Wikiwand AI

Talk:Climate change/Archive 94

From Wikipedia, the free encyclopedia

Archive 90Archive 92Archive 93Archive 94Archive 95Archive 96Archive 97

Chart for global warming since the industrial revolution

Chapter "Warming since the Industrial Revolution" could show a chart that shows global warming since the industrial revolution:

The chart shows decades of daily temperature anomalies against the preindustrial level. Temperature increases in time, leading to a stream of new warmth records.

  • Old years are blue. These are the coldest years at the bottom.
  • Recent years are red. These are the hottest year at the top.
  • 2023 shows as an exceptional hot year, breaking a lot of warmth records since June, months above +1.5C since Sep and even peaking above +2.0C in November.

A newer version is available which highlights the +1.5 and +2.0C level. Uwappa (talk) 17:49, 25 January 2024 (UTC)

@Uwappa: I love this Copernicus approach—for dramatizing temperature extremes in one year. However, this article already has longstanding incumbents File:Global Temperature And Forces With Fahrenheit.svg and File:Common Era Temperature.svg which show longer-term temperatures in greater perspective. The Copernicus focus approaches a violation of WP:NOTNEWS, and in this high-level article, the greater perspective is most appropriate. —RCraig09 (talk) 19:08, 25 January 2024 (UTC)
Yes, I know the drill.
2023 was just one extreme year. It was just one year, it may cool down in years to come. Nothing to worry about until the IPCC reports a multi decade average crossing the +1.5 and +2.0C limits. Climate is long term, so let us happily ignore the alarming warming of 2023 and satisfy for charts such as and with temperatures till 2020 .
A chart with rising temperatures of the last 80+ years would be way too recent. We should ignore https://climate.copernicus.eu/copernicus-2023-hottest-year-record describing
  • high sea surface temperatures smashing previous record
  • record low sea ice
  • an average global +1.48°C, hottest year on record
  • record number of days above the +1.5°C limit
  • first peak above +2.0°C.
Because... that would be news. Uwappa (talk) 05:18, 26 January 2024 (UTC)
Certainly 2023 temperatures are news worthy, but they are not yet significant to the overall story of climate change. Depressed temperatures the few years prior to 2023 made 2023 much more likely. The IPCC uses a 20 year average to estimate temperatures, because year to year there are significant variations. We need to be careful not to be alarmist and cherry pick data- remember that "an inconvenient truth" spent a ton of time focused on the recent active hurricane season, but that was an outlier and really discredits the film in retrospect. Put another way, would you have advocated for a version of this chart in 2022, prior to last year's temperature spike?
It's a good point that the graphs we are using should be updated though. I'll look to do that. Efbrazil (talk) 18:18, 26 January 2024 (UTC)
Yes, I would have advocated this graph in 2022, had I known of its existence at the time. The graph is a design marvel:
  • 80+ years of daily global temperature anomalies. That is a massive amount of data in one chart. It is an excellent design.
  • despite the massive amount of data, still easy on the eye. The bands of red, white, blue are easy to see. Old years are blue, those are the years at the cool bottom. Recent years are red, those are the years at the warm top. It is very easy to see that temperatures have been going up the last 80+ years. It truly shows decades of global warming in a split second.
  • Vertically the days of one month can be compared, e.g. blue Januaries of the 1940s against red Januaries of the 2020s. It is easy to see that it is not just one particular season that warms up, it is all months, all year round.
  • And yes it shows that 2023 was an exceptionally hot year as it jumps out of bandwidth of 'normal' warming. Time will tell if this was an 'just' and exception or the first year of accelerated warming. Please read the section "A warning for 2024 and beyond at BBC" with these quotes:
"It raises the possibility that 2024 may even surpass the key 1.5C warming threshold across the entire calendar year for the first time, according to the UK Met Office." and
"The year 2024 could be warmer than 2023 - as some of the record ocean surface heat escapes into the atmosphere"
"it highlights the concerning direction of travel, with each hot year bringing the world closer to passing 1.5°C over the longer term".
  • It is very significant to the overall story of climate change to approach the 1.5°C and 2.0°C limits. Those limits are key elements of the Paris agreement for a reason, see chapter Tipping points and long-term impacts. The chart shows that daily temperatures already crossed +1.5°C and even +2.0°C. Yes, that is 'just' daily temperatures, 'just' a warning, not a confirmed multi decade average yet. Yet effects of temperatures at +1.5°C won't wait for a multi decade average. In 2023 some short term effects were felt already.
For now I suggest to update the chapter "Warming since the Industrial Revolution":
  1. insert the the chart and describe that it shows 80+ years of global warming, with the pre industrial level as base line. Suggestion: create an image gallery after the first paragraph with , and .
  2. update the second paragraph "Multiple independent instrumental datasets show ... are rising by about 0.2 °C per decade", mentioning the +1.48°C average of 2023. Describe that 2023 could have been an exception or the start of accelerated future warming. Time will tell.
The next chapter, "Future global temperatures", already describes: "...a 20 year temperature average ... it expects that 1.5 °C limit to be exceeded in the early 2030s". That is fine already. Uwappa (talk) 20:28, 26 January 2024 (UTC)
The overall trend line for temperatures already predicted 2023, as 2021 and 2022 were depressed. See this graph, which is what the update will look like:
https://data.giss.nasa.gov/gistemp/graphs_v4/graph_data/Global_Mean_Estimates_based_on_Land_and_Ocean_Data/graph.png
The lowess line is 5 year based and it actually shows temperature increases slowing down currently. The 20 year average rate of increase has been remarkably steady and well correlated with greenhouse gas concentrations, which is why the IPCC uses that. It smoothes out the el nino and la nina effects in particular.
Like Craig, I disagree with adding the annual spark line chart to the other two charts Craigs already made. As Craig said, it excessively emphasizes the most recent year. Efbrazil (talk) 20:53, 26 January 2024 (UTC)
I am looking forward to the updated and .
Is there any chart available that shows:
  • yearly global temperature anomalies
  • the 20 year averages
  • the +1.5 and +2.0 °C limits?
I have not found such a chart on Wikipedia. I hope the average will look like the straight red temperature trend line till dec 2023, based on 30 year averages. Would it be a good idea to create such a chart and show that we have not yet crossed the +1.5 °C limit? Uwappa (talk) 19:12, 27 January 2024 (UTC)
