A Simplified Integrated Design Concept for Filters
Part 3: Design criteria differing with filter types
In
Part 3, we consider design criteria that might differ between filter types but
would be identical within each filter type. Part 2 covered design criteria that
should be consistent irrespective of the filter type.
Part
1 covered terms and terminology. As noted in Part 1, I have observed in manuals,
articles, and reports, as well as in presentations and conversations at
conferences, that the complexity of terminology itself leads to misperceptions
and confusion over expected performance and to unnecessary and inappropriate
distinctions in design procedures and criteria. This dynamic has led to
inconsistencies in design procedures, frequently within the same manual. This
can result to a bias toward particular systems, because design criteria drive
costs. It also complicates the design process for the practitioner, particularly
those in local government who review the drainage plans of the development
engineer. Presented in Part 1 was Table 1, summarizing the many names and the
widely varying design criteria. I suggested simplification; two scenarios were
offered and are repeated in Part 3 as Tables 2 and 3.
Engineers
do not necessarily realize these differences and potential conflicts because we
work within our own community, state, or province with an agreed terminology and
set of design criteria. However, as it becomes increasingly common to trade
experiences and field results across regions and borders, contradictions and
miscommunication are becoming more frequent. A common and simplified set of
terminology and design procedures is warranted. Certainly, the design procedures
should be consistent within a given manual.
Design
Criteria
The
objective is consistency in the sizing of filters and the design specifications
for a particular filter type. What is proposed is not that all manuals have the
same design criteria, but rather that the design criteria within the same manual
be consistent.
Regardless
of the filter type:
- Require surface vegetation
where climate permits
- Use Darcy’s Law to size
filter surface area
- Be consistent in the
specification of the hydraulic conductivity
- Specify a common operating
water volume
- Be consistent if specifying a
maximum drainage area
Varying
with filter type:
- Pretreatment
- Media
specification
- Media
thickness
- Operating water
depth
- Drawdown
time
The
first set of criteria was covered previously in Part 2. The second set is
covered in Part 3.
Pretreatment
Most
manuals are vague about pretreatment. Inconsistencies often occur in manuals:
for example, a manual specifies pretreatment for bioretention but not for the
dry swale, which is essentially bioretention with a slope. Some manuals specify
pretreatment but not the method of sizing. Many specify the sizing methodology.
However, often the procedure varies between filter types or between filters and
other treatment systems with pretreatment, such as wet basins. Presented in
Table 4 is a summary of the different methods for determining the size of the
pretreatment unit I have found in manuals. There appears to be little rationale
with some of the methods, particularly where different methods are used in the
same manual without apparent explanation.
Recommendation:
All filter types should include some level of pretreatment. However, it need not
necessarily be the same degree for all filter types. Presented is a suggested
framework for deciding the appropriate level of pretreatment, which in turn
provides a rationale for the manual author to vary the procedure within a
manual. The intent is to provide consistent thinking for pretreatment needed not
only between filter types but also across all treatment systems, including wet
basins where the forebay serves the same function.
Pretreatment
may be placed into four general categories of pollutant removal: removal of
gross solids; removal of gross solids plus coarse sediments; removal of both of
those plus fine sediments; and removal of all three plus dissolved pollutants.
The removal of gross solids or gross and coarse solids is sufficient for
vegetated systems. It has been observed that the vegetation and/or mulch likely
retain the sediment, protecting the filter surface from clogging. With vegetated
systems, there is also the need, in the interest of aesthetics, of keeping
litter out of the vegetation and facilitating the ease of its removal. The same
is true for sand. Gross solids and coarse sediment removal should be the primary
function of the forebay in wet basins as well. Sand and much of the coarse silt
is removed, generally 50 microns and greater.
Bare
filters and infiltration systems require more substantial pretreatment to remove
much of the fine silts as well, perhaps as small as 25 microns. Pretreatment to
remove essentially all sediment and dissolved pollutants may be appropriate for
bare infiltration systems in coarse soils where added protection of the
underlying groundwater aquifer is warranted. In such situations, it may be
feasible to provide a pretreatment unit ahead of the infiltration basin for
removal of the gross and coarse solids, and a specified filter media as a
treatment layer in the infiltration basin for removal of the silts and dissolved
pollutants. This concept is discussed further later in Part 3.
Media
Specification
Almost
all manuals now specify ASTM C33 fine aggregate for sand filters, as well as for
the sand in bioretention media blends. The most common inconsistency is to
specify a very explicit media composition for bioretention, yet to require for
dry swales only “planting soil,” as noted in Table 1. I have found more than 20
different specifications for media in bioretention filters and four different
specifications for the organic filter. It has not been established that
performance differs between these mixes.
Recommendation:
Tables 2 and 3 relate targeted pollutant to filter type. Continue the practice
of specifying ASTM C33 fine aggregate (concrete sand) or the equivalent for sand
in sand filters and blends as used in bioretention. Within a manual, be
consistent with the specifications for what are essentially the same treatment
systems: e.g., bioretention and dry swales. Consider specifying a treatment
surface layer for infiltration systems.
It
is recommended that the “minus 100” fraction be removed from the ASTM C33. This
step will likely prolong the life of the bed, particularly sand filters, and
reduce complications related to media freezing in cold climates. Moist sand
freezes, but openings remain for water passage if the fine particles are removed
from the ASTM C33 mix. The filtration rate is decreased by water frozen to the
sand particles, significantly if several freeze-thaw cycles occur over a few
days (Backstrom and Bergstrom 2000). However, even under these conditions, the
hydraulic conductivity will still likely exceed the design value if based on
accumulated sediment, discussed in Part 2. The greater concern is freezing of
the moist accumulated sediment. However, if frozen, it likely thaws with the
initial flow of water, whether a winter or spring melt or storm. Regardless, it
may be prudent to clean the filter surface each fall.
Some
manuals specify 100% loam soil for bioretention filters, but the
infiltration/filtration rate is too low, requiring a substantial area if the
total design water-quality volume (DWQV) is to be temporarily
stored.
Performance appears satisfactory with the sand and organic blend. Composition
might differ by climate: minimal clay in cold climates and less organic matter
in semiarid climates. These variations may alter performance. Clay is likely to
be important for dissolved phosphorus removal, and organic matter for dissolved
metals and toxic organics removal.
Use
of the media specification for sand and bioretention filters should be
considered for infiltration systems in particular situations: e.g., coarse
outwash soils, cracked basalt, and regions of karst geology. The sand filter
specification would be used where the removal of dissolved pollutants is not an
objective, and the bioretention specification where it is. As noted previously,
excavation of the A soil layer removes natural organic matter important for the
removal of dissolved pollutants, in particular metals and toxic organics such as
pesticides.
The
sand filter specification, or perhaps a somewhat coarser mix, might also be
suitable for cold climates to avoid freezing of the soil, as well as where there
is concern about mosquitoes or the formation of algal mats.
Media
Depth or Thickness
We
currently have a ménage of criteria, ranging from 1 to 4 feet for engineered
filters (Table 1). If the depth to groundwater can be viewed as the media
thickness for infiltration systems, the criterion ranges from as little as 3 to
as much as 15 feet across the many manuals.
The
original criterion for bioretention cells was 4 feet, but now is as little as 18
inches. Some manuals specify 12 inches for the lineal sand filter but 18 inches
for the basin. A few manuals specify a few inches of gravel atop a sand filter
for differing reasons: In filter vaults, it collects the litter. In a surface
basin, it inhibits the formation of algal mats. Algal mats may form because of
the slow entry of the last few inches of water. By having this occur within the
gravel, formation of the mat is avoided. The design also deters mosquitoes.
However, gravel may complicate removal of sediment from the top of the
sand.
Why
do most manuals specify 18 inches of thickness for the sand filter basin but
only 12 inches for the lineal filter? Twelve inches, and possibly less, is
sufficient for both. A vegetated basin could have turf sod over as little as 6
inches of sand, rather than 18 inches. The selection of 18 inches almost 30
years ago was made without the benefit of knowing the effect of media thickness
on performance. Studies with both water-treatment and stormwater filters show
that essentially all removal occurs on the filter surface and within the top 1
or 2 inches of fine media. Recent studies of stormwater filters indicate that a
thickness of 9 to 12 inches is as effective as one of 18 inches for total
suspended solids (TSS) removal. The new BayFilter appears effective with only 1
inch.
Recommendation:
Twelve inches is sufficient for sand filters, including a turf layer where used,
when the objective is sediment and particulate pollutant removal. A greater
depth may be necessary if an amendment is included, depending on its capacity.
For bioretention, different depths have been recommended depending on the
treatment objective, as shown in Table 4 (NCSUCES 2006). The depth needs of the
surface vegetation are also considered: perhaps 12 inches for turf grass, 18
inches for herbaceous shrubs, and 30 inches for trees. The same concept can be
used for dry swales, and infiltration basins, and
trenches.
Concerns
are frequently raised about contamination of groundwater; this is the common
rationale for specifying a minimum distance to the seasonal groundwater level.
This issue could be resolved by placing a filter atop the basin soil after
overexcavation below the design elevation as previously
suggested.
Where
the removal of dissolved pollutants is an objective, the volume of the media
matters, not just the surface area. In these situations Darcy’s Law defines the
minimum surface area needed to pass water. An additional step is necessary to
define the volume of media required, which is related to the performance
capability and capacity of the particular sorptive media (Minton 2005). I have
observed that some studies of filter media identify operating capacity focusing
only on short-term performance.
Media
volume also matters if temperature or stormwater volume reduction is an
objective. Methods are needed to define the volume for either objective,
combined with Darcy’s Law to define the surface area. The most appropriate
approach is some form of continuous simulation as previous
suggested.
Operating
Water Depth
Most
manuals specify a maximum water depth of 6 inches for bioretention cells but 18
inches for dry swales, although both are covered by vegetation. A few manuals
allow 12 inches for bioretention filters, and there has been discussion of
increasing this depth to 18 inches. The greater the allowable operating depth,
the smaller the filter area, as well as total facility area needed to
temporarily store stormwater during each storm. Regardless, there should be
consistency within a manual.
Some
manuals specify a maximum water depth for
sand filters ranging from 1 to 10 feet, but they commonly do not specify a
maximum depth for infiltration basins. The shallow maximum depth appears to be
favored by those concerned about either the compaction of the sand or the
accumulated thin sediment layer. I have been unable to find evidence that this
effect occurs. For either reason, both filters and infiltration basins should
have the same specification. The decision on maximum operating depth is not
trivial: 1 foot requires three times the filter surface area as 6
feet.
Recommendation:
Certainly 3 feet, but possibly as much as 6 feet, is fine for sand filters and
infiltration basins. A maximum depth of 12 inches, but possibly 18 inches, is
likely satisfactory for bioretention filters and filter swales where plant
survival is of concern. However, it may be prudent to limit the operating depth
to perhaps 3 feet if using a thinner media of 6 to 12 inches in sand filters.
Thinner filters may be more subject to the formation of “holes” by turbulence or
other factors, as observed in potable water filters. Energy dissipation is
important in this situation. A surface fabric may inhibit this condition.
Studies are needed.
Greater
depths for bioretention filters may raise concern for safety, given how these
systems are placed in developments. Data are need regarding the effect of water
depth on the compaction of sediment accumulation, its effect on hydraulic
conductivity, and whether this increases the maintenance
frequency.
Drawdown
Time
The
specification varies between manuals and frequently within a manual for
different filter types, ranging from 24 to 72 hours, in an apparently irrational
manner. Why should sand filters drain within 24 hours, but infiltration basins
within 48 hours? Yet bioretention units, where plant health is of particular
concern, are often specified at 72 hours. One would expect that bioretention
should be specified to drain more quickly than sand filters. Manuals commonly do
not provide reasons for particular drawdown times. Reasons given elsewhere
include drying sand media to desiccate bacterial or algae growth that contribute
to clogging, protecting surface vegetation, safety, and avoiding
mosquitoes.
Anaerobiosis
deprives plant roots of dissolved oxygen but also can result in the production
of natural organics by bacteria that are toxic to plants (Minton 2005). Higher
rates of dissolved oxygen use are likely to occur with vegetated systems because
of the organic matter. Longer drawdown times are likely acceptable in colder
climates. The reduction of soil oxygen takes longer, and plants can survive
longer under water. However, under such conditions, wetland plant types that can
tolerate these conditions would replace less-tolerant species. But these plants
are apparently not desired in bioretention filters.
The
tolerance of plant species to temporary soil saturation as a function of design
water depth, flood duration, and frequency is not well understood. Some manuals
identify the species that have some tolerance to the conditions expected in
treatment systems that temporarily pond. But the relationship to the duration of
flooding—i.e., 24 versus 48 hours—does not appear to be understood. It is
undoubtedly related to frequency: the less frequent the flooding, the greater
the tolerance to the length of an event. This suggests that design drawdown
could be greater in semiarid environments with its less-frequent runoff events,
although it is likely that native plants in such areas have less tolerance than
species native to humid climates.
