Despite the strength and a increasing interest in application of artificial neural networks (ANNs) to rainfall runoff
simulating, the deficiencies associated with traditional applications of ANNs in which the networks essentially function
as black box models is obvious. The objective of this work is therefore to enhance the ANN-based rainfall runoff
models' ability in the description of hydrological processes such as interception, infiltration, surface runoff, sub-surface
runoff and evapotranspiration by integrating it with TOPMODEL, which is a simple physically based rainfall-runoff
model and has become increasingly popular and widely used in a great number of applications in recent years. A new
integrated model named ANN-TOPMODEL is proposed in this study. Baohe River basin (2413 km2), located at the
upper stream of the Hanjiang Catchment in Yangtze River Basin, China, is selected as the study area for testing the new
model. The results show that the daily stream flows simulated by the new model are in good agreement with the
observed ones, while the daily stream flows simulated by TOPMODEL greatly overestimates or underestimates some
peak flows both for calibration period and validation period. Further more, the new model resulted in a Nash and
Sutcliffe efficiency coefficient value of 0.905 for validation period, which is significantly larger than TOPMODEL. The
results demonstrate that the proposed integrated model based on ANN and TOPMODEL is promising in daily stream
flow modeling.
|