TheClimateChanger

Climate scientists first laughed at a ‘bizarre’ campaign against the BoM – then came the harassment | Climate science | The Guardian

Just to highlight how robust these trends are, here is a comparison to observed surface air temperatures from 1927 to the present. One correction: the air temperature is rising slightly faster than the water temperature [2.1F per century for air, 1.9F per century for the lake temperature]. Top twenty coldest and top twenty warmest years (1927-2022): Looking at the top ten coldest Crib water temperature years, we can see there is a very strong correlation between air temperature and the nearby Lake Erie water temperature. The correlation is slightly less for the coldest years and I think that is because a cold spring and cold summer are more important for determining the water temperature ranking. For air temperature, the opposite is true - winters have more variability and can really build up large temperature deficits. Conversely, for the water temperature reading, it is not physically possible for it fall below 32F. So even the very coldest winters do not see the water temperatures drop below 32F. That's why years with cold summers [e.g., 1982 & 1992] are able to punch above their weight on the water temperature ranking. Even so, six of the ten coldest water temperature years are also among the top ten coldest air temperature years over the same time period. For the warm water years, the correlation is almost 1:1, because those years are mostly open water so these ice dynamics don't come into play. Among the warm water years, 9 of the top 10 [12 years, due to 3-way tie for 10th place] are also in the top ten for air surface temperature. The others are 12th, 14th, and 14th. Coldest Years [air temperature rank in fourth column) Warmest years [air temperature rank in fourth column]

And I have finished my analysis of water temperature readings at the Buffalo water intake crib (30' depth) from 1927 to the present. The water temperature off Buffalo is warming at 1.87F/century, with an r-squared value of 0.1706. The twenty warmest and twenty coldest years (as measured by annual mean water temperature) are as follows: The decadal mean water temperatures: This pattern closely mimics the annual mean air temperature at Buffalo, although water temperatures tend to run about 1.5-2 degrees warmer than the corresponding air temperatures. This is likely because the water temperature cannot drop below 32F, so they tend to run warmer in the winter - especially during cold winters. For comparison, the decadal annual mean surface air temperatures at Buffalo are given below:

As part of the analysis for Buffalo, I am also examining water temperature data for the Buffalo water intake crib, which dates back to 1927. The lake water temperature at Buffalo is warming somewhat faster than the surface air temperature. The largest changes are in the spring, and associated with earlier loss of ice from the east end of Lake Erie. The mean water temperature in the month of May has warmed about 4 degrees over the last century, per a linear approximation. As you can tell by some of the outliers, there used to be periods where this part of the lake was covered in ice deep into May (such as 1936 & 1971), but that does not happen anymore. The water temperature analysis is not yet complete, but the May scatter plot with linear trend line is shown below.

I've finished my analysis of Buffalo air surface temperatures, using all available data. The coldest decade was the 1830s with a mean temperature of 45.7F, albeit with only 2 years of data. The warmest decade is the 2020s with a mean temperature of 51.1F, albeit with only 3 years of data. The coldest decade with full data is the 1880s with a mean temperature of 46.2F. The warmest decade with full data is the 2010s with a mean temperature of 49.4F. A linear approximation for the warming in annual temperature is 1.6F per century over the period from 1830 to the present, with an r-squared value of 0.2776. There are significant seasonal variations in the rate of warming, ranging from no change in January to a rate of 3.0F per century in the month of May. Twenty coldest and twenty warmest years on record:

Fascinating stuff. Detroit used to have such a lovely summer climate. While the means aren't quite Yooper-esque, some of the individual months [especially warm-season months] were certainly Yooper-esque, like that 60.5F from August 1836. But it is certainly a northern Michigan climate, comparing very favorably to Traverse City's modern annual mean temperature of 46.9F [mid-19th Century Detroit being about 0.3-0.4F warmer, and Port Huron about 0.6F colder]. That 1836-1838 period was really something else. Just happened to correspond to the beginning of systemized weather records in Detroit too. Allegedly, this was assisted by the 1835 eruption of Cosiguina in Nicaragua, but curiously that eruption wasn't even as impressive as Pinatubo in 1991. Port Huron, Michigan:

Still a far cry from the winter of 1837, when the St. Clair River was still closed with ice on June 1, and ice was harvested until the Fourth of July. Thanks to the "Urban Heat Island Effect," conditions like these no longer occur.

