Estimation of genotype-by-environment interaction (GEI) is important in breeding programs because it provides critical information to guide selection decisions. In general, multienvironment trials exhibit heterogeneity of variances and covariances at several levels. Thus, the objectives of this study were (a) to find the best genetic covariance matrix to model GEI and compare changes in genotypic rankings between the best covariance structure against a compound symmetry structure, (b) to define mega-environments for turfgrass performance across the southeastern United States, and (c) to estimate genetic correlations between drought or nondrought and growing or nongrowing conditions to determine the extent of GEI under specific environments. Three nurseries with 165, 164, and 154 genotypes were evaluated in 2011–2012, 2012–2013, and 2013–2014, respectively. These nurseries were conducted at eight locations (Citra, FL; Hague, FL; College Station, TX; Dallas, TX; Griffin, GA; Tifton, GA; Stillwater, OK; and Jackson Springs, NC). The response variables were averaged turfgrass quality (TQ), TQ under drought (TQD), nondrought TQ (TQND), TQ under actively growing months (TQG), and TQ under nongrowing months (TQNG). This study demonstrated that (a) the best variance structure varied among traits and seasons, and changes in genotype rankings were dependent on GEI; (b) considering TQ and TQND, mega-environments formed between Jackson Springs and College Station, and between Citra, Dallas, and Griffin, whereas Stillwater, Hague, and Tifton represented unique environments across the southeastern United States; and (c) genetic correlations between drought or nondrought and growing or nongrowing conditions suggested that indirect selection can be efficient in multienvironment trials for contrasting environmental conditions.