While many applaud the fresh breeze of Green Infrastructure (GI) that is now proposed to become part of the Consent Decree for Washington, DC, many also question how well will it work? And how well will it work not just for the day it is built, but for the next year, ten years later, and ten years after that…
This question seems to be the main challenge to GI. And fairly so in my view – can GI be constructed to yield performance benefits that are equal to the tunnels GI would replace. Tunnels have very clear performance in year 1 and year 100 – capturing and conveying huge quantities of stormwater to a treatment facility. Stormwater that would otherwise flood untreated sewage and other wastes and refuse from city streets into our rivers and streams during major rain events.
We can therefore calculate the improvement by evaluating the water quality in these bodies of water after removing this existing source. We can calculate with precision exactly how much overflow will be captured, and at what point even huge tunnels will be at over capacity. This is a significant attribute of the tunnels or any remedy by the way – no matter how big you build them, there can still be a storm big enough to fill them and cause an overflow. There is no perfect solution.
Starting in 2009, DC Water wanted to eliminate or reduce the two proposed tunnels to capture overflow – one reasonably large tunnel along the Potomac and a smaller tunnel planned for the Piney Branch, a tributary of the Rock Creek, which itself is a tributary of the Potomac. Our first proposal, which we floated in 2010, was to extend the time lines in the Consent Decree for eight years to give us time to fully evaluate GI. At the end of the eight years, we would propose the alternate solution or return to the tunnels. Of course, if it was clear that GI would not work earlier than 8 years, we would return to the tunnel remedy even sooner.
Many were not impressed. There are many fans and advocates for GI – but waiting eight years just to get through the evaluation process and to a proposal, which in itself would have a multi-year implementation schedule, was simply too much time to wait. Under the current decree, all three tunnels are to be finished by 2025 – now just a decade away. The alternative put that deadline too far in the future, and too much in doubt until the evaluation was done. Timing is obviously a critical issue, and I will write more about that in the next post.
The other key question, regardless of timing, is the performance of GI. Can GI capture enough rainfall at the surface so the number and quantity of overflows to the Potomac are reduced in a fashion similar to the tunnels? The word we were using at first was to achieve “comparable” results. Environmental activists based in Washington (Of whom I consider myself one), wanted “equivalent” result.
My charge to our superb engineering team was to approach GI with the same level of engineering skill that is devoted to gray infrastructure. The team dug in – developing a series of technical memoranda reviewing the latest literature and analytical data on building, maintaining and measuring GI. Some of these technical memorandums can be found in the Full Green Infrastructure Modification Document.
• Appendix F: ‘Technical Memorandum No.2–Approach to Hydrologic and Hydraulic Modeling’
• Appendix G: ‘Technical Memorandum No.4–The District of Columbia’s Experience with Green Infrastructure’
• Appendix H: ‘Technical Memorandum No.5–Green Infrastructure Experience – Foreign and Domestic Case Studies’
• Appendix I: ‘Technical Memorandum No.6–Green Infrastructure Technologies’
• Appendix J: ‘Technical Memorandum No.7–Green Infrastructure Screening for the Potomac River and Rock Creek’
The first reality we faced was an initial answer “no.” To be blunt, our engineering team, led on this project by our fantastic Chief Engineer Leonard Benson and Carlton Ray, the talented Director of our Clean Rivers Program (the name for the project to nearly eliminate combined sewer overflows to our waterways) – concluded that after careful evaluation, DC Water simply would not be able to achieve comparable, much less equivalent, performance.
At least for the overflows directly to the Potomac River – because we simply did not have enough land or space to build GI in the cramped areas of Georgetown and Upper Northwest Washington DC to capture enough of the rainwater at the surface. The planned tunnel for the Potomac River was both huge (34 feet in internal diameter) and reasonably long – from North of Key Bridge to South of the Kennedy Center. It was planned to function like a huge underground cistern, and we had to build a complicated underground pumping station to push all the stormwater that would gush into it back to the surface and into existing sewers after the storm surge was past.
Some asked how we developed stormwater capture rate estimates for GI. That effort starts with GIS mapping (aerial, computerized maps) of the neighborhoods that would drain stormwater into the tunnels. From the GIS maps, we assessed every parcel that could be modified to improve its capacity to infiltrate and capture stormwater on-site. Ironically, the best opportunity for capture is created by past development that featured impervious cover e.g. parking lots, hard blacktop and other similar surfaces. By identifying those areas and replacing it with pervious cover (permeable pavement and the like), or adding bioretention and other amenities to capture the rainwater flowing off the site – we could “gain” stormwater capture foot by foot and lot by lot.
