Agriculture Reference
In-Depth Information
Figure 12 shows examples of the validation of the CTM comparing the simulated
daily river flow discharge with the observed one at Shimobuchi, Suda and Funato in
1990. Let Q sim represent amount of the simulated daily river flow (m 3 /s), let Q obs represent
amount of the observed daily river flow (m 3 /s), let N represent the number of data. The
annual mean relative error of the simulated daily river flow to the observed one is defined
as following equation:
(1)
Table 5 shows annual mean relative errors of simulated daily river flow to the
observed flow. It shows that the model simulates, with a desirable level of accuracy at
each point in most years, namely 21.9% at Shimobuchi, 24.8% at Suda, and 34.3% at
Funato in an average of 8 years.
Table 5. Annual mean relative errors of simulated daily river flow to observed daily river flow (%)
Point
1983
1984
1985
1986
1987
1988
1989
1990
Ave.
Shimobuchi
30.2
24.2
20.1
15.9
24.5
17.1
18.3
25.1
21.9
Suda
23.2
29.1
21.3
24.3
25.0
24.5
24.1
27.0
24.8
Funato
43.9
40.0
34.4
28.8
33.3
22.4
27.4
43.8
34.3
the component ratio of return flow from upstream
agricultural land in the diverted flow discharge at each diversion work. The return flow
￿￿￿￿￿￿￿￿￿ ￿￿￿
Note that
￿ ￿ ￿ ￿
￿￿ ￿￿￿
return flow should consist of only water irrigated in the upstream farmlan
the
￿ ￿￿￿
regarded as the irrigation water reuse ratio (%) at each diversion work. Figure 13 and
Figure 14 show the
￿￿ ￿￿ ￿￿￿￿￿￿￿
s
￿￿￿ ￿
￿ ￿￿￿￿￿￿￿￿￿￿￿￿￿￿￿
owest reach in the
river basin. The year 1987 has the lowest precipitation of 617 mm during the irrigation
period; that in 1988 was 1076 mm and that in 1989 was 1335 mm. It predicts that the
lower the precipitation, the higher the irrigation water reuse ratio at the lower reaches in
the basin.
Search WWH ::




Custom Search