Wikipedia

Wikipedia:Featured article candidates/Numerical weather prediction/archive1

< Wikipedia:Featured article candidates
The following is an archived discussion of a featured article nomination. Please do not modify it. Subsequent comments should be made on the article's talk page or in Wikipedia talk:Featured article candidates. No further edits should be made to this page.

The article was promoted by SandyGeorgia 19:13, 26 March 2011 [1].

Numerical weather prediction edit

Numerical weather prediction (edit | talk | history | protect | delete | links | watch | logs | views)

Toolbox
Nominator(s): Titoxd(?!? - cool stuff) 01:32, 9 February 2011 (UTC)[reply]

Thegreatdr and myself have been working on this article for a while, and following the reviews given by Feline Hymnic and Hurricanehink, I believe the article meets all the featured article criteria. Hopefully you find the article informative, as well as well-written. Titoxd(?!? - cool stuff) 01:32, 9 February 2011 (UTC)[reply]

  • "ensemble spread to be small enough". I see 5 relationships b/w ensemble members and observations on Warner 267. Are we saying that underdispersion is by far the most common problem? I'm no saying you have to add all 5; you only need to explain what you are saying and why you are saying it... Locke'sGhost 09:33, 9 February 2011 (UTC)[reply]
    Well, from what I understand, underdispersion is the most severe problem for extended forecasts (~ τ ≈ 720 hours) and thus it gets more emphasis. I tried to clarify that in the article's prose. Titoxd(?!? - cool stuff) 20:45, 9 February 2011 (UTC)[reply]
  • "Although these Monte Carlo simulations showed skill, Leith showed" ... Leith who? Locke'sGhost 12:24, 11 February 2011 (UTC)[reply]
    • Cecil Leith from NCAR. should I add the "from NCAR" bit too? Titoxd(?!? - cool stuff) 18:27, 11 February 2011 (UTC)[reply]
  • Gongbing Peng, Lance M. Leslie, Yaping Shao "Environmental modelling and prediction" has a nice flowchart of the modelling process (figure 3.2, p. 81), with discussion starting around p. 78. Would it be useful to add something like this...? The whole article is pretty difficult to digest; needs some guidance...? Please don't fall into "a reviewer suggested it; must add it". Your thoughts solicited. Locke'sGhost 13:03, 11 February 2011 (UTC)[reply]
    If a flow diagram of the process helps the reader into understanding how NWP works, I don't see it shouldn't be added. That said, that diagram seems a bit too simple for my tastes, since it glosses over objective analysis and initialization, and it is not very clear about how model output is recirculated in the model. Titoxd(?!? - cool stuff) 18:16, 11 February 2011 (UTC)[reply]

Sources comments

  • Something has gone haywire with the formatting in ref 18; an url seems to have escaped.
  • Retrieval date formats need to be consoistent (sometimes you have "Retrieved", other times "Retrieved on")
  • It is not necessary to add a retrieval date for online links if the material is available in print form, but if you choose to do this, you need to be consistent and do it in every case.
  • Print sources (newspapers, journals, books etc) should be italicised. See ref 33 The Washington Post, look for others
    • This seemed to be the only occurrence, and it was fixed. Thegreatdr (talk) 20:36, 11 February 2011 (UTC)[reply]
  • CUP should be spelt out
    • I think CUP is short for Cambridge University Press, which appears to be indicated on a web search, but the book's actual information gives no clues. Thegreatdr (talk) 20:32, 11 February 2011 (UTC)[reply]
  • Ref 42. A reader wishing on check this would have to search through your list to find what "Pielke" referred to. Where short citations are used, details of the book should be listed separately, in a bibliography section.
  • Cited works should not appear in "Further reading" (Pielke again)

Otherwise, all sources look scholarly and reliable. Spotchecking was difficult when most of the sources have very wide page ranges, and I am totally unfamiliar with the material, so not much done. Brianboulton (talk) 20:07, 11 February 2011 (UTC)[reply]

Sources issues all satisfactorily resolved. Brianboulton (talk) 22:25, 12 February 2011 (UTC)[reply]

Comments:

  • Heck of a job tackling a difficult (to me, at least) subject. It took me a lot of careful reading to understand, especially because of the jargon involved, but it gets the message across. This is the kind of article where the wikilinks have to do a lot of the heavy lifting in regards to addressing the jargon; there simply isn't a way to express the terms without using many more words. That said, I have to wonder if the topic wouldn't be better suited if the history section were placed further down the page. That way, you can introduce the concepts first and explain some of the jargon without throwing the reader directly into the mix. I found myself finding answers to questions I was asking in the history section as I read further into the article. It's something to consider.
  • I did a quick copy edit, so please take a look at what I've done to make sure I haven't screwed anything up too badly. It was only about halfway through that I realized this might have been written in British English, accounting for the collective noun/possessive agreement issues I found, so rather than stop halfway through and try to find everything I changed, I just kept going.
  • The article frequently uses the word "fields"; both "three-dimensional fields" and "wind fields" appear -- could you explain this?
    These are scalar fields and vector fields, sometimes in 3D space. (We could also link to flow velocity, but that article makes no sense whatsoever to most readers.) Titoxd(?!? - cool stuff) 21:01, 12 February 2011 (UTC)[reply]
  • When talking about the 1950 prediction, could you drop in a reference that refers to the technique they did, one sentence before No. 5 appears? I can't tell if No. 5 is supposed to cover that as well, or just the sentence about how it reduced computer time.
  • Could you cite Richardson's quote, please?
    • It is covered by the reference at the end of the next sentence, but made that explicit nonetheless. Titoxd(?!? - cool stuff) 21:01, 12 February 2011 (UTC)[reply]
  • What is a "thermotropic model"? You explain the barotropic model, but this one pops up in the second-to-last paragraph of the history section.
  • You say Hinkelmann was the first to produce a "reasonable" forecast. What made JNWPU's work unreasonable? Was it because it was based on a simpler idea?
    • JNWPU's work was reasonable, but it was not based on the primitive equations. Hinkelmann's work was reasonable and based on the primitive equation model. Titoxd(?!? - cool stuff) 21:05, 12 February 2011 (UTC)[reply]
  • What does "high-frequency noise" mean in this context?
    Essentially, minuscule variations in the initial condition. For the sake of example, if we approximate the state of a theoretical one-dimensional atmosphere as a sum of sines and cosines (essentially a Fourier series), we will find that with a few terms, we have a very smooth atmosphere. The more terms we add, we are able to distinguish more and more detail. However, as the picture in the Fourier series article shows, the solution becomes very spiky. When we take the derivative, those spikes correspond to huge changes in slope in a small distance, so the derivative becomes huge. Since the models depend on the value of the derivative, the whole thing explodes into a useless mess. This can be best explained with the graph at the start of [2].
    In reality, the situation is more complicated, since you could have initial conditions that don't match up and produce a velocity field that is not divergence-free or all other sorts of numerical issues. I tweaked the wording to try to clarify this. Titoxd(?!? - cool stuff) 21:31, 12 February 2011 (UTC)[reply]
  • I've switched your wikilink from radiation to solar radiation -- my first thought was radiation from atomic tests, but I knew that couldn't be correct.
  • What is convective rain? Is there a wikilink for this?
    • It's just jargon for normal rain affecting convection processes. Reworded. Titoxd(?!? - cool stuff) 21:46, 12 February 2011 (UTC)[reply]
  • How does a global forecast model (mentioned in the last sentence of the history section) differ?
    • Essentially, global forecast models simulate the whole Earth, and thus tends to have lower resolution. Titoxd(?!? - cool stuff) 21:31, 12 February 2011 (UTC)[reply]
  • "initialize and model soil and vegetation types" How do you initialize a type?
    • I'm not really sure how to address this: initialization is just telling the computer what kind of initial conditions and in which domain the problem is going to be solved. As such, initializing the soil and vegetation types is just punching in a variable in the computer model. I'm not sure how to clarify this without going completely off-topic. Titoxd(?!? - cool stuff) 10:29, 13 February 2011 (UTC)[reply]
  • How does modeling oceanic processes differ? The climate section leads me to believe there are additional problems involved, but I don't know.
    It's essentially a completely different problem. While air and water are both fluids, water is 1,000 times as dense as air, and such behaves completely differently. Oceanic circulation has a much higher Rossby number, so the Earth's rotation is an extremely-dominant factor in the circulation; additionally, problems such as the thermohaline circulation are completely inapplicable to atmospheric dynamics. Our article in ocean dynamics is the best place to explain this, and would be the primary article for this information.
    As for this article: flow of air over the ocean affects the ocean surface, and its modeling history is explained in this article. I added a couple of {{main}} links to the page, and again, I'm not sure what to add to satisfy this concern without veering too much off-topic. Titoxd(?!? - cool stuff) 10:29, 13 February 2011 (UTC)[reply]
  • Under limited area modeling, I think you first need to define what a "limited area" is. Under the history section, I got the impression forecasters started out by modeling regions, which seem to me like a limited area. If there's a difference, it'd help if it was spelled out.
    • Yes. The first models were limited-area; then came global models. Modern limited-area models use global models as inputs, and then use different numerical schemes to resolve more physical processes over a limited area. Titoxd(?!? - cool stuff) 21:31, 12 February 2011 (UTC)[reply]
  • Under ensemble modeling, "showed skill" isn't very clear to me ... does this mean he was on the right track but didn't quite have it?
    • "Showed skill" just means that the forecast is better than what one would gather by just looking at the climatological conditions in the area. Linked to the first occurrence of forecast skill to try to clarify this. Titoxd(?!? - cool stuff) 21:53, 12 February 2011 (UTC)[reply]
  • What is "initial probability density"?
  • Is METAR report redundant?
  • Could you provide a citation for ship-supplied weather information, as you have with pilot reports?
  • Is there a wikilink for prognostic chart?
    • No, but the article defines it: "The visual output produced by a model solution is known as a prognostic chart, or prog for short.[45]" Titoxd(?!? - cool stuff) 21:50, 12 February 2011 (UTC)[reply]
  • Is there a wikilink for spaghetti diagrams?
    • No, but the article defines it: "Ensemble spread is diagnosed through tools such as spaghetti diagrams, which show the dispersion of one quantity on prognostic charts for specific time steps in the future." Titoxd(?!? - cool stuff) 21:50, 12 February 2011 (UTC)[reply]
      • The one previously within wikipedia for spaghetti plot has been upgraded, and it is now linked. A redirect now exists from spaghetti diagram. Thegreatdr (talk) 00:12, 27 February 2011 (UTC)[reply]
  • That's about it from me; if you have any questions, comments or concerns about what I've written here, please drop a note on my talk page, and I'll get back to you as soon as you can. Good luck with the article! JKBrooks85 (talk) 13:35, 12 February 2011 (UTC)[reply]
  • The changes look good to me. I'm sure folks with other viewpoints will want to chime in, but I'm more than happy to support this article as comprehensive and worthy of FA status. JKBrooks85 (talk) 11:22, 13 February 2011 (UTC)[reply]

Comments. Looks good, but some niggles at the top that suggest the whole text needs TLC:

  • "atmospheric-dynamics-based forecast models"—please always consider reversing long gobbledy nominal groups of the type that strictly speaking need multiple hyphens. Why not this: "forecast models based on atmospheric dynamics".
  • "The development of limited area, or regional, models"—is that an equative "or", or an either-or? It goes bump-bump, too. Hyphen required. Why not "The development of limited-area (regional) models"?
  • "are run to help create forecasts for nations worldwide."—A bit clunky, the "nations" bit. It couldn't be just "worldwide forecasts", could it?
  • Pet peeve: scientists who write in telegram language. If there's an "of" after the nominal group, put a "the" before it (mostly works): "The use of model ensemble forecasts ...".
  • "Because the output of atmospheric dynamics-based forecast models is not perfect near ground level"—so it is perfect further up? Hard to believe.
    • Well, perfect is the incorrect word to use. The models' output has to be corrected for terrain (which can cause rain shadows/orographic rain, surface rougness (affects wind speeds), among other assorted effects based on the presence of the atmospheric boundary layer. Switched with "needs corrections". Titoxd(?!? - cool stuff) 19:44, 14 February 2011 (UTC)[reply]
  • I guess this long sentence is ok without splitting it, but make it easier for the readers: "which post-process the output of dynamical models with the most recent surface observations and climatology using statistical techniques" -> "which use statistical techniques to post-process the output of dynamical models with the most recent surface observations and climatology".
    • I added a bit about forecast points and split the sentence. Titoxd(?!? - cool stuff) 19:44, 14 February 2011 (UTC)[reply]
  • "forecast skill"—I haven't gone to the link-target, but skill sounds bizarre when applied to a model. "power to forecast", "forecasting accuracy" ... err, surely there's another way.
    • The problem is that forecast skill is a clearly-defined term in weather forecasting, and it stands for the ability of a model to predict the weather in comparison to another baseline (usually climatology). As such, I'm hesitant to change it since it could unexpectedly change the meaning of the prose. Titoxd(?!? - cool stuff) 19:44, 14 February 2011 (UTC)[reply]
  • "Until the end of the 19th century, weather prediction was entirely subjective and based on empirical rules." Is there tension between "subjective" and "empirical rules"? So often we see subjectivity pitted against the evidence-based. I trip up at this "contrast" here.
    • I see what you mean. The contrast is actually between subjective empiricism and scientific forecasting. I added a bit of prose that I believe clarifies this. Titoxd(?!? - cool stuff) 19:44, 14 February 2011 (UTC)[reply]
  • Start of "History": past tense / present tense dissonance. "proposed that the atmosphere was governed by the same principles of thermodynamics and hydrodynamics that were studied in the previous century" -> "proposed that the atmosphere is governed by the same principles of thermodynamics and hydrodynamics that had been studied in the previous century" (he was proposing a universal truth, yes?). Soon after, present tense is used for such proposed truths, which is fine.
  • "Richardson produced a 6-hour forecast for the state of the atmosphere over two points in central Europe by hand, taking at least six weeks to do so." Perhaps "Richardson produced by hand a 6-hour forecast for the state of the atmosphere over two points in central Europe, taking at least six weeks to do so." Or maybe even "Richardson took at least six weeks to produce by hand a 6-hour forecast for the state of the atmosphere over two points in central Europe."
The italics were, of course, an artefact of this review, to point out the change suggested. Tony (talk) 23:44, 14 February 2011 (UTC)[reply]
    • Switched to the first alternative, but took out the italics, since they are being used in the article to denote definitions of ancillary terms. Titoxd(?!? - cool stuff) 19:44, 14 February 2011 (UTC)[reply]
  • Again, what is shoved right at the end of a long sentence seems to belong earlier, but I can't work out a better way: "The first successful numerical prediction was performed in 1950 by a team composed of the American meteorologists Jule Charney, Philip Thompson, Larry Gates, and Norwegian meteorologist Ragnar Fjörtoft and applied mathematician John von Neumann, using the ENIAC digital computer." Unsure ... "The first successful numerical prediction was performed using the ENIAC digital computer in 1950 by a team composed of the American meteorologists Jule Charney, Philip Thompson, Larry Gates, and Norwegian meteorologist Ragnar Fjörtoft and applied mathematician John von Neumann.
    • I switched it to your suggestion, but having the two adverbial clauses "using … in" sounds weird to me for some reason. Titoxd(?!? - cool stuff) 19:44, 14 February 2011 (UTC)[reply]
  • "could be performed on the relatively primitive computers available"—maybe "... computers of the day" stops the impression that better computers were around, but they couldn't access them. Unsure.

