WHAT'S NEW: See 'HISTORY' below for most recent update information. NOTE 1: Frequently-Asked-Questions.txt provides more info. NOTE 2: See http://www.ncdc.noaa.gov/oa/rsad/blendedseawinds.html for more on data access methods, registration to receive notices on product updates, etc. SUMMARY: Blended vector sea surface winds (at 10-m above sea level) and surface windstresses (stresses felt by the water). The wind speeds are blended from multiple satellites [up to six, including Scatterameters (QuikSCAT), SSMIs, TMI and AMSR-E] observations, on a global 0.25-degree grid and time resolution of 6-hourly, as four 'instantaneous' global snapshots per day at UTC/GMT 00, 06, 12 and 18Z. The blended speeds are then decomposed into (u,v) components using the NCEP Re-analysis 2 (NRA-2) wind directions interpolated onto the blended speed grid. The 6-hourly (u,v) instantaneous fields are provided at our data site, and the speed itself can be simply computed as w=(u^2+v^2)^0.5, which is exactly the blended speed from the multiple satellites. Daily fields are also provided as simple avearges in time of the 6-hourly data, and monthly fields are provided as averages of the daily fields. Because the vector means do not exactly equal to the scalar/speed means, the scalar means are thus also provided in the daily and monthly datasets, as (u,v,w). The scalar/speed means correspond to the means of the blended speed from mutiple satellite speed observations, and may be useful for wind energy indistries as an example. The vector means can be computed as the magnitude of the mean u and v, and may be useful for oceanographers as an example. Computations for the wind stresses follow the same methodology. Particularly in this version, the 6-hourly satellite-scaled global 0.25-degree grid wind stresses are computed as: taux_s = -[(w_s/w_m)**2]*taux_m tauy_s = -[(w_s/w_m)**2]*tauy_m where '_s' indicates satellite-scaled values and '_m' indicates NRA-2 model values interpolated to the satellite grid. Wind speeds are at the time stamps and taux/y are 6-hour averages starting from the time stamps. The satellite wind speed blending scheme is a Gaussian like 3-D (in x,y,t) interpolation. To reduce the data gaps for the early years (when fewer satellite were available) and to reduce aliases for the instantaneous fields for the later years (when more satellites are availabe), the time window and spatial radiaus are chosen as +/- 6 hours and 62.5 km, respectively. Sampling feasibility study can be found at http://www.agu.org/pubs/crossref/2006/2006GL027086.shtml Products are available from 09 July 1987 - present. The 6-hourly 4 snapshots per day are at UTC 00, 06, 12 and 18Z. The center time of the daily averages is thus UTC 09Z. The grid cells are: lon=0.0:0.25:359.75 (im=1440) lat=-89.75:0.25:89.75 (jm=719) Note: If you have trouble getting the data, the following scripts may help: By lftp - see ftp://eclipse.ncdc.noaa.gov/raid1c/seawinds/SI/my_lftp.sh By wget (with constraints for subsetting) - see ftp://eclipse.ncdc.noaa.gov/raid1c/seawinds/SI/my_wget.sh ******************************** DATA DIRECTORY STRUCTURE: ./seawinds/SI/ contains topographic data & reading program (<0 below sea level). Subdirectories ./uv/ and ./stress/ under ./seawinds/SI/ contains the vector winds and windstress, respectively: --------------- ./uv/ contains the vector winds, with sub-directories for different time resolutions of ./uv/6hrly/ [6-hourly, 4 global snapshots (u,v) at UTC 00, 06, 12 and 18Z] ./uv/daily/ [averages of the above 4 time points, thus with a center time 09Z; also provided is the scalar mean, (u,v,w)] ./uv/monthly/ [averages of the above daily data; also provided is the scalar mean, (u,v,w)] ./uv/clm/ [climatological monthlies; also provided is the scalar mean, (u,v,w)] For each of the time resolutions, there are two subdirectories that contains data in netCDF and IEEE - big endian, such as: ./uv/6hrly/netcdf/ ./uv/6hrly/ieee/ (same for the ./daily/ and ./monthly/ datasets) Then under each data format-subdirectory, the data are organized by decades for the 6-hourly and daily data, such as ./uv/6hrly/netcdf/1980s/ ./uv/6hrly/netcdf/1990s/ ./uv/6hrly/netcdf/2000s/ (same for the ./ieee/ data) The data file names are in the formats of uvYYYYMMDD.nc (or .be) for the 6-hourly and daily data, and uvYYYYMM.nc (or .be) for the monthly data. Support reading routines are included in the time-resolution subdirectories [read*.f (FORTRAN), read*.m (MATLAB), W*.gs (GrADS), etc], with verification data in text (verify_SI_uv_6hrly.txt, etc); these files are self-explaining. ./stress/ has the same sub-dir structures as ./uv/. **************************** Please Note: For better sevice to data users, please register to our user list (email to huai-min.zhang@noaa.gov) so that we can update you on new development. We also ask you to acknowledge us in your use of the data to help us justify continued service. This may be done by including text such as: The wind data are acquired from NOAA's National Climatic Data Center, via their website http://www.ncdc.noaa.gov/oa/rsad/blendedseawinds.html. We would also appreciate receiving a copy of the relevant publication. These are research products and thus are considered as experimental in nature. ***************************** CONTACTS for seawinds: Huai-Min.Zhang@noaa.gov (e-mail) Phone: (+1) 828-271-4090 or ncdc.satorder@noaa.gov Phone: (+1) 828-271-4800 ***************************** Updated 3 July 2007 ***************************** HISTORY - reverse chronically: Version 1.2 (July 2007): The wind speeds/magnitudes are the same as Version 1.1 (blended only from satllite data). However, wind directions are shifted by 6-hours in time in the NRA-2 renalysis fields, thus affects the decompositions of wind speeds into the u-v components. It was found that the 'time stamps' in the NRA-2 output files are not for the fields themselves, but they are the forecast starting time, or as the 'reference times'. The output fields themselves are 6-hour forecasts valid for the times 6-hours after the 'reference times' for winds etc and 6-hour averages for flux fields. These are taken into account in this new version (Version 1.2) of the blended fields, and are valid as 'instantaneous' snapshots at the time 'stamps' imbedded in the netCDF files for the wind speeds and u-v components. Version 1.1 (April 2007): 1) This version used the individual satellite retrievals from the Remote Sensing System, Inc (http://www.remss.com/) that were released in September 2006. 2) Vector winds and windstresses are provided in this version, with the wind speed from blending the multiple satellite observations and wind direction from the NCEP Reanalysis-2 interpolated to the blended satellite speed grid. 3) To save storage space, the 6-hourly (4 instantaneous global snapshots per day) data only contain the (u,v) components, since the speed/magnitude can be computed as sqrt(u^2+v^2) for the instantaneous field, which is exactly the blended speed from the multiple satellites. However, for the (daiyly/monthly) mean speed, there are commonly two ways to compute the means, as vector mean or scalar mean, and they generally do not equal to each other. For comparison purpose, the scalar mean speed (w) is also provided together with means of the (u,v) components in the daily and monthly mean fields [i.e. as (u,v,w) or (taux,tauy,tau)]. The scalar means correspond to the means of blended satellite wind speed. Version 1.0 (August 2006): This version used the individual satellite retrievals from the Remote Sensing System, Inc (http://www.remss.com/) before their major updates in September 2006. Also, only wind speed data were provided for the 6-hourly, daily and monthly resolutions. This version was moved offsite and can be obtained by the above 'CONTACTS'. (Note versioned - 12-hourly or lower resolution, 2005)