Wednesday, November 01, 2006

The Analysis

Ok I know, this article will be really boring, but potentially it is one of the most important. Basically all I’m writing here is how I got the data, and how I did my analysis. It’s important because I want to tell people about the unbiasedness of my work. It is also important so that should someone want to replicate what I did to prove that I was honest/dishonest, then they can do so and I encourage people to do so. This is the best part of science. So if you are interested in how the analysis was done, then read on, otherwise the next post will have the first stage of some of the results.

Data is taken from the Australian Bureau of Meteorology (ABM). The ABM divide Australia into a number of rainfall districts as shown here

For our analysis we took one weather station from each district. The chosen weather station was one that had weather predictions up to the current date and also had results that went back the furthest in time. This resulted in 102 weather stations in total. On occasions, where a certain long lasting weather stations stopped and another one took its place on the previous stations ending, both stations were taken in the sample. The analysis however in this instance still treats these two stations as two independent weather stations, and analysis is done separately on both.

Following this, weather stations on islands and the 2 stations on Antarctica were removed from the sample as it was estimated that they might not represent Australia as accurately as possible. Due to information regarding urban heating, stations that had approx. 100,000 people or more were removed from the sample as well, and we were left with 82 weather stations around Australia to use.

From each these stations the average monthly temperature (or rainfall etc.) was calculated from the time the measurements started until current. Hence we have different average temperatures per month per station (eg. The average temperature in KALUMBURU in January, the average temperature in KALUMBURU in February etc.).

From these averages we calculated the deviations from the mean for every month of every year for each station. These were then summed to get the average deviation from the mean for every station for every year. The average deviation for all stations was then recorded and this is shown in the graphs and analysis. Variables at this stage included in the analysis are Average Maximum and Minimum Temperature, Aver Temperature at 9am and 3pm, and Average Rainfall per month.

No comments: