Atmospheric Soundings and Numerical Models

[baroclinic_instability]

Over the last month:

Over the last eight years:

Satellite derived data assimilation has improved so much over the last 5-10 years that the radiosonde network is far less important nowadays. If wxbadger made it over here, he could explain it a lot better.

Posted Today, 07:36 AM

Over the last month:

Over the last eight years:

Satellite derived data assimilation has improved so much over the last 5-10 years that the radiosonde network is far less important nowadays. If wxbadger made it over here, he could explain it a lot better.

[baroclinic_instability]

Posted by am19psuam19psu, on 18 November 2010 - 07:36 AM, said:

Over the last month:

Over the last eight years:

Satellite derived data assimilation has improved so much over the last 5-10 years that the radiosonde network is far less important nowadays. If wxbadger made it over here, he could explain it a lot better.

I believe that's basically the crux of the matter, at least at this point. I remember reading a paper of the initialization of the EC, but it referenced the rest of the models as well, satellite derived data vastly, vastly dwarfs the amount of data that's collected via land based and balloon/aircraft observations. I remember reading something on the order of millions upon millions of observations in excess of what could possibly be ingested into a model run. The paper was essentially on the issues surrounding model initialization compute time, which for the EMC back when the paper was written (circa 2007 or so) was nearly up to the amount of time required to actually run the model.

[baroclinic_instability]

Thanks for the information. This begs the question then....will they ever end atmospheric soundings for use in numerical models?

[baroclinic_instability]

From Ed Mahmoud, previous thread.

The twice a day soundings from a limited number of stations? The SPC will request 18Z soundings when severe weather is expected, there must be some benefit to that. Local NWS office complains sometimes about being between the CRP and LCH soundings in their discos, although they do get some ACARS data.

I get to see actual balloon soundings on the interwebs a couple hours after the fact, I sometimes like to compare the 0 hour NAM and GFS skew-Ts from the NIU make your own sounding site to the actual soundings when playing amateur severe weather prediction dude, to see which model best fits the soundings in nearby states, and ACARS isn't on the internet.

[Ed Lizard]

Cut-n-paste from other thread...

Deleted, someone cut-n-pasted first

[baroclinic_instability]

Cut-n-paste from other thread...

Deleted, someone cut-n-pasted first

I guess what I meant to say was the regular large scale soundings across the country for use in numerical guidance. If the data here are correct, it seems a waste to keep sending up soundings unless warranted under certain circumstances like severe wx, for instance.

[Ed Lizard]

How does the HPC, in their model diagnostic discussions, know, for example, the NAM initialized the 500 mb low over Utah 5 DM too strong? Just pulled that particular one out of the air, but there is an HPC discussion. They will discuss initialization errors, and I assume some of that knowledge comes from actual soundings.

I could be wrong about that, of course.

[baroclinic_instability]

How does the HPC, in their model diagnostic discussions, know, for example, the NAM initialized the 500 mb low over Utah 5 DM too strong? Just pulled that particular one out of the air, but there is an HPC discussion. They will discuss initialization errors, and I assume some of that knowledge comes from actual soundings.

I could be wrong about that, of course.

I understand what you are saying...but if the satellite derived data is this good, they could just plot the satellite derived soundings instead.

[isohume]

Satellite derived data assimilation has improved so much over the last 5-10 years that the radiosonde network is far less important nowadays. If wxbadger made it over here, he could explain it a lot better.

GOES soundings are decent as long as there is no cloud cover. NESDIS uses them a lot to account for changing PWs close to and associated with high precip events. However, the RAOB network can't be replaced insofar as measuring data within clouds...which is normally where the active weather events are located and the most dynamic. The RAOB network will be quite important for many years to come.

[baroclinic_instability]

GOES soundings are decent as long as there is no cloud cover. NESDIS uses them a lot to account for changing PWs close to and associated with high precip events. However, the RAOB network can't be replaced insofar as measuring data within clouds...which is normally where the active weather events are located and the most dynamic. The RAOB network will be quite important for many years to come.

Yes I agree, I was mostly posting my responses as questions...in that I doubt the validity of the data since I can attest to how poor the models still perform over oceans. I don't think it is coincidence models perform better with weather systems after they are observed by the sounding network. I am not sure how the data in the report was developed, but I just find it difficult to believe a numerical weather model would be able to perform equally as well without observational data such as balloon soundings.

