CDO

Midhun Thomas wrote:

Hi all,
I just started using Cdo and I am fascinated by its versatality. I am looking currently at correlation of wind with total precipitation over a grid region for 30 years (daily values). For this I am using detrending both these variables using detrend and then going for timcor. I hope, the overall approach is not wrong. However, I have three issues
1. Should i detrends the variables using cdo detrend every year separately . A important doubt

2. Detrend of precipitation over a grid region can also lead to negative values . Is that generally allowed
3. I am wanting to see the general strength of relationship over a long time.
There seems to be no uniform way in the detrending method (linear etc.) . Some are detrending alone, some subtracting anomalies as well and some smooth and detrend

I wish advanced regression, correlation and detrending technique other than linear are in cdo as well

Hello,

It is not necessary to detrend the variables separately for each year. You can detrend the entire time series together to capture the long-term trend. Detrending precipitation can indeed lead to negative values, and it is generally allowed. Negative values indicate a decrease in precipitation compared to the long-term average. To assess the overall strength of the relationship over a long time, you can calculate the correlation coefficient using the detrended variables.

This will give you an idea of the magnitude and direction of the relationship. While Cdo offers basic detrending methods like linear, it does not currently include advanced regression and correlation techniques. You may need to explore other statistical software or programming languages for more advanced analysis options.

Thank you!

My another doubt is , regarding detrending. I find many especially those who are just interested in getting the statistical correlation to just for detrened (linear detrendin) irrespective of the actual trend. Some might for smoothening some might go for removing subtracting running means. Is there any typical rule that decides this. Or is this practice not uncommon?