@Uwappa: Your enthusiasm is impressive, but in general I think you'll have better results here if you focus on simplicity in charts and brevity in comments/suggestions. In answer to your specific 19:12 27 Jan suggestion: combining too many ideas in one chart (mixing past values, averages of past values, projections to future values, comparisons to thresholds) also risks running afoul of WP:SYNTH... in addition to requiring frequent updates. —RCraig09 (talk) 21:34, 27 January 2024 (UTC)
Thank you. Please feel free to simplify this chart, remove future values and update it yearly:
Uwappa (talk) 10:35, 28 January 2024 (UTC)
Thanks! Regarding the data, can you comment on the source more? Nice to have it going back to 1850, but I would like to know how reliable that is. I have been pulling from the public NASA dataset instead, located here, the rebaselining to the IPCC 1850-1900 average:
https://climate.nasa.gov/vital-signs/global-temperature/
Regarding the trend line, there is an issue that the line stubs out 10 years before today, which leads to the impression that temperatures are lower than they are. I have been working on calculating a 20-year lowess smoothing line instead, which will extend to the limits of the data set. Efbrazil (talk) 17:49, 29 January 2024 (UTC)
Great, you are welcome! The straight yellow line since the 70s surprised me. Based on news reports I expected a curve going up, accelerated heating. But no, you were right with your 'remarkably steady'.
The early ending (and late starting) yellow line and continuing-till-present blue dots tell the right story: While 2023 got close to the 1.5C limit, it will take another decade before we know the average for 2023 conform the IPCC definition. I suggest to leave the chart the way it is and explain the blue-yellow year-range difference in the text. That will counter a lot of speculation currently in the news: 2024 will break the +1.5C limit.
Link to source at: commons:file:Global_temperature_anomalies_with_20_year_average.png. I used the NASA CSV for the first version of the chart, but discarded it as it misses 1850-1879 and 2023 was way off +1.48C.
Can you recommend a url of an IPCC CSV file? I'll be happy to draw a fresh chart based on an other source data though it won't change the chart much. In a previous version of the chart, a line connected the blue dots and that line was very similar to the black line in . The line was also similar to the first chart version, based on NASA data. Uwappa (talk) 20:28, 29 January 2024 (UTC)
In the interests of avoiding wasted effort, I point to the large number of charts that are already in Commons categories: Hockey_stick_temperature_graph, Global_warming_graphs and Climate_change_diagrams. I'm not seeing what would be accomplished by having another chart. Also, the possible WP:SYNTH issue I mention above, hasn't been dealt with. Separately, I think that the smoothed average, being more indicative of meaningful global warming, should be more dominant (darker) than the choppy annual values.RCraig09 (talk) 21:11, 29 January 2024 (UTC)
Well, well, well RCraig09. OK, so let me address WP:SYNTH. Be careful. You may want to skip my answer as it might rock your boat again, but anyway, here we go:
  1. WP:SYNTH is about reaching or implying a conclusion not explicitly stated by any source. The chart does not present any conclusion. It is a chart presenting data, not text describing a conclusion. The text for the image caption is yet to be defined. Any conclusion would be in the mind of the reader, not in the chart. That is similar to a reader arriving at a personal conclusion after reading a text. Я не занимаюсь контролем над разумом.
  2. WP:OI (scroll down a little bit, it is just below WP:SYNTH) encourages Wikipedians to upload own images. It makes me happy to see that the idea of dots and lines is so well received, you already proposed it for . Good cooperation, well done!
  3. There is no rule that images should be based on only one source, just as there is no rule that text of one paragraph should have only one source. An image like , with data from several sources is fine. is based on data from just one source, which renders the synthesis concern, well eh, ... Fill in the dots yourself.
  4. Computing a 20 year average is a routine calculation, primary school level. It does not qualify as original research. Please read WP:CALC.
Thank you for your comment about the dominance of the yellow line for the 20 year average. The yellow line is now blue and the dots are smaller. Is the line now sufficiently dominant (darker) for you?
Uwappa (talk) 23:41, 29 January 2024 (UTC)
The main issue was what value would such a chart add to those already in existence, and which currently-used chart it should replace. I'm hoping to avoid your doing numerous versions of a chart, if it doesn't supersede an existing graphic (which could also add various horizontal lines, including 1.5°, and 2.0°(?), and 2.7°(?) where do we stop?). You're definitely right about WP:CALC, but the SYNTH issue has to do with the purpose/meaning of adding miscellaneous dominant horizontal line(s) whose relevance must be explained somewhere: adding explanations to the chart itself clutters it, and adding to captions cannot be readily policed if it is added by multiple editors across multiple articles in multiple languages. PS is sourced to a single page. —RCraig09 (talk) 04:46, 30 January 2024 (UTC)
So you are worried about my time? That should not be your worry. I suggest you focus on spending your own time well. Your suggestions on improving the chart have been welcome. Thank you.
  • Are you worried that an inferior graphic will be replaced? Are you worried that it may replace one of your graphics? I have not made such a suggestion. Replacing inferior graphics should not be a worry. It should something to strive for.
  • Are you worried that the +1.5 and +2.0 C limits are miscellaneous, hard to explain? Really??? See 4th paragraph of the article. News about those limits is hard to miss. It seems ridiculous but I have added article 2.1.(a) of the Paris agreement as a source for the two horizontal lines.
  • Are you worried that a future caption may describe: The 20 year average is approaching the +1.5 limit of the Paris agreement? Sorry, but that caption is not there yet and it would not be a new conclusion. Feel free to write the future caption yourself.
  • Are you worried that global warming will pass +2.0C in the near future? Well, that is worry I do share, especially since Nov 2023 peaked above +2.0C already. For now I suggest we stop at 2.0C, the upper limit of the Paris agreement. Worry about adding new limits only when a new agreement sets new limits. I do not expect that to happen any time soon.
  • Are you worried that this chart may get popular in many language versions of Wikipedia? Well, I take that as a compliment. Feel free to police the caption in all languages if it does get popular. My advice: Have a bit of faith in your fellow Wikipedians.
I am not a psychologist, but I do suggest you stop worrying about future edits that might never happen. Cross a bridge when your get there.
Let us go back to the main issue:
Chapter "Warming since the Industrial Revolution" could show a chart that shows global warming since the industrial revolution.
Any new suggestions to further improve the chart? Uwappa (talk) 09:34, 30 January 2024 (UTC)

Best chart candidates

I would like to say that this needs input of other editors, but this discussion is now almost hopelessly impenetrable for anyone else. May I suggest updating the section directly above this one (Graphics/charts) with all the newly proposed charts, then moving that section to the end of talk page? It now looks as if you two have very different idea on what makes a chart good, and it would be great if other editors had more of a chance to compare all the proposals side-by-side. InformationToKnowledge (talk) 10:06, 30 January 2024 (UTC)

Go for it in a new topic! But please leave discussion above intact. Proposed candidates are:
  1. It shows the long term trend since the 1940s with an incredible amount of daily temperatures anomalies against the pre industrial level. The blue-white-red colour coding for old->recent shows the long term trend: it is getting warmer. It also highlights the recent 2023, the hottest year on record with its many days of 2023 peaking above the Paris levels.
  2. a newer version at https://climate.copernicus.eu/copernicus-november-2023-remarkable-year-continues-warmest-boreal-autumn-2023-will-be-warmest-year#85df0848-1996-45c2-aced-a5061edc6b30 It is an updated version of option 1 with data till end Nov 2023 and highlighting the +1.5 and +2.0C limits.
  3. a simpler design with a long term focus on the 20 year average against +1.5 and 2.0 limits. It does show yearly average anomalies since 1850, but years are not the focus. There is no highlight of the most recent year. It lacks daily anomalies, so it does not show any day-based peaks above the limits. It tells: while years jump up and down, the increase of the 20 year average is quite steady since the 1970s, almost a straight line. Recent year averages are close to the +1.5C limit but no year has crossed the limit yet. The latest 20 year average is a decade old and roughly half a degree from the +1.5 limit.
Uwappa (talk) 14:24, 30 January 2024 (UTC)
Bottomline: given already prominently shows warming since the industrial revolution, doesn't add substantive content beyond horizontal line(s)—especially in an already-crowded highest-level article. —RCraig09 (talk) 16:56, 30 January 2024 (UTC)
Added 'limit' to +1.5 and +2.0 C Y-axis labels in Uwappa (talk) 07:14, 31 January 2024 (UTC)
Alright, I have added a subheading to potentially help other editors participate in the selection process. This is my opinion to date.
  1. There is certainly enough space in the article for one more temperature-related chart. Consider that Climate change#Differences by region is currently lacking a chart next to it, which certainly sticks out. Normally, I would suggest moving down to that section, as its caption makes clear that it is already concerned with regional changes ("In recent decades, new high temperature records have substantially outpaced new low temperature records on a growing portion of Earth's surface").
  2. However, this comes second to rectifying a much larger omission: the lack of the SSP graphics. After a reader sees the temperature rise to date, they would have to scroll all the way down to Climate change#Reducing and recapturing emissions in order to find out how much warming is actually expected! This is completely unacceptable, so we need to place the SSP temperature projections from now to 2100 at the start of that section ASAP!!
  3. If we do that, we probably wouldn't have enough space to keep in Climate change#Future global temperatures. I think we might be able to both move that graphic upwards and downwards, to "Differences by region". We would just have to use a paragraph or so to explain how new cold records can still occur in spite of the overall warming trend. For now, this point does not appear to be made anywhere in the article, which is quite an omission. (And yes, the language of "Differences by region") talks about the present, and the graphic technically shows the future, but at this point, the +1.5C is already less than a decade away even according to the IPCC, so it wouldn't matter as much. Further, seeing +1.5C next to +4C in terms of regional temperature change and then immediately seeing the graph explaining that one of those is inevitable and another is very much not is likely to be very effective at conveying the idea.
  4. If we can manage this (or if we just remove outright, if it comes to that - after all, we already have the page image to show that warming is different depending on the area), then there would be free space at the start of Climate change#Warming since the Industrial Revolution to place one of the options. I find that first two options are both far' more valuable than the third one. I would strongly agree that provides nearly all of the same information - and it happens to look a lot better as well.
  5. (This and the following suggestions may or may not workable under WP:IMAGEOR.) In theory, we can move the +2.07 label from the first option to the second. Doing this would leave us with all the strengths and no weaknesses of the two.
  6. If allowed, I would also like to highlight the previous peak - the one which apparently occurred in February/March of 2020, around 1.8C. Digging into Copernicus posts from around that time would probably reveal a graph with such a label.
  7. Also in theory, but we can simply highlight the degree limits on the left side of : i.e. larger font, different colours (yellow-orange for 1.5 and shades of red for 2 and 2.5 instead of the current white for lower values, etc.) and/or subtly change the colour of their horizontal lines on the grid to match whatever colour is assigned to them.)
  8. Finally, I would STRONGLY prefer that any graph which aims to highlight how close we are to the 1.5C and 2C limits also includes the SSP graphs until around 2050-2060, thus showing when we are actually expected to officially reach those temperatures. Simply highlighting the limits and leaving our readers to extrapolate the trend mentally is NOT a good idea. The only scenario where I would prefer option 3 to option 1 is if WP:IMAGEOR makes this impossible to do for one, but not the other.InformationToKnowledge (talk) 08:25, 31 January 2024 (UTC)
Having considered all the options above, I simply ended up going ahead and doing what I considered necessary with those graphics, and several others that were clearly missing. I hope you all like it. InformationToKnowledge (talk) 20:31, 31 January 2024 (UTC)
The black graph is not readable at Wikipedia's thumbnail size * 1.35, so will require some tweaking before it's ready for prime time. Agree with ITK (8) about not making our readers extrapolate in their minds . —Femke 🐦 (talk) 20:48, 31 January 2024 (UTC)
Added to : +1.5C expected in 2033. Uwappa (talk) 23:06, 31 January 2024 (UTC)

Moving red+orange map from lead

Since there are three graphics in the lead, causing spillover into the "Terminology" section, I propose moving the longstanding map graphic down to the specifically appropriate /* Differences by region */ subsection that now (3 Feb 2023) has no graphic at all. —RCraig09 (talk) 18:31, 3 February 2024 (UTC)

I have thought about this too, but on the other hand, it is a really good image to present to our readers immediately. I don't remember that "most readers only see the lead" finding in detail, but @EMsmile can probably name it.
My idea was to keep that map where it is, and instead add to "Differences by region". Its entire first paragraph only really talks about the distribution of energy imbalance, so that graphic would be fitting. InformationToKnowledge (talk) 19:33, 3 February 2024 (UTC)
I also subscribe to the most-readers-only-look-at-the-pictures theory... Here, doesn't focus on overall global warming (the article's main topic) nearly as well as the line chart does. Meanwhile, is dreadfully techy, with parenthetic pairs, quantitative ranges, jargon, all requiring substantial geekxplanation for our lay audience to decipher. —RCraig09 (talk) 19:54, 3 February 2024 (UTC)
I don't have a link at my fingertips regarding what readers read but I am 99% sure that I read somewhere that "most readers only read the lead" (or something like that). Pictures are important as well of course but I think the text of the lead is very important. I quite like the image that I2K proposed and wouldn't rule it out on the basis of being "dreadfully techy". It does convey that climate change is actually quite complex and goes way beyond just an increase in temperatures.
Currently we have 3 images in the lead (which is actually a lot). Two of them are about temperature only, the third one is a collage of 3 effects of climate change. I am not sure if it shows the best 3 possible examples. E.g. coral bleaching has several causes, not just the marine heatwaves; nutrient pollution is another important one. And can the bulk of our readers relate to corals, other than those who like scuba diving and snorkelling? Doe the bulk of the people understand the impact of bleached corals on fisheries and biodiversity loss? I doubt it. The 4-image collage that we use at effects of climate change could be repeated here (although it also uses the coral example). But overall, that probably adds too many images to the lead.
I would favour something along the lines of (if needed, a simpler version of it) but something that does make a reference to the water cycle and the Earth's energy balance would be good, not just graphs of temperature going up. EMsmile (talk) 10:19, 5 February 2024 (UTC)
= Globe + 50 years of warming = Global warming. Good, keep it! Uwappa (talk) 11:11, 5 February 2024 (UTC)
OK, and what are your thoughts on what should be in the (currently imageless) "Differences by region"?
When I first started making major edits here, I placed your (Copernicus) graph to where is right now, and moved that bar chart to "Differences by region", but this has been reverted. My second proposal was to use , although there are objections to it as well. The third proposal I have is , because that section does mention melting ice. (Though not as much as it should - there is actually nothing in it about melting mountain ice caps or the mountains in general, now that I think about it.) That graphic is also already .svg, which should remove one of the major objections raised about the other graphics. InformationToKnowledge (talk) 12:26, 5 February 2024 (UTC)
There is no need to have an image in each an every section. Yes, would work well in "Differences by region", but it does too fine a job illustrating global warming in the lead.
I do love , but... I do not work for Copernicus. It is not my design.
About the other images currently in the lead:
  • could go to "2.2 Warming since the Industrial Revolution". Suggestions: replace 20 year smoothing by IPCC's 20 year average. Highlight +1.5 and +2.0 limits. Extrapolate 20 year average to +1.5C in 2033. De-emphasize yearly averages, use just dots. See for inspiration.
  • Move to "5. Impacts": and
Uwappa (talk) 13:44, 5 February 2024 (UTC)
I strongly urge everyone to focus on one topic, and avoid tangents and digressions and diversions. To resolve issues, discussions should converge, not diverge. —RCraig09 (talk) 17:51, 5 February 2024 (UTC)

