A Tower of Babel
A systems-based approach toward an integrative management strategy
So just what is hydrodynamic separation? This descriptor for certain treatment products has recently proliferated. Yet no one seems to have provided a definition. The term hydrodynamic separation was initially applied to vortex separators such as the Downstream Defender and Vortechs. But this little group has expanded to include small manufactured wet vaults such as Stormceptor, BaySaver, and CrystalStream devices. And what is vortex separation?
Common to the early evolution of new fields of technical application is the multiplicity of duplicative and overlapping terms. Consolidation occurs in time with the disuse of duplicative and ill-defined terms. Perhaps this occurred with automobiles in the early 20th century and computers in the late 20th century.
In our field we have several names for essentially the same treatment system, different systems with the same name, words without apparent definition, and the misapplication of words. Does not hydrodynamic or vortex convey the impression that the labeled treatment system has some extra-special characteristic that improves its performance?
Consider swales: We have biofilters, grassy swales, vegetated swales, grass channels, landscaped swales, wetland swales, bioretention swales, dry and wet swales, and enhanced dry and wet swales. No wonder we have confusion over “what is best.” Several of these swales are sized based on the peak of a design event, while others are sized based on the volume of the design event. The latter are, in effect, basins or filters, depending on the particular system.
In contrast, we also have what are essentially the same systems but with different names. Consider the organic filter presented in some manuals and the bioretention system with underdrains (i.e., filter). The bioretention media is a mixture of 50% sand and 50% organic media, the latter a loam soil, compost, or a mixture. The media depth ranges from 18 to 48 inches. Cover is commonly a mix of grass, shrubs, and trees, although one specification is for grass only. In contrast, the organic filter has a media mixture of 50% sand and 50% organic matter, usually peat. It has an 18-inch layer of the mixed media placed above an 18-inch layer of sand. A grass cover may or may not be specified. Although the two systems are essentially identical, we have very different names for them and, more importantly, two distinctly different sizing procedures. The organic filter is sized as a filter using Darcy’s Law, whereas the bioretention filter is often sized as an infiltration system even when it has underdrains.
Currently, bioretention is a name for a treatment system. But what do we mean by bioretention? Presumably that biological organisms play a role in pollutant removal. The plants remove nutrients and metals for growth. Bacteria transform ammonia into nitrogen gas. Thus defined, bioretention also occurs in wet ponds and constructed wetlands as well. Therefore, a better use of the term bioretention is as a general descriptor (i.e. mechanism) of pollutant removal—what wastewater engineers call a unit process, a concept that will be shortly introduced and described. Bioretention thus used is akin to sedimentation, flotation, and filtration. Just as we have used the word retention to describe the holding back of water, bioretention describes the holding back of pollutants by plants and bacteria.
A New Framework
The proliferation of terminology has lead to misperceptions of performance, uncertainty as to applicability of particular treatment systems, and the inconsistent, confusing, and sometimes illogical application of design criteria. What is needed is a simplified framework with concurrence on a common name for the same treatment system to which design criteria are applied consistently. In this article I offer such a framework. It is offered as a beginning point for a dialog on this subject.
We begin with Figure 1. Figure 1 represents a hierarchical portrayal of the functioning of treatment systems as well as their organization. Basic principles serve as the foundation for the performance of all treatment systems. We may for convenience divide these into engineering, biological, and chemical principles (Minton 2005). These principles underlie how it all works. Examples are the settling of fine particles as described by Stokes’s Law and Newton’s Law, photosynthesis in plants, the chemistry of precipitation and sorption of pollutants to sediments in stormwater as well as to filter media, and the movement of water though a basin.
Principles define how unit processes work. The concept of unit processes and operations shown in Figure 1 is borrowed from wastewater and chemical engineering but simplified to meet our needs. Unit processes as defined here are the mechanisms of pollutant removal: for example, sedimentation, flotation, coagulation, filtration, denitrification, and screening. We may convey the role of biological organisms by putting the qualifier bio in front of some of these terms—for example, biofiltration, biosedimentation, and bioretention. The first example, biofiltration, refers to the removal of particulate pollutants by the filtering action of plants such as biofilms on wetland plants remove colloidal material. An example of biosedimentation is grass in a grass swale, which slows the water, allowing particles to settle to the bottom of the swale. As proposed here, bioretention refers to and is interchangeable with the term biological processes, essentially representing a group of unit processes.
Unit processes occur within unit operations, the next level in the pyramid of Figure 1. A unit operation is the treatment unit: the “box” in which one or more unit processes occurs. For example, in a sand filter itself, excluding the pretreatment unit, we have sedimentation and filtration. These are two different unit processes. Depending on the media, we may have a third unit operation, sorptive filtration with the removal of dissolved pollutants. Sorptive filtration is distinguished from classic physical or mechanical filtration.
One or more unit operations is a system. A swale usually has one unit operation, the swale itself, but it is a system as well. A sand filter as a system has two unit operations: pretreatment and the filter. A wet pond with a forebay can be said to have two unit operations because the forebay serves a particular function, which is to ease maintenance. By recognizing each element as a separate unit operation with a particular function, the engineer may consider more cost-effective solutions. Rather than a forebay in the pond, the engineer might choose a swirl concentrator. I suggest that the term system is a more apt term than treatment train. Confusion exists with the use of the descriptor treatment train. It is commonly applied to a system consisting of two or more separate structures—for example, a swale followed by a wet pond—but is not applied to two distinct treatment elements within the same structure—for example, a sand filter vault that includes both the filter bed and the pretreatment forebay. In contrast, we have the multichamber treatment train (MCTT) in which the pretreatment unit and filter bed are in the same vault. The application of the term unit operation eliminates this inconsistency. Table 1 provides my observations on the unit processes that occur in selected unit operations.
Although I have drawn the concept of unit processes from the wastewater industry, I have modified it to better suit our needs. In wastewater treatment, the term process is applied to biological and chemical processes, but the term operation is applied to physical processes. As a wastewater engineer early in my career, I found the distinction arbitrary and without explanation as to purpose or value. Rich (1961), who first fully applied the distinction to wastewater treatment, neither explained nor established its purpose or benefit. He borrowed from chemical engineering. In fact, Rich defined unit operations as physical processes and even identified some physical processes as unit processes rather than operations because of their chemical dependence (Rich 1963). Nor has the chemical engineering profession from which Rich borrowed the construct provided a rationale for the distinction.
An earlier text by engineers who were very influential in wastewater engineering defined all pollutant removal mechanisms as unit operations (Fair and Geyer 1958). The point of this brief historical trip is to establish that past categorization has been arbitrary and varying. Hence, let’s use a simplified construct that works for us. Unfortunately, the old concept by Rich has been put forth in a recent publication on stormwater treatment (WERF 2005). I highly recommend the simplified construct of unit processes and unit operations as defined here be adopted. There is no need for unnecessary complications that diminish the clarity and as a consequence the power of our framework.
