Last June the City of Austin ran a water conservation conference, and among the presentations was a review of Austin Water’s plans to incentivize, or perhaps require, building-scale “waste” water treatment and reuse for all projects housing over 250,000 sq. ft. of floorspace. As that presentation was winding down, the current assistant director of Austin’s Watershed Protection Department came by and whispered, “You were ahead of your time.”

He and I had met in 1986, exactly because I had just written “The Decentralized Concept of ‘Waste’ Water Management”, the first of many versions of that and similar works setting forth the idea that, by organizing the system to treat – and reuse to the maximum extent practical – the “waste” water as close to its source as practical, we would produce an infrastructure model that would be more fiscally reasonable, more societally responsible and more environmentally benign. He was in a business at that time, selling building-scale wastewater treatment units, for whom that basic idea was rather central. As he moved on through stints at various agencies and consulting firms, somewhere along the way he seemed to have “lost” that vision. I recall a conversation on his back patio, about 10 years ago when he was with a mainstreamer national consulting engineering firm, when we discussed how cities would manage water, he opined that my “vision” of decentralizing down to the building or campus scale would never be embraced, rather cities would always stay with the conventional centralized, pipe-it-“away” scheme. So it was, I can only guess, he felt a mea culpa moment, as some version of the very vision I’ve espoused these last three decades was being displayed in front of us as the direction Austin Water is moving.

However … As the presentations, and subsequent discussions with Austin Water folks, made clear, the means by which they expect to implement building-scale reuse is by using the tools of conventional centralized systems. Most particularly the inherently unstable activated sludge treatment technology, which is practically the “knee-jerk” choice of the mainstreamers, pretty much because it is deemed the “reasonable” choice in conventional centralized systems. Which brings us to the idea of appropriate technology for the scale of the system.

A central tenet of the decentralized concept, set forth in that original 1986 paper, is that the nature of the technologies used to assemble the system should recognize that, with distributed systems, there would be many more treatment units to police, so to minimize the total O&M liability, these systems would need to employ “fail-safe” technology. As I set forth in those early writings on this subject, there is a difference between “fail-safe” and “reliability”. A system can be reliable if it has the capability, when properly operated and maintained, to consistently and reliably produce the advertized effluent quality. But “fail-safe” means that the inherent nature of the technology is such that it can maintain reliability in the face of non-optimal operating conditions, because the technology is robust, inherently resistant to “upsets”.

Activated sludge technology is inherently unstable because it depends for its treatment action on very few trophic levels of microorganisms living in concentrations far higher than found anywhere in nature (a trophic level is a rung on the food chain—organisms on a higher trophic level eat organisms on a lower trophic level), thus it is a very truncated ecology that is not inherently sustainable.  The process can only be kept “on track” by maintaining proper operating conditions with constant inputs of energy to aerate the wastewater and monitoring the process to maintain a proper food/microorganism (F/M) ratio. Typically maintaining the F/M ratio requires frequent withdrawals of sludge from the system, on a time scale measured in hours. So failure to pay close enough attention leads to an “upset” in very short order. Therefore, while the process can be reliable, as long as proper operating conditions are maintained, it is not “fail-safe” because it is so sensitive to adverse conditions. Such a process is not really what you want to depend upon in a context like a building-scale reuse system.

So for highly distributed systems, like these building-scale reuse systems, we need to be using “fail-safe” technologies. Before proceeding, note that I always put “fail-safe” in quotes. Nothing is ever completely fail-safe. No matter how robust a technology may be, it will always require proper operation and maintenance if it is to be expected to continue to perform reliably. Again, there are certain technologies that, by dint of their very nature, are rather more immune to adverse conditions than the inherently unstable activated sludge technology, and for which the timing of O&M procedures is not so critical.

While various versions of constructed wetland technology may have merit – the Hassolo on Eighth project in Portland, Oregon, is an example of a project-scale reuse system that employs this technology – for my money the recirculating “sand” filter should be the “workhorse” technology of the decentralized concept. I put sand in quotes, because while the original version of this basic technology did use sand media in the filter beds, modern versions of it use “packed beds” of gravel media, geotextile fabric media, Styrofoam bead media, foam rubber media, etc. To cover all the different media that might be used, a more generic name for this technology is recirculating packed-bed filter.

