GISdevelopment.net ---> AARS ---> ACRS 1990 ---> Geology Disaster

Modelling Landcover and rainfall data for estimating ground water table fluctuations

Om Prakash Dubey
Civil Engineering Department University of Roorkee,
Roorkee - 247 667, India


Abstract
The landcover patterns are governed by environmental parameters e.g. soil, topography, climate etc. Rainfall is the major parameter influencing the ground water recharge. A linear model utilizing the land cover and rainfall has been developed to yield the ground water storage fluctuation, which can be utilized to estimate the ground water table fluctuations. The land cover data was obtained from the integrated analysis of the remote sensed data (landsat F.C.C. blow up at 1:250,000 scale, TM F.C.C. blow up at 1:250,000 scale) and field survey. The model was calibrated from 6 years data and then fluctuation was predicted and compared with the observed data.

Background
As a result of external and internal forces e.g. Rainfall, pumping etc. the ground water table fluctuates. The ground water table fluctuations are essential information for planning various developmental activities e.g. drinking water and Agriculture etc. Conventional method of collecting fluctuation data is difficult, costly and may not be representative. Remote Sensing techniques an alternate data set containing a wealth of information at one place regarding surface characteristics, subsurface indicators and topography etc. The rainfall recharges ground water (Chon 1964). The rainfall recharges the ground water depending upon the surface and subsurface characteristics. The land cover is influenced by surface and subsurface characteristics and is a good indicator (Dubey 1989). The land cover data can be conveniently evaluated from remote sensing but in the present study rainfall data has been collected from the meteorological department.

The Model
Theland cover operator model (Dubey et al. 1984) may be written as

AIR = GS --------------------(1)

where G is the land cover operator system matrix, I is the depth of rainfall vector, R is the recharge coefficient vector, and GS is the change in the ground water storage in a particular time. The above equation can be simplified as

AR = G --------------------(2)

where G is the change in ground water storage per unit rainfall

Paramenter Evaluation
Land cover Data
The land cover data has been evaluated from the visual analysis of landsat F.C.C. at 1:250,000 scale described below.
  1. From the survey of India topographical map at 1:250,000 scale, a base map is prepared.

  2. Priliminary analysis of F.C.C. at 1:250,000 scale on the basis of image charcteristics and development of the interpreation keys.

  3. Anaysis of T.M. F.C.C at 1:50,000 scale in selected areas in the same manner as in 2 above.

  4. identify the doubtful areas for field experimentation.

  5. Field traverses in the doubtful area and collecting the ancillary data.

  6. Finalization of the land cover map.

  7. Preparation of a land information system consisting of Rainfall and land cover (Cell size 5mm x 5mm)
Model Calibration
The model was calibrated in a part of Indogangetic plain of about 200 Sq. Km. bounded between Yamuna and Hindon rivers. Using data from 1973-1978 the model (equation 2) was calibrated, and vector R has been evaluated by (O.P. Dubey et al. 1988):
  1. Conventional Inversion

    R = A-1 G ----------------------(3)

  2. GLI

    R = A- G = Vp A ^-1 U TP -----------------------(4)

  3. MPI

    R = A+ G = A* (A* AA*)-1 A*G ----------------------------(5)

    Where,
    A-1 = Conventional inverse of A operator
    A- = GLI of A operator
    A+ = MPI of A operator
    VP = P eigen vectors accociated with rows
    U p1 = P eigen vectors accociated with columns
    A ^-1 = Diagonal matrix P non zero eigen values of A.
Estimation of the fluctuations
Once the model was calibrated the ground water table fluctuations can be predicted using equation (2) above if land cover and rainfall is known.

Results and Discussinon
once the model calibrated the ground water table fluctuations can be predicted using this technique only.

For the year 1980 the observed and predicted values of the ground water table fluctuations for some stations has been tabulated in Table I. The results compare well.

Ground water level fluctuations
TABLE I
Location Fluctuations
(in m)
Obsered Predicated
Badshahibag
Behat
Mirijapur
Raipur
Sidhau liquadim 		9.94
2.72
4.14
4.17
3.12		 5.00
2.54
4.50
4.30
3.01

Acknowledgments
The author is thankful to IIRS, NRSA, GWIO, CBIP etc. for their help in carrying out the study, his friend for live discussions and shri Ashok K. Sharma for neat and timely typing of the manuscript

References
  1. Chow Ven te. 1964. Hand Book of Applied Hydrology. McGraw Hill Book Co.

  2. Dubey O.P.; Sri Niwas ; Awasthi A.K. ; 1984. Analysis of Remotelysensed Data For ground water studies of piedmont Zone , Proc. 5th ACRS Kathmandu , PP 6-1-E-6-4.

  3. Dubey O.P.; Sri Niwas; Parvez Ahmed; 1989 Deciphering Ground Water using Remotesensing Techniques . Proc. International conference on development of modern techniques to ground water studies; hyderbad.

  4. Dubey o.p. sri Niwas; Awasthi A.K. ; 1989 , predicting Water Table Depth Using remotelysense Data . 10th Asian conference on Remote sensing, Kullalumpur.