The derivation of the framework should use the following principles. The name of a unit operation or system conveys its basic characteristic: for example, filter swale rather than dry swale. The name should not be so general as to be applicable to unit operations or systems other than itself. Examples that have too broad an application are bioretention and MCTT. Divisions of unit operations or systems should be made sparingly only when a clear reason is present. Certainly no framework works perfectly. Compromises will be seen in the following framework.
Families
Systems with common key characteristics are grouped into families, located at the top of the pyramid in Figure 1. Five broad families are proposed for stormwater treatment:
- Basins
- Swales
- Filters
- Infiltrators
- Screens
A figure is presented for each of the above families (Figures 2 through 6). Some of the families are further divided into subfamilies before further division to unit operations and systems. Examples of systems are given in each family. In some cases, the unit operation and system have the same name. But there is a distinction, which is illustrated with examples. As a unit operation, a sand filter is just the filter bed and its container. As a system, the sand filter is the pretreatment unit and the bed, whether they are in the same container or separate containers. As a unit operation, the grass swale is just the swale. It may also be just the swale as a system, but sometimes the swale is immediately preceded by a unit operation that removes coarse sediments and/or litter. A complication of the hierarchical framework represented in Figure 1 is that to construct a system one uses unit operations from more than one family. Notwithstanding this weakness, I have found the proposed framework to work.
Provided in Figures 2 through 6 are terms I recommend be discontinued for simplicity and clarity. Our goal is to have one name for what are essentially the same systems.
Family Basins
A basin is a family of unit operations in which water is detained, the duration differing with the type of basin. The family is illustrated in Figure 2. Treated stormwater is discharged via a weir or orifice to a pipe or ditch, eventually reaching a surface water body such as a stream or lake. A portion of the stormwater may infiltrate through the bottom of the basin, but the aggregate volume infiltrated over time is less than what is discharged to surface waters. In considering performance, credit may be given for infiltration depending on whether the interest is concentration or loading reduction.
The family is divided into three broad subfamilies: wet, wet extended detention, and extended detention. Extended detention basins become completely dry after each storm, with the lowest outlet at the bottom of the basin. The lowest outlet of a wet basin is commonly several feet above the bottom for a wet pond but only a foot or so above the bottom for a constructed wetland. The term wet is used rather than retention, a commonly used descriptor as wet contrasts with dry. Some would argue that detention contrasts with retention. However, whether the pond is wet or dry, water is detained or retained. The only difference is the duration. Retention is used by some to describe infiltration basins. But why is water detained if it discharges to a receiving water body but retained if it discharges to aquifers?
The wet extended detention basin is a hybrid of the two extremes. It is common to split the design volume between the active and dead storage, suggesting a performance between the fully wet and dry systems. The concept evolved to improve the performance of an extended detention basin. The benefits of the wet extended basin over a full wet basin have not been stated. A wet basin retains twice the volume of water between events, thereby improving the removal of dissolved pollutants and clays. Two possible benefits are provided by the hybrid: less thermal impact, and safety with its shallower permanent pool depth. These distinctions justify the division. The word pond is used in Figure 2 for extended detention rather than basin, the more commonly used term, because a basin can be a vault or pond. A distinction is made between vaults and ponds or wetlands because the latter two remove dissolved pollutants.
The wet swale is included in the basin family rather than the swale family. The name is lengthened to wet extended detention swale, although retaining the name wet swale seems reasonable. The system is placed in this family because its size is based on a design storm volume rather than flow rate. It is essentially a very shallow and long wet extended detention basin with a more substantial slope (1% to 5%) than other basins. It is proposed that this particular configuration be simply identified as a variant of the wet extended detention basin, just as the perimeter sand filter is shown as a variant of the sand filter in the filter basin.
Included in the family are manufactured vaults such as those from Stormceptor, BaySaver, and CrystalStream. These basins are grouped in the unit operation called “standard” for want of a better term. Each has internal elements to improve hydraulic efficiency and to inhibit resuspension of removed sediments. But so does a vault designed by the engineer. A wet vault need not be proprietary. As any vault designed by an engineer must include baffles, the term baffle box is dropped. It does not differ significantly from oil/grit separator vaults, for example.
The oil/water separator (OWS) is a special vault with elements to enhance the removal of petroleum-related compounds. OWS separators could be divided into three separate types of separators or unit operations: the simple vault (oil/grit separator and API separator), those using laminar or coalescing plates (CPI), and those that rely solely on sorptive media, of which there is only the EcoSep. As with many systems, a manufactured unit can be purchased or the engineer can prepare a design and purchase a coalescing plate pack. A fourth type substitutes open plastic balls for the coalescing plates.
We drop the term water-quality inlet: It is superfluous as well as misleading inasmuch as the system is not usually placed on a drainage inlet but at the site discharge point like other treatment systems. Most often these are pretreatment units. It should also be noted that the term oil/grit separator has been applied by several to vortex separators and small values like the Stormceptor and CPI separators. I propose that the term oil/grit separator be dropped, as any vault of similar size (e.g., CrystalStream) can remove oil and grit at similar levels of efficiency. That is, the oil/grit separator is no different than a common pretreatment vault with baffles. It is essentially a grossly undersized API separator. This confusion is apparent in a fact sheet by the USEPA (1999).
Swirl concentrators are also vaults but with the unique attribute of a tangential inflow, the purpose of which is to improve the hydraulic efficiency of a simple round vault—for example, Hydro International’s Downstream Defender, Flo-Gard’s Dual Vortex,and AquaShield’s Aqua-Swirl. The descriptors vortex separation and hydrodynamic separation are not used for reasons given at the end of this article. It is recommended that these terms be dropped as general descriptors.
There are several wetland configurations. We are familiar with the marsh and micropool wetlands, the later essentially an extended detention basin with small wet pools at the entrance and exit. The hummock wetland is an Australian development consisting of alternating deep open-water pools and lateral shallow benches with wetland vegetation. This configuration has been found to exhibit improved hydraulic efficiency, better nitrogen removal, and lower mosquito populations than the other configurations and is therefore identified as a distinctly different unit operation. A wet pond-wetland has large areas of deeper water typically in the center. The distinction between this configuration and a small wet pond with a safety bench covered by wetland vegetation is admittedly obscure. A system not presented in Figure 2 is the pocket wetland, identified for relatively small drainages. It is not included as it can be considered as simply a small version of any of the other configurations that are shown.
The subsurface flow (SSF) wetland differs from the others in that the substrate is gravel, thereby allowing the stormwater to flow through the root system. This concept is adopted from wastewater treatment, where it is also called the SSF wetland. This term is proposed in lieu of the descriptor gravel wetland, which some are using. The engineer can design the SSF as an earthen basin or use the StormTreat, the one manufactured system that uses the concept. SSF wetlands have benefits with respect to avoidance of mosquitoes and are possibly less susceptible to freezing.
