Jim Urban, FASLA, is an expert on urban trees and soils — as well as being intimately involved in the development and design of the Silva Cell system. Author of the industry bible on urban forestry, Up By the Roots, Urban has also proven to be a fountain of wisdom for various inquiries about trees, soil, bioretention, and Silva Cells throughout the years. Below you’ll find a selection of his most helpful Q&A exchanges.

Are Silva Cells limited in their volume of stormwater retention?

Silva Cells, as with all suspended pavements approaches, were originally designed for trees, but it was quickly realized that adding stormwater to the soil would be a benefit to both the tree and the problem of stormwater management.

I do not see many designers actually starting with the drainage area and volume of water required to be treated and working forwards to determine how much treatment volume is needed. One of the reasons is that if we actually do this, often other limitations such as funding, or space for the system, or conflicts with other structures, set the limit of the size of system that can be designed. So, what is needed is to find the most efficient treatment system that takes up the least amount of space. Silva Cells are often part of that answer.

In a stormwater application do Silva Cells require the soil in the planting space to be set low next to the pavement?

No, this is not correct.

The distance from the pavement to the top of the soil in the area that receives the water is set by the hydraulic requirements to get the water into the area with enough head and flat space to allow the water to properly infiltrate into the soil. The Silva Cells can be and most often are lower than that elevation and function properly. The systems where the elevation of the receiving soil area is very low is most often the result of trying to capture water from the street curb. If the water comes in at the curb gutter line elevation, the soil must be at least 3-6 inches lower. With a 6″ curb, a 2% cross slope on the sidewalk and a 1-2% slope on the gutter line, the soil level on the sidewalk side of the treatment bed will be well over a foot lower. Depending on the gutter grades and how carefully the designer understands the hydrology, the receiving soil areas can be as much as 18″ deep.

On the other hand, if curb water is not going to be included, then the soil can rise to an elevation that works for the head and water collection required for that drainage system. The surface elevation of the treatment soil must be level or close to level in order to function well. In an otherwise sloped landscape around the bed, large differences in the elevations of the surrounding pavements can become significant. None of that geometry problem is related to the fact that Silva Cells may be under the pavement adjacent to the exposed treatment bed.

In an ideal world, built just for stormwater and trees, all the required soil would be exposed to the surface soil with no need to put any soil under the pavement. In a world designed to meet all the compromises of a dense urban environment; however, most of the stormwater treatment and tree planting soil might want to be under pavement. The usual answer in most designs is somewhere in between. The size of the water receiving area must be designed to allow the accumulation of the required water volume.

Overall infiltration time is calculated into the surface area of the soil for the drainage area designed. The soil surface area can be smaller as the head of water over the soil is increased in order to receive the required water volume, but that quickly drives the soil surface deeper below the paving, not desirable as mentioned above. In most urban sites, the exposed soil area is mainly there to be the fore-bay of the system to trap floatables and sediments. In a well-designed urban system, the majority of the water may be directed into the soil under the pavement by using a raised inlet connected to an exfiltration pipe within the sub paving soil. This allows both for the filtering of the coarse material in a relatively smaller soil opening on the surface, and the treatment and retention or detention of large amounts of water taking up the least amount of space. Of course, the hydrology calculations to make all this happen are critical but often not correctly designed.

Are other rainwater treatment options better than Silva Cells? 

Other ideas, such as porous paving or green roofs, are all part of the solution. Success will come only with the use of all options, each in their appropriate place, and often combining ideas.

For example, pervious pavers should have a much wider application potential. They should almost always be recommended on top of Silva Cells as the ideal way to get more water into soil that supports both the stormwater goals and tree canopy goals. Pervious paving allows the reduction in size or even the elimination of the need for the fore-bay feature of the treatment system. Green roofs do treat the roof portion of the rainwater but miss the rain that lands elsewhere. In intense urban areas, every surface must be put into play when treating rainwater.

What about the tree trunk flare growing so large that it changes the hydrology of the surface in the open soil bed?

