Atlantic (NEA) mackerel (Scomber scombrus), one of the most widespread and commercially important fish stocks in the North Atlantic. This is
despite the fact that an estimate of recruitment is an important requirement for the provision of advice to fishery managers. The work
here addresses this by compiling catch rates of juvenile mackerel from bottom-trawl surveys conducted between October and March during
1998–2012 and applying a log Gaussian Cox (LGC) process geostatistical model incorporating spatio-temporal correlations. A statistically significant
correlation between the modelled catch rates in adjacent quarters 4 and 1 (Q4 and Q1) demonstrates that bottom-trawl surveys in winter are
an appropriate platform for sampling juvenile mackerel, and that the LCG model is successful in extracting a population abundance signal fromthe
data. In this regard, the model performed appreciably better than a more commonly used raising algorithm based on survey swept-area estimates.
Therefore, the LCG model was expanded to include data from the entire survey time-series, and a recruitment index was developed for use in the
annual ICES stock assessment. We hypothesize that catchability is positively density-dependant and provides supporting evidence from acoustic
observations. Various density-dependant transformations of the modelled catch rateswere furthermore found to improve the correlation between
the derived annual recruitment index and recruitment estimated by backcalculation of adult mackerel data. Square root transformation led to the
strongest correlation, so this is recommended for further analysis of mackerel abundance. Finally,we provide maps of spatial distributions, showing
that the most important nursery areas are around Ireland, north and west of Scotland, in the northern North Sea north of 598Nand, to some extent,
also in the Bay of Biscay.