ML-for-B-E / nevergrad

A Python toolbox for performing gradient-free optimization
https://facebookresearch.github.io/nevergrad/
MIT License
2 stars 0 forks source link

Run a first experiment of irrigation planning for a sugarbeet crop in Porto-Novo #33

Closed akouete-kpakpo closed 2 years ago

akouete-kpakpo commented 2 years ago

I have this plot for the variety 601 of sugarbeet crop

image

Unsatisfying as:

image

image

Have to understand why

Ethel2003 commented 2 years ago

When i did my first simulation of sugarbeet based on the notebook of Allard I have the same level of leaf area index as Allard. I could say the fact that you don't have the same thing as Allard is due to the parameters of crop.

akouete-kpakpo commented 2 years ago

thanks for your suggestion @Ethel2003 . Indeed, my previous plot shows a sugarbeet simulation in Lome. But in Allard's tuto, the crop locates in Wageningen, Netherlands. This is what I have by fixing the location: now I have the same range of values for the leaf area index as in the tuto notebook

image

akouete-kpakpo commented 2 years ago

The difference is explained by the rain intensity during the crop (source of data plotted: NASA Power database)

Lome, Togo

image

Wageningen, Netherlands

image

akouete-kpakpo commented 2 years ago

I think that

root zone soil moisture

image

is more explicit than

Leaf area index

image

to correlate crop ecosystem evolution with irrigation events

Ethel2003 commented 2 years ago

Okay. It is more comprehensible

akouete-kpakpo commented 2 years ago

solved in #36

it was a config bug. Now all good

image