Geospatial Data Analysis for Study of Suspended Sediments in Govind Ballabh Pant Reservoir, Singrauli Coalfield, India

Abstract
The Govind Ballabh Pant (GBP) reservoir covering an area of about 450 sq.km is the resultant reservoir of Rihand dam constructed in the year 1962 on Rihand River near Renukoot town in Sonebhadra district of Uttar Pradesh in India for hydropower generation. The catchment area of the reservoir is 13,385 sq.km. The availability of water in GBP reservoir and abundance of steam coal in nearby Singrauli coalfield have offered an ideal location for various large scale industries like thermal power plants, aluminium plant, chemical plant, cement and coal mining projects around the periphery of this reservoir. As sequel to industrialization in this region, not only the physical environment is being subjected to ever increasing interference, but the effect has also lead to progressive shrinkage in storage capacity and pollution of the GBP reservoir. The reservoir being the ‘Life Line’ of all the industries situated around it, a detail study on suspended sediments and its pollution level in the reservoir is very much warranted for effective reservoir management.

Comprehensive geospatial data analysis using satellite data of IRS-IC: LISS-III of the year 1998 & 2001 along with ground collateral data, has been attempted to assess the dispersal pattern of suspended sediments in GBP reservoir and sources thereto. Digital satellite data were processed using Geomatica v.8.2 digital image processing software and the ground collateral data were captured and analysed using GIS. Data of all the four bands of IRS-IC satellite were initially analyzed individually to determine the efficacy of different bands in mapping of suspended sediments. It has been observed that there is a positive functional relationship between the concentration of suspended sediments and the spectral signatures recorded in the visible wavelength bands-1 & 3 and near infrared band-4. Data of band-3, however, has a distinctive edge over other two bands. Based on the histogram distribution of spectral signatures of band-3, four concentration levels of the suspended sediments were identified and nomenclatured as viz. ‘high’, ‘high to moderate’, ‘moderate to low’ and ‘low’.

It is evident from the study of satellite data of the year-1998 that suspended sediments are mainly concentrated in western part of the reservoir, where number of thermal power plants, ash ponds as well as coal mining projects is situated. ‘High’ suspended sediment zone is present near Renusagar, Anpara, Singrauli and Vindhyachal Power Plants primarily due to discharge of the ash ponds located on the periphery of the reservoir. The Balia and Matwani streams flowing through Jayant and Dudhichua coal mining projects are also discharging its loads in western part of the reservoir. However, the discharge loads from these two streams are negligible as compared to the overflow load from the ash ponds. Eastern fringe of the reservoir is almost free from suspended sediments except near Rihand Power Plant.

IRS data of the year-2001 on the other hand reveals that besides the siltation in western part of the reservoir, the eastern fringe of the reservoir has also started getting silted showing moderate to high siltation. This is mainly because of progressive depletion of vegetation cover in and around the catchments area causing great deal of topsoil erosion resulting in discharge of suspended sediment load in the reservoir.

The clay minerals composition is an indicator of identifying the provenance of the sediments/-accumulated silt in the reservoir. Trace elements studies allow assessment of pollution level in the reservoir. An attempt has been made to integrate the available analytical data of the suspended sediments/– accumulated silt using GIS for quantitative assessment, it provenance, as well as status of pollution level. The kaolinite/illite(K/I) ratio study indicate that in western part of the reservoir sediments had high K/I ratio (3.2 to 6.8) which is indicative of Gondwana provenance of the sediments whereas in the eastern part, K/I ratio is low (1.7 to 2.22) indicative of gneissic provenance of the sediments. The chemical analysis of the reservoir sediment indicates that the metal and trace element content of the sediments is slightly higher than the average natural level of Pb. As, Se, Sn, Cd and B. This may be due to chemical composition of rocks exposed around the reservoir. Mercury concentration is also marginally higher than average natural value in almost all parts of the reservoir except in the intermediate part. However, a very high mercury pollution of the sediments was observed close to the dam where Dongia nala meets in the reservoir. The high concentration of mercury is due to discharge of the Kanoria chemical’s chloro-alkali plants in Dongia nala.

Periodic geospatial data capture and analysis offers an efficient tool for mapping of suspended sediment dispersal pattern and its provenance, pollution level and follow up remedial measures to prevent further aggravation for efficient reservoir management.

Introduction
Govind Ballabh Pant (GBP) reservoir is a resultant reservoir of Rihand dam, which covers an area of about 450 km2. The dam was constructed in the year 1962 on Rihand River for hydropower generation and is located near Renukut in the state of Uttar Pradesh in India. Availability of water in the reservoir and steam-grade coal in the nearby Singrauli coal basin have offered ideal locations adjoining the reservoir for various large scale industries like coal mining projects, thermal power plants, aluminum plant, chemical plant, cement plant, etc. About 7500 MW of electricity is being generated at present by the five thermal power plants located in the area. As a result of industrialization, the physical environment of the region has been subjected to growing interference. This has also lead to progressive shrinkage in the storage capacity of the reservoir. The reservoir being the ‘life line’ of all industrial activities in the area, a detail study of its suspended sediments and pollution level is very much necessary for efficient reservoir management.

Estimation of suspended sediments in GBP reservoir using conventional methods is expensive and time consuming. On the contrary, mapping of these sediments using multi-spectral satellite data in combination with collateral data is time-cost effective. In the present study, comprehensive geospatial data analysis was carried out using IRS-1C satellite data of the year 1998 & 2001 together with collateral data to assess the dispersal pattern of suspended sediments in GBP reservoir and its sources thereto.

Figure-1: PAN sharpened LISS-III FCC of Singrauli Coalfield based on IRS-1C data
Study Area
The GBP reservoir lies between latitudes 24O00’00” & 24O12’43’’ North and longitudes 82O38’00” & 83O00’00” East. Location map of the reservoir is given in Figure-1. The catchment of the reservoir is about 13,385 km2. The average annual inflow and gross storage capacity of the reservoir is 5138 acrefeet and 8600 acre-feet respectively. The normal full reservoir level is 264 m and average tail water level is 190 m. The regional geology of the study area comprises Precambrian metamorphic represented by quartzite, phyllite & schist and Gondwana Formation represented by shale & sandstone.

Data Source
In the present study, IRS-IC satellite data (LISS-III, Path-102, Row-55) of the year 1998 & 2001 together with collateral data such as Kaolinite/Illite (K/I) ratio and concentrations of different metals & trace elements in the sediment samples collected from the reservoir (EdF & CdF, 1992) have been used to assess the dispersal pattern of suspended sediments in GBP reservoir as well as the level and source of pollution in it.

Relationship between Suspended Sediments & Reflectance
In satellite remote sensing, the sensor records electromagnetic (EM) energy, which is reflected, scattered or emitted by the objects on the surface of the earth. Thus, reflectance is a function of the wavelength of incident energy and the physical & chemical properties of any object.

The EM energy incident on a water body is partly absorbed, partly reflected and partly scattered. Any primary signal from the water body is due to the total reflectance and back-scattered energy caused by the impurities in the water; these signals are used for water quality analysis. Further, the penetration depth of EM energy is influenced by the mineral content and/or size of the suspended sediments (Whitelock et al, 1978). It has been demonstrated that turbid water is more reflective than clear water both in the visible and near-infrared regions of the EM spectrum (Moore, 1978). Further, the measured signal is dependent on the wavelength used, the size & shape of the particles present, and their reflectance, absorption & refraction properties. A decrease in particle size results in an increase in reflective area. Hence, reflectance of water bodies can be studied to assess the dispersal pattern of suspended sediments in it.

Data Processing & Analysis
Digital satellite data were processed using Geomatica digital image processing system. Data processing was carried out to correct for geometric distortions, to calibrate the data radiometrically and to eliminate noise. Data of all the four bands were initially analyzed individually to determine the efficacy of different bands in mapping of suspended sediments. It was observed that there is a positive functional relationship between the concentration of suspended sediments and the spectral signatures recorded in band-3 (0.62 to 0.69 mm, in the Red region of EM spectrum) and band-4 (0.77 to 0.89 mm, in the near-infrared region of EM spectrum). Data of band-3, however, has a distinctive edge over the other band. Based on the histogram distribution of spectral values in bands-3&4, unsupervised classification was carried out using K-means clustering (Tou & Gonzalez, 1974). Four turbidity levels of the suspended sediments were identified namely ‘high’, ‘moderate to high’, ‘low to moderate’ and ‘low’.

Figure-2: IRS-1C (B:5,4,3) data of the year 1998 showing distribution of suspended sediments in GBP reservoir.
Analysis of IRS-1C data of the year 1998 indicates that the suspended sediments are mostly concentrated in the western part of the reservoir where a number of thermal power plants and open-pit coal mining projects are situated. The ash ponds of the thermal power plants are located along the west bank of the reservoir. Moreover, the vegetation cover along the west bank of the reservoir is generally thin due to industrialization. This has caused a great deal of topsoil erosion resulting in discharge of suspended sediment load into the reservoir during the monsoon period. ‘High’ suspended sediment zone is present near Renusagar, Anpara, Singrauli and Vindhyachal power plants (locally known as super thermal power stations or STPS) (Figure-2). This is primarily due to the discharge from the ash ponds located in the periphery of the reservoir. The Balia stream (nala), flowing through Jayant and Dudhichua coal mining projects, is also discharging its sediment load during the monsoon period into western part of the reservoir. In pre- & post- monsoon periods, the load carrying capacity of Balia nala is considerably reduced; silts are deposited along the streambed only. Eastern part of the reservoir is almost free from suspended sediments except near Rihand power plant, which is mostly due to ash pond. The discharge of sediment load by the stream is negligible as compared to the overflow load from the ash ponds. Analysis of IRS-1C data of the year 2001 reveals that in addition to siltation in the western part of the reservoir, the eastern fringe has also started getting silted showing moderate to high turbidity. This is because of the discharge from the ash pond of Rihand power plant. Moreover, there is progressive depletion of vegetation cover in and around the catchment area causing a great deal of topsoil erosion resulting in discharge of suspended sediment load into the reservoir. The central part of the reservoir, however, shows low turbidity (Figure-3).

