For the first time, mesoscale simulations of shocked explosives are validated with hot spot temperature data. Following recent experiments, we simulate a 25 mu m aluminum flyer impacting a powder bed of octahydro-1,3,5,7-tetranitro-1,3,5,7-tetrazocine (HMX) with an initial speed of 1.4 to 3.6 km s(-1). The powder bed has a density of 0.56 g cm(3), or 29 % crystal density, and an average grain size of 4 mu m. Simulations are performed in the multi-physics code, ALE3D. We employ a comprehensive HMX material model that accounts for the chemistry, unreacted and product equation-of-states (EOSs), elastic-plastic response, as well as melting and melt viscosity. Simulations show compaction, reaction, and gas jetting during the initial stages of shock in the powder bed. Shock reflection at the glass window interface increases hot spot temperatures in HMX product gas in excess of 7000 K at higher impact speeds which is readily observable using pyrometry techniques. We extract a temperature histogram from simulations at discrete times, using it to calculate an effective emission spectrum and effective temperature, T-eff, for comparison to data. Simple T- and T-4-based methods inadequately weight the higher temperature regions and result in a low T-eff that is in poor agreement with data. However, calculating T-eff based on Planck's Law with a gray body assumption demonstrates good agreement with hot spot temperature data. A modified Arrhenius model is fit to the normalized reaction rates predicted by mesoscale simulations and inferred temperatures range from 1544 K to 1779 K; these temperatures are approximately 4x lower than the peak hot spot temperatures but 2-3x higher than the bulk shock temperatures. The agreement with experimental results validates our mesoscale simulations as useful tools for elucidating the role of hot spot mechanisms in the shock initiation of heterogeneous explosives.