An information system for water resources by ngs20854


									           Hydrologie Applications of Space Technology (Proceedings of the Cocoa Beach
           Workshop, Florida, August 1985). IAHS Publ. no. 160, 1986.

           An information system for water resources

           R. LEWELA & Y, SUCKSDORFF
           National Board of Waters, Helsinki,                     Finland

A computer based information system for water resources has
been developed in the Hydrological Office of the National
Board of Waters in Finland. The system contains a central file
of all data on watercourses and drainage basins and is compat-
ible with other environment file systems having geographical
coordinates or drainage basin numbers as co-ordinating enti-
ties. In addition to the central file all the files of the
National Board of Waters will be combined to the system. A
digital terrain model digitized from maps of scale 1:200 000
is also used. The physiographical factors of a drainage basin
are interpreted from pictures obtained from Lansat or SPOT sa-
tellites and weather satellites are used to interprète rapidly
changing entities such as snow covered area, freezing and
thawing of the lakes and the temperature of surface water.
   The system will cover the whole of Finland by 1986.

There are many factors which have contributed to the develop-
ment of an information system for water resources in Finland.
Firstly, there has always been a need for different combina-
tions of hydrological data for many purposes,, for example for
research work, for hydrological and water resources planning
and for monitoring natural changes. Secondly, in recent years
many hydrological registers have been stored in computer
memories Instead of manually. Thirdly, a very good program for
the collection, processing, recording, combining and output of
digital data and digital data bases has been developed by the
National Board of Survey in Finland. The same institute pre-
pared the base map and digital terrain models of Finland. Fi-
nally, it has recently become possible in Finland to use sa-
tellite pictures operationally, eg. for Interpreting the phys-
iographical factors of drainage basins.
   In the following the system is described and some results
obtained during its development and utilization are presented.

General description of the system
The system is presented In Fig. 1. It consists of a central
file of all drainage basins and a program which collects data
from different files, combines and processes it and presents
the processed data in tabular and graphical form.
   All the registers of the National Board of Waters will be
combined into the system. Other registers having geographical
coordinates or basin numbers can also be linked. Basin char-
acteristics are interpreted by means of satellite images. The
percentage of cultivated land, foresttypes, drained bogs,
bogs and lakes in a drainage basin will be interpreted from
Landsat images. Weather satellite images will be used for the
interpretation of rapidly changing entities, such as freezing
448 R.Lewmela s      Y.Sucksdorff

              Hydrological registers:                      Drainage basins:
                water level, runoff, snow,                   border lines
                lakes, ice cover, water                        x and y coordinates
                temperature etc.                             names
              Other water registers :                        outlets
                water quality, quality of
                precipitation, bio,
                poisons in water, observers etc.

              Other registers:
                digital terrain models, waste
                waters etc.

              Land-use satellites:
                Landsat, SPOT

              Weather satellites

              Other data with
              x- and y-coordinates
                                           \u v
                                    Landinformation system program
                                      creation of data files
                                      selection of data from files
                                      combination of data

                                    Material produced
                                      in tabular form
                                      in graphical form
                                         thematical maps

                                      research, planning, monitoring

             FIG.l       General description       of the information   system   of
             water       resources.

and thawing of lakes, snow cover distribution and the temper-
ature of surface water in lakes.
Drainage basins in Finland
Finland is divided into 74 drainage basins. These main basins
are each divided into three smaller area classes, making 9^
areas in a£l. The arenas of the smallest sub-basins are thus
from 20 km to 30 km . In addition the coastal areas of Finland
are divided into areas of about the same size as the drainage
   The divides of the drainage basins are drawn onto maps of
scale 1:50 000. These maps are reductions of the base map of
Finland, which is originally to the scale 1:10 000 (1:20 000 in
Northern Finland). The divides are drawn with the help of to-
pographical contour lines, which are marked on the maps every
2,5 meters in the vertical direction (every 5 meters in
Northern Finland). Aerial photographs and field measurements
are used in difficult terrains, such as bifurcations and large
                           Information   system for water resources   449

