented, could identify soundalternatives for prairie grouse populations, advance ourlearning, and improve management policies to benefitprairie grouse populations. In this paper we presentbackground on the history and idea of adaptive manage-ment. We discuss why many adaptive managementstrategies have failed and illustratehow and why it is important todevelop an approach for implement-ing management policies for prairiegrouse within an adaptive frame-work. We use 2 case studies toillustrate our points, one for prairiesharp-tailed grouse (T. phasianellus)in Manitoba and one for greater sage-grouse (C. uropha-sianus) in Alberta.

History of adaptive managementBuzz Holling, Carl Walters, and Ray Hilborn devel-

oped the concept of adaptive management in 1974through a series of workshops (Ludwig and Walters2002). The concept was aimed at building models tounderstand uncertainties associated with naturalresources, and involved managers, policy-makers, andscientists in the process (Holling 1978, Walters 1986).Walters and Hilborn (1976) introduced the idea of adap-tive resource management in 1976. They pointed out thatexperimentation was the most reliable means of under-standing uncertainties in resource systems, and that com-paring alternative models should form the basis of man-agement, experimental design, and monitoring of theresource system (Holling 1978).

The adaptive policy process involves stakeholders andbegins by integrating existing knowledge and scientificinformation into dynamic models used to make predic-tions about impacts of alternative management practices(Holling 1978, Walters 1986, Walters 1997a). Practicesare re-evaluated and adjusted as new information isobtained from current management (Holling 1978,Walters 1986, Walter 1997b). Adaptive management isessentially learning by doing (Haney and Power 1996,Walters 1997b), placed in the framework of experimentaldesign with feedback from the learning stage back into

the doing stage. This concept is not new; most man-agers, scientists, and wildlife biologists are aware ofadaptive management. However, it is a poorly under-stood concept that is difficult to apply and often inappro-priately applied (Walters 1997b, Wilhere 2002).

Adaptive management definitions andmisconceptions

The 2 key components of adaptive management are 1)management is effectively set out as an experiment witha sound experimental design, and 2) a direct feedbackloop exists between science and management (Halbert1993). Essentially, adaptive management is the incorpo-

ration of the scientific method (experiments) into a man-agement framework (policy decisions). This differenti-ates adaptive management from traditional trial-and-erroror learn-as-you-go management (Hilborn 1992, Halbert1993). Managers and stakeholders involved in conserva-tion planning processes often disregard this fact and thinkof adaptive management simply as sound management ormanagement with a willingness to change (Wilhere2002). Others see it as flexible management, an opportu-nity to contest policies they consider objectionable(Wilhere 2002). Bormann et al. (1999:506) definedadaptive management as “an approach to managing natu-ral systems that builds on learning—based on commonsense, experience, experimenting, and monitoring—byadjusting practices based on what was learned.”

Adaptive management can be implemented at variouslevels, the simplest of which is called reactive manage-ment, similar to trial-and-error management (Hilborn1992). In this case, changes are driven by stimuli exter-nal to the management system, such as politics, lawsuits,and public opinion (Bormann et al. 1999). As a result,conflicts often arise due to multiple stimuli, resulting inhaphazard management (Bormann et al. 1999, Roe andVan Eeten 2001). Many wildlife management decisions,both old and current, have been made based on reactions,and most of them resulted in crisis management, withone policy often trying to correct crises created by earliermanagement decisions (Bormann et al. 1999, Roe andVan Eeten 2001). Reactive management is unreplicatedand lacks statistically valid experimental design, often

Although adaptive management has become a popu-lar idea among management agencies and can be use-ful when implemented correctly, in practice it is oftenused as a buzzword and never implemented.

9944 Wildlife Society Bulletin 2004, 32(1):92–103

producing unreliable information (Hurlbert 1984, Wilhere2002). This often results in implementation of singlepolicies or strategies that are assumed to be suitable with-out evaluating the policies or comparing alternativestrategies or controls.

In contrast to reactive management, passive adaptivemanagement involves long-term monitoring and learningfrom a gradually evolving management strategy (Walters1986, Hilborn 1992), but typically lacks scientific rigor(i.e., no replication, no controls, no randomization;Hurlbert 1984, Wilhere 2002). This relatively simple andinexpensive approach involves measuring responses rela-tive to what happened in the past (learning from experi-ence; Roe and Van Eeten 2001), but understanding ofcausal relationships is limited (Bormann et al. 1999).This process often becomes reactive trial-and-error man-agement, when funds committed to monitoring areremoved (Bormann et al. 1999). Passive adaptive man-agement can be useful in systems that have a high degreeof natural variability (Halbert 1993); large enough naturalperturbations can be measured and correlated to distur-bances. However, if processes other than managementare causing the variability (i.e., environmental variabilitylike weather patterns), it can make causal relationshipsdifficult to discern.

Finally, active adaptive management is a rigorous formof the concept (Walters 1986, Hilborn 1992) that occursthrough parallel learning (Bormann et al. 1999). Thisapproach differs in that management policies aredesigned, replicated, and tested against each other.Management is essentially achieved through a series ofcontrolled experiments (Walters 1997b, Bormann et al.1999, Roe and Van Eeten 2001), identifying cause-and-effect relationships between management activities andchanges within the system (Wilhere 2002). This pro-motes rapid learning and the formation of optimal poli-cies that will guide future management (Wilhere 2002)and prevents the broad application of any single policythat may or may not work, possibly preventing the alter-ation of future management (Walters 1997b).

The primary difficulty with implementing an activeadaptive management strategy is replication, making itdifficult to use in a single or unique system (Walters1986), (e.g., small endangered populations; Boyce 1993).One option may be to implement various managementstrategies in separate subpopulations (Boyce 1993), pos-sibly comparing treatment effects between areas withsimilar characteristics. In this case, it would be difficultto control for spatial variation, but replication of treat-ments may be achievable.

