|
|
Post by Steelernation on May 5, 2023 0:28:56 GMT -5
The trend of cool Aprils continues, it was the sixth April in a row below the 1991-2020 average mean. Avg Max Anomaly: -3.9C Avg Min Anomaly: -2.7C Mean Anomaly: -1.4C Lowest Max: -5.2C (5th) Highest Min: 2.6C (13th) God that is one horrible climate |
|
|
|
Post by chesternz on May 5, 2023 10:01:41 GMT -5
Bangkok, Bang Na agromet.
--------------------------------
Highs
Max: 38.7 C
Mean: 35.6 C
Min: 31.1 C
Lows
Max: 29.8 C
Mean: 28.1 C
Min: 23.7 C
No reliable data on rainfall, but annecdotally there was only one major rain event the whole month. The BOM gives a total of 5 mm for BKK metro this April. Pretty average April (albeit drier than the norm), but it got pretty hot towards the end of the month with downtown just cracking 40 C one day and managing a 30.0 C low one morning.
|
|
|
|
Post by Cheeseman on May 6, 2023 7:19:52 GMT -5
April 2023 Summary for Madison, WIIt was a rather warmer and drier than average month - but not a consistently warm month, as the majority of days were cooler than long-term averages. The averages were boosted significantly by the anomalously warm second week, which brought four consecutive days in the 80s - the high of 83 F (28 C) on the 12th made it the earliest warm day since 2012! Proving this is still Wisconsin in April, we still had a high in the 30s and two events with measurable snowfall after that. Thankfully, we avoided the snow system on the 30th through May 1 that brought snow by the foot to the UP. With no real major precipitation events and a nine-day streak without a drop, as well as dew points near freezing on the warmest days, April definitely produced some elevated fire danger conditions, leading to a 3,000 acre wildfire in west-central Wisconsin.  The Year So Far The above-average streak that's been going on the entire time I've lived in Madison continues. Will May continue the streak?
|
|
|
|
Post by Steelernation on May 6, 2023 12:22:43 GMT -5
The above-average streak that's been going on the entire time I've lived in Madison continues. Will May continue the streak? December and March were below average |
|
|
|
Post by tommyFL on May 6, 2023 18:17:44 GMT -5
The above-average streak that's been going on the entire time I've lived in Madison continues. Will May continue the streak? December and March were below average Bear in mind that Cheesehead believes that 1869-present averages are more representative of the current climate than 1991-2020. |
|
|
|
Post by Benfxmth on May 6, 2023 18:23:25 GMT -5
Even if unintentional, using older normals than the current climate period can be utilized as a Boteving tactic, just like very short POR averages.
|
|
|
|
Post by greysrigging on May 6, 2023 18:41:50 GMT -5
Even if unintentional, using older normals than the current climate period can be utilized as a Boteving tactic, just like very short POR averages. I think even the last 30 years normals is too short of a POR... and if one is happy to use historical records ( be they max, min or rainfall ) one should also use the long term averages. |
|
|
|
Post by Steelernation on May 6, 2023 21:32:48 GMT -5
I think even the last 30 years normals is too short of a POR... and if one is happy to use historical records ( be they max, min or rainfall ) one should also use the long term averages. Records are what could happen so it makes sense to have a long por for them, averages are what is “typical” so makes sense to have a representative period for them, ie. 91-20. Usually using a the full por back to the 1800s for averages is a climate change denier tactic—“the climate isn’t changing so the 1800s are relevant for averages and the same as today” |
|
|
|
Post by Cheeseman on May 6, 2023 21:38:36 GMT -5
December and March were below average Bear in mind that Cheesehead believes that 1869-present averages are more representative of the current climate than 1991-2020. 