We’ve used models of the land surface and sewer lines in the District to predict how much stormwater needs to be removed using GI to achieve the CSO control needed. The models predict how much water runs off from the various surfaces in the District, the amount that infiltrates, and the amount that ponds in low areas. GI features are added to the model to represent how runoff is stored or infiltrates in the GI practices such as bioretention cells or porous pavement. The amount of runoff that needs to be removed from the system to provide the degree of CSO control required can be calculated using the model, and the amount of GI needed to control this amount of runoff can then be estimated based on the capacity of the types of GI practices anticipated in the sewersheds.
Our best modeling indicated we would be able to find enough land that drains to the Piney Branch tunnel to be able to eliminate the tunnel entirely with the use of GI. Our modeling indicated we could not for the Potomac – the quantities of stormwater to capture are just too great.
That did not stop the team! They offered the first major revision to the plan with several bold steps. First, we would eliminate only half the Potomac Tunnel and redesign the part that remained. This would guarantee significant capture of volume in the tunnel. Second, we would separate the combined sewer into storm and sewer lines in a small part of the area that drains to the Potomac – so that stormwater would no longer cause overflows in the first place. Third, we would eliminate the northern stretch of the tunnel and use GI to capture a substantially reduced volume of stormwater. Again, we had done the GIS analysis and hydrologic and hydraulic modeling and have identified the parcels necessary to achieve this outcome. Still included was a deep tunnel pump station to lift the combined flow from the smaller, but still huge, tunnel back to existing sewers. And we asked for a 7-year extension to get the work done – until 2032.
This is the proposal we announced to the public in January of 2014 and were confident that it could achieve comparable results to the tunnel-only remedy.
After a 90-day comment period, DC Water received nearly 500 comments. While the vast majority were positive, there were still challenges identified: need for equivalent performance not just comparable; concern over the 7 year extension; concern over the viability of the land needed and the deep tunnel pump station.
By the way, a concern DC Water has had throughout was the construction of the tunnel in any design. The level of disruption that is needed for this tunnel is dramatic – connected to both the tunnel construction itself and the drop shafts that are needed to convey the combined sewage that would otherwise overflow into the river, down into the tunnel – not to mention the huge pump station to get the combined slop back to the surface sewers. Anyone who knows the waterfront in Georgetown – incredibly expensive, historic and teeming with activity – understands intuitively how difficult this project would be to pull off.
The National Park Service, responsible for much of the land where we would need to build, including a recently finished and spectacular Georgetown Waterfront Park along the Potomac, is very concerned. They require an Environmental Impact Statement under NEPA before we can build anything. There is a reasonably good chance that the huge tunnel will not make it through that review.
Once again, our crack engineering team went to work. How could we reduce the time frame of the project, improve the performance, and reduce the complexity of the resulting work.
I remember again the next dramatic meeting in my office when my team presented the next iteration of the plan. In short, we would maintain key elements of the January 2014 plan – eliminating the northern portion of the tunnel and separating a portion of the lines. Then they described the even cooler part.
The underground pump station would be eliminated. The planned tunnel would be lengthened considerably. In fact, long enough to enable it to cross the Potomac and connect with the huge Anacostia Tunnel we are building on the other side of the River and the other side of Town.
Lady Bird TBM just wrapped up her work and will be lifted out of the Anacostia Tunnel in the next few weeks. Gray infrastructure performance is very clear from Year 1 to Year 100 – capturing and conveying huge quantities of stormwater to a treatment facility.
What? Yes – by connecting a much longer Potomac Tunnel directly to the Anacostia Tunnel, the captured combined sewage would now have a place to go, and once connected to the Anacostia Tunnel, will be conveyed all the way down to Blue Plains for treatment. Improved overall capture, no need for a pump station – gravity would convey the flow. Not only vastly reduced complexity, but vastly reduced maintenance costs and a huge reduction to our power bill in a storm.
Plus we can get this entire package done in five additional years rather than seven. Five additional years yields a 25 year remedy – which is precisely the length of time almost all Consent Decrees are currently designed to achieve.
This analysis supports my view that the debate today is not about grey versus green infrastructure – but in fact what is the best combination of both – that should be adopted by any particular community. Now, after years of trial and error design and modeling, and meaningful and helpful public input, we think we have got it right for Washington, DC!
I offer my profound congratulations and respect to the crack team that did the work. This evolution demonstrates engineering skill at is best – and renders silly the refrain from some that engineers are not innovative and creative. Our engineers have been innovative from the onset – re-conceiving, redesigning, recreating a remedy that will both improve our waterways, improve the quality of life in our neighborhoods, and enable DC Water to help lead the nation in Green Infrastructure.
My hat is off to the entire team, led by the following people, but – as the above photo clearly shows – supported by so many others: Leonard Benson (Chief Engineer), Carlton Ray (Director, Clean Rivers Project), Bethany Bezak, (Manager, Green Infrastructure), John Cassidy (Consulting Engineer, Greeley and Hansen).