I don't think this is currently handled in mor - i.e. it is possible to have negative lagoon width - see https://github.com/RegMeasures/HapuaModel/wiki/Notes-on-logic-rules-in-mor.updateMorphology
I guess this is not a particularly big issue because at this stage it will be closed in riv.assembleChannel but it should probably be handled for completeness (and mass balance).