By Anthony J. DePasquale, Douglas Leatherman, and Randy Thomas

Molly Ann’s Brook, a tributary of the Passaic River, flows through an urban area in Paterson, Haledon, and Prospect Park, NJ. Inadequate channel capacity led to repetitive flooding in neighborhoods along a 2.5-mi. stretch of the brook. In the summer of 1999, the Philadelphia District US Army Corps of Engineers designed and constructed a flood protection project to alleviate the problem, incorporating a deeper and wider channel to provide flood protection from a 50-year storm event.
      Although erosion protection is paramount when designing flood protection projects in any environment, it becomes even more critical in urban settings because of the proximity of structures, as was the case with Molly Ann’s Brook: In several areas, structures comprised the banks. In addition, features of the project’s location, such as limited right of way, undulating and outcropping bedrock, and high stream velocities caused by stormwater runoff from surrounding developed areas, imposed many constraints on design and construction.
      Conventional solutions, such as riprap, wire-mesh gabions, cabled concrete mats, and more than 3,000 lin. ft. of precast concrete retaining walls, had been used to protect the banks and contain the floods on adjacent areas of the brook. For two especially difficult locations, however, the corps chose less conventional geocell approaches to ease construction and reduce costs. Although cellular confinement has been around for almost 20 years - developed in the late 1970s as a cooperative research effort between the corps and Presto Products Company of Appleton, WI–the corps has rarely used geocells in flood protection projects and had never used it for constructing earth-retention walls along streambanks.
      In one location in Molly Ann’s Brook, design velocities approaching 20 ft./sec. along 300 ft. of a sharp bend in the channel would traditionally have required a 54-in.-thick layer of riprap. Seeking a more economical alternative, the corps first considered using cabled concrete mats, previously used on a separate phase of the project. The final design, however, called for a single, 4-in. layer of Presto’s perforated Geoweb cellular confinement system (geocell) filled with low slump concrete and underlain with a nonwoven geotextile, which was adequate to protect the bend while still reducing labor, excavation, and materials costs.
      In the other location, a geocell system was used to construct a 400-ft.-long gravity wall, an alternative to a more expensive reinforced concrete retaining wall. The corps used perforated, 8-in.-high geocell layers filled with compacted soil or aggregate to construct the wall along the right bank of the brook. Designed to withstand flash flooding and high velocities during its 50-year design life, the wall ranges from 9 to 14 ft. high and is protected from scour by a 4-in. concrete-filled geocell along its base.
      The corps estimates the methods used saved $80,000 over more traditional solutions. This article takes a close look at how geocell technology was incorporated in the construction of the channel lining and gravity wall.

Channel Lining

In the channel bend, polymeric tendons and 0.5-in.-diameter steel bars anchored the geocell sections to the 1:2 channel sideslope. The factor of safety for slope stability controlled the spacing of the anchor system. The tendons were spaced every 19 in. or every third cell, running perpendicular to the brook. Varying lengths of 0.5-in.-diameter rebar, spaced between 8 and 48 in., anchored the tendons to the slope, depending on the depth to bedrock. If rock was located within 1 ft. of the final grade, rock anchors were specified.
      Perforations were required in the interior geocell walls for two reasons: to allow the concrete within the 4-in. depth to adequately bond between the cells and to increase the frictional resistance between the concrete and the cell walls. The internal bond was established by using the perforated cell walls and filling the cells with a reduced-size aggregate in the concrete. The perforations also allowed a high factor of safety against the dislodging of individual concrete-infilled cells in the channel, which is exposed to freeze/thaw cycles and high-velocity storm events.
      Anticipating bedrock, the corps determined that any outcropping would be left in place and the geocell system would be formed around the outcrop. Weep holes were placed on 8-ft. centers near the bottom of the slope to alleviate pore water pressure and uplift.
      Installing the geocell system required a minimal crew of three. After laying the geocell on the graded slopes from top to bottom and anchoring it, the contractor used a bucket on a tracked backhoe to fill the geocells with concrete. Although the original design called for a trowel finish, the smooth surface caused some safety concerns - students from a nearby high school who cut across the channel could potentially slip on a smooth surface - and a rough-raked finish was used instead. This change in the concrete finish had a negligible impact on the channel capacity and allowed the contractor to complete the job in significantly less time.
      To allow equipment access along the streambed, the entire cross-section was completed in two phases. First the channel sideslopes were graded, followed by the streambed and low flow channel sections. The ability to complete the cross-section in phases and not be concerned with cold joints was helpful. The flexibility of the geocell system was beneficial in constructing the low flow channel; the material had to be flexible enough to cover the 8-ft.-wide x 2-ft.-deep trapezoidal channel and maintain the integrity of the cells. The cells also act as expansion joints throughout and prevent future uncontrolled cracking. The 300 lin. ft. of channel lining was completed in less time than it would have taken an equal length of 21-in.-thick riprap.

