Abstract: Traditional flow measurement methods (e.g. flow meter method, buoy flow measurement method, index flow rate method, specific drop area method, etc.) for most small and medium-sized rivers at present are featured with large investment and large time consumption. In association with traditional sectional flow calculation theory and the latest surface velocity measurement technology, we propose a method of determining the stage-discharge relation of river section via inputting large sample data of surface velocity and sectional water level. This method needs only a small amount of observed flow data at medium and high water levels to calculate flow rate and fit the stage-discharge curve. Compared with traditional methods, the proposed method reduces the workload of flow rate measurement required for fitting the stage-discharge curve, and cuts down the errors brought about by measurement, hence enhancing precision. It is also convenient for operation with low cost.

Key words: large sample data model, small and medium-sized rivers, flow measurement, water level, stage-discharge relation curve