SMB model retraining and new plots

New LOGO and LSYGO SMB reconstruction models retrained with new Tré la Tête SMB data. Some plots  have been updated with the new data and some improvements. New CV and ensemble SMB model files uploaded.

code/smb_validation.py

@@ -816,7 +816,8 @@ def main(compute, reconstruct):
                             year_end_file = end_ref
 
                         # We store the cumulative SMB for all the glaciers with their RGI ID
-                        cumulative_smb_glaciers = np.concatenate((cumulative_smb_glaciers, np.array([glims_glacier['ID'], np.sum(SMB_nn)])))
+#                        cumulative_smb_glaciers = np.concatenate((cumulative_smb_glaciers, np.array([glims_glacier['ID'], np.sum(SMB_nn)])))
+                        cumulative_smb_glaciers.append([glims_glacier['ID'], np.sum(SMB_nn)])
 
                         # We store the simulated SMB 
                         store_file(SMB_nn, path_smb_all_glaciers_smb, "", "SMB", glimsID, start_ref, year_end_file+1)
@@ -832,6 +833,7 @@ def main(compute, reconstruct):
                 # End glacier loop
 
                 # We store the cumulative glacier-wide SMB of all the glaciers in the French Alps
+                cumulative_smb_glaciers = np.asarray(cumulative_smb_glaciers)
                 store_smb_data(path_smb_all_glaciers + 'cumulative_smb_all_glaciers.csv', cumulative_smb_glaciers)
 
             else: