Thursday, May 16, 2013

News Bits

I thought I'd take a moment to report three items of interest making news this past week.  

1. Quantum Leap.  First, as discussed at the Capital Weather Gang, a major infusion of computer power is in the works for the National Centers for Environmental Prediction (NCEP), which is receiving support from the Disaster Relief Appropriations Act of 2013 (a.k.a., the Sandy Supplemental) to increase their computer capacity from 213 teraflops to 2,600 teraflops by the 2015 fiscal year.  The Capital Weather Gang called it a game changer, so I'll call it a quantum leap.  No matter what it is called, it is a badly needed infusion of computer power that will hopefully be leveraged to produce a significant advance in weather prediction capabilities for the nature. 


2. Strategic Planning.  Curiously, as NCEP gets this infusion of computer power, this week also saw the release of Forecast for the Future: Assuring the Capacity of the National Weather Service, a report from the National Academy of Public Administration.  I gave it a quick skim, however, and didn't see anything too earth shattering.  Given that the NWS is still living off the vapors of the Modernization effort of the 1990s, lets go after a quantum leap not just in terms of NCEP computing but across the agency.  Such an investment would not only benefit the public, but also a growing private-sector weather enterprise.  

3. 97% again.  As reported in the Guardian, another study has come out showing that ~97% of climate scientists/climate papers/climate abstracts conclude that climate change is caused by human activity.  More specifically, the latest study by Cook et al. (2013) finds that "among abstracts expressing a position on [anthropogenic global warming], 97.1% endorsed the consensus position that humans are causing global warming."  I find it remarkable that 97% keeps coming up in these surveys.  There must be a conspiracy...


  1. According to the article, that still leaves 83 papers for the conspiracy theorists to grasp onto . . . and you just know we'll be seeing those folks on Fox news.

  2. Interesting post - the compute capability upgrades especially interest me. My days of running Climate modeling studies were exclusively on vector supercomputers of the 80's and 90's. Have Climate and Weather models now been adapted to run on linux compute clusters? I know that Cray is still around. But, I'm very curious about the amount of effort that has gone into adapting the models to take advantage of "grand challenge" type hardware platforms that we find weapons codes running on in the national laboratories.

    1. Nearly all weather and climate models today can and are run run on linux clusters. The NCEP operational systems are still running on an IBM Power6 (probably running AIX?), but I believe their forthcoming system is a linux cluster.

      An iPad has about the same computer power as the Cray Y-MP I did my first modeling on in the late 80s and early 90s.

    2. My first "supercomputer" was a CDC Cyber 205. It had 1 million words (8 MBytes) of memory and its footprint filled a large room. It's amazing that my laptop has 4 GBytes RAM and fits on my lap. :)