NORTH CAROLINA’S CLIMATE AND IT’S SIX PRIMARY WEATHER STATIONS
The National Weather Service has designated six weather stations as Primary Local Climatological Data sites (PLCD) in North Carolina.¹ They are located in Asheville, Charlotte, Greensboro, Raleigh, Wilmington and Hatteras, representing the breadth of the state from east to west. Their locations are depicted on the above map.
These “primary” stations are what the National Weather Service also refers to as “First Order” stations and are the highest quality weather stations in the United States, being subject to the most stringent quality control measures. There generally are three kinds of weather stations in descending order of quality and reliability, First Order (also known as Primary), Secondary and then COOP. First Order stations are monitored by professionals and reliably record hourly (and even sub-hourly) observations and are subject to additional and more rigorous quality checks, including manual quality checks, than all other stations. There are only 274 First Order Stations in the country and they have more instrumentation and include more data than all other types of weather stations.²
This article examines the historical climatological data from these six Primary weather stations by looking at their average temperature, average maximum temperature and average minimum temperatures over the 100+ years of records. First the historical temperatures for the three categories are determined by adding up all the annual temperatures and dividing by the number of years of records available. Then the average temperatures for the past 30 years are calculated. Next, it is determined whether that reading is higher than the average; if so, then whether or not another 30-year period starting before 1950 also had a higher than average temperature in that category is determined.
The results are set out below but in summary, this review found that in only three out of 18 instances were there rises in average temperatures that also did not have a corresponding rise in the average temperature being measured in a 30-year period starting prior to 1950.
Here is an image of the installation that includes the Primary/First Order weather station in Raleigh, North Carolina (id #13722) located at the Raleigh-Durham Airport (RDU)³.
Here is a close-up look at another Primary weather station⁴:
The RDU Primary station has been in place since 1944 and is professionally managed and monitored, while recording hourly and sub-hourly observations. It has a 100% record of days recording and reporting the weather data discussed in this article and presently records 95 weather parameters. Through the National Oceanic and Atmospheric Association (NOAA) ThreadEx program⁵, historical weather observations from as far back as 1887 by its predecessor station are included in the weather records provided.⁶ Use of weather stations in the NOAA ThreadEx program is an accepted, recognized method of assessing regional climate trends in the United States.⁷
This article examines historical climate data from the six Primary/First Order weather stations in North Carolina and assesses whether significant, enduring and unprecedented temperature trends are occurring, or not. This article also will discuss why the weather records from these six stations represent a viable and accurate reflection of historical climate trends in these regions of the state rather than the methods employed in prior climate change studies of using many, unidentified lesser quality and less reliable weather stations.
North Carolina is known to be divided into three regions, the Mountains in the West, the Piedmont region in the middle and then the Coastal area to the East.
A Primary/First Order station is located in each of these three regions. However, prior climate change studies in North Carolina have not focused on utilizing these six weather stations in order to examine regional climatic trends. An examination of past climate change studies reveals that they conclude that sweeping conclusions on climate trends for the entire state are not possible due to the geographic and climatic variety present throughout the state and that such assessments should be done on a more local, regional basis.
HISTORY OF CLIMATE CHANGE STUDIES IN NORTH CAROLINA
There have been three major climate change studies in the past couple of decades.
Two leaders of the State Climate Office of North Carolina (SCONC), published a 2002 paper in Environment International titled Analysis of Climate Trends in North Carolina (1949–1998).⁸ This study found that North Carolina has one of the most complex climates in the country and that most of the climate trends found were local, not statewide. The authors of the study stated, “Local scale climate analysis can more accurately represent the complex climate that exists in North Carolina …”
The study found on a statewide basis that while temperatures in the last 10 years of the study were warmer than usual, they were not warmer than the temperatures experienced in the 1950’s. Maximum temperatures had remained fairly constant and there was no widespread pattern showing increasing minimum temperatures across the state.