Another approach for adaptive management of smallpopulations is the Before-After-Control-Intervention

(BACI) design, originally identified by Green (1979).With this design, variability in the system is monitoredprior to intervention. In adaptive management applica-tions, the key elements of a BACI design include replica-tion, controls, and monitoring both before and after man-agement intervention. In cases where management poli-cies will take place at the scale of the entire population(i.e., small populations), replication is impossible.However, management actions can be compared using aninformation–theoretic approach. By framing treatmentsin the context of multiple working hypotheses(Chamberlin 1890, Anderson et al. 2000), hypotheseswould consist of a series of a priori candidate models,management options outlined during policy planningmeetings (Walters 1997b, see below). If a particularstrategy achieved a biologically meaningful goal (Reedand Blaustein 1995), such as an increase in populationsize of a predefined magnitude, then the strategy shouldbe considered successful, regardless of the statistical sig-nificance of the experiment. With this comparativeapproach, the adaptive management process is still usedand biological goals can be tested.

Reasons adaptive management failsAdaptive management can fail for a variety of reasons

(Walters 1997a, Moir and Block 2001). The major flawis that the process rarely progresses from the modeldevelopment stage to the design and implementation offield experiments. Walters (1997a) participated in 25adaptive management planning processes, only 7 (28%)of which resulted in large-scale management experi-ments; 2 (8%) had well-planned experimental designswith controls and suitable replication. Walters (1997a)suggests that experiments often are opposed by people

Female sage-grouse sitting on her nest under silver sagebrush (Artemisiacana) surrounded by grass and forbs. Photo by Cameron L. Aldridge.


protecting self-interests in management bureaucracies,and proponents of adaptive management need to beforceful and expose these groups and their interests topublic scrutiny. This will help to keep the adaptive man-agement process on track and maintain the sustainabilityof public resources. Value conflicts arise from the neces-sity of involving all stakeholder groups with an interest inthe resource in the decision-making process, each poten-tially possessing different values, morals, and opinions(Walters 1997a).

Adaptive management needs effective implementationof experiments, which may be expensive or risk-prone(Macnab 1983, Halbert 1993, Walters and Green 1997,Walters 1997a) compared with baseline options, especial-ly when threatened or endangered species are involved.Public agencies by nature are risk-averse and manage forthe status quo (Halbert 1993). However, if risk-aversestatus-quo management (Halbert 1993) is not benefiting athreatened species, then adaptive management presents aviable option. Identifying uncertainties early in the plan-ning stages (see below) will eliminate individual or mul-tiple management options that may pose risks (Walters1997b).

Even if experimental designs are implemented, adap-tive management often fails because the information feed-back loop (monitoring and evaluation) is broken (Moirand Block 2001). Thus, learning is inhibited and there isno evolution of management policies. This loop typicallyis broken because managers are looking (by necessity) forshort-term responses and feedback from managementpolicies (Moir and Block 2001). To avoid such conflicts,the realization that responses to management may belonger term needs to be identified by stakeholders duringinitial policy planning to avoid such conflicts.

Below, we present 2 case studies in which attemptswere made to implement adaptive management strategiesfor prairie grouse. Both attempts failed to successfullyimplement adaptive management; however, we feel thereare important lessons to be learned from our experiences.

Case Study 1: Manitoba sharp-tailedgrouse

In Manitoba a pilot project for sharp-tailed grousehabitat management using principles of adaptive manage-ment has been underway for a number of years. The pro-gram is a partnership approach involving a local nonprof-it organization, the Sharptails Plus Foundation, in associ-ation with the University of Manitoba and the ManitobaDepartment of Conservation.

Estimates of sharp-tailed grouse in Manitoba indicateda decreasing population, although they are still at rela-

tively high levels when compared to other NorthAmerican populations (Froese 2002). Habitat alterationwas suggested as the cause of the long-term decline(Baydack 1996). In more southerly areas of theprovince, agriculture had eliminated historically impor-tant cover types, particularly shrubland and woodland; incentral areas, fire suppression and reduced grazing hadcaused shrubland and woodland to increase above his-toric levels (Berger and Baydack 1993). In central areassuch as Manitoba’s Interlake region, sharptails werethought to have thrived historically in diverse vegetativecover, roughly equivalent to an equal one-third mix ofgrassland, shrubland, and woodland (Bird 1961). Thishistorical composition has changed throughout theprovince, most notably with respect to shrubland compo-sition, which is nonexistent in many areas (Berger andBaydack 1993).

To address the effects of habitat alterations on sharp-tailed grouse populations, a private, nonprofit organiza-tion, the Sharptails Plus Foundation (SPF), was created inManitoba in the early 1990s. It is volunteer-based, com-prised of relatively influential citizens whose overall goalhas been to enhance the habitat condition for sharp-tailedgrouse and thereby influence population levels through-out the province. Although the primary focus of theorganization is on sharp-tailed grouse, members recog-nize that their work will affect many other species andcommunities, hence the “Plus” designation in their name.In the early 1990s, Sharptails Plus began to implement avariety of habitat-related programs across Manitoba withthe help of a Technical Advisory Committee of ecolo-gists, agrologists, and land managers. Their PrivateLands and Pilot Projects Program was developed with theintent that it be implemented using concepts of adaptivemanagement.

The program was based on several important underly-ing concepts. Cooperation with agriculture was essentialfor achieving effective sharp-tailed grouse habitat man-agement. Landowners were seen as having reasons fordoing what they do, and they are, in effect, the real habi-tat managers in much of Manitoba. Sharptails Plus rec-ognized that understanding the motivation for landownerdecisions about habitat and wildlife on their property wasimportant. Similarly, Sharptails Plus recognized that bio-logical data generally were lacking about how to bestmanage a parcel of land to meet sharp-tailed grouseneeds. Finally, landowners needed to continue to managetheir holdings, regardless of whether the “final” biologi-cal answer could be provided. The Technical AdvisoryCommittee of Sharptails Plus noted that this created anopportunity to implement an adaptive management pro-gram that utilized the best available local knowledge

from landowners and biologists to reach prescriptionsthat could be tested against each other. Steps inSharptails Plus’s Private Lands Program included the fol-lowing:

1. Interested landowners in target locations were identi-fied.

2. STP Technical Advisory Committee (biologists,agrologists, land managers) assessed existing habitatalong with individual landowners.

3. Habitat objectives were established for each location inconsultation with landowners, generally based on thepercent cover of woodland, shrubland, grassland, andcropland, compared to the historic one-third levels.

4. Independently, the STP Technical Team andlandowners developed their own habitat-managementtreatment recommendations for each location basedon their individual reasoning.

5. Land parcels were randomly allocated to each “com-peting” management option, resulting in an overallAdaptive Farm Management Plan.

6. Management treatments were monitored, evaluated,and refined annually using both biological and eco-nomic measures. Budgetary constraints preventedradiomarking grouse to track their response to chang-ing vegetation over time and limited monitoring toattempts to locate sharp-tailed grouse leks and nestsites. If leks were located near treatment sites, atleast 5 counts of sharptail abundance were recordedannually in spring within 0.5 hour of sunrise, andthese counts were compared over time and amongtreatments. Vegetative measurements were takenbefore and after treatments in spring, summer, andfall at every treatment site, using a nested-blockdesign as described in Froese (2002). Measurementswere dependent on the treatment applied and includedvegetation composition, structure, height, rate ofgrowth, percent kill of unwanted species, and others.Economic analyses of treatments also were performedannually and compared over time (Froese 2002).

A variety of management treatment options wereavailable for consideration at every location, including:rotational and pulse grazing, delayed mowing and hay-ing, leaving increased edge, planting shelterbelts andwoodlots, prescribed burning, mechanical removal ofvegetation, chemical removal of vegetation, and foodplots.

The program has been operational in Manitoba at 4locations (near the communities of Plumas, Lundar,Chatfield, and Vita) for the past 5 years. Preliminaryresults revealed the following trends:

• Landowners seem generally satisfied with the out-comes, which have resulted in increased forage usefor their cattle and an apparent increase in sharp-tailed grouse abundance, although increased monitor-ing (i.e., daily or twice daily) of activity on leks inspring and use of radiomarking has been suggested.

• Most landowners considered an indicator of successto be whether a sharp-tailed grouse lek was actuallyestablished on their properties, often a difficultexpectation to achieve.

• Biological indicators of success related to attainmentof desired cover composition (percent cover in wood-land, shrubland, grassland, cropland), which general-ly has been possible.

• Management treatment options proposed by biolo-gists were no better in achieving vegetative covercomposition “success” than those suggested bylandowners.

• Monitoring of economic indicators suggested thatincentive programs would be necessary to offsetincreased costs to farm operations. However, esti-mated costs of treatments likely were higher thanwould be expected due to the cost of smaller-scaleexperimental procedures used to formulate estimates.

On the basis of the success of the program to date asperceived by landowners, the Private Lands and PilotProjects Program will continue in Manitoba into the fore-seeable future. Although the Sharptails PlusFoundationis finding that financial requirements continue to be rela-tively high compared to their available resources, theybelieve the program is useful to ensure local support forconservation measures and to increase the area of cover-age within the province. The program is recognizing thatdedicated landowners are the real key to success in man-aging sharp-tailed grouse on private lands in Manitoba


Two-day-old sage-grouse chick in good escape cover. Photo byCameron L. Aldridge.

and, perhaps more importantly, that dedicated landownersare at least as successful as biological managers in know-ing the best practices for their landholdings. Therefore, apressing need exists to encourage the release of this valu-able local knowledge onto the prairie landscape so as toenhance management of wildlife habitats for future gen-erations.

This case study represents a good initial attempt atmanaging declining sharp-tailed grouse numbers and wassuccessful in achieving some of its goals. The programbegan using a collaborative planning process (Walters1997b) and set out to improve habitat for sharp-tailedgrouse by creating habitats similar to historical timeswhen sharp-tailed grouse were more abundant, with anequal one-third mix of woodland, shrubland, and grass-land. Most management techniques appeared to be suc-cessful at achieving vegetation goals; however, theprocess was limited from the beginning, when funds forgrouse evaluations could not be allocated. Realistically,the program should be evaluated by how successful itwas at creating habitats beneficial to, and used by, thebirds—hence, increasing population numbers. In addi-tion, other measures of biodiversity could have beenincorporated into the monitoring programs so as to pro-vide additional indicators of success for managementdecision-making.

Case Study 2: Implementation failuresof Alberta sage-grouse adaptive

managementThe greater sage-grouse in Alberta has experienced

some of the highest rates of decline of any known sage-grouse population (66–92% over the last 30 years;Aldridge and Brigham 2001, 2003), and the species islisted as endangered at both the provincial and federallevels. Research completed in 1999 suggested that limit-ed residual cover and litter buildup was limiting the pop-ulation through poor chick survival and low nest success(Aldridge and Brigham 2001, 2002). To document therole of range management, it was recommended that avariety of stocking rates be used in 2000–2001 in anattempt to enhance litter buildup and residual cover forsage-grouse and that the response of the vegetation andsage-grouse be monitored.

The goal of this management experiment was to assessthe effect of grazing on nesting success of sage-grouse.We (C.L.A. and M.S.B.) designed the experiment withreplicate treatments using 3 stocking rates (all variouslevels lower than current rates) and had many sites thatmaintained current stocking rates (controls). We chosesites (treatments and controls) with similar grazing histo-

ries and previous stocking rates, season, and duration ofgrazing. Treatments were randomly allocated (boundedby landowner participation) and replicated over differentpastures, allowing us to compare management treatments(stocking rates) against each other and controls.Individual nest sites located within each treatment unit(pasture) would be grouped and the size of the treatmentsand controls kept similar, albeit limited to the size of cur-rently utilized pastures. However, the effect of treatmentsize also could have been incorporated into the experi-mental design to test for an effect of treatment size (seestep 4 below). A nested design was used to avoidpseudoreplication (Hulbert 1984), with each additionalnest site within a treatment increasing our power todetect differences. Below, we highlight why thisapproach has failed, then describe how the process stillmay be implemented by adopting a collaborative adaptiveprocess (Walters 1997b).

In spring of 2000, with fears in the ranching commu-nity about the pending Canadian Species At Risk Act(SARA) and its potential consequences for management,scientists suggested that Alberta Sustainable ResourceDevelopment (SRD) and the Canadian Wildlife Service(CWS) hold public forums with local ranchers. The 2organizations were to work under the umbrella of theCWS Endangered Species Habitat Stewardship programto approach landowners who might be interested in par-ticipating in adaptive sage-grouse range-managementexperiments. Due to a lack of funds from governmentagencies in summer 2000 and the sensitive and politicalnature of the pending endangered species legislation, thisuncoordinated approach failed and only added to locallandowner concerns about outsiders attempting to man-age their land.

In 2001 SRD collaborated with the AlbertaConservation Association (ACA) to develop management


A new oil well being drilled in prime endangered sage-grouse habitat insoutheastern Alberta. Photo by Cameron L. Aldridge.

strategies. However, at this point there was no involve-ment by Alberta Public Lands, the agency charged withmanaging the public lands that constitute >80% of cur-rent sage-grouse habitat in southern Alberta. The ACAhad a habitat management program, called Native PrairieStewardship, expanding into the area, which providedopportunity for collaboration. The program’s focus wason developing and implementing sound range manage-ment to improve residual cover and litter for wildlife(important for nest success and chick survival of bothsage-grouse and sharp-tailed grouse). The programworked with individual ranchers to develop range-man-agement plans for each participating ranch and thenhelped to focus new management initiatives, assistinglandowners with cost of implementation.

This program had limited monitoring and evaluationcomponents, both important parts of the initial designprocess, as well as management policies themselves. Asthe program continued into 2003 under the premise ofenhancing and protecting critical habitat for both sage-grouse and sharp-tailed grouse, some post-hoc songbirdcounts were added in an attempt to evaluate the effective-ness of the program. However, without suitable evalua-tions, there is no way to ensure that management initia-tives achieve desired ecological goals.

At roughly the same time, the Nature Conservancy ofCanada (NCC) began signing conservation easements inthe area and developing management plans with thenotion of protecting sage-grouse habitat. Like the NativePrairie Stewardship program, limited evaluation is associ-ated with the NCC program. Both programs developconservation management plans that are private landown-er agreements and do not allow others access to the man-agement plan or actions. By creating independent ranch-management plans void of experimental treatments andpreventing access to management policies, currentlyradiomarked sage-grouse on these ranches could not beused to understand how grouse respond to managementinitiatives, preventing an adaptive approach involvingevaluation and subsequent evolution of managementstrategies.

As a result of these activities and proposed manage-ment, Alberta Public Lands also became involved insage-grouse management. They began their own initia-tives to address the decline of sage-grouse, unconvincedthat grazing was related to those declines. This resultedin 5 different agencies developing management strategiesfor sage-grouse and effectively overwhelming landownerswith sage-grouse issues, causing them to close theirdoors to ongoing research. As a result, SRD coordinateda public meeting with landowners, researchers, the ACA,and Public Lands in spring 2002. This resulted in the

formation of a Provincial Recovery Action Group (RAG)designed to bring together science, management, indus-try, and landowners in a collaborative effort aimed atimplementing adaptive management strategies for therecovery of sage-grouse in Alberta. Walters (1997b) sug-gests that this is the first step in the policy-planningprocess necessary for the successful implementation ofadaptive management (see summary below). Thisprocess will incorporate evolving economic and socialconcerns as well as new biological information pertainingto the system, an important process in adaptive manage-ment (Haney and Power 1996).

Delays in implementation continued as Alberta PublicLands refused to embrace the uncertainty about grazing,insisting there was not enough information to begin adap-tive trials. Adaptive management should embrace thisuncertainty, allowing for further understanding of the sys-tem through the implementation of a series of soundexperimental management policies. However, PublicLands undertook its own research, including backgroundrange inventories, developing a sagebrush-soils classifica-tion, performing water-impediments studies, and per-forming a historical grazing-practices study. While retro-spective studies may provide some insights, managementexperiments were postponed. Walters (1997b) points outthat if policy-design processes begin by attempting toidentify all scientific uncertainties about a system, thenthat process will fail on the simple fact that there is aninfinite number of uncertainties. The goal of adaptivemanagement is to embrace these uncertainties throughexperiments, increasing our ability to learn about the sys-tem.

The RAG is now showing signs of progress. Afterappointing a moderator familiar with the adaptive man-agement process in late 2002, the group is approachingmanagement following methods similar to Walters’(1997b) policy-planning process, albeit 3–4 years afterthe initial investment of funds by 5 different agencies andstakeholders. The RAG was able to bring together indus-try and conservation representatives, managers, scientists,and landowners for the planning process.Representatives and their organizations have all boughtinto the process, most committing long-term to the RAG.Currently, the group is developing models to predict theoutcomes of certain policies and management options,aimed at identifying and understanding uncertaintieswithin the system (see steps 2 and 3 below). While adap-tive management most often fails prior to implementingexperiments (Walters 1997a), by following an adaptivepolicy-design process similar to that outlined by Walters(1997b; see below), the RAG will be able to identifysound management options and will have the opportunity


to implement those options as a series of experimentswith suitable monitoring and evaluation components.

These 2 prairie grouse examples illustrate how impor-tant it is to have local involvement and “buy-in” at thepolicy-planning stages of adaptive management(Freyfogle 1998) and how easily the process can failwithout it. Below, we discuss how one might implementa successful adaptive management policy and avoid thesame mistakes we encountered during the policy-plan-ning stages, prior to implementing any managementexperiments.

Implementation of successful adaptivemanagement

Implementing an adaptive management policy beginswith stakeholder gatherings and identification of all ele-ments, variables of interest, management acts, objectives,indicators, time horizons, and spatial extents (Holling1978). This is often one of the most difficult steps toundertake, but a concerted effort to involve all stakehold-ers at the beginning of the process will reduce thechances of failing at later stages and provide a sense ofownership to the decisions for each group involved.Ensuring that conservation groups, local landowners,industrial representatives, and managers stay committedto adaptive management processes over the long term canbe difficult, especially given the turnover of individualswithin most agencies and the fact that local landownershave limited time to commit to such efforts. This iswhere having an independent facilitator familiar withadaptive management and policy planning can reduce thelength of the process and time commitments. This mod-erator needs to be meticulous in recording details ofprocess developments, obtaining commitments fromagencies and organizations at the beginning of theprocess. Timelines for implementing management strate-gies, including monitoring, evaluation, and review of theadaptive process, and commitment of funds from groupsinvolved must be identified and outlined up front in man-agement or conservation plans being developed. Thiswill provide structure to the adaptive managementprocess and ensure long-term involvement of individualsand commitment of financial resources from each group,even when personnel changes occur. We believe that ifindividuals and organizations involved with managingprairie grouse populations adopted similar methods,learning about prairie grouse populations would beadvanced and habitat management problems could beaddressed more effectively.

Once stakeholders come together, the first step is todefine policy options and identify policy performance

measures (Walters 1997b). This is where managementoptions are identified and uncertainties about those out-comes and consequences are outlined. These options canvary over space and time; thus it is useful to define aballpark scale for treatment comparisons. If managementis directed at a single species, one must consider thescale(s) at which the species is likely to respond to themanagement. For example, if one implemented cattleexclosures to enhance prairie grouse nesting habitats,careful consideration would have to be given to the sizeof patch managed that may elicit responses by grouse.Implementing grazing practices can be difficult at scalesother than those at which pastures are typically managed.Thus, one must think about the scale at which manage-ment occurs, as well as the scale at which a biologicallymeaningful response is likely to occur. The goal of adap-tive management should be to seek out untested optionsand to evaluate new methods, not to find a cookbookbest-prescription. Policy planning should include per-formance measures for each strategic option (Walters1997b). This likely will include measures such as eco-nomic costs and benefits, ease of implementation, likeli-hood of success, and others, as well as the original bio-logical goals (Walters 1997b).

The second step in the policy-planning stage is toidentify major uncertainties by trying to predict the out-comes of policy alternatives (Walters 1997b). The candi-date set of policy options identified in the first step isused to generate models and simulations to predict eachpolicy’s impact on desired performance measures(Walters 1997b). These predictions should not be consid-ered as simple static comparisons; they need to be tempo-ral predictions, because uncertainties in the importance oftime scales will inevitably arise (Walters 1997b). Thisprocess should remove policies not directly relevant tothe questions at hand and those likely to fail to beginwith. In the sage-grouse case study, the Native PrairieStewardship Program attempted to improve range habitatconditions for livestock (and wildlife) by creating severalwatering sources in upland habitats to more evenly dis-tribute cattle. While this might have reduced grazingimpacts in important mesic sage-grouse brood-rearinghabitats, increasing cattle activities in upland areas couldhave negatively affected nesting success for sage-grouse.Thus, this management strategy would likely have beenidentified early during the planning process as problemat-ic and been modified or abandoned altogether.

The third step in Walters’ (1997b) process involvesscreening of models to define a good set of policy treat-ments. This is the process of weeding out policies thatare not worth testing because they may have hidden pit-falls or deleterious cumulative impacts that would pre-


vent the achievement ofmanagement goals. Themodeling processes iden-tified in step 2 will helpto identify these hiddenpitfalls.

Step 4 involves parti-tioning the landscape intoexperimental units atscales appropriate to theuncertainties (Walters1997b). Walters suggeststhat 2 questions be posedat this stage: 1) What por-tion of the land should bedevoted to each of thebasic treatment regimes?and 2) How large shouldeach experimental unitbe? This is also the stageat which uncertaintiesabout the size of experi-mental treatments can be tested and included as an experi-mental unit. For our sage-grouse grazing-intensity exam-ple, we proposed to implement several treatments withdifferent levels of grazing pressure for comparison.Concurrently, or as an independent experiment, tests forthe optimal treatment patch size could be conducted. Theamount of replication would depend on the size of experi-mental units, potentially limiting the number of manage-ment policies that could be tested. With an endangeredprairie grouse population, the population could be dividedinto an experimental unit and a control. However, this iswhere implementing the BACI design (Green 1979) mayalso be useful.

Next, the temporal and spatial scales at which keyresponses should be monitored must be considered(Walters 1997b). This is where science and managementoften are in direct conflict, but where they should be inharmony (Johnson et al. 1997). The priority should be todevelop a sound and replicated experimental design thataddresses, as directly as possible, the question(s) at hand.One must also remember that the time scale is an issueand answers to management questions may not be avail-able immediately after implementation. This is wherelong-term commitments of individuals, organizations,and funds, as identified early in the planning stages, canensure the successful identification and thus implementa-tion of adaptive management.

