High Resolution WRF (Weather Research and Forecasting)/microscale code simulations are carried to predict and characterize stratospheric Optical Turbulence (OT) layers induced by jet streams and gravity waves under various local atmospheric conditions. This information in turn is used to improve prognostic parameterizations of eddy mixing coefficients and diagnostic parameterizations of optical turbulence for the tropopause and the lower stratosphere regions. Non-homogeneous, anisotropic, non-Kolmogorov patchy shear-stratified stratospheric turbulence requires that a fine mesh be used to resolve stiff velocity and temperature gradient profiles. Our approach is based on vertical nesting and adaptive vertical gridding in nested WRF/microscale codes. We perform effective ensemble forecasting by using initial and boundary conditions from both GFS and high resolution T799L91 ECMWF datasets. This methodology is applied to the analysis of field data from the Hawaii 2002 campaign and TREX Campaign (Terrain-induced Rotor Experiment), Owens Valley, CA, 2006. We obtain local distributions of simulated optical turbulence (C n2) in the upper troposphere/lower stratosphere using explicit simulations and parametrization formula that show strongly laminated structures with thin layers of high values of refractive index. These layers are characterized by steep vertical gradients of potential temperature and are located at the edges of relatively well mixed regions produced by shear instabilities and wave breaking.