fix for double counting halocarbons #201
Conversation
deleted the total forcing from halocarbons so it is not added to the total forcing.
|
this addresses issue #200. |
|
I pulled out a random year of 2000 and plotted the difference between the old version of Hector and the new fixed Hector. We see a significant change in the total harlocarbon forcing, which has been deleted, resulting in a change in total radiative forcing, but we also see a small change in the forcing from CO2. |
|
The CO2 change is presumably a feedback from previous years? Halocarbons changed, changing temperature, changing the terrestrial and ocean uptakes, changing CO2? |
|
That sounds right to me. The direction of the feedback change looks right, at least with respect to ocean uptake. Higher temperatures should lead to less ocean absorption, which leaves more in the atmosphere. Does terrestrial uptake go in the same direction? |
There was a problem hiding this comment.
This looks fine to me. I suggest you add a note somewhere in the forcing module to the effect that forcings in this list have to be mutually exclusive; there can't be any subtotals. Calculating subtotals for different groups of pollutants seems to be a thing that people do in these models (GCAM has several such, for example), and if we don't put a note in there warning against it, someone is bound to do it again someday.
We should also check SO2. Computing direct, indirect, and total components for it seems to be a common practice, so we should make sure we're not doing that.
|
good idea @rplzzz, I added a comment in the code. I don't see a problem with SO2. We have 2 entries, one for indirect and one for direct with no total SO2. |
no subtotals in forcing calcuation.
|
Just to be complete - atmospheric CO2. |
This applies the bug fix from Hector's 2.0 branch to Pyhector: JGCRI/hector#201 After compiling the wrapped Hector in its subdirectory, test output has been updated by running ./source/hector input/hector_rcp26.ini for all RCP scenarios and copying over the updated output streams to Pyhector's test. The wrapped Hector needs to be patched in `hectorwrapper/hector`: git pull https://github.com/JGCRI/hector pull/201/head
deleted the total forcing from halocarbons so it is not added to the
total forcing.