Changes between Version 30 and Version 31 of ticket/1851/General

Timestamp:

2017-02-27T14:25:31+01:00 (7 years ago)

Author:

frrh

Comment:

--

ticket/1851/General

@@ -228 +228 @@
 Unfortunately it very definitely IS. 
 
-I've just proved it by overriding the value written by stuffing it full of a fixed value (9.5 as it happens). When I look in the restart_trc files 
-qsr_mean contains nothing but 9.5 !
+I've just proved it by overriding the value written by stuffing it full of a fixed value (15.0 as it happens). When I look in the restart_trc files 
+qsr_mean contains nothing but 15.0 !
 
 So that nails that for starters - the value appearing in the dump is entirely dependent on the restart frequency (not the frequency of dump writing 
@@ -235 +235 @@
 
 Now we need to see if this actually makes a difference to the rest of the model evolution or if this is just a stand-alone effect.
-
-
-
+So comparing a 2x1-day run with qsr_mean set to 15.0 and a 2x1-day run with it set to 95.0, we compare restart_trc files at day 2. 
+
+The only differences are in the qsr_meam values which are uniformly 15 in one file and 95 in the other. All other fields are identical. So this really is a pointless thing to have in the restarts and is just confusing matters by being no reproducible. Ideally we'd ditch it. 
+
+ 
+ 
+
+
+