Geology Reference
In-Depth Information
Fig. 6.24 The observed versus the LLR predicted daily runoff at the Brue catchment a time series
plot of the training data set. b scatter plot of the validation data set
6.5.4 Data Driven Rainfall-Runoff Modelling
with Neuro-Wavelet (NW) Model
For this purpose, a multi-layer feed-forward network type of arti
cial neural net-
work (ANN) and discrete wavelet transfer (DWT) model were combined together to
obtain a neuro-wavelet (NW) model. The discrete wavelet transfer model is func-
tioned through two sets of
filters, namely high-pass and low-pass
filters which
decompose the signal into two sets of series namely detailed coef
cients (D) and
approximation (A) sub time series respectively. In the proposed NW model, these
decomposed sub series obtained from DWT on the original data directly were used
as inputs of the ANN model. The NW structure employed in the present study is
shown in Chap.
4
. Antecedent rainfall and runoff information of the Brue catchment
was decomposed to three series of detailed coef
cients (D) and three series of
