Rainfall-runoff paradox from a natural experimental catchment

As a part of the Chuzhou hydrological experimental system,the No. 1 experimental catchment,Nandadish,with drainage area of 7897 m2 sits on the andesite bedrock with Quaternary deposit of 2. 46 m in average. Various runoff components,surface runoff and subsurface runoff including interflow from unsaturated zone,groundwater flow from saturated zone are physically measured using special designed troughs. Several combined types of runoff components are identified as the SR type with surface runoff dominated,SSR type with subsurface runoff dominated and other intermediate types. Examples show that surface runoff accounts for 65% of total runoff for SR type,while the subsurface runoff accounts for 90% in SSR type. In July,the main rainy season,in total,the subsurface runoff contributes an amount of 54. 5% of total runoff while groundwater flow accounts for 33. 0%. Most 18O data of surface runoff is quite different from that of precipitation. Within the rainfall-runoff process with duration of about 1 400 minutes,averaged 18O of precipitation is-1. 210%,while that of surface runoff is-1. 132% for Hydrohill catchment(512 m2) ,-1. 065% for Nandadish catchment and-0. 801% for Morningflower(4573 m2) which is a catchment with thin layer of rock debris on bedrock. It challenges the assumptions involved in current isotopic hydrograph separation,i. e.,the 18O of surface runoff will not always equal to that of event precipitation and,the evaporation fractionation during the pathway of runoff components could not always be ignored. Event rainfall produces runoff but such runoff contains an amount of water not from the event rainfall,such a paradox exists in all of our experimental catchments. The total old water involved in event runoff accounts for 16% for the SR type while 64% for SSR type.