I hope the trunk flares grow large, as this indicates the tree is very healthy. The design must recognize how a tree grows and then put the tree in a location that works both for the tree and the hydrology. I have seen designers plant the tree in the same location as the slot that lets water into the soil fore-bay. This will plug the system as the trunk flare rises out of the ground. The best design would have the lower receiving soil and slot to let in curb water be at only one end of the open soil area, and the tree planted at the other. Unfortunately, I almost never see this. We must have some misguided notion that the tree and other features must be always symmetrical.

Clearly all the above ideas require that the designer think about all the issues and then design the system to respond to the hydrology and tree growth patterns. They must understand that the tree will change all the time.

How are trees compatible with the concept of the open soil area as a filtering fore-bay?  Will we just simply have to remove the trees after 10-20 years?

First, trees and stormwater treatment are entirely compatible. Forested wetlands are the most valuable of our upland stormwater treatment systems. Trees can easily accept the level of short-term flooding and sediment build up in these systems, which are overly dry, not overly wet, most of the time.

Related, won’t we have to dig up the soil every so often to clean out the system, killing the tree?

While fore-bays are critical to these systems, in intense urban areas the amount of sediment is often not as great as in suburban areas. This sediment can be removed without harming the tree by using air spade and Vactor truck technology. Vactor trucks are already used to clean out inlets that get silted in. We are already using air spades and Vactor technology in tree preservation work to remove soil from the base of trees. Some designers are working on concrete fore-bays that will segregate the majority of the coarse material, making it easier to clean. Assuming that the tree is provided with adequate soil under the pavement, I do not see why we cannot remove sediments above the base soil level, over time, as needed.

Won’t we eventually have to remove all the soil (and the tree) because the soil has become polluted?

Urban storm water is typically fairly clean of chemical pollutants. In a field study of the effectiveness of bioretention in California, the pollutant levels of the runoff from the parking lot site was so low that the researchers had to fabricate polluted water to make the study effective.

The levels of pollutants removed by these systems are measured in hundreds of parts per million. The tree can actually photo-remediate much of these pollutants and maintain soil organic matter that supports soil biology that also does a great job of immobilizing pollutants. All plants do some level of phytoremediation, some better than others. Trees are usually very efficient at this function. The levels of pollutant concentrations where the contaminant starts to harm the tree are measured in thousands to tens of thousands of parts per million. In all but the most polluted environments, the tree should be able to successfully grow in the slowly accumulating pollutant levels. I do not believe that we ever have to dig up this soil during the effective life of the tree.

Without a completely developed root system on all four sides, won’t trees fail? What about trees planted in deep soil — how does that affect stability? [This question was answered in tandem with L. Peter MacDonagh, formerly of Kestrel Design Group.]

Currently, most trees planted in pavement do not have access to an adequate amount of soil and therefore have poorly developed root systems on all sides, which does negatively affect stability (Johnson et al). Having said that, the most common cause of urban tree trunk failure is actually girdling roots inducing trees to snap off at their base at ground level during windstorms (Johnson et al). This phenomenon is caused by trees being planted too deep — in other words, burying the root ball, which compels the tree roots to move to the surface towards oxygen-rich soils. These roots continue circling the trunk and can weaken the stem catastrophically (Watson et al).

Urban trees planted without girdling roots will have a much lower chance of failure, as will trees that have a developed root ball, even if that root ball is not even on all four sides. A deeper, more well-developed root system is an order of magnitude better for tree stability. In fact, the separation of a tree’s root plate from the underlying soil profile is rarely seen with trees planted in paving (Smiley 2009).

Many people I speak to are concerned about the effect high winds have on tree stability, but this is not something we’ve observed many problems with. This is largely because wind speeds within metropolitan areas are dramatically reduced by surrounding buildings (Seely et al). By checking a meteorological website, you should be able to find local wind speeds for your area. Those speeds will usually be well below the failure thresholds for most trees.