It's fascinating to me how close that Smithsonian technique gets to approximating the true mean. I noticed in later years, they doubleweighted the 9 PM reading. So the formula became [7 + 2 + 9 + 9] / 4, where 7, 2 & 9 are the hours of the temperature reading, and that seems to eliminate most of the slight excess. I looked at Buffalo from 2016, and here's the comparison. True mean is 51.1F, average of 7 AM, 2 PM, 9 PM readings is 51.5F, average of 7 AM, 2 PM, and 9 PM with a doubleweighting of the 9 PM reading is 51.2F. Maximum Minimum Mean 7 AM EST 2 PM EST 9 PM EST AVERAGE SPEC. AVG 33.9 19.9 26.9 25 31 27 27.7 27.5 37.7 21.7 29.7 27 34 29 30.0 29.8 48.8 31.0 39.9 35 45 39 39.7 39.5 52.3 33.9 43.1 39 49 43 43.7 43.5 68.8 49.0 58.9 54 66 57 59.0 58.5 77.8 57.4 67.6 64 75 67 68.7 68.3 82.8 65.4 74.1 70 81 73 74.7 74.3 84.6 66.5 75.6 70 83 75 76.0 75.8 76.3 57.6 67.0 61 74 65 66.7 66.3 62.6 46.2 54.4 50 60 53 54.3 54.0 53.1 37.0 45.1 41 51 45 45.7 45.5 36.2 26.1 31.2 30 33 32 31.7 31.8 51.1 51.5 51.2

The 115th anniversary is approaching for a little known snowstorm that occurred on April 30, 1908. Most of the state saw accumulating snow, with about 6" near Columbus and 9" near Youngstown. From the April 1908 Ohio Monthly Weather Review:

I just don't know how anyone can doubt climate change. Look at the period from October 1836 to April 1837 in the Allegheny Arsenal [Lawrenceville, Pittsburgh] records. And the means of 48.2F and 46.9F aren't far from the coldest in the modern record [48.0F, in 1976]. But these are from a site 500' lower in elevation than the PIT Airport. But look at what a cooler climate, plus a major volcanic eruption in Nicaragua, was capable of doing. I feel like people don't realize how cold it used to be, and how in the natural climate state, we would probably be just several big volcanic eruptions in rapid succession from a new glacial advance. October 1836: 38.2F [The coldest in the modern record is 45.9F, or 7.7F warmer than this reading!] November 1836: 30.8F [The coldest in the modern record is 33.1F, or 2.3F warmer than this reading!] December 1836: 22.4F [Would tie 1963 for 2nd coldest in the modern record] January 1837: 16.8F [Would be the 2nd coldest in the modern record] February 1837: 25.6F [Would tie 20th coldest in the modern record] March 1837: 27.7F [Would be 2nd coldest in the modern record -- current 2nd place at 31.0F] April 1837: 31.3F [The coldest in the modern record is 43.8F, or 12.5F warmer than this reading!] Can anyone alive today even imagine an April with a subfreezing mean temperature at the latitude of Pittsburgh today? Unfathomable.

