Hi all,
I recently updated my ROMS repositories from revision 761 (Sat, 28 Mar 2015) to revision 838 (Sat, 04 Mar 2017) then I got inconsistent results between these two revisions which has confused me for a while.Hopes I can get some tips here.
Attached are 2 monthly averaged plots of two revisions about bottom temperature&velocity at June 2013 .The current patten is almost the same while newer revision shows a faster current and warmer bottom water，especially at the coastal zone.
Hernan replaced the cpp option "SPLINES" with 3 more flexible cpp options "SPLINES_VVISC","SPLINES_DIFF" and "RI_SPLINES", writing "Users should get identical results if activating the new options described above." (Search https://www.myroms.org/projects/src/ticket/681 for more details.) Following this I changed nothing but my header file associated with this revision ticket,that is to say I was using the same grid file,input file,forcing file,boundary file,etc except for the header file.And then I got the different results as mentioned above:(. The newer revision of conservative, parabolic splines reconstruction of vertical derivatives seems to build a larger vertical diffusion&viscosity and lead to a stronger vertical mixing,which conflicts to this revision ticket.
I'm still not clear what brings this different but I'll keep trying. Is there any suggestions?
Thanks in advance,
Haiyun
Inconsistｅｎｔ results after updating for the option SPLINES
Re: Inconsistｅｎｔ results after updating for the option SPLIN
There have been two years of changes. You can try the versions just before and after that SPLINES change to confirm that that is not what changed your answers. If you want to track down what did change, you can do the equivalent of "git bisect" to see when the change(s) occurred.
There are at least two schools of thought on this:
1. Get the changes as they come so you never get that two year gap.
2. Wait to update until you absolutely have to. One friend is taking this path and he's so far behind that if he has to update, he'll have to charge some project for the two months it will take him to get running again. He's hoping to retire before it comes to that.
There are at least two schools of thought on this:
1. Get the changes as they come so you never get that two year gap.
2. Wait to update until you absolutely have to. One friend is taking this path and he's so far behind that if he has to update, he'll have to charge some project for the two months it will take him to get running again. He's hoping to retire before it comes to that.
Re: Inconsistｅｎｔ results after updating for the option SPLIN
Hi Kate,
Well, the figures above are exactly from the versions before and after SPLINES changes.
I carefully read and compare ROMS codes of these two versions just to find all "SPLINES" cpp options are splited and replaced by "SPLINES_*" in the new version, while the algorithm remains the same. I've attached the head files, please take a look and correct me if there is any mistake.
These two opinions are reasonable. I'll accept the result of new versions till I get an idea about this question.
Thanks,
Haiyun
Well, the figures above are exactly from the versions before and after SPLINES changes.
I carefully read and compare ROMS codes of these two versions just to find all "SPLINES" cpp options are splited and replaced by "SPLINES_*" in the new version, while the algorithm remains the same. I've attached the head files, please take a look and correct me if there is any mistake.
These two opinions are reasonable. I'll accept the result of new versions till I get an idea about this question.
Thanks,
Haiyun
