Geology Reference
In-Depth Information
N
X
N
1
1
ð
Q
m
Q
o
Þ
Mean Absolute Percentage Error
MAPE
¼
100
ð
6
:
2
Þ
;
i
¼
Q
o
E
1
E
2
Efficiency
E
¼
ð
6
:
3
Þ
;
E
1
where
X
N
i
¼
1
ð
2
Q
o
Q
o
Þ
E
1
¼
ð
6
:
4
Þ
and
X
N
i
¼
1
ð
2
E
2
¼
Q
0
Q
m
Þ
ð
6
:
5
Þ
r
1
N
X
N
2
Root Mean Squared Error
RMSE
¼
i
¼
1
ð
Q
m
Q
o
Þ
ð
6
:
6
Þ
;
P
i
¼
1
ð
Q
o
Q
m
Þ
Mean bias error
MBE
¼
ð
6
:
7
Þ
;
N
Variance of the distribution of differences S
d
which expresses the variability of
(Q
o
-
Q
m
) distribution about MBE
P
i
¼
1
ð
2
Q
o
Q
m
MBE
Þ
S
d
¼
ð
6
:
8
Þ
N
1
where, Q
m
is the modelled or estimated runoff by a data based model, Q
o
is the
observation runoff
Q
m
is the average of the estimated runoff,
Q
o
is the average of
the observed runoff and N is the number of observations.
6.4 Data Selection Approaches in Data Based Rainfall-
Runoff Modelling
The data selection models like GT, Entropy Theory, BIC and AIC were applied
here in the context of rainfall- runoff modelling. In this chapter analysis is per-
formed mostly on the daily information from the Brue catchment.