Recommendation:
A manual might have two drawdown times: One for vegetated and one for bare
surface filter types. Vegetated systems perhaps should have a drawdown time not
exceeding 24 hours to minimize the likelihood of anaerobic conditions. In
contrast, there is no such restriction for bare filters: 72 hours may be just
fine. The use of a gravel layer, as previously described, may allow for a
drawdown time longer than 72 hours in semiarid areas, as it negates the issues
of mosquitoes and algal mats that occur with longer drawdown times. Subsurface
sand filters likely require a shorter drawdown time, particularly in wet
climates than surface filters. Subsurface filters are less exposed to
evaporation by wind and direct sunlight.
Climate
plays a role in specifying the drawdown time. Media supporting vegetation in
cold climates likely takes longer to become anaerobic, possibly allowing a
longer drawdown time than in semiarid and humid climates. Sand filters in humid
regions likely require a more rapid drawdown time than those in semi-arid
regions to provide sufficient time for drying between events.
One
manual for sand filters specifies 48 hours for design but 72 hours for
maintenance. A design hydraulic conductivity of 3.5 feet per day becomes 2.3 as
the drawdown time increases from 48 to 72 hours. The throughput rate is
correspondingly reduced by about one-third. In turn, the volume performance goal
(VPG), commonly 90%, is not met, having been reduced to 60%. However, it may
average 90% over the maintenance cycle, with higher throughput rates early after
cleaning. This may be a reasonable approach.
Advertisement
Summary
for Part 3
Let’s
simplify our terminology by using names that more explicit: filter
swale
rather than dry
swale;
bioretention
filter
rather than bioretention.
Let’s use consistent sizing procedures and design criteria for each filter type
within a BMP manual.
Summary
for the Series
The
objective is consistency in the sizing of filters and the design specifications
for a particular filter type. What is proposed is not that all manuals have the
same design criteria. Rather, that the design criteria within a manual be
consistent.
Author's Bio: Gary R. Minton, Ph.D., P.E., is an independent consultant on stormwater treatment with Resource Planning Associates. He is the author of the book Stormwater Treatment: Biological, Chemical, and Engineering Principles.
July-August 2008
A Simplified Integrated Design Concept for Filters
Part 3: Design criteria differing with filter types
In
Part 3, we consider design criteria that might differ between filter types but
would be identical within each filter type. Part 2 covered design criteria that
should be consistent irrespective of the filter type.
Part
1 covered terms and terminology. As noted in Part 1, I have observed in manuals,
articles, and reports, as well as in presentations and conversations at
conferences, that the complexity of terminology itself leads to misperceptions
and confusion over expected performance and to unnecessary and inappropriate
distinctions in design procedures and criteria. This dynamic has led to
inconsistencies in design procedures, frequently within the same manual. This
can result to a bias toward particular systems, because design criteria drive
costs. It also complicates the design process for the practitioner, particularly
those in local government who review the drainage plans of the development
engineer. Presented in Part 1 was Table 1, summarizing the many names and the
widely varying design criteria. I suggested simplification; two scenarios were
offered and are repeated in Part 3 as Tables 2 and 3.
Engineers
do not necessarily realize these differences and potential conflicts because we
work within our own community, state, or province with an agreed terminology and
set of design criteria. However, as it becomes increasingly common to trade
experiences and field results across regions and borders, contradictions and
miscommunication are becoming more frequent. A common and simplified set of
terminology and design procedures is warranted. Certainly, the design procedures
should be consistent within a given manual.
Design
Criteria
The
objective is consistency in the sizing of filters and the design specifications
for a particular filter type. What is proposed is not that all manuals have the
same design criteria, but rather that the design criteria within the same manual
be consistent.
Regardless
of the filter type:
- Require surface vegetation
where climate permits
- Use Darcy’s Law to size
filter surface area
- Be consistent in the
specification of the hydraulic conductivity
- Specify a common operating
water volume
- Be consistent if specifying a
maximum drainage area
Varying
with filter type:
- Pretreatment
- Media
specification
- Media
thickness
- Operating water
depth
- Drawdown
time
The
first set of criteria was covered previously in Part 2. The second set is
covered in Part 3.
