Analysis of Water Quality Pattern near the Lake Hinuma using Passive Image Data
Takashi Hoshi
Department of the Computer and Information Sciences.
Faculty of Engineering, Ibaraki University
Tel (81)-294-38-5133,Fax(81)-294-38-5282,
E-mail: hoshi@cis.ibaraki.ac.jp
Kazuhito Tanii
Graduate School of Science and Engineering, Ibaraki University
E-mail: tanii@cis.ibaraki.ac.jp
Abstract
The lake Hinuma is receiving the effect of the coastal water as it is close to the ocean.
Therefore the aquatic environment unlike general freshwater lake is shown. In this research, water quality pattern of the lake
Hinuma was analyzed using satellite image data from the passive sensor. This paper describes results of the
image classification and the principal component analysis in the Hinuma region.
Introduction
Lake Kasumigaura area (167.6km 2 ) is located in the
Ibaraki Prefecture of Japan (1), and is taken seriously
as a source of the water use, because it is used for the
drinking water. In case of Hinuma water system, it is
not suitable for drinking and the scale is small, so it is
not noticed in comparison with the lake Kasumigaura.
The lake Hinuma in Ibaraki Prefecture flows into the
coast along the Pacific Ocean from Nakagawa River.
The environment of soil, vegetation and crops have
changed delicately in the water system in river, lakes
and marshes along to the coast in order to receive the
influence of the salinity. The utilization of image data
of remote sensing has also handled to the result of
small-scale from the large scale. Though environment
around the lake Hinuma is not very much noticed up to
now, at present it should be also noticed in every region.
First, this is analyzed by the collection of field study
data for the lake Hinuma in Ibaraki Prefecture. Next,
the pattern classification is carried out using satellite
image data extracted from the passivity sensor. One of
the advantages of image data of remote sensing is to
make possible to observe simultaneously wide area.
This technology is effective for pattern analysis of water
current object that changes every moment. Maximum
likelihood method of the supervised classification and
Ward’s method of the unsupervised classification is used
for the classification technique. The results of the
pattern classification of satellite image are compared
and are examined while result of statistical analysis of
this field information is made to reference. Elements of
the water quality for the statistical analysis of field
information of the lake Hinuma are water temperatures,
water depth, turbidity, salinity, dissolved oxygen, COD,
etc. Except for these, the variable of clay, silt, sand,
moisture, specific gravity and all nitrogen is also used
for the statistical analysis in respect of the soil test of the
sediment change. The study area of chapter 2, the
analysis of water quality data of chapter 3, the sediment
of Hinuma of the chapter 4, and analysis method of
existing field study data are described. The chapter 5
describes thematic map preparation near the lake
Hinuma using the remote sensing technique.
Classification pattern of this thematic map examines if
there is any relation with the existing result of data
analysis.
The study area
Shishido river water empties into Lake Hinuma, and
it can not be measured the degree of the inflow in the
river with the small scale. The runoff from Lake
Hinuma flows in Nakagawa River, and then, it flows
into the coast. In this reason, the lake water is not
used for drinking water in order to include the thin
salinity. There is a field of algae in Lake Hinuma, and it
is nutritional lake that becomes a field of egg-laying and
water fish rearing of the fish. The major features is as
follows, area : 9.4km
2 , lake length : 20km, average water
depth : 2.1m, water surface altitude : 0m. The water
depth observation was carried out using a sound locator
and GPS in the early 1997. The geologic profile was
made from field study and existing holing materials.
The temperature of surface water was measured at
20cm depth from the water surface. Then, water
temperature of the low rise water was measured from
the lake bed in 20cm place. Turbidity, dissolved oxygen,
electric conductivity data except for the water
temperature was also measured from the material that
sampled water from two points of surface water and
lake bed of the superscription in shipboard. Soil test is
analyzed by the bottom sediment survey of moisture,
water specific gravity and grain size pattern. These data
are mainly used for statistical analysis.
The analysis of water quality data
The water depth of the lake Hinuma
To survey the water depth, 150 sites are selected from
9.4km
2 area of the Lake Hinuma, and it was carried out
by Mr. Ohshima et al. The water depth contour line was
obtained on the basis of this water depth data. This is
shown in the map of the 1/20,000 scale (Figure-1).
The correlation coefficient between water quality
variables
The data of the water surface and bottom by 18 sites in
Lake Hinuma was observed in the ebb tide in May 1997.
The observing data types are water temperatures, water
depth, turbidity, dissolved oxygen, COD, electric
conductivity, salinity and transparency. From these,
the correlation coefficient between 5 variables of water
temperature, water depth, turbidity, dissolved oxygen
(DO) and COD was obtained. Correlation coefficients
between main variables are as follows (2)
- EC-COD and water temperature -DO
(0.75³r³0.70)
- DO-turbidity (r = -0.63)
- water temperature-turbidity and water depth –DO
(r=-0.49)
- water depth - turbidity, water depth - water
tempera-ture, water depth - COD and water
depth - EC
(0.43³ ½r½ ³0.32)
- COD....water temperature (r=0.23)
- EC-water, EC-temperature, EC-DO, EC-turbidity
and turbidity -COD (0.10³½r½³0.02)
Fig.1 Water depth map of the Lake Hinuma
The principal component analysis
Main component score of surface water and bottom
layer water was plotted in the Z1-Z2 plane, because
70% at accumulation contribution ratio of primary main
component (Z1) and second main component (Z2) were
exceeded. As this result, it is divided into three groups
of fresh water effect area:a1, stagnation brackish
water:a2, salt water effect area:a3.
Figure-2, 3 shows
these three groups of a1,a2 and a3 in Lake Hinuma
map.
Fig.2 Samples to plotted in the Z1,Z2 plane
Fig.3 Groups of a1-a3 in Lake Hinuma
The sediment of Hinuma
The soil data analysis
The measurement of the specific gravity and
moisture, and particle size distribution test were carried
out after the picking material aired. Then, the element
of the statistical analysis was processed by six variables
in addition to clay, silt and sand which were the
composition element of the soil, in addition to moisture,
specific gravity and all nitrogen. It is calculated by the
correlation coefficients between six variables of the clay,
silt, sand, moisture, specific gravity and all nitrogen.
Correlation coefficient between soil variables
The correlation coefficients between main variables
are as follows.
- clay-silt (r=-0.82), all nitrogen-moisture (r=0.74)
- clay-all nitrogen, sand-moisture (r=0.63)
- sand-all nitrogen (r=0.58)
The absolute value on the correlation coefficient
between 2 variables except for the superscription was
also under 0.5 on anyway.
