wx2d: A PyRAF Routine to Resample Spectral Images

Single rows in STIS rectified spectral images are being analyzed in science programs where the highest obtainable spatial resolution is needed. Here we discuss a problem that limits the accuracy of rectified spectral images that contain spatially unresolved components. Because the point spread function is undersampled along the slit, interpolation between the rows of the distorted spectral image to produce an image linearized in the spatial dimension results in errors in the redistribution of the flux. The STScI pipeline and the STSDAS routines calstis and x2d use bilinear interpolation to rectify spectral images. Here we introduce the STSDAS routine wx2d, which iteratively subdivides pixels in the cross-dispersion direction using “average” interpolation of the flux instead of point interpolation, then combines the subpixels into pixels aligned with the spectral trace to form an image that is linearized in the cross-dispersion direction. We compare the products of x2d and wx2d and show that the latter produces substantially smaller errors in the peak flux row of a spatially compact source.