2015 Climate Change in North Carolina Report
This report, by the State Climate Office of North Carolina, looked at a much longer time period, investigating weather records back to the late1800’s.⁹ The report found no warming trend in the 113-year study period starting from 1895. While there was some warming since the mid-1970’s, it was found to be not unprecedented and similar to the warming observed from 1910–1950. The 2015 SCONC report stated that in North Carolina there is “no meaningful trend that is associated with global warming.”
While the report noted a rise in minimum temperatures, this was only in urban areas and was not found in rural areas. The report concluded that this possibly was due to urban development, but it was not due to global warming. The SCONC summed up its view on climate change in 2015 by pointing out that it also had analyzed dozens of other temperature data “(including seasonal patterns, days with extreme high and low temperatures, a variety of temperature thresholds, degree days, growing season data, frost data), and one common thread joined them all: “… the trends are generally flat …” The State Climate Office concluded that, “… local climate variability is so high in North Carolina that significant trends are difficult to deduce.”
2020 Climate Study — North Carolina
This study was done in response to an order of the Governor of North Carolina after his 2016 election.¹⁰ The above two prior climate change studies were not mentioned in the study, nor is there any mention of the history of climate changes studies in the state. It addressed climate issues statewide, and also in the three main regions of the state, Western Mountains, the Piedmont and the Coastal Plain. The 2020 study set forth conclusions as to climate trends for the entire state, but ambiguously concluded, as to average annual temperatures, that recent temperatures had increased statewide “about” 1° since 1895 (but nowhere in the report is the actual figure reported nor use of the word “about” defined). Further, it is not made clear whether this “about 1°” figure was compared to the existing average annual temperature in 1895 or compared to the average annual temperature for all the years since 1895.
The study also avoided the important question of whether, for example, there also was a similar “about” 1° rise in average annual temperature for the 30-year period from 1921–1950, or some other earlier period. The study did not engage in comparing recent periods of time with earlier ones and with at least 30 years of climate records.
This is quite important for two reasons.
First, NOAA has issued guidance that when trying to determine what the climate is like during a period of time (what NOAA refers to as a “climate normal”) at least 30 years of weather observations should be used.¹¹ The USDA has even stated that at least 50–100 years of data of overall average temperatures are needed to assess changes in the climate.¹² NOAA’s use of at least 30 years of data is longstanding and based on the instructions of almost a century ago by the International Meteorological Organization to calculate climate normals by using 30-year periods. NOAA also relies on the general rule in statistics that at least 30 numbers are needed to obtain a reliable estimate of their mean or average.¹¹
Second, comparison of a recent 30-year “climate normal” to a past 30-year climate normal yields information as to whether a recent climate trend not only is significant and enduring, but whether it is unprecedented. While the 2020 study acknowledged that the first half of the 20th century contained decades with “some of the warmest periods in North Carolina’s history,” in computing the state’s average temperature, the 2020 study only compared the single 10-year time-period of 2009–2018 with another single decade, 1930–39. The report found a difference of “about 0.6°” but then moved on without further discussion or “30-year climate normal” to “30-year climate normal” comparisons with other prior temperature phases in the 1800’s or 1900’s. However, as discussed above, NOAA standard practice dictates that when comparing climate normals from periods of time to use 30-year, not 10-year, periods.¹¹ Further, climate records through the 1950’s are deemed significant by the climatic research community as they occurred during a pre-industrial period reflective of natural (that is, non-manmade) climate conditions. As the lead author of the 2020 North Carolina Climate Study wrote, “Because industrial development was minimal until about midway through the 20th Century, the earlier record reflects mainly natural variability.”¹³
For a more detailed discussion of these climate studies as well as a discussion of them and actual historical climate records in the context of climate conditions of most interest to gardeners, see A GARDENER’S LOOK AT THE CLIMATE IN NORTH CAROLINA BACKYARDS: HISTORY OF CLIMATE STUDIES AND PIEDMONT REGION CLIMATE — 2021 UPDATE.¹⁴
The 2020 study’s conclusions of state-wide climatic trends to date, however ambiguous and inconsistent in the time periods used, are undercut by two basic concepts.
