Added final merging of created files and further documentation to get_era5_forecasts.sh.

Jupyter_Notebooks/get_era5_forecasts.sh

 #!/usr/bin/env bash
 
-# This bash-script reads the 2m temperature from ERA5 short-range forecasts files, crops the data to the largest target domain (see Dataset ID 1-3) and then performs a merging step with the ground truth and the persistence forecast data provided by from the postprocessing of a trained model (see prostprocess-step of workflow.)
+# ********************************************* doc-string *********************************************
+# This bash-script reads the 2m temperature from ERA5 short-range forecasts files,
+# crops the data to the largest target domain (see Dataset ID 1-3) and then performs a merging step
+# with the ground truth and the persistence forecast data provided by from the postprocessing
+# of a trained model (see prostprocess-step of workflow).
+# More details on the ERA 5 short-range forecasts is documented under the following link:
+# https://confluence.ecmwf.int/pages/viewpage.action?pageId=85402030
+# The forecast-files here must contain the 2m temperature (2t) and should be organized as follows:
+# --- indir
+# |   |--- [<YYYY>]
+# |   |   |--- [>MM>]
+# |   |   |   |--- fc_06
+# |   |   |   |   | [<YYYYMMDDHH>_{06..12}_sf_fc.grb
+# |   |   |   |--- fc_18
+# |   |   |   |   | [<YYYYMMDDHH>_{06..12}_sf_fc.grb
+#
+# Thus, data must be available in grib-files for each forecast hour (between lead time 06 and 12 hours) in
+# yearly and monthly subdirectories. Furthermore, separate subdirectories exist for the short-range forecasts
+# initialized at 06 and 18 UTC.
+# ********************************************* doc-string *********************************************
 
 # Parse script argument
-yr=$1
+yr=$1                                                  # year for which data is available (2019 in the manuscript)
 
 # some directory parameters
 indir=<path_to_era5_short_range_forecasts>/${yr}       # directory where the ERA5-short range forecasts are located
@@ -18,7 +37,7 @@ fi
 hh_s=6                        # start of lead time range in hours
 hh_e=12                       # end of lead time range in hours
 
-declare -a hh_list=(18)       # initialization hour of ERA5 forecasts to evaluate (either 6 or 18)
+declare -a hh_list=(6 18)       # initialization hour of ERA5 forecasts to evaluate (either 6 or 18)
 
 for hh in ${hh_list[@]}; do
     hh0=$(printf "%02d" ${hh})