Pretreatment
Most
manuals are vague about pretreatment. Inconsistencies often occur in manuals:
for example, a manual specifies pretreatment for bioretention but not for the
dry swale, which is essentially bioretention with a slope. Some manuals specify
pretreatment but not the method of sizing. Many specify the sizing methodology.
However, often the procedure varies between filter types or between filters and
other treatment systems with pretreatment, such as wet basins. Presented in
Table 4 is a summary of the different methods for determining the size of the
pretreatment unit I have found in manuals. There appears to be little rationale
with some of the methods, particularly where different methods are used in the
same manual without apparent explanation.
Recommendation:
All filter types should include some level of pretreatment. However, it need not
necessarily be the same degree for all filter types. Presented is a suggested
framework for deciding the appropriate level of pretreatment, which in turn
provides a rationale for the manual author to vary the procedure within a
manual. The intent is to provide consistent thinking for pretreatment needed not
only between filter types but also across all treatment systems, including wet
basins where the forebay serves the same function.
Pretreatment
may be placed into four general categories of pollutant removal: removal of
gross solids; removal of gross solids plus coarse sediments; removal of both of
those plus fine sediments; and removal of all three plus dissolved pollutants.
The removal of gross solids or gross and coarse solids is sufficient for
vegetated systems. It has been observed that the vegetation and/or mulch likely
retain the sediment, protecting the filter surface from clogging. With vegetated
systems, there is also the need, in the interest of aesthetics, of keeping
litter out of the vegetation and facilitating the ease of its removal. The same
is true for sand. Gross solids and coarse sediment removal should be the primary
function of the forebay in wet basins as well. Sand and much of the coarse silt
is removed, generally 50 microns and greater.
Bare
filters and infiltration systems require more substantial pretreatment to remove
much of the fine silts as well, perhaps as small as 25 microns. Pretreatment to
remove essentially all sediment and dissolved pollutants may be appropriate for
bare infiltration systems in coarse soils where added protection of the
underlying groundwater aquifer is warranted. In such situations, it may be
feasible to provide a pretreatment unit ahead of the infiltration basin for
removal of the gross and coarse solids, and a specified filter media as a
treatment layer in the infiltration basin for removal of the silts and dissolved
pollutants. This concept is discussed further later in Part 3.
Media
Specification
Almost
all manuals now specify ASTM C33 fine aggregate for sand filters, as well as for
the sand in bioretention media blends. The most common inconsistency is to
specify a very explicit media composition for bioretention, yet to require for
dry swales only “planting soil,” as noted in Table 1. I have found more than 20
different specifications for media in bioretention filters and four different
specifications for the organic filter. It has not been established that
performance differs between these mixes.
Recommendation:
Tables 2 and 3 relate targeted pollutant to filter type. Continue the practice
of specifying ASTM C33 fine aggregate (concrete sand) or the equivalent for sand
in sand filters and blends as used in bioretention. Within a manual, be
consistent with the specifications for what are essentially the same treatment
systems: e.g., bioretention and dry swales. Consider specifying a treatment
surface layer for infiltration systems.
It
is recommended that the “minus 100” fraction be removed from the ASTM C33. This
step will likely prolong the life of the bed, particularly sand filters, and
reduce complications related to media freezing in cold climates. Moist sand
freezes, but openings remain for water passage if the fine particles are removed
from the ASTM C33 mix. The filtration rate is decreased by water frozen to the
sand particles, significantly if several freeze-thaw cycles occur over a few
days (Backstrom and Bergstrom 2000). However, even under these conditions, the
hydraulic conductivity will still likely exceed the design value if based on
accumulated sediment, discussed in Part 2. The greater concern is freezing of
the moist accumulated sediment. However, if frozen, it likely thaws with the
initial flow of water, whether a winter or spring melt or storm. Regardless, it
may be prudent to clean the filter surface each fall.
Some
manuals specify 100% loam soil for bioretention filters, but the
infiltration/filtration rate is too low, requiring a substantial area if the
total design water-quality volume (DWQV) is to be temporarily
stored.
Performance appears satisfactory with the sand and organic blend. Composition
might differ by climate: minimal clay in cold climates and less organic matter
in semiarid climates. These variations may alter performance. Clay is likely to
be important for dissolved phosphorus removal, and organic matter for dissolved
metals and toxic organics removal.
