Potential bug in the assigment of precipitation subcolumns

```
This will occur if use_precipitation_fluxes=.false. and NCOLS>1, which is not a 
configuration that has been thoroughly tested. I seems that the fix is easy, 
just use column_prec_out instead of column_frac_out.
```

Original issue reported on code.google.com by `alejandr...@metoffice.gov.uk` on 16 Aug 2013 at 1:17