So, I was sharing some information about Pennsylvania climate from the 19th century. Thought I should share this information here too. These Allegheny Arsenal (Lawrenceville, Pittsburgh, PA) records are just absolutely lit. You can probably ignore the data from 1825-1827, which is oddly much warmer than the rest of the records. The period from October 1836 through April 1837 is absolutely insane. And these are from 704', adjacent to the Allegheny River (or about 500' below PIT airport). The overall means [around 51F] aren't too far from the 1991-2010 mean for PIT airport [51.8F], but again these are 500' lower in elevation and somewhat to the south within the boundaries of the present-day city of Pittsburgh [Lawrenceville was a separate borough at the time]. The elevation difference would account for about 1.5F of warming relative to the airport, even ignoring the change of latitude. The mean at PIT airport probably would have been about 49-49.5F at this time. But even this seemingly minor cooling would rewrite the record books - 48.2F in 1836, 46.9F in 1837, 49.1F in 1840, 48.9F in 1842, 47.8F in 1856 would all be among the five coldest in the modern records. For the record, these were calculated by three daily readings - 7 am, 2 pm, and 9 pm. Probably local solar time, as I believe this predates even standard time zones [and far predates daylight saving time]. I was curious how these would compare to modern means computed from max & min, and I looked at the data for 2011 using the hourly averages reported in each monthly LCD for 7 am, 2 pm, and 9 pm (standard time) and determined the difference is negligible [in that year, this method would have produced a warming bias of 0.3F for the yearly mean]. Obviously, you'd want to take a look at more data to confirm, but it looks like a small bias - and a positive one at that. So these readings are not being artificially cooled by the methodology. The biggest warm bias is in the early summer, because the 7 am standard [8 am EDT] / 7 am local solar time reading would be a couple hours after sunrise and miss the actual low. Anyways, the 1836-1837 is absolutely insane. This is what has been stolen from us snow lovers. People don't realize how cold it was, and how big of a difference just a couple degrees can really make. I mean we were probably perilously close to tipping into an ice age with these conditions. October 1836: 38.2F [The coldest in the modern record is 45.9F, or 7.7F warmer than this reading!] November 1836: 30.8F [The coldest in the modern record is 33.1F, or 2.3F warmer than this reading!] December 1836: 22.4F [Would tie 1963 for 2nd coldest in the modern record] January 1837: 16.8F [Would be the 2nd coldest in the modern record] February 1837: 25.6F [Would tie 20th coldest in the modern record] March 1837: 27.7F [Would be 2nd coldest in the modern record -- current 2nd place at 31.0F] April 1837: 31.3F [The coldest in the modern record is 43.8F, or 12.5F warmer than this reading!] The April 1837 reading probably appears impossible at first glance, but given the temperatures of the preceding six months, there must have been a massive snow cover in place at the end of March 1837 that gradually melted off throughout April offsetting the extreme sun angle. There's a lot of support for this extreme cold. Cosiguina erupted in 1835, dropping global temperatures by .75C. June 28, 1836 is the only time snow occurred in Sydney, AU when one inch fell. People have no idea how cold it used to be, or how cold the natural climate should be. We should be teetering on the edge of a new ice with a series of major volcanic eruptions. Next time someone complains about how cold it is in spring, just remember it used to be possible to have subfreezing temperatures in April at the latitude of Pittsburgh less than 200 years ago.

Just for the record, for the year of 2011, the Smithsonian mean for PIT would have been 53.1F, or just 0.3F warmer than the true mean. Like I said, this was an incredible approximation in the absence of a max/min thermometer.

Obviously, to get a better comparison of the Smithsonian Institution means compared to the true means, you would need to look at a longer time frame and more locations. But, in general, the differences are surprisingly small. Remember these Smithsonian people were smart - they may not have had modern technology, but they were resourceful and able to make do with what they had. Generally, there appears to be a slight warm bias in the Smithsonian means - this is particularly the case in the early summer, when the 7 am reading [8 am EDT] would be too late to capture the true minimum temperature. In general, there is already a couple degree rise by that point. This effect is, of course, negligible in the cold season when 7 am EST is around sunrise. It is worth noting that, a few of these old records, are said to have taken the first reading at sunrise. In those cases, the above effect would not be true and there may even be a slight cool bias in those records. However, the effect again would be very small. The main point is you can look at these old records obtained from three measurements and directly compare them to modern true means without the need for any adjustment since the error is not in one direction, small in degree, and generally a small warm bias. With the small warm bias, the old records are actually underestimating the amount of warming that has occurred.

