GIS for Aquifer Monitoring and Modeling - From Field Surveys to Simulation Models: A case study of Kaluvelly Pondicherry basin, South India

Pierre Jorcin
GIS Analyst
Water Hrvest
pierrejorcin@hotmail.com

BACKGROUND: In the Kaluvelly Pondicherry basin of Southern India, groundwater resources are now threatened by a fast and heavy depletion. As there are no perennial rivers in the area, the irrigation tanks, fed by the monsoons, are not sufficient to cover the agricultural needs for more than a single crop. This has given rise to the overexploitation of the coastal aquifers. The water quality is being badly affected by salinization leaving people without proper drinking water supply in many places. Moreover, because of high permeability, a fast intrusion of seawater through the fresh-saline water interface is likely to happen and risks needs to be evaluated.

However, the geological and morphological context indicates potential high capacity zones where extraction could be concentrated. The primary reports available highlighted the need for a better understanding of the aquifer system and its dynamics. In a context of future growth and sustainable development, a groundwater modelling study is required to support integrated water management plans with all stakeholders.

The project presented here aims to study the quantification of the groundwater resource of a multi-layered aquifer system for a predictive evaluation with a GIS model. The hydrogeological study will target an identification of best favourable areas for exploitation.

In building such a model, there are 3 main steps –

Building a geographic database from intensive field surveys
Modeling geology by tracing contour maps and cross sections
Creating a 3D model and calculate the aquifer capacity

Building a geographic database from intensive field surveys

The first step is to set up a GIS database where specific parameters can be entered and analyzed. To answer the conditions of the project, highly accurate data need to be collected and processed, such as groundwater table and its fluctuations, water extraction, natural recharge potential, etc… Therefore, intensive field surveys will be carried out to collect first hand calibrated data, in order to provide reliable data following the scientific requirements. This method allows a direct control on the data and a validation of up to date information. On the other hand, it demands an important investment in terms of time and human resource. However, the quality of the GIS built from such field surveys gives lots of prospect for developing research projects and activities related to water management.

The hydro geological model will be built from the GIS database created by Harvest GIS Unit during previous related projects. A census of data availability and accuracy has been carried out, identifying data needed for building the model. Base maps and thematic maps featuring hydrological parameters are already part of the GIS. The drainage network system has been mapped from various data source, such as cadastral maps, engineering reports, and GPS field surveys. The lakes and irrigation tanks with their feeder and surplus outlet channel were marked in detail. Therefore, the run-off pattern and watershed limits can help to identify the most suitable areas for rain water harvesting. Regarding groundwater data, Harvest for several years has been conducting regular monitoring programs, collecting monthly data on groundwater quality, especially salinity rate. Groundwater level is being measured from selected observation bore wells, according to specific parameters such as the aquifer tapped, the volume of water extracted in the immediate surrounding area, etc…Water level contour maps are generated in GIS to study the fluctuation of the groundwater table and its current evolution in time and season wise, together with studies of the rainfall pattern. In Pondicherry Kaluvelly basin, in the main deeper aquifer, groundwater table is found to be 50 meters below the mean sea level, at 15 kilometers from the sea shore. Exact altitude is an essential parameter to trace the piezometric level of the aquifer. For this reason, the observation wells have been located using a differential GPS with centimeter accuracy in longitude, latitude and altitude. (FIGURE 1 )

For water balance evaluation, an intensive field survey has been carried out for a detailed census of more than 6000 bore wells over an area of 250 sq.km, with a handle GPS of 15 meters accuracy. All technical characteristics have been entered in GIS as attribute data, targeting mainly an estimate of the water extracted daily by pumping.

The ground surface elevation is known from an existing Digital Elevation Model (DEM) built up thanks to a grid of altitude benchmark data collected from various survey programs using a differential GPS, with centimeters accuracy in X, Y and Z. The DEM will be the reference for the realization of the geological model, to measure the absolute depth of the aquifers in reference to mean sea level.

A lot of investment has been put into these monitoring programs. By collecting data directly from the field, the GIS can be designed and calibrated from the beginning of its creation. Data integration is planned according to the need of the geographic analyses, and made to answer only requested criteria. To minimize the investments put in surveys, each campaign is planned at multiple scales, with various levels of precision. Advantage of intensive field surveys is in getting first hand calibrated data ready for specific processing. Accurate data can help to make rapid scientific assessments on various themes, whereas the same study would take more time if done through several steps using second hand information. For a local organization like Harvest, working to implement research and development projects, it helps a lot to empower the structure by building its capacity, and not relying on other sources to validate important information.

Modeling geology by contours mapping and cross sections generation

Pondicherry Kaluvelly basin geology consists of a crystalline bedrock overlaid by a sedimentary series of several layers from Cretaceous and Tertiary, with alluvium from Quaternary. The hydro geological context is a complex multiple layers aquifer system, with an altercation of sandstone, limestone and clay.

