The hydraulic conductivity values, boundary conditions, water levels, specific storage, specific yield, initial heads and recharge data were used for transient state calibration. The transient state calibration was done for the period from the year 2000 to 2006 (2557 days). The storage coefficient values were varied iteratively so that a reasonably good match was obtained between calculated and observed values. The hydraulic property values obtained after transient state calibration are shown in Table 1. The computed and observed water level graphs for selected observation wells after calibration is shown in Fig. 2. From the figure, it could be seen that calculated head values are on par with observed values. The graph lies in 95 per cent confidence level. This is a clear indication of proper model development with relevant water table data.
Table 1. Hydraulic properties of the layer after calibration
Sl.NO
|
Model Properties
|
Layer 1
Laterite
|
1
|
Hydraulic conductivity in longitudinal direction Kx,md-1
|
55
|
2
|
Hydraulic conductivity in lateral direction Ky, md-1
|
55
|
3
|
Hydraulic conductivity in vertical direction Kz, md-1
|
3
|
4
|
Specific storage Ss, m-1
|
0.0005
|
5
|
Specific Yield Sy , %
|
20
|
6
|
Effective porosity, %
|
35
|
7
|
Total porosity, %
|
40
|
Computed water table contour map of the study area obtained from the model is shown in Fig.3. From the figure, it could be seen that in coastal region the water table elevation was very low ranging from zero to 6 m. At the extreme west side, it is zero because of Arabian Sea as boundary. Water table in the entire study area ranges from –6.2 to 9.5 m (green to red shade). Tirurangadi—Chemmad area has highest water table (red shade).
The velocity vector of the study area is given in the Fig. 4. At the southern coastal side of the study area, the direction of flow was towards Arabian Sea. At Tanur block region the flow was north easterly. And a chance of saline water intrusion was present in the area where the Kadalundi River joins to Arabian Sea and there was a fluctuation in the flow towards Arabian Sea. The flow direction was towards east which means towards inland in Parappanangadi, Chettipadi, Ariyallur and Vallikkunnu area which indicated a chance of saline water intrusion, which was coincides with the result of water table contour map.
After calibration, it needs to be validated the model which is accomplished by testing the model with data which are not used for the calibration. The validation is done for the period from 2007 to 2009 (3653 days). Observed head Vs calculated head graph is shown in Fig. 5. From the figure, it could be seen that the values are in 95 per cent confidence interval.
Prediction was done from 2010 to 2019 (3653) Fig. 6 represent the water table contour map after increasing the pumping rate by 5, 10 and 15 per cent respectively. There is a decline in water table in the coastal stretch. Thus a hydraulic gradient is developed which will lead the movement of saline water to the coastal fresh water aquifers. Computed head Vs time graph after prediction is shown in Fig. 48 to Fig. 51. From the figure, it could be seen that water level was decreasing when the pumping rate was increased up to 15 per cent. In between the days 3653 and 4653, there indicated a sudden decrease in the elevation of water table. After 5653 days, water table elevation line becomes constant which may be because the water table reached the bed rock region and these results were in agreement with the finding of Sajeena (2015).
Contaminant Transport Modelling
Salt water intrusion study was carried out using visual MODFLOW with MT3D and this provided the observed concentration Vs calculated concentration graph, concentration Vs time graph and concentration contour map as output. Computed Vs observed concentration map for the selected four observation wells (wells of Tanur, Tirurangadi and Parappanangadi blocks) are shown in Fig. 7. It could be seen that computed and observed values are on par at 95 per cent confidence interval. Concentration contour map of the study area obtained from the model after calibration is shown in Fig. 8. From this figure, it could be seen that intrusion was present up to 3.2 km along the coastal stretch from the northern boundary of the area (Kadalundi estuary region) which extended from 0.5 to 1.2 km laterally from the coast.
Calibration values are taken in to consideration for validation of the model. The model was validated with the water quality data of the period from 2007 to 2009. Scatter plots for computed Vs observed concentration for selected observation wells are shown in Fig. 9 and it could be seen that the Root Mean Square Error (RMSE) value for almost all the wells during validation are reasonably low and are within the acceptable limits. Concentration contour map after validation is shown in Fig. 10 and it was observed that there is a chance of saline water intrusion in to the coasts of Parappanangadi and Cheeramangalam area which is extended to a distance of 1km from the point where the sea water presence after calibration.
Validated values of calculated and observed heads and concentrations are shown in Table 2 and Table 3. From these tables, it is evident that normalized RMS is less than 10. The model values are matching with observed values. Hence the model is ready for prediction.
The validated model was run for predicting the trend of saline water intrusion for next 10 years (3653 days) by assuming increase of pumping rate by 5, 10 and 15 per cent. Fig.11 showed that the predicted concentration contour map for next 10 years by increasing the pumping rate by 5, 10 and 15 per cent and it could be observed that, if the pumping rate is increased from 5 per cent to 10 per cent, there is a clear increase of saline water intrusion in the coastal stretch of the river basin. There is no change in saline water intrusion when pumping rate is increased from 10 per cent to 15 per cent which indicate that water table reached the bed rock. By 2019 vertical extent of saline water increases which ranges from 3.2 to 4.5 km along the coast from the northern boundary (Kadalundi estuary region) which extends laterally from 1 km to 1.9 km from the coast.
Table.2. Validated values of observed and calculated head
SL.NO
|
Well Name
|
Observed Head
|
Calculated Head
|
RMSE value
|
1
|
BW 189
|
38.22
|
38.24
|
0.021
|
2
|
BW 190
|
0.69
|
0.76
|
0.072
|
3
|
OW 5A
|
0.43
|
3.41
|
2.98
|
4
|
OW 6A
|
64.22
|
68.29
|
4.07
|
5
|
OW 178
|
-5.64
|
-5.31
|
0.331
|
6
|
OW 35
|
-3.70
|
-1.50
|
2.20
|
SL. NO
|
Well name
|
Observed value
|
Calculated value
|
RMSE value
|
1
|
BW 190
|
20.51
|
20.66
|
0.15
|
2
|
BW 189
|
20.66
|
20.66
|
0.001
|
3
|
OW 5A
|
20.55
|
20.66
|
0.11
|
4
|
OW 6A
|
16.25
|
16.15
|
0.97
|
Table. 3. Validated values of observed and calculated head
From all the above given results it can be concluded that there is a clear indication of saline water intrusion in the coastal stretch of the Kadalundi river basin due to drastic increase of pumping rates. The main objective of the study was to simulate the saline water flow condition in the river basin and to predict the advancement of saline water into the coast. The graph of observed values Vs calculated values remain around 95 per cent confidence level, which indicate that calculated and observed values are almost same. In the head Vs time graph and concentration Vs time graph, the calculated and observed lines remain closer. These two parameters indicate that simulation of the model was done in the right way. Concentration contour map after increasing the pumping rate by 5, 10 and 15 per cent help to predict the advancement of saline water in to the coast. Analysis of predicted contour maps indicate that by 2019 saline water will intrude into the land about 1 to 1.9 km from the coastal line.