PMIP 2 Boundary Conditions

[Check the FAQ, if you want to print this page nicely...]

The following are guidelines for the boundary conditions for the PMIP 2 coupled models experiments.

We require each group to document precisely what it does!

Analysis will be performed over the final 100-200 years of all the simulations,
after the stage when the trends are small.

Pre-industrial control run

Notes

The minimum requirements for PMIP 1 were that we all used the exact same change in forcing.

For PMIP 2, we recommend that any group embarking on new control simulations use the above recommendations. It is important to have pre-industrial CTRL simulations.

1. Vegetation will be provided for anyone interested.
2. Orbital parameters values are for 1950 AD.
3. Each group has its own spinup methodology, which we ask they document.
   
   The model should be run long enough for any trends to be small, and the control run should be at least as long as any anomaly integrations.

Mid-holocene 6k BP

Notes

1. Insolation 
   
   Make sure you check the insolation related tables at the end of this page!
   
   
   • Insolation is the most important change to be prescribed in the boundary conditions for the 6kyr BP experiment. The PMIP 2 runs must use the exact same forcing!
     
     
   • The orbital parameters are given by Andre Berger (Long-term variations of daily insolation and Quaternary climatic changes, JAS, 35, 2362-2367, 1978).
     
     
   • The date of Vernal Equinox:
     
     It is necessary that we all use the same reference date for the 6kyr BP experiment, if we want to compare "comparable" climates. We then recommend that you set the 21st of March at noon (e.g. 21.00) as the date of your vernal equinox for the 6kyr BP simulation (for 360 as well as for 365-day year).
     
     Indeed, defining a calendar at 6kyr BP is arbitrary. Nevertheless it becomes very critical when we want to compare 2 climatic periods. We then need to know how to phase the 6kyr BP insolation pattern with the insolation pattern for the present-day climate. This is not trivial because the time intervals between equinoxes and solstices varies with the orbital parameters.
     For example, the number of days between the winter solstice and the vernal equinox changes from 89 days nowadays to 93 days 6kyr BP (for a 365-day year).
     Fixing the present date for the vernal equinox or for the winter solstice, at 6 kyr BP, then leads to 2 calendars differing by 4 days around the vernal equinox. A drift of 4 days in the insolation pattern is important: it leads to differences in insolation of the order of magnitude of the changes induced by the change in orbital parameters!
     
     
   • The calendar:
     
     Should we use a celestial calendar to compute monthly means at 6ka?
     
     Following the relationship between seasons and insolation (Joussaume, S. and P. Braconnot, 1997: Sensitivity of paleoclimate simulation results to season definitions. Journal of Geophysical Research (Atmosphere), 102, 1943-1956) would yield more accurate results, but may be difficult to implement in some of the models, with a risk of error which may be difficult to detect...
     
     The following tables list the first day and the length of the angular months, depending on the year length, with vernal equinox fixed to be March 21^st. The dates are relative to the present calendar months.
     
     126 ka has been included to help see the difference between two climatic periods.

     File type	Length of year (days)
     	365	360
     HTML	www	www
     TEXT	txt	txt

      
     
     Each group will have to tell if it is possible or not to introduce such months' limits in their model or post-processing software. A subset of daily values will be available.
     
     
   • Checking insolation changes:
     
     Before you start your simulations, we strongly advise that you check your computed insolation values. Indeed, Andre Berger has warned us about the various approximations used by the GCMs that may induce differences in insolation that are not negligible with respect to the Milankovitch forcing. We thus provide you with tables for 1) the present- day and 2) 6 kyr BP minus present-day.
     We think that differences larger than 10% between your calculations of 6 kyr BP minus present-day and ours should be corrected because they may significantly alter the model-model comparisons.
     If your calculations differ that much from our tables, we recommend that you carefully check your insolation code and send us a mail message reporting on the differences you found. We can then help clear up the differences.

Last glacial maximum 21k BP

Notes

1. We will use a new ice sheet reconstruction, ICE-5G, provided by Dick Peltier. The largest difference between ICE-5G and the previous version (ICE-4G), used in the first phase of PMIP, is that it takes into account new information about the extent of glacial maximum ice sheets in Eurasia provided by the QUEEN project.
   The data will be provided on the same grid as ICE-4G. Boundary conditions will then be adapted to the model grid using the same protocol used in PMIP 1.
   
