Flammulated, Boreal, and Great Gray Owls in The

Chapter 12 Information Needs: Boreal Owls

Gregory D. Hayward, USDA Forest Service, Rocky Mountain Forest and Range Experiment Station, Laramie, WY 82070

INTRODUCTION servation strategies. In the western United States (excluding Alaska), our knowledge is sufficient to Most humans are reluctant to make decisions with- predict, on a broad scale, where boreal owls are likely out thorough knowledge of the consequences of to occur and thereby indicate to land managers those decisions. Therefore, a desire for further re- where management of the species should be a con- search is almost universal in any complex manage- cern (see Map 2). In Alaska and east of the Rocky ment arena. The value of further study is deter- Mountains, particularly in the Great Lakes region mined, in part, by the cost in time and resources to and New England, understanding of the species dis- obtain new information and the cost of making in- tributional status is very weak. Knowledge of dis- correct decisions without the desired knowledge. In tribution is growing rapidly, however, and land the case of managing forest lands and conserving management agencies have the capability to obtain boreal owls, our knowledge is so limited that well the necessary information in several years. Methods directed research on distribution, habitat use at many for determining boreal owl distribution (not abun- scales linked with studies of local demography, for- dance) using playback surveys have been developed est history, and interactions with competitors and and are being used by managers. predators can further our understanding at little expense compared to the potential benefits of man- Response of Boreal Owls to Stand-Level aging forests without adverse effects on boreal owls Habitat Changes and the larger subalpine ecosystem. In this chapter, I will assess the strength of exist- Current knowledge indicates that boreal owls in ing knowledge in forming a conservation strategy. some geographic areas use mature and older forest Then I will discuss some strategies for obtaining habitats for critical life functions. Land managers, needed information. This chapter will include few however, must predict the response of particular citations because it relies on Chapters 9 and 10 which boreal owl populations to various management ac- review literature on this species and some of the for- tions in order to assess the consequences of alterna- ests it inhabits. tives. Our knowledge is not sufficient to complete this task. STRENGTH OF EXISTING KNOWLEDGE Our understanding of habitat use by boreal owls TO FORMULATE A does not allow a ranking of habitats in terms of qual- CONSERVATION STRATEGY ity. Knowledge of habitat use based solely upon ob- servations of habitat use will not be sufficient to Throughout this section I approached each topic guide habitat management. Rather, the relative pro- as a question. Is our understanding of this topic suf- ductivity of owl populations among different habi- ficient to support the development of a sound, na- tats must be discerned in order to provide a ranking tional-scale conservation strategy for the boreal owl? system with which to weigh alternative land man- For example, "is the response of boreal owls to stand agement schemes. level habitat change understood in sufficient detail Even habitat use data are unavailable for most to formulate a conservation strategy?" geographic areas within the species' range. Geo- graphic variation in habitat quality must be docu- Distribution mented and developed into a ranking among habi- tats. Furthermore, proximate and ultimate factors Current knowledge of boreal owl distribution is determining the observed pattern of owl success not sufficient to support the development of con- among habitats must be understood in order to pre- dict habitat use patterns for locales other than those Understanding how owl movements change when studied. In other words, the relationship between individuals encounter landscape boundaries of dif- the owl and its prey, primary cavity nesters, poten- ferent types and the role of corridors in dispersal tial predators, and competitors and habitat relation- will facilitate building models to predict the impact ships of each must be addressed to formulate a pre- of different harvest prescriptions on foraging suc- dictive model useful in evaluating management al- cess. Coordinated investigations examining the re- ternatives. lationship between habitat condition and prey abun- dance would increase the predictive power of such Effects of Landscapsscale Changes on a model. Home Range Use Demography Boreal owl studies have not examined the influ- ence of habitat characteristics at the landscape scale Knowledge of boreal owl demography is not suf- on home range size or the success of individuals. ficient to estimate persistence of either local popu- The literature demonstrates that under some circum- lations or metapopulations. Reproductive rates of stances boreal owls occupy home ranges that exceed boreal owls and the factors determining productiv- 2,000 ha and therefore indicates that management ity are documented for populations in Europe, and for boreal owls must consider extensive landscapes. productivity has been observed for several popula- The available studies do not aid managers in evalu- tions in North America. Age-specific survival and