Talk:Numerical weather prediction

Possible merger

The information from this page needs to be merged with another recently created Numerical Weather Prediction page

Numerical Weather Prediction

Done. nick 15:42, 11 May 2005 (UTC)

Mayhap the technique of inserting 'bogus' data' to fill in the gaps left by observations should be briefly mentioned...although I'm not really sure if it's relevant enough to be inserted...anywho...Dr Denim 12:38, 31 May 2006 (UTC)

Eta vs NAM

Please revert back to the 10 Jan 06 version. The North American Mesoscale (NAM) run at the National Centers for Environmental Prediction (NCEP) is not a model, it's a placeholder. Currently the model designated as the NAM is the Eta model. Some time in the second quarter of CY06 NCEP will begin running the NMM-WRF as the NAM.

Incidentally, the Global Forecast System (GFS) is also a placeholder. Currently the AVN model is designated as the GFS, but if NCEP ever replaces the AVN model with a global version of WRF, or any other model, it will still be designated as the GFS.

If the article is ever to go to GAC...

...it needs inline citations. Thegreatdr (talk) 19:18, 13 January 2008 (UTC)

Global models (list)

Both this article and the "Atmospheric model" article contain a section "Global models" which is a list of models. Theses are similar but look to have diverged a little. Not nice. So I suggest that this list be maintained in just one article (not both), and that the other article contains a cross-reference (e.g. wikilink) to that first article. If that is the right way to go, my guess is that this (NWP) article might be the better home, as some models are (presumably) more than just atmospheric. Feline Hymnic (talk) 23:30, 1 January 2009 (UTC)

If there is no difference between the Atmospheric model article and this article, the better idea would be to suggest their merger, making Atmospheric model a redirect to this article. That would require a vote amongst the editors, and some time. Thegreatdr (talk) 23:35, 1 January 2009 (UTC)

But aren't they different (although overlapping) topics? "Atmospheric model" is, presumably, just about the atmosphere, whereas NWP is more general and would include other coupled systems (e.g. oceans), wouldn't it? So wouldn't it be better to keep them separate (although well-defined, and well-focussed on their respective primary topics)? Feline Hymnic (talk) 23:53, 1 January 2009 (UTC)

That would certainly be easier. If kept separate, the list of forecast models could disappear from this article. Thegreatdr (talk) 00:00, 2 January 2009 (UTC)

Progress so far

We're down to two sections which remain reference-less, and a few other passages which need refs. Some external links likely need to be weeded out. Other than a little more content, significant progress has been made improving this article during this past week, and I can see GAN on the horizon. Thegreatdr (talk) 18:49, 1 January 2011 (UTC)

For the initialization section, do we need to write about individual objective analysis techniques, such as successive correction, optimal interpolation, or the 3D/4D variational assimilation methods? Titoxd^{(?!? - cool stuff)} 20:50, 2 January 2011 (UTC)

For FA, maybe. For GA, probably not. Really, all that stuff should be gone into detail within the article about objective analysis techniques (if wikipedia has one). Thegreatdr (talk) 20:56, 2 January 2011 (UTC)

Interestingly, wikipedia does not have an objective analysis article. There is definitely room for one as a subarticle to this article, as well as any article while requires mention of providing initializations for computer programs. Barnes interpolation is the closest article I could find which addresses this issue. Thegreatdr (talk) 23:52, 2 January 2011 (UTC)

For the citation looked for in the lead, we already have a reference for it below (Cox). Did you want another ref placed up there as well? Thegreatdr (talk) 20:58, 2 January 2011 (UTC)

Oh, I didn't see that one. I don't think that an additional citation in the lede is necessary. Titoxd^{(?!? - cool stuff)} 21:23, 2 January 2011 (UTC)

Referencing for the current content of this article is finally done (I think). Whenever we are comfortable with its content, we should be able to GAN it. Thegreatdr (talk) 01:10, 3 January 2011 (UTC)

I'm happy with it. One question, though: How much history of climate models should we mention here? We have a Global climate model, and I'm not sure how much should go here instead of there. Titoxd^{(?!? - cool stuff)} 02:13, 3 January 2011 (UTC)

This is the parent article of that one. A simple summary should suffice, an additional line or two. We already have one line about it. It actually may not be a bad idea to mention air quality models too, if the info could be found. I don't think hydological modeling has reached this point...though last I heard, it's on the cusp of it. Thegreatdr (talk) 06:04, 3 January 2011 (UTC)

But are the latter two really "weather" models? They are fluid dynamics models, certainly, but those seem too far detached from weather forecasting to be in this article, IMO. Titoxd^{(?!? - cool stuff)} 06:25, 3 January 2011 (UTC)

