Hi,
We are seeking guidance for the CLM BGC spin-up. We have reviewed the previous posts on this forum regarding this topic, but our confusion remains unresolved. Therefore, we would like to pose some questions regarding the BGC spin-up.
We need to perform CLM BGC simulations for the future (2015-2100) with our own climate forcing and land use land cover change data at the CONUS scale with 1/8 degree resolution. Our own LULCC info were generated based on Global Change Analysis Model (GCAM) output downscaled by DEMETER model.(The GCAM model is publicly available at GitHub - JGCRI/gcam-core: GCAM -- The Global Change Analysis Model. The Demeter model is publicly available at GitHub - JGCRI/demeter: A land use land cover disaggregation and change detection model)
Since these are BGC runs, we need to perform a spin-up to reach equilibrium conditions. In our understanding, the initial equilibrium should be achieved for the preindustrial conditions of 1850 by using the 1850 surface dataset, fixed atmospheric CO2, fixed N deposition, fixed aerosol deposition, and by recycling climate forcing. Secondly, we perform transient historical simulations from 1850-2015, followed by future simulations from 2015-2100 with transient CO2, N deposition, aerosol deposition, climate forcing, and land use timeseries file.
However, we don't have our own GCAM-based LULCC data for the historical period. We are uncertain about the reasonableness of performing a spin-up (1850, AD + final spin-up) and transient simulation (1850-2015) using the surface dataset and land use timeseries from the NCAR CCSM input data repository, and then conducting future simulations using our own GCAM-based land use timeseries file. Our guess is that we won't be able to use the restart file from the end of the historical transient simulation for our future simulation due to inconsistencies between the two LULCC datasets, i.e., NCAR's historical and our own GCAM-based future land use timeseries. Is there any namelist option to bypass this inconsistency? We are aware of check_dynpft_consistency, but our understanding is that it only avoids checking for consistency between pct_nat_pft from the landuse_timeseris file and pct_nat_pft from the surface dataset instead of between two land use time series.
On the other hand, we have noticed that some researchers do spin-ups for the year when the actual simulation takes place. For example, they perform a spin-up using 2000 surface data (and other data such as nitrogen deposition) before starting a 2000-2015 run. We are unsure about the potential differences in the spin-up results between these two methods and whether such differences could have a significant impact on BGC processes.
We are eager to receive any guidance or suggestions regarding our spin-up process. Thank you very much for your assistance!
We are seeking guidance for the CLM BGC spin-up. We have reviewed the previous posts on this forum regarding this topic, but our confusion remains unresolved. Therefore, we would like to pose some questions regarding the BGC spin-up.
We need to perform CLM BGC simulations for the future (2015-2100) with our own climate forcing and land use land cover change data at the CONUS scale with 1/8 degree resolution. Our own LULCC info were generated based on Global Change Analysis Model (GCAM) output downscaled by DEMETER model.(The GCAM model is publicly available at GitHub - JGCRI/gcam-core: GCAM -- The Global Change Analysis Model. The Demeter model is publicly available at GitHub - JGCRI/demeter: A land use land cover disaggregation and change detection model)
Since these are BGC runs, we need to perform a spin-up to reach equilibrium conditions. In our understanding, the initial equilibrium should be achieved for the preindustrial conditions of 1850 by using the 1850 surface dataset, fixed atmospheric CO2, fixed N deposition, fixed aerosol deposition, and by recycling climate forcing. Secondly, we perform transient historical simulations from 1850-2015, followed by future simulations from 2015-2100 with transient CO2, N deposition, aerosol deposition, climate forcing, and land use timeseries file.
However, we don't have our own GCAM-based LULCC data for the historical period. We are uncertain about the reasonableness of performing a spin-up (1850, AD + final spin-up) and transient simulation (1850-2015) using the surface dataset and land use timeseries from the NCAR CCSM input data repository, and then conducting future simulations using our own GCAM-based land use timeseries file. Our guess is that we won't be able to use the restart file from the end of the historical transient simulation for our future simulation due to inconsistencies between the two LULCC datasets, i.e., NCAR's historical and our own GCAM-based future land use timeseries. Is there any namelist option to bypass this inconsistency? We are aware of check_dynpft_consistency, but our understanding is that it only avoids checking for consistency between pct_nat_pft from the landuse_timeseris file and pct_nat_pft from the surface dataset instead of between two land use time series.
On the other hand, we have noticed that some researchers do spin-ups for the year when the actual simulation takes place. For example, they perform a spin-up using 2000 surface data (and other data such as nitrogen deposition) before starting a 2000-2015 run. We are unsure about the potential differences in the spin-up results between these two methods and whether such differences could have a significant impact on BGC processes.
We are eager to receive any guidance or suggestions regarding our spin-up process. Thank you very much for your assistance!
