FP 076 Quality assurance of hydrological data

1
FP 076Quality assurance of hydrological data
Dr. Takeo Kinosita Representative, Suimon
Kankyo Tokyo, Japan
2
1. Introduction
The author already discussed about such
problems2)3) particularly at the Steering
Committee Meetings of Global Runoff Data Centre
in Koblenz and International Association of
Hydrological Sciences workshop at Sapporo in
2003.4)
3
2. Elements of hydrological observation
The most important elements of hydrology,
namely, precipitation, water level and discharge,
are discussed in this report for water resources
development and management.
4
Photo 1
Raingauge in snowy area in Shirataki, Hokkaido
5
Photo 2
water gauge (Float type) in Gyotoku, Tottori
6
Photo 3
Drift rod
7
Fig.1
raingauge
Various windshield
8
Photo 4
Field Calibration of the tipping bucket raingauge
9
Fig.2
cable connection
10
Photo 5
Drained water frozen
Raingauge, sensor frozen, Shimo-shiobara, Tochigi
11
Photo 6
Raingauge on the roof not preferable
12
(2) Staff gauge
2)Where a few staff gauges are set at the
riverside, each staff should overlap 0.5m with
both the upper staff and the lower staff like
Photo 7. In many cases, this rule is forgotten.
13
Photo 7
Saff gauges, 0.5m overlapping,Kashima, Shizuoka
14
3) The highest staff must be set higher than the
levee. Otherwise the extra-ordinary flood
cannot be observed. If such a high water is
observed, it will be perhaps selected as the
design flood in the future.
15
Fig.3
Reed switch water gauge
16
(2) Low-water observation
2) In the coarse-grained riverbed, river water
easily infiltrates into the ground. Contrarily,
ground water naturally comes out if the water
table is high.
(3) High-water Observation
17
Daisenji, Joganji River
18
4. Data assurance
Data processing is explained in the WMO Guide
to Hydrological Practices.1) The Japanese
hydrologists also decided data processing
procedures by their own experiences.
19
4.1 Precipitation data
Correlation
A, B and C names of stations good high
correlation bad low correlation In this example,
A maybe doubtful. Fig.4 Correlation detection
20
(No Transcript)
21
(4) In order to discriminate an abnormal
precipitation from the normals, it must be
investigated whether the abnormal precipitation
exceeds the sum of the average and three times of
the standard deviation of the yearly maximum
precipitation series. See Fig.5.
22
Fig.5
criterion of abnormal
23
Fig.6
Left Complicated precipitation contour lines,
because including erroneous data. Right Smooth
precipitation contour lines, because erroneous
data are removed.
24
4.2 Water level data
(1) A sudden rise water level is suspicious. (2)
A sudden fall is generally implausible. (3) A
spike noise sometimes appears. (4) The value 0m
must be distinguished from missing.
25
4.3 Discharge data
H
Fig.7 Loops of the water-level discharge
relation
26
Fig.8
Surface gradient correction
27
(3) Outlier check based on uniform flow
assumption
A rating curve is formulated as Qa
(Hb)2 where Q is the discharge, H is the water
level, a and b are determined by observed data.
In the case of many observed data, the least
square method can be applied to determine a and
b. In the case of the outlier check, under the
assumption of uniform flow and the steep river
bank, a rating curve can be elongated within some
range.11) aV(1p)2B2/(4A) Hb
2R/(1p) where V is the mean velocity over the
observation section, p is the power of the
uniform flow formula (p of Mannings formula is
2/3), B is the width of the flow section, A is
the flow area, R is the hydraulic radius.
28
Fig.9
Slight bent on H- Q relation
29
4.4. Discharge hydrograph
(1) Correlation Qu(t)and Qd(t) respectively
denote the upstream discharge and the downstream
discharge at time t. Travel time of a flow
between two stations is T. Supposing Qu(t-T) is
measured on the abscissa and Qd(t) is on the
ordinate, the plottings stand on a clear curve as
shown in Fig.10. If any plotting is at a
separate place, it perhaps is erroneous.
30
Fig.10
downstream discharge
Qd(8)
in the case of T3
upstream discharge with the travel time.
Discharge hydrograph correlation
31
(3)The peak of the downstream discharge C appears
on the depletion curve of the upstream discharge,
shown in Fig.11. If the peak C exists at the
different place, the observation system must be
carefully investigated.
32
Fig.11
Comparison of discharge hydrographs at the
upstream and the downstream
33
discharge
discharge
good
erroneous
time
time
discharge
discharge
good
erroneous
time
time
Qu upstream discharge
Qd downstream discharge
34
5. Conclusion
35
Acknowledgement
The author would like to express hearty thanks
to hydrological field observers who are working
for accurate data acquisition both in hot
sunshine season and in frozen winter.