Environmental Engineering Reference
In-Depth Information
major pollution sources
[8]
, but this is not practical in concrete-lined channels,
designed to minimize sediment accumulation.
We developed and trialed an alternative passive sampling method based on
solid-media. The method uses polythene mesh bags of granulated activated carbon
(GAC) media to absorb pollutants present in industrial runoff. We applied this
method to determine the effectiveness of the City of Kingston's E&E program in
reducing the amount of pollution discharged from industries into receiving
waters.
Method
This study was conducted at in industrial areas within the City of Kingston, 25 km
south-east of Melbourne, Australia. Sub-catchments with clearly defined drainage
to a single outfall were initially selected from a drainage plan. Sub-catchments
containing land zoned residential, traversed by major roads, or with significant
areas currently under construction were excluded. Visual inspection excluded a
further two drains which were either submerged or could not be accessed safely,
leaving 20 sub-catchments for assessment, each draining to a single outfall.
Polyethylene mesh bags measuring 400 mm × 300 mm edged with 40 mm hook-
and-loop fastening strips were filled with 500 mL of granular activated carbon
(GAC). Polyvinyl chloride (PVC) mounting plates fitted with 40 mm hook-and-
loop strips were fastened to drain surfaces with construction adhesive. Bags of
media were then secured to PVC mounting plates with hook-and-loop strips either
within or immediately below each drain. After exposure for 1 week (including five
business days), bags were sampled and replaced with fresh media. Concentrations
of accumulated hydrocarbons (C6-C36) and heavy metals (As, Cd, Cr, Cu, Fe, Ni,
Pb, Zn) were determined by Ecowise Environmental (Scoresby, Victoria) according
to methods WSL-030 and WSL-032 respectively.
Daily rainfall records were provided by Melbourne Water for the Clayton
Retarding Basin, Braeside Park, and Mentone gauges. Weekly rainfall was esti-
mated as mean 24 h totals for the three gauges, summed for the week preceding
collection. Weekly rain frequency was calculated as the sum of days on which mean
rainfall exceeded 0.2 mm.
Where concentrations were below detection limits (DL), values were imputed by
Helsel's robust method
[5]
, using UnCensor v5.1
[6]
. Data were log
10
transformed
prior to parametric statistical tests. After three sampling collections, principal com-
ponent analysis was conducted on heavy metal and TPH media concentrations to
identify redundant parameters. Principal components were extracted from a corre-
lation matrix, retained where eigenvalues were over 1. Parameters highly correlated
with Cu (Fe, Ni, Cr) or consistently below detection limits (As, Cd, Pb) were not
determined after the third week.
Sub-catchments were paired by two-step cluster analysis of catchment charac-
teristics and pollutant concentrations collected during the first sampling stage only.
