Köppen Climate Classification Changes: 1981-2010 to 1991-2020

 "Indeed it has been said that democracy is the worst form of Government except for all those other forms that have been tried from time to time" - attributed to Churchill, but likely originated elsewhere

The venerable Köppen Climate Classification system is much maligned and derided, but like the famous quote above, it is the best system out there that is applicable across the Earth. If you have a better system for the entire Earth, I am happy to hear about it.

Climate Normals:

Every 10 years, the U.S. National Center for Environmental Information (NCEI) updates the 30-year climate normals for thousands of stations across the U.S. On May 4, 2020, the new 1991-2020 climate normals were released. These replace the 1981-2010 normals previously in effect. Note that there are 20 overlapping years in these two periods. They essentially dropped the 1980s and replaced them with the 2010s.

Köppen Classifications:

There are many ways to describe the climate of a place. The descriptions can range from quite technical to quite informal. The venerable Köppen Climate Classification System (Köppen 1884) is easily the most famous and is described in every climate textbook written in the last 100 years. The system uses monthly and annual temperature and precipitation to classify all portions of the earth into one of 5 major categories and 30 minor categories. There are many critics of the system and many alternative classification systems have been developed but nothing has come close to the widespread acceptance of the Köppen Climate Classification System. 

The five main categories of the Köppen Climate Classification System are as follows:

A – Topical climate: All months have a temperature greater than 64.4°F.

B – Dry, arid, or semiarid climate: Potential evapotranspiration exceeds precipitation. Criteria on a sliding scale based on average annual temperature.

C – Mesothermal (a.k.a., Mid-latitude) climate: At least 1 month above 50°F and at least 1 month below 64.4°F.

D – Microthermal (a.k.a., Continental) climate: At least 1 month above 50°F, at least 1 month below 64.4°F, and at least 1 month below 32°F.

E – Tundra climate: All months below 50°F. Note: climate classification tundra and ecological tundra are different.

IMPORTANT: Climate types A, C, D, and E, are first and foremost defined by temperature. Type B is based on precipitation, or lack thereof. If a station meets the type B dryness criteria, it supersedes whatever the A, C, D, or E category was set to be.

There are many subcategories based on a variety of temperature and precipitation factors. The Encyclopedia Britannica entry for the Köppen Climate Classification System has an excellent description of the major and sub categories. (Note: they use 26.6°F as the cutoff between C and D climate types whereas the traditional cutoff is 32°F).

Maps:

Figs 1 & 2 below show Köppen Climate Classifications for 1991-2020 (stations and gridded) and Figs 3 & 4 show Köppen Climate Classifications for 1981-2010. Fig. 5 shows which stations changed during the two periods.

Fig. 1: Köppen Climate Classification of all stations with 10+ years of temperature and precipitation data for the 1991-2020 period.

Fig. 2: Köppen Climate Classification of 2.5 arc-second grids of temperature and precipitation data for the 1991-2020 period.

Fig. 3: Köppen Climate Classification of all stations with 10+ years of temperature and precipitation data for the 1981-2010 period.

Fig. 4: Köppen Climate Classification of 2.5 arc-second grids of temperature and precipitation data for the 1981-2010 period.

Fig. 5: Köppen Climate Classification changes between the 1981-2010 period and the 1991-2020 period. Note that there are slightly fewer stations since it was required that a station had published normals during both periods. 

Notable Changes:

Of the 5,543 stations with normal data published for both periods, 695 showed a change between the two periods. Exactly 200 of those changes were "minor," and typically meant that rainfall patterns changed from one part of the year to the other. That leaves 495 substantive changes. Fig 6 shows a partial list of those stations with notable changes.

Fig. 6: Selected list of stations with notable changes in the Köppen Climate Classification categories between the 1981-2010 and 1991-2020 periods.

