TheClimateChanger

One of only two years (out of 136) at Harrisburg, where the minimum temperature has not dropped below 65F.

Appears to be a good chance that this month will be the first to not have at least one low temperature below 55 at DuBois-Jefferson County Airport (dating to 1963), and one low temperature below 50 at Bradford Regional Airport (dating to 1957). While it hasn't been a super hot month, it has been consistently warm - particularly at night - leading to elevated minima. DuBois [days with minimum <55F] Bradford [days with minimum <50F]

I am working, so I didn't have time to respond. 1) I did not intend to cherrypick a date. I ran the analysis after seeing well-known Alaskan climatologist Brian Brettschneider's map earlier today on Twitter. I wasn't even aware there was a negative trend in snowfall. 2) Brian's map ran from 1973 - a period of 50 years. I selected 1950 to integrate more data, and specifically chose that date because (1) that's around the time most observations moved to airports from city centers; and (2) observations prior to that timeframe include a lot of missing data and generally seem less reliable to me. I did not realize the 1960s were an unusually snowy period in eastern Pennsylvania. It was one of the least snowy decades at Pittsburgh. 3) Yes, the linear trend line is not a good fit. But that is, indeed, the best fit linear trend line calculated by Excel for each of those datasets. The r-squared value is very low, because the data is so noisy. You are not going to generate a strong fit with such noisy data.

The trend at Williamsport is even worse. A loss of about 26" per century. Wow - I had no idea the situation was so bleak on that side of the state.

Wow! I redid this trend for Harrisburg, PA, and snowfall has plummeted at a rate of 20" per century since 1950 for Harrisburg. That is wild.

For context, here is the graph for Pittsburgh since 1950. The first few years are from AGC, the rest at PIT. I used 1950, as I feel that's the start of reasonably reliable snowfall records. Not to mention, the early years were all downtown.

So far, we have been fairly lucky, although snowfall has trended downwards over the past several decades. Since 1950, I calculate the trend as a loss of 7" per century, so a fairly slow decline. But looking at the map posted by Brian Brettschneider, it looks like the oranges and reds aren't looming too far to our south. Do you guys think we will eventually reach a level of warming where snowfall in Pittsburgh plummets on the order of 25 or 50 percent?

I had responded earlier, noting the daily high shift from the actual observations. This site reported a 7 am observation time - which was pretty rare in the 1930s. Most sites observed temperatures between 5 and 8 pm, which resulted in "double counting" of high temperatures, relative to a midnight-to-midnight day. Also missed the occasional midnight daily lows that occur. What appears to happen is the algorithm used to clean these records up assumed the high temperature was being assigned to the day on which it was measured, and so it shifted them all back since the high actually would have occurred the prior day. However, it appears that this observer was already making that shift, ascribing the max temperature at 7 am to the prior day and the low to the day on which it was observed. So the algorithmic shift incorrectly bumped all the highs to the day prior to when they occurred.

The next few days will likely drop those figures a bit, but modeling shows hotter weather moving in by middle of next month and perhaps through the end of the month. Therefore, I felt pretty comfortable that the final figures should be fairly similar to the current ones (and possibly, a bit higher depending on how the heat performs later this month). The fact that upper 40s in northern Pennsylvania is a big deal now just shows how these changes have been normalized. There have been years where the mean minimum at Bradford for the entire month of July was in the 40s, including as low as 47.3 in 1971.

Has the NWS shared the 2022-2023 snowfall map yet?

Records are from Jefferson County Airport.

Runaway "urban heat island effect" leading to record or near record high minimum temperatures this month, even in some of the most remote places in the Commonwealth, such as Jefferson and McKean Counties. Just not sure how we are going to be able to stop all of this urban waste heat from polluting our forests, streams and oceans. DuBois Bradford Mount Pocono Williamsport Comment: I have my doubt about the validity of the 71.1 from 1901. LOL. Scranton/Wilkes-Barre Harrisburg Reading

It's crazy to see how this urban heat island effect keeps growing and even affecting places hundreds of miles from any city, such as McKean County, Pennsylvania.

Made at a surprise run at the 50s this morning, but fell just short with a low of 60.

If the 1936 heat wave occurred this year, Bastardi would be downplaying it by saying its not that hot on a per capita basis. Looking at the actual data, July 1936, as a whole, was actually around or slightly below current July normals in most places where people actually lived. And outside of a one week intense heat wave, downright chilly (especially at night) compared to recent Julys. Funny how it was only hot in places where no one lived. I suspect this temperature pattern would not score highly on Bastardi's population density metric. Toronto (downtown), Canada Excluding 7-days from 7/8 to 7/14: 79.6/58.6 Mean: 69.1F Boston, Massachusetts No heat wave to knock out, only a few discrete hot days scattered early-mid month. New York, New York Excluding 7-days from 7/8 to 7/14: 83.4/64.8 Mean: 74.1F Buffalo, New York Excluding 7-days from 7/8 to 7/14: 76.5/61.3 Mean: 68.9F Rochester, New York Excluding 7-days from 7/8 to 7/14: 78.9/59.0 Mean: 68.9F Hartford, Connecticut Excluding 7-days from 7/8 to 7/14: 81.0/60.4 Mean: 70.7F Harrisburg, Pennsylvania Excluding 8 days from 7/8 to 7/15: 83.8/63.9 Mean: 73.8F Philadelphia, Pennsylvania Excluding 7 days from 7/9 to 7/15: 83.4/66.9 Mean: 75.1F Scranton/Wilkes-Barre, Pennsylvania Excluding 7 days from 7/8 to 7/14: 82.7/59.0 Mean: 70.3F Williamsport, Pennsylvania Excluding 7 days from 7/8 to 7/14: 85.9/56.9 Mean: 71.4F Erie, Pennsylvania Excluding 7 days from 7/8 to 7/14: 78.8/61.8 Mean: 70.3F Washington, District of Columbia Excluding 8 days from 7/8 to 7/15: 85.3/67.3 Mean: 76.3F Hagerstown, Maryland Excluding 7-days from 7/9 to 7/15: 83.4/59.3 Mean: 71.4F Baltimore, Maryland Excluding 7-days from 7/9 to 7/15: 85.0/67.9 Mean: 76.4F Detroit, Michigan Excluding 7-days from 7/8 to 7/14: 82.6/58.9 Mean: 71.4F Lansing, Michigan Excluding 7-days from 7/8 to 7/14: 81.2/56.2 Mean: 68.7F Alpena, Michigan Excluding 7-days from 7/8 to 7/14: 80.9/48.0 Mean: 64.5F Akron, Ohio Excluding 7-days from 7/8 to 7/14: 84.0/58.0 Mean: 70.9F Toledo, Ohio Excluding 7-days from 7/8 to 7/14: 82.2/61.9 Mean: 72.1F Morgantown, West Virginia Excluding 7-days from 7/9 to 7/15: 85.9/56.5 Mean: 71.2F Elkins, West Virginia Excluding 7-days from 7/9 to 7/15: 80.7/57.3 Mean: 69.0F Clarksburg, West Virginia Excluding 7-days from 7/9 to 7/15: 84.9/59.3 Mean: 72.1F Milwaukee, Wisconsin Excluding 8-days from 7/8 to 7/15: 77.0/63.7 Mean: 70.3F Chicago, Illinois Excluding 8-days from 7/8 to 7/15: 80.9/66.2 Mean: 73.5F

It's kind of funny. Ryan Maue, on Twitter, was comparing low temperatures at Phoenix to those observed at Tuscon to illustrate the supposed UHI effect. However, he did not mention that Tuscon is 1500' higher in elevation. Wonder why he didn't compare the lows at Sky Harbor Airport to those at Furnace Creek? After all, nobody lives in Death Valley. Sky Harbor Airport is actually closer in elevation to Death Valley than to Tucson. Rather unsurprisingly, the low temperatures in Death Valley are significantly higher than those observed at PHX. Not because of more urbanization at the former than the latter, but rather the difference in elevation!

Wow - those lows! Clearly, it must be the urban heat island effect!

I did a little comparison between PHX and Grand Canyon Airport (GCN) - the latter of which is quite literally in the middle of nowhere when it comes to human population density. Unfortunately, records for GCN were only available back to 1997 for GCN, but that still allowed for about a quarter century of data. From 1997-2022, I calculated a warming trend of 7.3F/century for daily maxima at GCN, 4.3F/century for daily minima, and 5.7F/century for daily mean temperature. For the same period at PHX, I calculated a warming trend of 10.0F/century for daily maxima, 7.8F/century for daily minima, and 8.8F/century for daily mean temperature. While the relatively short period of analysis makes drawing definitive conclusions a bit challenging, a couple of observations. It is warming at an alarming rate at both sites; however, PHX does show a residual warming trend of about 3.0F/century compared to GCN. This residual may be related to a growing urban heat island effect. However, Dr. Spencer's research would suggest that, at some point, the UHI effect will essentially "max out" and stop being a significant contributor to the trend. Second, maximum temperatures have climbed more than minimum temperatures at both sites over this interval. This is inconsistent with UHI effect. Mr. Maue has some graphics showing a collapse in the diurnal temperature range (DTR) at Phoenix since the mid 20th century. This data suggests that trend has reversed somewhat since the late 1990s, and DTR has increased by a small bit since that time. Lastly, I feel it is disingenuous to compare the PHX trend to the globally averaged trend and suggest all of the excess is due to UHI. It should be clear that the regional climate around PHX is warming considerably faster than the global average.

Ryan Maue seems to be suggesting that most of the warming in Phoenix is due to the urban heat island effect - what do you guys think?

All I would say is that if a relatively normal, or some would say cool, July today results in mean temperatures similar to 1988 - which was front page, headline news everywhere for its heat and drought - what is going to happen when the conditions that created the drought and heat wave of that year recur in this new era? Makes me wonder if we aren't destined for some absolutely shocking summertime heat waves in the near future.

They've also bumped the normal temperatures up so much in the most recent release that it makes comparisons difficult. Look at Toledo, Ohio, it shows up as normal on the map; but if the month ended today, it would be the 24th warmest [out of 151 years] -- decidedly warmer than the long-term average. Prior to 2010, it would have been the 18th warmest [out of 138 years], as six of the warmer Julys have occurred in the 13 year period from 2010 to 2022. You take this same "normal" month and place it in the 1970s, 1980s, or 1990s and it would suddenly be considered a hot July.

Actually, the number of 85+ days, 90+ days, and 95+ days in the month of July all show a positive trend over that timespan - albeit rather small. On the order of 2 additional days each century for the first two categories, and 1 additional day every 1,000 years for the latter. It's very rare to reach 95+ at Pittsburgh International. No use examining the trend for 100+ days, when there have only been three such days in recorded history at the airport. Elsewhere on the forum, I got flamed for suggesting the 1988 and 1995 temperature records were somewhat inflated. But it's worth noting the hygrothermometer in place during that era (the HO-83 model) has a known warm bias, which largely led to its replacement in the ASOS units. Funny how the only 3 100+ days at PIT happened to occur in that narrow 10-ish year window while that unit was in place. Given the bias, only the 103F would probably be 100+ with the current equipment. The other two dates would have likely been 97-99F range. Source for some context: https://climateaudit.org/2007/08/22/the-ho-83-hygrothermometer/ So while I would agree that there has been no significant increase in extreme heat, I don't agree that it is necessarily getting less extreme. The actual data show the occurrence of extreme heat in July has been fairly steady since the early 1950s, but with a very small positive trend. I do think humidity and dew points are on the rise overall, though, but it's more difficult to obtain that data. Obviously, that wasn't the case for the start of this summer, but since the beginning of July it's been rather muggy.

I should clarify. I'm not a skeptic of the urban heat island effect per se, but rather I'm a skeptic of the theory that it is a significant driver of the global changes that have been observed. Actually, noted climate change skeptic, Dr. Roy Spencer, had an interesting analysis of the urban heat island effect back in March. His chart, which passes the eyeball test to me, suggests the effect is greatest as population increases at low density and then gradually decreases (although still positive) as population density increases. Interestingly, the chart also suggests that the effect has been decreasing over time. In other words, there was a more significant positive difference in temperature between densely populated urban areas and remote rural locations in the late 19th century than in the 21st century. Source: https://www.drroyspencer.com/2023/03/ While not framed this way by Dr. Spencer (for obvious reasons, given his skepticism), this actually gives credence and support to a negative temperature trend due to UHI for certain locations - namely those that have been super densely populated since the late 19th century (i.e., places like New York and Chicago). If the difference in temperature was greater in the 1800s in densely populated urban sites than zero population density locations, then it stands to reason that the UHI effect has actually imparted a small negative component to the trend for such sites. In other words, if the UHI effect has decreased for a given population density and those sites have not experienced an appreciable increase in population density over the period, then there must be a negative trend component attributable to the effect. This might seem to be a counterintuitive finding - and Dr. Spencer expressed surprise towards it - but several compelling theories were raised in the comments. My thinking is it's probably related to improved siting of urban stations. In the 1800s, the urban stations were almost exclusively sited in window sills (yes, you read that correctly) or on rooftops. Now they are mostly ground-based, often in an area of green space or parkland, or at an airport with more open grass than the dense urban core. But some other interested theories were offered in the comments - perhaps increased tree canopy cover, better building materials / insulation preventing loss of heat to the environment, and/or the advent of skyscrapers has led to more heat being trapped higher in the atmosphere as opposed to at the level of the thermometer. Those theories would suggest the UHI effect itself has been decreased, whereas my initial thought was simply improved temperature recording - i.e., there hasn't been a decrease in UHI, but rather some of the temperature difference which was previously ascribed to UHI was actually just a result of poor station siting and not actually UHI effect. In reality, it may be a combination of both of these factors.