1991-2020 is weather, not climate. There are tons of cyclical climate-affecting processes, such as the Atlantic Multidecadal Oscillation with a period of about 70 years, that work on a much longer time scale than that. Also, 1869-present includes the years of 2021, 2022, and 2023 - not included in 1991-2020. What makes 1991-2020 superior to, say, 1993-2022 (the actual last 30 years if we're using that as our comparison period) other than the fact 1991 and 2020 end in specific digits? It was outdated as soon as 2021 was out and we ought to have switched to 1992-2021 normals if that's your logic. And if I'm going to use a restrictive 30-year period, pretending nothing before 1991 is relevant to the climate of my location, then I should compare my extremes to that period as well - which would have made my December 23 high of -3 F (-19 C) the oldest December high on record because 1989 doesn't count, and my March the snowiest on record because 1959 doesn't count. That seems disingenuous as hell! Even if unintentional, using older normals than the current climate period can be utilized as a Boteving tactic, just like very short POR averages. The term "Botev" is thrown around a lot on here by people who don't understand what it means. What do I have to gain by making my climate seem colder than it is (winter averages are colder in the full POR than in the 1991-2020 set)? Last I checked, boteving was supposed to be about misrepresenting your climate to make it seem more ideal in accordance with your own preferences - not less. I think even the last 30 years normals is too short of a POR... and if one is happy to use historical records ( be they max, min or rainfall ) one should also use the long term averages. Records are what could happen so it makes sense to have a long por for them, averages are what is “typical” so makes sense to have a representative period for them, ie. 91-20. Usually using a the full por back to the 1800s for averages is a climate change denier tactic—“the climate isn’t changing so the 1800s are relevant for averages and the same as today” If data from the 1800s aren't relevant to modern climate, then records that have stood since then ought to be viewed as similarly impossible (such as the -36 F in Green Bay in 1888), and similarly excluded from the data set - right? |
|
|
|
Post by Steelernation on May 6, 2023 21:44:03 GMT -5
1991-2020 is weather, not climate. There are tons of cyclical climate-affecting processes, such as the Atlantic Multidecadal Oscillation with a period of about 70 years, that work on a much longer time scale than that. Also, 1869-present includes the years of 2021, 2022, and 2023 - not included in 1991-2020. What makes 1991-2020 superior to, say, 1993-2022 (the actual last 30 years if we're using that as our comparison period) other than the fact 1991 and 2020 end in specific digits? It was outdated as soon as 2021 was out and we ought to have switched to 1992-2021 normals if that's your logic. And if I'm going to use a restrictive 30-year period, pretending nothing before 1991 is relevant to the climate of my location, then I should compare my extremes to that period as well - which would have made my December 23 high of -3 F (-19 C) the oldest December high on record because 1989 doesn't count, and my March the snowiest on record because 1959 doesn't count. That seems disingenuous as hell! Ah yes, you know better than every single weather agency out there.It’s inconvenient for weather agencies to update normals every month or every year, that’s why they do it every 10 years. Otherwise 1993-2022 would be more relevant. And 30 years is a plenty long time for climate, 5 or 15 years is too short. Big difference between averages and records as they show 2 different things therefore as my previous post says a full por for records and a relevant por for averages makes sense. Even if records can’t happen nowadays, one doesn’t know that 100% and they still happened therefore they should count. Many places still set monthly/all time records. However when it’s now 2-3 f warmer on average than a 100 years ago, that’s not relevant for an average that represents todays climate. |
|
|
|
Post by Cheeseman on May 6, 2023 21:52:56 GMT -5
1991-2020 is weather, not climate. There are tons of cyclical climate-affecting processes, such as the Atlantic Multidecadal Oscillation with a period of about 70 years, that work on a much longer time scale than that. Also, 1869-present includes the years of 2021, 2022, and 2023 - not included in 1991-2020. What makes 1991-2020 superior to, say, 1993-2022 (the actual last 30 years if we're using that as our comparison period) other than the fact 1991 and 2020 end in specific digits? It was outdated as soon as 2021 was out and we ought to have switched to 1992-2021 normals if that's your logic. And if I'm going to use a restrictive 30-year period, pretending nothing before 1991 is relevant to the climate of my location, then I should compare my extremes to that period as well - which would have made my December 23 high of -3 F (-19 C) the oldest December high on record because 1989 doesn't count, and my March the snowiest on record because 1959 doesn't count. That seems