skip to primary navigationskip to content

Climate & Plant Cover Simulations

Accessing the Penn State University (PSU) Palaeoclimatic Simulations: Introduction, Instructions and Basic Glossary for Users 

This text can be downloaded as a MS-Word file, and it is recommended that you do so. 


Each climate simulation run started with GENESIS GCM, a global circulation model that provided the boundary conditions for a high-resolution regional circulation model (REGCM2) that covers Europe and adjacent oceans and seas. The GCM runs are available but are of limited interest for most Stage 3 purposes. 

The currently available output of RCM simulations consists of sets of colour maps and tables of numerical data resulting from the Phase 4 1999-2000 runs. For each set the initial GCM and RCM outputs are both available, but the latter will usually be the appropriate ones for most Stage 3 applications (plant cover, fauna, archaeology, etc. 

Obtaining the simulation plots and tables

The process of downloading simulations as maps or in tabular form may seem a bit intimidating at first, but the instructions below should be helpful. Although above word processing level, all of it is quite routine; if you run into difficulties you will most likely find adequate assistance in your own place. Any questions should be directed to David Pollard

The palaeoclimate simulation maps and tables are available by anonymous ftp from the PSU server. When you click the link, you enter an FTP site at Penn State's Earth and Mineral Sciences Environment Institute. You can move around in this site just like in any other web page. 

There are README files in each directory which describe the contents of that directory. If you print out those README files, you can use them as a menu to help navigate through the abbreviated file names in each directory. 

Overall README File (uppermost directory)
Plot files and gridded data files are available for the "final" Genesis GCM/RegCM2 RCM Stage 3 experiments run in 2000 AD. The subdirectories containing the files are:

DATA: data files for RegCM2 regional model results
PLOT: plot files for RegCM2 regional model results
DATAGCM: data files for Genesis global model results
PLOTGCM: plot files for Genesis global model results
README: files in each subdirectory describe the files

The plot files with .pdf suffix are viewable on the web and do not necessarily need to be downloaded if you have Adobe Acrobat or a similar application. Click on the file name, and the plot should appear on your screen. All files are downloadable from the web to your computer. On a PC or SUN work station, right-click on any file name. A menu will pop up. Select, "Save link as..." and save the file on your computer.

The Biome3.5 Maps

The Biome 3.5 simulations themselves can be found in the palaeo-climate archive at the address above, but their interpretation in terms of vegetation units is in doubt. They have been sensitivity tested by Mary Jo Alfano, however they are not easily exploited by the uninitiated. Brian Huntley and Judy Allen, using Mary Jo's tests, are working on a practical solution and will as soon as possible produce help for those authors who need vegetation patterns. In the interim, the Biome 3.5 outputs should not be used for publication.

Caution: Outdated PSU Simulations
The coloured maps you have seen at meetings and in NEWS reports, have been sets of four panels per sheet which, for a large number of climatic variables called FIELDS, displayed:

a) the modern equivalent;
b) a cold event with small ice-sheet;
c) a cold event with large ice-sheet;
d) a warm event with small ice-sheet.

These sets from the Phase 2 and 3 simulations should not be used for interpretation and publication.

The 1999-2000 Phase 4 Output

This is the 1999-2000 and final Stage 3 output which is now available for research purposes. It covers four climate states; (1) the modern equivalent, (2) a warm Dansgaard/Oeschger (D/O) event and for comparison (3) the Last Glacial Maximum at ca. 21 ka BP, as well as (4) a Stage 3 cold D/O event in two versions:

  • Modern Interval (MODERN):
    The present equivalent of the variable: Input: from observed climatological data and CO2 = 345 ppmv.
  • Warm Interval (WARM):
    Input data: Sea surface temperatures (SSTs) from CLIMAP, GLAMAP Atlantic and 40 ka North Atlantic cores (SST Panel); small Fennoscandian RCM ice sheet (van Andel and Arnold); CO2 = 200 ppmv.
  • Late Glacial Maximum (21 ka):
    Input data: SSTs from CLIMAP and GLAMAP Atlantic (SST Panel); GCM and RCM ice sheet 21 ICE-4G model; CO2 = 200 ppmv.
  • Cold Interval (COLD):
    Input data: SSTs from CLIMAP, North Atlantic GLAMAP and 30 ka North Atlantic cores (SST Panel); a GCM ice sheet based on the Younger Dryas (John Andrews) and a small Fennoscandian one (van Andel and Arnold); an RCM ice sheet based on a large Fennoscandian one (Arnold and Lambeck); CO2 = 200 ppmv.
  • Cold Interval (COLDC):
    Input data: As COLD above, with ad hoc SSTs, Bay of Biscay sea-ice and longer sea-ice duration.

The Phase 4 Workshop 1 rejected simulations COLD and COLDC which, although important in understanding glacial climate modelling, are not adequately supported for their use in climate and landscape models and publications. Michael Sarnthein and co-workers intend to extract more SST and sea-ice data from IMAGES Project and other cores, but acceptable cold D/O simulations will not be available for a few years. Instead, the Late Glacial Maximum simulation will serve as an extreme cold climatic state, the Modern Interval being the warm extreme.

NOTE: Because of chronological problems the Project has from its inception regarded the simulations as typical for a cold and a warm event without committing itself to any specific cold or warm event or date reflected in the Greenland ice and Atlantic sediment cores.

Users should decide for themselves whether they need to label the Warm Interval with calendar dates, but if they do so, the input SST data, which are from a warm event at about 40 cal ka BP the logical choice based on the GISP2 core chronology, but the use of land-based named interstadials such as Hengelo or Denekamp does seem unwise.

Codes and description: Phase 4 Output

The Phase 4 output provides useable simulations for three climatic states with many variables (labelled FIELDs).

Climatic states:

Modern (MODERN)
Warm Interval (WARM) with small Fennoscandian ice sheet
Late Glacial Maximum (21 ka)


Winter (December, January, February): DJF
Summer (June, July, August): JJA
Spring (March, April, May; MAM) where important
Autumn (September, October, November; SON) where important.


Above the surface are given in sigma coordinates (s = pressure/surface pressure; s = .995 is approx. 995 mb (millibars):

For the RCM output: For the GCM output:
  km (approx.)
  km (approx.)
= 11
= 10
= 5
= 5
= 1.7
= 1

Codes for Simulations

The full list of available GCM and REGCM2 fields numbers over 150, many of those not of direct interest to the palaeo-environmental aspects of the Stage 3 Projects. Below we list those that are most likely to be of use for non-specialists in the Project.

First part of file name Field
grid land-ocean-ice mask
topog topography, metres above sea-level


chill wind chill (in degrees Fahrenheit)
t lowest-level (~40 m) air temp. (ºC)
dt difference between lowest-level air temp. and MODERN (ºC)
tmax maximum monthly lowest-level air temp. (ºC)
tmin minimum monthly lowest-level air temp. (ºC)
trange range of monthly lowest-level air temp. (tmax-tmin) (ºC)
tran24 diurnal range of lowest-level air temp. (ºC)
tsoi soil temp., upper ~1m
ts2 2m level air temp. (ºC)

Precipitation, Hydrology

relhum relative humidity (0-1)
pme precipitation minus evapotranspiration (mm/day)
precip precipitation (mm/day)
dprecip difference in precipitation from MODERN (mm/day)
evap evapotranspiration (mm/day)
eop evapotranspiration/precipitation ratio
pbin3 fraction of time that precipitation > 3 mm/day
runoff surface runoff (mm/day)
snowd number of days/year with snow cover (1 to 365)
snowh snow depth (actual, not liquid equivalent) (cm)
wsoi soil moisture relative to pore space (0-1m), whole column (~3 m)


wind lowest-level (~40 m) wind speed (m/s)
wbin6 fraction of time that lowest-level wind speed > 6 m/s
u210 wind vectors, sigma (pressure/surface pressure) = 0.210
u510 wind vectors, sigma (pressure/surface pressure) = 0.510
u815 wind vectors, sigma (pressure/surface pressure) = 0.815


cloud cloudiness fraction of whole sky (0-1)


aetopet actual/potential annual evapotranspiration ratio (BIOME3.5)
biome biome types (BIOME3.5)
gdd0 annual growing-degree days above 0 degrees C
gdd5 annual growing-degree days above 5 degrees C
lai seasonal maximum Leaf Area Index (BIOME3.5)
npp Net Primary Productivity, gC/m2/yr (BIOME3.5)
npppft NPP of selected Plant Functional Types in gC/m2/yr (BIOME3.5)

Availability of Maps

Maps are available as follows:

  • Single panel colour plots of most variables of all Phase 4 (1999-2000) experiments
  • Two-panel plots of the most relevant variables from the Warm State and the Late Glacial Maximum (21 ka)
  • Two-panel plots of the most relevant variables from the Warm and the Modern States

Tables of Numerical Data

Plots/maps are fine tools for analysing general conditions on a regional basis, but on occasions you may need to extract the actual numerical data which are presented in tables. The tables contain the same information as the plots, but in numeric format (explanation in the README files). If extensive use of this type of data is required, it may be best to import the tables into programs such as EXCEL. This may require some restructuring, depending on the package being used.

Comments on the Significance of Some Variables

Most variables (FIELDs) are easily understood. The ground temperature gives information on the condition of the ground (e.g. permafrost), whereas the air temperature is more relevant for understanding hominid and mammalian responses. Wind chill is also important in this respect, because it indicates the temperature stress if there is no shelter from the wind. The minimum/maximum temperatures show the diurnal stress. Normally, minimum temperatures will occur during the night and so may be an important influence on the need for shelter.

For the hydrological cycle, the surface moisture balance (given by PME, precipitation minus evaporation) is important because it gives a clear idea of the water stress. Similarly, the soil moisture content is an important indicator of moisture stress on vegetation (although deeper layers within the soil also should be used). Therefore, it potentially also indicates stresses on fauna. Runoff is another important indicator of the vigour of the hydrological cycle.

Snow depth (SNOWD) is approximate and depends on estimates of how compacted the snow becomes. Still, it is probably a good, intuitive guide to snow cover; It is more easily grasped than the main model variable for snow which is the amount of water yielded by the snow if it were to be melted.

Cloud cover is split into low, middle and upper level clouds which are averaged to yield total cloud cover. Low clouds tend to be the most effective means of cooling the climate (by reflecting sunlight), whereas high clouds are more effective at warming climate (by trapping infra-red radiation). Thus knowledge of cloud cover will help explain the temperature variations. It also has an important impact on vegetation, because it is a major factor in controlling photosynthetically active radiation. Cloud cover may have little direct effect on hominids, unless they suffered from seasonal affective disorder (SAD) or rickets!

The wind variable (WIND) gives the total wind strength, whereas the two components of wind (U and V) provide wind directions. WIND is also used to calculate wind chill factor.

Many other variables are available from the model. The selection contains the most relevant ones. The other variables may help for some studies or may be used to help explain the causes of changes in the main set of variables.

It is worth adding that temperatures and winds are probably the most reliable aspects of model predictions; precipitation has greater uncertainty and clouds are the most uncertain of all. If your research critically depends on cloud cover, we strongly recommend that you discuss your work with the modellers in order to obtain a clearer understanding of the reliability of the results.