Chester County PA - Analytical Battle of Actual vs. Altered Climate Data

[ChescoWx]

So let's go ahead and use the adjusted NCEI average annual temperatures for Chester County PA from NCEI this century. These are the adjusted altered figures Charlie....what does this tell us?  Even after adjustments by NOAA PHL is indeed warming almost 70% faster than Chester County PA - the PHL UHI is a problem even NOAA shows it....

[chubbs]

20 hours ago, ChescoWx said:

Bad assumption only using the warmest station which is Phoenixville 1E - to better determine the true Chesco average you should add in the nice mix we have today of high and low elevations to capture the true average Chester County average temperature. Heck the averages I derive from these stations is in reality supports almost exactly the calculated average temperatures posted by NCEI on their site for their recent Chester County PA average temperatures. If NCEI develops an average temperature for the county that is statistically the same as what I calculate from the raw stations. See below since 2004. NCEI must think the mix of stations I use is strong correct? 

I have no problem with your current set of stations. As you say it provides good coverage and matches NCEI. The problem isn't the current stations, its the difference between the current stations and the older stations. There isn't a single current station in the county that is as warm as any of the 3 county stations in 1945. You can't get our climate history right without properly accounting for the changes in county stations with time. Note also that consistency is much more important than coverage when tracking long term changes in temperature. 

[chubbs]

On 5/22/2026 at 12:42 PM, ChescoWx said:

So let's go ahead and use the adjusted NCEI average annual temperatures for Chester County PA from NCEI this century. These are the adjusted altered figures Charlie....what does this tell us?  Even after adjustments by NOAA PHL is indeed warming almost 70% faster than Chester County PA - the PHL UHI is a problem even NOAA shows it....

The difference in slope is not statistically significant. That said, wouldn't be surprised if PHL had a some warming enhancement from UHI; but, based on comparisons to other sites it is small compared to the region-wide warming in recent decades. More data is needed to make the case though. Note that the Avondale USHCRN station has a higher slope than PHL. Are you saying that the warming there is due to UHI?

[ChescoWx]

No matter how you slice it using either the raw or altered adjusted data. The gap is clearly widening between PHL and Chester County

[chubbs]

17 hours ago, ChescoWx said:

No matter how you slice it using either the raw or altered adjusted data. The gap is clearly widening between PHL and Chester County

But it does matter how you slice it; and, you continue to cherry pick favorable comparisons and ignore the rest. I stated above that the slope difference between NCEI and PHL between 2000 and 2025 was not statistically different. Expanding the yaxis doesn't change that result.

Looking over a longer period and re-normalizing the yaxis, we see that the delta between Chester County and the Philadelphia Airport has stayed the same since 1970. The delta goes up and down but without a long-term trend; 2000 just happens to be a minimum. Doesn't matter whether you use NCEI or high quality sites like Coatesville, Avondale USCRN, and DEOS. The answer is the same. 

 

[ChescoWx]

3 hours ago, chubbs said:

But it does matter how you slice it; and, you continue to cherry pick favorable comparisons and ignore the rest. I stated above that the slope difference between NCEI and PHL between 2000 and 2025 was not statistically different. Expanding the yaxis doesn't change that result.

Looking over a longer period and re-normalizing the yaxis, we see that the delta between Chester County and the Philadelphia Airport has stayed the same since 1970. The delta goes up and down but without a long-term trend; 2000 just happens to be a minimum. Doesn't matter whether you use NCEI or high quality sites like Coatesville, Avondale USCRN, and DEOS. The answer is the same. 

 

But again the facts remain - either way we look at the data both raw and altered still shows a widening gap since 1970 - no reason to think that gap will not continue to widen. 

[rygar]

this thread is like an Alexa vs Google Home conversation

[chubbs]

On 5/25/2026 at 9:40 AM, ChescoWx said:

But again the facts remain - either way we look at the data both raw and altered still shows a widening gap since 1970 - no reason to think that gap will not continue to widen. 