disingenuous as hell! Ah yes, you know better than every single weather agency out there.It’s inconvenient for weather agencies to update normals every month or every year, that’s why they do it every 10 years. Otherwise 1993-2022 would be more relevant. And 30 years is a plenty long time for climate, 5 or 15 years is too short. Ah, so 1991-2020 is more relevant than 1993-2022 simply because "muh WMO" or whatever met agency says so. Never mind that it is excruciatingly easy to make a 1993-2022 POR or whatever other desired POR in 30 seconds on NOWData. Excellent appeal to authority - and great job of ignoring the part about climate drivers acting on cycles with periods longer than 30 years as well. I find it easy to imagine a particularly warm or cold decade throwing off a 30-year set of averages - then when that anomalous decade leaves the 30-year data set, it suddenly appears as though the climate has changed significantly, when really all that happened was a regression toward the mean. It would take a much stronger arctic blast to produce that -36 F today than it would have in 1888, due to climatic changes and increased UHI over that time frame. So are events from 1888 relevant to today's climate, or are they not? Only when they fit a certain agenda, I see. |
|
|
|
Post by Steelernation on May 6, 2023 22:03:56 GMT -5
Ah, so 1991-2020 is more relevant than 1993-2022 simply because "muh WMO" or whatever met agency says so. Never mind that it is excruciatingly easy to make a 1993-2022 POR or whatever other desired POR in 30 seconds on NOWData. Excellent appeal to authority - and great job of ignoring the part about climate drivers acting on cycles with periods longer than 30 years as well. I find it easy to imagine a particularly warm or cold decade throwing off a 30-year set of averages - then when that anomalous decade leaves the 30-year data set, it suddenly appears as though the climate has changed significantly, when really all that happened was a regression toward the mean. It would take a much stronger arctic blast to produce that -36 F today than it would have in 1888, due to climatic changes and increased UHI over that time frame. So are events from 1888 relevant to today's climate, or are they not? Only when they fit a certain agenda, I see. I’ll use whatever the official normal period is and that’s 91-20, but I agree with you 94-23 is more relevant. And if nearly every weather agency in the world agrees on a 30 year period I think they’re on to something, I trust them more than an armchair weather nerd. The climate is warming regardless of long term climate drivers so I ignored that, they don’t matter. I’ll agree with you one anomalous decade dropping out is an issue but it’s better than including data from 100 years old in the “normal”. The climate is different than 80 or 100 or 120 years ago, get over it. Records represent what is possible, normals represent what you can expect. Therefore, a record in 1870 is relevant because it happened but 1870 isn’t relevant for averages because the climate is different now so that’s not what you can expect. Not hard to understand. And no I don’t have an “agenda”. For some reason you’re the only one not using official normals, therefore you have an agenda, whether it’s making your climate seem worse than it is, making yourself feel better that it’s always above average, or you don’t want to use the updated normals because that’s what the liberals use |
|
|
|
Post by Cheeseman on May 6, 2023 22:19:21 GMT -5
Ah, so 1991-2020 is more relevant than 1993-2022 simply because "muh WMO" or whatever met agency says so. Never mind that it is excruciatingly easy to make a 1993-2022 POR or whatever other desired POR in 30 seconds on NOWData. Excellent appeal to authority - and great job of ignoring the part about climate drivers acting on cycles with periods longer than 30 years as well. I find it easy to imagine a particularly warm or cold decade throwing off a 30-year set of averages - then when that anomalous decade leaves the 30-year data set, it suddenly appears as though the climate has changed significantly, when really all that happened was a regression toward the mean. It would take a much stronger arctic blast to produce that -36 F today than it would have in 1888, due to climatic changes and increased UHI over that time frame. So are events from 1888 relevant to today's climate, or are they not? Only when they fit a certain agenda, I see. I’ll use whatever the official normal period is and that’s 91-20, but I agree with you 94-23 is more