Geology Reference
In-Depth Information
hybrid models [
14
,
18
,
25
,
28
]. Tan et al. [
26
] and Piri et al. [
21
] used ANNmodels for
modeling daily pan/open water evaporation in dry countries. Kisi and Cimen [
15
]
used SVM for evapotranspiration modeling, utilizing several meteorological data
from three stations (Windsor, Oakville, and Santa Rosa) in central California, USA.
Recently, Kişi
i and Tombul [
18
] have used a fuzzy genetic (FG) approach for mod-
eling monthly pan evaporation. Abdullah et al. [
1
] used a hybrid of Arti
ş
cial Neural
Network-Genetic Algorithm for (ET
0
) estimation in Iraq. El-Sha
e et al. [
8
] suggest
that the ANN model is better than the ARMA model for multi-lead ahead prediction
of evapotranspiration. Evapotranspiration is a complex natural process in which
many environmental variables interact in a nonlinear fashion. The advantage of data-
based models is that they do not require comprehensive physical details of the system.
Therefore, in data-scarce situations, researchers tend to use such data-based models.
The application of ANNs comparisons with standards for evapotranspiration esti-
mation studies has been less frequent than comparison on data-based models in
evapotranspiration modeling. Therefore, this chapter has two objectives. The
first is
application of traditional and widely applied evapotranspiration models in two totally
different environments and comparison of their results with the ANN model. The
second application is the implementation of data selection procedures in evaporation
modeling input space and then applying soft computational models such as ANNs,
ANFIS, SVM, and wavelet-based hybrid data-based models (with different combi-
nations of inputs) on evaporation data.
7.2 Study Area
In this chapter, meteorological data from Santa Monica Station (USA), Brue
(United Kingdom), and Sistan Region (Iran) were used for modeling purposes.
7.2.1 Santa Monica CIMIS Station
The daily climatic data from the automated weather station at Santa Monica Station,
California, USA (latitude 34
W), operated by the California
Irrigation Management Information System (CIMIS), were used in this study, mainly
for comparison of ET estimation using traditional equations with neural network
models. The CIMIS is a project developed in 1982 by the California Department of
Water Resource and the University of California at Davis to assist California
°
30
′
N, longitude 118
°
29
′
s irri-
gators to manage their water resources effectively for better productivity. The Santa
Monica Station is one of over 120 automated weather stations in the state of Cali-
fornia managed by CIMIS, which is located at an elevation of 104 m. The daily
climatic data for the Santa Monica Station was downloaded from the CIMIS web
server (
http://www.ipm.ucdavis.edu/WEATHER/wxretrieve.html
)
. The daily data
from 1st January 2000 to 31st December 2005 was used for this topic. The total
'
