at least three environments or locations

I think three and more

For GGE biplot and AMMI analysis, the same set of genotypes tested at least at six environments ( at three locations for two tears) is essential.

I did with three and more

At least 3 environments for statistical analysis point of view but it is always good to have more locations and at least two years data per location. The minimum number of locations should form a good representation of the full set of the target environments or recommendation domain for the program.

I think for season, at least three years and eco, two environments, for country like Iraq. For country with mega environment like USA, Canada, Russian etc.. Six environments with two years.

Cheers

To determine the required numbers of genotypes and environments depends upon how those variables are different from each other. Widely diverged genotypes and/or environments need more numbers as compared to those of some similarities. Genotypes and environments interactions need enough years to determine significant or insignificant interactions. When years are calculated in temp.,humidity, sunny days, dusty days, clouds,......the study will need at least five to six years.

Of Prof Dr Medhat Elsahookie, the proper answer.