Or basically, if that is too long - how can I statistically compare data recorded from 5 different locations over 5 different times periods? How do I approach this problem?
Hi there, I've recently been asked to come up with an idea on how I could make this statistical comparison and overcome any issues/challenges over the reliability and validity of the results. I'm interning for a development organization and a phd Climatologist wants to write an original paper and has given me this to look over and help his work load. Ironically, I'm a bit of statistical amateur, at least comparatively, and I took this on because it looks interesting, but he'll be requiring a seriously advanced or at least highly suitable explanation. Nonetheless:
The premise: Between the years of 1900 and 2011 in Egypt, investigate if there has been any significant change from the dates in which seasonal environmental events have traditionally occurred. Thus I would be taking maybe 10 fixed variables, all dates in which 10 significant environmental events have be known to occur throughout the year. These "seasonal environmental events" are a sort of imparted farmer wisdom that stretches very far back. Many of these farmers are now complaining that these events are all completely unpredictable now, an example would be of a famous great wind that would always come at the beginning of January, literally, they say, as far back as their ancestors moved to that particular region. Another example is that of great storms out at sea in which fisherman, as long as they can remember, have always kept to strict proceedings throughout the year due to the cyclical timing of environmental events.
What we want to investigate is, firstly, to compile all relative meteorological data over the 110 year span, maybe in 5 key categories (temperature, air pressure etc) and assess the findings. But, the plot thickens...
The issue/s: From 1900 to 2011 the national observatory has moved 6 times, and pretty large distances. How can I overcome this using to statistics? For instance, we assemble all the data, how can we statistically validate if there has been significant change and its reliable? Obviously, the in such differing locations will all have very different averages in temperature and air pressure etc...so how can one overcome this, and any other problems that I have failed to foresee, in order to gain accurate findings?
Any help or advice would be really great as I'm quite unfamiliar with immediately recognizing statistical problems. Cheers.