Enable permafrost by default in Hector

[bpbond]

Following #720 , this PR changes the model's INI files so that permafrost carbon is nonzero—i.e., enabled—by default. The parameterization follows that of Woodard et al. 2021.

It also fixes a number of code issues and bugs, fully described below.

Adding permafrost produces a warmer model overall:

Specifically, for SSP2.45 we see +100 ppm CO2 and +0.4 °C with permafrost enabled. The permafrost pool loses about 250 Pg C, and the thawed pool peaks about ~180 Pg C. These results are broadly consistent with Figure 3 in Woodard et al. (which used a much older version of the model and different parameterization).

[github-actions[bot]]

Differences in Hector outputs

Hello, this is leeyabot!

The current pull request's outputs do not differ from 3.1.1 (0953efdd).

[bpbond]

Several specific bug fixes are broken out in the figure above, including:

• "no soil refreeze" - this was causing the test failures in #721 . When the climate cools, new permafrost forms; the carbon for this is preferentially pulled from the thawed permafrost pool, but if that runs out, from the soil pool (i.e., carbon that was never previously permafrost can freeze). This process (i) produces numerical differences in the climate when run globally versus in multiple biomes, and (ii) has no significant (in a science sense) climate impact. So in 0953efddeb05dbe87ab7a9860344ed473bb2bf6b I disabled it, and will open an issue to come back to this later.
• "decompose labile" - the code was incorrectly computing respiration based on the static, not labile, fraction of thawed permafrost. Fixed in d0fdcd24a7905a60d0a25875698bdf0e3a188b86.
• "correct CH4 routing" - methane produced by thawed permafrost was being routed to the methane component, but also to the atmosphere as CO2. This double-counting was fixed in 8a11c741c2ba14dd8fe5d9c1a9aeb2bd2b69a863.

Here's a better visualization of the science impact of these changes, plotted as the difference from run (B) above:

The labile/static fix is the only one with any appreciable impact, cooling the climate as expected (because fpf_static is 0.74). Fixing the double-counting also cools the climate slightly, again as expected.

[bpbond]

Finally, note that this PR implements permafrost in a single global biome. Following the more realistic protocol of Woodard et al., with a permafrost biome that warms 2.2x as fast as the global average, produces a much warmer climate:

But as @kdorheim and I have decided, it's much simpler to have the default 'out of box' setting be a single global biome. The INI files have commented-out sections that show users how to implement the Woodard configuration.

[codecov-commenter]

Codecov Report

All modified and coverable lines are covered by tests :white_check_mark:

:exclamation: No coverage uploaded for pull request base (dev@d55e3af). Click here to learn what that means.

:exclamation: Current head 76d15d8 differs from pull request most recent head 87c5134. Consider uploading reports for the commit 87c5134 to get more accurate results

Additional details and impacted files

```diff @@ Coverage Diff @@ ## dev #722 +/- ## ====================================== Coverage ? 80.15% ====================================== Files ? 61 Lines ? 5992 Branches ? 0 ====================================== Hits ? 4803 Misses ? 1189 Partials ? 0 ```

:umbrella: View full report in Codecov by Sentry.
:loudspeaker: Have feedback on the report? Share it here.

[bpbond]

Talking with @melatg0 I noticed that atmospheric CO2 is 800 ppm in 2000 --- 😕 ?!?

[kdorheim]

Talking with @melatg0 I noticed that atmospheric CO2 is 800 ppm in 2000 --- 😕 ?!?

Hi Ben. Okay, so we've run into this issue in the past, and it has to do with the units (remember how confusing the atmospheric C vs CO2 was so we added the CONCENTRATIONS_CO2() variable). You are plotting the atmos_co2 variable which is in Pg C vs CO2_concentration which is in ppmv CO2. Which means that the values you and @melatg0 were looking at are atmos_co2 786.4845 Pg C or CO2_concentration 369.2416 ppmv CO2 which is a little high but reasonable

[bpbond]

Haha thank you...again!

[kdorheim]

Since [CO2] is kind of a catch-all (changes in ocean carbon cycle & terrestrial carbon cycle) would also cause [CO2] to change I lean towards adding a pf variable to the old new test, it shouldn't increase the size too much... what do you think?

[bpbond]

Since [CO2] is kind of a catch-all (changes in ocean carbon cycle & terrestrial carbon cycle) would also cause [CO2] to change I lean towards adding a pf variable to the old new test, it shouldn't increase the size too much... what do you think?

I think it's not needed, but doesn't hurt. Will do.

[github-actions[bot]]

Differences in Hector outputs

Hello, this is leeyabot!

The current pull request's outputs do not differ from 3.1.1 (87c5134a).