Repeating the narrow issue:
Since there are three graphics in the lead, causing spillover into the "Terminology" section, I propose moving the longstanding map—which shows global warming differences by region—down to the specifically appropriate /* Differences by region */ subsection that now (8 Feb 2023) has no graphic at all. Please focus your responses. —RCraig09 (talk) 18:17, 8 February 2024 (UTC)

  • Support per original post. Replies: The red/orange chart would still be fairly prominent (in the third desktop-screenful); there's no need to purposely emphasize the less-than-useful perception that GW+CC are complicated, and the red/orange map is a better section summary than a glacial ice chart.RCraig09 (talk) 18:17, 8 February 2024 (UTC)
  • Stongly Oppose This graphic starkly presents the impact on the globe, showing the key issue of very uneven warming. It is clear data from NASA that people can get their head wrapped around. The graphic has also been updated for 2023 data and includes a Celcius only version that has been localized into many languages. I also see no spillover into the terminology section on a laptop screen, although perhaps on giant desktops there is an issue. Further, 2/3rds of our viewers are on smartphone, for which this is entirely not an issue.
While I would rather see no cuts from the lead, if something must be cut, it should probably be the effects collage. That collage explodes into many images on smartphone and is a very partial list of impacts. We already have 2 effects collages later in the article, so the content could just be moved there. Efbrazil (talk) 18:40, 8 February 2024 (UTC)
GW(redirect) and CC (broadly) are the key issues for lead, not "uneven" warming "stark"ly (alarmist?) presented. —RCraig09 (talk) 18:59, 8 February 2024 (UTC)
Starkly meaning clearly. It brings the issue home to people as they can see where they are on the globe and see what the change has been for them over the last 50 years. The map helps drive the point home much more clearly than a chart or picture of a forest fire will. Also, different effects / temperatures in different areas is an issue that is discussed in the second paragraph of the lead. Efbrazil (talk) 19:33, 8 February 2024 (UTC)
I'm not sure a color-coded map distinguishing local warming of a degree or two makes global climate change—whose effects are not limited to those localities—clear. —RCraig09 (talk) 20:13, 8 February 2024 (UTC)
Areas of the antarctic have cooled while the artic has warmed by over 4 degrees celcius. The issue is not subtle. The image also makes it clear that land is warming most quickly. Finally, the map makes it very clear the issue is real, not just a graph or he said / she said or a forest fire taken out of context. This graphic is the most compelling graphic we have on the topic, and I can't imagine why you'd want to hide it. Efbrazil (talk) 21:22, 8 February 2024 (UTC)
I was going to say that the average reader isn't going to recognize land vs. sea, arctic vs. other, etc. ... but I decided to just add this to the image's caption so that they do understand what they're seeing. PS - The map's "alt" caption was actually more specific, already! PPS - I wasn't trying to "hide" it. —RCraig09 (talk) 03:53, 9 February 2024 (UTC)
That works, thanks! Efbrazil (talk) 15:31, 9 February 2024 (UTC)
If you are concerned about too many images in the lead, we should really be cutting the "effects of climate change" collage. The images can be merged into the impact galleries. Is that something you can get behind? Efbrazil (talk) 19:33, 8 February 2024 (UTC)
Image gallery = another issue. —RCraig09 (talk) 20:13, 8 February 2024 (UTC)
If the complaint driving this ask is too many images in the lead, then the question is how to address that issue. Effects of climate change is 3 separate images on smartphone, which is the vast majority of our users. So cutting those images makes a much bigger improvement to the issue than cutting the first image. Efbrazil (talk) 21:12, 8 February 2024 (UTC)
Though many climate-related articles start out with image collages, I'm OK with moving the small collage lower, per ordinary editing. —RCraig09 (talk) 03:57, 9 February 2024 (UTC)
I agree with the proposal by Efbrazil to move the collage "effects of climate change" to later in the article (after checking for overlap/repetition). Having 5 images in the lead (2 separate ones plus a 3-image collage) is too many for the lead of a Wikipedia article, in my opinion. In addition to it being really hard to pick out 3 "representative" images for the effects of climate change... EMsmile (talk) 21:40, 8 February 2024 (UTC)
Someone else should make that proposal. My personal preference is to leave all the images in, but if there is a consensus that a cut needs to happen, that's the cut I'd make. Efbrazil (talk) 15:36, 9 February 2024 (UTC)
I've moved those 3 images about effects of climate change from the lead down to the main text (seemed like there was a mild consensus forming on that). That set of 3 images does double up a bit with other photos later (there is another one on wildfire, drought and coral reefs) but perhaps that is OK like that. EMsmile (talk) 16:40, 10 February 2024 (UTC)


Graphics / charts

(Breaking out discussion re graphics and charts so it's not buried in the preceding section)
A. September only
B. All months
C. Ratios
D. Shares (total 100%)
E. Raw numbers, stacked

Per the above, it's important to prominently emphasize both the degree and the depth of warming that are the very basis of climate change, with "degree" being the more important of the two. Re degree, I'm thinking the second or third graphic should replace the first ("September") graphic.

What is everyone's preference?

