26 August 2015

Windsday

The NCAR Mesa Lab has a reputation as a particularly windy site. This video taken from my office window shows why.


The ML real-time measurements, courtesy of NCAR's Earth Observing Lab, quantify just how gusty it is outside.


19 August 2015

Setting up to work with GRIB2

We are still receiving requests for translation of data files from GRIB2 to GRIB1.  Archiving the same data in two different formats takes up disk space that could otherwise be used to provide other useful data.  That is why our policy is to offer data in its native format.

18 August 2015

Interoperable Access to NCAR Research Data Archive Collections


To enhance user experience and utility, the RDA now offers THREDDS Data Server (TDS) access for many highly valued gridded dataset collections (http://rda.ucar.edu/thredds).  TDS offered datasets with native formats in GRIB1 and GRIB2 are presented as aggregations,  enabling users to access an entire dataset collection that can be comprised of 1000’s of files, through a single virtual file.  Individual files can also be accessed from all GRIB1, GRIB2, and NetCDF format dataset collections found on the TDS. 

06 August 2015

Cite your data

Science should be reproducible.  Citing your data as clearly and unambiguously as possible is as important a part of science communication as explaining your methodology.

It's not enough to say that you used FNL or GFS or CALIPSO data.  The software that produced FNL and GFS analyses and forecasts evolve over time.  New parameters or methods are introduced.  Bugs are fixed.   Satellite processing algorithms similarly evolve.

These changes are often not well documented.   For someone else to understand (and possibly reproduce) your work, they need to know exactly which version of the data you used for your work.

03 August 2015

New Terms and Conditions of Use for Japanese 55-year Reanalysis (JRA-55) Products

As of July 31, 2015, the JMA has relaxed its Terms and Conditions of Use for Japanese 55-year Reanalysis (JRA-55) Products. The main change is that "JMA provides the datasets with no imposition of restrictions regarding the purpose of use and/or redistribution", especially for commercial purposes. This impacts RDA datasets ds628.0, ds628.1, ds628.2, ds628.3 and ds628.4 (JRA-55, JRA-55C, JRA-55AMIP). The new conditions are as follows:

Accessing RDA collections from Yellowstone

The CISL User Services Section will hold a 60- to 75-minute seminar and discussion for Yellowstone users at 10 a.m. Friday, August 7, in the Small Seminar Room at the Foothills Lab in Boulder (FL2-1001). The session will also be webcast via UCARConnect.

Three brief presentations are planned:

Accessing RDA collections from Yellowstone: Grace Peng (CISL/DSS)—A large subset of NCAR/CISL Research Data Archive (RDA) dataset collections can be accessed directly from the GLADE file spaces on Yellowstone. This presentation will provide an overview of the RDA resources and tools, including metadata access and subsetting services, that are available to Yellowstone users.

Simple “command file” syntax for task parallelism: Dick Valent (CISL/USS/CSG)—By using command file syntax to exploit characteristics of your workflow, you may be able to achieve task parallelism to make more efficient use of your Yellowstone allocation and reduce time to solution. Command file syntax can be used in forms from simple to complex. Dick will present the simplest form of command file.

Real-time, ensemble forecasting using CISL resources: Ryan Sobash (MMM)—A daily, real-time, ensemble forecasting system has been running at NCAR since late March using Yellowstone and other CISL resources. Ryan will talk about some unique computing challenges this type of system poses, solutions developed to maintain reliability and timeliness, and the development of an ensemble visualization system and website (http://ensemble.ucar.edu) to view the forecasts.

All Yellowstone users are welcome. If you plan to attend in person, please sign up here to give us an idea of how many participants to expect.

#HPC #bigdata