While experimenting with the "contributing area" approach of application model hland_v1 (following HBV96), we concluded that we should allow defining zones as "sealed areas". Then one could hopefully improve the model's skill in predicting small summer flood events without introducing an inverse relationship between flood generation and soil moisture.

My first idea was to let sealed areas work like FIELD or FOREST with zero field capacity. This simple approach would increase the amount of water available for runoff generation. However, it does not need to become direct runoff by necessity. For dry summer conditions and with a moderate fraction of sealed areas within a subbasin, runoff generated on such sealed zone would fill the subbasin-wide upper zone storage only a little. Hence, much of this runoff could percolate to the lower zone layer and become base flow instead of direct runoff.

I see four alternatives for handling the runoff generated on sealed areas:

1. Route it directly to the subbasin outlet
2. Pass it into the unit hydrograph
3. Pass it to a separate (linear) runoff concentration storage
4. Combination of alternatives 3 and 4

Alternative one would not work well for large subbasins. I would go for alternative two, which does not introduce any new parameters. If we realise that we need more flexibility, we could introduce another way to model the runoff concentration of sealed surface flow later.