• I expanded the Domains section with a small discussion of vertical coordinates, but now it probably needs a picture. Not sure what picture though. Titoxd^{(?!? - cool stuff)} 06:27, 3 January 2011 (UTC)
  • I made subsections for air quality, climate, and tropical cyclone modeling limited area models. Air quality is directly related/influenced by weather and nowadays (in the US) run off NAM model output, so there's no harm in adding information about its history, so I did so. Hydrology is affected much of the time by past weather conditions as much as current soil conditions and increasingly forecast rainfall. I'll hold off on including hydrological modeling, although I can imagine someone asking why it's not included in here during a GAN or FAC run; I wouldn't know how to argue well against its inclusion. Added some clarification in a couple spots, because the article was leaving certain information assumed, such as pressure decreasing with height making sigma/pressure a natural vertical coordinate. I feel like we should clarify barotropic model slightly better, but am not sure how. The domain section does not necessarily have to have an image, though if we chose it, for balance, we would likely need to show an image of LAM/regional model output. Thegreatdr (talk) 17:05, 3 January 2011 (UTC)
    • Well the way I see it is that those are affected by the weather, more than they affect the weather. Of course, more input would be welcome but I'm not sure anyone else is watching the talk page at the moment… :/ Titoxd^{(?!? - cool stuff)} 17:58, 3 January 2011 (UTC)
    • Maybe an article on the quasi-geostrophic vorticity equation? Or would that be too technical? Titoxd^{(?!? - cool stuff)} 17:59, 3 January 2011 (UTC)
      • Considering what the physics project contributes to wikipedia, there's no such thing as too technical. There's never any harm in a new subarticle. =) Should we go ahead and GAN this article now? There seems to be a week or so backlog at the present, so we'd still have time to make changes. I'm not sure we'd get more input until the article is reviewed in some form as it is. Thegreatdr (talk) 18:07, 3 January 2011 (UTC)

Suggestion: the picture File:NAM 500 MB.PNG at the top is almost indecipherable; massive information overload for the beginner! Even worse, it took me ages to see even which part of the world it is (I suspect, as a foreigner, I might see North America in very pale outline somewhere under there), let alone interpret any data on it. How about something from ECMWF such as [1], copyright permitting? Feline Hymnic (talk) 18:29, 3 January 2011 (UTC)

I believe that's the biggest problem with ECMWF images...copyright. I swapped it out for an image which was previously further down the article, which might be better suited within the lead. Thegreatdr (talk) 15:26, 6 January 2011 (UTC)

Jan 2011 GAN

I'm setting up this section as a drop-box for comments about the article, so that those with more knowledge can pick them up, assess them and act (or not) upon them.

