Using the CUJET3=DGLV+VISHNU jet-medium interaction framework, we show that dijet azimuthal acoplanarity in high energy A + A collisions is sensitive to possible non-perturbative enhancement of the jet transport coefficient, q(T, E), in the QCD crossover temperature T ∼ 150 − 300 MeV range. With jet-medium couplings constrained by global RHIC& LHC χ 2 fits to nuclear modification data on R AA (p T > 20) GeV, we compare predictions of the medium induced dijet transverse momentum squared, Q 2 s ∼ qL ∼ ∆φ 2 E 2 , in two models of the temperature, T , and jet energy E dependence of the jet medium transport coefficient, q(T, E). In one model, wQGP, the chromo degrees of freedom (dof) are approximated by a perturbative dielectric gas of quark and gluons dof. In the second model, sQGMP, we consider a nonperturbative partially confined semi-Quark-Gluon-Monopole-Plasma with emergent color magnetic dof constrained by lattice QCD data. Unlike the slow variation of the scaled jet transport coefficient, qwQGP /T 3 , the sQGMP model qsQGMP /T 3 features a sharp maximum in the QCD confinement crossover T range. We show that the dijet path averaged medium induced azimuthal acoplanarity, ∆φ 2 , in sQGMP is robustly ∼ 2 times larger than in perturbative wQGP. even though the radiative energy loss in both models is very similar as needed to fit the same R AA data. Future A+A dijet acoplanarity measurements constrained together with single jet R AA and v n measurements therefore appears to be a promising strategy to search for possible signatures of critical opalescence like phenomena in the QCD confinement temperature range.