relevant. And if nearly every weather agency in the world agrees on a 30 year period I think they’re on to something, I trust them more than an armchair weather nerd. The climate is warming regardless of long term climate drivers so I ignored that, they don’t matter. I’ll agree with you one anomalous decade dropping out is an issue but it’s better than including data from 100 years old in the “normal”. The climate is different than 80 or 100 or 120 years ago, get over it. Well, I think considering the effects of long-term climate drivers is important - otherwise, you risk running into what you even admit is an issue. The shorter POR you use, the easier it is for an anomalous period to drop out, and for someone to come up with a disingenuous or incorrect conclusion as a result of newly changed averages. I'd perhaps consider a 1948-present POR, considering that's the length of time the Madison station has been at its current site at the airport - but if I'm using that POR, you won't catch me comparing anything to something that happened in the anomalously warm 1930s either. Also remember that the WMO or any other met agency is run by people - humans, error-prone and clouded by their own agendas by nature; not omniscient deities with 100% accurate understandings of climate patterns. I'm not claiming my understanding is perfect either - far from it. But the appeal to authority is itself anti-scientific. Records are a part of climate though. If a record is no longer possible under current climatic conditions, it doesn't make sense to include it in discussion of the current climate. And because the full range of what's climatically possible isn't going to occur in an abbreviated 30-year POR, that's another argument for using a POR that's longer than that! I'm also not the "only one" not using 1991-2020 normals - there's another post very early in the thread that references a 1941-2023 POR. Oh, but anything that happened in 1990 or before is irrelevant to the modern climate! Nice tinfoil hat straw man at the end there too: "oh anyone to the right of me politically has to be a fascist climate-denying QAnon freak who's solely focused on 'owning teh libtardz' at all costs!" I don't use the "updated normals" for the sole reason that I believe them to be inaccurate - especially considering monthly "normals" do differ from raw monthly averages for the same POR - in other words, they are fake data. The only person whose credibility you've damaged here is your own. |
|
|
|
Post by Steelernation on May 6, 2023 22:32:19 GMT -5
Also remember that the WMO or any other met agency is run by people - humans, error-prone and clouded by their own agendas by nature; not omniscient deities with 100% accurate understandings of climate patterns. I'm not claiming my understanding is perfect either - far from it. But the appeal to authority is itself anti-scientific. Records are a part of climate though. If a record is no longer possible under current climatic conditions, it doesn't make sense to include it in discussion of the current climate. And because the full range of what's climatically possible isn't going to occur in an abbreviated 30-year POR, that's another argument for using a POR that's longer than that! I'm also not the "only one" not using 1991-2020 normals - there's another post very early in the thread that references a 1941-2023 POR. Oh, but anything that happened in 1990 or before is irrelevant to the modern climate! Nice tinfoil hat straw man at the end there too: "oh anyone to the right of me politically has to be a fascist climate-denying QAnon freak who's solely focused on 'owning teh libtardz' at all costs!" The only person whose credibility you've damaged here is your own. The warming trend is clear no matter how many years the comparisons use. If the 1980s were cold, 1980-2020 would still be warmer than 1870-1900 in the vast majority of places. So a decade dropping out isn’t that big an issue. ”the full range of what's climatically possible isn't going to occur in an abbreviated 30-year POR” that just proves my point for using a long por for records. And who’s to say they’re 100% impossible? The Texas cold snap in 2021 set records people thought were impossible. And heat records are much more likely to be set so would a heat record in 1910 be impossible? No, so then you’d have to use two different pors for record highs and record lows by your logic. And an average is showing what is typical, not what’s possible so it makes sense to use different pors for both. Greysrigging lives in Darwin, a