Nghiên cứu Sinh thái học Bầy Đàn Hải Ly tại Vườn Quốc gia Isle Royale: Ước tính Số lượng và Ảnh hưởng Săn mồi

ESTIMATES OF ABUNDANCE AND PREDATION - THE POPULATION ECOLOGY OF BEAVER IN ISLE ROYALE NATIONAL PARK by MARK C. ROMANSKI A THESIS Submitted in partial fulfillment of the requirements for the degree of MASTER OF SCIENCE IN APPLIED ECOLOGY MICHIGAN TECHNOLOGICAL UNIVERSITY 2010 © Mark C. Romanski 2010 This thesis, “Estimates of abundance and predation - the population ecology of beaver in Isle Royale National Park,” is hereby approved in partial fulfillment of the requirements for the degree of MASTER OF SCIENCE IN APPLIED ECOLOGY. School of Forest Resources & Environmental Science Signatures: Thesis Advisor _________________________________________ Dr.

Vucetich Dean of School _________________________________________ Dr. Gale Date __________________________________ ACKNOWLEDGEMENTS I am the beneficiary of people too numerous to mention who have labored on Isle Royale to study and understand this most unique and endearing ecosystem. First and foremost are those with the vision that began only to look at, learn from, and accept what the “island” offered to teach them in its own good time. Specifically, thank you to Dr.

Duward Allen, who initiated the study of wolves and moose and who impassioned Dr. Phillip Shelton to pursue his interests in beaver ecology on the island. Shelton’s efforts and long-term commitment to his work preserved a data set that is unique and unparalleled. Rolf Peterson’s uncanny ability to always capture the right data, I am indebted.

But then again, when one pursues with tenacity and detail, good things abound. I am fortunate to have had the opportunity to work with Rolf and learn from his example. Douglas Smith, who ensured despite his own enormously busy schedule with something called the Yellowstone Wolf Project that the beaver survey continued in the years following Dr. Thank you for your commitment as well.

I would especially like to thank my advisor, Dr. His ability to comprehend and exquisitely express the natural world through research is something I will seek to emulate as I move forward from this point. To John and my committee members, Dr. Casey Huckins and Dr.

Joseph Bump, you have my gratitude for the time you have committed and the enormous amount of patience you bestowed. I also owe much appreciation to Dr. Leah Vucetich who grappled with the administrative aspects of my education. Additionally, her intimate knowledge of Isle Royale helped my efforts.

iii To the employees of Isle Royale National Park, my co-workers and friends, I am thankful for each and every one of you. Your efforts, not as National Park employees, but as dedicated men and women who strive to protect and preserve a piece of heaven, inspires me daily. Thank you for your direct efforts to this project and your support these 15 years I have worked alongside you. A few names must be singled out: Jack Oelfke, Ann Mayo-Keily, Alex Egan, Katy Goodwin, Leah Ettema, Jean Battle, Steve Windels, James Hummel, Jon-Hudson Spencer, Paul Brown, Betsy Rossini, Phyllis Green, Erin Grivicich, Kerry Martin, Buzz Brown, Steve Martin and Sue Ruddy.

Thank you all! The longevity of this project would not have been possible without the financial commitments of the Isle Royale Keweenaw Parks Association. Specifically, Jill Burkland, who always gave when asked. Thank you Jill! Additionally, financial contributions came from the National Park Service Great Lakes Inventory and Monitoring Network and were inspired by the foresight of Bill Route, its Director. To my family, Azure, Birch, Acorn and Zephyr, thank for understanding, for your sacrifices, and most importantly your love.

Without it, none of this would have been possible. I missed you plenty and I plan to get that time back. To my parents, Karen and Victor Romanski, your continued support and unconditional love has made me a better person in all that I do and continues to give me strength. To all our friends who have supported me and my family, I will await the opportunity to return your good will and love.

iv TABLE OF CONTENTS ACKNOWLEDGEMENTS .x CHAPTER 1: Double-count surveys and unexpectedly misleading estimates of sightability for beaver colonies .3 FIELD METHODS & DATA PREPARATION .17 CHAPTER 2: The fourth dimension’s influence on top-down, bottom-up and abiotic regulation of beaver population dynamics .26 FIELD METHODS & DATA PREPARATION .32 Reconciliation of predation rate and growth rate .64 APPENDIX A – Historical narrative of beaver research on Isle Royale National Park, Michigan, USA.83 vi LIST OF TABLES Table 1-1. Statistics associated with the double count method for aerial surveys of active beaver colonies between 1990 and 2008. Population trends of wolves, moose, and beaver, and measurements of beaver predation by wolves as calculated using a linear equation to interpret scat occurrences in wolf feces between 1990 and 2009. Results of regression analyses to assess the influences of abiotic processes and top-down control on beaver population dynamics between 1962 – 2009.52 vii LIST OF FIGURES Figure 1-1.

Estimated probabilities of detecting an active beaver colony under various conditions. Estimated abundances of active beaver colonies on Isle Royale, 1962-2008. Map of Isle Royale, Lake Superior, North America, showing the distribution of the distinct forested communities and a typical distribution of active beaver colonies. The abundance of active beaver colonies, wolves and calculated predation rates for the period between 1962-2009.

Inter-annual variation in water availability for use and impoundment by beavers as represented by mean annual precipitation from 1962 to 2009 and the mean annual stream flow from pooled monthly averages between 1965 and 2003. Temporal correlation, using Pearson’s product-moment correlation coefficient, of beaver colony abundance under three, distinct, forest succession trajectories: eastern boreal forest region, central 1936 burn area, and the western northern hardwoods forest region. Simple linear regression analyses relating beaver colony growth rate to water availability represented as mean annual precipitation (cm) and mean annual stream flow (cubic meters per second). Evaluations of predation as a top-down control of beaver population dynamics using regression analyses.

Regression analyses to assess the relationship between the proportion of wolf diet that was beaver, determined by macroscopic inspection of wolf feces, and beaver colony and wolf abundance, and beaver colony growth rate.59 viii PREFACE The population dynamics of the beaver (Castor sp.) are fairly well studied and beaver are predominantly thought to have a strong top-down influence on the ecosystems they inhabit (see review in Rosell et al. Despite widespread acceptance of beaver as a keystone species and ecosystem engineer, the impacts of predation on the population dynamics of beaver are almost completely ignored (for exception see Potvin et al. Myself and the co-authors identified here have taken advantage of long-term data available for both beaver colony and wolf abundance on Isle Royale National Park to initiate assessments on the top-down influences of predation, bottom-up impacts of forest succession and the importance of abiotic processes as factors contributing to beaver population dynamics. Beaver colonies are typically surveyed in the fall of the year, when their food cache is visible.

Normally aircraft are used to accomplish this work. As part of ensuring an accurate count, investigators often employ a “double-count” survey. This survey involves two independent observers, each recording observations. These observations are then compared to one another and an estimate for the entire population is obtained, including those members of the population of interest not seen by either observer.

We have completed 5 double-count surveys on Isle Royale between 1990 and 2008 to increase the accuracy of our abundance estimate for beaver colonies. In two of those surveys, 1990 and 2002, a large fixed wing aircraft housing pilot and both observers was used. Specifically, this aircraft was a DeHavilland Beaver. In the remaining counts, ix 2006-2008, we used two smaller fixed-wing aircraft (Aeronca Champ or Aviat Husky) each with its own pilot and observers.

Unexpectedly, sightability, probability of detecting a beaver colony, was lower from the smaller aircraft, whose flight speed was slower, altitude lower and flight pattern was circling as opposed to the high altitude and straight flight line of the larger aircraft. Surprisingly, overestimation of sightability in the larger aircraft only became apparent after use of the smaller aircraft. Our results are important for others who employ the double-count method, in so much as, application of the double-count method can give scientists false confidence in meeting the assumptions associated with the double-count method. Predation plays a crucial role in the biodiversity and function of ecosystems (Duffy 2002, Duffy et al.

Generally, this importance is quantified in terms of mean abundance of consumers and resources available to these consumers. Less frequently investigated are the varying time scales at which consumer (predator) and resource (prey) abundances interact with each other. We manipulate the spatial-temporal aspects of the wolf, moose and beaver colony data collected on Isle Royale between 1962 and 2009, as well as, aspects of forest succession in each of three distinct forested communities to articulate the meaningful time scales at which top-down, bottom-up and abiotic processes influence beaver population dynamics. This approach is interesting for three reasons; (i) it incorporates the role of predation in beaver population dynamics which has largely been ignored previously, (ii) it allows us to quantify the influence of predation on the population of not only beaver, but also in a novel way, assess the impacts of beaver as a bottom-up influence on wolf population dynamics, and (iii) x contemplate restoration of beaver populations, both in North America and Europe, where in much of the beaver’s former range restoration occurred in the absence of predation.

Without elucidating what role predation plays in beaver population dynamics we may be operating under misconceptions regarding exactly what ecosystem processes we have restored. The story of beaver and beaver research in Isle Royale National Park began in 1790 and does not end here, but rather will continue because dedicated people care and this story is worth telling. I have attempted to provide a brief synopsis of this story to date in Appendix A. This snapshot is intended to give context to our efforts here and hopefully give the reader perspective regarding beaver ecology and its heretofore largely overlooked role in a unique ecosystem whose wolf-moose / predator-prey interactions have drawn attention away from this “ecosystem engineer”.

The work in Chapters 1 & 2 of my thesis, although I hesitate to say “my”, is largely our efforts to compile and extract a small aspect of the work done by Phillip Shelton, Rolf Peterson, and Doug Smith. These individuals kept the interest in beaver ecology on Isle Royale alive despite already demanding workloads. With the help of John Vucetich, who clarified our thoughts and provided important insights, I hope we have initiated a good start to what undoubtedly requires more concentrated effort to address key aspects of beaver ecology that perhaps can be more fully exacted on Isle Royale than anywhere else. xi CHAPTER 1: Double-count surveys and unexpectedly misleading estimates of sightability for beaver colonies.

1 Abstract An important method for estimating the abundance of wildlife is the double-count method, which involves comparing numbers of organisms seen by two independent observers. We used the double-count method to estimate sightability and number of North American beaver (Castor canadensis) colonies in Isle Royale National Park (Lake Superior, North America). In 1990 and 2002, we used the method with two observers flying in a DeHavilland Beaver. In 2006, 2007, and 2008, we used two observers in separate, smaller aircraft that flew transects and circles at lower altitudes.

Unexpectedly, sightability was lower from the smaller aircraft. Specifically, sightability had been 0.90]) for the smaller aircraft and 0.99]) for larger aircraft. However, other evidence suggests sightability had been grossly overestimated for the larger aircraft. Sightability when flying higher and straighter, required in the larger aircraft, may have been ~0.

Overestimation arose from violating an assumption of the double-count method – the assumption that all lodges are equally sightable. Disturbingly, that violation became apparent only after surveys from the smaller aircraft. The results are an object lesson where, the double-count method gave reason to think observers had a high rate of detecting colonies, when in fact it had been much lower.