[isohume]

Yes I agree, I was mostly posting my responses as questions...in that I doubt the validity of the data since I can attest to how poor the models still perform over oceans. I don't think it is coincidence models perform better with weather systems after they are observed by the sounding network. I am not sure how the data in the report was developed, but I just find it difficult to believe a numerical weather model would be able to perform equally as well without observational data such as balloon soundings.

I must have missed the question marks. lol

If the all the GOES sounding data was used as input to the models, the taxation on computing power (time) would render the output operationally useless. That kind of power is just not available presently. The RAOB network will take some time, if ever in our careers, to replace effectively.

[jntkwx]

Dendrite (Brian) and I wrote a paper for a Technical Communication class we were in (see attachment). We were looking at the feasibility of increasing the current RAOB spatial and temporal resolution. The overwhelming consensus (we even asked all those with red tags on what used to be Eastern), was that while an increase in both spatial and temporal resolution would be ideal, it is not a financially viable option, particularly in this economy.

We also found it interesting how many (if not all) meteorologists regard the 12Z and 00Z model runs with a greater degree of trust and/or accuracy (due to them obviously being initialized with RAOB data), due to the mere belief that forecasters DO tend to find it difficult to believe a numerical weather model could perform equally as well without observational data.

A_Report_on_Radiosonde_Temporal_and_Spacial_Resolution.pdf

[baroclinic_instability]

I must have missed the question marks. lol

If the all the GOES sounding data was used as input to the models, the taxation on computing power (time) would render the output operationally useless. That kind of power is just not available presently. The RAOB network will take some time, if ever in our careers, to replace effectively.

Haha yeah, questions posed using retorts I should say.

But yeah, I have not read the document in question (the graphs above), but it would be interesting to read what context it was in.

[kpantz]

Been out of the loop for some time - how does ACARS factor in these days? I'd have to think it's a good contribution in areas GOES can't really help (as long as the planes are in the air, obviously). Last info I read on it implied the coverage is pretty good, both spatially and temporally. Is it more a matter of just not having as many variables measured?

[isohume]

Been out of the loop for some time - how does ACARS factor in these days? I'd have to think it's a good contribution in areas GOES can't really help (as long as the planes are in the air, obviously). Last info I read on it implied the coverage is pretty good, both spatially and temporally. Is it more a matter of just not having as many variables measured?

Yeah, that's a big disadvantage to ACARS. Most ACARS report temp and wind...however very few report dewpoint. Td readings are being increased tho. Also, data is not normally available above 300mb and the overall volume is reduced around large storms. However, it's very good data and plenty of it makes it into model initializations. It's cheap too, about 10 cents a sounding.

[kpantz]

Yeah, that's a big disadvantage to ACARS. Most ACARS report temp and wind...however very few report dewpoint. Td readings are being increased tho. Also, data is not normally available above 300mb and the overall volume is reduced around large storms. However, it's very good data and plenty of it makes it into model initializations. It's cheap too, about 10 cents a sounding.

If Reed Timmer were a commercial pilot, he'd mount probe launchers on an A320 to close that gap.

[baroclinic_instability]

If Reed Timmer were a commercial pilot, he'd mount probe launchers on an A320 to close that gap.

Ugh, that name should be banned around here.

[isohume]

If Reed Timmer were a commercial pilot, he'd mount probe launchers on an A320 to close that gap.

I'm not sure who that is...but more power to him. lol

NCEP uses ACARS to help verify model intitializations, but mostly the ACARS data is used for near term forecasting. San Diego uses them to get a handle on the depth of the marine layer, helping to fcst inland cigs and vsby. They're also a good real time tool to help predict inversion breaks and gust potentials for svr wx. Many other fcst applications as well.

[kpantz]

Ugh, that name should be banned around here.

LOL. I couldn't help but mentally replace his name with Chuck Norris as I wrote what I did.

[baroclinic_instability]

I'm not sure who that is...but more power to him. lol

NCEP uses ACARS to help verify model intitializations, but mostly the ACARS data is used for near term forecasting. San Diego uses them to get a handle on the depth of the marine layer, helping to fcst inland cigs and vsby. They're also a good real time tool to help predict inversion breaks and gust potentials for svr wx. Many other fcst applications as well.