Geocell Gravity Wall

The optimal use of a geocell wall was in an area where the elevation of rock surface was known to vary greatly. Essentially, a gravity wall had to be designed for the greatest anticipated height, then modified in the field for varying heights depending on the elevation of the rock.
      The design of the wall was an iterative process between the corps and InterSol Engineering Inc., consultants for Presto Products. The wall was designed as a conventional gravity wall composed of 8-in.-deep geocell layers. Geocell layers extended from the face into the bank a distance of 3 to 7.5 ft. Each layer of geocell was set back 2 in. from the face of the underlying layer. The slope of the front face of the wall was 1:4. The design called for filling the cells with the native silty sand and gravel materials except for the outer five cells of the first five geocell layers; these cells were filled with 0.75-in. stone. The concern was that the exposure to constant flows could cause loss of material in the wall. The final stream grade was established, and the drawings reflected only the one visible rock outcrop along the 400-ft. length.
      The wall was designed to accommodate several required pipe penetrations for local stormwater outfalls. In addition, a foundation drain was included to collect any groundwater flows from behind the wall.
      Because the wall would frequently be subject to high water velocities, scour along the toe of the wall was a major concern. A 4-in.-thick layer of geocell (the same as used in the channel lining), 8 ft. wide, was included at the base of the wall and filled with concrete.
      To improve the appearance, a solid tan geocell material was used on the face of the exterior walls. In addition, the 2-in. setback in each layer provided space for the outer cells to be filled with topsoil and planted with Virginia creeper, a hardy perennial vine. The open front cells of each geocell section form horizontal terraces that capture rainwater while controlling groundwater evaporation, creating a natural environment for vegetation. Eventually, the entire wall face will have a vegetated appearance.
      Before excavation to the final stream grades, the exact location and elevation of the bedrock were unknown. Once the overburden was removed, the bedrock was exposed under the footprint of the wall in four major areas. Beginning at one end of the wall, layers of geocell were brought up one level at a time. When outcrops protruded in the path of the wall, the geocell was simply cut and the next layer started. As the layers were constructed, the outcrops were eventually encapsulated beneath additional layers of geocell. Outcrops that extended beyond the face were incorporated in the final structure. Because of the width of the channel in this area, impacts to channel capacity as a result of the rock outcrops were acceptable.
      Initially the design called for filling the cells with native materials, but because of the many large rocks in this material and the need to overfill the cells for compaction, the geocell infill was changed to 0.75-in. stone. Including stone infill alleviated the need for meticulous compaction and removal of excess materials and rendered the drain unnecessary. The effect of the free-draining stone, combined with the perforated cells, allowed the entire wall to act as a drain. Previous stability concerns regarding slow draining of the backfill after a flood event were alleviated.
      Inadvertent damage to some of the exterior cells during construction demonstrated the redundant nature of geocell wall construction. With the 8-in. horizontal cell depth, a breach of an outer cell exposed the next cell completely intact. Multiple cell breaches would be unlikely under even the worst conditions.
      Although the Molly Ann’s Brook project required only 300 lin. ft. of concrete-filled geocell channel lining and 400 lin. ft. of geocell gravity walls, the project should be a springboard to future use of geocells in flood protection and streambank erosion projects. The estimated cost for design and construction of the geocell wall was less than half that for an equal length of reinforced concrete retaining wall. The channel lining was completed at about one-quarter of the cost of a 54-in. riprap section of equal length.

Anthony J. DePasquale, P.E., and Douglas Leatherman, P.E, are with the Philadelphia District US Army Corps of Engineers in Cherry Hill, NJ. Randy Thomas, P.E., is with ACF Environmental in Richmond, VA.