Premium
On the Strategy of Estimating Regional‐Scale Transmissivity Fields
Author(s) -
Christensen Steen
Publication year - 1997
Publication title -
groundwater
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.84
H-Index - 94
eISSN - 1745-6584
pISSN - 0017-467X
DOI - 10.1111/j.1745-6584.1997.tb00068.x
Subject(s) - kriging , hydrogeology , aquifer , groundwater , hydraulic head , geology , inversion (geology) , pressure head , scale (ratio) , soil science , hydrology (agriculture) , environmental science , statistics , mathematics , geotechnical engineering , geography , geomorphology , mechanical engineering , engineering , cartography , structural basin
A case study of a leaky fluvioglacial aquifer concentrates on methods of estimating the zonated log 10 ‐transmissivities of a regional‐scale ground‐water model covering 450 km 2 . Mainly three estimation methods are discussed: (a) kriging based on local measurements and predictions, (b) hydrologic inversion (i.e., nonlinear regression) based on head data, and (c) hydrologic inversion based on head data and on prior estimates from kriging. (a) Due to the shortage of data which is usual in heterogeneous aquifers, the present study questions the practical value of forming zonal kriging estimates from local data whenever estimates are to be used as an input to ground‐water models. In some parts of the homogeneous aquifers local data are sufficient to make such estimates, (b) In this study zonal hydrogeological parameters can be estimated by inversion based on head data. However, inaccuracies in head data may seriously damage the reliability of estimated parameters and, as a consequence, the ground‐water model, (c) Using zonal kriging estimates as prior information in the regression reduces the width of the confidence intervals of the parameters with prior information by up to 75%. The study indicates, however, that using prior information in the estimation of the hydrologic model parameters only minimally reduces the uncertainty of the predicted hydraulic heads. For this specific case, the results suggest that in order to parameterize and identify the parameters of the ground‐water model one should concentrate on qualitative mapping of the hydrogeology, on sampling accurate head data and on subsequent estimation of the zonal parameters by inversion (or manual calibration).