I studied the response of wheat to different nitrogen rates and I sent the paper to a journal. The rates of nitrogen were 120, 180, and 240 kg N /ha. I recorded the grain yield. The reviewers asked me to:
1- use quadratic plateau modeling approach to check which dose will be best
2- what will happen after 240.
The data are in the attached file.
Instead of quadratic modelling, we can use best for models..
Take your data it would nice to have a zero fertility value.
Apply linear, power functions and have them compared based on the R square value.
The highest value of R square shows the best fit.
The data you present does not show much plateau effect just a diminishing return on input. To find a plateau effect try doubling your high rate. Your curve will be better by using a zero and intermediate rate.
You won't be successful with fitting a quadratic plateau model. One problem is that your data don't show a plateau. But also, with only three different x values, you won't be able to fit a nonlinear model with three parameters.
In your data, yield increases with N rate. It's a rather linear relationship. There's no value in adding even a quadratic term. You cannot say what will happen after 240, because there is no indication that the yield plateaus.
Thank you very much Dr Salvatore S. Mangiafico
Relay it is the best answer
- Badges
- Science topic
- Similar topics
- Medicine
- Physiology
- Biosignals
- Biomedical Signal Processing
- Statistical Analysis