Fix Marine VRFY Task in the SOCA ctests (#856)

#### What this PR includes:

- Fix test_gdasapp_soca_JGDAS_GLOBAL_OCEAN_ANALYSIS_VRFY

by updating `marine_gdas_plots.yaml` format and it related
`marine_eva_post.py`

- <del>As of now, ctests passed on `Hera Machine` and `Orion Machine` is
still needed to check</del>
- ctests passed on both `Hera Machine` and `Orion Machine`, see
#829

Fixes #829

Thanks to @CoryMartin-NOAA

ush/eva/marine_eva_post.py

@@ -29,7 +29,7 @@ def marine_eva_post(inputyaml, outputdir, diagdir):
             newfilename = os.path.join(diagdir, os.path.basename(filename))
             newfilenames.append(newfilename)
         dataset['filenames'] = newfilenames
-    for graphic in input_yaml_dict['graphics']:
+    for graphic in input_yaml_dict['graphics']['figure_list']:
         # this assumes that there is only one variable, or that the
         # variables are all the same
         variable = graphic['batch figure']['variables'][0]

ush/eva/marine_gdas_plots.yaml

@@ -58,6 +58,9 @@ transforms:
 
 graphics:
 
+  plotting_backend: Emcpy
+  figure_list:
+
   # ---------- Map Plots ----------
   # Map plot of OmBQC
   # --------