TheClimateChanger

It’s interesting, or, at least I found it interesting. But I was looking at the July numbers from Pittsburgh International (1953-present) and found a curious trend. From 1953-1986, only three years (out of 34) met or exceeded the current July mean of 74.1F. Those years were 1955, 1966, and 1980. Since 1987, 17 out of 36 years met or exceeded that temperature. So the current temperature went from a reading that would be reached less than once every ten years to one that’s reached or exceeded in more than 2 in 5 years. Since 2010, 7 of 13 Julys have exceeded 74.1F, or more than once every other year. I think some looking at this might chalk it up to land use changes or increase in urban heat island effect, but I’m just not sure if there’s been enough land use changes around the airport to explain this anomaly. What do you guys think? Are our Julys getting hotter?

So I looked into this, and for whatever reason, the official numbers at Brookville have the high temperatures shifted forward one day versus the raw observation. So they actually reported a high of 88 and low of 44 on the 16th, and a high of 73 and low of 37 on the 30th, and so the actual observed high was exactly the same as observed at AGC on those dates. Lows were much lower, but I would argue it's the AGC low temperatures that are too high [perhaps due to the rooftop exposure in place at the time]? The lows of 37 and 39 to close the month might seem unbelievable, but I would point out that New Cumberland Lock and Dam in northern West Virginia, just west of Beaver County, Pennsylvania, had lows of 36 degrees on both of those dates. In any case, it's the official Pittsburgh numbers that are the outlier that month. I would suggest part of it is from the observations being at AGC at the time, and a big part the rooftop exposure of the equipment resulting in higher minima than would have been observed on the ground. Regardless, 75.0F at AGC is not that impressive. That's 0.4F less than the month to date there. And just since the ASOS was installed at AGC in 1998, this figure has been exceeded 10 times, excluding this year (which is still in progress). Assuming it stays above 75F this month, that will be 11 times in 25 years. New Castle 1N - a GHCN station - had a mean of 71.2F, 1.9F below the 1991-2020 average. Like the Brookville data, this data also has been time-shifted from the observations, so there are weird diurnal range discrepancies. Morgantown, W. Va. was 72.8F, 1.3F below the 1991-2020 average. The mean minimum temperature of 58.2F is fourth lowest out of 140 years of records.

I have been seeing a lot of discussion about July 1936 on Twitter, and the incredible heat of that month. Looking at the data from Pennsylvania, it appears to have been a very pleasant month by modern standards. Outside of a fairly extreme 8-day heat wave, it was quite cool away from the cities and their inflated temperature readings. This is one heck of a way to run the allegedly hottest month on record. Looking at MPO, we see a mean temperature of 66.0F, roughly 5 degrees cooler than 2023. Looking at the Brookville FAA Airport site in Jefferson County, Pennsylvania - which is not threaded with the current Jefferson County Airport [KDUJ], but would be the best source of comparison -- we find a mean temperature of 67.7F, is nearly 5 degrees cooler than 2023 to date at DUJ. Also an unbelievably, refreshing mean monthly low of 49.6F - something never seen today in Pennsylvania at low elevations (maybe even high?).

Will be interesting to see how the last two weeks of the month play out. Very warm start to July. The current mean of 70.3F at Bradford would be good for second place (out of 66 years), behind only the incredibly scorching July of 2020. The current mean of 72.5F at DuBois would be good for 4th place (out of 59 years): NOTE: The late 1980s data is probably unreliable, since it was likely measured with an HO-83 thermometer, which has been shown to produce a warming bias of greater than 1 degree Celsius, especially in the summertime. So, in reality, all of the warmest months are probably from the 2010s and 2020s. Looking at some locations with longer "threaded" histories. Some of the earlier figures are probably too warm due to issues with equipment, exposure, and siting. But a similar story emerges. Due to the recent trends, I've included the corresponding ranking excluding post-2010 years, which I feel provides a better historical context for the heat. 10th warmest on record at Harrisburg, PA (out of 136 years) [prior to 2010, this would be ranked 4th warmest]: 9th warmest on record at Mount Pocono, PA (out of 114 years) [prior to 2010, this would be ranked 4th warmest]:

I've noticed this trend too. A lot of blame for warming placed on UHI, but then when you run the numbers, it's the rural, secluded sites surrounding by montane forests that have the most warming. I think the greenhouse warming effect may be more significant in places that used to radiate amazingly well, and it could be impacting mountainous, high elevation sites moreso than lower elevation sites through an increase in the conditions that lead to temperature inversions. That, or the urban sites changed locations and exposures a lot - from window sills, to rooftops, to modern ground-based measurements at suburban airport locations. I think there's little, if any, evidence of any significant urban heat bias in the temperature records. I would even go further and argue that the urban heat island effect itself is exaggerated, and to some degree, the result of substandard siting and exposure of equipment, coupled with the fact that the dense urban cores tend to be located at lower elevations than the surrounding suburbs - a natural result of cities popping up around bodies of water, whether it be the coastal Plain or a broad valley at the confluence of multiple rivers. Obviously, the city resting hundreds of feet below the outlying areas will be hotter. When I ran the numbers last year based on airport records, I found little dependence on population of the surrounding county. Instead, almost all of the variance could be attributed to elevation and latitude. Airports, in particular, tend to be at higher elevations than the cities they service (for obvious reasons), or, alternatively, immediately adjacent to large bodies of water that have a cooling effect relative to locations even a few blocks inland. I'm certainly a skeptic of the claim that most of the warming is due to urban heat and land use changes. Remarkable claims require remarkable evidence, and I don't see any evidence for that claim. I do see a lot of evidence of greenhouse gas warming, though.

This piqued my curiosity, so I decided to take a look at the numbers for DCA and it's the same exact story as Pennsylvania. 15th warmest (out of 152 years), so that is 90th percentile temperature. We live in a world where 90th percentile heat at the hottest time of the year is considered to be a major victory, because we are so used to it being 95-99th percentile each year. It's just the nature of the world we are living in. DCA is easily beating years like 1988 & 1995, from my youth, that were considered to be unbelievably hot. Even just a tiny fraction of a degree below 1999 to date.

It certainly has been anything but cool. July is already the hottest month of the year, and the records show many areas are running close to the 90th percentile for the first two weeks of the month [i.e., warmer than 9 out of 10 years]. Even at PIT, where the thermometer has oddly been running a bit cooler than most surrounding sites, it's 20th warmest out of 78 years at the airport site, so roughly 75th percentile. 75-90th percentile for heat at the hottest time of the year is certainly not cool, by any stretch of the imagination. No matter how it's distorted. I like to look at actual data, not conjecture about lack of "extreme heat" - whatever that means. That's a subjective thing. It's been very hot. There were 17 heat-related hospitalizations at the Ed Sheeran concert this past weekend in Pittsburgh, even though the air temperature was only lower 80s.

All I know is I've had to run the air conditioning since late June. There was a cool spell in late June, but it was marred by dense smoke and so I had to keep the windows shut due to air quality. Since then, it's mostly been low/mid 80s and humidity, with lows in the low/mid 60s. We used to get occasional crisp days in the 70s, where you would go to an outdoor concert or event and it would be in the low/mid 60s and you'd actually be chilly without a jacket or long sleeves. We don't get that anymore.

Yes, I have a lot of confidence in those readings. It's measuring air temperature, not rocket science. They used sheltered mercury thermometers that were properly calibrated, and carefully observed the temperature three times daily. These aren't records from a random Joe, but War Department records from the Signal Service Corps generally taken by army surgeons. The tri-daily mean actually produces a warming bias compared to the average of maximum and minimum, so if anything it was cooler than reported. In any case, it's a LOT easier to produce spuriously high temperatures than low temperatures. The readings also line up with each other, as well as our understanding of climate change and impacts of land use changes and urbanization over time. And they are in line with the earliest Weather Bureau readings which show a lot of cool temperatures, especially in 1875.