tropical place with a tiny annual temp range and tiny deviations. Not a good comparison to a temperate zone climate. Fort Collins’ lows have warmed by over 5 f, Darwin doesn’t even have a 5 f range in average highs. WMO people are humans, yes, but they know a lot more than we do. You don’t get sick and say “this is what’s wrong with me”, you go to the doctor because they know more than you do about your illness. So I think going with the authority isn’t an issue unless you just blindly follow them. AW argued for 1870-2020 normals because he didn’t believe in climate change so in his mind the 1870s and 2010s would’ve been exactly the same. You’re similarly right wing so it’s not far fetched to think that’s a possibility for you too even though you believe in climate change. Seems to me if you understood the climate is changing you wouldn’t want to use data from a colder period in your averages. |
|
|
|
Post by Cheeseman on May 6, 2023 23:02:15 GMT -5
Also remember that the WMO or any other met agency is run by people - humans, error-prone and clouded by their own agendas by nature; not omniscient deities with 100% accurate understandings of climate patterns. I'm not claiming my understanding is perfect either - far from it. But the appeal to authority is itself anti-scientific. Records are a part of climate though. If a record is no longer possible under current climatic conditions, it doesn't make sense to include it in discussion of the current climate. And because the full range of what's climatically possible isn't going to occur in an abbreviated 30-year POR, that's another argument for using a POR that's longer than that! I'm also not the "only one" not using 1991-2020 normals - there's another post very early in the thread that references a 1941-2023 POR. Oh, but anything that happened in 1990 or before is irrelevant to the modern climate! Nice tinfoil hat straw man at the end there too: "oh anyone to the right of me politically has to be a fascist climate-denying QAnon freak who's solely focused on 'owning teh libtardz' at all costs!" The only person whose credibility you've damaged here is your own. The warming trend is clear no matter how many years the comparisons use. If the 1980s were cold, 1980-2020 would still be warmer than 1870-1900 in the vast majority of places. So a decade dropping out isn’t that big an issue. Let's dig into that claim as far as it pertains to my location - for 1931-1960, the annual mean temp here was 46.9 F, compared to 45.4 F for 1961-1990. In fact, the 1931-1960 July mean of 73.0 F is still significantly warmer than the 72.0 F seen in 1991-2020, let alone the 71.1 F of 1961-1990! I'm not denying an overall warming trend - but clearly climatic variations are more significant than merely that, and it would be easy to use the tidbits I just provided to make a disingenuous "global cooling" claim. Austin and San Antonio both recorded their all-time record lows in 1949. Oklahoma City's all-time record low dates back to 1899. I'm not denying the lows seen in 2021 were significant - but there is a difference between "rare" and "unprecedented". Furthermore, there are places with very old record highs - Miami's only 100 F (38 C) high on record was in 1942. My location's all-time record high dates back to 1936, as is the case in much of the Midwest - and extreme highs seem to have been tempered a bit since then, making exceeding that mark unlikely. If it's not possible in the "current" climate averages period you're using, it's not a part of the current climate and shouldn't count as a record within the context of the current climate. To include these records implies that they are still relevant to the modern climate - with which I would agree, and the averages from 1936 and 1942 as a whole are still relevant as well IMO as they provide additional context that may be missing from a shorter POR. So where's the cutoff at which it becomes okay to use full POR averages? Better to stick to a consistent standard, rather than claiming one of us is boteving and the other is in the clear - either we should both be using 1991-2020 normals, or it's okay for both of us to use full POR averages. Defining trusting authority versus blindly following authority is bound to be nebulous - but I'd contend you have been advocating for blindly going along with the 30-year normal period precisely because it's what a certain met agency does. Is there some inherent property that makes 30 years better than 20 and 40? Or because I believe that more data collected over a longer period of time will allow for a better understanding of what is truly normal for a given climate, and because I recognize that short-period normals can easily be swayed by anomalous years or let alone decades, as we've discussed. It's just like if I were to discuss the average amount of change I find in a day. Would I go by my average over the past four days, which have been anomalously fruitful in that regard with a $4.50 jackpot this morning; or would I go by my average over the past several months, which is a lot closer to $0.50 than to $5? |
