Abstract

Topic: Moose management and monitoring


Using Statistical Population Reconstruction to Fill Temporal Gaps in Aerial Survey Estimates of Moose Abundance and Demographics

Sergey S Berg1, Christopher Stocken1, Morgan Swingen2, Ron Moen3, Steve Windels4, Michelle Carstensen5, Seth Moore6, Anna Weesies6, Tiffany Wolf7, William J. Severud8

  1. Department of Computer and Data Sciences, University of St. Thomas, Saint Paul, Minnesota, USA
  2. Treaty Authority, Duluth, Minnesota, USA
  3. Natural Resources Research Institute, University of Minnesota Duluth, Duluth, Minnesota, USA
  4. National Park Service, Voyageurs National Park, International Falls, Minnesota, USA
  5. Minnesota Department of Natural Resources, Wildlife Health Program, Forest Lake, Minnesota, USA
  6. Grand Portage Band of Lake Superior Chippewa, Department of Natural Resources, Grand Portage, Minnesota, USA
  7. Department of Veterinary Population Medicine, University of Minnesota, Saint Paul, Minnesota, USA
  8. Department of Natural Resource Management, South Dakota State University, Brookings, South Dakota USA

Background: Estimating moose abundance over the large spatial scales at which many management decisions are made often relies on aerial surveys conducted during the late autumn and winter months when plentiful snow allows for visual detection of animals from above. These aerial surveys, however, present numerous logistical challenges that include insufficient snow cover, high costs, adverse weather conditions, personnel availability, and other external factors. Statistical population reconstruction (SPR) may provide a reliable and cost-effective means of estimating moose abundance and other demographic parameters during years when aerial surveys are not conducted by integrating survey data from other years with survival information from radiotelemetry studies. Objectives: Our objective was to quantify how the amount of available aerial survey and radio telemetry data, as measured by the number, age, sex of monitored animals each year, influences the accuracy and precision with which SPR can be used to fill temporal gaps in demographic estimates typically derived from aerial surveys. Methods: We used a Monte Carlo simulation to generate aerial survey and radiotelemetry data for moose populations experiencing a wide range of demographic conditions and survey intensities. We then used a combination of Pearson correlation and relative absolute deviation (RAD) to quantify the accuracy and precision of model estimates of overall abundance and bull- and calf-to-cow ratios. Results: Given sufficient radio-telemetry data, model estimates of overall abundance and demographic ratios were highly correlated with simulated values and had relatively low RAD when the number of years of missing aerial data were few. As the number of years without aerial surveys increased, the accuracy and precision of these model-derived estimates decreased steadily. As expected, model performance was also strongly influenced by the amount and precision of available radiotelemetry and aerial survey data. Conclusions: We encourage wildlife managers to explore using SPR models to fill any temporal gaps that may arise in aerial survey estimates of moose demographics due to insufficient snow cover, high costs, or other factors. We also recommend that agencies perform cost-benefit analyses to assess if the reduced cost associated with foregoing aerial surveys is worth the corresponding loss in accuracy and precision.