I know some people will look at the Smithsonian Institution numbers I've shared from the 19th century, and say "but those are averages of 7 am, 2 pm, and 9 pm" readings and not averages of maximum and minimum temperatures, and my response would be that's there's little difference between the two. First, the times are on local solar time - which for most of Pennsylvania is similar to local standard time. So 7 am and 2 pm are roughly high and low, and 9 pm is roughly in the middle of the two. For the sake of comparison, let's look at the Pittsburgh readings for 2011. January was 24.2F. From the LCD, the 7 am (standard) mean was 21, the 2 pm (standard) mean was 28, and the 9 pm (standard) mean was 25F. The Smithsonian mean would be reported as 24.7F, or 0.5F warmer than the true mean. February was 31.8F. From the LCD, the 7 am (standard) mean was 29F, the 2 pm (standard) mean was 36F, and the 9 pm (standard) mean was 32F. The Smithsonian mean would be reported as 32.3F, or 0.5F warmer than the true mean. March was 39.2F. From the LCD, the 7 am (standard) mean was 33F, the 2 pm (standard) mean was 45F, and the 9 pm (standard) mean was 39F. The Smithsonian mean would be reported as 39.0F, or 0.2F cooler than the true mean. April was 53.3F. From the LCD, the 7 am (standard) mean was 48F, the 2 pm (standard) mean was 59F, and the 9 pm (standard) mean was 53F. The Smithsonian mean would be reported as 53.3F, or exactly the same as the true mean. May was 62.9F. From the LCD, the 7 am (standard) mean was 57F, the 2 pm (standard) mean was 70F, and the 9 pm (standard) mean was 63F. The Smithsonian mean would be reported as 63.3F, or 0.4F warmer than the true mean. June was 70.0F. From the LCD, the 7 am (standard) mean was 65F, the 2 pm (standard) mean was 78F, and the 9 pm (standard) mean was 71F. The Smithsonian mean would be reported as 71.3F, or 1.3F warmer than the true mean. July was 76.9F. From the LCD, the 7 am (standard) mean was 70F, the 2 pm (standard) mean was 85F, and the 9 pm (standard) mean was 77F. The Smithsonian mean would be reported as 77.3F, or 0.4F warmer than the true mean. August was 72.8F. From the LCD, the 7 am (standard) mean was 66F, the 2 pm (standard) mean was 81F, and the 9 pm (standard) mean was 72F. The Smithsonian mean would be reported as 73.0F, or 0.2F warmer than the true mean. September was 65.4F. From the LCD, the 7 am (standard) mean was 60F, the 2 pm (standard) mean was 71F, and the 9 pm (standard) mean was 64F. The Smithsonian mean would be reported as 65.0F, or 0.4F cooler than the true mean. October was 52.8F. From the LCD, the 7 am (standard) mean was 47F, the 2 pm (standard) mean was 59F, and the 9 pm (standard) mean was 52F. The Smithsonian mean would be reported as 52.7F, or 0.1F cooler than the true mean. November was 46.9F. From the LCD, the 7 am (standard) mean was 42F, the 2 pm (standard) mean was 53F, and the 9 pm (standard) mean was 47F. The Smithsonian mean would be reported as 47.3F, or 0.4F warmer than the true mean. December was 37.5F. From the LCD, the 7 am (standard) mean was 34F, the 2 pm (standard) mean was 42F, and the 9 pm (standard) mean was 38F. The Smithsonian mean would be reported as 38.0F, or 0.5F warmer than the true mean.

I'm still not finding a urban heat island effect. Like this data for Grampian in Clearfield County, has a mean of 45.7F, with individual years ranging from 42.7F [1875] to 48.6F [1878]. If I compare this to the modern records for nearby DuBois, we find a mean of 47.8F. Again, 2.1F doesn't sound like a ton. But it's massive in this context. 45.7F is the third coldest annual mean in the DUJ records dating back to 1963. And DUJ is 400' higher in elevation than this site, and somewhat further north. There are 10 years in this data set that are colder than anything observed at DUJ since 1963. We would need a high-end VEI 8 just to have a chance to experience what is a natural climate in this state for a few years.

