There are three criterion that were used to evaluate the error of the results. Mean error (ME), mean absolute error (MAE) and root mean square error (RMSE). Error estimation was computed using Equations (33)–(35) (Sadeghi Tabas et al. 2016):
33
34
35where 
, indicates the observed and simulated heads respectively while ‘m’ is the number of monthly time steps and ‘n’ is the number of piezometers. These mentioned errors are computed in Table 4.
Table 4
ME, MAE and RMSE errors
| MLPG method (m) | FDM method (MODFLOW) (m) | |
|---|---|---|
| ME | –0.08 | 0.159 |
| MAE | 0.573 | 1.434 |
| RMSE | 0.757 | 1.197 |
| MLPG method (m) | FDM method (MODFLOW) (m) | |
|---|---|---|
| ME | –0.08 | 0.159 |
| MAE | 0.573 | 1.434 |
| RMSE | 0.757 | 1.197 |