Abstract
The importance of small scale variations in currents on the predictability of spatial distributions of fish eggs is investigated using dense observations of surface currents. The currents were measured with HF radar and used to drive an advection–diffusion model of ichthyoplankton concentration. We first demonstrate that the model produces acceptable agreement with observed egg fields. We then use the predicted egg fields as a basis for comparison with model runs made with currents subsampled in space and filtered in time. Significant error was found for spatial sampling intervals as small as 3 km, even though most of the variance in the currents occurred at much longer length scales. This was primarily due to the loss of the rich small scale variability in horizontal divergence. Nevertheless, most of the error is due to the misfit in the egg fields at larger length scales; that is, small scale forcing is necessary in this system for the larger scale features to be reproduced. This study thus suggests that considerable caution should be exercised before assuming that circulation models, even very sophisticated and detailed ones, capture enough of the variability of marine systems to make accurate forecasts of plankton distributions.