Separately: I've changed the ocean heat content chart to a red color scheme since ~red=warm.RCraig09 (talk) 17:09, 22 January 2024 (UTC)

Agree with replacement, the choice of September is arbitrary and could be construed as cherry picking data to make a point. I'm partial to the second graphic as it's not just another hockey stick graph. Seeing the cold record decline is helpful. One thing that isn't clear to me from the second graphic is why the warm and cold trace lines extend forward to current year, but don't extend backwards to the beginning of the time period (that issue applies to the September graphic as well though). Efbrazil (talk) 18:04, 22 January 2024 (UTC)
Additionally, it is not clear to me why "cold records" start so elevated over warm records. Shouldn't they start at a similar level, since the baseline would be pre-industrial? Efbrazil (talk) 18:23, 22 January 2024 (UTC)
The solid lines in the first two graphics are ten-year trailing averages, so the ten-year-average doesn't exist until the tenth year (here, 1959). Separately, I'm not sure "why" cold records start higher, but I-don't-make-the-news-I-only-report-it; though I note there was a cooling trend during mid-century. The charts aren't temperature graphs per se, but are based on new high records and new low records that have nothing to do with comparison to a pre-industrial baseline. —RCraig09 (talk) 18:39, 22 January 2024 (UTC)
I could theorize that localized smog was prevalent at that time and led to localized cooling that led to a lot of cold records, while modest heating was spread globally and so produced fewer records. The overall temperature graph doesn't go down by much in that time, so you'd need a localized effect. That's just a theory though; maybe records for cold temperatures just got better or something. See this for instance:
https://climate.nasa.gov/explore/ask-nasa-climate/3071/the-raw-truth-on-global-temperature-records/
Would be nice to see an official explanation. Efbrazil (talk) 19:00, 22 January 2024 (UTC)
To correctly show a ten year average you should be plotting against the central year the average is based on, correct? So the average should begin 5 years after the start of data and end 5 years before the end of data. Efbrazil (talk) 00:20, 23 January 2024 (UTC)
Suggestions to make less "busy":
  • use small light blue dots for cold records.
  • for warm records also use dots, light red, no black border. No need for 2 shapes as colour is already differentiating.
  • remove red/blue background colour below lines
  • replace dashed grid lines by thin solid light grey lines. Alternative: light grey background, solid white grid lines.
  • move legenda for records outside graph, as already done for lines. Replace "Thick lines are" by a short red and blue line, like in graph "Ratios". Remove the "(plotted annually)"
  • simplify y-axes description to "Earth surface", the percentage is already in y axes values
Uwappa (talk) 00:12, 23 January 2024 (UTC)
A not busy alternative: A stacked bar chart
  • x-axis: years
  • y-axis: percentage of earth, blue: cold records, red: warm record
Old years will have a lot of blue, recent years will have a lot of red. Simplified example with just 3 years:
1
2
3
4
5
6
7
1950
1980
2020
  •   % surface with cold record
  •   % surface warm record

The chart could show 100% by adding categories as: colder than usual, normal, warmer than usual. The result will be a graph similar to , but with a meaningful y-axis.

Another idea, leading to a similar design: show a stacked bar per year showing how recent records were broken.

I've uploaded Version 3 of Chart B, which has less dominantly colored dots—to reduce busy-ness. Other suggested changes were either not necessary to reduce busy-ness, were hard to implement in SVG, or were not color-blind friendly.
Your suggestions make me think it might be good to have a stacked bar chart of percentages of new record highs (red) and new record lows (blue). It's like a more balanced, comprehensive alternative to Chart C.
We're getting lost in details and endless alternatives... To focus discussion, please let me know which chart's approach you all think is best used in this high-level article. —RCraig09 (talk) 05:26, 24 January 2024 (UTC)
This will probably rock your boat even more, but still: A chart in the chapter "Warming since the Industrial Revolution" should show more than number of record days.
Please show eh, well, ... warming since the industrial revolution. How much warmer has it been?
does exactly that. (newer version available with 1.5 and 2.0C emphasized ) Record days are easy to spot: the lowest/highest at the time.
A simplified alternative based on same data:
  • a stacked bar chart, with on x-axes: years
  • for each year the number of days per warming range, e.g. in buckets of half degree Celsius.
  • colour codes, dark blue for much colder, dark red for much warmer.
The number of record days will show in a different way. Only recent years will have bar segments with the hottest colours, e.g 2023 is the first year shortly peaking above +2C and will have the highest number of days above 1.5C. Uwappa (talk) 14:31, 24 January 2024 (UTC)
@Uwappa: My boat is definitely rocked. Over the years here, I've learned this community favors simpler graphics that clearly convey (usually) one concept per graphic. I've come around to agreeing with that view, since we're communicating with the general public in a layman's encyclopedia—not trying to impress our Ph.D. advisor. I've just added some relatively simple bar charts above, as Charts "D" and "E". If you have radically new suggestions, please place them in a separate sub-section. —RCraig09 (talk) 20:46, 24 January 2024 (UTC)
Chart E is really easy to read, decreasing blue, increasing red. Too easy, well done!
Small suggestion: reduce visual noise, replace dashed grid lines by solid lines in a lighter shade of grey. The chart will be ready to go and replace the current
I'll start a new section with the radical idea: a chart showing warming since the industrial revolution in the chapter "Warming since the Industrial Revolution". Uwappa (talk) 17:33, 25 January 2024 (UTC)
Craig- I'm copying this comment from up above because I think you missed it:
To correctly show a ten year average you should be plotting against the central year the average is based on, correct? So the average should begin 5 years after the start of data and end 5 years before the end of data. Efbrazil (talk) 17:13, 24 January 2024 (UTC)
@Efbrazil: Sorry for not responding earlier. I'm aware of both "centered" moving averages and "trailing" moving averages. I've always used trailing moving averages, as they visually convey that the chart is up to date. Yes, I realize that the Ocean Heat content chart uses centered moving averages, but that is only because one of the sources provided only that data.) In any event, it's a formal consideration to be tackled after the substantive decision re what's most appropriate to include in this article (I've just uploaded new bar charts "D" and "E", above), indirectly inspired by Uwappa.) —RCraig09 (talk) 20:54, 24 January 2024 (UTC)
The problem with 10 year trailing moving average is it shows the average as of 5 years ago as being the average today, which is simply incorrect. Lowess smoothing or something like that is good, although Excel doesn't natively do that. One thing you can do in Excel is use forecast.linear to smooth the last 5 data points in either direction, then use a moving average in the middle. Efbrazil (talk) 18:36, 25 January 2024 (UTC)
Hmmm... I'm not fully grasping Lowess-ing, and Excel doesn't generate the curves (my spreadsheets generate the actual curves)... I am debating whether to center-justify (as in File:1955- Ocean heat content - NOAA.svg) or simply amend legends to read "...trailing moving average" to resolve any ambiguity. 19:13, 25 January 2024 (UTC) The question is likely moot for this particular article, since the bar chart seems to be the chosen replacement for the September-only graphic. —RCraig09 (talk) 19:22, 25 January 2024 (UTC)
The way forecast.linear works in excel is that it fits a linear line to data you have, then uses that to estimate data points outside of that interval. It works well for extending a moving average to the extents of the data as the central data point in the linear fit will be the same value as the moving average.
Partly for my own understanding, I figured things out in Excel formula world. Say you are doing a 9 year moving average on a data set that goes from B2 to B101 (100 years).
This will give you the moving average for cells B6 through B97 in a way that lets you paste the formula anywhere:
=AVERAGE(INDIRECT(ADDRESS(ROW()-4,2)):INDIRECT(ADDRESS(ROW()+4,2)))
This will give you the forecast for cells B2 through B5:
=FORECAST.LINEAR(ROW(),B$2:B$10,ROW(B$2:B$10))
And this will give you the forecast for cells B98 through B101:
=FORECAST.LINEAR(ROW(),B$93:B$101,ROW(B$93:B$101))
You can then put that all together into a giant hideous if statement like this:
=IF(ROW()<6,FORECAST.LINEAR(ROW(),B$2:B$10,ROW(B$2:B$10)),IF(ROW()>97, FORECAST.LINEAR(ROW(),B$93:B$101,ROW(B$93:B$101)), AVERAGE(INDIRECT(ADDRESS(ROW()-4,2)):INDIRECT(ADDRESS(ROW()+4,2)))))
Although separate columns might be better so you can keep track of things. Really terrible that Excel in 2024 doesn't support the most basic of trend line options. Welcome to monopoly world. Efbrazil (talk) 20:23, 25 January 2024 (UTC)
I'm not comfortable presenting a chart trace with "ends" that are based on fewer data points than everything between the ends. Before I jump in to studying Excel in this regard, can you tell me (yes/no): are your four coding examples exactly literal? This coding seems simpler. —RCraig09 (talk) 21:20, 27 January 2024 (UTC)

Narrow discussion re natural-vs-observed chart

@Efbrazil and Uwappa: and others: Reactivating discussion about the natural-vs-observed chart . I think (a) the smoothed average should be more dominant than the choppy annual temp values, (b) the annual values should be dots, not continuous-choppy, and (c) the not-understood-by-laymen green confidence interval should be eliminated or made much less dominant. Focus; please start another (sub)section for digressions or tangents. 20:28, 9 February 2024 (UTC)

I agree. These 2 charts now overlap quite a bit.
My suggestion: merge the designs of those two charts:
  • yes, downplay years to small dots, maybe connected by a thin line
  • yes, show a dominant line, but stick to the IPCC compatible 20 year average
  • Swap the F and C scales, so recent years are easy to compare against +1.5C
  • show an extrapolation of the straight line since the 70s to the +1.5C limit
  • alternative: do show the green confidence interval, but remove the dark green line.
A combined chart would tell a complete story:
  1. since the mid 70s observed temperatures are beyond natural drivers.
  2. since the 70s the 20 year average is almost a straight line going up
  3. the year 2023 was close to the +1.5C limit, such hotter-than-average has happened before
  4. current expectation: it will take another decade before the average crosses the +1.5C limit.
Uwappa (talk) 22:22, 9 February 2024 (UTC)
Regarding the bullets above:
  • I do not want to switch yearly data from a clear line to points or something faded out for reasons I've gone over several times now, please see prior discussions for details.
  • Not sure what this means, there is already a 20 year average in the first chart.
  • I'd be fine swapping celcius and fahrenheit keys. Note there is already a celcius-only version of the first chart for localization purposes, so that wouldn't change.
  • I do not want to present data that does not exist and is not sourced. People can use their imagination.
  • I have a version in the works that added the confidence information to the key in response to an ask, I plan to include that in the next version.
Efbrazil (talk) 22:42, 9 February 2024 (UTC)
I, also, am against extrapolation and superfluous dark lines at "1.5" etc, and favor simplicity (in graphics... and on Talk Pages). Efbrazil, I don't know what "the key" is (a legend that would add complexity?). Efbrazil, I couldn't locate your response re dots for annual data, as there are 74 instances on this 281KByte Talk Page of the text string, "efbrazil (talk)". —RCraig09 (talk) 06:57, 10 February 2024 (UTC)
The key is the place where I describe what the 3 line colors are. I just add a fourth thing talking about what the shaded area means. Doesn't garbage up the chart much.
Regarding shifting one or more lines to points, sorry for being vague. I get grumpy repeating myself, which is one of several reasons why I was a lousy people manager. But that's besides the point.
Anyhow, one place we discussed it was in the section "Global temperature and forces graphics updated". Pasting my comments in here from there:
Good question about changing the annual line to points. I agree it could make sense if there weren't the "Natural drivers only" line, but with that line there I don't think the change is a good idea. They should either both be points or neither, and I don't think natural drivers should be changed to points and a trend line because the value is in showing how the annual fluctuations mirror the natural driver fluctuations (and that's how the IPCC presented the data).
I don't know that the black line being most visible is bad, as that's the only real data here. It used to be all we showed in this graphic. The trend line and natural drivers lines are additions meant to provide context and aren't real data. The IPCC just switched from a 30 year to a 20 year trend line for instance (from AR5 to AR6), so those lines are kind of fungible off shoots of the real data, which is the black line. The red trend line is topmost in z-order, for what that's worth, but you are right that the black line is higher contrast and so more visible. Efbrazil (talk) 16:24, 10 February 2024 (UTC)
Thanks for explaining, @Efbrazil: ● I think Uwappa and I are talking about emphasis (trends) and de-emphasis (confidence intervals), and not removal of items. ● Eyeballing, the annual temp line versus the natural drivers line do not in fact visually correlate on a year-to-year basis, so our (esp. lay) audience could not possibly notice an annual correlation re internal variability. ● Separately, the annual temp values are themselves averages of millions+ of measurements, and are thus not "real data" any more than smoothed data; it's the (smoothed) trends that are substantively important for this top-level CC article. ● Few IPCC publications are oriented toward a lay audience, so techy/academic details of IPCC graphic presentation should not govern us here. ● (my diff, FYI). ● Aside: consider the simplified labeling (no "key"), and many fewer tick marks, as in (no need to respond). —RCraig09 (talk) 18:14, 10 February 2024 (UTC)
Yeah, fair enough though that a 20 year average and a 1 year average are both showing real data. I guess we could show monthly or daily data if we wanted to. Since this is a climate change article I can see the argument for featuring the 20 year line and deemphasizing the 1 year line, like you say. Thanks for explaining that.
I'm still uncomfortable reducing annual temps to a scatter plot though. An annual line is how natural drivers are shown, is what is typically shown online / in the press, and is what I think our audience will want to see. Nobody talks about the 20 year average. Also, I do see a partial correlation between annual temps and natural drivers- a lot of the peaks and valleys are the same. Interesting that the lines are even more closely correlated using the data set for the graph you came up with, particularly volcanic influences. We could call out the issue in the caption.
Would you be happy with a revision that changed the emphasis towards the 20 year line? You've persuaded me on that count. Efbrazil (talk) 19:27, 10 February 2024 (UTC)
Yessireebob, see for an example of how annual values can straddle smoothed lines. Separately, I again urge a much lighter shade of confidence-interval-green so lay readers aren't distracted/confused. PS - Exploding the key/legend into separate labels alongside respective traces—like —simplifies the chart (the "xx-year smoothing" label is not needed here, and can be added to the Commons file description page). PPS - I'd zap the superfluous tickmarks per my 18:14 comment. —RCraig09 (talk) 21:33, 10 February 2024 (UTC)
Thanks Craig, I'll look to integrate all this next week. Efbrazil (talk) 19:46, 11 February 2024 (UTC)
New version uploaded. I tried several things and went with a version that is maybe more incremental than you would like. Changes:
  1. Celcius and Fahrenheit flipped
  2. Red line is much more bold now, black line much thinner
  3. Clarified in key that the black line is the annual average
  4. Clarified in key that natural drivers includes both the estimate and the error range
I tried moving the labels next to the lines but it made things worse IMHO. One issue is that you can't really point to an error range. Another problem is that the red and black lines are effectively in the same spot. Finally, putting labels next to lines would mean moving the natural line descriptions to the bottom, and I didn't want to do that as it either meant extending the chart or squeezing it or cutting the 2.0 C number (which is a target and should stay IMHO). While I agree that labels next to lines is a good goal in general, I think the existing key does the job and is clear with the text colored the same as the lines.
I also tried moving the 20-year smoothing to be the top line item in the key, but the trouble is that's based on the annual averages and I wanted to keep the "2023" text, to make it clear the chart is current. Given that, the whole thing reads better with the sequencing that is there.
Hopefully what's there is acceptable as a compromise. Efbrazil (talk) 21:22, 12 February 2024 (UTC)
Great improvements!
As the dark green and light green visually connect, you may want to try a lighter shade of red to de-emphasize the years, visually connecting years with the 20-years smoothing. I expect the thick red line can be thinner when combined with lighter red years. Uwappa (talk) 22:11, 12 February 2024 (UTC)
The temperature traces in the 16:11 12 Feb version are a mild improvement; however, the green natural drivers error range area is still pointlessly dominant for our lay audience and could be halved in darkness of tone. Separately, the growing legend/key complicates things for lay readers: ● "up to 2023" is superfluous, ● the entire "20 year smoothing" entry is techy detail that's unneeded outside academia (compare ), ● the "estimate and error range" phrase is confusing to laymen. Taken together, the green error range area, the still-too-dominant annual temp trace, and the verbose and techy legend/key, essentially double the complexity of what should be a simple-concept chart showing the important quantities: (1) smoothed temperature and (2) natural drivers. They are THE point of the chart. Happily, cutting the key to two simple entries allows the two entries to be placed alongside the two important traces (OK to keep less-dominant annual temp and less-dominant green error range area, without naming them in text). Separately: the unnecessary (0.1°) micro-tick marks remain, but the useful gridlines that once spanned the chart width have gone missing. Substantively, I thought the average 2023 avg temp was between 1.4 and 1.5 °C, but the chart shows it under 1.4. (Sincerely, Simon Cowell)RCraig09 (talk) 22:58, 12 February 2024 (UTC)
Thanks Uwappa! Good point to connect the 20 year and annual averages by using the same color. I played around with reds, but the difficulty with red is that very light red can't be rendered as a text color while remaining visible, and red on it's own draws the eye more than black on its own. I tried going with shades of gray but then the text isn't easy to read. It's best if the annual average is black because of that. To help a bit I just made the red line a tick darker.
Craig is sadly less positive. Point by point:
  • The green confidence darkness varies by monitor- much brighter and it won't show up a lot of the time. It is also important information to convey and I have already greatly lightened it from source material.
  • I do not want to remove "up to 2023" as people wanting to know what the annual trace line goes up to will find that valuable, plus it makes it clear how current the chart is.
  • 20-year smoothing is accurate- the technical term would be "LOESS line" or "LOESS smoothing", which is what NASA puts on their charts. There's a balance to strike between being too technical and being open to misinterpretation, and I think where the chart is now strikes the right balance.
  • I disagree with your characterization of "the important qualities" and "THE point of the chart". Annual temperature line is extremely important and, if you want simplification, is really the only line we should be showing. Error range is also important as when temperatures exit that range it means scientists are confident that human factors are responsible; natural drivers is a modeled number and it can be presented differently.
  • I removed the spanning lines as they are no longer necessary since celcius is on the right hand side. I removed the .1 tick marks from Fahrenheit but not celcius, as it's important to show exact how far along temperatures are. I had them on Fahrenheit mostly to match celcius.
  • Regarding temperatures, I am pulling data from NOAA and normalizing it to the 1850-1900 average. Time scale is 12 months. This is the source. This past year was 1.35 C but the 20 year line is only up to 1.2 at this time. If you have a better data source please share it. I was using NASA temperatures, but they don't go back to 1850. Thoughts?
Efbrazil (talk) 18:39, 13 February 2024 (UTC)
You are welcome.
  • I love the swapped C and F labels. It it now very clear that recent years are close to +1.5C. It now even triggers discussion on what the most recent temp should be, good!
  • An other idea to visually join the year lines: Use the same, or very similar red for both. But.. .make both a bit transparent so you can see them both, even when they cross. I think you could safely half the linewidth of the individual years, so the 20 year line gets the focus, while years remain visible. Your previous version of red (#FF0000) has my preference. You may create one legend entry for both year lines, with a straight thick and jumping thin line as legend icon.
  • You could move 2023 to a smaller tick on the very end of the x-axis, as done in . See its sources for temperatures that do go back to 1850.
  • For me the main story the chart is telling is: Until the 40s year temperatures were within natural range. Since the 70s the average temperature goes up, in a straight line. Since the 80s it is outside natural range. Individual years are already approaching the +1.5 limit.
  • For me the main things missing remain: where is the IPCC 20 year average? When is it expected to cross +1.5C? I know we disagree on this point.
Uwappa (talk) 20:04, 13 February 2024 (UTC)
@Efbrazil: Normalized data for 2023 from six sources is listed in the "Data for Version 16 (through 2023)" section (expandable text) on the file description page of . It reads: "1.44 1.4 1.44 1.44 1.43 1.48". Original Met Office source with six links to six datasets: https://climate.metoffice.cloud/temperature.htmlRCraig09 (talk) 22:14, 13 February 2024 (UTC)