   
2. Use the land-sea mask and change in topography supplied by Dick Peltier, consistent with the ice sheet in ICE-5G, rather than a simple lowering using present day bathymetry (change land height, don't make oceans shallower). The data set will tell what increase in land elevation has to be applied to each grid cell (land is then cells above LGM 0 level).
   Check narrow or low depth straits carefully!
   
   
3. Greenhouse gases
   
   Monnin, E., A. Indermuhle, A. Dallenbach, J. Fluckiger, B. Stauffer, T.F. Stocker, D. Raynaud, and J.M. Barnola, Atmospheric CO₂ concentrations over the last glacial termination, Science, 291 (5501), 112-114, 2001.
   
   Dallenbach, A., T. Blunier, J. Fluckiger, B. Stauffer, J. Chappellaz, and D. Raynaud, Changes in the atmospheric CH₄ gradient between Greenland and Antarctica during the Last Glacial and the transition to the Holocene, Geophysical Research Letters, 27 (7), 1005-1008, 2000.
   
   Fluckiger, J., A. Dallenbach, T. Blunier, B. Stauffer, T.F. Stocker, D. Raynaud, and J.M. Barnola, Variations in atmospheric N₂O concentration during abrupt climatic changes, Science, 285 (5425), 227-230, 1999.
   
   
4. River routing
   
   Sandy Harrison has provided river pathways estimates at 5-minute resolution using HYDRA (Harrison, Bartlein, Coe and Sickel, in prep.).
   
   
5. Ice sheet melt
   
   Provide a simple solution for models in which something has to be done.
   For models with only a global closure of the water budget, introduce this excess fresh water in the global estimate to ensure water conservation in the model.
   
   
6. Each group has its own spinup methodology, which we ask they document.
   
   If initializing from modern conditions (Option 1):
   1. Option 1a: we recommend doing a 100-year Haney forced run, restoring the SSTs to those of a slab model LGM anomaly SSTs.
   2. Option 1b: or 100-year Haney forced run, restoring to CLIMAP LGM anomaly SSTs.
   3. Option 1c: or ocean-only stage restoring SSTs to slab or CLIMAP LGM
   
   
   4. Option 1d: or just run it (brute force approach)!
   
   
   For both Options 1 and 2:
   Run the OAGCM (initialized from either the end of Option 1, or from the cold state of an existing LGM simulation) for long enough for any trends to be small.
   

Trend diagnosis

The models will have to be run for long enough for any trends to be small (at least 100 years of spinup).

Use the annual mean time series of the following variables to summarize large-scale model behaviour and trends:

• Check open/closed throughflows
• Global average of surface temperature
    trend in global annual mean should be no larger than -0.05K/Century.
• Global 2m air temperature
• Global sea surface temperature
• Global ocean temperature and salinity (volume), and in each basin
• Min and max THC strength
• Flow to southern ocean at 20° South
• Northern and southern hemisphere sea ice volume
• Evolution of forest area
• Other vegetation-related parameters?

About insolation computation

In the following, we provide tables and information concerning insolation in order to help you check your insolation code.

All the results we give have been obtained using:

• The orbital parameters given above.
• A solar constant value of 1365 W/m².
• A calendar based on the 21st of March at noon (21.00) for the date of the vernal equinox.

All the values of insolation are given in W/m². They are given at every 10 degrees of latitude (no latitudinal band average is done!). All the computations follow the method proposed by Berger (JAS, 1978) and are based on an expansion accurate to order e**3 for the computation of the true longitude (lambda, angle defining the Earth position relative to the Vernal Equinox).

Dates of Equinoxes and Solstices

Insolation monthly means

The insolation values for monthly means depend on the length of the year and on the reference date used (21.00 March vernal equinox in the following table).

Note: the text files you can download in the tables below have unix-style end of lines. If you are using Windows, you can view them directly in your browser, or you can download them and display their content with WordPad.

Insolation mid-month values

The insolation mid-month values are obtained as daily mean insolation values in W/m² and are computed at fixed true longitudes with longitude increments of 30°, starting from the vernal equinox (Berger, JAS, 1978) ... i.e. around the 20^th of each month.

Using this definition, we have:

These tables of mid-month values:

• Allow direct comparisons of insolation at equinoxes and solstices.
• Avoid any problem of calendar, either between 0 and 6 kyr BP or between 360-day or 365-day years.