Volume Errors

I'm simulating a 1 year simulation of a simple (<100 segment) reservior with 6 branches, 7 inflows, and 1 withdrawal structure. I keep getting the following warning: Computational warning at Julian day = x.xxx spatial change = 0.11621489E+07 m^3 temporal change = 0.11641099E+07 m^3 volume error =-0.19610035E+04 m^3 starting around this point volume error starts to accumulate to large values and the water surface elevation deviates from the data (typically the model is underestimates the elev/volume). i've made sure that all my flows and volume changes are balanced. and i've confirmed that my volume-elevation table is accurate. i've tried to adjust the timestep (DLTMAX) down during the onset of the volume errors, but it doesn't help much (often, below a critical DLTMAX, the model goes from a gross underestimate to a gross overestimate). any ideas as to how to correct this error? thanks in advance, harry hzahakos@qeallc.com