Too high veg carbon in East US and China

[rosiealice]

Vegetation carbon is too high overall.

https://ns9560k.web.sigma2.no/datalake/diagnostics/noresm/kjetisaa/NorESM_Key_Simulations/n1850.ne16pg3_tn14.noresm3_0_beta09.Run4.2026-01-06/OBS_comparison/ANN/n1850.ne16pg3_tn14.noresm3_0_beta09.Run4.2026-01-06_compare_FATES_VEGC_ESACCI_ANN_0221-0230.png

[rosiealice]

Is GPP/NPP/LAI too high here?

https://ns9560k.web.sigma2.no/datalake/diagnostics/noresm/rosief/n1850.ne16pg3_tn14.noresm3_0_beta09.Run3.2025-12-30/seasonal_cycle/cfluxes/n1850.ne16pg3_tn14.noresm3_0_beta09.Run3.2025-12-30_cfluxes_EasternU.S..png

[rosiealice]

Is it turnover?

[kjetilaas]

Some point plots in South East US, which all show large positive vegC biases, while GPP and LAI looks ok:

[rosiealice]

https://ns9560k.web.sigma2.no/datalake/diagnostics/noresm/rosief/n1850.ne16pg3_tn14.noresm3_0_beta09.Run3.2025-12-30/clim_maps/ANN/

[rosiealice]

[rosiealice]

So there is maybe not enough size related mortality here? Need to discuss how to efficiently calibrate this...

[kjetilaas]

Some point plots from China. It seems like the dominant tree type is evergreen needleleaf extratropical (NEET). LAI seems to be up to 2x the obs, GPP ok or somewhat too low, vegC is 3-4x too high.

Note that obs are present day and simulations (incl. LU) are 1850, so not directly comparable...

[JessicaNeedham]

Based on discussion above we started a mini PPE to explore reducing the Veg C bias in the SE US and in China. Since GPP is not that biased we focused on mortality parameters for the two dominant PFTs - needleleaf evergreen extratropical tree and broadleaf deciduous extratropical tree.

<style> </style>

Simulation	Parameter	PFT	Value
1	bmort	NEET	0.006
2	bmort	NEET	0.008
3	bmort	BDET	0.006
4	bmort	BDET	0.008
5	size mort	NEET	180
6	size mort	NEET	170
7	size mort	BDET	165
8	size mort	BDET	155
9	default parameter file

PPE 1

• 4x5 degree
• 2000s climate, CO2, land use
• bmort and ip_size_senescence
• 40 year spin up and then start the ensemble
• 9 ensemble members (including default)

Since size-independent and size-dependent mortality are expected to have different CO2 fertilization responses we also run a PPE from 1850 to present day.

PPE 2

• ne 16
• 185 year spin up then start ensemble in year 1850
• 1850 - 2023 with transient climate, CO2 and land use
• 5 ensemble members (2,4,6,8 from above table plus the default)

[JessicaNeedham]

~50 years into the ensemble (90 years total) it looks like bmort of the broadleaf deciduous extratropical tree has the largest impact on Veg C, with little change on GPP.

Sim 3 - updated bmort

Default

[JessicaNeedham]

Will run these for longer to allow any impacts of the size-dependent mortality parameters to be assessed. Historic runs are also still running.

[JessicaNeedham]

140 years in the run with higher values of bmort for the BDET seems to have a much reduced bias in SE US and a slightly improved bias in China.

Size distributions looks reasonable I think, although there aren't observations to compare with.

[JessicaNeedham]

There were some issues with the biomass datasets in the diagnostics package which have now been fixed (see NorESMhub/xesmf_clm_fates_diagnostic#55).

There is quite large variation among the different datasets. But looking at ESA we now see high biases in the tropics (200 years in):

With increased size-mortality that bias goes down (350 years in):

The high biomass bias in the Eastern US and China remains.