UM_Bristol

Here we give a brief summary of our experiments. We plan to have several phases of experiments.

Single testing phase, where the primary goal is to test the functionality of our models (to resolve different forcings) to participate in CMIP7. And as several of us are responsible for a specified forcing, we will document them in different categories.
Combined testing phase, where various forcings are configured together to make sure the models are working fine.
CMIP7 experiments phase, where we make use of CMIP7 forcings. These will be submitted.

Single testing phase

1. Historical greenhouse gas concentrations

experiment	description	running length	source
xqcha	test run	/	<- xpuoo
xqchb	control run	1890-2090	<- xpzab <- xpwca
xqchc	CMIP6 GHGs	1890-2090	<- xqchb
xqchd	CMIP6 GHGs + CMIP6 O₂ conc	1890-2090	<- xqchc

2. Historical CO₂ emissions

experiment	description	running length	source
xqche	emission test	1650-1850	<- xqchb
xqchf	SRESA2 emissions, with TRIFFID bug	1750-2100	<- xqchb
xqchh	SRESA2 emissions, bug fixed	1750-2100	<- xqchf
xqcht	CMIP6 emissions, with TRIFFID bug	1750-2014	<- xqchb
xqchi	CMIP6 emissions, bug fixed	1750-2014	<- xqcht
TBC	CMIP7 emissions		<- xqchb
xqchk	CMIP6 emissions, veg timestep bug fixed	1750-2014	<- xqchi

3. Solar forcing

xqcpa: This job is based on a standard HadCM3B job, tdaag, whose setup is HadCM3-MOSES2.1-TRIFFID_dyn, Pre-industrial. Job xqcpa further includes CMIP6 Total Solar Irradiation data (annual mean) and run for 450 years (1850-2299).

solar file set as BC4:/user/home/mf22281/um_updates/varying_TSI_CMIP6.dat

please refer to [[CMIP7_fcg_solar

Solar forcing for HadCM3]] for more information.

To be notice, other jobs related to solar forcing need to be run later, including 1) uses monthly CMIP6 TSI data, 2) uses CMIP6 spectrum data (different solar irradiation value for different spectrums), 3) combine CMIP6 monthly/annual TSI and spectrum data, 4) use CMIP7 data (?)

4. Volcanic forcing

5. Land-use forcing

experiment	description	running length	source
xqcna	test dyn TRIFFID, static Land Use (LU)	1850 - 1880	<- xqavc
xqcnb	dyn and compressed LU, BUG	1850 - 1880	<- xqcna
xqcnc	dyn and uncompressed LU, BUG	1850 - 1880	<- xqcna
xqcng	dyn and compressed sequencial LU, new mods, BUG FIXED	1850 - 1880	<- xqcnc
xqcnh	dyn and uncompressed sequencial LU, new mods, BUG FIXED	1850 - 1880	<- xqcng
xqcni	dyn and uncompressed Hyde LU, new mods, BUG FIXED	1850 - 1880	<- xqcng

6. Sulphur cycle/aerosol scheme

7. Ozone in HadCM3 and how to include CMIP forcing

Combined testing phase

experiment	description	running length	source
xqchz	emissions + solar forcing	1750-2014	<- (xqchi,xqcpa)
xqchy	emissions + solar + land-use	1750-2014	<- (xqchz,xqcni)
xqchx	emissions + solar + land-use + GHGs	1850–2022	<- (xqchy,xqchd)
xqhug	aerosol run (sulphate direct effect)	TBC	<- xqhuc

xqhug

Aerosol (sulphate) run based on the updated-mods spin-up xqhuc. Only the direct effect (ARI) is enabled; the indirect effect (ACI) is switched off.

Aerosol indirect effect crashes

L_USE_SULPC_INDIRECT_SW and L_USE_SULPC_INDIRECT_LW cause systematic negative theta/pressure crashes every year or two despite compiling correctly. Root cause not yet identified.