Note that in Figure 2 the same term is often used for both the unit operation and the system. The distinction is this: A pond or vault identified as a unit operation is just the single unit with no pretreatment element, for example. As a system, a pond or vault may contain a forebay. I view this as a separate unit operation as it has a distinct purpose apart from the pond, as previously described.
Family Swales
Swales (Figure 3) are unit operations with a distinct slope to facilitate the transit of stormwater from entrance to exit and very shallow water depths, usually a few inches at the maximum design depth. The majority of the treated stormwater is discharged to a surface water body rather than to groundwater. As with basins, infiltration may occur, but the volume is less than what is discharged to surface waters. In considering performance, credit may be given for infiltration depending on whether the interest is concentration or loading reduction. All unit operations in this family are sized to treat up to the peak flow of the design storm.
We have two subfamilies: swales and strips. With swales, the stormwater transverses the length, whereas with strips, the stormwater transverses its width. The former has concentrated flow whereas the later has sheet flow.
All swales and strips are vegetated given that a bare soil is to be avoided. Swales are categorized according to the type of vegetation: grass, wetland, and landscape. The same is true for strips. As implied by the name, a grass swale is covered by grass; a wetland swale is covered by wetland plants; and a landscape swale has a mix of grass, shrubs, and possibly trees. A wetland swale or strip has a very gradual slope, 1% or less. As a consequence, and depending on the soil type and climate, the soil remains wet throughout much of the year, thereby supporting wetland plant species.
The distinction between a landscape strip and a buffer is ground topography. The ground is leveled for a landscape strip, generally in a commercial development where strips are most commonly used. The buffer retains its original topography and is an integrated component of a residential development that abuts a stream, wetland, or lake.
As with the basin family, the same terms are sometimes used for both the unit operation and the system. But as with a pond or a vault, a swale as a unit operation is just the swale. As a system it may include a pretreatment unit to ease maintenance.
Several terms are dropped from current usage including biofilter, vegetated swale, grass channel, and bioretention swale. The dry swale is not included in this family, but rather with the filter family, as filtration is the dominant unit process.
Family Filters
A filter is a unit operation with these characteristics: underdrains, engineered filter media, and the majority of the treated stormwater discharging to a surface water body—that is, it is not infiltrated to the groundwater aquifer. Types of media include sand, peat, leaf compost, perlite, zeolite, fabric, metal-oxide–coated sand, and loamy soil. The unit may have only one medium or be multimedia, either as layers or mixed. The family is illustrated in Figure 4.
The family is divided into fine- and coarse-media filters. The distinction is made because the latter provides lesser but in most cases acceptable performance, and because the designs differ significantly. Fine media generally refers to sand filters, for which ASTM C33 fine concrete aggregate is commonly specified. A substantial head is required to obtain desired filtration rates to minimize the surface area of the filter. Coarse-media filters are commonly manufactured. The beds are thinner and the detention time is minutes in contrast to hours for fine-media filters, whose beds are usually thicker.
Fine- and coarse-media filters can be distinguished by the rate of filtration per unit area. The throughput of fine-media filters is about 0.02 gallon per square foot (0.06 liter per second per square meter), whereas it is 2 gallons per minute per square foot (6 liters per second per square meter) or greater for the coarse-media filters. It follows that coarse-media filters require a much smaller footprint. Examples of coarse-media filters are presented in Figure 4.
While most manufactured filters commonly use coarse media, the BayFilter has sand like the public-domain sand filter. The StormFilter also uses fine sand in some applications and therefore is found under both fine and coarse media in Figure 4. Some manufacturers offer media finer than their standard coarse media, although it is still coarser than what is used in sand filters. The dividing line between fine and coarse remains to be defined.
A third subfamily is catch basins or drain inlet inserts, for which there are many manufacturers.
Sorptive media listed in Figure 4 can be mixed with sand to remove dissolved pollutants, in particular metals or dissolved phosphorus. For this distinction, these filters are placed in their own subfamily of amended sand filters.
Bioretention is an amended sand filter and therefore is placed with this group. Most significantly, the term is dropped as a filter name. As noted previously, its media composition is not dissimilar to the older organic filter. The term multichamber treatment train is dropped, as it also refers to an amended sand filter with a composition similar to bioretention units. The name multichamber treatment does not meet two of our criteria: that names indicate the nature of the system, and that the name not be applicable to other treatment systems with the same characteristics. The descriptor multichamber is applicable to many other treatment systems. A wet pond with a forebay is a multichamber, as is a sand filter with its pretreatment unit. Our goal is to use names that most explicitly convey the primary nature of the system, which, for a MCTT, is amended sand filter.
Filterra is classified as an amended sand filter paralleling the recommendation that the term bioretention not be used as a unit operation. The media of the Filterra is a mixture of sand and gravel, with the surface mulch essentially an amendment. Its composition is between fine and coarse media. But as sand and mulch control the flow rate, it is placed in the fine-media grouping.
The same design parameters should be used to determine the surface area of the bed for the sand and amended sand filters (Minton 2005). The common design parameters are hydraulic conductivity, maximum drawdown time, and operating head. A different design procedure is currently used for the MCTT, particularly with respect to the pretreatment unit. The pretreatment unit is much larger than what is used with the standard sand or amended filter. The objective, in combination with laminar plates, is to extend the time before cleaning of the bed surface must occur, from perhaps annually to several years. Perhaps this approach should be adopted for all fine-media filters. But these differences are not sufficient to warrant its unique name or identification as something other than an amended sand filter.
It may be desirable to distinguish surface filters by the absence or intentional presence of vegetation. Although not common, sand filters have been intentionally designed with a grass cover. Anecdotal information indicates grass extends the maintenance cycle. This also appears to be the case with the bioretention filter. The design hydraulic conductivity may differ depending on whether the surface is vegetated. Information is lacking at this time to make the distinction.
Sand and amended sand filters are divided into several system types such as basin, lineal, and cell. It could be argued that basin and cell filters are the same thing, but a lineal filter is also in effect a basin. But the separation is made to accommodate current usage and to differentiate systems by size, configuration, and usage. The use of the descriptor basin with this family conflicts with its use in the basin family but is unavoidable. Vessel and container are unappealing terms.
Included in the filter family is the filter swale, placed here rather than with the swale family because its primary mode of treatment is filtration. I suggest that the proposed terminology is more explicit than dry swale and bioretention swale.
Family Infiltrators
Infiltrators are all unit operations in which the design water volume is infiltrated into the soil, with the ultimate receptor being the groundwater aquifer. The family is illustrated in Figure 5.