ating different mixes of forest age classes or provide factors influencing survival are virtually unknown. a sound basis for predicting local or regonal varia- Likewise, age-specific dispersal and the factors in- tion in home range use. fluencing immigration and emigration are not un- derstood. The interaction between environmental Effects of Regional-Scale Habitat Changes conditions (forest structure, prey populations, preda- on Movement Patterns tors, competitors, landscape patterns) and demo- graphic parameters must be understood prior to Annual and seasonal movement patterns of bo- developing comprehensive management plans for real owls in North America are virtually unknown. boreal owls. This does not mean that management Without this knowledge the demographic interac- is not possible without data on the demography of tions among populations forming the larger every target population. Rather, patterns of change metapopulation cannot be discerned. Therefore, the in demography across the species' range should be regional effects of local management actions cannot understood so that a reductionist approach to man- be addressed. For example, current knowledge of agement is unnecessary. owl movements is not sufficient to determine Methods to monitor the trend of boreal owls in whether management in montane forests below the selected populations will be necessary to evaluate subalpine zone will influence movements between management practices and facilitate adaptive man- populations. agement responses. Development of methods other than playback surveys has begun (Hayward et al. Effects of Stand-Level Habitat Changes on 1992) but these methods must be validated and re- Foraging Behavior fined if they are to be used effectively by manage- ment. Boreal owl foraging behavior and habitat use is very poorly understood. For instance, the response Dynamics of Primary Plant Communities of foragng owls to landscape boundaries (meadow or clearcut edges, changes in tree density) have not Knowledge of the successional dynamics of sub- been observed. The interaction between prey abun- alpine and boreal forests, while not complete, is ex- dance and habitat structure in determining prey tensive and will facilitate prediction of future forest availability is also poorly understood. The impor- conditions, particularly as that knowledge relates to tance of understanding the foraging ecology of bo- the dominant tree species. Forest ecologists under- real owls cannot be over emphasized. Studies in stand many of the factors related to the occurrence Europe and North America indicate populations are of major natural disturbance agents (fire, insects, frequently food limited. Prey abundance and forag- windfall, and disease). Factors influencing seed ing habitat condition together determine population crops, regeneration and survival of trees, gap dy- trend for many populations. namics, and material cycling have been studied. The 149 stochastic nature of disturbance agents have been knowledge of forest dynamics. Instead, in the sec- incorporated into conceptual models of forest dy- tion that follows, I provide direction in research phi- namics at the stand and landscape scale. losophy and ideas for research approaches. My goal Understanding of the dynamics of forest species is to outline the scope of investigative approaches other than the dominant tree species is less complete. that will be necessary to obtain the knowledge high- The sera1 development of the detritus system, so lighted above. important to subalpine forests, is not understood. Similarly, the dynamics of fungi and lichen popula- Integrate a Variety of Research Tools tions that are important to small mammals have re- ceived little attention. Knowledge of understory- A research program designed to obtain the knowl- plant population dynamics is incomplete. These ar- edge necessary to build a conservation strategy for eas of forest ecology need further attention in order boreal owls will require the integration of several to understand the dynamics of small mammal popu- research approaches. Modeling, field experiments, lations. and observational studies must be integrated so the The population dynamics (production, survival) knowledge stemming from each approach contrib- of tree cavities used by boreal owls is also poorly utes toward corroborating or refuting key hypoth- understood. Biologists do not understand the char- eses. It will be necessary to coordinate efforts using acteristics of forest stands (other than tree size and each of these tools in several strategically selected presence of primary cavity *excavators)that deter- geographic locales in order to elucidate the ecology mine the production of tree cavities. The role of tree of boreal owls in particular environmental settings disease (fungi, insects), wind damage, and moisture and describe how the species' ecology varies across conditions in cavity production are not understood. environmental gradients. In addition to employing Likewise, what factors determine the lifespan of tree a range of research approaches, studies must be car- cavities has received limited attention. ried out across a range of spatial scales. Below I will describe how I envision the employment of these History of Distribution and Composition research tools to obtain the information necessary of Forest Communities to manage boreal owls over a broad geographic area.

Forest ecology has recently directed more atten- Modeling tion toward the historical ecology of forests. Despite Boreal Owls in the Subalpine Forest.-The biologi- these efforts, synthesis of data on the historic abun- cal and ecological topics identified in the first por- dance and distribution of subalpine and boreal for- tion of this chapter represent interesting research ests is not sufficient to indicate how current trends topics on their own but become significant when compare to the past. In particular, knowledge of integrated as a body of knowledge to understand pattems in distribution and abundance of older age the subalpine forest system in which the boreal owl classes of these forests has not been synthesized. exists. The envirogram depicted in figure 5, Chap- To aid in conservation planning for boreal owls, ter 9 represents a hypothesis of how the components historic ecology must include investigations exam- of the subalpine forest relate to the ecology of bo- ining patterns in these forests thousands of years ago real owls. This envirogram should be elaborated into as well as in recent centuries. The value of the his- a model describing the processes that link the bo- toric information lies in the perspective it can pro- real owl to other components of the subalpine forest vide on the potential variation in boreal owl distri- system. The model should predict how changes in bution in the past and therefore the range of condi- subalpine forests (e.g., biomass of logs, abundance tions that the species faced. of woodpeckers, or size class distribution of trees) would impact boreal owl populations. Such a model A RESEARCH DIRECTION would begin as a qualitative representation of the system generating specific hypotheses that could be The assessment of current knowledge points to tested through field research. As particular linkages major topics that must be addressed in future re- are examined, portions of the model should be re- search designed to provide the basis for a conserva- fined into quantitative expressions of the system. tion strategy. I will not set priorities among these Without such a predictive model, research at the themes because many of the topics are parallel. For stand and watershed scales will lack focus and likely instance, knowledge necessary to rank habitat qual- not concentrate on key characteristics of the owl/ ity will not improve habitat management without forest system. A well developed and tested model would ultimately provide managers with a tool to layers built into the Gap data bases, the predicted evaluate the consequences of management on the distribution of boreal owl populations would be es- boreal owl through the impacts on the forest sys- timated based on a series of geographically tuned tem. owl habitat models. Individual-Population Demographic Models.- The metapopulation model would be a valuable Although reliable estimates of mortality rates, im- tool for management and research. Managers could migration rates, and emigration rates are not avail- immediately employ the geographic data in biolog- able, preliminary modeling of boreal owl cal evaluations or other impact assessments. They demography will be useful in focusing research and could also use the model to predict the location of refining the assessment of conservation status. Pre- unidentified owl populations and test those predic- liminary estimates of demographic parameters tions through surveys. Research would use the (which are available - see Chapter 9) would per- model to aid in setting priorities among research mit examination of life history sensitivities. Based topics based on a sensitivity analysis. The model on these analyses, and refinements of the models as could also aid in choosing study areas based on more precise demographic data are gathered, man- which owl populations are predicted to be key popu- agers will become aware of aspects of the owl's life lations for the persistence of boreal owls. history that are more sensitive to change. Through use of these demographic models, assessments of Experimental Studies Using Management persistence probabilities will be refined and man- Treatments agement alternatives more easily compared. Experimentation is the workhorse of modern sci- Examination of individual-population demo- ence. Experiments provide the most straight-forward graphic models would aid in understanding the rela- methods to examine cause-and-effect and to deter- tive quality of different habitats for boreal owls. mine the processes responsible for patterns observed Habitat quality is most directly expressed through in natural history investigations. Unfortunately eco- population trend (quality habitat leading to h > 1). logical experiments employing appropriate Sam- Through the joint examination of local demography pling designs and proper controls are difficult to and habitat characteristics, poor and good quality design and execute (Hairston 1989), especially in habitat may be identified, providing important in- studies of large, low density mobile vertebrates. formation for managers (see Small Scale Coordinated efforts between management and re- Demography below). search can overcome some of the logistic difficulties Spatially Explicit Metapopulation Models.