Your "raw data" doesn't look like the raw data for individual stations that I posted upthread. Why? The station population that you are averaging is cooling with time which distorts the data. Here is a plot of Chester County stations minus Coatesville, the most stable station between 1970 and 2025. Positive values indicate that the station is warmer than Coatesville, negative cooler.

The station population cools you are using with time. Over 55 years Coatesville goes from being the warmest station to an average station. The 19 station average warms by 1.5F between 1970 and 2025. That is larger than the "raw data" increase for PHL shown in your chart. Indicating that the positive slope in your chart is due solely to the change in station population. Reposting my table which shows that the raw data from individual stations doesn't show any difference in warming between Chester County and PHL. When you factor in the bias from the changing station population, your analysis agrees with mine. 

[ChescoWx]

2 hours ago, chubbs said:

Your "raw data" doesn't look like the raw data for individual stations that I posted upthread. Why? The station population that you are averaging is cooling with time which distorts the data. Here is a plot of Chester County stations minus Coatesville, the most stable station between 1970 and 2025. Positive values indicate that the station is warmer than Coatesville, negative cooler.

The station population cools you are using with time. Over 55 years Coatesville goes from being the warmest station to an average station. The 19 station average warms by 1.5F between 1970 and 2025. That is larger than the "raw data" increase for PHL shown in your chart. Indicating that the positive slope in your chart is due solely to the change in station population. Reposting my table which shows that the raw data from individual stations doesn't show any difference in warming between Chester County and PHL. When you factor in the bias from the changing station population, your analysis agrees with mine. 

Inventing and mixing and matching  data I see - Coatesville is not the most stable station for those 55 years....heck they only have 13 years at 1SW and 25 years at another location 2W. That only adds up to 38 years over a 55 year span. There has been no Coatesville COOP data for the last 20 years!! 

[forkyfork]

why does chester county matter so much

[ChescoWx]

5 minutes ago, forkyfork said:

why does chester county matter so much

Because it is where I live....LOL!

[chubbs]

3 hours ago, ChescoWx said:

Inventing and mixing and matching  data I see - Coatesville is not the most stable station for those 55 years....heck they only have 13 years at 1SW and 25 years at another location 2W. That only adds up to 38 years over a 55 year span. There has been no Coatesville COOP data for the last 20 years!! 

I knew you would find some reason to dismiss.  My stability assessment is based on comparing Coatesville to NCEI and other regional stations without station changes. Coatesville tracks NCEI closely. Yes there are three stations; but, the three stations are well matched for annual average temperatures; and, there are no major station changes. You would get the same answer if you used NCEI instead of Coatesville.

 

[Cobalt]

Do we have equal right to open up a thread for each of the other 3142 counties in the US?

[Santa Claus]

imagining the managing director of a multibillion dollar catastrophe bond fund at a reinsurance house making trades based on the weather station at scumhill road in chester county pennsylvania

[ChescoWx]

13 hours ago, Cobalt said:

Do we have equal right to open up a thread for each of the other 3142 counties in the US?

Please do Cobalt that would be great! Thanks!!

[ChescoWx]

As we get into the summer months lower elevation (<450 ft asl) areas of Chester County PA will see at least one heat wave (defined as 3 or more consecutive days with highs at 90 degrees or above) while the relatively higher ridge locations (>600 ft asl) only average a heat wave every couple of years or so. Below sorted by highest to lowest elevation is a ranking of all reporting stations since 2000 with their recorded heat waves. Our lowest elevation spot at Phoenixville averages a whopping two to four times the number of heat waves as almost all other locations across the county.

[ChescoWx]

Even the longest running continuous and warmest station in Chester County at Phoenixville has a clear trend toward less of the so called "heat waves" 

[ChescoWx]

On 5/29/2026 at 1:29 PM, chubbs said:

I knew you would find some reason to dismiss.  My stability assessment is based on comparing Coatesville to NCEI and other regional stations without station changes. Coatesville tracks NCEI closely. Yes there are three stations; but, the three stations are well matched for annual average temperatures; and, there are no major station changes. You would get the same answer if you used NCEI instead of Coatesville.