The land information program
The National Board of Waters has obtained a program called
PINGIS (Finnish Geographical Information System) (Keisteri
and Tuhkanen, 1982) from the National Board of Survey for col-
lection, processing, recording and output of all spatially ori-
ented data. The capability of the program of handling posi-
tional information also makes it possible to link different
registers together, if the information can be located with
   Other common links can be used to combine registers, too,
for example the numbers of the drainage basins.
   The units to be processed in the map data base are points,
lines, areas, connections and texts. A line goes from node (end
point) to node, and the nodes are linked to all their lines.
Areal features are also linked to all surrounding lines.
Digitizing of drainage basin maps
The drainage basin maps are digitized with a digitizer and a
microcomputer. All the data is then sent to the master computer
and data bases (drainage basins) are generated with the FINGIS-
program. Corrections are made with a graphical working station
with the- help of the same program.
   Digitizing starts by orientation of the map by giving the
coordinates of the corners of the map and digitizing them.
After this every divide line and the sites of the release points
and drainage basin numbers are digitized. When digitizing, dif-
ferent meny orders are used, indicating the line type (there are
five different line types depending on the drainage basin), out-
let, text etc. The numbers of the maps and the drainage basins
are also fed to the microcomputer.
   In order to make the drainage basin register as compact as
possible, extra digitized points of drainage divides are filter-
ed. The digitizing speed is normally 3-2 coordinate pairs per
second, at which speed the number of digitized points is on
average 9.2 points per kilometer. Filtering is done with a prog-
ram which calculates ellipses along a digitized line using the
points i and i+2. If the point i+1 is within the ellipse it is
filtered, otherwise it will be the first point when forming a
new ellipse. The perpendicular axis of the ellipse is a.' given
tolerance, which represents the accuracy of digitizing plus the
accuracy of the basin divide in the horizontal direction. Seven
drainage basins were digitized several times, with a mean areal
error of 0.14 %. This corresponds to a tolerance of about 50 m
(1 mm on a map of scale 1:50 000). This is the tolerance used
in filtering. A greater tolerance could be used, but in that
case the accuracy of the basin divide should be known exactly
although in fact it varies. The filtering degree is 73 % when
using a tolerance of 50 meters, leaving 2.5 points/km. To dig-
itize, filter and transmit one map in the scale 1:50 000 to the
main computer takes„about four hours. The area covered by one
map sheet is 600 km .

Height and slope of the watershed
Two digital models were used to calculate the average height,
average slope and slope direction. Both models were con-
structed by the National Board of Survey in Finland. The first
model was digitized from maps having a scale of 1:10 000 and a
vertical distance between contour lines of 2.5 m. The covering
450 R.Lemmela   &   Y.Sucksdorff

net has a length of 50 m in this model, but a length of 100 m
is used in the calculations. The second model, called the "GT-
model", is digitized from maps having a scale of 1:200 000 and
having a vertical distance between contour lines of 20 m. The
net size is 200 m x 200 m.
    The models were tested in seven drainage basins with areas
ranging from 10.9 to 40.2 km . Errors were calculated for some
of the above mentioned parameters and differences between the
models for some others.
    The difference in the average height of a drainage basin
between the models was on average 1.7 m and the standard devia-
tion was 1.5 m. The root mean square error of one point in the
"GT-model" is 2.15 meters when compared with the height of the
nearest contour line. From these numbers it is concluded that
both models can be used when calculating the average height.
    The mean slope of a drainage basin was calculated in two
ways. First, the local slope was calculated with help of a
square of 3x3 points. Then the mean slope of the whole drainage
basin was calculated as an average value of these points. The
difference between these values was on average 58 % when the
two models were compared, so the average of local slopes cannot
be used. Instead the slope from every divide point to the out-
let point was used and the distanceweighted mean of these
values was calculated. The differences between slope values
of the models in three areas were 0.8 % of the slope, varying
from 8.0 to 13.0 %. The standard deviation for the slope was
calculated theoretically and was 17 % for the 1:200 000 model
and about 9 % for the 1:10 000 model. When all the basin
points were used the corresponding figures were 23 and 14 %,
respectively. When all the points were used the difference in
average slope (GT-model) varied from 1.0 % to 3-3 %• The
average slope then became smaller because the slope becomes
smaller with decreasing distance from the outlet point. In
conclusion both models can be used when the average slope of a
basin is calculated. Better results are reached when all the
points are used, but it is also possible to use only the di-
vide points. The height of the outlet point should be dig-
itized from a base map and not taken from the terrain model.
   The average slope direction should be calculated from
local slope directions as a percentage of for instance eastern,
southern, western and northern slopes in a basin.
Basin characteristics determined using satellite data
Basin characteristics, such as cultivated land, forests, bogs,
drained bogs and lakes will be interpreted from satellite
images. Research has been carried out with Landsat images and
an operational system has been started this year (1985). Kuitti-
nen and Kauppi (1981) used images from Landsat-2 and compared
digitally interpreted land use types with base maps and the na-
tional forestry inventory number 6. When they compared land use
interpreted from Landsat images with land use interpreted from
base maps (16 points per km ) they found the following corre-
lations: tilled ground outside densely populated 2 areas R2 =
0.95, tilled ground in densely populated areas R = 0.84, for-
ests and forest-supporting bogs together R2 = 0.95 and forests
R 2 = 0.64. Bogs had a correlation of R2 = 0.35 because most bogs
had a stand of forest.
    The amount of growing stock was also interpreted from satel-
                            Information   system for water resources   451

