On refinement of Water Resources Mapping and development of Flood Forecast Method, using Remote Sensing and Ground Data

On refinement of Water Resources Mapping and development of Flood Forecast Method, using Remote Sensing and Ground Data

Sumiya. G
National Remote Sensing Centre, State Committee for Nature and Environmental Protection, Ulaanbaatar, Mongolia

Abstract
In this paper, the primary results obtained in Mongolia on the application of Remote Sensing Data to Surface Water Hydrology are Introduced. The applicable visual interpretation methods of the Remote Sensing Images for water resources and snow cover mapping are also outlined. The Spacecraft Photographs are useful for Water Resources Mapping particularly, potential ground water assessment and the Satellite Multispectral Image for Operational Hydrology i.e., an assessment of whole watershed characteristics at real-time for purposes of monitoring snow cover dynamics and flood forecasting. The forecast method presented here is applicable to snow-melt runoff forecast of a mountain river the regime of which is predominantly conditioned by snow-fed as the case of the mountain watersheds is western part of Monglia.

I. Introduction
In recent years, Remote Sensing Data is being employed more increasingly as one possible answer to hydrologic complex problems in many countries of the world [2,3,4,6,9,11]. The Remote Sensing Data can provide the comprehensive hydrologic Information of rivers and lakes of a concrete region for short time and desired frequency and greatly improved space and time coverage. Particularly, a quite new and comprehensive information on water resources and hydrologic regime of the rivers and the lakes can be obtained from the remote ungaged river catchments located in the Mountain and the Gobi Desert Regions.

Possibility of Remote Sensing Data use in Surface Water Hydrology depends on the required resolution and information frequencies of the existing Space Technology. We started use of detailed quantitative assessments of the natural objects, phenomenons and process for investigation and analysis when the up-to-date Interactive Image Processing System with high capacity computer facilities was introduced in 1985. The studies of end factors and exofactors of the water bodies dynamics in different physiographic and climate conditions in relation with their spectral reflectance characteristics are being carried out with help of the Digital Data Processing System. Consequently, we are able to obtain digital information and compare the derived results of the observations made at the test sites by ground equipment and from aeroplance and space flat forms. These give hydrologists an opportunity to estimate and predict further changes of the water bodies behaviour.

II. Data and Methodology
For water resources assessment were used the remote sensing data in conjunction with the ground observation data of the Hydro meteorological Stations and available information, published reports and field survey data [5,9]. For the water resources mapping the various Remote Sensing Data were used to obtain direct and indirect characteristics on water resources and hydrologic regime for refining the previously compiled maps and field survey data. In other words they were used for obtaining the qualitative assessments.

Multiband images at scale of 1:250,000 to 1:1000,000 in visible and near infrared regions, obtained from the spaceraft “Soyuz”, the space laboratory “Salyut” and Meteorological Satellite “Meteor” were used for water Resources Thematic Mapping. The above image data were used as a photo base map and drainage pattern, structure, density and drainage boundaries in the run-off dispersion area, lake shorelines and locations of the present glaciers and snow pack were defined and refined on this photo base. As well as, run-off dispersion area, specific area for the ground water formation, boundaries of the Artesian Basins, ground water recharge and discharge areas were determined on the images. As is known, the water bodies are identified by their definite tone contrast with the surrounding objects on bands of 0.8 – 1.1 mm images where they appear relatively darker. Due to impossibility of the ground water availability to be identified directly on the space-acquired inmates, we used a Landscape Type Scheme developed by appropriate geo-morphological elements in relation with the different litho logy, hydro geological formations and processes.

Thus crystalline base origination and structure of a large area in band of 0.6 – 0.7 mm, tectonic cracks and line structures in band of 0.8 -1.1 mm and Quaternary Alluvial deposit in band of 0.5 – 0.6 mm were clearly delineated. In determination of the large Artesian Basins and Hydrogeological formations were used regional space images with 300-1000 m spatial and spectral resolutions at nadir. The interpretation of the images was resolutions at nadir. The interpretation of the images was resolutions at nadir. The interpretation of the images was carried out visually according to the established indicators or clues; shapes, pattern, tones, texture and structure of the clues; shapes, pattern, tones, texture and structure of the natural objects [5,6]. The Remote Sensing Data was selected, widely depending on the natural objects to be identified and also physiographic, climate condition and season when the images were acquired.

For the study of Khovd River snow-melt run-off were used METEOR TV images with resolution of 0,2 – 1.2 km in the period of April and May, 1972-1982 and data on air temperature and precipitation of the Meteorological Stations as Yalalt (2148m), Uigil (171 om), and Knovd (1405m) located in the catechment and conventional river discharge data of the Hydrological Stations for the same period of 1972 – 1982. In addition, Hypsographic Curve of the catchment and Pyhsiographic and Landscape Maps of the study region were used as well.

The Khovd River catchement is situated in the Western Part of Mongolia and is related to a river type with spring – summer snow-melt floods and 60% of its annual run-off constitute snow-melt run-off [Fig.2]. The catchemtn area is 59,932 sq. km. The spring floods of the Khovd River usually begin in early April reaching their peaks in June-July and end up in late August.

Figure 2. Drainage network scheme of the Khovd River Basin
The established in the result of the study W = f (P,%) relationship was analysed by the ground records on precipitation, air temperature, river run-off and snow cover condition at the Hydro meteorological Stations located in the Khovd River catchment.

In selecting images consideration should be given to cloud condition, sun elevation and coverage of the catchments area at nadir. So the frequencies of the could free days in the Altai Mountain Region estimated by the Satellite Data during the period from April to May were 42 – 52 percent. The Satellite TV Image in visible band of 0.5 – 0.6mm and near infrared band of 0.7 – 1.1 mm were particularly suited to investigation and analysis of the snow-melt dynamics. The geographical fastening to the Topographic Base Map performed by the relief features e.g., mountain ranges, depressions, lakes and drainage network, which were easily and mapping of the snow cover were carried out visually on the basis of the specially developed interpretation indicators [5,6]. Because of impossibility of detecting absolute values of he spectral reflectance characteristics of the ground objects on the images by visual interpretation we used the relative interpretation indicators such as photo tone, texture, structure and image specifics [Fig. 4]. Measurement of the snow-covered area was made by using the ordinary methods i.e. plainer or special pattern (grid).

Figure 4. Scheme of snow covered area in Khovd River catchment, METEOR - 28 Satellite image aquired on 26 April, 1978
III. Results
Mangolian researchers-hydrologists created the Thematic Map of both the surface and the ground water resources of Mongolia at scale 1: 1000,000 [Fig. 1] using the Remote Sensing and Ground Data. Furthermore, in the course of interpretation of Remote Sensing Data the target areas for ground water prospection were identified, outlining the places where the ground water depth is close to the surface. On the Meteor Satellite Images the regional lineaments of the Gobi Altai and the Khangai Mountain Ranges stretched to the western direction could clearly detected and was revealed that the lineament – dominated areas located along the latitudinal direction were particularly favourable sites for higher yield wells and springs.

Figure 1. Surface water resources distribution scheme of Mongolia (mm) from the original map at scale 1:10000000 refined by Remote Sensing Data
And also were conducted researches on the classification of river channel types in different conditions of geological structure, vegetation and soil covers and landscape and an attempt by other scientists to estimate water potentials under sand masses was made. As result, they have established that there in the saline lands in the Gobi Deseart and Steppe Regions could be collected 60-80 thousand cubic metre of water and 10-12 thousand cubic metre of water under sand masses of the Great Lakes Hollow.

The river over floods caused by heavy rainfall and intensive snow-melt are considered as a natural disaster in Mongolia. Therefore we have tried to develop the method of forecasting the spring flood ( snow-melt ) run-off volume of the high mountain glaciered and snow-fed rivers for the flood control and the economic benefits in water supply, water energy and irrigation to enhance the agricultural output in the study area. On the basis of the experiences and obtained results on the development and application of the methodology for snow-melt flood forecasting of the snow-fed mountain rivers, achieved in other countries of the same natural conditions [ 5,10,11] was made an attempt to establish a forecast relationship between snow cover percentage of the catchment ( P, % ) at fixed moment of time during the snow-melt period and the flood run-of volume ( w, m3) for whole flood duration [ Fig. 3].

