Data Sources 

Updated: 18 July 2003

There are three DODS servers for accessing the data directly over the internet from your model, visualization software, or other programs: Additionally, data are now being made available by FTP for those who cannot access any of the servers above because of firewall, compiler, or other issues.  Access to the FTP server is tightly controlled, and should be use as a last resort only if you have tried and cannot access the data sets interactively from your programs (see the section below on GSWP-2 and DODS for guidelines on using the DODS-enabled libraries for your favorite programming language or applications).  For information on accessing the FTP site, see the bottom of the page.

GSWP-2 and DODS

The final input data sets for the participating GSWP-2 LSSs will be served over the Internet via several Distributed Oceanographic Data System (DODS) servers. DODS provides a protocol for remotely accessing and subsetting large data sets (see: http://www.unidata.ucar.edu/packages/dods/). DODS-enabled clients exist via the NetCDF protocol for FORTRAN and C, and a growing number of popular graphics and analysis packages are DODS-capable. In a nutshell, the “open” statement, which typically accesses a file on local disk via its pathname, instead accesses the DODS server via its URL using HTTP addressing. DODS includes metadata, so that served data sets are self-describing. DODS reduces strains on local disk resources, and, when properly used, is not taxing on the network or the server hardware. In order to reduce trans-oceanic networking delays, we anticipate serving the data from multiple DODS servers, operating in mirror mode, in North America (COLA, USA), Asia (University of Tokyo, Japan), and Europe (IPSL, France). If this approach is intractable to one or more modelers, we will consider alternate solutions, probably involving shipping of data on magnetic media.

The GrADS-DODS server (GDS) software, in use at the North American and Asian data servers, allows for dual-mode access to compressed NetCDF data in the CF convention (including compression by gathering) — either as the original one-dimensional spatial vector of data (for access by LSSs), or as a repopulated grid (for easy display purposes). All DODS functionality is available through GDS, as well as the ability to encapsulate computational expressions and commands which direct the data server to perform calculations (using the GrADS computing engine), and return only the results to the requesting client. More information on the GrADS-DODS server is available at the GrADS web site.

The GrADS-DODS server (GDS) software also allows you to access the complete time series of any variable as if it were all in one continuous file.  In fact, there is a separate file for each month for both the monthly and 3-hourly data sets.  This preponderance of files is hidden from the GDS user because of GrADS' templating ability, using time information in the directoy and file names to concatenate separate data files into a continuous stream.

Each of the GSWP-2 DODS data repositories will have a different URL. Users should choose the server closest to their location for optimal network performance. The URLs of the GSWP-2 DODS servers are:
For example, to access the land-only (compressed by gathering) SoilClass file in North America with a DODS-enabled client, the path would be:
http://www.monsoondata.org:9090/dods/gswp/vector/fixed/SoilClass
A catalog of the data sets on the DODS server are available by accessing the directories from a web browser:
http://www.monsoondata.org:9090/dods/gswp/vector/
http://www.monsoondata.org:9090/dods/gswp/vector/fixed/
http://www.monsoondata.org:9090/dods/gswp/grid/
http://www.monsoondata.org:9090/dods/gswp/grid/fixed/

The catalogs list the available variable files and links to several documentation files for each variable files, including a web-formatted plain-language description of the file and metadata called info, a DODS data descriptor structure file called dds, and a NetCDF-style data attribute structure file called das.

The European server is a conventional DODS server at IPSL.  The directory structure is the same as on the North American server, except that the multi-file nature of the time-varying fields is now clearly evident and requires you to access each month by a separate address.   Direct access to the data using DODS-enabled software is through the URL path:
http://dods.ipsl.jussieu.fr/cgi-bin/nph-dods//gswp/
Because the European server is a conventional DODS server, you can view the data files directly from the web.  Data sets may be downloaded using the Lynx browser, or using the wget command:
wget -r http://dods.ipsl.jussieu.fr/gswp/
However, only the vector (compressed by gathering) files are accessable using the conventional DODS server.  The repopulated grids are not accessible from the European site.

There are four directories for the GSWP-2 forcing data. The one-dimensional land-only data are in the vector directory, and the fully-populated two-dimensional global grids are in the grid directory (with water and Antarctic points set to the missing data value). The time-invariant fields are in two subdirectories called fixed. The files in the vector directory are smaller and will download more quickly. The NetCDF metadata, when combined with the file gswp/grid/fixed/Landmask, have all of the information necessary to navigate the compressed data vectors. So it is advisable to run the land surface models using this compressed data, rather than downloading the complete 2-D grids and masking out the missing data points. The 2-D grids are provided for quick visualization or analysis applications.  There is an additional directory called field which contains NetCDF formatted versions of several observational data sets, for validation or calibration purposes.

Below are listed the GSWP-2 input data fields by thier ALMA variable name, and the path to access each one on the DODS servers.

Fixed Fields

File: gswp/vector/ or gswp/grid/

SoilClass

fixed/soilclass

SoilDepth

fixed/Soildepth

Clay

fixed/clay

Sand

fixed/sand

Silt

fixed/silt

Organic

fixed/organic

Elevation

fixed/elevation

Slope

fixed/slope

W_fieldcap

fixed/w_fieldcap

W_wilt

fixed/w_wilt

W_sat

fixed/w_sat, fixed/w_sat_cea84

W_bpower

fixed/w_bpower_cea84

W_sat_hydc

fixed/w_sat_hydc, fixed/w_sat_hydc_cea84

W_sat_matp

fixed/w_sat_matp_cea84

Albedo_vi

fixed/albedo_vi

Albedo_ir

fixed/albedo_ir

Albedo_soil

fixed/albedo_soil

VegClass

fixed/vegclass_igbp, fixed/vegclass_sib, fixed/vegclass_bats

classFrac

fixed/classfrac

RootDepth

fixed/rootdepth_50, fixed/rootdepth_95

vegFrac

fixed/vegfrac_uk

Cti

fixed/cti

Grid only

File: gswp/grid/

Landmask

fixed/landmask

Monthly Fields

File: gswp/vector/ or gswp/grid/

LAI

lai_uk

grnFrac

grnfrac_uk

NDVI

ndvi_uk

FPAR

fpar_uk

Z0Surf

z0surf_uk

DisplH

displh_uk

Albedo

albedo_csu

3-hourly Fields

File: gswp/vector/ or gswp/grid/

Tair

tair_cru

Qair

qair_cru

Wind

wind_ncep

SWdown

swdown_srb

LWdown

lwdown_srb

Psurf

psurf_ecor

Rainf

rainf_gswp

Rainf_C

rainf_c_gswp

Snowf

snowf_gswp



FTP access to GSWP-2 data

Data are now being made available by anonymous FTP for those who cannot access any of the servers above because of firewall, compiler, or other issues. Access to the forcing data and fixed fields is at ftp://cola.gmu.edu/gswp/data/. Access to the Multi-model analysis is at: ftp://cola.gmu.edu/gswp/data/GSWP2_B0/. GrADS control files that unpack the NetCDF data files are at: ftp://cola.gmu.edu/gswp/ctl/ as is the necessary "pdef" file: ftp://cola.gmu.edu/gswp/ctl/gtd.filepdef.