In this family, we have these unit operations: basins, trenches, swales, vaults, cells, and the older dry wells. What we now call bioretention becomes an infiltration cell. No doubt the descriptor rain garden will continue to be used, given its appeal to landscape architects and homeowners. The descriptor infiltration cell does not have the same “zing” as rain garden when promoting low-impact development (LID) site design. The dry swale is an infiltration swale if it does not have underdrains.
Note that the descriptor infiltration is put in front of many of the systems to distinguish them from their noninfiltrating cousins. It is a bit cumbersome, however. Infiltrator vaults are typically manufactured and assembled onsite. Figure 5 lists some examples. These products may also be in the basin family as they are wet or extended detention vaults where the soil has inadequate infiltration capabilities.
Included in the infiltrator family is the infiltration swale, placed here rather than with the swale family because its primary mode of treatment is infiltration. In this case we replace the terminology of bioinfiltration (as a system), dry swale, and bioretention swale, the latter used here where soil has a sufficient infiltration rate to obviate the need for underdrains (filtration swale).
The terms bioretention and bioinfiltration are dropped as descriptors of particular systems but become descriptors of unit processes.
Shall we use the word porous or permeable or pervious? All three are widely used, but porous appears to be the more commonly used descriptor. They have the same common definition. I have no preference, but the descriptor porous is used in Figure 5. Open-cell paver is a better descriptor than turf cell, because the cell need not contain turf. The term cementious is a bit academic and has little meaning to the practicing engineer.
As with filters, we might distinguish infiltrators by the absence or intentional presence of vegetation.
A few manuals include partial infiltration. This refers to systems that infiltrate the majority but not all of the design water-quality volume due to low infiltration rates. Some of the design-storm volume flows to surface discharge. The distinction between a partial infiltration system and its counterpart in the other families is vague and derives primarily from intent. The distinction is that with partial infiltration systems the majority of the stormwater is infiltrated, but less than the design-storm volume. This is in contrast to the basin and swale families in which infiltration is considered coincidental, defined as ranging from minor to less than half of the design-storm volume.
Family Screens
The screen family illustrated in Figure 6 consists of unit operations with large openings. Their primary if not sole intent is to remove gross solids such as litter, leaves, and plastics. Almost all of the systems described with the first four families also remove gross solids but are not primarily intended to do so. However, some may perform as well as some devices found in the screen family—for example, some of the drain inlet inserts listed under the filter family. The majority of screens are manufactured. Of course, trash racks used as elements in systems in the other families can be made with the basic materials of concrete and steel.
This family illustrates the difficulty with the placement of some of the manufactured products: in this case the CDS Technologies continuous deflective separation unit. It is not placed with the swirl concentrators in the basin family because its original purpose was litter control. It is offered as a sediment removal device, however.
Closing Observations
So what is hydrodynamic separation? The terminology is used by chemical and mining engineers and in medicine. Their common definition of hydrodynamic separation is the removal of particles by gravity from a moving fluid. The fluid can be air or water. The particles may be more or less dense than the fluid. An alternative term is density separation. Chemical engineers also narrowly define hydrodynamic separation as removal of a particle from a fluid that experiences a very abrupt change in its direction of flow. If we apply the first definition to stormwater treatment, which seems most appropriate, wet ponds and constructed wetlands are hydrodynamic separators, as are extended detention basins and filters—yes, filters, because most of the removal is by sedimentation on and within the filter media. Hydrodynamic separation applies to flow-through grass swales. In effect, the term is meaningless. I conclude that the descriptor hydrodynamic separation should not be used to describe the small group of treatment systems to which it is currently applied.
The descriptor hydrodynamic separation could apply to the more narrow definition of removal by an abrupt direction change in flow, but only if it is the dominant removal mechanism. However, this unit process is not generally employed in our systems. A possible exception is the Downstream Defender, in which there is an abrupt change in direction as water flows from the outer chamber into the center beneath the dip plate. Although studies indicate the dip plate improves performance, it is not clear that it is due to the directional change. Regardless, it is not the dominant removal mechanism.
And vortex separation? Other terms have been used for these devices: swirl concentration, hydrodynamic separation, hydrodynamic vortex separation, rotary flow,teacup separation, and tangential sand traps. The descriptor swirl concentration was commonly used in the US prior to the mid-1990s, replaced by vortex separation and hydrodynamic separation, terms favored by British researchers. The USEPA used the descriptor swirl concentration in sponsored research in the late 1970s, in reference to the ability of the rotary motion to concentrate the sediment toward the bottom center. The original systems meant for the treatment of combined sewer overflows included a center well withdrawal pipe, not present in stormwater products: hence the reason for the term concentrator. In the late 1990s, the USEPA adopted vortex separation as the title of a fact sheet. Perhaps this was the beginning of the term’s widespread use in the US and Canada. Most manufacturers of the devices do not use the descriptor. Three use the descriptor swirl concentration and two use hydrodynamic separation.
The common definition of vortex is “a swirling mass of water forming a vacuum in the center, into which anything caught in the motion is drawn inward by the whirl or powerful eddy.” The intent is to complement gravity with inertial separation for enhanced removal. But with the low vortex condition at the velocities experienced in our devices, inertial separation is not a significant factor. The primary benefit of circular motion is simply improved hydraulic efficiency; the motion significantly reduces short-circuiting that occurs in round or rectangular devices without baffles. I propose that the descriptor vortex separation, while not incorrect, overstates the benefits of vortex motion and that swirl concentration is the more appropriate terminology.
Recommendations
The key recommendation is to simplify and clarify. Let’s not use descriptors that imply a process or expectation of performance that does not exist, like vortex separation. Let’s use descriptors that are as explicit as possible, such as infiltration cell rather than bioretention. Explicit terms convey directly the nature of the unit operation. We should use the more apt descriptor of swirl concentration or swirl separation in lieu of vortex separation.
Let’s not use descriptors whose definition applies so broad as to be meaningless. Such is the case with hydrodynamic separation. Its current usage, applied to a small number of manufactured products, leads to misperceptions and needless complexity. Use of the term should be limited to the narrow definition presented in this article.
We need to establish a single name for what is essentially the same unit operation, differing on some variant of the design criteria such as media specifications. An example is the amended sand filter. We should not use different names simply because of a different media specification or application. Grouping of the variants under a common name will lead to consistent sets of design criteria. The names of unit operations should be clear descriptors, conveying to the engineer the essence of the unit operation—for example, amended sand filter rather than bioretention.
Advertisement
The proposed framework is offered as a beginning point for discussion. It is hoped the framework will be adopted over time by authors of our state, provincial, and local stormwater manuals. I encourage the reader to send suggestions, additions, and criticisms of the proposed framework and use of terminology to mintonrpa@cs.com.
We need a formal, logical, and simplified framework akin to what was developed by the great Swedish scientist of the 18th century, Linnaeus, for biological species. His achievement followed several centuries of confusion and overlap of terminology. Hopefully, it won’t take us that long.
