"How Unusual Is That?"  Hypothesis Testing with Online Data

 


We occasionally hear "everyone talks about the weather but no one does anything about it."  We, of course, can't change weather at our whim but may be doing so unintentionally.  In several recent years Seattle had drought conditions, one followed by the heaviest day of rain on record.  Previous years have featured early cold snaps while others have mild temperatures through late Autumn.  Forecasts have become less reliable even though meteorologists have more data than ever before.  Is this variability and unpredictability unusual?  Are the warnings of global warming and the resultant climatic change actually coming true?  Inquiring minds want to know!

Luckily, we don't have to be meteorologists to explore our musings about weather phenomena.  We are familiar with the basic data because of its infusion in school curricula and ubiquity in the media.  The Internet features scores of weather sites, The Seattle Times devotes an entire page each day to weather, news reports always include weather, and The Weather Channel is a popular feature of cable TV.  Temperature, dew point, precipitation amount, wind speed, and air pressure are given in known units and are fairly easy to both understand and sense.  Current or forecast conditions often affect our activities or plans, therefore we have a natural interest. 

We begin our inquiry by identifying what we want to know and refining the question we want to ask.  This is the hardest part for our students because they often have little experience formulating questions of their own asking and crafting them into ones that can be investigated.  Understanding the kind of relationship we are investigating helps us identify our method and needed data.  If we started with "I wonder how unusual the temperatures have been lately?" we could restate the question as "How do the temperatures over the first two weeks of November compare to normal temperatures?"  Point your Web browser to Live from Earth and Mars (http://www-k12.atmos.washington.edu/k12/grayskies/nw_weather.html) and you will find a tool to plot graphs of your own design from data of your own selection.  For our purposes, choose "Temperature/Dewpoint" for "SeaTac Airport" for the timespan "Last Two Weeks" using "Lines."    Click "Go plot the selected data!" and you should see the following:

Figure 1


Graph plotting utility


Figure 2

Sample Plot
<>
The resulting plot can then be compared to the averages over the last several years (known as "period of record") looking at a different part of the site (http://www-k12.atmos.washington.edu/k12/grayskies/Our_climate.cgi?waseat) yields a plot of averages and extremes.  If it's highs that interest you, compare to the yellow line.  If it's the lows, look compare to the green line.  Our plot of recent temperatures does not, of course, challenge either extreme--but what if it did?  Does THAT mean our weather has been unusual?  What if we were close--perhaps having a series of very warm (or very cool) days is whatÕs unusual and for that we need the day-by-day temperatures to compare!  If you click on the current data plot (Figure 1) you will see the data from which the graph was constructed (and a whole lot more!).  You would then need to extract the high temperatures (if that is what your question is about) and compare to the daily temperatures for the period of record--not only difficult statistically but also a lot of data.  If you want to perform a scaled-down comparison try the average highs from the Western Regional Climate Center (http://www.wrcc.dri.edu/cgi-bin/cliMAIN.pl?waseat).  Click "Daily Tabular data" and use the data in the "TMAX" column.  You can even save the data in a file (click in the window, pull down under "File" to "Save as...", rename as "data" and save (in Text format) to a location where you can find it.  Open your spreadsheet program (preferably Excel) and "Open" the file you just saved.  The Text Import Wizard should launch and allow you to select "Delimited" then "Next" and check "space."  Click "Finish" to complete the import of the data in spreadsheet columns.  Now you have the data where you want it!
Figure 3



Comparisons can also be done within the same period and station that allow you to explore other connections.  Perhaps you are wondering how the temperature relates to solar radiation.  Maybe you wonder if air pressure and wind speed have any relationship.  Produce a set of plots and see if any connection exists.  If so, explore further with other dates and locations to see if your hypothesis is consistent.

In summary, consider your excursion into weather and climate data a search for patterns.  Whether you perform complex mathematical modeling with the data you find or merely compare pre-formatted plots the potential for revealing interesting (and perhaps to date undiscovered) relationships is huge.  There's a diamond in the rough waiting for you--all you have to do is dig!

Here are a few questions to consider that you can explore using the given tools.  What is the relationship between:

 

I hope whatever students (or you!) consider leads to more questions.

 

Richard T. Edgerton, Ph.D.

Seattle Public Schools

rtedgerton@seattleschools.org