• "The first successful numerical prediction was performed in 1950 by..." What does "successful" mean in this context? In whose view was this "successful" (and by implication, previous ones failures")? Feline Hymnic (talk) 18:39, 3 January 2011 (UTC)
  • Richardson's forecast is generally considered to be a failure, and Charney's forecast to be the first reasonable forecast. See Refs. 2, 3, 7, 9 for more details. Titoxd^{(?!? - cool stuff)} 19:03, 3 January 2011 (UTC)
• "barotropic vorticity equation". "Barotropic" is linked. Good. "Vorticity equation" feels like it ought to be linked. But to what? There are articles "vorticity equation", which has a little human-oriented (as distinct from mathematician-oriented) text and also "barotropic vorticity equation" which looks less human-oriented, but presumably closer to the actual topic. Feline Hymnic (talk) 18:47, 3 January 2011 (UTC)
  • I fixed the wikilink for barotropic vorticity equation, per your comment. You're right though...the subarticle needs more general description itself. I think by successful it is meant that it is the first computationally sound run of a forecast model, not that it verified. Tito can correct me on this if I've misunderstood. Thegreatdr (talk) 19:01, 3 January 2011 (UTC)
    • It was the first sound model (as in it didn't explode and produce 145 mb pressure changes :P) and it also sort of verified. There were some errors in areas of high baroclinicity but the general evolution of the flow seemed well-forecast for its time. Titoxd^{(?!? - cool stuff)} 19:05, 3 January 2011 (UTC)
• Lead mentions limitations on predictions: observations (etc.) and the PDEs. But isn't there discussion amongst researchers about "hydrostatic assumption" (or something like that) being a limitation? Feline Hymnic (talk) 19:08, 3 January 2011 (UTC)
  • Hmm. Doc can correct me on this one, but what I understand is that hydrostatic balance is generally assumed in global models to make them less computationally expensive, and it is not assumed in mesoscale models. Titoxd^{(?!? - cool stuff)} 19:10, 3 January 2011 (UTC)
    • Years ago, I believe this was true. I don't think neglecting latent heat processes is assumed in any models (global or otherwise) nowadays, hence talk of "convective and gridscale feedback" in the GFS, Canadian, and NAM models. I do believe that the mesoscale, limited area, regional models were the first to throw purely adiabatic assumptions out (meaning no addition of heat into the system from any source, whether it be the sun or convection) during the 1990s due to their focus on convection. Besides, computational power increased to the point that this assumption no longer needed to be made, as far as I understand it. I also don't believe that hydrostatic balance is assumed in the guidance anymore. I made some changes to the barotropic vorticity equation subarticle to explain what those assumptions would mean to the real atmosphere. I hope it is understandable. Thegreatdr (talk) 19:22, 3 January 2011 (UTC)
• There are three consecutive headings, all at the same level: "Climate models", "Limited area models" and "Ensemble models". But there's some orthogonality here (and, I suspect, at least one missing item), isn't there? If "L.a. model" is a heading then shouldn't GCM (in either guise of "General circulation model" or "Global climate model") also be present as an equal heading, in a sort of "in contrast to..." way of thinking? And, in similar fashion, isn't "Ensemble model" likewise an "in contrast to..." something else? Feline Hymnic (talk) 19:17, 3 January 2011 (UTC)
  • The ensemble model would be in contrast with a deterministic model (which are most of the models mentioned in the large first subsection) Titoxd^{(?!? - cool stuff)} 19:19, 3 January 2011 (UTC)
    • Perhaps my added subheadings are confusing the matter. I wanted to present that section in a logical order. First came models which were poorly resolved due to simplified equations per the lack of computational power. As the models became global in scope, climate models were developed. Once global models were run on a regular basis with increasing resolution, regional/limited area models were run at even higher resolution. Would eliminating the subheaders avoid this confusion? Thegreatdr (talk) 19:31, 3 January 2011 (UTC)
      • It's fairly chronological, so I like it. Titoxd^{(?!? - cool stuff)} 19:38, 3 January 2011 (UTC)
        • And I just saw you removed them. It's ok with me either way. Titoxd^{(?!? - cool stuff)} 19:41, 3 January 2011 (UTC)
          • I introduced them in order to try to provide some structure to the article while we were editing. If we kept those headers, we'd likely need another header for the initial model developments, and probably a summary section above that in order to remain consistent. They really aren't needed, and the initial summary forced by another header would produce more duplication within the text of the article, so I got rid of the headers. This did not change the current organization of that section. Thegreatdr (talk) 19:46, 3 January 2011 (UTC)
  • [outdent] Personally I like the principle of including subheadings. The issue I was trying (in my own clumsy way) to identify was that there were at least two strands running through, but that they didn't seem to be clearly identified or de-lineated: namely a strand of scale (limited area vs.climate) and a strand of type (single deterministic vs. ensemble). My vote is think about identifying those strands (any others?), structuring the article by mentioning them more explicitly, and using headings to help us apply and maintain that structure. Feline Hymnic (talk) 20:15, 3 January 2011 (UTC)
    • I've readded the subheaders per your recent comments. Thegreatdr (talk) 15:30, 6 January 2011 (UTC)
• "Edward Epstein recognized in 1969 that the atmosphere could not be described only with a deterministic forecast...". Could a met. person please read that sentence very carefully, and try to be a non-met person whilst doing so? Is the word "only" in the correct place, or is there a better place for it? What term is the "only" qualifying? Feline Hymnic (talk) 19:30, 3 January 2011 (UTC)

  Bad English on my part. Only should be "completely". Fixing. Titoxd^{(?!? - cool stuff)} 19:31, 3 January 2011 (UTC)

  Tried to fix it up further. All a deterministic forecast run describes is a run of a forecast model at its full resolution. Reworded the sentence to be more comprehensible to the lay reader. Thegreatdr (talk) 19:38, 3 January 2011 (UTC)

• "...processed by data assimilation and objective analysis methods..." and "...equations are initialized from the analysis data...". Am I right in thinking that the word "analysis" has a special technical meaning in NWP? If so, does it need describing or defining? Feline Hymnic (talk) 20:45, 3 January 2011 (UTC)
• "...locations usable by the model's mathematical algorithms (usually an evenly spaced grid)...". As a lay reader, I can understand "evenly spaced grid" for a small enough region that can be approximated by a flat-land. But what about as that scales up to a sphere? The lay reader may wonder how such a grid is constructed; it won't be latitude and longitude, but what is it? And are there different types of grid? Model-dependent? Finite vs. spectral? Feline Hymnic (talk) 20:51, 3 January 2011 (UTC)
  • How does the grid handle the convergence of meridians at the poles? Titoxd^{(?!? - cool stuff)} 22:13, 3 January 2011 (UTC)
    • Indeed. The previous version said "usually an evenly spaced grid". The recent edit says "usually a grid spaced by latitude, longitude, and elevation". (Let's leave aside elevation here.) If it is long+lat it certainly can't be evenly-spaced across the globe, can it? And yet, wouldn't a reasonable model have some evenness of spacing. Is the commonly used grid (or grids?) rectilinear near the equator and if so what happens as one approaches the poles? Or are different grids (tessellations?) used? Something triangular-ish? Or hexagonal-ish like a soccer ball? Or...? Feline Hymnic (talk) 22:57, 3 January 2011 (UTC)
      • I've added wiklinks (and a ref) for the two types of global model grids used, geodesic and icosahedral. Hopefully these articles which help explain it better. Evenly spaced grids on most guidance can only be said along latitude lines/parallels. Thegreatdr (talk) 15:21, 6 January 2011 (UTC)

• [outdent] Pardon my ignorance but I'm still confused on the "evenly spaced" thing with the current sentence "The grid used for global models is geodesic or icosahedral, which is spaced by latitude, longitude, ...". It seems there are two fundamentally different types of gridding:

1. something with even-spacing characteristics over the whole globe (geodesic and icosahedral seem to fit that idea)
2. something based on lat./long. which would have very different characteristics at the equator (wide-spaced long.) and poles (long. spacing tending to zero)

To me, those two types seem very different from each other. Am I missing something obvious? (Perhaps some models use an evenly-spaced geodesic grid, while others use an uneven lat./long. grid?) Feline Hymnic (talk) 23:43, 6 January 2011 (UTC)

Technically, none of the guidance is on an evenly spaced grid, which is why it is good that you made your initial comment. The boxes are narrower near the pole than they are near the equator. It shows up this way in our workstations at work. Whether you're talking geodesic of icosahedral grid, this will be true. Your initial impression that the grid was not evenly spaced was correct. The latitude lines are evenly spaced from the equator to the pole. The longitude lines converge at the poles, so they get smaller east-west with increasing latitude. Thegreatdr (talk) 02:57, 7 January 2011 (UTC)

• I've just been glancing through some of this discussion ("Global models (list)") above from a year ago, where we drew the distinction between NWP and "atmospheric model", mentioning that NWP includes other things, of which one example would involve ocean coupling. But checking the article as it currently stands, only one mention, and a passing one at that, is made of this topic. Isn't a little more needed about the influence of the ocean on the atmosphere and to introduce the topic of modelling this also? Feline Hymnic (talk) 22:51, 3 January 2011 (UTC)
  • The history section now talks about the development of ocean wave models, and the initialization section talks about the inclusion of sea ice and sea surface temperature within the model guidance. Thegreatdr (talk) 16:28, 8 January 2011 (UTC)
• In the history section: "Using a hydrostatic variation of Bjerknes's primitive equations, In his book, Weather Prediction by Numerical Process, Richardson produced...". That doesn't look right. But I don't know what was intended or what it should say.

  Richardson described his predictive scheme in WPNP, and it was based on the primitive equations described by Bjerknes. He also employed the hydrostatic balance approximation we discussed previously. It probably needs clarifying somehow. Titoxd^{(?!? - cool stuff)} 21:45, 4 January 2011 (UTC)

  So do you mean something like "Using a hydrostatic variation of Bjerknes's primitive equations, Richardson produced..."? (What I've done there is simply chop out the phrase "In his book, Weather Prediction by Numerical Process" (which also incorrectly had a capital "In".) Having done that reduction, you may then want to pad it up a bit (if necessary, if the existing citations in the sentences are insufficient). Feline Hymnic (talk) 22:21, 4 January 2011 (UTC)

  That should work. Titoxd^{(?!? - cool stuff)} 22:37, 4 January 2011 (UTC)

  [outdent] Someone left a comment on my own talk page: "...Why does it say that the first attempts were in the 1920s, and then go on to mention Richardson's work, which he did during the First World War? Something's no right." I don't know the history or meteorology enough to comment on this. The article doesn't actually mention the first world war, but might he have made an earlier attempt to do this during the war (i.e. before the 1920s)? Feline Hymnic (talk) 23:29, 6 January 2011 (UTC)

  Richardson started work on WPNP during WWI, but did not publish the book until 1922. So the answer depends on when you say the attempt was being made: At the time of calculation, or at the time of publication. Titoxd^{(?!? - cool stuff)} 04:52, 7 January 2011 (UTC)

• Somebody added a request for clarification of the singular vector/bred vector ensemble error probability distributions. I added those to differentiate the two models, but the differences between the two approaches are rather beyond the scope of an introductory article, and venture well into "hell if I know" territory for almost everybody but numerical analysts. Would it be better to remove that bit? Titoxd^{(?!? - cool stuff)} 06:42, 7 January 2011 (UTC)
  • Yeah, that would be better placed within the article on ensembles, if we have one. Thegreatdr (talk) 10:43, 7 January 2011 (UTC)
    • Rather than either/or, how about both/and?

