Algorithmic Analysis of the Climate of Pittsburgh, PA


Friday, December 04, 1998

Background

A temperate zone, such as that in which Pittsburgh is located, has four seasons: Spring, Summer, Fall, and Winter. While Spring and Fall generally have cool temperatures, Summer is warm and Winter is cold.

In order to encode four distinct values in binary, we need two "bits"[1]. Two bits allow for four values because we can choose either 1 or 0 for the first bit, then for each of those two choices, we have two choices for the second bit. Therefore the possible values with two bits are 00, 01, 10, and 11. We can assign one season per value, so Spring might be 00, Summer might be 01, Fall might be 10, and Winter could be 11. In this way two bits can differentiate between four possible states, or in this case seasons.

[1] A bit is a binary unit of information that can have only two states, in this case referred to as either 1 or 0.

Pittsburgh

I propose that Whomever created the weather for Pittsburgh used a one bit climate system. Unlike a two bit climate, with one bit you can only have 0 or 1, which only allows for two possible states (or seasons) for each day of the year.

Now in the computer industry, before the creation of color or grayscale computers, everything was black and white, or one bit. A zero would encode black, and a one white (or the other way around, the point is the same.) In order to make it appear that gray was being displayed on the screen, a technique known as "dithering" was used. With "dithering", black and white pixels are put next to each other in certain patterns in order to average out to gray, such as in the following image (Borrowed from Robert Berger's explanation of monitor gamma):

I would like to put forward the proposition that, unlike most temperate regions which use a two bit climate system, that Pittsburgh instead uses a single bit to encode its climate. You see, Pittsburgh assigns Summer to one value, let's say 0. Winter gets the other, in this case 1. In order to approximate Spring and Fall, a 50% dithering pattern is used. Thus a week during Spring might look like either 0101010 or 1010101. So when the entire season is taken as a whole, Spring and Fall display average temperature for a temperate region. However, any single day can only be one of two possible states-- Summer or Winter.

Conclusion

I am proposing that this is the reason that Pittsburgh temperature seems so variable, and Fall and Spring so erratic. During these seasons, any given day in Pittsburgh can seem like Summer never left, or that Winter descended early, sometimes switching within hours. Using my proposed system, however, it can be seen that Pittsburgh climate can follow the proper temperate patterns overall, but at the same time seem completely and totally erratic on a day to day basis.