One, the study should have heeded the teachings and admonitions of the prior climate studies identified above that “… local climate variability is so high in North Carolina that significant trends are difficult to deduce.”⁹ The 2020 study did not, like the prior studies before it, acknowledge the difficulty in arriving at sweeping assertions of “climate change” across the state. The problem with this approach was articulated by the authors of the 2002 study who wrote that, “North Carolina has one of the most complex climates in the U.S. Factors such as complex topography in the west and warm waters off the east coast combine with weather patterns to produce a region with locally variable weather and climate. Local scale climate analysis can more accurately represent the complex climate that exists in North Carolina, and offers new insights into precipitation and temperature patterns.”⁸
Consequently, this article focuses on climatic records and trends contained in the records of observations at the six Primary, or First Order, weather stations described above and which are geographically disbursed across the state such that all three regions, the Mountains, the Piedmont and the Coastal Area are represented. These stations, in addition to being at the top of the pecking order among weather stations in terms of quality, also have the advantage of containing weather data from as far back as the late 1800's.
Two, the weather stations mostly used in the 2020 climate change study are not as reliable and accurate as the First Order stations. They also have been in use mostly only in recent years and so cannot offer the data necessary to compare to prior NOAA climate normals/periods of time from the late 1800’s and the first half of the 20th century.
The 2020 study does not identify the individual weather stations it used (all weather stations have unique identifying numbers assigned by the National Weather Service). However, the 2020 study acknowledges that it included and primarily relied on the use of so-called “COOP” stations which, as mentioned above, are the least reliable type of weather station.
COOP stations (part of a Cooperative Observer Program), while more numerous, are monitored by volunteers (such as a homeowner with a station placed on his/her property) which can result in temporal gaps in weather data records and other quality control issues. While volunteer COOP observers can play a role in providing weather information for forecasting and basic weather information, periods of missing weather recordings can preclude reliance on them in determining long term trends. None of the three North Carolina climate change studies discloses whether any minimum station quality qualifications and criteria were employed or whether any COOP stations were disqualified from use.
Here is an image of a COOP weather station, (id #311535), located in Cary, North Carolina:
It is located in the front yard of a home in a residential neighborhood in Cary and with temperature observations beginning in just2001. It monitors three weather conditions, temperature, precipitation and snowfall. It is reported as missing 251days of temperature observations.¹⁵
This is the Falls Lake Weather COOP station (id #312993). It is in the backyard of a home in a residential neighborhood.
A closer look reveals that it is a weather apparatus manufactured by the Davis instruments company which can be purchased online.
This station has been in use just since 2000 and is reported to have 505 missing days of temperature data in its average temperature reports.¹⁶
Let’s look at one more COOP station in the state, this one near the coast in southeastern North Carolina, Shallotte, Brunswick County, station id# 317813. Once you make your way past the old cars and the shed, in a small clearing the COOP station can be found.
This station has been in service only since the year 2000 and is missing 267 average temperature readings, with 13 missing days of readings as recently as the past five years.¹⁷
Let’s now take a look at all the COOP weather stations in one county. Here is the list of COOP stations in Wake County, North Carolina as provided by NOAA.¹⁸
Of the 14 COOP type weather stations in Wake County¹⁹, six have not been functional this century and three stopped functioning in 2010, 2012 and 2022. Of the remaining five stations, three stations have start dates since 2000 and one since 1993. So, for example, three stations with only 20 years of weather data, and missing hundreds of days of data, presumably were included in the calculations by prior climate change studies to determine whether there have been climate changes since 1895. As explained above, NOAA’s guidance is that when trying to determine what the climate is like during a period of time (what NOAA refers to as a “climate normal”) at least 30 years of weather observations should be used.¹¹ Lastly, unlike, for example, the RDU First Order station in Wake County, none of these Wake County COOP stations have an uninterrupted report of complete climate records from 1887.