Wikipedia reports the following for West Chester (Blodget's means are in parentheses, along with the difference): January: 30.7 [29.7] [-1.0] February: 31.6 [31.5] [-0.1] March: 40.9 [37.3] [-3.6] April: 51.2 [48.7] [-2.5] May: 61.5 [59.9] [-1.6] June: 70.2 [69.1] [-1.1] July: 74.9 [74.0] [-0.9] August: 73.2 [71.5] [-1.7] September: 66.4 [64.3] [-2.1] October: 55.1 [53.9] [-1.2] November: 44.6 [42.4] [-2.2] December: 33.8 [32.9] [-0.9] Annual: 52.8 [51.3] [-1.5] Yeah, this looks similar to the other data I've seen. Looks like spring has the most warming, followed by autumn. It warms more rapidly into spring nowadays, and summer's warmth is more resilient. Does not decline as quickly into fall. The 1.5F difference doesn't sound like much, but it's a big change over the course of the year. And the change is more dramatic for cold years. The modern data set shows around 50F being the lowest, but there were several years in the high 40s in the numbers published by Blodget - including a 47.8F in the notoriously cold year of 1875. For comparison sake, the threaded records for PHL report 50.1F for the mean that year - by far, the coldest annual mean in the threaded records. Second place is 51.8F in 1904. I'd be willing to bet 47.8F is downright impossible for an annual mean in West Chester today. Unless there was a supervolcano eruption that lofted huge quantities of sulfur into the atmosphere.

People have no idea how cold it really used to be. I mean can you imagine any low elevation site in the State of Pennsylvania having a year with an annual mean of 40.0F nowadays? Compare that to today: Or even look at the Signal Service records for Pittsburgh. This doesn't look that much cooler than today - but these were taken 500 feet lower than the modern airport records. That's enough of an elevation change to easily account for 2 degrees or more. Even still, there are some doozies mixed in there. 46.9F in 1827, 47.8F in 1856. The modern record cold year (again 500' higher in elevation) is 48.0F in 1976. If there were records at the airport location, it could have been 45 or 46 degrees in those years.

And here's data for Bucks County from 1790-1887. Looks like a mean of around 51.9F, ranging from 48.8F in 1816 (year without a summer) to 56.3F in 1828. We can also see that the year without a summer wasn't an old wives tale. Mean temperatures were 57F in May, 64F in June, 68F in July, 66F in August, and 62F in September. Very chilly.

Just comparing it to your post 1890 data, it seems like it used to be significantly cooler in Chester County in the 1850s-1880s.

And for Pocopson, in Chester County, he had 21 years of data [1853-1873]. Mean annual temperature was 51.9F, ranging from 49.9F in 1868 to 54.7F in 1853. Unfortunately, this data set ends before the very cold year of 1875. Regardless, I feel like these temperatures are lower than anything reported in recent decades.

How do your records compare to those published in Lorin Blodget's seminal work on the Climate of Pennsylvania, published in 1889? He had 33 complete years of data for West Chester from 1855-1887, plus all of 1888 except for December, and computed an annual mean of 51.3F, ranging from a minumum of 47.8F in 1875 to a maximum of 53.7F in 1870.

Maybe I'm off base here, but, to me, it always seems the urban heat island effect is blown way out of proportion. Sure if you're measuring temperatures in a dense city, particularly over a dark rooftop or asphalt, it's going to be somewhat higher. But other than that, there doesn't seem to be any strong relationship between temperature and population. Just look at the temperatures this month around here, where is the urban heat island effect? PIT: 54.4F [Allegheny Co., pop: 1.25M] [Elev: 1201 feet] [lat: 40.50N] PHD: 54.0F [Tuscarawas Co., Ohio, pop: 93K] [elev: 892 feet] [lat: 40.47N] ZZV: 54.2F [Muskingum Co., Ohio, pop: 86K] [elev: 899 feet] [lat: 39.94N] DUJ: 53.4F [Jefferson Co, Pa., pop: 44K] [elev: 1804 feet] [lat: 41.18N] HLG: 55.8F [Ohio County, W. Va., pop: 42K] [elev: 1194 feet] [lat: 40.17N] MGW: 56.5F [Monongalia Co., W. Va., pop: 106K] [elev: 1227 feet] [lat: 39.65N] JST: 52.9F [Cambria Co., Pa., pop: 133K] [elev: 2274 feet] [lat: 40.31N] AOO: 55.0F [Blair Co., Pa., pop: 123K] [elev: 1467 feet] [lat: 40.30N] Here is data from the second most populated county in Pennsylvania, at the second busiest international airport in the State, versus data taken from a bunch of small airfields in mostly rural counties. Where is the urban heat island effect in this data? Almost like temperature varies by latitude and elevation, and not by population.

Just had a brief snow shower here.