PERFORMANCE OF SOIL-CEMENT PLATING IN THE TEXAS PANHANDLE By: D.N. Clute, P.E.; M.S. Garsjo, P.E.; K.R. Worster, P.E. National Design, Construction, and Soil Mechanics Center, USDA-NRCS, Fort Worth, Texas Abstract Soil-cement that is installed for wave protection (armor against waves that would otherwise cause shoreline erosion) is typically placed in horizontal layers. In 1979 the USDA - Soil Conservation Service (SCS) in Texas deviated from this conventional method of placing soil-cement in horizontal layers and began plating the front slope of dams being built in the Texas Panhandle with soil-cement placed in layers oriented parallel to the slope of the dam. Several dams were constructed in the Texas Panhandle over a 10 year period from 1979 to 1989 with thin soil-cement plating placed on the front slope for wave protection. Thin soilcement plating was also used to armor some auxiliary spillways and waterways or chutes needed for controlling onsite erosion from localized storm water runoff. Inspections of the thin soil-cement plating were made in 1985, 1991, 1996, and 2005. This paper documents the findings of the teams that inspected the soil-cement. The information in this paper should be useful for designers when designing soil-cement structures, selecting the materials, specifying the method of compaction and curing methods for soilcement used for wave protection and protection against erosion from localized storm water runoff. Introduction The USDA Natural Resources Conservation Service (NRCS), formerly the Soil Conservation Service (SCS), completed a long term study to document the performance of thin soil-cement plating on 15 flood control dams. The dams that were studied were built from 1979 to 1989 and are located in three watersheds in the Texas Panhandle (Figure 1). 1
These dams are typically less than 50 feet high with 3.5:1 (horizontal:vertical) front and back slopes. Photo 1 shows one of the dams in the study group. Photo 2 shows a aerial view of another of the dams in the study group. Both dams shown are typical of the dams in the study group. Based on NRCS Technical Release 69 (TR 69) 1, a significant wave height of less than three feet would be typical for these dams. Had rock riprap been used for wave protection, TR 69 would have required a layer of graded rock riprap approximately two feet thick with an approximate D 50 rock size of one foot. Soil-cement was deemed more economical than rock riprap since rock was not readily available in the area. Photo 1 A dam that is typical of those in the study group 2
Photo 2 Aerial view of a dam that is typical of those in the study group Thirteen of the 15 dams have no permanent pool so the threat of wave damage only occurs on those sites for a few days at a time when they retain floodwater. The thin soilcement plating was installed primarily to protect against this intermittent wave damage along the shoreline. The thin soil-cement plating was also used to armor some auxiliary spillways against erosion that would have occurred from localized storm water runoff and to construct soil-cement plated waterways or chutes needed for controlling onsite erosion from localized storm water runoff. Prior to 1979 the SCS had constructed soil-cement wave protection in the conventional horizontal lift configuration where 8-inch thick, 6 to 8-foot wide horizontal lifts or layers were placed on the front slope of dams above and below the waterline (Figure 2). SCS designers recognized that orienting the soil-cement layers parallel to the slope (Figure 3) would reduce the required soil-cement quantity by more than half that needed for the conventional horizontal lift configuration. In question was the performance and durability of the thin layers of soilcement. It was decided that thin soil-cement plating would be installed and monitored to determine if it would provide the needed protection and be durable in the Texas Panhandle environment. 3
Figure 2 - