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

[chubbs]

5 hours ago, ChescoWx said:

By the way Charlie that Phoenixville chart includes Ghastly ghost data from 1896 thru 1914.....wonder if that fake data impacted your charting??

5 hours ago, ChescoWx said:

Nope - Phoenixville 90 plus days are indeed decreasing

Once again you are jumping to the wrong conclusion. West Chester, Philadelphia and Kennett Square all agree well with the Phoenixville "ghost" data in the 1896-1914 period. All 3 stations show that the area had relatively few 90F days 100+ years ago. Makes sense, the "ghost" data is an IEM re-analysis product that uses nearby raw data.

The chart also shows that Phoenixville had a big spurious spike in 90F days starting in 1925. Starting your regression line in 1915 gives the bad Phoenixville data a big impact. Other than the 1925 and 1950 period, West Chester, Phoenixville and Philadelphia are all in reasonable agreement. The area has had a slow and irregular increase in 90F days. Of course the data isn't perfect: Phila transitioned from city to airport, West Chester cooled in 1970, and Phoenixville was too warm from 1925 to 1950 and too cool in the 1990s. All of those data problems worked to limit the increase in 90F days. The real increase may be even larger than the plot indicates.

As I stated above your analysis is biased by changes in the station population. As an illustration the 12 DEOS stations added between 2008 and 2014 only average 11 90F days per year. Much less than the Phoenixville, West Chester and Coatesville COOPs that dominate the county historical data. All you are showing is that rural parks are cooler than built up towns.

[chubbs]

47 minutes ago, donsutherland1 said:

Doesn't Atglen only go back to 2019? If your sample mix is changing over time, how do you know what seems to be a trend isn't, in fact, an artifact of the changing sample?

Typical Chesco, very misleading. Starts in 2010, a hot summer and per his table, many of the stations didn't operate over the entire period. Philly airport also shows a declining trend starting in 2010.

[ChescoWx]

1 hour ago, donsutherland1 said:

Doesn't Atglen only go back to 2019? If your sample mix is changing over time, how do you know what seems to be a trend isn't, in fact, an artifact of the changing sample?

2012 for Atglen

[ChescoWx]

18 minutes ago, chubbs said:

Typical Chesco, very misleading. Starts in 2010, a hot summer and per his table, many of the stations didn't operate over the entire period. Philly airport also shows a declining trend starting in 2010.

LOL!! so the cooling trend must be real if the UHI PHL Airport agrees!!

[donsutherland1]

1 minute ago, ChescoWx said:

2012 for Atglen

Thanks. That's still subsequent to your starting point. So, it appears that one is, in fact, dealing with a changing sample mix, which makes it tougher to discern the actual trend.

[ChescoWx]

34 minutes ago, donsutherland1 said:

Thanks. That's still subsequent to your starting point. So, it appears that one is, in fact, dealing with a changing sample mix, which makes it tougher to discern the actual trend.

If we exclude it the trend of course remains...

[donsutherland1]

3 hours ago, ChescoWx said:

If we exclude it the trend of course remains...

What are the trends for the last 30 years?

[chubbs]

The Avondale USCRN station which started in 2007 shows a slight upward trend. KMQS starting in 2008 is also increasing slowly. Given the noise in the 90F data these short-term trends may not be significant. Most of the other stations that are collecting data today started in 2012 or later, not 2010. Starting a 90F trendline in hot year like 2010 is misleading. Bottom-line - Considering how noisy the 90F data is there isn't much evidence that the long-term increasing trend established by Phoenixville and West Chester has changed. 

[ChescoWx]

8 minutes ago, chubbs said:

The Avondale USCRN station which started in 2007 shows a slight upward trend. KMQS starting in 2008 is also increasing slowly. Given the noise in the 90F data these short-term trends may not be significant. Most of the other stations that are collecting data today started in 2012 or later, not 2010. Starting a 90F trendline in 2010 is misleading. Bottom-line - Considering how noisy the 90F data is there isn't much evidence that the long-term increasing trend established by Phoenixville and West Chester has changed. 

Actually at Avondale 2N almost flat very slight upward trend mainly thanks to last year....after this low 90 plus season it will change! The clear data of remaining validated stations is a downward trend in 90 plus days!