As noted above, the 2020 study did not identify which weather stations it used. However, it explained that it obtained its climatic data from “NOAA’s National Centers for Environmental Information (NCEI) Climate Divisional Dataset (nClimDiv), version 2.” However, it should be noted that this dataset was created due to the recognition that historical climatic records based on the COOP system of weather stations contained many significant errors. NCEI itself acknowledged that various factors reflecting the unreliability of the COOP type station such as changes in location and measurement technology resulted in, “…large errors in calculating long-term climate trends.”²⁰ As a result, NCEI developed a “temperature data homogenization algorithm” in an attempt to reduce, but not eliminate, the errors in climate trend calculations. However, whatever type of algorithm was devised by modern day climate scientists, it still cannot account for the paucity of COOP stations that have been operating in North Carolina for long, historical periods of time. As we saw above, just in Wake County alone, there are no COOP stations operating continuously since 1895, with several only functioning for short periods of time and some only since the year 2000. Also, an algorithm cannot make up for missing data, with many if not most COOP stations having significant missing temperature records.
In a confounding twist, the 2020 study acknowledged that it used a different data set for all of its graphics depicting temperatures and precipitation. Instead of using the dataset identified just above (nClimDiv) from which it obtained its climate information, it instead turned to, “NOAA NCEI’s Global Historical Climatology Network-Daily (GHCN-D), version 3.” The study did not reconcile how it used climate information from one dataset for its conclusions, but then used data from a different data set to illustrate these conclusions in its graphs and charts. As to use of the GHCN-D network, however, NOAA acknowledges that two-thirds of the stations in this network only record precipitation. NOAA further states that anyone using this dataset should also cite a published article which found, according to NOAA, that, “In general, the stations providing daily observations were not managed to meet the desired standards for climate monitoring. Rather, the stations were deployed to meet the demands of agriculture, hydrology, weather forecasting, aviation, etc.”²¹ The 2020 study did not cite this article.
EXAMINATION OF CLIMATE DATA FROM NORTH CAROLINA’S PRIMARY/FIRST ORDER WEATHER STATIONS
A review of the average temperatures, average maximum temperatures and average minimum temperatures for each of these six Primary weather stations was undertaken as far back in time as possible through 2021. The methodology is set out in detail at endnote 22.
This review resulted in 18 total average temperature categories, three for each station. The examination then looked at the results for the last 30 years (1992–2021) to see if the temperatures were higher, lower or the same than the overall average temperature category. If it was higher for the past 30 years, then the records were examined to see if another, prior 30-year period prior to 1950 existed which also reflected a rise in the average, maximum or minimum temperatures. If so, this would show that the rise in average temperatures in the last 30 years had occurred previously. To the extent “climate change” must be defined as a climatic trend that is significant, enduring and unprecedented, then a finding of prior rise in temperature during a previous 30-year period would mean that a recent rise in the past 30 years is not unprecedented.
This review found that in only three out of 18 instances were there rises in average temperatures that also did not have a corresponding rise in the average temperature being measured in a 30-year period starting prior to 1950. A list of the six Primary weather stations and of the 18 average temperature categories can be found at endnote 23. The only three findings of rising temperatures unaccompanied by prior corresponding rises in time were:
- Greensboro (station GSOthr) Average Minimum Temperatures for the last 30 years are higher than the average with no corresponding rise in a 30-year time period before 1950.
- Wilmington (ILMthr) Average Maximum Temperatures for the last 30 years are higher than the average with no corresponding rise in a 30-year time period before 1950.
- Hatteras (HSEthr) Average Maximum Temperatures for the last 30 years are higher than the average with no corresponding rise in a 30-year time period before 1950.
This leaves the 15 other climate categories at the six Primary weather stations which either were not higher than the historical average or, if they were, there was a 30-year period before 1950 which also had temperatures also were higher than the historical average.²³ Examples of these include:
- Charlotte (CLTthr) Average minimum temperatures for last 30 years is not higher than the historical average.
- Charlotte (CLTthr) Average temperature for last 30 years was higher than the historical average but the average temperature during 1921–1950 also was higher than the historical average.
CONCLUSION
The six highest quality, most reliable weather stations in North Carolina, which are located in locations from west to east and representative of the diverse and varied local climatic regions of the state, do not support the conclusion that when looking at average temperatures, as well as average maximum and minimum temperatures over a lengthy period of time, there is any significant, enduring and unprecedented change in climate in these locations. Only the three temperature rises in the weather stations noted above reflected a rise in average temperature not matched by a previous rise in temperature for a similar period prior to 1950, with the other 15 climate categories examine not demonstrating this.
By Tom Packer, North Carolina State Extension Certified Master Gardener, President of the Gardeners of Wake County, Board Member of the North Carolina Cooperative Extension Agricultural Programs Foundation, Moderator of “Gardening Carolina” Facebook Group and longtime backyard farmer in North Carolina and Northern California. Other horticulturally related articles by Mr. Packer can be found at https://medium.com/@tpacker25. All this article’s content is that of Mr. Packer alone as an individual and not as a member of or on behalf of any organizations or groups with whom he is affiliated.
ENDNOTES
[1] See, National Weather Service Instruction 10–1004, May 17, 2020,Operations and Services, Climate Services, NWSPD 10–10, Appendix C. https://www.nws.noaa.gov/directives/sym/pd01010004curr.pdf. The six Primary Local Climatological Data Stations (PLCDs) in North Carolina are Asheville Regional Airport (AVL), Hatteras Billy Mitchell Airport (HSE), Charlotte Douglas International Airport (CLT), Greensboro Regional Airport (GSO), Raleigh-Durham Airport (RDU) and Wilmington New Hanover Airport (ILM). These stations are regarded as “Primary” because of the extensive and manual quality control checks they are subject to, which other weather stations are not. (Per communication with NOAA representative on April 27, 2021).
[2] These high quality stations go through a 3rd, manual quality control procedure. Additional elements are computed and an Annual Edited LCD (an annual summary of weather data) is created. As an example, Wet Bulb (WB) is observed but Dew Point (DP) and Relative Humidity (RH) are computed. Daily average WB, DP and RH are computed from the hourly values that are observed or computed. Most of the secondary (Non-First Order) ASOS stations (~900) may have additional elements computed but do not have the 3rd manual QC of an Annual Edited LCD. [Personal communication from NOAA representatives with author — July 20, 2020 and April 27, 2001].
[3] While the RDU weather station is a First Order station, identification #13722, according to a NOAA representative all major airport weather stations are also assigned a COOP station number, here, RALEIGH AP #317069, even if they are not COOP-type stations. So, there is only one weather station at RDU, it is a First Order station, and there is not a separate COOP station at the site.
[5] See, https://threadex.rcc-acis.org/ re NOAA’s ThreadEx records (click on “About.”) and see https://drive.google.com/file/d/1gMgMQ1X8R7PQa3AFXR-OmyhW3ye55y4V/view?usp=sharing
[6] APPENDIX B — Accessing NOAA Daily Temperature and Precipitation Extremes Based on Combined/Threaded Station Records, NATIONAL WEATHER SERVICE INSTRUCTION 10–1004, MAY 17, 2020, “This new ThreadEx data set provides a consistent basis for the reporting of daily extremes for the longest period of time meaningful. With the ThreadEx effort, maximized, consistent, updated daily extremes will be available for government, partner, and general public (especially media) use.”
[7] Use of Historical Data to Assess Regional Climate Change, Yuchuan Lai and David A. Dzombak, Journal of Climate, American Meteorological Society, July 15, 2019, pages 4299–4320, https://doi.org/10.1175/JCLI-D-18-0630.1
[8] https://drive.google.com/file/d/1MI4vnDo9vrH-YRSbwpIQqjhYnB5PnHeS/view?usp=sharing
[9] https://web.archive.org/web/20150420025024/http://www.nc-climate.ncsu.edu/climate/climate_change and https://drive.google.com/file/d/1Y4YgPJ7L0-YW7b3k1vGGTggEXw8AOCFs/view?usp=sharing
Note: This report was found in a website archive search site. It was captured by the website archive service on April 20, 2015 and prior dates. This report no longer appears on the State Climate Office of North Carolina’s website.
[11] See: https://www.ncdc.noaa.gov/news/defining-climate-normals-new-ways
[12]See: https://drive.google.com/file/d/1MA7gw5GgsVCusiA8JwzZaUqvsFqRN6AV/view?usp=sharing and http://web.archive.org/web/20201114005450/https://planthardiness.ars.usda.gov/PHZMWeb/AboutWhatsNew.aspx
[13] Temporal Variations of Extreme Precipitation Events in the United States: 1895–2000, Kenneth E. Kunkel, David R. Easterling, Kelly Redmond, Kenneth Hubbard, American Geophysical Union (AGU), GEOPHYSICAL RESEARCH LETTERS, VOL. 30, NO. 17, 1900, September 09, 2003. https://doi.org/10.1029/2003GL018052
[15] http://xmacis.rcc-acis.org/ — Seasonal Time Series / Avg temp / Mean / por -2021 / Annual / Average of Months / Daily Station Selection 311535; https://docs.google.com/document/d/1aiKaB_l-NA3lzJvcemYK-UgcXN9xgwtG/edit?usp=sharing&ouid=115138207048984797333&rtpof=true&sd=true
[16] http://xmacis.rcc-acis.org/ — Seasonal Time Series / Avg temp / Mean / por -2021 / Annual / Average of Months / Daily Station Selection 312993; https://docs.google.com/document/d/1X14m8krigTkwsTErnPQcDDMMc2nXMDaE/edit?usp=sharing&ouid=115138207048984797333&rtpof=true&sd=true
[17] http://xmacis.rcc-acis.org/ — Seasonal Time Series / Avg temp / Mean / por -2021 / Annual / Average of Months / Daily Station Selection 317813; https://docs.google.com/document/d/1WF_-XW4ZEI2kzY22cad-oxk9D7C-ysz6/edit?usp=sharing&ouid=115138207048984797333&rtpof=true&sd=true
[18] Communication from NOAA representative with author on January 21, 2021.
[19] The station identified as 317069 NSW COOP Raleigh AP is not included in this count of COOP stations as this is, per endnote 3, First Order station KRDU #13722.
[20] https://www.ncei.noaa.gov/access/monitoring/dyk/nclimdiv-tmax-tmin
[21] https://www.ncei.noaa.gov/access/metadata/landing-page/bin/iso?id=gov.noaa.ncdc:C00861;view=iso; https://www.ncei.noaa.gov/metadata/geoportal/rest/metadata/item/gov.noaa.ncdc:C00861/html
See, “Publications citing this dataset should also cite the following article: Matthew J. Menne, Imke Durre, Russell S. Vose, Byron E. Gleason, and Tamara G. Houston, 2012: An Overview of the Global Historical Climatology Network-Daily Database. J. Atmos. Oceanic Technol., 29, 897–910. doi:10.1175/JTECH-D-11–00103.1.”
[22] The data for the average temperature, average maximum temperature and average minimum temperature for all of the six Primary stations was obtained by using there three searches and inserting the name of the station:
- http://xmacis.rcc-acis.org/ : single station/seasonal time series/ avg/temp/mean/ por — 2021 / Station:_________
- http://xmacis.rcc-acis.org/ : single station/seasonal time series/ min temp/mean/ por — 2021 / Station: ________
- http://xmacis.rcc-acis.org/ : single station/seasonal time series/ max temp/mean/ por — 2021 / Station: ________
The Applied Climate Information System (ACIS) was developed and is maintained by the NOAA Regional Climate Centers (RCCs). It was designed to manage the complex flow of information from climate data collectors to the end users of climate data information. See, http://www.rcc-acis.org/
[23] See, https://drive.google.com/file/d/1-RdfKjFn1rWrcUrekM6ztNauadPaZrnJ/view?usp=sharing