-Based in producing large scale treatments and measuring on the distribution and life history of boreal owls, the effects. Hayward et al. (1993) hypothesized that boreal owls Experiments will be necessary to determine the occur in the United States as a strongly structured response of major prey populations to various tim- metapopulation. This hypothesis must be tested us- ber harvest techniques and to rank the quality of owl ing observational field studies documenting the de- habitats. Some of these experiments will require gree of demographic linkage among boreal owls landscape scale treatments that are routinely carried occupying separate patches of habitat. Based on this out by management. In order to use timber sales as hypothesis, a metapopulation model for the Rocky treatments, however, researchers must determine Mountains, Blue Mountains, and Northern Cascades (using a well defined sampling protocol) how treat- should be built. This model would integrate with ments are assigned to experimental units (stands or field studies on boreal owl dispersal, local larger land units) and what treatments will be em- demography and habitat use in an interactive sense. ployed. Field results would be used to parameterize the The link between arboreal lichen and boreal owls model while the model would be employed to fo- depicted in the Envirogram (figure 5, Chapter 9) cus field studies geographically and toward particu- provides an example of the power of experimenta- lar questions. The model would be tested and re- tion in developing knowledge necessary to manage fined as field data accumulated. boreal owls. The envirogram leads to the hypoth- The Geographic Information Systems (GIs) data esis that red-backed voles (a major winter and sum- base necessary to develop such a model would be mer prey of boreal owls) will be more abundant in expensive and time consuming to develop. How- forests with greater biomass of Bryoria sp. and re- ever, the U.S. Fish and Wildlife Service's Gap Analy- lated lichen. This hypothesis can be tested through sis program could provide the necessary data at the a series of laboratory (feeding trial) and field (lichen proper scale. Using the vegetation and topographic removal and addition) experiments. As demonstrated by Korpimaki (1988,l 989), nest data necessary to answer questions at this scale will boxes can be employed to facilitate experiments with require an extensive system of nest boxes facilitat- boreal owls. Hayward et al. (1993) employed nest ing studies in a range of geographic settings. Hay- boxes in a small experiment examining the relative ward et al. (1992) describe such a system that is simi- importance of forest structure vs. cavity availability lar to the programs established by successful scien- in nest site selection by boreal owls. Linkages be- tists in Europe (e.g., Korpimaki 1981, Sonerud 1985). tween habitat structure and nesting of boreal owls Research at the regional scale will largely involve should be further explored through nest box experi- integrating knowledge obtained at finer scales into ments. metapopulation models and then testing the predic- tions of those models. Empirical information on dis- Observational Studies persal of juvenile and adult boreal owls will be criti- Many of the gaps in our knowledge cannot be filled cal to developing any metapopulation model. Exer- through experimental investigation. The historical cising the metapopulation models to predict future ecology of forests used by boreal owls will be learned conditions for boreal owls will require sound, spa- through careful "sleuthing." Hypotheses may be tially specific information from land management posed and observations (pollen sediment patterns, plans. Development of sound metapopulation mod- etc.) used to determine the legitimacy of those hy- els that examine patterns of boreal owl occurrence potheses. Similarly dispersal patterns of boreal owls over multiple Forest Service Regions will be critical will be learned through "simple" observation. Stud- in coordinating management and determining broad ies of dispersal must discern how dispersing birds research priorities. react to edges, what habitats they will fly across, etc. Observational studies will also be essential in pro- lnvestigate Geographic Variation viding parameter estimates for modeling efforts at Understanding patterns of variation in boreal owl many scales and in forming hypothesis to be tested habitat use, demography, population regulation, through experimentation. As a consequence of the movement patterns, and food habits will produce large home ranges of individual boreal owls, and many of the insights necessary to manage the spe- because of the difficulties encountered in large scale cies. In European boreal owls, productivity and experiments, most field research on boreal owls will movements vary geographically. This variation has involve observational investigations. led to interesting hypotheses concerning the role of various biotic and abiotic factors in boreal owl life Beyond Tools history.