Walters (1997b) encourages the use of AdaptiveEnvironmental Assessment Modeling (AEAM; Holling1978, Walters 1986) as the sixth and final step in his pol-

icy-planning process. Adaptive EnvironmentalAssessment Modeling is a stakeholder-involved evalua-tion process that uses simulation modeling to predictfuture conditions (Holling 1978, Walters 1986).Maintaining the involvement of a wide range of stake-holders gives each of them ownership in the plan andstrengthens the adaptive management process.

For both of the prairie grouse case studies we present-ed, attempts to implement sound adaptive managementexperiments failed. This is not uncommon. Mostattempts to implement adaptive management in a varietyof different systems tend to fail in the planning stages,prior to implementing experiments (Walters 1997b).However, by following the policy-planning steps we havepresented, common mistakes in the planning stages can beavoided and the most appropriate management policiesmeeting biological, economic, social, and political goalscan be identified, advancing the adaptive managementprocess to the experimental phase, where managementpolicies are implemented. By definition, adaptive man-agement is an evolving process. To complete the fulladaptive cycle, management strategies must be evaluatedfollowing the monitoring strategies outlined in the plan-ning process. As new understandings occur, managementstrategies need to evolve and incorporate new information.

ConclusionsTo our knowledge, experimental adaptive management

has yet to be successfully used as a tool to advance learn-


Aerial display of the Sharptails Plus Foundation project at the Narcisse Wildlife Management Area, Manitoba.Photo by Patrick J. Caldwell.

ing and management of prairie grouse populations. Morebroadly, adaptive management seldom is successful withabout 75% of policy-planning processes failing to imple-ment adaptive management as experiments and less than10% possessing sound experimental design with suitablereplication and controls (Walters 1997a). Although adap-tive management has become a popular idea among man-agement agencies and can be useful when implementedcorrectly, in practice it is often used as a buzzword andnever implemented. Ludwig and Walters (2002) suggestthat this may be a defensive measure by bureaucratsattempting to demonstrate that change is occurring with-out actually changing anything. Adaptive managementsets out to embrace uncertainties within a particular sys-tem, resulting in an increased understanding of the sys-tem by managing through experiments (Holling 1978,Halbert 1993). This allows appropriate common-sensepolicy decisions to be identified (Ludwig and Walters2002).

Adaptive management can be expensive, but balanceshave to be achieved between contributing available fundstoward habitat protection and allocating funds for adap-tive management (Wilhere 2002). For adaptive manage-ment to succeed, there must be a commitment to long-term monitoring and evaluation that promotes learningand feedback into management strategies. Walters andGreen (1997) suggest that approximately 20% of fundsallocated to any management plan need to be set asidefor monitoring and evaluation. Unfortunately, most habi-tat conservation plans lack sufficient monitoring to evalu-ate their success (only about 5% of plans created in 1999had a suitable monitoring component [Kareiva et al.1999]). Large government conservation initiatives typi-cally do not promote the inclusion of monitoring in man-agement plans. For instance, in 2000 the CanadianWildlife Service implemented a 5-year, $45 millionHabitat Stewardship Program aimed at protecting at-riskspecies and their habitats through conservation actions.However, guidelines for the program prohibit the use offunds to evaluate the strategies. The importance of moni-toring and evaluation in management is frequently mis-understood, and failure to document consequences ofmanagement often results in the failure of many adaptivemanagement plans. The most successful informationtechnology companies in the world (XEROX, Kodak,IBM, and AT&T) recognize the importance of researchand development, typically reinvesting about 10% of cor-porate earnings back into research and development (Gill1997).

The management of prairie grouse needs an adaptiveframework, and we recommend using Walter’s (1997b)adaptive policy-planning process as a guideline to ensure

that the process does not fail in the planning stages andthat sound management options are identified. Thisprocess will enhance the probability of sound manage-ment experiments being implemented with suitable moni-toring and evaluation components. Both case studies wepresent here are excellent examples of how lack of acoordinated effort will give rise to independent and possi-bly harmful management policies that lack scientificrigor and appropriate monitoring and evaluation compo-nents. However, the Alberta sage-grouse example alsoshows that following an adaptive policy-planning process(the RAG) can prevent the most common problemsstalling the adaptive process (involving stakeholders,identifying and implementing sound experimental poli-cies) and allow policy alternatives to be modeled andcompared, thus identifying uncertainties. This processwill allow management to be implemented as sound sci-entific experiments, the goal of adaptive management.

Conservation plans developed by groups of stakehold-ers to identify management priorities have been undertak-en for many prairie grouse populations (C. E. Braun,Grouse Inc., personal communication), an important ini-tial step in the adaptive management process. However,like most habitat conservation plans, prairie grouse planstypically lack suitable monitoring and evaluation pro-grams. Even if working groups successfully identify andimplement conservation actions, uncoordinated imple-mentation of actions has made it impossible to identifyappropriate methodologies to assess impacts on popula-tion numbers, as with the Gunnison sage-grouse conser-vation plan (Anonymous 1997). Structuring managementpolicies as adaptive experiments should be a priority forindividuals and organizations managing prairie grousepopulations.