lite pictures and was compared with results from the national
forestry inventory. The results were 10770 m /km 2 from Landsat
pictures and 10910 m3/km 2 from the forestry inventory. The in-
terpreted area was 800 km 2 . In smaller areas the error could
probably be greater, but this could not be estimated because
results from the national forestry inventory method used in Pin-
land cannot be used in areas smaller than 500 km 2 .
   Research is continuing and operational interpreting from both
TM- and MSS-images will start in 1986. Physiographical factors
for all river basin parts (area 20...30 km 2 ) will be interpreted.
It is planned to carry out a reinterpreation after 10...20 years
in order to monitor changes in physiography of the drainage

The use of weather satellite data
In addition to the above-mentioned permanent or slowly changing
data it is planned to file data from the interpretation of rap-
idly changing entities seen in weather satellite images. NOAA
images have been used to determine the snow covered area and
provisionally to determine the freezing and thawing of lakes.
The temperature of surface water will also be estimated from
NOAA images.

Effect of basin characteristics on the quantity of water
Lemmela and Kuittinen (1976) summarized the effects of basin
characteristics on the quantity and quality of water. The re-
sults for quantity are mostly from small catchment areas (from
0,07 to 122 km 2 ) but when the water information system is in
operational use results should be available for all sizes of
basins. In Finland for example (Mustonen, 1965) the effect of
catchment area is such that the maximum runoff in general de-
creases with increasing area. The effect of the shape of the
catchment area on maximum runoff is dependent on the location
and length of the watercourse. Increase in altitude and land-
slope usually increases the maximum runoff. Lakes and artifi-
cial waterstorages have a moderating effect on water level vari-
ations. Furthermore, when the drainage area and the lake percent-
age increase, the date of high water is postponed. Reduction of
river flood areas increases the quantity of high water in the
lower reaches of the watercourse and the date of high water oc-
curs earlier. With increasing permeability of soil the discharge
is levelled. However, the permeability of soil may be consider-
ably reduced by frost. Increase and decrease in the volume of
growing stock has been found to affect evaporation and runoff.
Annual runoff was increased by 7 mm when the cubic volume of
growing stock in the area was decreased by 10 cu.m/ha,whereas
annual evaporation was increased by a good 10 mm when the cubic
volume of growing stock increased by 10 cu.m/ha. Human activity
in recent years has been concentrated primarily on the drainage
of forests, arable land and peatlands and on regulation of
watercourses. It has been found that the drainage of arable land
increases minimum runoff. Bog and forest drainage has been found
to increase the mean runoff, and particular note has been made
of the relatively high increase in minimum runoff. According
to investigations carried out in southeastern Finland, the
spring maximum runoff was increased on average by 30 % and the
summer maximum runoff by 130 % as a result of forest
452   R.Lemmela   S   Y.Sucksdorff

Kiisteri, T. and Tuhkanen, T., 1982, Digital Map Data Base and
   Application Programs Developed at the National Board of
   Survey in Finland: XI Cartographic Conference, ICA, Warsaw,
   p. 1-18.
Kuittinen, R. and Kauppi, L., 1981, Vesistôalueiden fysiogra-
   fisten tekijoiden inventointiin sopivat tiedostot ("Data
   bases suitable for inventoring the physiographical factors in
   river basins", in Finnish only): Vesihallituksen monistesarja,
   No. 1981:47, p. 1-24.
Lemmelâ, R. and Kuittinen, R., 1976, Effect of Climatic and
   Basin Characteristics on the Quantity and Quality of Water,
   and an Inventory of these Characteristics: Nordic Hydrological
   Conference 1976, Reykjavik August 29th - September 1st, Pre-
   prints of papers, Nordisk Hydrologisk Forening, p. 1-19.
Mustonen, S., 1965, Meteorologisten ja aluetekijôiden vaikutuk-
   sesta valuntaan, MVT 12. Helsinki. (English abstract: Effects
   of météorologie and basin characteristics on runoff.)

To top