Junshan,

One of the best discussions on infiltration to sinkholes or
sinkhole areas is given by Bill Jones in Karst Waters Institute Special
Publication 4 Karst Hydrology Atlas of West Virginia. Generally,
everything that does not evaporate or transpire, or is not trapped in
depressions on the bedrock will ultimately go into the groundwater. To
account for this, I estimate everything that does not evaporate and
model the infiltration as follows. I choose a runoff curve number
appropriate for surface runoff. This goes into the sinkhole directly and
on to groundwater through an orifice or culvert in the invert of the
sinkhole. I create another watershed for the subsurface flow and choose
the curve number to account for the rest of the infiltrating water that
does not evaporate. For the "underground watershed" I choose a longer
time of concentration because this water must pass through the soil
layer. You must choose this based on measurements, experience or even
SCS soil data. If the watershed feeding the sinkhole has non-karstic
formations, then you reduce the underground watershed area by that
amount on the theory that water must travel through much longer pathways
there and must have a time of concentration so long that it will not
impact flood flow.

Warren Campbell

Warren.Campbell@HSVCITY.COM