This paper proposes complexity reduction and performance enhancement schemes for the iterative row-column soft decision feedback algorithm of Cheng et. al., one of the leading algorithms for turbo equalization of two dimensional intersymbol interference (ISI) channels with additive white Gaussian noise. For complexity reduction, we propose sorting and truncation of the soft decision feedback probability configurations. On the 3??3 ISI channel, this technique reduces complexity by more than 95% with only marginal performance loss. For performance enhancement, we propose post processing the final log-likelihood ratio estimates by thresholding and clustering, followed by a localized squared Euclidean distance (SED) search against the received 2D signal. The SED search yields up to 0.4 dB performance gain, with modest additional complexity.