Use
of the media specification for sand and bioretention filters should be
considered for infiltration systems in particular situations: e.g., coarse
outwash soils, cracked basalt, and regions of karst geology. The sand filter
specification would be used where the removal of dissolved pollutants is not an
objective, and the bioretention specification where it is. As noted previously,
excavation of the A soil layer removes natural organic matter important for the
removal of dissolved pollutants, in particular metals and toxic organics such as
pesticides.
The
sand filter specification, or perhaps a somewhat coarser mix, might also be
suitable for cold climates to avoid freezing of the soil, as well as where there
is concern about mosquitoes or the formation of algal mats.
Media
Depth or Thickness
We
currently have a ménage of criteria, ranging from 1 to 4 feet for engineered
filters (Table 1). If the depth to groundwater can be viewed as the media
thickness for infiltration systems, the criterion ranges from as little as 3 to
as much as 15 feet across the many manuals.
The
original criterion for bioretention cells was 4 feet, but now is as little as 18
inches. Some manuals specify 12 inches for the lineal sand filter but 18 inches
for the basin. A few manuals specify a few inches of gravel atop a sand filter
for differing reasons: In filter vaults, it collects the litter. In a surface
basin, it inhibits the formation of algal mats. Algal mats may form because of
the slow entry of the last few inches of water. By having this occur within the
gravel, formation of the mat is avoided. The design also deters mosquitoes.
However, gravel may complicate removal of sediment from the top of the
sand.
Why
do most manuals specify 18 inches of thickness for the sand filter basin but
only 12 inches for the lineal filter? Twelve inches, and possibly less, is
sufficient for both. A vegetated basin could have turf sod over as little as 6
inches of sand, rather than 18 inches. The selection of 18 inches almost 30
years ago was made without the benefit of knowing the effect of media thickness
on performance. Studies with both water-treatment and stormwater filters show
that essentially all removal occurs on the filter surface and within the top 1
or 2 inches of fine media. Recent studies of stormwater filters indicate that a
thickness of 9 to 12 inches is as effective as one of 18 inches for total
suspended solids (TSS) removal. The new BayFilter appears effective with only 1
inch.
Recommendation:
Twelve inches is sufficient for sand filters, including a turf layer where used,
when the objective is sediment and particulate pollutant removal. A greater
depth may be necessary if an amendment is included, depending on its capacity.
For bioretention, different depths have been recommended depending on the
treatment objective, as shown in Table 4 (NCSUCES 2006). The depth needs of the
surface vegetation are also considered: perhaps 12 inches for turf grass, 18
inches for herbaceous shrubs, and 30 inches for trees. The same concept can be
used for dry swales, and infiltration basins, and
trenches.
Concerns
are frequently raised about contamination of groundwater; this is the common
rationale for specifying a minimum distance to the seasonal groundwater level.
This issue could be resolved by placing a filter atop the basin soil after
overexcavation below the design elevation as previously
suggested.
Where
the removal of dissolved pollutants is an objective, the volume of the media
matters, not just the surface area. In these situations Darcy’s Law defines the
minimum surface area needed to pass water. An additional step is necessary to
define the volume of media required, which is related to the performance
capability and capacity of the particular sorptive media (Minton 2005). I have
observed that some studies of filter media identify operating capacity focusing
only on short-term performance.
Media
volume also matters if temperature or stormwater volume reduction is an
objective. Methods are needed to define the volume for either objective,
combined with Darcy’s Law to define the surface area. The most appropriate
approach is some form of continuous simulation as previous
suggested.
Operating
Water Depth
Most
manuals specify a maximum water depth of 6 inches for bioretention cells but 18
inches for dry swales, although both are covered by vegetation. A few manuals
allow 12 inches for bioretention filters, and there has been discussion of
increasing this depth to 18 inches. The greater the allowable operating depth,
the smaller the filter area, as well as total facility area needed to
temporarily store stormwater during each storm. Regardless, there should be
consistency within a manual.
Some
manuals specify a maximum water depth for
sand filters ranging from 1 to 10 feet, but they commonly do not specify a
maximum depth for infiltration basins. The shallow maximum depth appears to be
favored by those concerned about either the compaction of the sand or the
accumulated thin sediment layer. I have been unable to find evidence that this
effect occurs. For either reason, both filters and infiltration basins should
have the same specification. The decision on maximum operating depth is not
trivial: 1 foot requires three times the filter surface area as 6
feet.