March-April 2007
A Tower of Babel
A systems-based approach toward an integrative management strategy
So just what is hydrodynamic separation? This descriptor for certain treatment products has recently proliferated. Yet no one seems to have provided a definition. The term hydrodynamic separation was initially applied to vortex separators such as the Downstream Defender and Vortechs. But this little group has expanded to include small manufactured wet vaults such as Stormceptor, BaySaver, and CrystalStream devices. And what is vortex separation?
Common to the early evolution of new fields of technical application is the multiplicity of duplicative and overlapping terms. Consolidation occurs in time with the disuse of duplicative and ill-defined terms. Perhaps this occurred with automobiles in the early 20th century and computers in the late 20th century.
In our field we have several names for essentially the same treatment system, different systems with the same name, words without apparent definition, and the misapplication of words. Does not hydrodynamic or vortex convey the impression that the labeled treatment system has some extra-special characteristic that improves its performance?
Consider swales: We have biofilters, grassy swales, vegetated swales, grass channels, landscaped swales, wetland swales, bioretention swales, dry and wet swales, and enhanced dry and wet swales. No wonder we have confusion over “what is best.” Several of these swales are sized based on the peak of a design event, while others are sized based on the volume of the design event. The latter are, in effect, basins or filters, depending on the particular system.
In contrast, we also have what are essentially the same systems but with different names. Consider the organic filter presented in some manuals and the bioretention system with underdrains (i.e., filter). The bioretention media is a mixture of 50% sand and 50% organic media, the latter a loam soil, compost, or a mixture. The media depth ranges from 18 to 48 inches. Cover is commonly a mix of grass, shrubs, and trees, although one specification is for grass only. In contrast, the organic filter has a media mixture of 50% sand and 50% organic matter, usually peat. It has an 18-inch layer of the mixed media placed above an 18-inch layer of sand. A grass cover may or may not be specified. Although the two systems are essentially identical, we have very different names for them and, more importantly, two distinctly different sizing procedures. The organic filter is sized as a filter using Darcy’s Law, whereas the bioretention filter is often sized as an infiltration system even when it has underdrains.
Currently, bioretention is a name for a treatment system. But what do we mean by bioretention? Presumably that biological organisms play a role in pollutant removal. The plants remove nutrients and metals for growth. Bacteria transform ammonia into nitrogen gas. Thus defined, bioretention also occurs in wet ponds and constructed wetlands as well. Therefore, a better use of the term bioretention is as a general descriptor (i.e. mechanism) of pollutant removal—what wastewater engineers call a unit process, a concept that will be shortly introduced and described. Bioretention thus used is akin to sedimentation, flotation, and filtration. Just as we have used the word retention to describe the holding back of water, bioretention describes the holding back of pollutants by plants and bacteria.
A New Framework
The proliferation of terminology has lead to misperceptions of performance, uncertainty as to applicability of particular treatment systems, and the inconsistent, confusing, and sometimes illogical application of design criteria. What is needed is a simplified framework with concurrence on a common name for the same treatment system to which design criteria are applied consistently. In this article I offer such a framework. It is offered as a beginning point for a dialog on this subject.
We begin with Figure 1. Figure 1 represents a hierarchical portrayal of the functioning of treatment systems as well as their organization. Basic principles serve as the foundation for the performance of all treatment systems. We may for convenience divide these into engineering, biological, and chemical principles (Minton 2005). These principles underlie how it all works. Examples are the settling of fine particles as described by Stokes’s Law and Newton’s Law, photosynthesis in plants, the chemistry of precipitation and sorption of pollutants to sediments in stormwater as well as to filter media, and the movement of water though a basin.
Principles define how unit processes work. The concept of unit processes and operations shown in Figure 1 is borrowed from wastewater and chemical engineering but simplified to meet our needs. Unit processes as defined here are the mechanisms of pollutant removal: for example, sedimentation, flotation, coagulation, filtration, denitrification, and screening. We may convey the role of biological organisms by putting the qualifier bio in front of some of these terms—for example, biofiltration, biosedimentation, and bioretention. The first example, biofiltration, refers to the removal of particulate pollutants by the filtering action of plants such as biofilms on wetland plants remove colloidal material. An example of biosedimentation is grass in a grass swale, which slows the water, allowing particles to settle to the bottom of the swale. As proposed here, bioretention refers to and is interchangeable with the term biological processes, essentially representing a group of unit processes.
Unit processes occur within unit operations, the next level in the pyramid of Figure 1. A unit operation is the treatment unit: the “box” in which one or more unit processes occurs. For example, in a sand filter itself, excluding the pretreatment unit, we have sedimentation and filtration. These are two different unit processes. Depending on the media, we may have a third unit operation, sorptive filtration with the removal of dissolved pollutants. Sorptive filtration is distinguished from classic physical or mechanical filtration.
One or more unit operations is a system. A swale usually has one unit operation, the swale itself, but it is a system as well. A sand filter as a system has two unit operations: pretreatment and the filter. A wet pond with a forebay can be said to have two unit operations because the forebay serves a particular function, which is to ease maintenance. By recognizing each element as a separate unit operation with a particular function, the engineer may consider more cost-effective solutions. Rather than a forebay in the pond, the engineer might choose a swirl concentrator. I suggest that the term system is a more apt term than treatment train. Confusion exists with the use of the descriptor treatment train. It is commonly applied to a system consisting of two or more separate structures—for example, a swale followed by a wet pond—but is not applied to two distinct treatment elements within the same structure—for example, a sand filter vault that includes both the filter bed and the pretreatment forebay. In contrast, we have the multichamber treatment train (MCTT) in which the pretreatment unit and filter bed are in the same vault. The application of the term unit operation eliminates this inconsistency. Table 1 provides my observations on the unit processes that occur in selected unit operations.
Although I have drawn the concept of unit processes from the wastewater industry, I have modified it to better suit our needs. In wastewater treatment, the term process is applied to biological and chemical processes, but the term operation is applied to physical processes. As a wastewater engineer early in my career, I found the distinction arbitrary and without explanation as to purpose or value. Rich (1961), who first fully applied the distinction to wastewater treatment, neither explained nor established its purpose or benefit. He borrowed from chemical engineering. In fact, Rich defined unit operations as physical processes and even identified some physical processes as unit processes rather than operations because of their chemical dependence (Rich 1963). Nor has the chemical engineering profession from which Rich borrowed the construct provided a rationale for the distinction.
An earlier text by engineers who were very influential in wastewater engineering defined all pollutant removal mechanisms as unit operations (Fair and Geyer 1958). The point of this brief historical trip is to establish that past categorization has been arbitrary and varying. Hence, let’s use a simplified construct that works for us. Unfortunately, the old concept by Rich has been put forth in a recent publication on stormwater treatment (WERF 2005). I highly recommend the simplified construct of unit processes and unit operations as defined here be adopted. There is no need for unnecessary complications that diminish the clarity and as a consequence the power of our framework.
