Broadband data from the NRDC deployment of stations in Kazakhstan at Karkaralinsk (KKL) and Bayanual (BAY) are subjected to detailed waveform modeling to understand the nature of regional wave propagation and to determine the nature of the explosion time function and sources of low- and high-frequency “tectonic release.” Minor modification of a crustal model inferred from DSS studies in the region is very effective in explaining the arrival times, amplitudes, and waveshapes of the Pn and PmP arrivals at both stations. The SH-wave data are remarkable, however, for showing a large, high-frequency arrival after the expected and observed arrival of SmS from the explosion source. These secondary S arrivals imply a secondary source 1 to a few kilometers to the south of the shot point. Shear waves from this source dominate the seismic data at high frequency but are minor at low frequency where shear waves and surface waves from the inferred relaxation of the explosion cavity dominate the wave field. The secondary S-wave source is inferred to represent explosion-driven faulting with no net moment. Significant overshoot in the RDP is required to match the P-wave data. It is found that von Seggern-Blandford and Haskell source functions are incapable of producing the observed overshoot. However, the Mueller-Murphy source model is very effective in matching the clear overshoot effects seen in the P waves. Spall effects are also considered but appear to be minor and serve to require even more overshoot in the RDP. The excitation and dispersion of Rayleigh waves in the period band of 1 to 10 sec is seen to be a sensitive, but nonunique, function of both VS and VP in the upper crustal layer. High Poisson ratios in this layer are needed to explain the Rayleigh waveform.