I took interpolation techniques on one of my interesting research sites using R-programming to predict and estimate long year-based mean monthly and annual rainfall and temperature using KED, IDW, and OK methods. As a result, I found the predicted values, which were equal/same with the actual value for most of the months and annual, specifically for KED using DEM_90m as a covariate, and OK. The problem was more worsen in the case of OK, and I attached herewith the findings. So, I need your valuable recommendations on this one.
Thank you for your time!!
NB: I took the same procedure on the other catchment, and found good results using the same interpolation techniques.
Quite possible , eventually that speaks off the precision of interpolation ...afterall thats the purpose of any interpolation to predict as close to original values ...
Anoop Kumar Srivastava
Dear Dr., Thank you so much for your brief explanation! here is the question; I'm not sure whether you saw or not the attached table, the value of correlation (R) was negative, and how does Pearson correlation (R) become negative?
Thank you again!
Only a statistical test can show you if you experimental and simulated values are significantly different or not.
The answer to the question depends on what do you mean by equal/same?
How many decimal places?
After-all you are discussing averages of observations! Is there any heteroskedasticity?
Why are you bothering to interpolate?
Dear Hirpo sorry for late response. Correlation value is negative, do you mean , at higher observed value , predictive value is lower or vice versa, please have a check at dat , whether correlation oefficient is really negative . No of observations through enclosed table is too low for precision of such statistical tools , where closeness to observed values is involved...
thank you all for your generous participation and feedback. I solved it!
- Badges
- Science topic