In this work, we present results from a combined numerical and experimental study aimed at investigating the role of competing mechanisms understood to govern dislocation glide in polycrystalline aluminum under ultra-high strain-rates and temperatures approaching the melting point. Pressure-and-shear plate impact (PSPI) experiments are conducted to investigate the effect of temperature on the dynamic flow stress at very high shearing rates and extreme temperatures. The results from experiments show that in all cases, the flow stress of aluminum saturates at increasing high levels of plastic strains at stress levels that decrease with increasing temperature. Numerical simulations are performed to correlate the experimentally observed temperature and strain rate dependence of flow stress at small and large plastic strains using the Austin-McDowell dislocation-mediated plasticity model parametrized to normal plate impact experiments conducted in an earlier study by Zaretsky and Kanel (2012). Extensions to the model are made to better represent the dynamic behavior of pure aluminum at larger plastic strains as observed in elevated temperature split Hopkinson Pressure bar (SPHB) experiments of Samanta (1971) and Lindholm and Yeakly (1965), and the combined pressure-and-shear plate-impact experiments conducted in the present study. The main theoretical extension is the introduction of a rate- and temperature-dependent dynamic recovery function and evolving obstacle spacing, which effectively reduces the mean free path and rate of accumulation of dislocations at large plastic strains. The numerical predictions of the revised and re-calibrated plasticity model are brought into agreement with the experimental observations, by recovering the strong agreement between the observations of an anomalous increase in dynamic yield stress at incipient plasticity with the original model and correlating well with the flow stress levels at larger plastic strains and elevated temperature