Addendum (02 December 2005): Some figures are shown in another page [link].
For Global Soil Wetness Project, Phase 2 (GSWP2; Dirmeyer et al. 2002), a new set of "standard" forcing data called "B1" is being prepared, and the version called "B1 beta" (documented on 23 Aug. 2005 at http://www.jamstec.go.jp/frcgc/research/p2/masuda/gswp/b1beta/b1beta.html) has been produced. Because of limitation of available 3-hourly ERA40 data, "B1 beta" did not included the files for the spin-up period (July 1982 to December 1985) except for shortwave and longwave radiation. To fill the gap, a set of forcing data is produced based on the essentially same input as for B0. It also covers part of the period overlapping with "B1 beta". (In the current release, we include the first year of the production period, January to December 1986, mainly for the purpose of comparison). Therefore we do not call this version "B1 beta", but tentatively "B0x". (If it is confusing we will rename it.) We tried to make "B0x" having similar characteristics as "B1 beta", but obviously the similarity is not perfect. In particular, one of the problem reported about B0, occasionally excessive variability of temperature, still appears in "B0x" though the occurrence has probably become rarer.
These GrADS CTL files require a binary data file "gtd.filepdef" which is the same one for the forcing data files available from the DODS server at COLA.
In our storage media, these files are first compressed by "gzip" and then archived by "tar". (Thus "*.tar" files contain many "*.nc.gz" files.)
According to comparison with synoptic observations (reported by Kenji Tanaka and Kazuaki Yorozu), ERA40 tends to underestimate wind speed at 10 m while NCEP R2 tends to overestimate it. Nevertheless in "B1 beta" we adopted ERA40 wind speed without adjusting it to match with observations, mainly because of difficulty to assess detailed local conditions. To spin-up models for "B1 beta" run, we consider that the wind speed should have magnitude comparable to Wind_b1beta. Thus we made a "hybrid" wind speed data Wind_b0x, based on NCEP R2 (actually the forcing data Wind_ncep which was part of B0). We do not claim that this is certainly nearer to reality than B0.
Monthly 3-hourly avarages at each grid box of both ERA40 and NCEP R2 is calculated. The reference period is from January 1986 to October 1985. The ratio, Wind_era / Wind_ncep, is used as the correction factor to Wind_ncep to produce Wind_b0x. In order to avoid excessive correction, values less than 0.1 m/s (in either data sets) are replaced with 0.1 m/s before calculating the ratio.
The procedure and the input data are essentially the same as B0 (Zhao and Dirmeyer, 2003, Section 2.3). As in "B1 beta", correction of variability based on the diurnal range is avoided. Thus, the temperature data of "B0x" (Tair_b0x) is NCEP R2 shifted by the difference of monthly values between elevation-adjusted CRU and NCEP R2. Thus, the variability within months contained in NCEP R2 is retained in "B0x".
Warning:
The problem found in "B0" temperature data
about excessive temperature variability
at some locations (reported by Kenji Tanaka
and included on 13 November 2003 in
http://www.iges.org/gswp/faq.html)
is also sometimes found in "B0x" as well,
though probably less often.
In the first case of Tanaka's CORECTT.pdf,
52.5oW, 65.5oN in Greenland,
Tair_b0x has stronger "spikes" than B0 (Tair_cru).
In other five cases including another location in Greenland,
the variability of Tair_b0x is smaller than B0
and seems reasonable, though not always similar to "B1 beta".
We release "B0x" data without modifying values of problematic cases,
though we are going to check the occurrences and report them.
Our production (Psurf_b1beta and Qair_b1beta) is the same as in B0 (Zhao and Dirmeyer, 2003, Sections 2.4 and 2.5) except that it uses our version of temperature Tair_b0x.
We started from the sum of Rainf_gpcc and Snowf_gpcc which were prepared for sensitivity experiments P3 based on B0. Those are NCEP R2 forecast precipitation rate scaled to match with CRU (up to 1985) or GPCC (from 1986 on).
Separation of rain and snow is made similarly to "B1 beta" by the formula of Yamazaki (2001), based on the "B0x" temperature, humidity and pressure data.
Correction of raingauge undercach is made similarly to "B1 beta" by Motoya's formula (which was utilized as part of B0 production as well). The wind speed in the formula should be the value at 2 m above ground, and it is approximated from the 10 m wind speed Wind_b0x multiplied by 0.7. based on the assumption of a logarithmic profile
As in "B1 beta", merging with GPCP data is not done.
Convective rain rate (Rainf_C_b0x) is calculated as Rainf_b0x * Rainf_C_ncep / Rainf_ncep . No attempts have been made to adjust it to match ERA.