Results showed that with the decreasing of raingauge density the biases (indicated by RMSE) of rainfall field estimates increase and the potential variability in rainfall field due to random sampling in raingauge location is exaggerated. By contrast, biases in model hydrographs are significantly smaller than that in rainfall field. When the raingauge governing area is <10 km2/gauge, the biased rainfall field shows no detectable effect on hydrographs. Incomparably lower RMSE in hydrograph indicates that surplus and deficit rainfalls at different locations were compensated in model simulation. In term of reliable hydrograph simulation, obviously, the criterion for raingauge density is not as high as that for rainfall estimate. When gauge governing is <20 km2/gauge, both the rainfall and discharge were successfully (±10% error) estimated in term of total volume. Accordingly, we suggested that covering area ~20 km2/gauge is acceptable for raingauge deployment to constrain the inherent variability in rainfall field and hydrograph simulation in mountainous watersheds.