Awhile back, in “The Engineer and the Fisherman”, I explored the unique challenges posed by trying to stabilize urban riverbanks using natural or vegetated measures. This post generated a lot of interesting feedback – much of it through the Urban Design Network LinkedIn group. This interest combined with the fact that the subject is so broad and complex presents an opportunity to deepen the conversation and more specifically target some of the engineering issues contained in these designs.
Approaching such a complex issue is daunting, but as with most engineering problems the discussion can be simplified by breaking the original topic into understandable pieces. This approach is the basis of my “Framework for Understanding Details”. Under this framework, we segregate constraints into one or more of three categories – programmatic, environmental and physical constraints – with the intersection of the categories representing the available design choices. This relationship is represented in the Venn diagram below.
Programmatic constraints are the limitations imposed by the project owner. Looking at the topic of naturalized riverbank design through the lens of this framework, there are two likely programmatic constraints that should concern the designer. First – and especially in urban settings – it is very likely that the project owner wants a bank that will not move in the long term. When some bank movement is acceptable, the amount is usually small and more or less only serves as an indication of the level and urgency of necessary maintenance. In these cases, the bank is usually returned to its originally designed configuration with routine maintenance.
Second, by targeting a naturalized riverbank, the project team adds a programmatic constraint that severely limits the group of available designs. Native vegetation provides the primary long term scour resistance needed to hold these banks in place, but its presence is more an indication of the underlying bank structure than it is the cause of that structure. In the same way that a building’s architecture is supported by a tailored structural design, many factors must be just right before vegetation can be sustained over a long period of time. Of course, the degree of this constraint can vary. There is a big difference between a design that uses natural elements as the primary structure and a design that has the desired outward appearance but is supported by a more conventional engineering approach below ground level.
These programmatic constraints are at least to some degree at odds with each other. Natural riverbanks are flexible, moving and changing over time – a feature that does not mix well with urban development. Project owners spend huge amounts of money developing property and they hire engineers to give them a level of assurance that their investments will be protected through storms, floods, earthquakes and whatever else time might bring. Very few options are left after constraining the design to solutions that use natural elements and create a relatively inflexible bank, and this is after only one category of constraints.
Environmental constraints are the limitations imposed by the interactions between the project and its environment. Not surprisingly, the driving interaction on riverbank projects is that with the river. From the river’s view, most projects occupy a very small amount of space – sitting high on the bank. This can mean that there are global stability problems that can’t be dealt with in the scope of a typical project. Many of these problems radiate from deep within the river’s main channel and can only be dealt with by creating a last line of defense – often soil improvements or deep foundations – at the face of the development. The sketch below illustrates one possible way that a global problem could affect the stability of a project.
These sorts of topography considerations are one of many factors that contribute to a river’s flow characteristics, which in turn define the stresses that the river imposes on its banks. Variations in the bank line can result in an infinite number of scenarios in this regard, but longitudinal stress – the stress imposed by water moving along the bank or down the river – and perpendicular stress – the stress imposed by waves moving up the bank – are the two parameters that will combine to drive the design. These values depend on the precise conditions for each project and vary from site to site, but generalizations can often be made for specific stretches of a river. For example, bank design on large river systems that carry shipping traffic may be controlled by the waves thrown by passing barges, while design on smaller creeks or streams might be controlled by the speed of the current during times of high water.
Many of the river related habitat and scour issues that our communities face today result from designs that did not account for their impact on downstream systems. The limitations imposed by downstream interactions are very much environmental constraints and need to be quantified and addressed as such. That said, most vegetated bank designs result in slower, less channelized flows and positively affect downstream properties. This leaves global bank stability, river flow related shear stress and wave related shear stress as the key environmental constraints for typical naturalized riverbank designs.
Physical constraints – the limitations imposed by such aspects as location and constructibility – make up the third category in the framework. The most global of these constraints for this conversation involves slope. Over time, many of the riverbanks that support our cities have been steepened to create more developable land. Returning these banks to their natural state usually requires returning them to something closer to their natural slope, creating two challenges.
Just as banks were originally steepened to create more developable land, flattening these banks requires giving some of that land back to the river. This is possible in some cases, but can be problematic on sites that host buildings or other features that are slated to remain. It can also be difficult to transition the re-flattened bank to meet adjacent properties. These transitions often have to be made gradually to avoid creating scour pockets and can lead to a scenario where most of the site is transition and very little is actually a fully naturalized design.
Along with topography and slope considerations, plant establishment time is a very real limitation that must be addressed in order for a naturalized design to succeed. Most natural riverbanks rely on native vegetation for scour protection, without it these banks would wash away. But even in the best cases it takes several years for a newly planted bank to develop the root system necessary to provide this protection, meaning that an interim solution must be integrated into the design. Ideally, this interim measure is biodegradable so that the support and structure it offers will lessen as the vegetated system gets stronger.
Coir mat* – made from coconut shell fibers – can be a useful cover in these situations as it is natural, biodegradable and available in many forms. Like most erosion control blankets, coir mat can be found in both woven – think burlap – and nonwoven – think furnace filter – forms. The woven version is strong and provides good resistance to shear stress. At the same time, though, it has large openings and on its own will allow soil to wash away. Nonwoven coir mat has small openings and is very effective at preventing soil loss, but it is not very strong and would likely rip under the shear stress imposed by swift river currents. Because each type of coir matting has a strength that supports the other’s weakness, an obvious solution is to layer both woven and nonwoven matting on the bank. In this application, nonwoven matting is placed directly on the graded bank to provide soil retention, then woven matting is laid over the top to provide shear resistance.
Of course, even the best matting design will be of little help if it isn’t properly attached to the bank. The attachment system has to be simple and constructible while at the same time providing resistance to horizontal and vertical movement across the entire bank. After working through several designs, the process that I feel best meets these goals is to lay the matting on the bank, stake it in place with a close grid of long, square stakes, then weave twine around and between these stakes. In this configuration the stakes offer resistance to horizontal movement and hold the blanket in place during construction while the twine offers vertical support, preventing the mat from floating or warping. The photo at the beginning of this post shows one example of this type of installation, taken at Portland’s South Waterfront.
With all of this as background, the following is a revised version of the previously presented Venn diagram summarizing the programmatic, environmental and physical constraints that are imposed on a typical naturalized riverbank design. Each project will include many more constraints than those discussed here, but this set is meant to be a fairly standard collection for this type of project. In order to succeed, a riverbank design must address all of these constraints.
You can see that while there are many design options that can be categorized as naturalized or vegetated solutions, there are really very few choices that will work with all of the constraints laid out here. This is not meant to imply that designers shouldn’t try to solve this problem though. The civil engineering profession exists to develop creative solutions to problems like this one and civil engineers have to some degree sidelined themselves in recent years by avoiding the risks necessary to confront these challenges. By dissecting and analyzing the project constraints, we can better understand both the steps we need to take to address them and the implications of pushing these boundaries.
Have I overlooked any common constraints that you have encountered on these types of projects? Please feel free to post them along with any other thoughts you might have on the topic in the comments section.
*The original version of this post recommended jute mat, which is also a good product for this application, but is not as strong as coir.