# Digitalizing precipitation data from the image of weather map

I have a set of very old precipitation anomaly maps, from 1995-2006 for Brazil. The data of these maps, is to the best of my knowlegde, lost.

I would like to have them digitalized, ie, to NetCDF or another similar gridded format. Is there any tools recommendend for this job?

The figure below is a recent plot, similar to my files.

• Are you sure the data is lost? Have you checked the standard data sources? – user967 Oct 4 '16 at 20:40
• Just three points to remember when it comes to interpretation: In arid to semi-arid areas the spottiness of the rainfall distribution doesn't average out on any useful time-scale, and certainly not over an 11 year time-series. Secondly, in mountainous areas like the Andes the local physiographic effects totally eclipse the true regional variation. Thirdly, the map you show looks suspiciously as if it has been generated by a 'contour package', which is often totally misleading because it treats the rainfall in a small rain-guage as an areal maximum or minimum - which it isn't!. – Gordon Stanger Oct 5 '16 at 4:47
• @BarryCarter One can always be surer, but I have put some hours in it and could not find data at this resolution going back to 1995. – kurast Oct 5 '16 at 12:48
• Fair enough. I take it climatedataguide.ucar.edu/climate-data/… and things like ftp.cpc.ncep.noaa.gov/precip/CPC_UNI_PRCP/GAUGE_GLB/V1.0/1985 were unhelpful? – user967 Oct 5 '16 at 14:16

You can convert the image to NetCDF data using gdal_translate. The command line looks something like:
gdal_translate -ot Int16 -of netCDF jpeg_filename nc_filename