Figure 3. Forecast relationship between drainage snow cover percentage and run-off volume at the Bayannuur gauging site of Khovd River
The analysis of the relationship show that the values of the snow-covered areas of a drainage basin, estimated by the Satellite Images may closely related to the snow-melt flood run-off volumes in the case if there is no substantial precipitation during the snow-melt period in the drainage basin until July, Thus we could estimate the spring flood run-off volume with accuracy of ± 5 – 25%, verifying the established by the study forecast relationship with an independent data on snow-melt run-off of 1983. It was proved in this way, that the spring flood run-off volume of the Khovd River could reliably be predicted in advance of 1 – 1.5 months, estimating snow-covered areas of the catchment on the Satellite derived image.

IV. Conclusion

The updated Water Resources Map and the proposed snow-melt runoff prediction method are being widely used in Water Resources Development and Management of the Decision Making, water Monitoring or Planning Organizations.
As result of the study, has been established the empirical relationship which is applicable to the forecasting the snow-melt flood run-off of the Khovd River, estimating the snow-covered areas at fixed dates on the TV Images received from the Meteorological Satellites.
Inspite of unsufficient satellite images suitable for analysis, an enough close relationship was established by a fewer numbers ( n=6) of dots.
Further, research should be carried out to improve the proposed relationship with detailed processing the ground observations on snow water equivalent, snow depth, air temperature and precipitation during the snow-melt period to get more numbers of dots.

References

Adyasuren TS., Badarch M. Osobennosti identichnosti nekotorykh prirodnykh obiektov na kosmicheskikh snimkakh. – Trudy Instituta Meteorologil 1 Gidrologii, 1983, no. 9, s-110 – 116.
Chernov, V. YU Kulinikhan, G.V.K. metodike polucheniya fotokart s izobrajeniem granits snejnogo pokrova v gorakh po snimkam s meteorologicheskikh ISZ. – Trudy SANII, 1983, vyp. 85/166/, s. 50-54.
Kalinin, G.P. Kurilov, YU V., Kolosov, P.A. Kosmicheskie metody v gidrologii, 1977, 182s.
Kupriyanov, V.V. , Prokacheva, V.G. Sptnikovaya informatsiya 1 izuchenie vod sushi. – Trudy GGI, 1976, vyp. 238, 144s.
Kurlilova, Yu. V., Poplavskaya, L.K. Deshifrovachnye priznaki meteorologicheskikh sputnikov. – Trudy IV Vsesoyuznogo Gidrologicheskogo Siezda, 1976, Tom 2, s. 412-419.
6. Kurilova, yu. V. Issledovanie snejanogo pokrova po sputnikovym dannym. – vodnye resursy, 1975, No2, s. 50-60.
Prokacheva, V.G., Chmutova, N.P. O prognozirovanil obyoma vesennego polovodya s pomoshiyu sputnikovol informatsii. Issledovanie zemli iz kosmosa, 1982, no2, s. 78-82.
Sadov. A.V., Revzon, A.L. Aerokosmicheskie metody v gidrologii 1 injenerno1 geologii. “Nedra”, 1979, 204 s.
Sumiya, G. Kartografirovanie vodnykh resursov Mongolii s ispolzovaniem kosmicheskoi informatsii. – Trudy Instituta Meteorologii 1 Gidrologii. 1983, No8, s. 134-141.
Sumiya, G., Adyasuren, Ts. O prognozirovanii obyoma stoka vesennego polovodya r. khovd s ispolzovaniem sputnikovol informatsii. – Trudy institute Meteorologii 1 Gidrologii, 1986, vyp, III, s. 183-189.
Tsarev, B.K. Otsenka klimaticheskikh sezonnol snegovol granitsy v Ferganskol doline po dannym sputnikovol informats11. – Trudy SANII, 1983, vyp, 85/166/, s 54-65.