58
Tecnología y Ciencias del Agua
, vol. VIII, núm. 2, marzo-abril de 2017, pp. 51-60
Kan
et al.
,
Daily streamflow simulation based on improved machine learning method
•
ISSN 2007-2422
purpose of comparing performances of differ-
ent models, the PKEK and NU-PEK models are
applied to daily RR simulation in the Chengcun
catchment. It can be remarked that the PKEK
model is able to simulate generally much better
runoff hydrograph than other models. From the
application and analysis of the PKEK model, we
can conclude that:
The PKEK model have all the advantages
of the previously proposed NU-PEK model,
including better calibration and validation
non-updating simulation accuracy, automatic
calibration without much impact from human
experiences, good compromise between the
sufficiency and parsimony of the input data,
and good compromise between the simulation
accuracy and topology complexity, and so on.
Compared with the original NU-PEK model,
the performance of the PKEK model becomes
much better for the daily RR simulation. The
improvement in the goodness-of-fit to the
observed discharge of the PKEK model is
Figure 4. Observed and simulated hydrographs of the NU-PEK model: (a) Calibration; (b) Validation.