 

Welp except for what you have stated before were the multiple moves from a relatively more urban locale to rural at Coatesville 1SW and of course the increase in elevation with the 1983 move to Coatesville 2W and then over to KMQS.......

[ChescoWx]

Maybe an even better example of the Urban Heat Island Problem at PHL Airport is to compare the airport to NWS COOP observations at Octoraro Lake on the SW border of Chester County at about 40 nautical miles due west of the airport. Octoraro Lake is a relative low elevation spot (260 ft asl) and that NWS COOP station is clearly seeing decreasing heat waves while PHL continues to have increasing "heat waves". The only material difference is rural vs UHI

 

 

[chubbs]

13 hours ago, ChescoWx said:

Welp except for what you have stated before were the multiple moves from a relatively more urban locale to rural at Coatesville 1SW and of course the increase in elevation with the 1983 move to Coatesville 2W and then over to KMQS.......

The Coatesville move from urban to rural occurred in 1948, well before the 1970-2025 period in the chart. Yes the Coatesville stations have different elevations, but there is little impact on annual average temperatures, hence the good agreement between the Coatesville series and NCEI over the 1970 to 2025 period.

[chubbs]

14 hours ago, ChescoWx said:

Even the longest running continuous and warmest station in Chester County at Phoenixville has a clear trend toward less of the so called "heat waves" 

The chart shows more heat waves at Phoenixville, when you consider the spurious warmth at Phoenixville in the 1930 to 1940 period that is well documented in this thread. 

 

[chubbs]

13 hours ago, ChescoWx said:

Maybe an even better example of the Urban Heat Island Problem at PHL Airport is to compare the airport to NWS COOP observations at Octoraro Lake on the SW border of Chester County at about 40 nautical miles due west of the airport. Octoraro Lake is a relative low elevation spot (260 ft asl) and that NWS COOP station is clearly seeing decreasing heat waves while PHL continues to have increasing "heat waves". The only material difference is rural vs UHI

 

 

But your Phoenixville chart shows increasing heat waves in the same 1978 to 2025 period. There are other explanations for the discrepancy. Station changes at Octoraro Lake being the most likely. You've been making the same cherry-picked comparisons with PHL for a decade now. Invariably the story changes when additional stations are included. Its about time for a KPTW update

[ChescoWx]

4 hours ago, chubbs said:

But your Phoenixville chart shows increasing heat waves in the same 1978 to 2025 period. There are other explanations for the discrepancy. Station changes at Octoraro Lake being the most likely. You've been making the same cherry-picked comparisons with PHL for a decade now. Invariably the story changes when additional stations are included. Its about time for a KPTW update

But of course during that 1978 to 2025 period Phoenixville 1E is warming at almost a 50% faster rate than the rest of Chester County. The Phoenixville area is about as least representative of the entire county with about 60% of Chesco lying at over 400ft above sea level

[chubbs]

20 hours ago, ChescoWx said:

But of course during that 1978 to 2025 period Phoenixville 1E is warming at almost a 50% faster rate than the rest of Chester County. The Phoenixville area is about as least representative of the entire county with about 60% of Chesco lying at over 400ft above sea level

The mistake you are making is assuming that your viewpoint and calculations are correct without looking at the complete weight of evidence.. Over the 1978 to 2025 period Phoenixville warms at the same rate as Coatesville and NCEI Chesco. Its a red flag that you don't match Phoenixville or Coatesville or NCEI.

No surprise either, the station network you are using is cooling with time. Well documented in this thread and shouldn't be this difficult to understand. You are giving "unrepresentative" Phoenixville a 25% weight in 1978 but only 6% weight in 2025. Devault, almost as warm as Phoenixville, also gets a 25% weight in 1978. Coatesville the coolest station in 1978 is an average station in 2025.  You are using a different representation of the County in the 1970s and 2025. Different again in 1945 with Coatesville, West Chester, and Phoenixville all in much warmer locations.