He is on "Storm Chasers" on Discovery, making a mockery of meteorology.

[isohume]

LOL. I couldn't help but mentally replace his name with Chuck Norris as I wrote what I did.

I googled his name and got this in a wiki...

"Timmer has been criticized by the research meteorological community for not conducting research into tornadoes, but rather seeking only fame through overly-dangerous encounters to be broadcast on TV and to sell his recent book"

[baroclinic_instability]

I googled his name and got this in a wiki...

"Timmer has been criticized by the research meteorological community for not conducting research into tornadoes, but rather seeking only fame through overly-dangerous encounters to be broadcast on TV and to sell his recent book"

Haha, pretty much sums it up. He is one of the guys who chases and makes money off Discovery Channel but then complains that too many people chase when they are trying to conduct research Well DUH. That is their fault for turning storm chasing into a national phenomena and inviting agencies such as the Weather Channel along for the ride, broadcasting for all to see.

[baroclinic_instability]

LOL. I couldn't help but mentally replace his name with Chuck Norris as I wrote what I did.

Haha nice. We need to unleash Chuck on him.

[kpantz]

Haha, pretty much sums it up. He is one of the guys who chases and makes money off Discovery Channel but then complains during V2 that too many people chase when they are trying to conduct research Well DUH. That is their fault for turning storm chasing into a national phenomena and inviting agencies such as the Weather Channel along for the ride, broadcasting for all to see.

I know I took this off topic, and I don't really want to keep it off topic, but Timmer isn't part of Vortex2. Different team/business.

OK, the ex-moderator in me says we can bring this back on topic.

So...has anyone ever retrieved a radiosonde?

[baroclinic_instability]

I know I took this off topic, and I don't really want to keep it off topic, but Timmer isn't part of Vortex2. Different team/business.

OK, the ex-moderator in me says we can bring this back on topic.

So...has anyone ever retrieved a radiosonde?

Ok good point, but there are others on VORTEX2 part of that show who complained about that exact same thing, including Wurman, if CD3 is to be believed, which I do believe. Ok sorry back on topic.

I have not retrieved a radiosonde

[baroclinic_instability]

Dendrite (Brian) and I wrote a paper for a Technical Communication class we were in (see attachment). We were looking at the feasibility of increasing the current RAOB spatial and temporal resolution. The overwhelming consensus (we even asked all those with red tags on what used to be Eastern), was that while an increase in both spatial and temporal resolution would be ideal, it is not a financially viable option, particularly in this economy.

We also found it interesting how many (if not all) meteorologists regard the 12Z and 00Z model runs with a greater degree of trust and/or accuracy (due to them obviously being initialized with RAOB data), due to the mere belief that forecasters DO tend to find it difficult to believe a numerical weather model could perform equally as well without observational data.

I actually just saw this, interesting information. I would be interested in seeing potential future studies on how a spatial increase (i.e. more RAOB sites, maybe say double the coverage) would effect numerical model performance as well as an increased temporal resolution (keep the existing sites but double the times from 0Z/12Z to 0,6,12,18Z). That is, would an increased temporal resolution be more effective than an increase in the spatial coverage, or vice versa.

[usedtobe]

As I understand it, satellite thermal and moisture profiles are sort of lookin a slices of the atmosphere as opposed to sounding which give the the change the temperature and moisture as the balloon rises and drifts away. Therefore, if you are interested in stability you get a better picture from the soundings. That's why I doubt in my lifetime, they will get rid sounding altogether.

[Ed Lizard]

As I understand it, satellite thermal and moisture profiles are sort of lookin a slices of the atmosphere as opposed to sounding which give the the change the temperature and moisture as the balloon rises and drifts away. Therefore, if you are interested in stability you get a better picture from the soundings. That's why I doubt in my lifetime, they will get rid sounding altogether.

Isn't there a graph of the GFS prediction for Hurricane Katrina run with and without G-IV soundings, same time and otherwise same inputs, and how much better the forecast was with high density soundings in the area?

[usedtobe]

Isn't there a graph of the GFS prediction for Hurricane Katrina run with and without G-IV soundings, same time and otherwise same inputs, and how much better the forecast was with high density soundings in the area?

I don't know.

[Ed Lizard]

I don't know.

Interweb search.....

http://celebrating200years.noaa.gov/transformations/aircraft/#enter