CGRG Bibliography of Canadian Geomorphology
Search Results
Author : Desbarats, A.; Logan, C.; Hinton, M.; and Sharpe, D.
Date : 2004.
Title : Geostatistical mapping of specific capacity in the Oak Ridges Moraine aquifer.
Publication : Geo-Engineering for the Society and its Environment. 57th Canadian Geotechnical Conference and the 5th joint CGS-IAH Conference. Hilton Quebec Hotel, Old Quebec. October 24-27, 2004.
Issue :
Page(s) :
Abstract
Most regional aquifer studies are hampered by a lack of data with which to characterize spatial variations in hydraulic conductivity. However, specific capacity measurements from public water well records represent a vast yet under-utilized source of surrogate conductivity information. One reason these data have been neglected is the general perception that they are of poor quality. Indeed, short-duration pump testing, inaccurate recording of results in the field and transcription errors in the public water well records may introduce noise in specific capacity data sets. Nonetheless, the objective of this paper is to demonstrate, using a geostatistical approach, that specific capacity data are not overwhelmed by random errors and can, in fact, be used to produce meaningful maps of aquifer productivity. To illustrate the approach, the study focuses on a portion of the Oak Ridges Moraine aquifer, in the area of Goodwood, Ontario. The deposits of the Oak Ridges Moraine (ORM) form an upper phreatic aquifer system confined locally by the Halton Till and underlain by the regionally-extensive Newmarket Till aquitard. With an average total thickness of about 50m, these deposits consist of rhythmically interbedded fine sands, silts and locally important coarse layers. Because of the sandy nature of the sediments and their elevated topography, the ORM is an important recharge area. Available well data for this study consist of 1335 specific capacity measurements following a highly skewed distribution with a median of 5.1m3/day.m and a coefficient of variation of 1.58. Before modeling these data, it is important to recognize that they are the product of a censoring process: sand layers that are too thin or insufficiently permeable are drilled through until a first “pay” sand is encountered. Geostatistical modeling is performed using a quasi-3D approach with screen elevation coded into a sequence of indicator variables as “0” for cased layers, “1” for screened layers and “missing value” for layers beneath the bottom of the well. Specific capacities are modeled using a service variable equal to the product of the layer indicator variable and the specific capacity of the screened interval. Variogram analysis in the areal plane, for each layer, reveals well-defined spatial correlation structures with ranges over 1 km and principal directions corresponding to known depositional patterns. This in itself shows that specific capacity data are not dominated by spatially uncorrelated random noise. By kriging the indicator and service variables layer by layer, one obtains vertical profiles of the probability of a “pay” sand in each layer and its corresponding specific capacity. From these profiles, it is possible to construct a map of depth to the most probable “pay” sand and the corresponding map of specific capacity. Such maps represent invaluable tools in the prospecting for regional groundwater resources.
Bibliography of Canadian Geomorphology