Solution: keep indirect effect switched off for now. Only direct effect (L_USE_SULPC_DIRECT=.TRUE.) is used.
RECONA hand-edits required

Turning on the SO2 tracer (item 79) via the UMUI panel may not propagate correctly to the RECONA file, and the SO4 mode prognostics are absent from the start dump entirely. The correct entries are:
```
### Tracer fields ###
&ITEMS ITEM=79,  DOMAIN=1, SOURCE=6, USER_PROG_RCONST=0.0 &END
### Atmos user-prognostic fields ###
&ITEMS ITEM=101, DOMAIN=1, SOURCE=3 &END
&ITEMS ITEM=103, DOMAIN=1, SOURCE=3 &END
&ITEMS ITEM=104, DOMAIN=1, SOURCE=3 &END
&ITEMS ITEM=105, DOMAIN=1, SOURCE=3 &END
&ITEMS ITEM=106, DOMAIN=1, SOURCE=3 &END
```
ITEM=79 must use SOURCE=6 (initialise to constant 0.0), not SOURCE=3 — the field does not exist in the original dump and RECON will abort with Stash code 79 not found on input file. Items 101 and 103–106 use SOURCE=3 (read from ancil) and must be inserted after ITEM=222. Use the aero_recon post-script to apply both fixes automatically: ~/um_updates/post_scripts_puma2/aero_recon <jobid>.

CMIP7 experiments phase

Test on HadCM3

experiment	description	running length	source
xqgtz	copy from xqgra, Tuning HadCM3	/	xqgra
xqgta	Equilibrium TRIFFID, under preind	40	xqgtz
xqgtb	Spin-up, Dynamic TRIFFID, under preind (varying solar, volcanic, and land-use)	200	xqgta
xqgtc	Spin-up, Dynamic TRIFFID, under preind (fixed solar, volcanic, and land-use)	165	xqgtd
xqgtd	Emission-driven test, without emissions	173	xqgtz
xqgte	Emission-driven test, cmip7	173	xqgtd
xqgth	copy from xqgtb, add GHGs	173	xqgtb

However, spun-up results for the carbon cycle are not satisfying. We modified some mods (mainly carbon cycle mods) and redo the spin-up.

experiment	description	running length	source
xqhua	updated mods, preindustrial Equilibrium TRIFFID spin-up	40	xqgta
xqhuc	updated mods, preindustrial Dynamic TRIFFID spin-up (fixed forcing)	200	xqgtc
xqhud	updated mods, Emission + GHGs	5	xqhue
xqhue	updated mods, CMIP7 Emission (calibrated based CEDS obs)	5	/
xqhuf	updated mods, Emission + Vol	5	xqhue
xqhuh	updated mods, GHGs	5	xqhuc
xqhui	updated mods, Test soil params to UM default, V crit off	5	xqhuc
xqhuj	Eqbm run, test plant stress (off)	40	xqhua
xqhuk	Eqbm run, test plant stress (off) and CCN (on)	40	xqhua
xqhul	CMIP7 Emission + GHGs + LU + Solar + Vol	173	xqhud
xqhum	CMIP7 all forcings (+ aerosol), debugging	TBC	<- (xqhul, xqhug)
xqhuu	check stash	/	xqbtn <- xqbmg
xqhuz	Tuning version of HadCM3C	/	xqhra
xqjlb	Dynamic TRIFFID spin-up, tuned params from xqjch	200	xqhuc
xqjlc	Dynamic TRIFFID spin-up, tuned params from xqjcj	200	xqhuc

xqjlb / xqjlc

Dynamic TRIFFID spin-up experiments using tuned terrestrial carbon cycle parameters from the soil ensemble selection. Both start from the updated-mods preindustrial spin-up xqhuc and run for 200 years with fixed forcing.

xqjlb: uses the parameter set from xqjch (V_CRIT_ALPHA = 0.361)
xqjlc: uses the parameter set from xqjcj (V_CRIT_ALPHA = 0.314)

These test whether the equilibrium-TRIFFID tuning results hold under dynamic TRIFFID spin-up. See CMIP7 soil ensembles for validation results.

xqhum

All-forcings run combining xqhul (Emission + GHGs + LU + Solar + Vol) with the sulphate aerosol scheme from xqhug. Currently being debugged.

[!NOTE] Document bugs and fixes here as they are identified.