High-resolution temporal and spatial resolution data--usually created with regional-scale models—have traditionally been proprietary and costly to obtain. Freely available global model data suffers from either lower spatial or temporal resolution, or both. Low spatial resolution fails to realistically represent wind speeds in complex terrain. Low temporal resolution fails to capture the full diurnal cycle of wind behavior.
The NCAR Global Climate Four-Dimensional Data Assimilation (CFDDA) Hourly 40 km Reanalysis was developed in 2009-2010 by the Research Applications Laboratory (RAL) to provide the most accurate boundary layer wind estimates available at that time. CFDDA used 28 sigma levels, with 19 between the surface and 700 hPa, a four-fold improvement over the contemporary NWP models. The dataset spans 21 years, 1985-2005, providing hourly atmospheric parameters, including winds, on 28 vertical levels on a global 40 km grid.
Unlike most global reanalyses, CFDDA output is available hourly to fully capture the diurnal range of weather phenomena. For instance, if the land-sea breeze or mountain-valley circulation in a region peaks at around 14:00 local time, and that does not coincide with the 00, 06, 12, or 18 UTC times of the saved reanalysis, then the strength of the phenomena will be underestimated in that region.
The fidelity extends beyond wind speeds. CFDDA has been rigorously validated for rainfall and other metrics. Please read Verification and Validation of the global CFDDA Dataset for the full details.
CFDDA project partners, Defense Threat Reduction Agency (DTRA), NCAR RAL and NCAR Mesoscale Microscale Meteorology (MMM) divisions are offering this dataset to the public for free with minor restrictions. NCAR Research Data Archive (RDA), hosted by the Computational and Information Systems Laboratory, provides data support. It is available at http://rda.ucar.edu/datasets/ds604.0/.
MethodologyNCAR’s Real Time Four Dimensional Data Assimilation (RTFDDA) and PSU/NCAR mesoscale model (MM5 version 3.6) were used on two 40-km horizontal polar stereographic grids and sigma-following vertical coordinates. Surface, upper air and satellite measurements were continuously assimilated with observation nudging. Domain “seams” were joined with Cressman-type interpolation.
Links:CFDDA ds604.0 Main Page
AGU Fall Meeting 2014 Poster announcing the availability of CFDDA
CFDDA Data Curation poster
CFDDA as a case study for Data Attribution poster
List of CFDDA Documentation