"Climate normals," or "climate means," are based on 30-year averages of meteorological data. Recently, daily and monthly 30-year data for literally hundreds of sites in the US, for 1971-2000 and 1981-2010, have been put online where they are easy to access here. The data originally come from the National Climatic Data Center. An online application for extracting data from these sites, by state, is available HERE. There are some interesting questions that can be asked: "Are there any discrepancies or errors in the data? Are there statistically significant differences between urban and rural areas? How does climate vary around my state? Are there significant changes in my state between the 1971-2000 and 1981-2010 time periods?"
The image below shows the first page of monthly data for Pennsylvania. This image shows that not all data are available for all stations. It is not at all clear why the first station in this list contains data only for precipitation, as it might be reasonable to assume that air temperature would be the mostly widely available measurement.
When considered as a science fair project, these data provide an interesting example of how to form
and use a hypothesis.
Suppose you choose the research question "Are there significant changes (in what parameter?) in my state between
the 1971-2000 and 1981-2010 data sets?"
There is very little justification for trying to guess, ahead of time, a "yes" or "no" answer to this or the other questions
that have a "yes" or "no" answer! If you choose
one answer or the other as a hypothesis, the purpose of that hypothesis is not to prove your "guess" right or wrong,
but only to provide a question
that can be tested and answered quantitatively. In the end, there will either be a
statistically significant difference or there won't. There is no "right" or "wrong" outcome and either is equally valid
and interesting.
This Excel bubble graph shows, for the Pennsylvania 1981-2010 data set,
[(30-year site temperature mean) - 30-year statewide temperature mean)]. The diameter of the bubbles is proportional to this
value. Filled bubbles are above the statewide average and open bubbles are below.
As expected, temperatures are above the statewide average in the southern part of the state,
particularly around the more densely populated
urban areas of Philadelphia in the southeast and Pittsburgh in the southwest.
Can these differences be attributed to urban heat islands? If so, or if not, how can you tell? Are the maximum
or minimum temperature patterns different from the means? Are the patterns different for different seasons? Will there be
differences between this plot and the same plot
for the 1971-2000 data set? These data can be used to examine all these and many other interesting climate questions for your state
or region!