Examine Boreal Owl Ecology at a Range lnvestigate Many Questions on the of Scales Same Site As indicated earlier, a sound conservation strat- A complex web of interactions is described in the egy will require understanding the response of bo- ecological web depicted in figure 5, Chapter 9. Un- real owls to habitat change on a range of geographic derstanding this web will require the integration of and temporal scales. Research approaches must dif- studies involving many disciplines -plant ecology, fer across scales involving different mixes of model- fire ecology, mycology, mammalogy, and commu- ing, experimentation, and observational studies. nity ecology to name just a few. Focusing investiga- Studies that examine hypotheses stemming from tions from many disciplines on common sites will models built from an expansion of the envirogram facilitate understanding of the links in this web. presented in Chapter 9 will largely focus on ecologi- These studies should be developed using techniques cal patterns and processes at relatively fine scales and approaches employed in population and com- (forest stands and small watersheds encompassing munity ecology. one to several square km). At this scale, some ex- For example, population and community level in- perimentation will be possible. vestigations of small mammals should be conducted At a broader geographic scale, the response of bo- on the same sites as studies of owls. An understand- real owls (home range characteristics, population ing of small mammal population dynamics and habi- density, productivity movements) to landscape pat- tat selection will require information on forest struc- terns must be examined. Individual-population dy- ture and dynamics including understory vegetation, namics models that examine the sensitivity of vari- fungi, and lichen. The necessary coordination among ous life history characteristics will help focus this disciplines is obvious. research. Gathering the population and behavior Coordinating interdisciplinary research is not Korpimaki, E. 1981. On the ecology and biology of trivial, however. Much has been written about the Tengmalm's Owl (Aegolius funereus) in south- value of interdisciplinary studies but far less pub- ern Ostrobothnia and Suomenselka, western lished resulting from such work. Initiating coordi- Finland. Acta Universitatis Ouluensis Series A nated investigations to understand boreal owl ecol- Scientiae Rerum Naturalium Biologics 118:13. ogy may not be easy. By initiating some studies at Korpimaki, E. 1988. Costs of reproduction and existing, long-term ecological research sites, some success of manipulated broods under varying of the logistic problems could be solved. food conditions in Tengmalm's Owl. Journal of Animal Ecology 57:1027-1039. Small-Scale Demography Korpimaki, E. 1989. Breeding performance of Tengmalm's Owl Aegolius funereus: effects of Although ignored for a while in applied wildlife supplementary feeding in a peak vole year. Ibis research, demographic studies are becoming a cor- 131:51-56. nerstone upon which conservation programs are Lofgren, O., B. Hornfeldt, and B. G. Carlsson. 1986. built (Dennis et al. 1991, Verner et al. 1992). Exami- Site tenacity and nomadism in Tengmalm's nation of population trend and analysis of the sen- Owl (Aegoliusfinereus (L.)) in relation to cyclic sitivity of h to variation in survival and reproduc- food production. Oecologia (Berlin) 69:321-326. tion have been the focus of some investigations. In Sonerud, G. A. 1985. Nest hole shift in Tengmalm's some efforts, estimation of h for whole populations Owl Aegolius funereus as defense against nest was sought. predation involving long term memory in the A potentially more efficient approach would esti- predator. Journal of Animal Ecology 54:179-192. mate demographic characteristics for small groups Verner, J., K. S. McKelvey, 8. R. Noon, R. J. of individuals and look for patterns relating Gutierrez, G. I. Gould, Jr., and T. W. Beck, demography to characteristics of the environment. technical coordinators. 1992. The California These investigations could be approached in an ob- Spotted Owl: a technical assessment of its servational or experimental framework. In either current status. United States Department of case, a system of nest boxes as described in Hay- Agriculture Forest Service Pacific Southwest ward et al. (1992) will facilitate obtaining some of Research Station, Albany, California, USA. the required data. Without nest boxes it will be ex- General Technical Report PSW-GTR-133. tremely difficult to obtain a sample of owls suffi- cient to answer most questions. Of course, nest boxes cannot be used to answer some questions concern- ing nesting habitat and must be employed judi- ciously to avoid biasing results. Studies in Europe by Korpimaki (1981), Sonerud (1985), Lofgren et al. (1986) and others have demonstrated the power of nest boxes in studying this species.

REFERENCES

Dennis, B., P. L. Munholland, and J. M. Scott. 1991. Estimation of growth and extinction param- eters for endangered species. Ecological Mono- graphs 61:115-143. Hairston, N. G. 1989. Ecological experiments: purpose, design, and execution. Cambridge University Press, Cambridge, New York, USA. Hayward, G. D., P. H. Hayward, and E. 0. Garton. 1993. Ecology of Boreal Owls in the northern Rocky Mountains, USA. Wildlife Monographs 124:l-59. Hayward, G. D., R. K. Steinhorst, and P. H. Hay- ward. 1992. Monitoring Boreal Owl popula- tions with nest boxes: sample size and cost. Journal of Wildlife Management 56:776-784.