The objective is to get a precise knowledge of the aquifer capacity based on the knowledge of the geological formations characteristics, such as their depth, thickness and volume. For this purpose, a 2 dimensions model can be built within the GIS to draw the aquifers geometric structure.

A surface geological map of the project area was available from the Geological Survey of India at 1: 250 000 scale, published in 1984. This map has been checked on the ground to verify its accuracy at the scale of our study, and eventually redefine the limits of the formation outcrop areas. The map has been found to be of excellent quality, with all boundaries matching the GPS points taken on the ground where the soil configuration changes. Because of erosion factor, limits have been slightly modified and retraced in two places. Detailed lithological data available from drilling reports are compiled together to build a database for the whole studied area. Among a wide range of data gathered for the whole sedimentary basin, a sampling set has been selected based on their precision and validity. Data source are various, mainly the Central Ground Water Board deep observation wells, the Tamil Nadu Water and Drainage Board, the Public Work Departments, and the Pondicherry Agro Service and Industries Corporation, all re-interpreted for the sake of the project. 280 logs were located using a 15 meters resolution handle GPS, over an area of 700 sq.km, with a average density of 5 samples per 10 sq.km grid, a maximum density of 35 samples per 10 sq.km grid, and a maximum gap of 5 km between two samples for the few areas of lower density sampling. (FIGURE 2)

The first step is to enter data in the GIS, and process the data to obtain a uniform classification of the formations stratigraphy from the detailed soil and lithological samples. For each and every formation, the depth of the top and bottom are defined and this will be the input value for the model. Values are converted in absolute values in reference to mean sea level, free from the ground surface elevation fluctuations. This is done by importing values from the DEM, as locations of the wells have been taken by GPS. The model will cover information on the geology from the ground surface down to the rock basement at 500 meters depth.

The methodology follows a step by step process, with a validation of results at every stage.

The model consists of surface contour maps of the probable bottom of every geological formation. The maps are traced using interpolation method combined with cross sections drawing and analyses. The methodology applied here makes use of computer processing techniques with GIS tools and standard scientific methods for geology layer mapping. Primary surface contour maps are generated from the set of sampling data and are corrected step by step by tracing cross sections and analyzing their characteristics. Results of the contour maps are checked following a spatial panel distribution and re-modified step by step to fit with the actual geological regional pattern. GIS is the combination of tools, hardware, software and human resources mobilized to process thematic layer data, and can not be limited to automatic computer processing. Cross sections are generated automatically and then re-traced according to the global understanding of the geology. These results are then re-entered in the GIS to create contour maps entirely calibrated by the user. To achieve this task, a new surface contour map is generated using calibrated new samples points at selected location with the required attribute data. Additional samples points are added in order to modify the shape of the surface model based on our interpretation and scientific knowledge. (FIGURES 3 and 4)

The method chosen for contour mapping is the Natural Neighbour Interpolation Technique. The Natural Neighbour method is a technique based on geometric parameters that uses area-weighting regions generated around each point of the sample data set to determine new values for every grid node. A network of centroid is created from the input data set by connecting edges from adjacent points. From this generated area for each data point, new values are assigned to the overlying grid cells. The new grid value is calculated as the average of the surrounding point values proportionally weighted according to the intersecting area of each point. This method is very effective for interpolation of geographic data showing a clustered spatial distribution. It gives the best results for our project according to the input data type and the results expected from the model. This interpolation technique is particularly appropriate to the methodology followed here, where results are re-calibrated by the user. As it is based on geometric parameters, it allows the user to directly give orientation to the results by working on the input data points.

The output grid cell size is defined according to data entry resolution and results expected from the study. A resolution of 100 meters is selected, as it is the highest resolution of the data spatial distribution. A high resolution will allow the transfer of a 2D model to a 3D model without loosing accuracy in the process. The vertical resolution is in centimeter range thanks to the GIS interpolation tools and precise original set of points. Results of decimeter accuracy are expected for the hydrological study. Accuracy is higher in areas where the density of samples is more important, mainly for the upper layers of the aquifer system. For the deeper formations, less data were available, as few very deep bore wells have been drilled. However, for the deeper formation, according to the depth and thickness of the formation, a larger margin error can be tolerated. The top of the upper layers are generated from the outcropping spread areas marked on the surface geological map, together with altitude data imported from the Digital Elevation Model. The whole sedimentary series has a 2 to 5% slope directed to the sea, following a North-West to South-East inclination. The sedimentary beds dip and thicken towards the sea, and towards the south.

3D Modeling and aquifer capacity calculation

The results of the surface contour mapping of the bottom of the geological formations are the base used to build the 3 dimensions model. The 2D layers created in GIS can be converted to 3D layers. The result will be 3D layers of the formations showing their geometry and structure. The extent, the depth, thickness, and spatial variations of the geological layers can then be analyzed and studied.

