Keep in mind, that heritability is not a property of a plant or of a plant character, but a statistical parameter of the studied data collection, see my answer to Francisca Addae-Frimpomaah (How can I calculate heritability..). If heritability is calculated within one single trial, the outcome depends on the tested genotypes and the residual error. If the genotypes differ very much in performance and the residual error is very low, the heritability can be close to 1. If the tested genotypes have similar performance, the heritability might be close to 0. Heritability, if calculated across your two environments, may be quite low, if the ranking of genotypes is different or close to zero, if the ranking is not reproducible. It can be close to one, if the ranking is the same. A crude, but robust simple way to calculate heritability across your two environments consists in calculating the correlation between the genotype means of the two environments, which is defined as Cov(xy)/(sx * sy), since Cov(xy) represents genetic variance and sx * sy the geometric mean of the phenotypic variance in both environments.