Integrated stock assessment models are widely regarded as the gold standard assessment methodology but successfully implementing them in non-standard fisheries can be challenging. Inland fisheries such as those in the Murray-Darling Basin, Australia are examples of such non-standard fisheries, with strong historical fishing pressure, intense ecological degradation and large-scale alteration of hydrological conditions. Each of these conditions presents a potential challenge to the implementation of a standard integrated stock assessment model in terms of the underlying assumptions in population dynamics. To test the applicability of standard integrated stock assessment models we conducted simulation experiments on hypothetical inland fisheries using an operating model-estimation model framework based on standard “Stock Synthesis” software. We altered specific components of the operating models then ran estimation models which did not account for these alterations. Our experiments included changes in carrying capacity over the model duration, catch misreporting, non-constant (spikes in) natural mortality, and autocorrelation in recruitment deviations. Our results showed that relatively small changes in the above experiments can result in biased estimation model outcomes except for autocorrelated recruitment which increases the variability of assessment outcomes at high levels of autocorrelated recruitment. Simple combinations of variations are also unpredictable with interactions between factors resulting in non-linear biases in estimated stock biomass. These results suggest that standard integrated assessment models should be used with extreme caution in non-standard fisheries such as the Murray-Darling Basin where key assumptions or population dynamics may not be accurately represented.