|
|
|
Post by Steelernation on May 6, 2023 23:24:35 GMT -5
Look at 1870-1900. Of course there’s going to be fluctuations, it’s not a linear increase. But given how different 1931-1960 and 1961-1990 were for those stats, that adds to my point that it makes sense to not lump them all together into one average. I’m gonna ignore your stuff about records because you’re clearly not understanding my point that averages and records represent two different things. I think everyone should use 30 year nórmale. But Darwin is so stable and has probably barely warmed so I can see why Greysrigging uses the full por. It’s much less problematic but I’m still on the side of 30 year normals. I’ll admit I don’t know enough to know why 30 is better than 20 or 40. I do know it’s better than 100 or 10. I also think they’re smoothed normals are bullshit so I don’t blindly trust them in general. A 10 f below average month skews a 30 year normal by only 0.3 f and there’s 10 f anomalies in the other direction too that will cancel that out so anomalous months don’t make much difference. Your change analogy doesn’t work because that’s assuming totally random fluctuation with no trend. A better example would be if your income steadily increased but you used your income from the first year you got hired as part of your average. Cheeseman |
|
|
|
Post by greysrigging on May 7, 2023 2:54:23 GMT -5
I'm happy to use the 1942- 2022 POR that the BOM publishes on their climate data site.
Mind you I recognise that there are 1991-2020 normals published by others for Darwin Airport ( ie on the Wikibox ) that show a definate warmer ( and wetter trend )... Yep its real ( the warming ) ... and to see this trend in a place that has varied little in over 80 years is concerning.
Our trend is that we have way more extreme heat days.... doesn't show up particulary in the means on a monthly basis because we also have more cloudy wetter days that suppress the max temps somewhat.
But simply; in the 21st century we have doubled and often tripled the number of +35c max temps recorded in a calendar year. Up until about 2000, about 10 such days per year. Now its about 30 such days per year....
2022 - +35c = 27
2021 - +35c = 44
2020 - +35c = 43
2019 - +35c = 45
2018 - +35c = 23
2017 - +35c = 24
2016 - +35c = 29
2015 - +35c = 21
2014 - +35c = 22
2013 - +35c = 14
2012 - +35c = 28
2011 - +35c = 4 ( record breaking wet season... over 3000mm at my place )
2010 - +35c = 19
2009 - +35c = 22
2008 - +35c = 20
2007 - +35c = 15
2006 - +35c = 20
2005 - +35c = 25
2004 - +35c = 25
2003 - +35c = 20
2002 - +35c = 26
2001 - +35c = 12
2000 - +35c = 12
|
|
|
|
Post by Benfxmth on May 7, 2023 5:36:26 GMT -5
Steeler is 100% right here. While not as much the case for stable climates with little temp changes (e.g. in the tropics), 30-year normals are sufficient to reflect a climate's averages in majority of places. The argument about warm and cold decades skewing 30-year averages is understandable, but on the other hand, with a 100+ year POR, in a climate that's significantly changed since (e.g. the places that are warming faster than average due to AGW + UHI), those years before said warming trend would skew down averages.
It's true that smoothed averages don't quite line up with raw averages, but on the other hand, raw averages are too flukey to be used for day-to-day averages, so maybe the best compromise would be to use smoothed averages for daily averages, and raw averages for monthly ones.
|
|
|
|
Post by cawfeefan on May 7, 2023 6:49:14 GMT -5
I agree with using 30 year normals as a baseline, and in most cases you can't discern an anomalous month from it. This just makes the March 1993 snow event in the SE US even more remarkable, as March is officially (but barely) the snowiest month in places like Birmingham, Alabama using 1991-2020 averages. (correct me if I'm wrong about this assumption since I'm not the most well versed with the weather there) But yeah, usually anomalies don't make much difference and get cancelled out.
