... information.¹

Parameters to include: image location, various flux and shape measures, and error estimates for them. For a starting point we will base this on the INT WFC pipeline catalogue products, with selective addition from the SDSS and 2MASS pipelines.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... solution.²

The annual growth rate of the science storage requirements is estimated as follows: a typical WFCAM science frame will be 4 devices of $\times$ $2048\times2048$ pixels, $\times2\times2$ interlaced, $\times4$ bytes per pixel $\approx250$ Mbyte. Assuming 75% UKIRT time is devoted to WFCAM with 70% fill factor (weather, faults, other science programmes) we will have to deal annually with $\approx$ 200 nights of science data. Assuming $\sim200$ frames per night (an average of one every 3 mins over a 10hr night) then the pixel archive grows by $\approx$10 Tbyte/yr. All other data (derived from pixels or elsewhere) will be $<<10$% the size of the pixel archive.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... persistence³

Including how to deal with bright stars, ghostimages, trails - one possiblity it to use input predictions for positions of bright stars, galaxies, globular clusters, and solar system objects.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.