In study of muon neutrino disappearance at 810 km, the NOvA experiment finds flavor mixing of the atmospheric sector to deviate from maximal ($sin^2theta_{23} = 0.5$) by 2.6 $sigma$. The result is in tension with the 295-km baseline measurements of T2K which are consistent with maximal mixing. We propose that $theta_{23}$ is in fact maximal, and that the disagreement is harbinger of environmentally-induced decoherence. The departure from maximal mixing can be accounted for by an energy-independent decoherence of strength $Gamma = (2.3 pm 1.1) times 10^{-23}$ GeV.