This is a fascinating story. I hope the nomination succeeds. Please try to find an independent copy-editor who's used to unwinding scientists' text. I love the concept that at a certain point of time, forecasts were quicker to perform that the forecast period! Tony (talk) 12:29, 14 February 2011 (UTC) PS, as usual, the text on the opening scematic is TINY. I'm enlarging it, but the task is challenging; readers shouldn't have to download the original image to get it (especially those on slow connections). Can we acknowledge NOAA in the caption? Whatever it is ... it's even hard to determine from the external link at the description page. And can't we have the full colour version? Tony (talk) 12:33, 14 February 2011 (UTC)[reply]

That is the full-color version. :) I guess we could get the graphics lab to give us a better picture, but I have no idea how long it would take. Re: reviewers: Who would you suggest? Usually when Hurricanehink needs a copyeditor, he asks me, and in this case, he has already provided one for the article's GAN. So that takes out both of WP:WPTC's copyeditors, I guess. Feline Hymnic provided both a copyedit and a content review. I'm not really sure who to ask. Titoxd(?!? - cool stuff) 19:44, 14 February 2011 (UTC)[reply]
I feel so unappreciated. I'll take a look, though... Juliancolton (talk) 21:46, 14 February 2011 (UTC)[reply]
The full colour version is on the NOAA's website: see the external ref. Tony (talk) 23:44, 14 February 2011 (UTC)[reply]
Ah, I see. They changed it from when I had last looked at that page (a while ago, I admit). Swapped it. Titoxd(?!? - cool stuff) 00:44, 19 February 2011 (UTC)[reply]
  • There is a misleading comment in the lead about "imprecision of the partial differential equations". There is nothing imprecise about the PDES. This section is poorly worded. The important point about chaos is explained better in the main article, but rather inappropriately in the "Ensemble" section. Poujeaux (talk) 13:49, 15 February 2011 (UTC)[reply]
    • I was hoping you were correct about this discussion lying within the partial differential article, but I don't see it. Can you point us to the passage? Or is there a different article which discusses the equation's imprecise solutions we should be linking to instead? Realized you were talking about this article's content, not the PDE article's content. The lead has been fixed. Thegreatdr (talk) 14:17, 15 February 2011 (UTC)[reply]
  • Thanks, that revision is a vast improvement. Although it is good, I still don't feel that the article quite merits FA status, but I don't feel sufficiently strongly to write 'oppose' in bold. The structure is a bit odd, for example ensembles coming up as a subsection of history and then as a separate section. The balance doesn't look right, with too much history in relation to the other topics. Also, starting with the history means that some subjects get explained twice, for example initial conditions and primitive equations. Poujeaux (talk) 13:31, 16 February 2011 (UTC)[reply]
    • All the ensemble information now lies within the ensemble section per your comments. I created a new subarticle named History of numerical weather prediction which covers all the historic content previously within this article, which allowed me to shorten the historical content in this article significantly. Efforts were made in this process to minimize duplication within the article, per your comments. See what you think. Thegreatdr (talk) 15:17, 16 February 2011 (UTC)[reply]

I disagree with there being a split article for the history. It sounds like a content fork. There were no problems with the article length. --♫ Hurricanehink (talk) 19:19, 16 February 2011 (UTC)[reply]

I sort of agree with Hurricanehink. I'm not sure the section needed to be split off, when rearranging the article would suffice. Titoxd(?!? - cool stuff) 19:21, 16 February 2011 (UTC)[reply]
I actually do agree with the reviewer that there was a strong bias/POV towards history in this article, which is now covered by the subarticle. It's not a content fork because we replaced a long section with a short summary: the two are not in disagreement. We've done this numerous times before within the TC and met projects during GAN and FAC. During FAC, this was done during the tropical cyclone and surface weather analysis reviews, so it's not exactly precedent setting. We'll wait to see what Poujeaux thinks about the change. It does reduce redundancy. Thegreatdr (talk) 21:56, 16 February 2011 (UTC)[reply]
We could have a subarticle on the history, but the historical summary at this point seems a bit short. Maybe bringing in a bit more of the previous content (e.g. the explanation of the different kinds of models introduced, linking to Atmospheric model#Types) would satisfy everyone's concerns. Titoxd(?!? - cool stuff) 22:05, 16 February 2011 (UTC)[reply]
I think what you've brought back so far is perfect. I'd hesitate bringing more back before getting feedback from the reviewer who brought this up. They indicated that previously, the history section led to "some subjects getting explained twice." Thegreatdr (talk) 23:20, 16 February 2011 (UTC)[reply]
The feedback was positive, so I think we have the right proportion of history within the article now. As for inclusion of model types, we could do that, but then we'd need to explore what types of models are used within the ocean and air quality, to see if we included them all. I'm going to wikilink to Types of atmospheric models in the See Also section. Thegreatdr (talk) 15:23, 17 February 2011 (UTC)[reply]
To me, the balance and structure of the article look really good now. At the moment you have a glitch with duplication of the last para of the lead and the last para of the history section. Poujeaux (talk) 15:28, 17 February 2011 (UTC)[reply]
I can't think of a way of avoiding that, although one of the other editors might. Now that the history section is basically a lead of the History of numerical weather prediction article, we essentially have two summary sections within the article...one with dates for the history section and one without for the main numerical weather prediction article. Because the lead is a summary of the article below, there will be some duplication between the lead and article content. Maybe someone can think of a way of rewording it so it's not so similar. Thegreatdr (talk) 15:34, 17 February 2011 (UTC)[reply]
I partially rewrote that part to make it less similar. What do you guys think? Titoxd(?!? - cool stuff) 18:36, 17 February 2011 (UTC)[reply]
It is still a bit similar, but as RJH says below, you don't need the discussion of chaos in the history section. You could cut the "Even with..." and "This limitation..." sentences from the history section. Poujeaux (talk) 11:36, 18 February 2011 (UTC)[reply]
Done. Thegreatdr (talk) 18:35, 18 February 2011 (UTC)[reply]

Support—My primary concerns have been addressed and this article appears to be close enough to FA quality that I'm lending my support. Thanks for your work on this article.—RJH (talk) 21:31, 21 February 2011 (UTC) Comment—Well it's a decent article, but I have a few issues.[reply]

  • Why does the "Climate modeling" section consist of history? Doesn't this belong in the History section? Ditto for the next two sections. When I look at the "main articles" for those sections, they mostly consist of content other than history. This doesn't seem like an appropriate application of WP:SS and they may need to be re-worked.
    • Fixed. Thegreatdr (talk) 18:35, 18 February 2011 (UTC)[reply]
      • Only for the "Climate modeling" section. The same issue occurs with the "Ocean surface modeling" and "Limited area modeling" sections. Sorry, I should have made that clearer.—RJH (talk)
        • Tito has done some improvements here. The sections have a less historic feel now and more content. Thegreatdr (talk) 17:44, 20 February 2011 (UTC)[reply]
          • Are we looking at the same thing? The two sections remain almost entirely history-based. I don't mind the history being there, but the section should also have two to three paragraphs summarizing their main article links.—RJH (talk) 18:29, 21 February 2011 (UTC)[reply]
            • Three paragraphs would be overkill. I added more content to the ocean dynamics section. Titoxd(?!? - cool stuff) 19:03, 21 February 2011 (UTC)[reply]
  • In the History section, there is an explanation for why the forecast accuracy decreases with time. I'm not clear why it belongs there, rather than in the Ensembles section that covers the same topic. It seems like a side bar topic in the history, so I'd suggest a relocation.
    • Fixed. This reduces duplication of information within the article. Thegreatdr (talk) 18:35, 18 February 2011 (UTC)[reply]
  • The images don't begin with "File:" and lack alt text. But the licensing appears okay.
  • Is "statical methods" jargon, or should it be "statistical methods"?
  • The "Limited area modeling" section uses a spaced en-dash, while the remainder uses unspaced em-dashes. Please stick to one.
  • There are many instances of ', and' rather than just 'and'. In a number of cases, the command and the 'and' are redundant. Please try to re-write some of these sentences, or split them up.
    • I fixed several of these.—RJH (talk)
  • Added 2/21/2011. There are a few minor issues with the references:
    • You appear to have a number of instances where the doi lookup points to the same location as your linked article title. For example, the "2007: 50th Anniversary of Operational Numerical Weather Prediction" is linked to "journals.ametsoc.org/doi/pdf/10.1175/BAMS-88-5-639". If you go to the doi page, the PDF file has the same link. As this is redundant, it just adds to the length of the page without providing significant value.
    • Several references are linked to a doi-based address rather than using the doi. For example: "Stochastic dynamic prediction", "History of Numerical Weather Prediction at the National Meteorological Center", "Representing Model Uncertainty in Weather and Climate Prediction" and "Ensemble Forecasting at NCEP and the Breeding Method".
      • I can't get the NCEP Ensemble reference to resolve via doi's. I'm trying to fix the rest. Titoxd(?!? - cool stuff) 19:47, 21 February 2011 (UTC)[reply]
      • I am leaving the urls of the references if the url field links to the pdf, and the doi links to the abstract. I personally think that is better, but if you disagree please let me know. Titoxd(?!? - cool stuff) 20:10, 21 February 2011 (UTC)[reply]
    • Many of the google book references are bloated. Only the 'id' and 'pg' parameters are needed for a successful page lookup.

