A Functional Relationship Between Watershed Size and Atlantic Salmon Parr Density

Abstract Atlantic salmon Salmon salar are at critically low levels in Maine rivers and maintaining current populations depends heavily on the stocking of hatchery-produced fry. Fry survival varies greatly not only among rivers but also within rivers. Better understanding of this spatial variability is needed to improve population recovery efforts. Quantile regression was used to examine how parr density varies along stream gradients as a function of the upstream cumulative drainage. Regression quantiles (τ  =  0.15–0.90) had significant negative slopes indicating that, in general, as cumulative drainage area increased, parr density decreased. The τ  =  0.90 quantile model, which may be viewed as an upper limit to parr density for a given cumulative drainage area, was density  =  [10⁁(1.0679 − 0.0013 · km2)] − 1. Quantile regression models were then combined with river-specific habitat data in a Geographic Information System to predict total Atlantic salmon parr production in several rivers. This analysis will help managers determine the potential of a stream reach or entire river to support Atlantic salmon parr and to prioritize and increase the effectiveness of management activities such as fry stocking and habitat restoration.