The use of 2D hydraulic models for the prediction of instream habitat has drawn a lot of attention, especially for complex hydraulic circumstances like those seen in braided rivers.It is generally believed that 2D models with higher spatial resolution and associated hydraulic modelling will provide more accurate forecasts of instream habitat. We apply a 1D model, two 2D models, and compare the predicted and measured water depths and velocities at two flows, as well as the anticipated habitat over a range of flows, to a portion of a braided river. The correlation between predicted and measured depths and velocities was higher for the 1D models compared to the 2D models.The subjectivity involved with 2D calibration and practical constraints on topographic characterization led to mistakes in anticipated water levels that might either cause braids to flow or cease flowing. Although there were discrepancies in the magnitudes, locations of maxima, and changes in gradient, all three models largely anticipated the same patterns in habitat (weighted useable area) variation with flow. Because there were as many discrepancies between the 1D and 2D habitat-flow relationships as there were between the two 2D models, the differences between the predictions of the 1D and 2D models could not be explained by the higher spatial resolution of the 2D models.The fact that they need more information does not automatically mean that they are better, and many scenarios do not call for the time and effort needed to create an effective 2D model. Their practical use is constrained by the challenge of obtaining sufficient and precise bed topography and the expertise needed to calibrate 2D models. The primary benefit of 2D models over 1D models is that, especially at high flows in braided rivers, they should be able to produce forecasts that are more accurate outside of the 1D modelscalibration range. However, better calibration and validation approaches are needed.