Acknowledgments. This research was supported finan-cially and logistically by the Alberta ConservationAssociation; Alberta Sport, Recreation, Parks andWildlife Foundation; Alberta Sustainable ResourceDevelopment; Cactus Communications (Medicine Hat,Alberta); Canadian Wildlife Foundation; ChallengeGrants in Biodiversity (University of Alberta); DucksUnlimited Canada (North American WaterfowlManagement Plan); Endangered Species Recovery Fund(World Wildlife Fund Canada and the Canadian WildlifeService); Esso Imperial Oil (Manyberries, Alberta);Manitoba Department of Conservation; Murray ChevroletOldsmobile Cadillac (Medicine Hat, Alberta); a NaturalSciences and Engineering Research Council of CanadaPost-Graduate Scholarship to C. L. Aldridge; theSharptails Plus Foundation; the University of Alberta; theUniversity of Manitoba; the University of Regina; and


Wildlife Habitat Canada. We thank D. Eslinger (AlbertaSustainable Resource Development), S. E. Nielsen andD. J. Saher (University of Alberta), W. D. Svedarsky, andone anonymous reviewer for improving previous versionsof this manuscript.

LLiitteerraattuurree cciitteeddALDRIDGE, C. L.,AND R. M. BRIGHAM. 2001. Nesting and reproductive activ-

ities of greater sage-grouse in a declining northern fringe population.Condor 103:537–543.

ALDRIDGE, C. L.,AND R. M. BRIGHAM. 2002. Sage-grouse nesting and broodhabitat use in southern Canada. Journal of Wildlife Management 66:433–444.

ALDRIDGE, C. L., AND R. M. BRIGHAM. 2003. Distribution, status and abun-dance of Greater Sage-Grouse, Centrocercus urophasianus, inCanada. Canadian Field Naturalist 117:25–34.

ANDERSON, D. R., K. P. BURNHAM, AND W. L.THOMPSON. 2000. Null hypothe-sis testing: problems, prevalence, and an alternative. Journal ofWildlife Management 64:912–923.

ANONYMOUS. 1997. Gunnison sage grouse conservation plan. ColoradoDivision of Wildlife, Gunnison, USA.

ANONYMOUS. 2000. Canadian sage grouse recovery strategy. TheCanadian Sage Grouse Recovery Team. Alberta Natural ResourcesService, Edmonton, Canada.

ANONYMOUS. 2002. Gunnison Sage-Grouse conservation plan summaryof evaluation. Bureau of Land Management, Gunnison, Colorado,USA.

APPLEGATE, R.A., R. K. BAYDACK, K. M. GIESEN, M.A. MORROW, M. C. PETERSON,AND M. A. SCHROEDER. 2000a. Greater prairie-chicken status surveyand conservation action plan 2000–2004. International Union forthe Conservation of Nature and Natural Resources, The WorldConservation Union, WPA/BirdLife/SSC Grouse Specialist Group,Gland, Switzerland.

APPLEGATE, R.A., R. K. BAYDACK, K. M. GIESEN, AND M. G. SCHROEDER. 2000b.Lesser prairie-chicken status survey and conservation action plan2000–2004. International Union for the Conservation of Nature andNatural Resources,The World Conservation Union,WPA/BirdLife/SSCGrouse Specialist Group, Gland, Switzerland.

BAYDACK, R. K. 1996. Habitat management for sharp-tailed grouse on pri-vate lands in Manitoba. Pages 20–24 in Proceedings of the SeventhInternational Grouse Symposium; 20–24 August 1996, Fort Collins,Colorado, USA.

BERGER, R. P., AND R. K. BAYDACK. 1993. Effects of aspen succession onsharp-tailed grouse in the Interlake region of Manitoba. CanadianField-Naturalist 106:185–191.

BIRD, R. D. 1961. Ecology of the aspen parkland of western Canada inrelation to land use. Queen’s Printer, Ottawa, Ontario, Canada.

BORMANN, B. T., J. R. MARTIN, F. H. WAGNER, G. W. WOOD, J. ALGERIA, P. G.CUNNINGHAM,M.H.BROOKES,P.FRIESEMA, J.BERG,AND J.R.HENSHAW. 1999.Adaptive management. Pages 505–534 in W.T. Sexton,A. J. Malk, R. C.Szaro, and N. C. Johnson, editors. Ecological Stewardship. A commonreference for ecosystem management. Elsevier Science Limited,Oxford, England.

BOYCE,M.S. 1993. Population viability analysis:adaptive management forthreatened and endangered species. Transactions of the 58th NorthAmerican Wildlife and Natural Resources Conference 58:520–527.

BRAUN, C. E. 1998. Sage grouse declines in western North America: whatare the problems? Proceedings of the Western Association of StateFish and Wildlife Agencies 78:139–156.

BRAUN, C. E., K. MARTIN, T. E. REMINGTON, AND J. E. YOUNG. 1994. NorthAmerican grouse: issues and strategies for the 21st century.Transactions of the 59th North American Wildlife and Natural

Resources Conference 59:428–438.CHAMBERLIN, T. C. 1890. The method of multiple working hypotheses.

Science 15:92–97.CONNELLY, J.W., M.W. GRATSON,AND K. P. REESE. 1998. Sharp-tailed Grouse

(Tympanuchus phasianellus).Account # 354 in A. Poole and F. Gill,editors. The Birds of North America. Academy of Natural Sciences,Philadelphia, Pennsylvania, and Amercan Ornithologists’ Union,Washington D.C., USA.

GILL, R. B. 1997. Wildlife research: looking backward to see our future.Pages 1–17 in Proceedings of the Colorado Chapter of The WildlifeSocieties’ annual meeting, 23 January 1997, Fort Collins, USA.

GREEN, R. H. 1979. Sampling design and statistical methods for environ-mental biologists. Wiley Interscience, Chichester, England.

FREYFOGLE, E. F. 1998. Bounded people, boundless lands. Island Press,Washington, D.C., USA, and Sheatwater Books, Covelo, California,USA.

FROESE, J. J. 2002. Determining effective strategies for conserving bio-logical diversity in Manitoba’s Interlake. Thesis, University ofManitoba,Winnipeg, Canada.

HALBERT, C. L. 1993. How adaptive is adaptive management?Implementing adaptive management in Washington Sate and BritishColumbia. Reviews in Fisheries Science 1:261–283.

HANEY,A.,AND R.L.POWER. 1996. Adaptive management for sound ecosys-tem management. Environmental Management 20:879–886.

HILBORN, R. 1992. Can fisheries agencies learn from experience.Fisheries 17:6–14.

HOLLING, C. S. 1978. Adaptive environmental assessment and manage-ment. John Wiley and Sons, New York, New York, USA.

HURLBERT, S. H. 1984. Pseudoreplication and the design of ecologicalfield experiments. Ecological Monographs 54:187–211.

JOHNSON, F.A., M. D. KONEFF, M. G.ANDERSON, R. O. BAILEY, R. K. BAYDACK,T. E.MARTIN, J.W. NELSON, J. K. RINGELMAN, AND C. RUBEC. 1997. Enhancingbiological performance of the North American waterfowl manage-ment plan. Transactions of the 62nd North American Wildlife AndNatural Resources Conference 62:377–385.

KAREIVA, P. M., S. ANDELMAN, D. DOAK, B. ELDERD, M. GROOM, J. HOEKSTRA, L.HOOD, J. FRANCES, J. LAMOREUX, G. LEBUHN, C. MCCULLOCH, J. REGETZ, L.SAVAGE, M. RUCKELSHAUS, D. SKELLY, H. WILBUR, AND K. ZAMUDIO. 1999.Using science in habitat conservation plans. American Institute ofBiological Sciences,Washington, D.C., USA.

LUDWIG, D., AND C. J.WALTERS. 2002. Fitting population viability analysisinto adaptive management. Pages 511–520 in S. R. Bessinger, and D.R. McCulloug, editors. Population viability analysis. University ofChicago Press, Chicago, Illinois, USA.

MACNAB, J. 1983. Wildlife management as scientific experimentation.Wildlife Society Bulletin 11:397–401.

MOIR, W. H., AND W. M. BLOCK. 2001. Adaptive management on publiclands in the United States: commitment or rhetoric? EnvironmentalManagement 28:141–148.

REED, J.M.,AND A.R.BLAUSTEIN. 1995. Assessment of nondeclining amphib-ian populations using power analysis. Conservation Biology 9:1299–1300.

ROE, E.,AND M.VAN EETEN. 2001. Threshold-based resource management:a framework for comprehensive ecosystem management.Environmental Management 27:195–214.

WALTERS, C. J. 1986. Adaptive management of renewable resources.McGraw Hill, New York, New York, USA.

WALTERS,C. J. 1997a. Challenges in adaptive management of riparian andcoastal ecosystems. Conservation Ecology 1(2):1. Available online athttp://www.consecol.org/vol1/iss2/art1/index.html (accessed 01August, 2002).

WALTERS, C. J. 1997b. Adaptive policy design: thinking at large spatialscales. Pages 386–394 in J. A. Bissonette, editor. Wildlife and land-scape ecology: effects of pattern and scale. Springer, New York, NewYork, USA.



WALTERS, C. J., AND R. GREEN. 1997. Valuation of experimental manage-ment options for ecological systems. Journal of Wildlife Management61:987–1006.

WALTERS C. J.,AND R. HILBORN. 1976. Adaptive control of fishing systems.Journal of the Fisheries Research Board of Canada 33:145–159.

WILHERE,G.F. 2002. Adaptive management in habitat conservation plans.Conservation Biology 16:20–29.

Cameron L. Aldridge (above) has been studying sage-grouse ecologyin Alberta for the past 8 years. He obtained a B.S. in zoology andecology from the University of Calgary and an M.S. in biology fromthe University of Regina. He is currently a doctoral candidate and anIzaak Walton Killam Predoctoral Scholar at the University of Alberta,where he is studying the population viability of sage-grouse in Alberta.His research interests include population ecology and wildlife man-agement of vertebrates, conservation biology, and endangered speciesmanagement. He has been a member of The Wildlife Society since1999, and is a past executive member of the TWS Alberta StudentChapter. Mark S. Boyce (at top, right) is professor of biological sci-ences and holds the Alberta Conservation Association chair in fish-eries and wildlife at the University of Alberta. He obtained a B.S. infish and wildlife biology from Iowa State University, M.S. in wildlifemanagement from the University of Alaska–Fairbanks, and M.Phil. andPh.D. degrees from Yale University. He is past president of the YaleStudent Chapter and the Wisconsin Chapter of TWS. He is presidentof the TWS Alberta Chapter for 2003–04. During 1996–97 he servedTWS as Editor-in-Chief of The Journal of Wildlife Management. Markis a certified wildlife biologist. His research interest is wildlife popula-tion ecology. Richard Kenith Baydack (below, right) has over 25 years

of research, teaching, and consulting experience in wildlife manage-ment and conservation. He has been on faculty with the University ofManitoba since 1979. Dr. Baydack is currently Associate Dean of thenew Faculty of Environment and has been acting head of theDepartment of Geography, graduate chair of the Faculty ofEnvironment, and Associate Director of the Natural ResourcesInstitute. He holds a B.Sc. (Honors) and a Master of NaturalResources Management degree from the University of Manitoba, anda Ph.D. in fishery and wildlife biology from Colorado State University.Professionally, Rich has been president of the Manitoba Chapter, vice-president of the Central Mountain & Plains Section, and chair of theBiological Diversity Working Group of The Wildlife Society. Currently,he is chair of the Program Committee for the 11th Annual Conferenceof TWS, being held in Calgary, Alberta in September 2004. Rick is acertified wildlife biologist.

Special section associate editor: Silvy

adaptive management of prairie grouse:how do we get … management of prairie grouse • aldridge et...

Documents