A channel-estimate-based decision feedback equalizer (CEB DFE) robust under impulsive noise is presented for single-input multiple-output (SIMO) underwater acoustic com-munications. Channel estimation is performed via the improved-proportionate M-estimate affine projection algorithm (IPMAPA), a linear complexity algorithm which is robust against impulsive interference and exploits channel sparseness. The superiority of IPMAPA to the normalized sign algorithm (NSA) and the normalized least-mean-square (NLMS) algorithm is demonstrated by processing data transmitted at 9 kbps over a 1.2 km shallow water environment contaminated by snapping shrimp noise.