The derivation of the framework should use the following principles. The name of a unit operation or system conveys its basic characteristic: for example, filter swale rather than dry swale. The name should not be so general as to be applicable to unit operations or systems other than itself. Examples that have too broad an application are bioretention and MCTT. Divisions of unit operations or systems should be made sparingly only when a clear reason is present. Certainly no framework works perfectly. Compromises will be seen in the following framework.
Families
Systems with common key characteristics are grouped into families, located at the top of the pyramid in Figure 1. Five broad families are proposed for stormwater treatment:
- Basins
- Swales
- Filters
- Infiltrators
- Screens
A figure is presented for each of the above families (Figures 2 through 6). Some of the families are further divided into subfamilies before further division to unit operations and systems. Examples of systems are given in each family. In some cases, the unit operation and system have the same name. But there is a distinction, which is illustrated with examples. As a unit operation, a sand filter is just the filter bed and its container. As a system, the sand filter is the pretreatment unit and the bed, whether they are in the same container or separate containers. As a unit operation, the grass swale is just the swale. It may also be just the swale as a system, but sometimes the swale is immediately preceded by a unit operation that removes coarse sediments and/or litter. A complication of the hierarchical framework represented in Figure 1 is that to construct a system one uses unit operations from more than one family. Notwithstanding this weakness, I have found the proposed framework to work.
Provided in Figures 2 through 6 are terms I recommend be discontinued for simplicity and clarity. Our goal is to have one name for what are essentially the same systems.
Family Basins
A basin is a family of unit operations in which water is detained, the duration differing with the type of basin. The family is illustrated in Figure 2. Treated stormwater is discharged via a weir or orifice to a pipe or ditch, eventually reaching a surface water body such as a stream or lake. A portion of the stormwater may infiltrate through the bottom of the basin, but the aggregate volume infiltrated over time is less than what is discharged to surface waters. In considering performance, credit may be given for infiltration depending on whether the interest is concentration or loading reduction.
The family is divided into three broad subfamilies: wet, wet extended detention, and extended detention. Extended detention basins become completely dry after each storm, with the lowest outlet at the bottom of the basin. The lowest outlet of a wet basin is commonly several feet above the bottom for a wet pond but only a foot or so above the bottom for a constructed wetland. The term wet is used rather than retention, a commonly used descriptor as wet contrasts with dry. Some would argue that detention contrasts with retention. However, whether the pond is wet or dry, water is detained or retained. The only difference is the duration. Retention is used by some to describe infiltration basins. But why is water detained if it discharges to a receiving water body but retained if it discharges to aquifers?
The wet extended detention basin is a hybrid of the two extremes. It is common to split the design volume between the active and dead storage, suggesting a performance between the fully wet and dry systems. The concept evolved to improve the performance of an extended detention basin. The benefits of the wet extended basin over a full wet basin have not been stated. A wet basin retains twice the volume of water between events, thereby improving the removal of dissolved pollutants and clays. Two possible benefits are provided by the hybrid: less thermal impact, and safety with its shallower permanent pool depth. These distinctions justify the division. The word pond is used in Figure 2 for extended detention rather than basin, the more commonly used term, because a basin can be a vault or pond. A distinction is made between vaults and ponds or wetlands because the latter two remove dissolved pollutants.
The wet swale is included in the basin family rather than the swale family. The name is lengthened to wet extended detention swale, although retaining the name wet swale seems reasonable. The system is placed in this family because its size is based on a design storm volume rather than flow rate. It is essentially a very shallow and long wet extended detention basin with a more substantial slope (1% to 5%) than other basins. It is proposed that this particular configuration be simply identified as a variant of the wet extended detention basin, just as the perimeter sand filter is shown as a variant of the sand filter in the filter basin.
Included in the family are manufactured vaults such as those from Stormceptor, BaySaver, and CrystalStream. These basins are grouped in the unit operation called “standard” for want of a better term. Each has internal elements to improve hydraulic efficiency and to inhibit resuspension of removed sediments. But so does a vault designed by the engineer. A wet vault need not be proprietary. As any vault designed by an engineer must include baffles, the term baffle box is dropped. It does not differ significantly from oil/grit separator vaults, for example.
The oil/water separator (OWS) is a special vault with elements to enhance the removal of petroleum-related compounds. OWS separators could be divided into three separate types of separators or unit operations: the simple vault (oil/grit separator and API separator), those using laminar or coalescing plates (CPI), and those that rely solely on sorptive media, of which there is only the EcoSep. As with many systems, a manufactured unit can be purchased or the engineer can prepare a design and purchase a coalescing plate pack. A fourth type substitutes open plastic balls for the coalescing plates.
We drop the term water-quality inlet: It is superfluous as well as misleading inasmuch as the system is not usually placed on a drainage inlet but at the site discharge point like other treatment systems. Most often these are pretreatment units. It should also be noted that the term oil/grit separator has been applied by several to vortex separators and small values like the Stormceptor and CPI separators. I propose that the term oil/grit separator be dropped, as any vault of similar size (e.g., CrystalStream) can remove oil and grit at similar levels of efficiency. That is, the oil/grit separator is no different than a common pretreatment vault with baffles. It is essentially a grossly undersized API separator. This confusion is apparent in a fact sheet by the USEPA (1999).
Swirl concentrators are also vaults but with the unique attribute of a tangential inflow, the purpose of which is to improve the hydraulic efficiency of a simple round vault—for example, Hydro International’s Downstream Defender, Flo-Gard’s Dual Vortex,and AquaShield’s Aqua-Swirl. The descriptors vortex separation and hydrodynamic separation are not used for reasons given at the end of this article. It is recommended that these terms be dropped as general descriptors.
There are several wetland configurations. We are familiar with the marsh and micropool wetlands, the later essentially an extended detention basin with small wet pools at the entrance and exit. The hummock wetland is an Australian development consisting of alternating deep open-water pools and lateral shallow benches with wetland vegetation. This configuration has been found to exhibit improved hydraulic efficiency, better nitrogen removal, and lower mosquito populations than the other configurations and is therefore identified as a distinctly different unit operation. A wet pond-wetland has large areas of deeper water typically in the center. The distinction between this configuration and a small wet pond with a safety bench covered by wetland vegetation is admittedly obscure. A system not presented in Figure 2 is the pocket wetland, identified for relatively small drainages. It is not included as it can be considered as simply a small version of any of the other configurations that are shown.
The subsurface flow (SSF) wetland differs from the others in that the substrate is gravel, thereby allowing the stormwater to flow through the root system. This concept is adopted from wastewater treatment, where it is also called the SSF wetland. This term is proposed in lieu of the descriptor gravel wetland, which some are using. The engineer can design the SSF as an earthen basin or use the StormTreat, the one manufactured system that uses the concept. SSF wetlands have benefits with respect to avoidance of mosquitoes and are possibly less susceptible to freezing.
