Page 238 -

9.14 CHAPTER NINE

gravel-type media with different gradations. The filters commonly are designed in stages
and for either vertical upflow or downflow or horizontal flow.
Recent designs used gravel filter material that decreases in size with flow direction.
The gravel size range is between 0.2 and 2 in. (5 and 50 mm), and flow velocities are in
the range of 0.02 to 0.08 ft/min (0.3 to 1.5 m/h). Pardon (1991) showed that greater than
90% removal can be obtained for particles 10/zm and greater, and 72% removal of par-
ticles between 2 and 5/xm, through use of vertical roughing filters. Roughing filters are
cleaned by flushing at high rates.
Other pretreatment methods for extremely turbid waters use single-stage gravel filters
located next to or within the source (mostly rivers or canals). These filters operate simi-
larly to roughing filters. Cleaning is normally accomplished manually because most of
the removal is at the top surface of the filter.

Preozonation. Concern with disinfection by-products in finished water has increased the
need to improve organic precursor removals through the treatment process. Research has
been conducted on ozone as a preoxidant ahead of slow sand filtration as a means of im-
proving organics removal.
In general, ozone use is greater in Europe than in the United States, and many plants
in Europe have ozone preoxidation and GAC adsorption before slow sand filtration. Mal-
ley et al. (1991) reviewed past research and performed pilot studies to evaluate the effect
on treatment performance of preozonation before slow sand filtration. Ozonation converts
nonbiodegradable organic matter to biodegradable forms, to enhance biological activation
of the filter media.
In this ozone treatment scheme, reductions occurred in ultraviolet (UV) absorbance
and trihalomethane formation potential (THMFP). It was also observed that ozone en-
hanced conditions in the filter water columns for removal of other objectionable matter.
However, the breakdown of organic matter also reduced filter run times. The studies
showed that ozonation by-products were removed through the biological slow sand
process.
Using ozone ahead of slow sand filtration may allow many communities to meet new
disinfection by-product regulations. However, treatment improvements may result in in-
creased operating and maintenance costs relating to shorter filter run times and operating
costs of the ozone system.

Granular Activated Carbon. The addition of granular activated carbon (GAC) to slow
sand filter media was initially tested in England at the Thames Water Utilities. Thames
Water currently operates seven slow sand treatment facilities with a combined capacity
of about 700 mgd (2,500 ML per day). Because of pesticide levels in the source waters
and strict regulations for pesticide removals, the utility determined that adding a GAC
treatment step could allow water quality goals to be met.
To avoid the relatively high cost of constructing GAC adsorbers, the utility explored
installing GAC within the filter bed. The "sandwich" bed used a 3- to 8-in. (75- to
200-mm) layer of GAC installed 4 to 6 in. (100 to 150 mm) below the sand surface. The
performance of the system was compared with that of conventional slow sand filtration
with respect to head loss, color removal, TOC removal, and THMFP. The results are pre-
sented in Table 9.4. Chlorine demand was also reduced with the use of GAC, and pesti-
cide levels were reduced to below standards.
Based on test results, Thames Water installed the GAC sandwich in several of its treat-
ment facilities.
Biologically activated carbon (BAC) with use of both preoxidation with ozone or other
oxidants and biological activation of the sandwich layer will enhance reduction of both