GOES navigation for Crystal-Face experiment sector

The GOES sector set up for the Crystal-Face experiment did not have navigation files. Gerry Heymsfield requested that navigation file be generated so he could actually use the GOES data in a quantitative way. I generated the navigation files in the usual tedious manner, making both a visible resolution navigation file and an infrared resolution file. The two gzipped files are available here:

Some questions arose as to how accurate these navigation files actually were, so I used my IDL map generating routines to make a map for the whole infrared setor and for a ROI of the visible sector centered on Florida (at 7392x2832 the visible sector is too large to generate a map in a reasonable amount of time). The code used to generate the visible ROI map outline follows:

PRO make_crystal_vis_map indir = '/Volumes/firewire_2/crystal/' navfile = indir + 'crystal_vis.nav.flt.z2.x07392.y02832' imagefile = indir + '0207311940dG08G8I01.tiff' sx = 3000 sy = 500 nx = 1000 ny = 1000 ex = sx+nx-1 ey = sy+ny-1 data = fltarr(2,7392,2832) openr, 1, navfile readu, 1, data close, 1 free_lun, 1 deg_lon = reform(data[1,sx:ex,sy:ey]) deg_lat = reform(data[0,sx:ex,sy:ey]) data = 0 j_plot_map, lat_array = deg_lat ,$ lon_array = deg_lon ,$ map = vmap ,$ tile_size = 1800 ,$ /usa,/hires visimage = read_tiff(imagefile) visimage = visimage[sx:ex,sy:ey] outdata=bytarr(3,1000,1000) outdata[2,*,*]=bytscl(visimage,min=50,max=450) outdata[1,*,*]=outdata[2,*,*] outdata[0,*,*]=outdata[2,*,*]>vmap write_tiff,'crystal_vis_test.tiff',outdata END

As can be seen, it is straightforward IDL coding - the same region is extracted from both the navigation file and the image data file. The only custom routine called, j_plot_map, generates map outlines based on navigation array. I converted the 10bit visible data to 8bit and embedded the map outline in the red channel to produce a 3 MB tiff file. Here is a zoomed up section of that image:

I estimate that the map outlines are off by about 10km in this region. GOES images shift around by several kilometers during the course of a day, so I would say the match between the navigation and image in this case is moderately good compared to results we usually see.

For reference, here are the 10bit image files that the navigations were based on and a photoshop file containing the map outline generated from the infrared navigation in one layer and the infrared data in another layer.