Thanks.—RJH (talk) 23:33, 17 February 2011 (UTC)[reply]

      • While that's true, previous experiences as FAC have led to such complete referencing format. If we can find all that information for the referencing, so much the better. This is the first time within the reviews I've been involved with that someone has brought up that we might reference things too well. I think I'm blushing... Thegreatdr (talk) 20:29, 21 February 2011 (UTC)[reply]
        • Like I said, it's only a minor matter. Thanks.—RJH (talk)
Yes, in fact all three of the 'applications' sections are quite historical in style. IMHO the article could do with significantly more on the current state of the art, from someone who is an expert in the field (without making it too technical). I still don't really think the article is up to FA standard. Poujeaux (talk) 11:42, 18 February 2011 (UTC)[reply]
The reason these feel historical is that they were split off from the History section per a request on the content review. Those sections definitely need revision since they were written with a different purpose in mind. I'll see what I can do, but I will have limited online access this weekend so it might take a bit of time to fix this. Titoxd(?!? - cool stuff) 00:03, 19 February 2011 (UTC)[reply]
Don't forget about Catch 22. One of the FA criteria is that the page is stable. This article has changed a lot in the last week or so as a result of this nomination! Poujeaux (talk) 13:30, 21 February 2011 (UTC)[reply]
That has more to do with edit wars and highly fluid topics, though. While this topic is about a fluid, it is not fluid per se... Titoxd(?!? - cool stuff) 18:22, 21 February 2011 (UTC)[reply]

Disambig/External Link check - 1 dab (Tellus); 0 dead external links. A few external redirects which may lead to link rot; see them with the tool in the upper right of this page. --PresN 00:27, 18 February 2011 (UTC)[reply]

I'm not sure that that dab is fixable, though. Tellus describes how there are two journals named Tellus: Tellus Series A: Dynamic Meteorology and Oceanography, and Tellus Series B: Chemical and Physical Meteorology. Both of them split from an older journal named Tellus. The reference that links to the dab page was published in the older, pre-split journal, so neither disambiguation target is correct. Titoxd(?!? - cool stuff) 01:16, 18 February 2011 (UTC)[reply]

Note for Laser brain, I have been meaning to look at this article-- as soon as I find time! Hopefully in the next few days. SandyGeorgia (Talk) 15:02, 21 February 2011 (UTC)[reply]

  • Missing DOIs as mentioned above. There should also be bibcodes for some of your cites (physics/astrophysics fields), particularly JGR, e.g. Bibcode:1996JGR...101.7419X, (use |bibcode= field) which are useful for physics articles. Have you checked arXiv too? Rjwilmsi 23:28, 21 February 2011 (UTC)[reply]
    • Added both the missing doi's, as well as bibcodes. ArXiv links seem to me to be too much, considering the bloated referencing concerns above, and that only one journal reference didn't have a bibcode (but had a doi). Titoxd(?!? - cool stuff) 19:06, 22 February 2011 (UTC)[reply]
  • Comment. I hope to find some time for reviewing this article within the next days but some immediate comments. My first thought was: Disappointing! This article is named "numerical weather prediction" and nothing about the calculation model in there! Whether prediction models are truly fascinating topics in weather prediction modeling, and essential to be described in more detail on an article about numerical weather prediction! Just mentioning the use of differential equations is meager, and that more complex differential equations can only be solved using numerical approximations is logical. What are the parameters that play a role in the numerical modeling? Temperature (atmospheric, ground, oceans), water vapor, cloud cover, terrain elevation, vegetation (esp. ice, snow, terrain moisture), surface and atmospheric winds, pressure, etc.? Total number of parameters? What are the models incorporated in the overall computation, e.g., diurnal temperature cycles, wind propagation models? How are the trajectories from the differential equations derived? What are current performance characteristics/requirements of computational models? What are the shortcomings of current models? Some quick inputs. Nageh (talk) 20:38, 23 February 2011 (UTC)[reply]
    But that is what the atmospheric model article is for. Many of the details you mention are best covered in that article, or in the article about a particular model itself. Titoxd(?!? - cool stuff) 20:41, 23 February 2011 (UTC)[reply]
I'm not expecting all the details. That would be impossible to do, even in the atmospheric model article. However, when one of the main sections of a topic is so short I do expect more details. Nageh (talk) 21:03, 23 February 2011 (UTC)[reply]
I was concerned about this problem last week, which is why the content was initially added to the Atmospheric model article. Have added that content into this article, which will likely need some additional expansion to include nonweather-related examples, such as those used in wave and air quality models. Thegreatdr (talk) 17:09, 24 February 2011 (UTC)[reply]
Cool! Looking forward to the progress. Nageh (talk) 17:37, 24 February 2011 (UTC)[reply]
I added a couple more lines dealing with ocean and air quality modeling, which forced additional upgrades to the parameterization and history of numerical weather prediction subarticles. I've also added how far a few of the global models are run into the future...I can't seem to find a reference for how long it actually takes to run them, though through daily use I'd say it's on the order of 2-3 hours for global models and one hour for mesoscale models. The quantities forecast by atmospheric models are the same from which they are initialized (outside topography). Do you think this needs to be restated? This article more than the others I've dealt with through other FAC experiences has caused significant upgrades to some of the subarticles, since they were not in very good shape (or did not exist) to begin with. Two of the subarticles have now successfully gone through the GAN process and passed, while another is currently on GAN. Neither Tito or myself were experts on this topic when we started its editing, though improving this article is changing that quickly. Is there anything significant we're missing from the parameterization section, the way it is currently written? The subarticle can deal with finer points on the topic...just want to make sure nothing major is missing for the purposes of this article. Thegreatdr (talk) 21:07, 24 February 2011 (UTC)[reply]
Sorry for it taking me so long to read the article (though I skipped the Applications section, for now). Let me say that I'm not an expert either. However, missing information on parametrization struck me as kinda obvious. I think the current state of the article deals fairly well with providing an introductory overview. (Nice job!) Here are a few issues that were no clear to me.
  • Section Parameterization, second paragraph: "Also, the grid size of the models is large when compared to the actual size and roughness of clouds and topography." This sentence looks somewhat misplaced to me. You were talking about clouds in the paragraph before. Can we remove the sentence?
  • Section Domains, second paragraph: "...become dependent variables, greatly simplifying the primitive equations.". Dependent variables doesn't immediately sound like the equations becoming simpler to me... but if the source says so.
  • "This follows since pressure decreases...". This? The correlation between the coordinate systems?
  • "This coordinate system receives that name since the independent variable σ is used to represent a pressure level (p) scaled with the surface pressure (p0) and in some cases the pressure at the top of the domain (pT).". Can you reword this sentence and avoid unnecessary variable names? "This coordinate system receives its name from the independent variable σ used to represent a constant pressure value." (If this is what is meant.)
  • Section Ensembles, paragraph "In a single model-based approach...": In the middle it says "It is common for the ensemble spread to be too small to include the weather that actually occurs, which can lead to a misdiagnosis of model uncertainty;". By whom? By forecasters, I assume.
  • "A spread-skill relationship sometimes exists, as spread-error correlations are normally less than 0.6.". Less than 0.6 doesn't sound like there is a relationship.

Nageh (talk) 21:45, 4 March 2011 (UTC)[reply]

Excellent! I hope I get to review the Applications section, too, within the next few days (but no promise). Best, Nageh (talk) 23:31, 4 March 2011 (UTC)[reply]
I gave it a quick read. It seems mostly fine. However, the models cited under section Wildfire modeling are very U.S. centric. Here is a link to an excellent and recent (2009) survey on wildfire models: Title: A review of wildland fire spread modelling, 1990-present, 1: Physical and quasi-physical models Nageh (talk) 15:35, 5 March 2011 (UTC)[reply]
I essentially wrote a new introductory paragraph based off that source, and now I'm not entirely sure we need the laundry list of models in the second paragraph of Numerical weather prediction#Wildfire modeling anymore. What do you think? Titoxd(?!? - cool stuff) 20:30, 5 March 2011 (UTC)[reply]
I agree. Did you want to wait for someone else's input before making the change? It seems like that would resolve the latest comments made. Thegreatdr (talk) 20:38, 5 March 2011 (UTC)[reply]
If you think it's ok to ax it then that's fine with me. That said, I'm not entirely sure how much of the second paragraph to remove, though. Pretty much everything after "These models include NCAR's Coupled Atmosphere-Wildland Fire-Environment (CAWFE)" seems redundant to me, but I'm not 100% sure whether the first two sentences of that paragraph should get attached as an introduction to the third paragraph. Titoxd(?!? - cool stuff) 20:50, 5 March 2011 (UTC)[reply]
I would also think so. If some of the models are particularly notable you may single them out but otherwise I'd leave them out. More detail can be provided in the main wildfire modeling article. Nageh (talk) 20:55, 5 March 2011 (UTC)[reply]
Nuked. How do you think it looks now? Titoxd(?!? - cool stuff) 21:09, 5 March 2011 (UTC)[reply]
I think it's fine now. Nageh (talk) 22:48, 5 March 2011 (UTC)[reply]
  • Support. I think the article provides a reasonably good overview on the topic now, so I think I can support it. Nageh (talk) 22:53, 5 March 2011 (UTC)[reply]
  • Am I the only one who thinks the image atop the page is too large?
  • Link to deposition is a dab page link. Any more of these? ... a second and final one at Tellus. Did anyone use the dab page tool?
  • "to determine its transport and diffusion" ambiguous.
  • what's a "spectral wave transport equation"? – Peacock.Lane 14:14, 25 February 2011 (UTC)[reply]
    • The issues you brought up should be mostly dealt with. The issue with Tellus is mentioned and discussed further up the review. This is one case where a dab may be the only reasonable link available. Thegreatdr (talk) 16:24, 25 February 2011 (UTC)[reply]
      • You wrote "The spectral wave transport equation" instead of "A spectral wave transport equation". Is there only one that is universally used? If so, the text as it stands now is a bit jarring. There needs to be a bit more cohesion b/w the spectral wave transport equation and its later relevance to numerical weather prediction, which is deferred until several phrases later. I will look at it and try to think of a suggestion.
        • As far as I can tell, yes, there is one spectral wave transport equation. Thegreatdr (talk) 01:16, 26 February 2011 (UTC)[reply]
      • Is there a typo in the doi of Bender, Leslie C. (January 1996)?
      • Why is a GCM referred to as a "code"? That word is not used elsewhere in this. Can we substitute "program" or "computer model", or (even easier) simply delete the word "code"? If that word is necessary, can we briefly define it, or...? – Peacock.Lane 01:19, 26 February 2011 (UTC)[reply]
        • Fixed. Used the term computer program, as you suggested. Thegreatdr (talk) 01:23, 26 February 2011 (UTC)[reply]
      • "can be configured for both weather and climate" I'm assuming this means they can be configured to make either a weather model or a climate model? Or does it mean that both weather and climate are plugged in to determine... weather? – Peacock.Lane 01:37, 26 February 2011 (UTC)[reply]
        • It means the same equations are used in both the short term weather forecasting and longer term climate models. It makes sense...it's not like the equations of motion used within the atmosphere change with time. Thegreatdr (talk) 01:43, 26 February 2011 (UTC)[reply]
      • So can we change "both...and" to "either...or"? – Peacock.Lane 01:55, 26 February 2011 (UTC)[reply]
      • Should the wikilink to deposition (chemistry) instead be to Deposition (aerosol physics)?
        • Either works. I switched to your link because that article is better developed than the chemistry one. =) Thegreatdr (talk) 02:52, 26 February 2011 (UTC)[reply]
      • "When ensemble spread is small..." and then "When many different forecast..". Do we have a problem with paragraph unity here? Should this be two paragraphs? Don't chop it up just because I say so. If they belong together in the same para, then the connection needs to be made a bit more clear.
        • Making it two paragraphs would look too "stubby". I've reworded a couple lines of that 4th paragraph, which should read better. Thegreatdr (talk) 04:39, 26 February 2011 (UTC)[reply]
      • Stochastic or "ensemble"... surely these are not synonyms?
      • Were Epstein's Monte Carlo simulations also a form of ensemble forecasting? If not, how were they different? ... actually, the more I think about it, the more I wonder if the "Epstein" paragraph (history) might perhaps simply be deleted, with perhaps one or two stray sentences relocated to the first para... thoughts..? – Peacock.Lane 04:31, 26 February 2011 (UTC)[reply]
        • I want to wait to hear from Tito on this point before removing any additional historical info from this page, to limit reversions. Thegreatdr (talk) 04:39, 26 February 2011 (UTC)[reply]
      • I rewrote the Ensemble section to improve coherence; see version temporarily in my sandbox. Also note it's a bit fuzzy at times because I have no domain knowledge; see mildly weasel-like words such as "combine" and "analyze". I'm also aware that 1) It's 5 paragraphs long (ask me if I care), and 2) perhaps para 4 and 5 might be switched (?)... there are different ways to order that last bit of info about multi-model approaches. – Peacock.Lane 12:07, 26 February 2011 (UTC)[reply]
      • (edit conflict) "Stochastic dynamic model" is the terminology that Epstein used to describe what we would call an ensemble in modern times. Since an ensemble is just a collection of forecasts (realizations), the Monte Carlo model just generated a bunch of forecasts with random (stochastic) perturbations to the initial condition. So no, they are not different. Titoxd(?!? - cool stuff) 12:13, 26 February 2011 (UTC)[reply]
        • That's what I kinda surmised, but I wasn't sure. But is Epstein's terminology still commonly used? I.... suspect not. If not, then equating the terms is a bit misleading not only on that level, but because the term "stochastic" has a broader (and far more common) meaning... but look at my sandbox for more (see immediately above). – Peacock.Lane 12:17, 26 February 2011 (UTC)[reply]
          • It's not common, but it is not wrong either, since that phrase describes exactly what Epstein did. He produced a stochastic ensemble of dynamical simulations using a Monte Carlo approach. Reading the sandbox, I would flip the last two paragraphs, and leave the original wording about Lorentz (saying that he discovered chaos theory is iffy—he presented the main formulation, but there have been statements that could be considered related to chaos since the 1800's.) but it otherwise looks ok. Titoxd(?!? - cool stuff) 12:24, 26 February 2011 (UTC)[reply]
      • The phrase "stochastic, or ensemble model" is used in the FAC version of this article within a passage that is outside of the contextual scope of the Epstein passage, implying that it is a valid synonymous relationship in current usage. That's what gives me pause... and the word "discovered" for Lorenz in my sandbox doesn't mean he discovered chaos theory; it says he discovered that it applies to the specific domain of weather forecasting. But if you see any value in the sandbox version, please do edit it and move it into the article. But do also watch out for imprecise terms like "combine" (as I warned above)... I am reviewing from the bottom of the page, so alas, I have quite a bit more of reviewing to do... sorry.
        • I see what you mean now. I've reworded that sentence, and copied your sandbox to the article itself. I also tried to tighten up the prose slightly. Titoxd(?!? - cool stuff) 20:12, 26 February 2011 (UTC)[reply]
  • You need to look at the link checker. Some links seem to be behind subscription firewalls; one timed out twice on me now. – Peacock.Lane 14:14, 26 February 2011 (UTC)[reply]
    • Those all work, except the NCDC one. In fact, the whole NCDC site is down for some reason, so I am assuming that is the problem. I'll check again later to see if it came back up. Titoxd(?!? - cool stuff) 19:28, 26 February 2011 (UTC)[reply]
      • The NCDC site will be back soon. They have taken it down temporarily for pre-planned Maintenance that was advertised on their site all week.Jason Rees (talk) 02:40, 27 February 2011 (UTC)[reply]
    • "Regional models use a compatible global model for initial conditions of the edge of their domain." You mentioned the model starts at the edges of the grid, then dropped this idea. Is there a reason why you start at the edges? Does the model work inward from there, or...? – Peacock.Lane 03:58, 27 February 2011 (UTC)[reply]
      • Regional models need a global model to initialize the boundary conditions at their edge, so systems can enter their domain. The physics/equations used within the regional/LAM dominate within the grid. Thegreatdr (talk) 04:40, 27 February 2011 (UTC)[reply]
    • I didn't see that explained in the FAC, but perhaps I missed it... moreover, adding a completely new section during FAc (wildfires) is kinda irregular... – Peacock.Lane 04:00, 28 February 2011 (UTC)[reply]
      • We've already had to add another section...parameterization...due to comments made during this FAC, and split what was previously one section into three. The fire weather modeling did seem relevant...it's not something I ran across until the last several hours. It can always be removed if it's not deemed relevant. Thegreatdr (talk) 05:37, 28 February 2011 (UTC)[reply]
    • I didn't see the earlier additions; sorry. I still think the passage "Regional models use a compatible global model" is a bit sketchy (unnecessarily so), as mentioned just two posts above this one.. I also want to see if User:Nageh's objections are satisfied (but perhaps I missed that bit, above)... I'm getting close to being finished with this FAc, and close to supporting. I do want to look at the lede, and... a bit more.. uh the relevance of the image captioned "A cross-section of the atmosphere over terrain " isn't really made explicit.
      • The image is a vertical cross-section of the domain used in high-resolution models, and shows how they deal with terrain. I tried to clarify that in the article. How does it look now? Titoxd(?!? - cool stuff) 21:25, 28 February 2011 (UTC)[reply]
    • MOS described as "Post-processing" in lede but not body of article. Could you explain? Does this need to be in the MOS section?
    • lede talks about "it was not until the advent of computer simulation in the late 1940s"; body talks about ENIAC in the 50s. – Peacock.Lane 11:33, 28 February 2011 (UTC)[reply]
      • I've made an attempt to resolve the regional model, MOS, and 1940s/1950s issue. See if this new wording is better. Thegreatdr (talk) 15:33, 28 February 2011 (UTC)[reply]
      • I've rewritten the lede, temporarily at User:Peacock.Lane/Sandbox]. Please, please, please, please read it very slowly and carefully! My biggest fear is that in a quest for brevity I may somehow have combined ideas that are categorically distinct... I also don't really see the connection between this article and wildfire modelling (in particular, neither your lede nor mine explains that at all), and I do not know the meaning of guidance ("used in the guidance") in the lede. – Peacock.Lane 01:48, 1 March 2011 (UTC)[reply]
          • I put your lede in the article and made a few changes to it, and I addressed the "guidance" bit, which is just weather jargon for a forecast/model. As for the wildfire section, the section points out that some weather models (such as WRF) are coupled with combustion codes to produce wildfire spreading simulations, but that resolving all the equations involved in combined combustion and weather processes takes us firmly into DNS land. I'll see if I can clarify that further. Titoxd(?!? - cool stuff) 02:25, 1 March 2011 (UTC)[reply]
      • "parameterizations for turbulent diffusion, radiation, moist processes (clouds and precipitation), heat exchange, soil, vegetation, surface water, and the kinematic effects of terrain". It's not clear to me that all of the things listed in that sentence are discussed/mention in the parameterization section. Perhaps some bits are paraphrased, or perhaps I missed something.. "Radiation"... umm.. the default mental image is of nuclear radiation; can we add an adjective? I also don't see the word "kinematic" elsewhere. – Peacock.Lane 03:53, 1 March 2011 (UTC)[reply]
        • Everything except turbulent diffusion was, so removed that bit. Added the adjective solar before radiation so no one confuses it with the "nucular" kind. =) Thegreatdr (talk) 00:48, 2 March 2011 (UTC)[reply]

Images

  • Caption on lead image: "Winds, etc...evaluate interactions with neighboring points."? That doesn't make sense to me as written - can you reword? In general, captions should be grammatically correct and clear, and sourced where necessary
  • Why not use {{PD-USGov-NOAA}} for NOAA images, other more specific tags for US government images where such tags exist?
  • File:Two_women_operating_ENIAC.gif is tagged as lacking author information
  • File:NOAA_Wavewatch_III_Sample_Forecast.gif - any more specific source information available, or a link? Nikkimaria (talk) 04:03, 2 March 2011 (UTC)[reply]
  • Conditional Support. Consider this a "Support" as soon as the image issues are resolved. Unwatching; ping me if needed. – Peacock.Lane 04:18, 2 March 2011 (UTC)[reply]
  • Support. Even though this is not quite my area of expertise (even though I have a BSc in physics and I have taken courses in computational physics, fluid dynamics, and introductory meteorology), there's nothing which sounds wrong to me, and it is written in such a way that an intelligent layman (say, someone in the last year of high school) would understand at least all the most important points. —A. di M. (talk) 17:11, 6 March 2011 (UTC)[reply]

Comments

  • "as input to initialize their models" seems slightly technical for the lead section.
    Rephrased. Titoxd(?!? - cool stuff) 19:47, 14 March 2011 (UTC)[reply]
  • "Similar mathematical models" - it might be just me, but when I see "similar" used, I think it refers back to something mentioned in the previous sentence, whereas here it seems to be saying that the models used for short-term and long-term forecasts are similar. I would be inclined to drop the word 'similar' altogether.
    Rephrased. Titoxd(?!? - cool stuff) 19:47, 14 March 2011 (UTC)[reply]
  • The 'moist processes' wikilink - the parenthetical 'clouds and precipitation' help somewhat, but is there more to it than that? Also, the links in that list are profoundly unsatisfying. Radiation, heat transfer, soil and terrain all told me nothing about weather prediction. Solar radiation would be a better link than radiation. I would delink clouds, soil and terrain. I also note that the image caption links radiation rather than solar radiation, and that it uses links to wind (not mentioned in the lead) and relative humidity and hydrology, all of which I found more informative than the text in the lead.
    I've delinked those. The image notes the outputs and the inputs of a numerical weather model; the text discusses only the things that need to be treated as parameters. Titoxd(?!? - cool stuff) 19:47, 14 March 2011 (UTC)[reply]
  • Giving nationality of the researchers mentioned would give useful context. e.g. Lewis Fry Richardson, Vilhelm Bjerknes, Norman Phillips
    This was originally in the article but was removed when the section got shipped off to the History subarticle and summarized. I am not sure whether others would consider this "bringing back content", as I was asked not to do previously in the FAC, and the value of such info is marginal (that can be verified easily from the biographical articles) so I won't do it. Titoxd(?!? - cool stuff) 19:47, 14 March 2011 (UTC)[reply]
    I disagree. Nationality of researchers is useful context, IMO, but I won't press the point here. Carcharoth (talk) 06:53, 23 March 2011 (UTC)[reply]
  • Should ENIAC and NOAA be given in full the first time they appear in the text, or are these initialisms commonly enough known not to need to do that?
    NOAA is common enough for it to not need to be spelled out. ENIAC's common name is ENIAC, in spite of it actually being called the Electronic Numerical Integrator And Computer, so it should also not need to be spelled out. Titoxd(?!? - cool stuff) 19:47, 14 March 2011 (UTC)[reply]
    As long as both are linked the first time they appear (I haven't checked). Carcharoth (talk) 06:53, 23 March 2011 (UTC)[reply]
  • "more physical processes" - I had to hover over this link to find out where it was going to take me. I think the way such links are made should give the reader some idea of what is being linked to.
    Rephrased. Titoxd(?!? - cool stuff) 19:47, 14 March 2011 (UTC)[reply]
  • "The development of limited area (regional) models facilitated advances in forecasting the tracks of tropical cyclone as well as air quality in the 1970s and 1980s." - cyclone should be plural, I think.
    That used to read "tropical cyclone tracks" but somehow somebody forgot to change the plural when they rearranged it. In any case, it's fixed. Titoxd(?!? - cool stuff) 19:47, 14 March 2011 (UTC)[reply]
  • "efforts began to initialize and model soil and vegetation types" - again, this word 'initilize' has cropped up. It might be explained later in the article, but I think it is a place where people will stumble. I think it means to input the starting conditions, but would have to finish reading the article to be sure.
    You are correct as to the meaning; however, I do not think that word is particularly foreign, and due to initialization being a specific technical process, I would be opposed to changing it. Titoxd(?!? - cool stuff) 19:47, 14 March 2011 (UTC)[reply]
  • "Sea ice began to be initialized" - again with the word 'initialize' - in fact the whole section is called 'Initialization', with no explanation of what this is.
    It was covered in the middle of the paragraph. I moved a sentence around to try to make this clearer. Titoxd(?!? - cool stuff) 19:47, 14 March 2011 (UTC)[reply]
  • "Essentially, an atmospheric model is" - I believe that manuals of style discourage the use of words like 'essentially' and 'basically' as too colloquial and informal. I think it should be dropped here, as the tone it promotes is of someone thinking 'how can I simplify things to explain this to the reader?' Just starting the section as "An atmospheric model is" should be adequate.
    Done. Titoxd(?!? - cool stuff) 19:47, 14 March 2011 (UTC)[reply]
  • "potential temperature scalar fields" - should there be a comma between temperature and scalar?
    No. the distribution of potential temperature is a scalar field. Titoxd(?!? - cool stuff) 19:47, 14 March 2011 (UTC)[reply]
  • "velocity vector field" - this appears to be an attempt to explain to a reader what the term 'velocity vector field' means by directing them to read two different articles and hoping they can work it out. Possible 'potential temperature scalar fields' is the same thing. I much prefer the constructions: 'scalar fields dealing with potential temperature' and 'vector fields dealing with velocity' (but velocity of what? wind velocity?).
    Air velocity. However, explaining scalar/vector fields is well outside the scope of this article. Besides, "x vector field" and "y scalar field" are the ways these things are referred to in the literature (this is especially true of the velocity vector field). As such, I don't think they should be reworded, as that might introduce ambiguity for technical readers. Titoxd(?!? - cool stuff) 19:47, 14 March 2011 (UTC)[reply]
  • "then it would be overturned, and the air in that vertical column mixed" - is it possible to link or explain 'overturned' here?
    Done. Titoxd(?!? - cool stuff) 19:47, 14 March 2011 (UTC)[reply]
  • "The process that determines the amount of solar radiation reaching ground level in rugged terrain or due to variable cloudiness occurs on the molecular scale" - are you sure 'molecular scale' is correct here? By this, do you mean absorption and re-radiation of solar radiation as it passes down or is scattered in the atmosphere? That might be more informative than 'molecular scale'. If not this, then how can rugged terrain and clouds be 'molecular scale'?
    Both. Clarified that in the prose. Titoxd(?!? - cool stuff) 19:47, 14 March 2011 (UTC)[reply]
  • "The horizontal domain of a model is either" - it might be an idea to first explain what a domain is, or to rephrase to have the explanation before going into details.
    Linked to Domain of a function. Titoxd(?!? - cool stuff) 19:47, 14 March 2011 (UTC)[reply]
  • "Richardson's 1922 model" - it has been a long time since the reader read about Richardson in the 'History' section. Maybe give his full name again here?
    Done. Titoxd(?!? - cool stuff) 19:47, 14 March 2011 (UTC)[reply]
  • I couldn't work out why, but I felt a bit lost in the 'Domains' section. The other sections were fine, but maybe this one could be polished up a bit more? Possibly what is missing is an introductory bit to the section explaining 2D and 3D models?
    But the section is the introductory bit to 2D and 3D models. I am not sure how to address this. Titoxd(?!? - cool stuff) 19:47, 14 March 2011 (UTC)[reply]
  • 'Model output statistics' - this section feels a bit short and underdeveloped (e.g. no dates provided here). I didn't learn much more here than had been stated elsewhere in the article. The second paragraph feels particularly poorly written. FWIW, it is only these two sections (this one and 'Domains') that I had problems with - the rest of the article seemed fine (apart from the very last section on wildfire modeling).
    I added the date. As for the second paragraph, please be more specific. Saying it sucks doesn't really help me figure out why or how to fix it. Titoxd(?!? - cool stuff) 19:47, 14 March 2011 (UTC)[reply]
    Possibly too much use of "MOS" as an abbreviation - some added variety is possible in the writing. Also, maybe review the main article that is linked, and see if more of a fuller summary can be brought over. Carcharoth (talk) 06:53, 23 March 2011 (UTC)[reply]
  • Nationalities of Epstein and Lorenz and Leith could be given.
    Same as above. Omitted to be consistent. Titoxd(?!? - cool stuff) 19:47, 14 March 2011 (UTC)[reply]
    As above, still disagree, but not worth sticking over this. Left unstruck so other reviewers can consider this point. Carcharoth (talk) 06:53, 23 March 2011 (UTC)[reply]
  • "Cecil Leith revealed that they produced" - revealed sounds wrong here.
    Fixed. Titoxd(?!? - cool stuff) 19:47, 14 March 2011 (UTC)[reply]
  • "Ensemble forecasting is an approach used since the 1990s" - this is confusing as you have just (in the previous paragraph) said that Epstein "proposed using an ensemble. That is an early example, as you say, but the text fails to move from the early history to the modern techniques - the jump is too sudden. Something like 'Following these early efforts, the modern techniques was developed by X in YYYY...", that sort of thing.
    What the article is trying to say is that only in the 1990s ensembles began to be used for actual forecasts. Clarified that in the prose. Titoxd(?!? - cool stuff) 19:47, 14 March 2011 (UTC)[reply]
  • "normally less than 0.6, and only under special circumstances range between 0.6–0.7" - it would help to say what scale this is - is it a probability from 0 to 1? Or a correlation value between 0 and 1? i.e. How good or bad is 0.6-0.7? It would also be nice to have a non-technical bit saying how good or bad weather forecasts are (certainly among the general public, I think the perception is still that weather forecasts are sometimes very unreliable and sometimes adequate - but this tends to be only hearsay and remembering whether last night's weather forecast was accurate or not).
    What you are describing is forecast skill; that is the next sentence. As for the first part, it is a correlation (and the previous sentence mentions that), 1 is perfect, and 0.6 is "meh". Thus the article says that there is a weak or missing spread-skill correlation. Note that all of this info is exclusive to ensembles. Titoxd(?!? - cool stuff) 19:47, 14 March 2011 (UTC)[reply]
  • "The relationship between ensemble spread and skill varies substantially" - suggest adding 'forecast' before 'skill' (i.e. wikilink as a whole phrase rather than piping behind 'skill').
    Fixed. Titoxd(?!? - cool stuff) 19:47, 14 March 2011 (UTC)[reply]
  • "In the same way that many forecasts from a single model can be used to form an ensemble, multiple models may also be combined to produce an ensemble forecast. Multi-model ensemble forecasting is an approach which uses many different forecast models to generate a forecast." - these two sentences appear to say the same thing.
    Fixed. Titoxd(?!? - cool stuff) 19:47, 14 March 2011 (UTC)[reply]
  • "a process known as super ensemble forecasting. This type of a forecast significantly reduces errors in model output." - copyediting needed here: hyphenated 'super-ensemble forecasting'? Remove the superfluous 'a' before 'forecast'.
    I have always seen it written as one word, so combined it. Removed the stray 'a'. Titoxd(?!? - cool stuff) 19:47, 14 March 2011 (UTC)[reply]
  • 'El Niño-Southern Oscillation' - is there a reason this is not linked? Also 'study its forcings' sounds strange unless you explain what a 'forcing' is.
    Linked both. Titoxd(?!? - cool stuff) 19:47, 14 March 2011 (UTC)[reply]
  • "decade to century time scale" - hyphenation needed somewhere here.
    Reworded the whole sentence. Titoxd(?!? - cool stuff) 19:47, 14 March 2011 (UTC)[reply]
  • "originally created by Syukuro Manabe and Kirk Bryan" - can you give the year here? Also, could give nationalities, though giving the location of the research institute (as you do) is good enough.
    Added the date. See above for nationalities, trying to stay consistent. Titoxd(?!? - cool stuff) 19:47, 14 March 2011 (UTC)[reply]
  • "smaller scale interactions" - hyphenation
    Done. Titoxd(?!? - cool stuff) 19:47, 14 March 2011 (UTC)[reply]
  • "wind blowing over the surface of an ocean and ocean's upper layer" - missing 'the'?
    Fixed. Titoxd(?!? - cool stuff) 19:47, 14 March 2011 (UTC)[reply]
  • 'spectral wave transport equation' - should this be a red-link?
    Yep, fixed. Titoxd(?!? - cool stuff) 19:47, 14 March 2011 (UTC)[reply]
  • "until the decade of the 1980s" - this is strange phrasing - it should be automatic to remove 'the decade of' from this phrase. Could you check to see if this too closely paraphrases the source?
    That was just me using a Spanish set phrase in an English context. The paraphrasing should be ok, considering that one paragraph is condensed into one sentence. Titoxd(?!? - cool stuff) 19:47, 14 March 2011 (UTC)[reply]
  • "showed skill" - again, give the full term 'forecast skill' (as commented previously)
    "Show skill" is actually a very common phrase in the literature, so I am hesitant to change it. Titoxd(?!? - cool stuff) 19:47, 14 March 2011 (UTC)[reply]
  • "statistical methods continue to show higher skill over dynamical guidance" - would it be asking too much to drop the technical language and say something simple like 'statistical methods continue provide better results than dynamical methods'?
    But doing so would cause problems for readers reading the references, which exclusively use the skill terminology, and "results" is too weasely of a word. Forecast skill has a particular definition and replacing it would introduce unneeded ambiguity. As such, we just used terms defined and used throughout the article. Titoxd(?!? - cool stuff) 19:47, 14 March 2011 (UTC)[reply]
  • 'Wildfire modeling' section feels like it is pitched too high - it is very technical and hard to understand.
    Again, this is too vague to be useful. What concepts are too technical? Titoxd(?!? - cool stuff) 19:47, 14 March 2011 (UTC)[reply]
  • Are these five examples the only applications of numerical weather prediction, or are they five examples?
    Only five examples, although they are rather main areas of investigation (aside from the obvious one, weather forecasting). Titoxd(?!? - cool stuff) 19:47, 14 March 2011 (UTC)[reply]
    Is it possible to give other examples? Carcharoth (talk) 06:53, 23 March 2011 (UTC)[reply]
  • Article might benefit from a closing section to summarise recent advances and future directions of research. Finishing on wildfires because that is last alphabetically under 'Applications' feels unsatisfying. Particularly as the language used in that section is particularly technical and full of jargon.
    That would be very close to WP:CRYSTAL for my comfort. The latest advances have really been in ensemble forecasting. Titoxd(?!? - cool stuff) 19:47, 14 March 2011 (UTC)[reply]
  • The 'See also' links to Frederick Gale Shuman and André Robert should be removed after being incorporated into the article. Alternatively, they should be annotated to explain why they have been provided for the reader.
    Moved to History of numerical weather prediction, where they fit better. Titoxd(?!? - cool stuff) 19:47, 14 March 2011 (UTC)[reply]

Overall, I enjoyed reading the article and think I learnt something. I did get a bit lost at times, though. Hopefully the above comments will help. I did like the images used, they were very helpful to look at. Carcharoth (talk) 03:57, 13 March 2011 (UTC)[reply]

Thanks for the replies. I've had time to look at the changes, but not time to review and strike the points I made above. I do think that you have met most of the points raised, though, so I'm going to enter a limited support until I've had time to re-read the article again. It is limited because I do think some of the newer sections still need a shake-down to make them more accessible and less technical. I'll try and say exactly what the most difficult parts of the article are, though I've mentioned the last secton already. Part of the reason for wanting a closing section was not wanting it to end on such a fiercely technical note as the closing sentences of the wildfire section. I'll quote them here:

"The cost of added complexity is a corresponding increase in computational processing, so much so that a full three-dimensional explicit treatment of combustion in wildland fuels by direct numerical simulation (DNS) at scales relevant for atmospheric modeling is not currently practical because of the limited skill of weather models at spatial resolution under 1 kilometer (0.6 mi). Consequently, even these more complex models parameterize the fire in order to calculate local fire spread rates using fire-modified local winds. Although models such as Los Alamos' FIRETEC carry prognostic conservation equations for the reacting fuel and oxygen concentrations, the computational grid cannot be fine enough to resolve the reaction rate-limiting mixing of fuel and oxygen, so approximations must be made concerning the subgrid-scale temperature distribution or the combustion reaction rates themselves."

The placing of references in those three sentences is a bit strange as well. But my main point is that most readers will feel lost after reading that, and it will drive out what they read earlier in the article. Better, in my opinion, to finish on a note that reminds readers what numerical weather prediction is - one technique used is to find a recent good quote from someone authoritative to sum things up. If that is too difficult to do, or too far outside the style of the article, then why not end with a summary of the five applications mentioned and other major or minor applications? That summary would usually go at the beginning of the 'applications' section, but there would be good reasons to shift it to the end here, to provide a readable finishing point for the reader. Carcharoth (talk) 07:10, 17 March 2011 (UTC)[reply]
  • While I see your point, wikipedia article structure allows for a summary at the beginning of the article (the lead), not the end. I have made some edits to the wildfire section to try to make it more understandable to the lay person, per your comments. Thegreatdr (talk) 16:58, 17 March 2011 (UTC)[reply]
    • The changes help a bit, but you have misunderstood what I said. I was saying put a summary of the applications section at the end of the article. Currently, there is no summary of the applications section at the beginning of that section - you are just launched straight into five sections detailing those five applications. And regarding the technical nature of that section, see the change made here. It wasn't until that change was made that I realised that slope and terrain are the same thing. Given that the wildfire section was only added during this FAC, I'm not comfortably moving from my limited support until some of the earlier reviewers have had a chance to look at this. Carcharoth (talk) 05:36, 18 March 2011 (UTC)[reply]
      • I have made further changes to that section to try to make it more accessible. That said, I still don't think that adding the summary of the applications section at the end of the article would be a good idea. It would be an unorthodox article structure (why not put the summary at the start of the section, like all the other articles do?), and it would ultimately be redundant, as it would summarize the things that were just being said. Titoxd(?!? - cool stuff) 09:27, 22 March 2011 (UTC)[reply]
        • The changes help, and I'll strike my objections above. Will have to accept that this is a very technical subject. Overall, the applications section does now seem to have more on wildfire modelling than anything else. Possibly because of the changes I suggested (sorry!). Maybe cut down the wildfire section or expand the other sections? Carcharoth (talk) 06:53, 23 March 2011 (UTC)[reply]
  • Support - following further discussion above, happy to enter a full support (and have unbolded my limited support above). Carcharoth (talk) 06:53, 23 March 2011 (UTC)[reply]

Retrieval dates aren't needed on journal articles, but no need to remove them since they're there. Are all those External links needed? SandyGeorgia (Talk) 18:38, 26 March 2011 (UTC)[reply]

I've pared them back, and moved half of those links to the Atmospheric model article, where they're better placed. Also fixed some intervening problems which have cropped up since your edit. Thegreatdr (talk) 22:10, 26 March 2011 (UTC)[reply]

Very nice, but please review throughout for WP:MOSDAT##Precise language (currently, etc.), check for redundancies of still vs. however, etc. SandyGeorgia (Talk) 18:53, 26 March 2011 (UTC)[reply]

Occurrences of the words, still, however, and moreover have been eliminated. Thegreatdr (talk) 22:18, 26 March 2011 (UTC)[reply]
The above discussion is preserved as an archive. Please do not modify it. No further edits should be made to this page.
Retrieved from "https://en.wikipedia.org/w/index.php?title=Wikipedia:Featured_article_candidates/Numerical_weather_prediction/archive1&oldid=1138468111"