The 8-14 day outlook valid for this past week from the beginning of them month was cold and wet. While there may have been more rainfall, temperatures have verified warmer than normal. The week 3 & 4 outlook also had cold and wet, and it looks like continued normal to warmer than normal weather for the next week. The 8-14 day again wants to bring cooler temperatures the following week, but I have my doubts.

Looks like we'll notch our first official 90 of July tomorrow. I realize some parts of the area have already seen 2 or 3 days of 90+, but none at PIT yet.

Oh, accidentally deleted the original post. Just showing the means at various sites in Pennsylvania, for the first 13 days of the month. Our cool summer has come to a fiery end. So much for all the long range models showing cool weather. This humidity has been relentless.

Confirmed. By average minimum temperature, it's been 10th warmest (of 129 years) at IPT, 8th warmest (of 136 years) at MDT, 4th warmest (of 58 years) at DUJ, and 5th warmest (of 65 years) at BFD.

Interesting. Here is my best guess for the lower Great Lakes, using 7 reference sites. Data before 1860 is limited, and subject to larger deviation from actual. Edit: Corrected error in data for 1857 and added 8th reference site. Top 5 Coldest: 1. 1842; 2. 1862; 3. 1839; 4. 1903; 5T. 1834; 5T. 1857 Top 5 Warmest: 1. 1949; 2. 1919; 3. 2005; 4T. 2021; 4T. 1933 Comparing to NCEI rankings for 1895-present only: Top 5 Coldest: 1. 1903; 2. 1926; 3T. 1927, 1916, and 1928 Top 5 Warmest: 1. 1949; 2. 1919; 3. 2005; 4T. 2021; 4T. 1933 For the entire Great Lakes basin, NCEI has: Top 5 Coldest: 1T. 1916 & 1926; 3. 1958; 4. 1928; 5T. 1915, 1917 & 1982 1903 is 8th coldest and 1927 is 9th coldest. My dataset has 1958 and 1982 tied for 6th coldest since 1895, and 1917 & 1915 tied for 11th coldest since 1895. Some variation, but this my dataset covers only a subset of the full GL basin. Top 5 Warmest: 1T. 1919 & 2005; 3. 1933; 4. 1949; 5. 2021 All match, except for order.

What data are used to generate these? June 1856 was not the warmest June on record in the Great Lakes. That is incorrect. Smithsonian records have mean temperatures of 68.0F at Chicago, 70.7F at Detroit, and 70.6F at Cleveland - and these are based on tri-daily measurements, which have a warm bias compared to the average of the maximum and minimum - particularly pervasive in early summer, when 7 am local solar time is hours after sunrise. That is not even close to years like 1919, 2005, 1933, and 1934. That has to be wrong. The more recent data looks similar to official government statistics, except that the most recent years appear to have been systematically cooled relative to the past - which makes little sense, considering most of the biases in the earlier data would result in a warming bias.

This is actually a misconception. Not only are the values smoothed and normalized, but advanced statistical tools are employed to take it into account changes in exposure, location, and trends. For example, the 1991-2020 30-year average monthly mean for ORD in July is 74.6F; however, the 1991-2020 normal is 75.4F. The latter number is much closer to the current 15-year (2008-2022) running average of 75.3F. A "normal" July would actually be 0.8F warmer than the average from 1991-2020. Similarly, the 1991-2020 30-year average monthly mean for CLE in July is 73.6F; however, the official 1991-2020 normal is 74.5F. The latter number is exactly even with the current 15-year (2008-2022) running average of 74.5F. A "normal" July is 0.9F warmer than a simple arithmetic mean.

Overperformed on the highs today. Reached 90 at both Zanesville & Morgantown. Forecast showed a 0% and 6% chance of that occurring.

Chicago was also a bit drier in the past, particularly from April through August. The current normal is just shy of 5" more than the old MDW norms, with the bulk of the difference occurring in the warm season. I thought maybe it was because MDW is simply drier than ORD, but NowData shows a normal annual precipitation of 40.88" for MDW.

Doesn't strike me as particularly warm outside of December and March, certainly nothing crazy. November 1877 December 1877 January 1878 February 1878 March 1878

On the whole, the 1800s were quite a bit colder. The official coldest yearly mean in Chicago history is 45.1F, from 1875. However, the mean temperature was 45F in 1855, 43F in 1856, 44F in 1857, 44.9F in 1860, 45.1 in 1862, 44.3F in 1863, 42.5F in 1864, and 44.3F in 1865, according to records published by the United States Weather Bureau in 1947. The normals for 1947, which at that time were based on the 50-year period from 1878-1927, adjusted to airport exposure, are given below. The normal temperature for December was 26.9F, for January 22.3F, for February 24.9F, and for March 34.5F. By comparison, the current normals for O'Hare are: December, 30.5F; January, 25.2F; February, 28.8F; and March, 39.0F. These values are 3.6F, 2.9F, 3.9F, and 4.5F warmer than the reported 1947 Midway norms.

They used to post excellent comparative data in the annual and monthly summaries. It looks like the downtown business district was about 2-2.5F warmer than MDW in the 1940s in general in the wintertime. Below is 1947, by way of example. It says the city was 32.5F in January 1947, but I calculate the mean to be 31.1F. Using my corrected mean, the courthouse site was 2.4F warmer than MDW and 1.5F warmer than the University of Chicago in January 1947. In February, it was 2.0F warmer than MDW and 1.4F warmer than the University of Chicago. In December, it was 2.2F warmer than MDW and 1.0F warmer than the University of Chicago. For the annual means, the differences were much smaller... 1.1F warmer than MDW and 0.8F warmer than the University of Chicago. Annual mean maximum temperature was actually 1.5F cooler than MDW, but the mean minimum temperature was 3.8F warmer.

