It appears that the normalized depth-area curves (and other related curves) used in the SWMM model have some errors. This is likely a carry over from how the curves were developed from legacy versions. This issue was first reported by John Barton at MSDGC. I have been able to confirm these errors by comparing the depth-area curves used in SWMM with ones derived by integrating the depth-width curves (which have been confirmed to be correct). Where appropriate, the numerically integrated curves from the depth-width curves will be used to correct these errors by replacing the relevant curves (i.e., A v. Y, R v. Y, Y v. A, S v. A).

