Using redd counts to estimate salmonids spawner abundances: A Bayesian modelling approach

Redd (nest) counts are widely used for monitoring salmonid populations. Because of their relative low cost, redd counts allow a good spatial coverage over a long period of time. In this paper we develop a generic Bayesian hierarchical model relating Atlantic salmon, Salmo salar L. spawner abundance and redd counts which accounts for the associated sources of uncertainty. Three rivers of a large catchment are considered and variability across these rivers in the spawners to redds relationship is taken in account. We use short time series, where data on both redds and spawner numbers are available to estimate the parameters of the relationship. We then use the same relationship to estimate spawner abundances for a longer time series with only redd counts available. It is concluded from this exercise that redd counts can be a valuable source of information for monitoring spawning abundances of salmonids and should not be dismissed a priori. However, the uncertainty of spawner abundance estimates derived from redd count data alone may be large and therefore must be assessed. Ancillary data informing the redd counting process can provide great improvements in the precision of the abundance estimates.