1. Leave the text almost as it is, but try to find suitable wikilinks (I've just done the vector breeding one, and, incidentally, the "Mathematics" wikiproject agreed to my suggestion about renaming that article). At least the passing reference here alerts folk to the principle that there are different ways of doing the modeling.
2. Let a future article on ensembles go into more detail. Feline Hymnic (talk) 11:45, 8 January 2011 (UTC)

• • • • Look at you, coordinating across projects. Well done. =) Good job on finding the bred vectors article. When I searched for vector breeding, I couldn't find it. Per your recent coordination, and some additional thinking, I've added the physics (fluid dynamics) and math projects to the talk page, since both are relevant here. Thegreatdr (talk) 15:49, 8 January 2011 (UTC)

Do not like the recent revision, in separating out the sea ice and SSTs into their own tiny sections. First off, those section aren't likely to grow from their current tiny size, which would prevent GANing and FACing. Secondly, they are initialized in global atmospheric models, climate models, and likely ocean modeling, so why not keep them somewhere above the subsections since they apply to all of models mentioned (except maybe air quality)? I've gone ahead and reverted the edit. Thegreatdr (talk) 20:59, 8 January 2011 (UTC)

OK, fair enough! (Hence my comment in my edit summary when I did that.) Feline Hymnic (talk) 21:17, 8 January 2011 (UTC)

I've been reflecting a little more on this, and bear in mind that I am a not a meteorologist at all. I suspect that there is still something that needs clarifying in the article. In my earlier edit (attempt to move sea-ice and SST out of section "initialization" and into section "modelling") I was probably assuming that a model run would be treating those items as variables that would themselves evolve through the run of the model. But I suspect what you're saying is that an initial state for them is fixed, and that they are assumed to remain constant throughout the run. (In practice they'll change slowly, but do models take the pragmatic approach that any such change can be ignored during a run? Are they something vaguely like boundary conditions in (what little I remember of) some differential equations?) Might the article benefit from distinguishing (probably) three things: (1) what items evolve (the usual atmospheric state stuff) (2) what items are assumed to remain constant (perhaps sea-ice, SST?) (3) the information of how initial data is assimilated at the start of a run. Or something like that. Feline Hymnic (talk) 10:50, 9 January 2011 (UTC)

Will have to look this up, but within Atmospheric models when sea ice and SST were originally used for initialization, it was originally climatology. Then it became observed fields that didn't change through the forecast runs. I'm not sure this is true anymore, though. With increased computational power comes more possibilities for these fields to be forecast like the atmospheric fields. In global climate models, SSTs and sea ice would have to be forecast with GCMs, because the fields change markedly through the year, and for them to be able to forecast ENSO state and global warming, they couldn't keep them stagnant or tied to climatology. Tropical cyclone models would need to have SSTs as a forecast field due to upwelling caused by stalled/stationary systems. Will have to check into Atmospheric models though. Thegreatdr (talk) 13:23, 9 January 2011 (UTC)

It took a while to find using several different combinations of search terms, but so far it looks like the GFS and Navy NOGAPS global models produce SST forecasts within the short range period (out 84 hours). If they are forecasting this variable, the Canadian, UKMET, and ECMWF likely do as well. I did find that NCEP produces a sea ice model for Alaska, which runs through 384 hours, and has since the late 1990s. Thegreatdr (talk) 21:10, 9 January 2011 (UTC)

What's left?

So what is left to do before a GAN/FAC run? Titoxd^{(?!? - cool stuff)} 04:46, 12 January 2011 (UTC)

I don't know. I think we responded to Feline's useful input. Bitewise, we've quadrupled the size of the article since we began. I significantly improved one of the subarticles, sea surface temperature, over the weekend, and tried to make the barotropic model wikilink/subarticle more understandable. We could probably safely GAN this now. Any other input? Thegreatdr (talk) 09:50, 12 January 2011 (UTC)

I posted the GAN. I think that we could probably try our luck at FAC as soon as the GAN is complete, assuming everything goes well. Titoxd^{(?!? - cool stuff)} 00:21, 13 January 2011 (UTC)

look at that spike Titoxd^{(?!? - cool stuff)} 09:56, 13 January 2011 (UTC)

I've reviewed the article for GAN. In addition to my comments, be sure you address the above comments by Feline. --♫ Hurricanehink (talk) 04:21, 17 January 2011 (UTC)

GA Review

This review is transcluded from Talk:Numerical weather prediction/GA1. The edit link for this section can be used to add comments to the review.

Reviewer: ♫ Hurricanehink (talk) 04:18, 17 January 2011 (UTC)

• The lede should be split into two paragraphs

  How about that? Titoxd^{(?!? - cool stuff)} 05:38, 17 January 2011 (UTC)

• You have two consecutive sentences starting with "Bjerknes" - try rewording one of them

  Fixed. Titoxd^{(?!? - cool stuff)} 05:38, 17 January 2011 (UTC)

• "to numerically estimate" - avoid split infinitives
  • "to explicitly resolve" - that's used in "Domains"
  • "to definitively predict " - that's used in "ensembles"
  • "to not be" - that's also used in "ensembles"
    • Fixed all those. Titoxd^{(?!? - cool stuff)} 05:38, 17 January 2011 (UTC)
• I notice a lot of the history uses passive voice. Sometimes, the sentence could and should be reworded to clarify what has been done.
  • "was entirely subjective and based" - that is an example where there is improper parallelism, in that subjective would be the adjective to "was"'s being verb, but "based" makes "was based" a passive voice clause.
• "His forecast calculated that the change in surface pressure would be of 145 millibars (4.3 inHg)" - the "would be of" is throwing me off - is that proper writing?
  • Fixed. Titoxd^{(?!? - cool stuff)} 05:38, 17 January 2011 (UTC)
• "Soon after, in 1966" - is that really soon after 1959?
  • I think it is :) but removed it anyways. Titoxd^{(?!? - cool stuff)} 23:05, 17 January 2011 (UTC)
• Who/what organization introduced the first global computer model?
  • That whole paragraph is all the National Meteorological Center (now known as the HPC). How would I reword it without being redundant? Titoxd^{(?!? - cool stuff)} 05:38, 17 January 2011 (UTC)
• "In 1956, Norman Phillips developed a mathematical model which could realistically depict monthly and seasonal patterns in the troposphere, which became the first successful climate model." - is "realistically" the best word there?
  • Well, it's the first climate model with any sort of accuracy, so what other words would be adequate here? Titoxd^{(?!? - cool stuff)} 05:38, 17 January 2011 (UTC)
• "Also, swell was not modeled realistically." - it's a really small sentence - any way to combine it with another one, or expand on it?
  • Merged it into the previous sentence. Titoxd^{(?!? - cool stuff)} 05:59, 17 January 2011 (UTC)
• The "Initialization" paragraph is really long. Can you split it?
  • Split and rearranged. Titoxd^{(?!? - cool stuff)} 05:59, 17 January 2011 (UTC)
  • "Models are initialized using this observed data." - any reason for the italics?
    • This is to use as a definition of the term, in case we create a redirect like Initialization (meteorology). Titoxd^{(?!? - cool stuff)} 05:59, 17 January 2011 (UTC)
• Also, how come there is history stuff in the "Initialization" paragraph? There is already a history section
  • Yes, but the dates mentioned in the initialization section pertain to the times where model initialization was improved through the addition of new variables. Titoxd^{(?!? - cool stuff)} 06:06, 17 January 2011 (UTC)
• "A model, in this context" - that's a poor way to start a paragraph. Try introducing the subject better
  • Took out the awkward phrase, which was a relic to a version that predated our recent edits. Thegreatdr (talk) 22:49, 17 January 2011 (UTC)
• "the set of equations used to predict the known as the physics" - is there a word missing?
  • Stray clause. Fixed. Titoxd^{(?!? - cool stuff)} 17:44, 17 January 2011 (UTC)
    • I further refined the sentence. Thegreatdr (talk) 22:54, 17 January 2011 (UTC)
• The writing feels rather dull in the first paragraph of the "Computation" section. So many of them start with "the x did this", or "this y did that".
  • See if the new version is better. Thegreatdr (talk) 22:54, 17 January 2011 (UTC)
• "This makes sense" - ehh, pretty unencyclopediac writing
  • Reworded. Thegreatdr (talk) 22:56, 17 January 2011 (UTC)
• "Extremely small errors in the initial input, such as temperatures and winds, within numerical models double every five days" - it seems the commas are out of place. The sentence just reads weirdly
  • It looks like this has already been addressed. Thegreatdr (talk) 22:58, 17 January 2011 (UTC)
• "involving multiple forecasts created with and individual forecast model" - something missing there? Is "and" supposed to be "an"?
  • Yes. It has been fixed. Thegreatdr (talk) 23:00, 17 January 2011 (UTC)
• "The ensemble forecast is usually evaluated in terms of the ensemble mean of a forecast variable, and the ensemble spread, which represents the degree of agreement between various forecasts, known as ensemble members, in the ensemble system." - you use "ensemble" five times in one sentence! Plx fix, k thx
  • Okay. That was funny. It has been reworded. Thegreatdr (talk) 23:03, 17 January 2011 (UTC)

All in all, a good read. There are just some minor writing things that I'd like to see worked on. As such, I put the GAN on hold. Let me know if you have any questions. --♫ Hurricanehink (talk) 04:18, 17 January 2011 (UTC)

Looks good! And I was especially impressed how you limited the number of "ensemble" usages in the one sentence from five to two. I'm happy to pass it now. --♫ Hurricanehink (talk) 00:28, 18 January 2011 (UTC)

So we have the green plus sign

Do we go for the bronze star? Titoxd^{(?!? - cool stuff)} 02:15, 18 January 2011 (UTC)

We could. I wish others from the project would have helped out. Oh well. Did you want to go through a peer review first, or is that unnecessary? Thegreatdr (talk) 23:25, 23 January 2011 (UTC)

Just a quick note of congratulations to Thegreatdr and Titoxd for achieving this. Good stuff! Feline Hymnic (talk) 23:36, 23 January 2011 (UTC)

Random to-do's

• A reviewer in the FAC asked for links to prognostic chart and spaghetti diagram. While the first one probably doesn't have enough content to have a separate article, the second one could be made into a stub with a picture rather easily. Titoxd^{(?!? - cool stuff)} 19:51, 13 February 2011 (UTC)
  • Done. Thegreatdr (talk) 00:09, 27 February 2011 (UTC)

Question about a sentence in the lead

The following sentence in the lead raises some questions in my mind:

...the chaotic nature of the partial differential equations that govern the atmosphere...

First, the use of the word 'govern' here seems odd. I would think that the equations model the behavior of the atmosphere, rather than govern it. Second, are the equations themselves chaotic, or is it the fact that any small imprecision in the input causes a chaotic variation in the output? Thank you.—RJH (talk) 20:16, 21 February 2011 (UTC)

You're right, govern was inappropriate there. Otherwise, both issues you mention cause the errors within forecast guidance to grow with time...initial uncertainty and the solutions from the partial differential equations, where various assumptions are made. Some studies indicate that with increasing resolution, the "assumptions" made in the forecast guidance (the statistically-based model parameterizations) also introduce increasing amounts of error. Thegreatdr (talk) 20:25, 21 February 2011 (UTC)

(edit conflict) Well, in respect to the first point, "governing equations" in engineering and science is the term used to describe the model used to simulate some process. That should probably be reworded since it's a bit jargony. As for chaos, it's both: the second part (infinitesimal imprecisions in the input resulting in large, finite changes in the output) is essentially the definition of a chaotic process, so the equations themselves are said to be chaotic. Titoxd^{(?!? - cool stuff)} 20:26, 21 February 2011 (UTC)

Okay, thank you.—RJH (talk) 20:43, 21 February 2011 (UTC)

"Governing equations" is correct; see e.g.[2]. Also it is indeed the chaotic nature of the coupled equations that causes error in the initial conditions to grow. The system is chaotic regardless of the magnitude of the initial error. Short Brigade Harvester Boris (talk) 21:07, 21 February 2011 (UTC)

FA promotion

Congratulations! Good work! Nageh (talk) 17:21, 3 April 2011 (UTC)

How is a cool, interactive, timeline of the exact article linkspam? I was going to donate this technology to Wikipedia. Granted, this article doesn't have a lot of dates in it. I am not running any ads on this page, it just displays the article in a unique way. Jroehl (talk) 01:58, 10 August 2011 (UTC)

The link you were posting does not even mention anything about numerical weather prediction. It's a general timeline about U.S. history. Titoxd^{(?!? - cool stuff)} 02:01, 10 August 2011 (UTC)

Titoxd,

The link I posted was:

http://wikitimelines.net/?tl=ECL6A1A2X&todo=showtl

And I even tested it, and it came up correctly. The timeline about U.S. history is the default timeline if no parameters are passed in the URL. I dont know how you got the wrong timeline. And thanks for your informative and speedy reply Titoxd. Jroehl (talk) 02:10, 10 August 2011 (UTC)

Just the US History here, too. Nageh (talk) 13:17, 10 August 2011 (UTC)

Does wikipedia strip off parameters passed in a URL when it is an external link such as the following:

http://wikitimelines.net?tl=ECL6A1A2X&todo=showtl

Something seems to be inserting a "/" before the question mark when the page is rendered such as:

http://wikitimelines.net/?tl=ECL6A1A2X&todo=showtl

Which nullifies the parameters and brings up the default timeline, which is the history of the United states. I cant figure that out. Jroehl (talk) 15:10, 10 August 2011 (UTC)

Never mind, the URL is:

http://wikitimelines.net/tl3.fwx?tl=ECL6A1A2X&todo=showtl

I shouldnt be working this early in the morning. I am making stupid mistakes. Jroehl (talk) 15:23, 10 August 2011 (UTC)

And yet...

It is impossible to solve these equations exactly, and small errors grow with time (doubling about every five days).
And yet we can predict to a tenth of a degree the temperature in 80 or 90 years. Imagine that. — Preceding unsigned comment added by 24.250.206.199 (talk) 04:58, 10 August 2011 (UTC)

The temperature in 80 or 90 years is a matter of climate, which can, in some respects, be interpreted more accurately decades in advance than weather can be predicted 24 hours out. The weather is different from city to city, usually street to street, and very often from house to house; it tells us things like the transient temperature, weather conditions at a specific point (whether it's raining, snowing, overcast, clear, etc). Climate is much more broad, and isn't so much a matter of trying to pinpoint where it'll rain next as looking back in the past, figuring out what's going on now, and (largely) extrapolating it well into the future. Juliancolton (talk) 13:10, 10 August 2011 (UTC)

the pic on the right

"height or pressure" you mean "height AND pressure" right? — Preceding unsigned comment added by Odarcan (talk • contribs) 21:56, 10 August 2011 (UTC)

Numerical modeling

Milutin Milanković should be mentioned in this article because he initiated numerical modeling of the climate.--Свифт (talk) 09:21, 4 April 2012 (UTC)

Data assimilation?

I understand that data assimilation is a major component of NWP. But contrast the state of this NWP article (very good) and the DA article (distinctly less good). Could some of the NWP-knowledgable editors turn their attention to the DA article, please? Feline Hymnic (talk) 00:19, 19 October 2013 (UTC)

Older definition of weather prediction

/there are some semantic issues though/

The probability of the occurrence of a weather phenomenon (e.g. precipitation)
is the expected value (of that phenomenon occurring)
of a large number of virtual outcomes,
for certain periods of time (quanta of time)
and the amount of a specific measurable effect of that weather phenomenon (over a threshold)
over each unit of time. — Preceding unsigned comment added by 2.84.222.127 (talk) 19:27, 16 August 2015 (UTC)

External links modified

Hello fellow Wikipedians,

I have just modified 3 external links on Numerical weather prediction. Please take a moment to review my edit. If you have any questions, or need the bot to ignore the links, or the page altogether, please visit this simple FaQ for additional information. I made the following changes:

• Added archive https://web.archive.org/web/20100710080304/http://www.geog.ucla.edu/~yxue/pdf/1996jgr.pdf to http://www.geog.ucla.edu/~yxue/pdf/1996jgr.pdf
• Added archive https://web.archive.org/web/20110401013742/http://www.atmos.washington.edu/~dargan/591/diag_cloud.tech.pdf to http://www.atmos.washington.edu/~dargan/591/diag_cloud.tech.pdf
• Added archive https://web.archive.org/web/20110823082059/http://nldr.library.ucar.edu/collections/technotes/asset-000-000-000-845.pdf to http://nldr.library.ucar.edu/collections/technotes/asset-000-000-000-845.pdf

When you have finished reviewing my changes, you may follow the instructions on the template below to fix any issues with the URLs.

This message was posted before February 2018. After February 2018, "External links modified" talk page sections are no longer generated or monitored by InternetArchiveBot. No special action is required regarding these talk page notices, other than regular verification using the archive tool instructions below. Editors have permission to delete these "External links modified" talk page sections if they want to de-clutter talk pages, but see the RfC before doing mass systematic removals. This message is updated dynamically through the template {{source check}} (last update: 5 June 2024).

• If you have discovered URLs which were erroneously considered dead by the bot, you can report them with this tool.
• If you found an error with any archives or the URLs themselves, you can fix them with this tool.

Cheers.—InternetArchiveBot (Report bug) 14:07, 20 May 2017 (UTC)

External links modified

Hello fellow Wikipedians,

I have just modified one external link on Numerical weather prediction. Please take a moment to review my edit. If you have any questions, or need the bot to ignore the links, or the page altogether, please visit this simple FaQ for additional information. I made the following changes:

• Added archive https://web.archive.org/web/20121022215636/http://www.metoffice.gov.uk/research/areas/data-assimilation-and-ensembles/ensemble-forecasting/MOGREPS to http://www.metoffice.gov.uk/research/areas/data-assimilation-and-ensembles/ensemble-forecasting/MOGREPS

When you have finished reviewing my changes, you may follow the instructions on the template below to fix any issues with the URLs.

This message was posted before February 2018. After February 2018, "External links modified" talk page sections are no longer generated or monitored by InternetArchiveBot. No special action is required regarding these talk page notices, other than regular verification using the archive tool instructions below. Editors have permission to delete these "External links modified" talk page sections if they want to de-clutter talk pages, but see the RfC before doing mass systematic removals. This message is updated dynamically through the template {{source check}} (last update: 5 June 2024).

• If you have discovered URLs which were erroneously considered dead by the bot, you can report them with this tool.
• If you found an error with any archives or the URLs themselves, you can fix them with this tool.

Cheers.—InternetArchiveBot (Report bug) 14:45, 6 October 2017 (UTC)

WP:URFA/2020

• Numerous unsourced statements, including an entire section
• Grammatical errors throughout
• References are outdated in some areas
• Some sections are in need of an update material/coverage-wise as well

• Surely something has happened since the 1990s that can be added to the history section
• I see plenty of new scholarly sources out about this topic

I am listing this article at WP:FARGIVEN due to its numerous issues. Noah^Talk 23:01, 4 December 2021 (UTC)

Likely copyright violation detected. Needs further checks. Noah^Talk 23:05, 4 December 2021 (UTC)