Tuesday, 8 January 2013: 2:15 PM
Ballroom B (Austin Convention Center)
Land surface air temperature (SAT) is one of the most important observational variables in climate change research. For a variety of applications, the SAT diurnal cycle and day-to-day variation are also needed, but global historical long-term hourly data do not exist. The three-hourly SAT data have been developed by adjusting the six-hourly reanalysis SAT data with the in situ monthly data in several previous efforts. However, the diurnal cycle from reanalyses (even after the monthly mean bias correction) are still unrealistic. We have developed a global hourly 0.5-degree SAT dataset based on four reanalyses (i.e., Merra, ERA-40, ERA-Interim, and NCEP/NCAR-1) and CRU TS3.1 data from 1948-2009. Our data development makes use of two new ideas, including: a) the temporal downscaling of reanalyses from their native resolution to hourly using the MERRA hourly data; and b) the adjustment of daily maximum and minimum temperature data from reanslyses using different approaches. Through our procedure, the adjusted hourly SAT data from each reanalysis have exactly the same monthly Tmax and Tmin as the CRU TS3.1, and hence are a much better representation of the diurnal cycle. In this presentation, we will evaluate reanalysis SAT data using the CRU TS3.1 and other in situ data, provide the details of our data development method, and analyze the new global hourly 0.5-deg SAT, such as the differences between 24-hour average versus (Tmax +Tmin)/2, and the trend differences in monthly mean Tmax and Tmin from CRU versus in the maximum and minimum values of monthly averaged diurnal cycle from our new product.
- Indicates paper has been withdrawn from meeting
- Indicates an Award Winner