Note that in Figure 2 the same term is often used for both the unit operation and the system. The distinction is this: A pond or vault identified as a unit operation is just the single unit with no pretreatment element, for example. As a system, a pond or vault may contain a forebay. I view this as a separate unit operation as it has a distinct purpose apart from the pond, as previously described.
Family Swales
Swales (Figure 3) are unit operations with a distinct slope to facilitate the transit of stormwater from entrance to exit and very shallow water depths, usually a few inches at the maximum design depth. The majority of the treated stormwater is discharged to a surface water body rather than to groundwater. As with basins, infiltration may occur, but the volume is less than what is discharged to surface waters. In considering performance, credit may be given for infiltration depending on whether the interest is concentration or loading reduction. All unit operations in this family are sized to treat up to the peak flow of the design storm.
We have two subfamilies: swales and strips. With swales, the stormwater transverses the length, whereas with strips, the stormwater transverses its width. The former has concentrated flow whereas the later has sheet flow.
All swales and strips are vegetated given that a bare soil is to be avoided. Swales are categorized according to the type of vegetation: grass, wetland, and landscape. The same is true for strips. As implied by the name, a grass swale is covered by grass; a wetland swale is covered by wetland plants; and a landscape swale has a mix of grass, shrubs, and possibly trees. A wetland swale or strip has a very gradual slope, 1% or less. As a consequence, and depending on the soil type and climate, the soil remains wet throughout much of the year, thereby supporting wetland plant species.
The distinction between a landscape strip and a buffer is ground topography. The ground is leveled for a landscape strip, generally in a commercial development where strips are most commonly used. The buffer retains its original topography and is an integrated component of a residential development that abuts a stream, wetland, or lake.
As with the basin family, the same terms are sometimes used for both the unit operation and the system. But as with a pond or a vault, a swale as a unit operation is just the swale. As a system it may include a pretreatment unit to ease maintenance.
Several terms are dropped from current usage including biofilter, vegetated swale, grass channel, and bioretention swale. The dry swale is not included in this family, but rather with the filter family, as filtration is the dominant unit process.
Family Filters
A filter is a unit operation with these characteristics: underdrains, engineered filter media, and the majority of the treated stormwater discharging to a surface water body—that is, it is not infiltrated to the groundwater aquifer. Types of media include sand, peat, leaf compost, perlite, zeolite, fabric, metal-oxide–coated sand, and loamy soil. The unit may have only one medium or be multimedia, either as layers or mixed. The family is illustrated in Figure 4.
The family is divided into fine- and coarse-media filters. The distinction is made because the latter provides lesser but in most cases acceptable performance, and because the designs differ significantly. Fine media generally refers to sand filters, for which ASTM C33 fine concrete aggregate is commonly specified. A substantial head is required to obtain desired filtration rates to minimize the surface area of the filter. Coarse-media filters are commonly manufactured. The beds are thinner and the detention time is minutes in contrast to hours for fine-media filters, whose beds are usually thicker.
Fine- and coarse-media filters can be distinguished by the rate of filtration per unit area. The throughput of fine-media filters is about 0.02 gallon per square foot (0.06 liter per second per square meter), whereas it is 2 gallons per minute per square foot (6 liters per second per square meter) or greater for the coarse-media filters. It follows that coarse-media filters require a much smaller footprint. Examples of coarse-media filters are presented in Figure 4.
While most manufactured filters commonly use coarse media, the BayFilter has sand like the public-domain sand filter. The StormFilter also uses fine sand in some applications and therefore is found under both fine and coarse media in Figure 4. Some manufacturers offer media finer than their standard coarse media, although it is still coarser than what is used in sand filters. The dividing line between fine and coarse remains to be defined.
A third subfamily is catch basins or drain inlet inserts, for which there are many manufacturers.
Sorptive media listed in Figure 4 can be mixed with sand to remove dissolved pollutants, in particular metals or dissolved phosphorus. For this distinction, these filters are placed in their own subfamily of amended sand filters.
Bioretention is an amended sand filter and therefore is placed with this group. Most significantly, the term is dropped as a filter name. As noted previously, its media composition is not dissimilar to the older organic filter. The term multichamber treatment train is dropped, as it also refers to an amended sand filter with a composition similar to bioretention units. The name multichamber treatment does not meet two of our criteria: that names indicate the nature of the system, and that the name not be applicable to other treatment systems with the same characteristics. The descriptor multichamber is applicable to many other treatment systems. A wet pond with a forebay is a multichamber, as is a sand filter with its pretreatment unit. Our goal is to use names that most explicitly convey the primary nature of the system, which, for a MCTT, is amended sand filter.
Filterra is classified as an amended sand filter paralleling the recommendation that the term bioretention not be used as a unit operation. The media of the Filterra is a mixture of sand and gravel, with the surface mulch essentially an amendment. Its composition is between fine and coarse media. But as sand and mulch control the flow rate, it is placed in the fine-media grouping.
The same design parameters should be used to determine the surface area of the bed for the sand and amended sand filters (Minton 2005). The common design parameters are hydraulic conductivity, maximum drawdown time, and operating head. A different design procedure is currently used for the MCTT, particularly with respect to the pretreatment unit. The pretreatment unit is much larger than what is used with the standard sand or amended filter. The objective, in combination with laminar plates, is to extend the time before cleaning of the bed surface must occur, from perhaps annually to several years. Perhaps this approach should be adopted for all fine-media filters. But these differences are not sufficient to warrant its unique name or identification as something other than an amended sand filter.
It may be desirable to distinguish surface filters by the absence or intentional presence of vegetation. Although not common, sand filters have been intentionally designed with a grass cover. Anecdotal information indicates grass extends the maintenance cycle. This also appears to be the case with the bioretention filter. The design hydraulic conductivity may differ depending on whether the surface is vegetated. Information is lacking at this time to make the distinction.
Sand and amended sand filters are divided into several system types such as basin, lineal, and cell. It could be argued that basin and cell filters are the same thing, but a lineal filter is also in effect a basin. But the separation is made to accommodate current usage and to differentiate systems by size, configuration, and usage. The use of the descriptor basin with this family conflicts with its use in the basin family but is unavoidable. Vessel and container are unappealing terms.
Included in the filter family is the filter swale, placed here rather than with the swale family because its primary mode of treatment is filtration. I suggest that the proposed terminology is more explicit than dry swale and bioretention swale.
Family Infiltrators
Infiltrators are all unit operations in which the design water volume is infiltrated into the soil, with the ultimate receptor being the groundwater aquifer. The family is illustrated in Figure 5.
