crosscheck old and new results h1

[Vivian-Felde]

ISSUE:

• There is a difference in the new results of importance of spd and old for temporal analysis
• However, events are the same
• spds have similar patterns for most but the importance is now extremely low compared to scenario runs
• there are differences close to the cut-off ranges in spd from previous run - but this was expected because the first run was done with spd run with age cut off so the end around 0-3 and 7-9 ka were highly uncertain.
• there is missing importance of climate in some time slice

Actions:

• Have checked whether spd is normalised - they are NOT normalised to 1, which is correct
• have checked the summary tables - minor fix in group_by for regions of importance, though this fixed another issue, it did change the results
• checked results by running functions I made in working script with new input data (also gives new output)

TO DO:

• [x] check input table and output of temporal analysis (compare with old input and output) and varhp function - what is different?

[Vivian-Felde]

Europe example: RED = events, BLUISH = spd

[Vivian-Felde]

Hey @OndrejMottl I could not find out what is wrong with the spd function, it works fine when testing it on different subsets. The issue was traced back to the RC subsets and I checked several records, and many of these had missing estimates though there is enough RC dates within the 250 distance.

Now I think I found out what it is, and it is a really minor detail... Basically I did not use the named object with the distance, but I just set the distance parameter to 250 in the function. I change this now using the named object vector, tested the iwalk again, and it works as expected - at least giving calculations for all records that I checked was missing. So I have started a new run with spd calculations (though saving the new output in a new folder spd_temp as I made a test folder, but when it is done, I will delete the other one).