Thanks Craig! Looking at that data the difference between the sources is how they estimate temperatures back in the 1800s, so the issue is what the 1850-1900 average is, not so much the temperature trace since then. Since the IPCC is our bible I'll look at that tomorrow. Do you happen to know what source they use? Efbrazil (talk) 01:07, 14 February 2024 (UTC)
That Met Office source nicely aggregates and normalizes the six datasets. The source states "Dataset anomalies are calculated relative to a 1981 to 2010 baseline and offset by 0.69°C, which is the best estimate difference for that period from the 1850-1900 average given in the IPCC sixth assessment report." (It's in the caption to the chart titled "Annual global mean temperature difference from pre-industrial conditions", and you can get down to the data itself by clicking on, you guessed it, "Get the data". :-D ) —RCraig09 (talk) 04:23, 14 February 2024 (UTC)
Averaging the data sets sound reasonable, but I don't know that simple averaging is the best way to do things, and it seems like original research since there's no official source behind that averaging. What I really want to know is what the IPCC did. Do you know?
What I did previously is use the NASA dataset (which starts in 1880) and then offset it by what the IPCC says about averages in current times vs the 1850-1900 time frame. The IPCC says this in AR6 WG1 SPM pp5 A.1.2:
Global surface temperature was 1.09 [0.95 to 1.20] °C higher in 2011–2020 than 1850–1900, with larger increases over land (1.59 [1.34 to 1.83] °C) than over the ocean (0.88 [0.68 to 1.01] °C). The estimated increase in global surface temperature since AR5 is principally due to further warming since 2003–2012 (+0.19 [0.16 to 0.22] °C).
So, given that, it's easy to do the offsets. I could go back to that methodology, but somebody complained that the data I was showing didn't go back to 1850, which is why I switched to the NOAA data set. Trouble then being that if your data goes back to 1850 you can't really fudge the baseline to match the IPCC- you need to use the actual average from 1850-1900.
Thoughts? Efbrazil (talk) 00:43, 15 February 2024 (UTC)
I was never advising averaging the six datasets. Since the main purpose of the present chart is to compare observed temp versus natural drivers, I think that any of the 1850+ datasets would be fine, especially since they all have greater than 99.1% correlation (see correlations). —RCraig09 (talk) 05:56, 15 February 2024 (UTC)
I agree with Craig09. Just pick any as differences are too small to be relevant for the chart. uses https://climate.metoffice.cloud/formatted_data/gmt_HadCRUT5.csv as a source with 2023 at +1.44C. If I remember correctly Ed Hawkins used HadCRUT for his climate spiral. Uwappa (talk) 07:48, 15 February 2024 (UTC)
The issue though is that the data sets do differ in the 1850-1900 range. If you average NOAA 2011-2020 compared with NOAA average from 1850-1900 you get a temperature anomaly of 1.02 C, but the IPCC says it should be 1.09 C. I don't know what data set the IPCC is using to say 1.09 C. Efbrazil (talk) 16:38, 15 February 2024 (UTC)
The 1850 temperatures in are -0.08, 0.09, and -0.23, implying a range of 0.09-(-0.23)=0.32 °C and an average of -0.07 °C. True, these numbers are noticeably different from the "0" line (vertical center of the "natural forces" area for 1850). My quick Google-search of site:ipcc.ch didn't yield a simple answer to which dataset the IPCC is using (all charts were more complicated: example). However, the main point of your chart is to show divergence in recent decades, not precision in 1850. So I'm not worried about the particular choice of dataset. —RCraig09 (talk) 18:27, 15 February 2024 (UTC)
OK, is what I did is go back to the NASA dataset and then offset according to the IPCC declaration on the temperature delta in the years 2011-2020. Note graph now starts in 1880 (when NASA data set starts). Like you say, we shouldn't be getting hung up on the years 1850-1880 but we should be honoring the IPCC definition of anomaly. This also allows me to show the time axis in 20 year increments, instead of 50 year increments. Efbrazil (talk) 18:46, 16 February 2024 (UTC)
to Efbrazil's data choice. —RCraig09 (talk) 21:54, 16 February 2024 (UTC)
Please have a look at this path to the CSV:
https://www.ipcc.ch/report/ar6/wg1/figures/summary-for-policymakers/figure-spm-1/
> datasets
https://catalogue.ceda.ac.uk/uuid/0b2759059ad6474098e40dad73e0a8ec
> Download
https://data.ceda.ac.uk/badc/ar6_wg1/data/spm/spm_01/v20221116
> panel b
https://data.ceda.ac.uk/badc/ar6_wg1/data/spm/spm_01/v20221116/panel_b
> gmst_changes_model_and_obs.csv
https://dap.ceda.ac.uk/badc/ar6_wg1/data/spm/spm_01/v20221116/panel_b/gmst_changes_model_and_obs.csv?download=1
Is Column 8 (H) what you are looking for? It is documented as:
Global Surface Temperature Anomalies (GSTA) relative to 1850-1900 ... Uwappa (talk) 20:19, 16 February 2024 (UTC)
Thanks! I had forgotten about that data source. The nature part of the chart is in fact from that data set, but of course it ends at 2000. I needed temperature data going up to present (2023). I wasn't comfortable merging data sets, although maybe that would be best (using IPCC data to 2020, then NASA data for the last 3 years). I'd like to leave the graph as it is now if you are OK with that, as I don't think changing data sets again will help much. I've decided that whether the data starts in 1880 or 1850 doesn't much matter. Efbrazil (talk) 21:26, 16 February 2024 (UTC)
An alternative to visually connect the years and 20-years lines:
Use the same red for both, just make the years very thin, e.g. line width 0.5 Uwappa (talk) 17:40, 16 February 2024 (UTC)
I tried that, but it is helpful having them be different colors in the key (so the text color is all you need to identify the line). I think what is there now works pretty well. Efbrazil (talk) 21:28, 16 February 2024 (UTC)
+ I can prepare a substitute version if that would make things easier. —RCraig09 (talk) 18:13, 13 February 2024 (UTC)
I don't think so; I understand your concerns- we are disagreeing mostly on substance mostly here. The changes you have suggested that I have not adopted are ones that I don't think improve the graphic. Efbrazil (talk) 18:55, 13 February 2024 (UTC)
Good, looking forward to the swapped F and C scales. Yes end of January you have expressed you objection to replace the black line by points. We also discussed the 20 year average versus smoothing before. Source for the 2033 estimate documented , since 31 January.
RCraigg09, to find previous answers search for 'annual values to dots'. Uwappa (talk) 14:37, 10 February 2024 (UTC)

Should the effects collage be restored to the lead?

Revisiting the choice of images for the effects in the lead

Long-term chart

Tidying up list of sources?

Tried to improve reading ease of one sentence in the lead

Proposed text change in lead for mitigation, up or down vote please

Related Articles

Timelines

Top Qs

Fact Checks