Figure-3: IRS-1C (B:4,3,2) data of the year 2001 showing distribution of suspended sediments in GBP reservoir.
The amount and composition of clay minerals, particularly the Kaolinite/Illite (K/I) ratio, in the suspended sediments/accumulated silt of the reservoir is an indicator of its provenance. Similarly, metal and trace element studies allow assessment of the pollution level in the reservoir. In the present study, an attempt has also been made to integrate the available analytical data of the suspended sediments/accumulated silt using MapInfo GIS; this has helped in quantitative assessment of the sediment’s provenance and status of pollution level in the reservoir.

The concentration levels of Hg, Ti, V, Cr, Ni, Cu, Zn, Cd, Pb, As, Se, Sn, F, B and Al were determined for the sediment samples collected from different locations of the reservoir. Spatial distributions of concentration values (‘Al’ normalized) were mapped using GIS.

Origin of Sediments
The origin of sediments deposited in the reservoir was determined by analyzing their clay mineral contents. Earlier studies (Friedman & Sanders, 1987) have shown that the clay mineral composition of sediments is a good indicator for identifying their provenance. Clay minerals are the products of rock alterations making silicate minerals; these reflect the mineral composition of the rocks and their weathering conditions.

The spatial distribution of K/I ratio, as determined from the clay fractions of the reservoir sediments, was analyzed using GIS (Figure-4). Following four sedimentological areas were identified in the reservoir on the basis of K/I ratio:

In the north-western part, the reservoir samples had a ratio exceeding 3 (i.e. 3.2 to 4.3), which is indicative of Gondwana provenance;
In the southern and eastern part, the reservoir samples had a very low K/I ratio (i.e. 0.7 to 1.6) indicating gneissic origin of the sediments;
On the western bank, areas with low ratios (1.7 to 2.2) were also observed indicating origin from gneisses;
The mud brought by Rihand river into the reservoir had ratio between 2.4 and 3.0 (with an average of 2.9); this indicates gneissic provenance of the sediments.

Figure-4: Clay minerals from different geological substrates of GBP reservoir.
Pollution Level of Sediments
The somewhat higher concentration of metals and trace elements may be attributed to the chemical composition of the rocks exposed surrounding the reservoir. Study of mercury concentration suggests a marginally higher value in almost all the parts of the reservoir except in the central part (Figure-5). Very high mercury pollution, i.e. (>) 1000 mg.kg-1, however, was observed close to Rihand dam where Dongia stream (nala) meets with the reservoir. The high concentration of mercury is due to the waste discharge from the chloro-alkali plant of Kanoria chemicals in Dongia nala.

Figure-5: Distribution of mercury (Hg) content in the sediments of GBP reservoir.
Conclusion
Periodic geospatial data capture and analysis offers an efficient tool for mapping of suspended sediment dispersal pattern in the reservoir. K/I ratios of the reservoir sediments indicate the provenance. Chemical analyses of the sediments show the pollution level of the reservoir. Integration of geospatial data using GIS helps in analyzing the existing status of pollution level of the reservoir as well as the source of pollutants. Follow-up remedial measures should be taken-up for efficient reservoir management based on such study.

Acknowledgement
The authors are grateful to Chairman-cum-Managing Director, Central Mine Planning & Design Institute Ltd., for his encouragement and kind permission to publish the paper in the conference. The views expressed in the paper are not necessarily those of the company to which the authors belong.

References

Electricite de France International and Groupe Charbonages de France, 1992. Environmental study of Singrauli area. Unpublished report.
Friedeman, G.M. and Sanders, J.E., 1978. Principle of Sedimentology. John Wiley & Sons.
Julius T. Tou and Rafael C. Gonzalez, 1974. Pattern Recognition Principles. Addison-Wesley Publishing Co.
Moore, G.K., 1978. Satellite surveillance of physical water quality characteristics. Proceeding of the 12th International Symposium on Remote Sensing of Environment, ERIM, Ann Arbor, MI. Vol. I, pp. 445-462.
Ritchie, J., Schiebe, F.R., and Mctenry, J.R., 1976. Remote sensing of suspended sediments in surface waters. Photogramm. Engg. & Remote Sensing. Vol. 42, No. 12, pp. 1539-1545.
Whitelock, C.H., Witte, W.G., Usry, J.W., and Gurganus, E.A., 1978. Penetration depths of green wavelength in turbid water, Photogramm. Engg. & Remote Sensing. Vol. 44, No. 11, pp. 1405-1410.