Cooperative Fish and Wildlife Research Units Program: Minnesota
Education, Research and Technical Assistance for Managing Our Natural Resources

Brownie tagging models are commonly used for estimating survival rates from multiyear tagging studies. The basic model, model 1, assumes that all tags have the same recovery rate. An alternative, model 0, allows newly tagged animals to have a different tag recovery rate than previously tagged animals. This feature might be necessary because new tags are less fouled and more visible than previously applied tags and thus have a higher reporting rate. Model 0 accommodates the different recovery rates through the use of additional parameters, which leads to larger standard errors than in model 1. Model 0 , a new model, also allows newly tagged animals to have a different tag recovery rate than previously tagged ones. It makes use of a known fouling time (the time it takes for tags on newly tagged animals to have the same visibility as tags on previously tagged animals) to divide the year into two parts. During the first part of the year the tags on newly tagged animals are more visible than those on previously tagged ones, while in the second part all tags have the same visibility. Dividing the year into parts and recording the recaptures in each part avoids the failure of the assumption that the reporting rate is constant for all tagged animals, achieves greater precision, and provides estimates of the survival rate at the end of the second year instead of at the end of the third year (as in model 0). The superiority of model 0 over models 0 and 1 is demonstrated for several cases using Monte Carlo simulation. Simulations were based on the queen conch Strombus gigas fishery of the Turks and Caicos Islands, British West Indies. For that fishery, if the tag reporting rate is altered by 25% or more by fouling, it is beneficial to use model 0 instead of model 1.