Looks like it could just be the use of different data. The monthly outlooks appear to rely on data from certain sources. The week 3-4 outlooks appear to have more human input. Based on this discussion, it looks like a cooler and wetter second half of July, with periods of smoke and haze from Canadian wildfires. Prognostic Discussion for Week 3-4 Temperature and Experimental Precipitation Outlooks NWS Climate Prediction Center College Park MD 300PM EDT Fri Jun 30 2023 Week 3-4 Forecast Discussion Valid Sat Jul 15 2023-Fri Jul 28 2023 This Week 3-4 temperature and precipitation outlook is for the period of July 15 - July 28, 2023, which is climatologically near the warmest two-week period of the year for much of the United States. June has been characterized by a tripole pattern of temperatures across the contiguous U.S. (CONUS): 1) record warmth in the southern CONUS, particularly in Texas, 2) below normal temperatures in the middle tier of states from California to the East Coast, and 3) above normal temperatures from the Pacific Northwest to the Great Lakes. This pattern is largely consistent with the weak teleconnection that is associated with El Niño during summer, which consists of zonally-elongated troughing of 500-hPa heights across the middle of CONUS with anomalous ridging both to the north and the south. Currently, Niño 3.4 sea surface temperature (SST) anomalies exceed +0.9 degree C, thus El Niño may continue to act as a weak background forcing. In fact, the Climate Prediction Center's (CPC) current Week 2 outlook is remarkably consistent with a weak El Niño teleconnection persisting. Other physical drivers considered for this outlook include the Madden Julian Oscillation (MJO), extratropical SSTs, soil moisture, the North Atlantic Oscillation (NAO), the Pacific-North American pattern (PNA), the long-term trend, and the ongoing Canadian wildfires. The Real-time Multivariate MJO (RMM) indices are currently weak, and they are forecasted to remain so over the coming weeks, with no coherent propagating signal. Thus, the MJO is not expected to have a meaningful impact on the Week 3-4 circulation pattern. Extratropical SSTs are below normal along the southern Alaska coast and the West Coast. They are above normal in the Gulf of Mexico and along the East Coast. The greatest soil moisture anomalies currently reside in the Mid-Mississippi River Valley, where much of the region is below the 10th percentile of climatology. However, conditions are likely to improve there as the current Weather Prediction Center's (WPC) quantitative precipitation forecast shows much of the region receiving 2 to 4” of rain over the coming week. The NAO is forecast to remain negative through Week 2, while the PNA is forecasted to be neutral. However, dynamical model guidance for Week 3-4 indicates that the NAO may persist in its negative phase while the PNA may trend toward positive values. These phases of NAO and PNA would increase the chances for troughing over eastern CONUS. With respect to the long-term trend, it is toward warmer temperatures across all of CONUS and Alaska. Finally, the Canadian wildfires have been a permanent fixture this summer. Smoke from these fires has intruded into CONUS regularly under anomalous northwesterly flow. Hazy skies produced by smoke generally reflect insolation, decreasing day time highs slightly. However, at night, smoke is a relatively inefficient longwave absorber/re-emitter and thus has little impact on nighttime lows. Therefore, the overall impact of the smoke is to lower temperatures slightly. The Week 3-4 temperature and precipitation outlooks take into consideration the El Niño teleconnection, the aforementioned physical drivers, and a blend of dynamical model guidance from the European Centre for Medium-Range Weather Forecasts (ECMWF), Climate Forecast System version 2 (CFSv2), Japan Meteorological Agency (JMA), and, to a lesser extent, the The Global Ensemble Forecast System version 12 (GEFSv12). The GEFSv12 was an outlier with respect to the other models. The forecasted 500-hPa height anomalies indicate ridging centered over the Aleutians and broad positive anomalies over Canada and western CONUS. Some troughing is centered over the Great Lakes in Week 3 that relaxes toward neutral anomalies during Week 4. With respect to western CONUS, the positive height anomalies are largest in the Desert Southwest and are near zero in the Pacific Northwest during Week 3. This overall pattern resembles the tripole, El Niño-like teleconnection described above. For the temperature outlook, above normal temperatures are forecasted in the southern CONUS. The Desert Southwest has the highest probabilities of above normal temperatures, being located under the ridge. Moreover, some internal, experimental tools are showing elevated probabilities of excessive heat in the region. A broad region of weakly enhanced probabilities of below normal temperatures are indicated from South Dakota southward and eastward to the Ohio Valley and Great Lakes. While most model guidance forecasted near zero anomalies in this region, the eastern CONUS has consistently verified below normal this summer due to persistent northwesterly flow and possibly due to the dampening influence of smoke on daytime highs. The smoke and haze is likely to continue, as the anticipated ridging over Canada promotes conditions favorable to wildfires and also provides northwest flow that advects the smoke into the eastern CONUS. Above normal temperatures are favored in the extreme northern tier of CONUS, where the impact of any ridging in western Canada will be most felt. Finally, most guidance agrees on enhanced probabilities of below normal temperatures in southwestern Alaska under northerly flow and above normal temperatures in northeastern Alaska under westerly flow. With respect to precipitation, northerly flow due to the presence of upstream ridging over the Aleutians promotes below normal precipitation over southern Alaska and above normal precipitation in northern Alaska. Below normal precipitation is likely over the Desert Southwest, under the influence of ridging. Above normal precipitation is forecasted from the Northern Plains to the East Coast, consistent with anomalous upper-level cyclonic flow in the region. Finally, for Hawaii, the multi-model ensemble from the Subseasonal Prediction Experiment (SubX) shows the highest probabilities of above normal temperatures in the northwestern islands, consistent with observed SST gradient. Below normal precipitation is favored in the southeastern islands and above normal precipitation is favored in the northwestern islands.

Really misleading about how cool it was, though. Climate report shows 3.1 degrees below normal, which makes it sound like it was a freezing month. But it would have only been 0.6F below normal in the 2000s or 0.8F below normal in the 1990s using the normals then in effect. They’ve inflated these “normals” so much it’s hard to compare to when we were younger.