The model is built by exporting data from the contour layers, at a predefined resolution. The 3D model is created by using a triangulation method from a regular and dense grid of points. The accuracy of the triangulated features relies on the grid spatial resolution, as well as on the precision of the elevation attribute data. It is here needed to mention that for computer processing, higher resolution grids require a good memory capacity to run the application, and limited systems may face difficulties to apply such parameters in their GIS. Resolution can therefore be re-defined following this constraint.

Over the whole Pondicherry basin, a particular area corresponding to the first aquifer of the region is targeted for a capacity evaluation. This top aquifer with the largest natural recharge area, namely Cuddalore sandstone, is studied more in depth to estimate its capacity. The 3D model allows direct volume calculation for a well defined area, using values of the top and bottom depths of the aquifer. In this case, the aquifer top corresponds to the ground surface as it is outcropping on a large extent. The same work is then overtaken to calculate the volume of the water saturated part of the aquifer, according to the water table measurements. This is done by entering water level values and generating the equivalent model. (FIGURE 5 and 6)

For a targeted area of 135 Sq.km, the volume of the Cuddalore sandstone formation is about 8,425 Million Cubic meters. With an average water table about 20 meters below mean sea level, the volume of this aquifer is about 5,550 Million Cubic meters. Finally, porosity parameters are applied to estimate the actual groundwater potential for the same selected area. For a volume of 5,550 Million Cubic meters, the aquifer capacity is about 830 Million Cubic meters for a porosity index of 15 %, and 1110 Mcum for a porosity index of 20%. These results can then be analyzed together with the groundwater extraction reports generated from the GIS database.

Conclusion

In this project, all thematic quantitative and qualitative data collected are part of an integrated database structured in a GIS. The administration of such a hydro geological database is an advanced and suitable tool for identifying existing problems, and searching adequate solutions for a better integrated water management. The result is an effective tool for groundwater analyses, with the possibility to generate scaled maps and cross sections for any particular studied area. By locating a point by GPS, the model is able to give a report on the geological patterns for this location up to 500 meters depth.

The resulting scientific assessments are for an important part based on the accuracy of the data processed. For a project at this particular scale, a huge amount of detailed data has been gathered and processed. This project has been able to set up an interactive database, which can be easily updated regularly, creating a direct link from the ground situation to the planning stage. The strength of the GIS is also its flexibility. The methodology followed here made use of combined automatic computer processing together with standards scientific methods for geological mapping. For the whole processing, from the field surveys to the modeling, different GIS packages have been used, according to their capacity to overtake every specific task. The GIS user has been mastering the complete processing work, from planning the data collection on the field and designing the database, to the spatial analyses. The next step will be to go further in hydro geological modeling by transferring GIS data to other professional specialized platform for advanced simulations.

One can see, through the results of the project, that the investments put by local agencies to collect data and process it themselves, allows the objectives to be realized, and also helps to build up a solid basis for long term GIS investigations and management. This process involving local agencies, and local communities to a minor extent, helps them to build up their capacities and knowledge and prepare valuable reports for decision making and planning. Knowledge and awareness will act as a key of empowerment and own governance for the different stake holders. Thanks to the emphasis put on research and development, the Pondicherry Kaluvelly basin is part of the UNESCO Hydrology for Environment Life and Policy (HELP) Basin Program.

REFERENCES:

Vincent A., Violette S., Marsilly G. DE, Boulicot G., Jorcin P., Krishnamurthy R.S., Sivasubramaniam K. (2006) “Infiltration quantification at the watershed scale using hydrological models, in a semi-arid context (Kaluvelly-Pondicherry coastal sedimentary basin, Pondicherry territories and Tamil Nadu state, India)”, European Geosciences Union General Assembly, Vienna, comm.
Vincent A., Violette S., Krishnamurthy R.S., Sivasubramaniam K., Jorcin P., Boulicot G. (2004), “Preliminary report on the hydro geological study of the Kaluvelly-Pondicherry coastal sedimentary basin (Pondicherry territories and Tamil Nadu state, India) HELP – Project”, In Progress Report program UNESCO HELP, 31 p.
Gassama N., Violette S., D’Ozouville N., Dia A., Jendrzejewski N. (2003), “Multiple origin of water salinization in a coastal aquifer in the bay of Bengal”, WOM, New Delhi, comm.
Gassama N., Violette S., D’Ozouville N., Dia A. (2002), “ Multiple origin of salinization in a coastal aquifer, South India, geochemical point of view”, Geochemica and Cosmochemica Acta 66(15A), pp. 265.
Violette S., Gassama N., Jendrzejewski N. (2001), “Coastal aquifer in the bay of Bengal – Is water salinization due to sea water intrusion or does it have multiple origin?” Proceeding of 1st international conference on salt water intrusion and coastal aquifers, Essouira, Morocco, 4p.