Reasons why the recirculating packed-bed filter technology is inherently “fail-safe” include:

  • This technology is an attached growth, rather than suspended growth, concept, with the treatment effect accomplished by organisms attached to the filter media, harvesting food from the pollution in the water as it flows on by. Attached growth is far less prone to “wash out” than suspended growth, so the treatment effect is inherently much more robust and stable.
  • The loading rates on recirculating packed-bed filters are quite low, on the basis of microorganism “density” relative to the food source – this imparts a high mean cell residence time in the system – which renders the process more resistant to “upsets” and so enhances the stability of the treatment process. Again it is quite robust.
  • Power is not required to maintain the treatment process. Rather power is only needed to move the “waste” water to the top of the filter bed, and the actual treatment process is passive, imparted as the water flows down through the media by gravity. So, in sharp contrast to the activated sludge process, loss of power does not result in loss of the treatment process. If a power outage were to occur, the biota would sit there, waiting for the flow to resume, with no impact on treatment quality.
  • Flow equalization is inherent in the treatment concept, with the filter beds being loaded at the same hydraulic application rate on the same schedule every day, without regard to how much or how little flow enters the system on any given day. This hydraulic steady state operation renders the process highly consistent and reliable. This is particularly important in buildings with the occupancy patterns of commercial and institutional buildings, with high activity during the day, on weekdays, and little through the night and on weekends.
  • The biology of the system is quite diverse, typically including many trophic levels of microorganisms, and some macroorganisms as well. This characteristic also renders the process inherently resistant to upsets, allowing it to readily accommodate situations where system loading is highly non-uniform, as it will be in this circumstance.
  • The only moving parts are the pumps that dose the filter beds and a passive valve that operates on water level. Again, loss of pump power over the “short term” would have no impact on the treatment process, and a malfunction of the valve can be accommodated for some time before the treatment process may be impacted, allowing the operator to fix the valve essentially at his/her leisure.
  • The pumps are installed inside sealed tanks, setting under water, so would impart no noise pollution.
  • The only odor production might be imparted as the water is distributed over the filter beds. These units are sufficiently well covered so that odors would not be obtrusive.
  • Sludge management is very unobtrusive. The major mode of sludge management is pumping the septic tanks that are the “front end” of the treatment unit. This is typically only required at multi-year intervals and is not time-critical – months could pass between observation of sludge level in the septic tank indicating pumping is needed and the pumping actually being executed without any significant impact on the treatment process.
  • The system, once set up, basically “operates itself” day-to-day. There is nothing to adjust, and only infrequent routine maintenance is required.
  • Operations and maintenance activities for this system are rather simple and straightforward. They can readily be conducted by personnel with minimal training. As long as the control system components are not themselves proprietary, it does not rely upon any one vendor for this service.
  • The major “failure” mode of this technology is clogging of the filter bed. This occurs very slowly, allowing time for the operator to respond essentially at his/her leisure. With insightful design, filter bed clogging can be remediated in very short order.

My approach to building-scale, or campus scale, treatment and reuse was informed by learning from Takashi Asano back about 1990 that the California Title 22 reuse rules basically specified a system composed of a “waste” water treatment unit followed by a water treatment unit, to produce very high quality effluent for “unrestricted” reuse, such as for toilet flush water. Of course, being mainstreamers, the folks who wrote those rules set the water treatment requirements in terms of conventional water treatment plants, entailing coagulation-sedimentation-filtration. But it immediately occurred to me that for small-scale implementation of this concept, perhaps the slow sand filter should be the water treatment plant part of the scheme, for similar reasons that the recirculating packed-bed filter is favored for the “waste” water treatment part of the system.

While sand filtration had been used for water treatment for centuries, the slow sand filter concept as we know it today was first used in Scotland in 1804, and was first implemented for public water supply in London in 1829. It became widely adopted – it was first used in the U.S. in Poughkeepsie, New York, in 1872 – and despite the introduction of more “modern” water treatment processes, it continues to be used for municipal water treatment, including by many large cities, particularly in Europe. Like the recirculating packed-bed filter technology, the slow sand filter is “low tech”, being rather simple to operate and maintain, rather robust and largely “passive”. Indeed, it is these characteristics that make it the go-to water treatment technology in “third world” settings, where operating capabilities may be quite limited. That, of course, also makes it a great choice for distributed systems.