 

 

[ChescoWx]

Since 1893 we have seen a cooling trend during the month of May here in Chester County PA

[ChescoWx]

Of note Spring of 2026 was the 14th warmest Spring since 1893 - conversely Winter 2025-26 was our 14th coldest! Our weather does have a way of balancing out!

Below is our Spring Temperature trends since 1893....only some slight warming evidenced by the data.

[ChescoWx]

Below is the Chester County PA area daily climate records for June since 1893. Also below is the June warmest high temperatures by state. Of note 32 states set their June record highs before 1960 this is similar to here in Chester County where 26 of our 30 June daily record highs all occurred prior to 1960.

[TheClimateChanger]

On 6/5/2026 at 4:07 PM, ChescoWx said:

Below is the Chester County PA area daily climate records for June since 1893. Also below is the June warmest high temperatures by state. Of note 32 states set their June record highs before 1960 this is similar to here in Chester County where 26 of our 30 June daily record highs all occurred prior to 1960.

This is what Martz’s June statewide high-temperature record map looks like if you present the most recent occurrence of each record, correct a few omissions/data errors, and add in the 2025 records that were set or tied across New England. With the updated/tied values included, 23 of 51 U.S. jurisdictions (50 states plus DC) have set or tied their statewide June high-temperature record in just the past 14 years, from 2012 through 2025. That is a huge chunk of the record book, especially considering that many of these records extend back into the 19th century.

A lot of the changes come from the brutal 2012 heat wave, when numerous states tied or set June records. Then 2025 added another cluster in New England. The result is a map that looks a lot less like “the old records still dominate” and a lot more like what you would expect in a warming world: the upper tail keeps getting pushed higher, and old records keep getting matched or exceeded.

Here is a table showing all June statewide record high temperatures set or tied from 2012-2025:

State / Jurisdiction	June Record	Year
Colorado	114°F	2012
Georgia	110°F	2012
Illinois	109°F	2012
Kansas	118°F	2012
Kentucky	111°F	2012
Maryland	106°F	2012
Missouri	112°F	2012
Nebraska	116°F	2012
New York	104°F	2012
North Carolina	107°F*	2012
South Carolina	113°F	2012
Tennessee	110°F	2012
Virginia	106°F	2012
District of Columbia	104°F	2012
Idaho	116°F	2015
Hawaii	96°F	2019
Oregon	119°F	2021
Washington	120°F	2021
Massachusetts	105°F	2025
Maine	101°F*	2025
New Hampshire	102°F	2025
Rhode Island	100°F	2025
Vermont	103°F	2025

[TheClimateChanger]

8 minutes ago, TheClimateChanger said:

This is what Martz’s June statewide high-temperature record map looks like if you present the most recent occurrence of each record, correct a few omissions/data errors, and add in the 2025 records that were set or tied across New England. With the updated/tied values included, 23 of 51 U.S. jurisdictions (50 states plus DC) have set or tied their statewide June high-temperature record in just the past 14 years, from 2012 through 2025. That is a huge chunk of the record book, especially considering that many of these records extend back into the 19th century.

A lot of the changes come from the brutal 2012 heat wave, when numerous states tied or set June records. Then 2025 added another cluster in New England. The result is a map that looks a lot less like “the old records still dominate” and a lot more like what you would expect in a warming world: the upper tail keeps getting pushed higher, and old records keep getting matched or exceeded.

A few notes on the edits:

Using Martz's own data, I changed the format to recognize the most recent date on which a given high was set. Showing only the earliest occurrence, of course, is going to favor earlier years.

For Maryland, I recognized the higher 106°F reading from Baltimore Science Center in 2012, which was omitted by Martz.

For North Carolina, I added the 107°F tie from 2012 at a cooperative station that was missing from Martz's data.  Note that several RAWS stations were even hotter, in the 107–109°F range, but I did not recognize those here.

For Maine, Fryeburg and other coop stations reached 101°F in 2025, matching the value Martz had listed. Rachel Carson RAWS reportedly reached 103°F, but again, I went with the lower conventional station mark.

For New York, I removed the absurd 105°F reading from 1919. No other station in New York was above 98°F that month. The site that supposedly reached 105°F is in the elevated northern interior, and on the same day several nearby stations reported highs only in the 88–92°F range. The same station also supposedly had a low of 50°F, implying a 55°F diurnal range. That reading is not credible. I used 104°F from 2012 instead, though even that may be too high. The next value would be 103°F, which was most recently reached in 2025.

There are also a few older records that are probably questionable, but I left them in place because the goal here was mainly to update Martz’s map rather than fully reconstruct every state record from scratch.

Pennsylvania’s 107°F from 1933 is almost certainly erroneous as it is far above surrounding observations. But I left it because the next-highest readings were either from that same year at other stations, including on the opposite side of the state, or from the same general era. In other words, changing it would require a deeper reconstruction rather than a simple update.

Wyoming is another interesting case. There is a reported 114°F reading from 1988 that Martz rejects apparently because it was several degrees higher than other Wyoming observations. But that one does not strike me as obviously impossible. Montana set its state record in 1988, Minnesota tied its June record in 1988, and the Wyoming site was near western Nebraska, where there were observations in the 110–112°F range. Even if the 114°F were 3°F too high, it would still match the 111°F record from 1919 shown here. Note that the 111°F record recognized by Martz was also the highest by several degrees that month.

So yes, there are likely other discrepancies and judgment calls buried in the record book. But the big picture does not depend on any one marginal station. Once you present the most recent tied occurrence, correct the obvious omissions/data issues, and add the 2025 New England records, the map changes dramatically.

[ChescoWx]

10 hours ago, TheClimateChanger said:

A few notes on the edits:

Using Martz's own data, I changed the format to recognize the most recent date on which a given high was set. Showing only the earliest occurrence, of course, is going to favor earlier years.

For Maryland, I recognized the higher 106°F reading from Baltimore Science Center in 2012, which was omitted by Martz.

For North Carolina, I added the 107°F tie from 2012 at a cooperative station that was missing from Martz's data.  Note that several RAWS stations were even hotter, in the 107–109°F range, but I did not recognize those here.

For Maine, Fryeburg and other coop stations reached 101°F in 2025, matching the value Martz had listed. Rachel Carson RAWS reportedly reached 103°F, but again, I went with the lower conventional station mark.

For New York, I removed the absurd 105°F reading from 1919. No other station in New York was above 98°F that month. The site that supposedly reached 105°F is in the elevated northern interior, and on the same day several nearby stations reported highs only in the 88–92°F range. The same station also supposedly had a low of 50°F, implying a 55°F diurnal range. That reading is not credible. I used 104°F from 2012 instead, though even that may be too high. The next value would be 103°F, which was most recently reached in 2025.

There are also a few older records that are probably questionable, but I left them in place because the goal here was mainly to update Martz’s map rather than fully reconstruct every state record from scratch.

Pennsylvania’s 107°F from 1933 is almost certainly erroneous as it is far above surrounding observations. But I left it because the next-highest readings were either from that same year at other stations, including on the opposite side of the state, or from the same general era. In other words, changing it would require a deeper reconstruction rather than a simple update.

Wyoming is another interesting case. There is a reported 114°F reading from 1988 that Martz rejects apparently because it was several degrees higher than other Wyoming observations. But that one does not strike me as obviously impossible. Montana set its state record in 1988, Minnesota tied its June record in 1988, and the Wyoming site was near western Nebraska, where there were observations in the 110–112°F range. Even if the 114°F were 3°F too high, it would still match the 111°F record from 1919 shown here. Note that the 111°F record recognized by Martz was also the highest by several degrees that month.

So yes, there are likely other discrepancies and judgment calls buried in the record book. But the big picture does not depend on any one marginal station. Once you present the most recent tied occurrence, correct the obvious omissions/data issues, and add the 2025 New England records, the map changes dramatically.

I love when climate zealots simply eliminate stations they deem absurd....LOL!