Lower number-density termination threshold #189
Conversation
…es to allow cohorts to demote successfully
|
It does make sense that we would only start to see answer changing results on the LM6 run. |
|
why—is that the long one? |
|
yeah, I figure the six month test is a long enough simulation for the termination mortality to really take effect. The other tests are only a few days at most, we simply don't have enough growth/mortality cycles to effect plan number densities by then. |
|
retested on lawrencium after merging; same results as before. @rgknox, now that this PR contains only the answer-changing code, could you take another look to make sure that what is there is what is intended to be there, run the rapid science check, and then assign to @bandre-ucar to test on ucar machines and pull? |
|
I performed a review of the code to see if it makes sense to me, and it makes sense to me. The changes seem to address what was intended. I also ran a Rapid Science Check (RSC) and the results are as expected. Since @ckoven changes are expected to be science changing, I was anticipating small but not-insignificant divergence from master, and that was what was generated. See below for a 1x1 brazil:
Make note of the changes in maintenance respiration. Note that while GPP stayed roughly the same, maintenance respiration responded, presumably somehow to the decreased frequency of "demotion termination". |
|
That's good to know. Those 'RSC' figures are pretty informative...
…On 13 March 2017 at 13:29, Ryan Knox ***@***.***> wrote:
I performed a review of the code to see if it makes sense to me, and it
makes sense to me. The changes seem to address what was intended.
I also ran a Rapid Science Check (RSC) and the results are as expected.
Since @ckoven <https://github.com/ckoven> changes are expected to be
science changing, I was anticipating small but not-insignificant divergence
from master, and that was what was generated. See below for a 1x1 brazil:
[image: bdead_threshold]
<https://cloud.githubusercontent.com/assets/5891980/23871507/f80423a8-07e7-11e7-9e4e-20bb02ba2d2e.png>
[image: ncohort_threshold]
<https://cloud.githubusercontent.com/assets/5891980/23871509/fad3c21e-07e7-11e7-8085-f74b28cd2518.png>
[image: cfluxes_threshold]
<https://cloud.githubusercontent.com/assets/5891980/23871511/fc88a930-07e7-11e7-8979-d3ee1d25dcd5.png>
[image: fluxes_diurnal_threshold]
<https://cloud.githubusercontent.com/assets/5891980/23871512/fe4b88dc-07e7-11e7-8a65-9fe3f4dcec32.png>
Make note of the changes in maintenance respiration. Note that while GPP
stayed roughly the same, maintenance respiration responded, presumably
somehow to the decreased frequency of "demotion termination".
—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
<#189 (comment)>, or mute
the thread
<https://github.com/notifications/unsubscribe-auth/AMWsQ9DoxQydtZ0r7TOaY-YOwPNYRr94ks5rlZkigaJpZM4MLxPO>
.
--
------------------------------------------------------------------------
Dr Rosie A. Fisher
Staff Scientist
Terrestrial Sciences Section
Climate and Global Dynamics
National Center for Atmospheric Research
1850 Table Mesa Drive
Boulder, Colorado, 80305
USA.
+1 303-497-1706 <(303)%20497-1706>
http://www.cgd.ucar.edu/staff/rfisher/
|
|
since these passed lawrencium already, passing this on to @bandre-ucar for testing on yellowstone and friends |
|
running tests |
Fixes to termination logic to allow demotion to occur successfully
reduces termination threshold due to small number density and also deletes calls to terminate_cohorts in CanopyStructureMod except for those that are after a call to fuse_cohorts.
this is on top of the code in PR #184, so will shrink to just ten or so lines of code once that is merged.
Fixes: #187
User interface changes?: No
Code review: changes done in consultation with @rgknox and @rosiealice
Test suite: ed tests on lawrencium
Test baseline: a5dc8da
Test namelist changes:
Test answer changes: climate changing
Test summary: [Remove 'PASS' fields]
[cdkoven@n0002 ed.lawrencium-lr3.intel.37ba5ce]$ ./cimeteststatus | grep FAIL
FAIL SMS_D_Lm6.f45_f45.ICLM45ED.lawrencium-lr3_intel.clm-edTest.clm2.h0.nc : baseline compare clm2.h0 (baseline: compare .base file with ed.lawrencium-lr3.intel.a5dc8da file)
FAIL SMS_D_Lm6.f45_f45.ICLM45ED.lawrencium-lr3_intel.clm-edTest.cpl.hi.nc : baseline compare cpl.hi (baseline: compare .base file with ed.lawrencium-lr3.intel.a5dc8da file)
FAIL SMS_D_Lm6.f45_f45.ICLM45ED.lawrencium-lr3_intel.clm-edTest.rtm.h0.nc : baseline compare rtm.h0 (baseline: compare .base file with ed.lawrencium-lr3.intel.a5dc8da file)
FAIL SMS_D_Lm6.f45_f45.ICLM45ED.lawrencium-lr3_intel.clm-edTest : baseline compare summary (baseline: compare .base file with ed.lawrencium-lr3.intel.a5dc8da file)