A geographic information system (GIS) raster technique has been developed a
nd used interactively with remediation designers to evaluate the optimum ex
tent of excavating soil contaminated by chlorinated solvents. The technique
and the results of its application are presented. The site was a former ch
emical storage plant for acids and solvents. Two distinct solvent plumes we
re detected within the ground using a photo-ionisation detector. The solven
ts were found to be dissolved in the groundwater and migrating in the gener
al direction of groundwater flow. A remediation strategy was proposed invol
ving the localised excavation of contamination 'hot spots' followed by the
implementation of a groundwater remediation system. A number of excavation
options were discussed and the GIS raster technique was developed to evalua
te these options in terms of contaminant removed and excavation cost.
The plumes were initially mapped using a triangular irregular network (TIN)
. These TIN models were rasterised to produce a regular grid of rectangular
cells, each cell having a value relating to the concentration of contamina
nt at that spatial point. The proposed excavation zones were then overlaid
on to the raster models as masks. The relationship between the value of con
taminant concentration of cells within the mask (or excavation zone) and th
e total value of contaminant concentration of cells within the solvent plum
e was used to determine the efficiency of the excavation.
The excavation options were compared taking into account the percentage of
the contaminant plume removed, the excavation area (soil volumes) and relat
ed costs. Once the GIS raster technique had been developed, it proved very
quick to rerun the analysis for the other excavation zones. The optimum exc
avation zone, based upon cost and contaminant recovery, was found for the s
ite. The technique helped by targeting the worst area of contamination and
provided the client with a cost-benefit analysis of the different remediati
on options. (C) 2001 Elsevier Science B.V. All rights reserved.