Bill Hark
EF5
Interesting article. I think the newer computers will help but I'd rather see a higher density of ground and vertical observations to input into the computer models.
Bill Hark
Bill Hark
Andrew Ryan
EF3
That's not a weighted statement at all...
But in all seriousness, I'm looking forward to this. It's about time we got better data assimilation for the GFS. Not that it will correspond to better forecasts, but it should reduce the amount of apocalyptic solutions.
Jesse Risley
Staff member
While it seems plausible that we can make some improvements in the models themselves (i.e., integration of physical processes in the model, computations, etc.), I have to wonder how far we can go without really eliminating deficiencies in upper air and surface data over certain regions with an understandable lack of ingestion points.
rdale
EF5
I think you've hit the head... About 15 years ago, Greg Mann (now at WFO DTX) did a study that showed dramatic improvements in lake-effect forecasts from a temporary UPA station on the north side of Lake Superior. Adding such a station is probably much cheaper than a supercomputer, but nowhere near as sexy in a press release.
Chris Nuttall
EF5
More observation stations MAY be cheaper in the short-term, but not in the long-term. Keep in mind that all of that equipment requires maintenance, which costs money from labor and training to do the work. Automated stations don't require people to take obs, but upper-air stations do. When you add the labor costs to them, those stations become quite costly in a hurry.
I have seen some proposals for movable UA stations that use fairly cheap equipment, but you still have to factor in transportation and labor costs.
I have seen some proposals for movable UA stations that use fairly cheap equipment, but you still have to factor in transportation and labor costs.
rdale
EF5
I've not priced a supercomputer lately but I wonder how many years of observing would cover just one
Let's even look at wind profilers... The "test" study was a resounding success, and I struggle to believe that they are a major (or even medium) financial burden, yet they've tried to cancel it and have done nothing to expand it for 20 years.
Let's even look at wind profilers... The "test" study was a resounding success, and I struggle to believe that they are a major (or even medium) financial burden, yet they've tried to cancel it and have done nothing to expand it for 20 years.
Jesse Risley
Staff member
Yeah, I agree that you have to consider TCO (total cost of ownership) when investing in any infrastructure, whether it be computers or observation stations. I'm sure that increasing the number of observation stations will come at a cost, though I'm in no position to debate how considerable that cost might be.
However, over the long haul, even assuming it would involve more expense, as you stated, what's the most bang for the taxpayer's buck? If we're really interested in bettering weather forecasting and improving the numerical models, it seems to me that the wisest investment would be finding a way to overcome the data ingestion holes, even if their is potentially a 25%+ greater cost to doing so. At some point it might be a fiscally improbable course to pursue, but I would like to think that serious considerations about doing so has more than a passing merit.
Since we are talking about federal financing, I'm sure economic trade-offs in this economy are playing much more of a role than not.
Andrew Ryan
EF3
I would imagine that the TCO for improved observation network would be similar to the TCO of the supercomputer. Obviously the observation network would be a better investment, but you have to look at from the general public taxpayer point of view.
"Hey! NOAA just spent $27 million on a supercomputer that will spit out accurate forecasts."
OR
"Hey! NOAA just spent $27 million dollars on weather balloons. WTF?! I thought we didn't use those anymore!"
Granted, I don't think the general public would be thrilled with either option, but they would be more impressed with the supercomputer.
"Hey! NOAA just spent $27 million on a supercomputer that will spit out accurate forecasts."
OR
"Hey! NOAA just spent $27 million dollars on weather balloons. WTF?! I thought we didn't use those anymore!"
Granted, I don't think the general public would be thrilled with either option, but they would be more impressed with the supercomputer.
Similar threads