So it is that I suggest that the “standard” treatment unit for building-scale or campus scale reuse projects be composed of a recirculating packed-bed filter for basic “waste” water treatment, followed by a slow sand filter, to produce a near-potable quality water. UV disinfection of the treated water completes the system. A schematic of this sort of system is shown below. Indeed, one of the presentations, by Amelia Luna of Sherwood Design Engineers, at the Austin water conservation confab last June noted this as a good candidate for this duty. Notably, system concepts she highlighted do use the recirculating packed-bed filter unit as the “waste” water treatment portion of the system.RBPF-SSF TREATMENT UNIT

Besides its inherent “fail-safe” nature, this “low-tech” recirculating packed-bed filter/slow sand filter treatment concept would entail significantly less energy use to run it, imparting a much lower carbon footprint. The blowers in an activated sludge plant run 24/7/365, and in the versions using a membrane rather than a conventional clarifier to produce the final effluent – the version certain to be used in a building-scale system – quite a bit of power is also required to force the water through the membrane. By contrast, the pumps in the “low-tech” unit run only intermittently, needing to impart only a modest lift of the water to the tops of the filter beds, so drawing far less power.

It is also quite likely that the installed cost of the recirculating packed-bed filter/slow sand filter unit would be somewhat lower than for the activated sludge unit of the same capacity. Confirming this awaits an opportunity to design a unit for an actual application, but the cost factors seem to favor the “low-tech” unit.

Because there is not a whole lot of economy of scale for installed costs of the “low-tech” facilities, the cost per gallon for a 1,000 gallon/day (gpd) unit would not be greatly increased over the cost per gallon for a 5,000 gpd unit, so this scheme could be just as readily used for smaller projects as for buildings having 250,000 sq. ft. of floorspace. For example, a 250,000 sq. ft. office building might house 1,250 persons, and a 50,000 sq. ft. office building might house 200 persons. The Texas on-site wastewater rules set forth a design flow rate criteria for office buildings of 5 gallons per person per day, imparting a design flow rate of 1,000 gpd for the 50,000 sq.ft. building, and 6,250 gpd for the 250,000 sq. ft. building. If the costs do scale fairly uniformly over such a range of design flow rate, the building-scale reuse scheme might be just about as cost efficient for the smaller building as it is for the larger building. This would make it feasible to cover a much larger segment of the commercial-institutional building market with project-scale reuse than just the “big box” buildings.

Indeed, we have the opportunity here to create Zero Net Water commercial and institutional buildings and campuses. With the water use “intensity” in these buildings – that is, the amount of water demanded relative to the size of the building – such buildings would have adequate roofprint so that building-scale rainwater harvesting (RWH) could provide the “original” water supply, for lavatories and building grounds irrigation, while the building-scale reuse system supplies the flush water. Again, this appears feasible for buildings much smaller than 250,000 sq. ft.

This strategy would be especially beneficial for managing the “nodal densification” proposed by the “Imagine Austin” plan. This suggests that various properties within already urbanized areas would be redeveloped at higher activity levels, or “density”. Implicit in this is that more water supply would have to provided for that “node” and more “waste” water would be generated there. If managed conventionally, this would surely require upsizing water and wastewater lines in that area, which would probably entail “upgrading” the existing lines. For example, the densification of Austin’s downtown area required the installation of a 60-foot deep tunnel to pipe the increased flow of “waste” water “away”. That is all expensive and disruptive. Meeting these increased demands instead with a Zero Net Water strategy – RWH for water supply and decentralized concept “waste” water reuse systems – would obviate all that. Instead of importing more water and draining “away” more “waste” water, the water supply would be gathered and looped around within the “node”, to serve a building or campus of buildings. Then too, growth would be adequately served without increasing demands on our conventional water supplies, which are becoming increasingly strained in this region. RWH would also reduce stormwater management issues in the denser development.

So to summarize, Austin Water should seriously consider appropriate technology as it determines if and how to incentivize, or require, building-scale reuse in commercial and institutional buildings. They should consider implications for cost and reliability. By pursuing this function with systems composed of appropriate technologies, it is more likely that the concept would be more trouble-free, and thus more widely accepted, even embraced, as it becomes clear that we can save water and money by engaging in the fundamental transformation of the form and function of our water resources infrastructure that the decentralized concept strategy accomplishes, creating a system that is more fiscally reasonable, more societally responsible, and more environmentally benign.

Explore posts in the same categories: Uncategorized


  1. Paul W Lawrence Says:


    You are one of the fortunate few to have lived long enough for the world to catchup with your ahead-of-your-time ideas. Bravo. I hope it all pans out in a robust fashion.


What do you think about this?

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: