The effect of large‐scale spatial variation of pelagic fish on hydroacoustic estimates of their population density in Windermere (northwest England)

• Published in: Ecology of Freshwater Fish
• Main Authors: Baroudy, E., Elliott, J. M.
• Format: Article in Journal/Newspaper
• Language: English
• Published: Wiley 1993
• Subjects: Ecology; Aquatic Science; Ecology, Evolution, Behavior and Systematics; Arctic; Arctic charr; Salvelinus alpinus
• Online Access: http://dx.doi.org/10.1111/j.1600-0633.1993.tb00098.x; https://api.wiley.com/onlinelibrary/tdm/v1/articles/10.1111%2Fj.1600-0633.1993.tb00098.x; https://onlinelibrary.wiley.com/doi/pdf/10.1111/j.1600-0633.1993.tb00098.x

Abstract From July 1989 to December 1992, an echo sounder provided monthly estimates, usually for both day and night, of pelagic fish densities in the north and south basins of Windermere, the largest natural lake in England. Sampling was along contiguous transects, 3 in the north basin and 5 in the south basin. It was impossible to separate records for Arctic charr ( Salvelinus alpinus ) from those for brown trout ( Salmo trutta ), but previous sampling by gill‐nets and anglers showed that charr formed about 90% of this mixed population in the north basin and 60–75% in the south basin. In each basin, estimates of population density for the combined transects did not agree with estimates obtained by treating the transects as a contiguous sample of 9 sampling units; only the latter, of course, provided a measure of the precision of each estimate. Analyses of the contiguous samples showed that the variance was significantly greater than the arithmetic mean in most samples, indicating that the fish were distributed nonrandomly in a patchy or clumped pattern. The relationship between the large‐scale spatial variance and mean for these samples was well described by a power function; the parameter estimates did not vary significantly between basins, day and night samples or years. As the power in this equation did not differ greatly from 2 (value for pooled data was 1.70 ± 0.11, n = 136), the variance was stabilized by a log transformation of the data, and the geometric mean, rather than the arithmetic mean, provided the best estimate of population density when some measure of precision was required. These conclusions may be applicable to other echo‐sounding estimates of population abundance, and similar comparisons should be made for pelagic fish in other lakes.