Hi! I'm comparing two 50-year SPEI (Standardized Precipitation Evapotranspiration Index) time series created from model and observed data (1950-1999) and averaged over a large area. Both time series are annual (i.e., 50 values each). My goal is to assess very simply the reliability of the model-based SPEI before analyzing future projections for mid- and late-21st century.
What is a good aggregation period on which to perform this correlation?
I tried the annual data (i.e., 50 values each), as well as 5- and 10-year averages (i.e., 10 and 5 values) but got higher correlation coefficients the longer the aggregation period, ranging from -0.02 (annual) to 0.2 (5-year) to 0.95 (10-year), which tells me I can get any correlation coefficient depending on the aggregating period.
I would appreciate any advice from more experienced users on how to choose a "good" aggregation period. As a note, my future projections represent 25-year averages (i.e., 2041-2065 and 2075-2099).
Thank you very much!
Toni
Hi
All that was a great time period was a closer study and the results are impressive
Toni Klemm
Will you say a bit more about the size of the large areas used for the average?
Thanks
As there is no specific limit for aggregation period. I think the one giving highest correlation can be chosen.
- Badges
- Science topic
- Similar topics
- Mathematical Sciences
- Statistics
- Probability
- Probability Theory
- Reliability