We are in the process of preparing the boundary conditions to be used in the model simulations. Some of the boundary conditions, such as the orbit and trace gases, are straightforward. However, preparing the land sea masks, bathymetry, and ice masks is not entirely automatic. We would therefore value your input in order to try and make these boundary conditions as realistic as possible.

Our starting point is the Ice 5g data provided by Dick Peltier. This is in the form of 1 degree x 1 degree maps of surface elevation/bathymetry, ice sheet thickness (we don't use this) and ice mask. They are provided every 500 years from the present to 17kyr BP, and then every 1000 years to 21kyr BP.

We have to process these fields to provide input into the Hadley Centre models, and specifically the FAMOUS model. This model uses two grids. There is a grid associated with the ocean model which is at 3.75 degree x 2.5 degree resolution, and a grid associated with the atmosphere and land surface which is at 7.5 degree x 5 degree resolution.

The following table shows a variety of results from our regridding process. If you don;t have time to look at all of them, please focus on entry 1 since this is the most important one that can be manually modified.

1 Land sea mask on the ocean model resolution of FAMOUS (3.75 x 2.5 degrees)	This is the most important of the input fields, as it defines the overall geometry of the model. The mask can only take the value 0 (for ocean) or 1 (for land). We prepare this field by firstly using the Peltier surface elevation map to produce a 1 x 1 degree land sea mask by making all elevations greater than zero to be land, and all other grid points are ocean. NOTE that this introduces some issues since it incorrectly handles land areas which are below zero, or expansion to lakes/seas due to changes in hydrological regime (e.g. the Caspian sea which was expanded in glacial times due to changes in drainage). The resulting 1 x 1 degree land sea mask is then regridded onto the ocean resolution (3.75 x 2.5) using one of a number of regridding algorithms that we use. We have chosen to the scheme provided by AMIP (atmospheric model intercomparison project). This regridding produces a mathematically correct representation but may introduce some "odd" artifacts such as isolated ocean points. It is possible to intervene at this stage and make manual changes. We have done this for the present day and 21kyr BP land masks but not any time period in between. Please take a close look at the resulting plots and advice us of whether you see any major errors. Of particular interest are: When does the Caspian sea regain its normal size (start of Holocene? There is apparent water around the Laurentide during deglaciation. Is this real or an artifact of using a definition of ocean is anything below sea level> Similar comments also refer to Fennoscandinavian ice sheet, but I think this may be real	FAMOUS ocean-based Land sea mask plots (small figures) Should fit on one screen. FAMOUS ocean-based Land sea mask plots (large figures) May require big screen to view properly
2 Land sea mask on the atmosphere/land grid of FAMOUS (7 x 5 degrees)	This is produced from the ocean model but using an algorithm intrinsic to FAMOUS. It can only take values of 0 or 1, where a land point is anything which has a non-zero land fraction. Therefore this land sea mask looks a bit odd, and is very extensive. We cannot change it manually and it is only really used for technical issues.	FAMOUS Land sea mask plots (small figures) FAMOUS Land sea mask plots (large figures)
3 Fractional Land Sea Mask at FAMOUS resolution (7 x 5 degrees)	FAMOUS is the only one of the HadCM3 suite of models that has a special "coastal tiling" scheme which allows the model to have a smoothly varying land sea fraction around the coastline. This is again calculated from the ocean model using an algorithm intrinic to FAMOUS and we therefore cannot simply change it. I show these for reference purposes only.	FAMOUS Land fraction plots (small figures) FAMOUS Land fraction plots (large figures)
4 Land Ice Mask at FAMOUS resolution (7 x 5 degrees)	Ice 5G also contains the land ice fraction and we simply have to regrid this onto the FAMOUS model resolution (7 x 5 degrees). Land ice fraction can only be 0 or 1, and is zero over the ocean by definition. I show these for reference purposes only.	FAMOUS Land ice plots (small figures) FAMOUS Land ice plots (large figures)
5 Surface Elevation on FAMOUS resolution (7 x 5 degrees)	These are computed from the Ice 5G surface elevation by using the same AMIP-based algorithm at the atmosphere/land resolution of 7 x 5 degrees. Elevations are set to zero for ocean points (as defined by the integer land-sea mask described in 2 above). We can manually modify this if desired. There are also several other orographically related fields (e.g. sub-grid scale variability) which need to be supplied but they are all derivatives of the elevation itself.	FAMOUS Elevation plots (small figures) FAMOUS Elevation plots (large figures)
6 Land Ice Mask at HadCM3 resolution (3.75 x 2.5 degrees)	As 4 but at HadCM3 resolution. Some time slice experiments may use this resolution.	HadCM3 Land Ice plots (small figures) HadCM3 Land Ice plots (large figures)
7 Surface Elevation at HadCM3 resolution (3.75 x 2.5 degrees)	As 5 but at HadCM3 resolution. Some time slice experiments may use this resolution.	HadCM3 Elevation plots (small figures) HadCM3 Elevation plots (large figures)
8 Changes in Land Sea Mask on the ocean model resolution of FAMOUS (3.75 x 2.5 degrees)	As 1 above, but showing the difference relative to present day.	Difference plots FAMOUS (ocean resolution) land sea mask plots (small figures) Difference plots FAMOUS (ocean resolution) land sea mask plots (large figures)
9 Changes in Land Sea Mask on the atmosphere/land resolution of FAMOUS (7.5 x 5.0 degrees)	As 2 above, but showing the difference relative to present day.	Difference plots FAMOUS land sea mask plots (small figures) Difference plots FAMOUS land sea mask plots (large figures)
10 Changes in fractional Land Sea Mask at FAMOUS resolution (7.5 x 5.0 degrees)	As 3 above, but showing the difference relative to present day.	Difference plots FAMOUS fractional land sea mask plots (small figures) Difference plots FAMOUS fractional land sea mask plots (large figures)
11 Changes in surface Elevation on FAMOUS resolution (7 x 5 degrees)	As 5 above but showing difference relative to present day.	Difference FAMOUS Elevation plots (small figures) Difference FAMOUS Elevation plots (large figures)
12 Changes in Surface Elevation at HadCM3 resolution (3.75 x 2.5 degrees)	As 11 but at HadCM3 resolution. Some time slice experiments may use this resolution.	Changes in HadCM3 Elevation plots (small figures) Changes in HadCM3 Elevation plots (large figures)
13 Bathymetry at FAMOUS ocean resolution (3.75 x 2.5 degrees)	These are computed from the Ice 5G surface elevation by using the same AMIP-based algorithm at the ocean resolution of 3.75 x 2.5 degrees. Bathymetries are set to zero for land points. We will have to smooth this raw bathymetry to make sure thatthe ocean model is numerically stable.	FAMOUS bathymetry plots (small figures) FAMOUS bathymetry plots (large figures)
14 Changes in Bathymetry at FAMOUS ocean resolution (3.75 x 2.5 degrees)	As 13 but at showing the changes in bathymetry.	Changes in FAMOUS bathymetry plots (small figures) Changes in FAMOUS bathymetry plots (large figures)