40655251_-_Statement_of_Work_1.pdf

Pacific Walrus Close-kin Mark-Recapture Preliminary Model Update

Introduction and Background

The U.S. Fish and Wildlife Service (USFWS) conducted a genetic mark-recapture project for five years (2013–2017) to estimate Pacific walrus (Odobenus rosmarus divergens, hereafter walrus) abundance and survival. Concomitantly, the U.S. Geological Survey (USGS) conducted age structure surveys from 2013–2015 to estimate walrus demographic rates and population trend, and USFWS continued these age structure surveys from 2016–2017.

The USFWS genetic mark-recapture project addressed relevant questions about walrus kin structure (Beatty et al. 2020) as well as estimated total abundance (Beatty et al. 2022). However, the financial (~$8 million), human capital, opportunity, and wildlife disturbance costs of conducting long research cruises for five years requires continued investment in alternative approaches to estimate population abundance and demographic rates. Thus, an alternative approach to estimate walrus abundance and vital rates at more frequent intervals is required.

Close-kin mark-recapture (CKMR) is an emerging approach to estimate abundance and demographic parameters that relies on the frequency of kinship relationships among individuals sampled from a population (Bravington et al. 2016). CKMR has been implemented with samples exclusively collected from dead animals, although it can also be implemented with samples collected from a traditional live mark-recapture design, or a combination of live and dead sampling. Thus, samples collected from the walrus harvest may be able to augment samples collected on research cruises from live animals, reducing the overall financial expenditure needed to generate estimates of abundance and demographic rates. We are seeking support to update a preliminary walrus CKMR model to include samples from harvested animals. The updated model can be used to inform sampling design for future research cruises that collect data on walrus demography to reduce their length and frequency. For example, including several hundred samples from the harvest each year may allow researchers to reduce the length of research cruises or reduce the number of years that research cruises are conducted. However, the effects of harvest samples on precision of abundance and demographic rates are unknown and we seek to quantify those effects with an updated preliminary walrus close-kin mark-recapture model.

Objectives and Products

The objective of the project is to develop an updated preliminary close-kin mark recapture model to include samples from harvested animals. Specifically, we are seeking information on the effects of sample size requirements, number of years to collect samples, and age classes to target for sampling. In addition, we are seeking information on the impact of seasonal differences in sample collection from harvested animals on model parameters and approaches to account for any possible biases. Finally, we are seeking information on the effects of spatially clumped sampling for harvested animals. We anticipate one product from this work:

1. A technical report submitted to the project manager at U.S. Geological Survey. The technical report shall contain:

a. An updated preliminary walrus close-kin mark-recapture model along with detailed justification of the model structure.

b. A detailed accounting of model assumptions.

c. All relevant statistical derivations for the updated preliminary close-kin mark-recapture model.

d. Computer code with sufficient annotations that someone without knowledge of the language can implement the code in other software.

e. Guidance on sampling design and sample sizes for future walrus field seasons.

Methods

The focus of the study is the Pacific walrus population, which inhabits the northern Bering Sea and Chukchi Sea. We propose to update a preliminary CKMR model to include samples from harvested animals. The initial preliminary CKMR model did not include samples from harvested animals due to technical limitations. The updated model will include both samples from live animals and harvested animals, leveraging information from all available data sources. In addition, the updated model will also consider different scenarios to evaluate the consequences of the harvest on precision of estimated model parameters. The number and demographics of harvested samples may vary substantially from year-to-year with nearly all adult males harvested in some years to nearly all adult females with calves in other years. We seek to evaluate the effects of this variation on precision of model parameters.

Timeline

We are seeking delivery of technical report to be submitted to the project manager by December 31, 2024.