I have been tasked with analyzing temperature time series from multiple weather stations. I have minute and hour resolution temperature data spanning three years at 4 stations and 35 stations respectively. The stations reside on ~1000 km^2 (southern tip of Vancouver Island) so they cover a relatively concentrated area.
I am looking for additional "statistical" techniques I can apply to the data. You might ask: what exactly do you want to glean from your data? Well, I am a complete amateur in weather and time series analysis. Pretty much anything that gives useful interpretation of data.
So far I have performed the following:
- Power spectral density on minute data
- Probability distribution plots on minute data
- Auto lag correlation and cross-correlation between stations on minute data
- Principal component analysis on hourly data
Techniques I have come across:
- Canonical correlation analysis
- Cluster analysis
- ?
Any suggestions for books/papers/websites explaining analysis and practical applications are also much appreciated. I have found "Statistical Methods in the Atmospheric Sciences" (2006) by Wilks to be quite informative.
Thanks!