Google Earth Files:

• 1981-2010 (Stations)
• 1991-2020 (Stations)
• Changes between 1981-2010 and 1991-2020

References:

Köppen, Wladimir (1884). Translated by Volken, E.; Brönnimann, S. "Die Wärmezonen der Erde, nach der Dauer der heissen, gemässigten und kalten Zeit und nach der Wirkung der Wärme auf die organische Welt betrachtet" [The thermal zones of the earth according to the duration of hot, moderate and cold periods and to the impact of heat on the organic world)]. Meteorologische Zeitschrift (published 2011). 20 (3): 351–360. 

Precipitation Concentration Index

Fig 1. Precipitation Concentration Index (PCI) of the U.S. (using 1991-2020 normals) and Canada (using 1991-2020 monthly averages).

The PCI

How do we quantify the seasonality of precipitation? The short answer is that there is no one satisfactory method. Nevertheless, it is worth trying to get a handle on how precipitation varies throughout the year. A method that I prefer is the use the Precipitation Concentration Index (PCI) (Oliver 1980). The PCI looks at monthly average precipitation over a specified time period (although you can computed it for an individual year) and assesses how uniform the precipitation is. If a station had the exact same average precipitation every month, we would say it is perfectly uniform. If the received all their annual precipitation in a single month, we would say is it the opposite of uniform. 

The PCI was developed to characterize areas with pronounced wet/dry seasons – particularly equatorial monsoon regions. Still, we can apply the calculation to any location with monthly precipitation. The calculation is as follows, take a month's precipitation and square it. Do this for each of the 12 months and add up the total. This is the numerator of the equation. Then, take annual precipitation and square that. This is the denominator of the equation. Now divide the numerator into the denominator and multiply by 100. This gives you the PCI. Fig. 2 shows the formula.

Fig 2. PCI formula.

If a station averaged exactly 5.0" every month of the year, we compute the PCI by squaring 5.0 (5.0^2 = 25.0) and doing this 12 times. When you add all those up you get 300. When you square the 60.0" annual precipitation, you get 3,600. To get the PCI, you solve: 300/3600*100 = 8.333. This is the lowest possible PCI score and indicates a perfectly uniform precipitation distribution.

According to Oliver, a PCI value < 10.0 represents mostly uniform precipitation distribution. A value between 11 and 15 represents moderate precipitation concentration. A value above between 16 and 20 denotes irregular distribution. Values above 20 represent strong irregularity. I sliced and diced these ranges a little to make the map in Fig. 1. Specifically, I used the following categories from most to least uniform: 8.33 to 8.5, 8.5 to 9.0, 9 to 10, 10 to 11, 11 to 15, 15 to 16.7, and 16.7 to 21.8.

As the map shows, the lowest values are from east Texas through Maine. The interpretation of this is that there is not much difference between average monthly precipitation totals. Of course a lot can happen in any given year, but when we aggregate the numbers, there's a lot of similarity between the monthly values. Conversely, areas in southern California have a pronounced wet  season and a pronounced dry season. Several summer months have almost no precipitation at all. Fig. 3 shows the monthly precipitation distribution for the U.S. stations with the lowest and highest PCI values. The PCI computation does not care about the total annual precipitation. It only cares about the distribution across the 12 months compared to it's own annual total.

Fig 3. Highest and lowest PCI values in the U.S. Brewerton Lock 23, NY, has a PCI of 8.37 and Morongo Valley North, CA, has a PCI of 21.76.

Looking at "major" stations only (see Fig. 4), we see that the lowest PCI values are all in the eastern Lower 48 with a primary concentration in the Northeast. For large cities, Charlotte, NC and Boston, MA, have the most uniform precipitation distribution. On the other side of the ledger, nearly all the stations with the highest PCI values are in southern California. A few Alaska stations are also on this list. For large cities, Los Angeles has the least uniform distribution .

Fig 4. Lowest and highest 25 PCI values in the U.S. for major stations. I took some liberty in deciding what was a major station.
References: 

Oliver, J.E. 1980. Monthly Precipitation Distribution: A Comparative Index. Professional Geographer. 32, 300-9.