Thursday, April 19, 2007
Enough of Western Australia, and onto south Australia. There are only 4 stations of worth in this area, and we will do each one individually as they are spaced out between each other a long way. Ceduna's temperature record isn't great but is good enough. Maximum and minimum temperatures go back to the end of the second world war, whilst time based temperatures start in the mid 50s, with the exception of Midnight, which contains only 16 years of data.
Analysis of maximum and minimum temperatures both show a significant increase (Max: t=3.3,p<0.01; Min: t=3.5,p<0.01).
Note that with maximum temperatures, before 1960 we had lower than normal temperatures and since then most years have been in the positive. The last 9 years have all been above par. Minimum temperatures show a similar pattern. But lets have a look at the time based temperatures.
All of Midnight, 3am, 6am, 9am, Noon, 3pm, 6pm and 9pm show no significant increase or decrease in temperature (Mid: t=1.8,p=0.08; 3am: t=0.5,p=0.6; 6am: t=-0.4,p=0.7; 9am: t=0.2,p=0.8; Noon: t=1.4,p=0.16; 3pm: t=2,p=0.055; 6pm: t=0.8,p=0.4; 9pm: t=1.4,p=0.16).
So how can it be that we have found an increase in maximum and minimums but no increase in any time? well lets have a look at maximum first.
It must be noted that there was almost a significant increase in temperatures at 3pm, but just fell above the 5% mark. Increases - although non significant - and not as strong - also occurred at Noon and 6pm, so we can probably say that Ceduna has been heating up during the day especially at the heat of the day around 3pm. There is also the case that maximum temperature data has been recorded an extra 10-12 years before time based data. When looking at the maximum temperatures in the early days, they were all very much negative. When looking at maximum data as well as temperatures from 3pm we find that they are highly correlated (r squared = 90%), so if early temperatures were accurate, it is a reasonable assumption that temperatures at 3pm in the 40s and early 50s were also below average.
Minimum temperatures also show lower then average temperatures from the mid 50's and beforehand - a time where the temperatures at certain times of the day were not recorded. Hence the most logical reason for the increase in minimum temperatures, but no increase overnight. It must be noted however, that since 1955 no significant minimum temperature increase was found (t=1.6,p=0.2)