Premium
Instability in aquifer identification: Theory and case studies
Author(s) -
Yakowitz Sidney,
Duckstein Lucien
Publication year - 1980
Publication title -
water resources research
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.863
H-Index - 217
eISSN - 1944-7973
pISSN - 0043-1397
DOI - 10.1029/wr016i006p01045
Subject(s) - aquifer , identification (biology) , set (abstract data type) , noise (video) , geology , instability , parameter identification problem , computer science , geotechnical engineering , algorithm , groundwater , mechanics , artificial intelligence , botany , physics , biology , model parameter , image (mathematics) , programming language
It is our conviction that we and other investigators have in the past not been fully cognizant of how troublesome measurement noise can be to aquifer identification methods. In the present work we set forth theoretical reasons why one may anticipate that the aquifer identification problem can be numerically unstable. This insight is underscored by numerical demonstration showing that under certain realistic assumptions concerning aquifer parameters and measurement error, transmissivity estimates provided by certain sensible identification schemes are seriously in error. By examination of the insensitivity of piezometric head to change in transmissivity we suggest that the numerical instability is fundamental to certain aquifer identification problems; it cannot be overcome by finding the ‘right’ identification method, just exacerbated by poor methods. This work provides procedures whereby other investigators can perform error and sensitivity analysis on their aquifer configurations and identification algorithms.