Recommendation:
Certainly 3 feet, but possibly as much as 6 feet, is fine for sand filters and
infiltration basins. A maximum depth of 12 inches, but possibly 18 inches, is
likely satisfactory for bioretention filters and filter swales where plant
survival is of concern. However, it may be prudent to limit the operating depth
to perhaps 3 feet if using a thinner media of 6 to 12 inches in sand filters.
Thinner filters may be more subject to the formation of “holes” by turbulence or
other factors, as observed in potable water filters. Energy dissipation is
important in this situation. A surface fabric may inhibit this condition.
Studies are needed.
Greater
depths for bioretention filters may raise concern for safety, given how these
systems are placed in developments. Data are need regarding the effect of water
depth on the compaction of sediment accumulation, its effect on hydraulic
conductivity, and whether this increases the maintenance
frequency.
Drawdown
Time
The
specification varies between manuals and frequently within a manual for
different filter types, ranging from 24 to 72 hours, in an apparently irrational
manner. Why should sand filters drain within 24 hours, but infiltration basins
within 48 hours? Yet bioretention units, where plant health is of particular
concern, are often specified at 72 hours. One would expect that bioretention
should be specified to drain more quickly than sand filters. Manuals commonly do
not provide reasons for particular drawdown times. Reasons given elsewhere
include drying sand media to desiccate bacterial or algae growth that contribute
to clogging, protecting surface vegetation, safety, and avoiding
mosquitoes.
Anaerobiosis
deprives plant roots of dissolved oxygen but also can result in the production
of natural organics by bacteria that are toxic to plants (Minton 2005). Higher
rates of dissolved oxygen use are likely to occur with vegetated systems because
of the organic matter. Longer drawdown times are likely acceptable in colder
climates. The reduction of soil oxygen takes longer, and plants can survive
longer under water. However, under such conditions, wetland plant types that can
tolerate these conditions would replace less-tolerant species. But these plants
are apparently not desired in bioretention filters.
The
tolerance of plant species to temporary soil saturation as a function of design
water depth, flood duration, and frequency is not well understood. Some manuals
identify the species that have some tolerance to the conditions expected in
treatment systems that temporarily pond. But the relationship to the duration of
flooding—i.e., 24 versus 48 hours—does not appear to be understood. It is
undoubtedly related to frequency: the less frequent the flooding, the greater
the tolerance to the length of an event. This suggests that design drawdown
could be greater in semiarid environments with its less-frequent runoff events,
although it is likely that native plants in such areas have less tolerance than
species native to humid climates.
Recommendation:
A manual might have two drawdown times: One for vegetated and one for bare
surface filter types. Vegetated systems perhaps should have a drawdown time not
exceeding 24 hours to minimize the likelihood of anaerobic conditions. In
contrast, there is no such restriction for bare filters: 72 hours may be just
fine. The use of a gravel layer, as previously described, may allow for a
drawdown time longer than 72 hours in semiarid areas, as it negates the issues
of mosquitoes and algal mats that occur with longer drawdown times. Subsurface
sand filters likely require a shorter drawdown time, particularly in wet
climates than surface filters. Subsurface filters are less exposed to
evaporation by wind and direct sunlight.
Climate
plays a role in specifying the drawdown time. Media supporting vegetation in
cold climates likely takes longer to become anaerobic, possibly allowing a
longer drawdown time than in semiarid and humid climates. Sand filters in humid
regions likely require a more rapid drawdown time than those in semi-arid
regions to provide sufficient time for drying between events.
One
manual for sand filters specifies 48 hours for design but 72 hours for
maintenance. A design hydraulic conductivity of 3.5 feet per day becomes 2.3 as
the drawdown time increases from 48 to 72 hours. The throughput rate is
correspondingly reduced by about one-third. In turn, the volume performance goal
(VPG), commonly 90%, is not met, having been reduced to 60%. However, it may
average 90% over the maintenance cycle, with higher throughput rates early after
cleaning. This may be a reasonable approach.
Summary
for Part 3
Let’s
simplify our terminology by using names that more explicit: filter
swale
rather than dry
swale;
bioretention
filter
rather than bioretention.
Let’s use consistent sizing procedures and design criteria for each filter type
within a BMP manual.
Summary
for the Series
The
objective is consistency in the sizing of filters and the design specifications
for a particular filter type. What is proposed is not that all manuals have the
same design criteria. Rather, that the design criteria within a manual be
consistent.