|
|
|
|
Post by Cheeseman on May 7, 2023 8:47:24 GMT -5
Look at 1870-1900. Of course there’s going to be fluctuations, it’s not a linear increase. But given how different 1931-1960 and 1961-1990 were for those stats, that adds to my point that it makes sense to not lump them all together into one average. I’m gonna ignore your stuff about records because you’re clearly not understanding my point that averages and records represent two different things. The annual mean here from 1870-1900 was 45.8 F - warmer than 1961-1990, and similar to the full-POR average of 46.1 F! In fact, the warmest January, February, May, and December here are all pre-1900. All of this decadal variability only goes to show exactly what I'm trying to say: a 30-year period is not long enough to get an accurate gauge of a location's climate, as anomalous periods CAN last a decade or more and really throw off a 30-year average! It's like I'm talking in circles with you. What would be valid would be to note, for example, that a particular month was warmer than average using the full period of record, but cooler than average compared to a recent period - which, you will note, I did do in my March 2023 summary. In the off chance that the 2000s-2020s end up having been an anomalously warm period just as the 1961-1990 normal period was anomalously cool, watch the number of "cooler than normal" months increase substantially in the 2030s-40s as the normals reflect only the warmest period. There's also a big difference between not understanding your point about using different PORs for averages and records, and believing it to be bullshit. Darwin has seen a marked increase in 35+ C days in recent years as Greys' post above mine notes, making that portion of your premise incorrect. He also understands that records and averages are both a part of climate, and that if one is willing to look at the entire POR for records, one should also look at the entire POR for averages. Comparing a random month in, say, 1910 to solely modern averages is silly. Smoothed normals are bullshit, they're faked data - so why do you use them and expect me to do the same? Your analogy fails in that regard, as nobody ever asks about the average amount of money someone makes in a year throughout their career. Of course the $8 per hour I made working retail jobs as a teenager is irrelevant to my income level now, as my job duties aren't comparable. I would also like to think I have gotten better at spotting and picking up change over the years, though that doesn't make it any less incorrect to claim my average over the past four very fruitful days as my overall average. If I actually kept those statistics, I could say for example "this past Monday's $0.74 is below my year-to-date average, but still above my 2011-present average" - which sounds remarkably similar to the comment I made in my March 2023 summary and mentioned above. Steeler is 100% right here. While not as much the case for stable climates with little temp changes (e.g. in the tropics), 30-year normals are sufficient to reflect a climate's averages in majority of places. The argument about warm and cold decades skewing 30-year averages is understandable, but on the other hand, with a 100+ year POR, in a climate that's significantly changed since (e.g. the places that are warming faster than average due to AGW + UHI), those years before said warming trend would skew down averages. It's true that smoothed averages don't quite line up with raw averages, but on the other hand, raw averages are too flukey to be used for day-to-day averages, so maybe the best compromise would be to use smoothed averages for daily averages, and raw averages for monthly ones. While my climate has as a whole warmed over the past 150-odd years due to the combination of AGW and UHI as you mentioned, I still don't think the late 19th century is wholly incomparable to the present day - once again, my warmest Jan, Feb, May, and Dec were all pre-1900, and my coolest July was as recent as 2009. Also note that in general, I don't use daily averages - in general we make monthly summaries and comparisons; and it would be disingenuous to claim, for example, that a given year had a hot summer because July 13 in particular was x number of degrees warmer than average, as well as meaningless to claim that a day with no precipitation is "0.12 inches drier than average" - so that consideration becomes moot. I agree with using 30 year normals as a baseline, and in most cases you can't discern an anomalous month from it. This just makes the March 1993 snow event in the SE US even more remarkable, as March is officially (but barely) the snowiest month in places like Birmingham, Alabama using 1991-2020 averages. (correct me if I'm wrong about this assumption since I'm not the most well versed with the weather there) But yeah, usually anomalies don't make much difference and get cancelled out. Your example only illustrates the problem with using an abbreviated POR. Why should March, a month with a mean temp of 56.0 F under the same set of normals, be the snowiest month of the year? Full-POR averages clearly illustrate January as Birmingham's snowiest month, which makes complete sense given the climate. |
|