In this family, we have these unit operations: basins, trenches, swales, vaults, cells, and the older dry wells. What we now call bioretention becomes an infiltration cell. No doubt the descriptor rain garden will continue to be used, given its appeal to landscape architects and homeowners. The descriptor infiltration cell does not have the same “zing” as rain garden when promoting low-impact development (LID) site design. The dry swale is an infiltration swale if it does not have underdrains.
Note that the descriptor infiltration is put in front of many of the systems to distinguish them from their noninfiltrating cousins. It is a bit cumbersome, however. Infiltrator vaults are typically manufactured and assembled onsite. Figure 5 lists some examples. These products may also be in the basin family as they are wet or extended detention vaults where the soil has inadequate infiltration capabilities.
Included in the infiltrator family is the infiltration swale, placed here rather than with the swale family because its primary mode of treatment is infiltration. In this case we replace the terminology of bioinfiltration (as a system), dry swale, and bioretention swale, the latter used here where soil has a sufficient infiltration rate to obviate the need for underdrains (filtration swale).
The terms bioretention and bioinfiltration are dropped as descriptors of particular systems but become descriptors of unit processes.
Shall we use the word porous or permeable or pervious? All three are widely used, but porous appears to be the more commonly used descriptor. They have the same common definition. I have no preference, but the descriptor porous is used in Figure 5. Open-cell paver is a better descriptor than turf cell, because the cell need not contain turf. The term cementious is a bit academic and has little meaning to the practicing engineer.
As with filters, we might distinguish infiltrators by the absence or intentional presence of vegetation.
A few manuals include partial infiltration. This refers to systems that infiltrate the majority but not all of the design water-quality volume due to low infiltration rates. Some of the design-storm volume flows to surface discharge. The distinction between a partial infiltration system and its counterpart in the other families is vague and derives primarily from intent. The distinction is that with partial infiltration systems the majority of the stormwater is infiltrated, but less than the design-storm volume. This is in contrast to the basin and swale families in which infiltration is considered coincidental, defined as ranging from minor to less than half of the design-storm volume.
Family Screens
The screen family illustrated in Figure 6 consists of unit operations with large openings. Their primary if not sole intent is to remove gross solids such as litter, leaves, and plastics. Almost all of the systems described with the first four families also remove gross solids but are not primarily intended to do so. However, some may perform as well as some devices found in the screen family—for example, some of the drain inlet inserts listed under the filter family. The majority of screens are manufactured. Of course, trash racks used as elements in systems in the other families can be made with the basic materials of concrete and steel.
This family illustrates the difficulty with the placement of some of the manufactured products: in this case the CDS Technologies continuous deflective separation unit. It is not placed with the swirl concentrators in the basin family because its original purpose was litter control. It is offered as a sediment removal device, however.
Closing Observations
So what is hydrodynamic separation? The terminology is used by chemical and mining engineers and in medicine. Their common definition of hydrodynamic separation is the removal of particles by gravity from a moving fluid. The fluid can be air or water. The particles may be more or less dense than the fluid. An alternative term is density separation. Chemical engineers also narrowly define hydrodynamic separation as removal of a particle from a fluid that experiences a very abrupt change in its direction of flow. If we apply the first definition to stormwater treatment, which seems most appropriate, wet ponds and constructed wetlands are hydrodynamic separators, as are extended detention basins and filters—yes, filters, because most of the removal is by sedimentation on and within the filter media. Hydrodynamic separation applies to flow-through grass swales. In effect, the term is meaningless. I conclude that the descriptor hydrodynamic separation should not be used to describe the small group of treatment systems to which it is currently applied.
The descriptor hydrodynamic separation could apply to the more narrow definition of removal by an abrupt direction change in flow, but only if it is the dominant removal mechanism. However, this unit process is not generally employed in our systems. A possible exception is the Downstream Defender, in which there is an abrupt change in direction as water flows from the outer chamber into the center beneath the dip plate. Although studies indicate the dip plate improves performance, it is not clear that it is due to the directional change. Regardless, it is not the dominant removal mechanism.
And vortex separation? Other terms have been used for these devices: swirl concentration, hydrodynamic separation, hydrodynamic vortex separation, rotary flow,teacup separation, and tangential sand traps. The descriptor swirl concentration was commonly used in the US prior to the mid-1990s, replaced by vortex separation and hydrodynamic separation, terms favored by British researchers. The USEPA used the descriptor swirl concentration in sponsored research in the late 1970s, in reference to the ability of the rotary motion to concentrate the sediment toward the bottom center. The original systems meant for the treatment of combined sewer overflows included a center well withdrawal pipe, not present in stormwater products: hence the reason for the term concentrator. In the late 1990s, the USEPA adopted vortex separation as the title of a fact sheet. Perhaps this was the beginning of the term’s widespread use in the US and Canada. Most manufacturers of the devices do not use the descriptor. Three use the descriptor swirl concentration and two use hydrodynamic separation.
The common definition of vortex is “a swirling mass of water forming a vacuum in the center, into which anything caught in the motion is drawn inward by the whirl or powerful eddy.” The intent is to complement gravity with inertial separation for enhanced removal. But with the low vortex condition at the velocities experienced in our devices, inertial separation is not a significant factor. The primary benefit of circular motion is simply improved hydraulic efficiency; the motion significantly reduces short-circuiting that occurs in round or rectangular devices without baffles. I propose that the descriptor vortex separation, while not incorrect, overstates the benefits of vortex motion and that swirl concentration is the more appropriate terminology.
Recommendations
The key recommendation is to simplify and clarify. Let’s not use descriptors that imply a process or expectation of performance that does not exist, like vortex separation. Let’s use descriptors that are as explicit as possible, such as infiltration cell rather than bioretention. Explicit terms convey directly the nature of the unit operation. We should use the more apt descriptor of swirl concentration or swirl separation in lieu of vortex separation.
Let’s not use descriptors whose definition applies so broad as to be meaningless. Such is the case with hydrodynamic separation. Its current usage, applied to a small number of manufactured products, leads to misperceptions and needless complexity. Use of the term should be limited to the narrow definition presented in this article.
We need to establish a single name for what is essentially the same unit operation, differing on some variant of the design criteria such as media specifications. An example is the amended sand filter. We should not use different names simply because of a different media specification or application. Grouping of the variants under a common name will lead to consistent sets of design criteria. The names of unit operations should be clear descriptors, conveying to the engineer the essence of the unit operation—for example, amended sand filter rather than bioretention.
The proposed framework is offered as a beginning point for discussion. It is hoped the framework will be adopted over time by authors of our state, provincial, and local stormwater manuals. I encourage the reader to send suggestions, additions, and criticisms of the proposed framework and use of terminology to mintonrpa@cs.com.
We need a formal, logical, and simplified framework akin to what was developed by the great Swedish scientist of the 18th century, Linnaeus, for biological species. His achievement followed several centuries of confusion and overlap of terminology. Hopefully, it won’t take us that long.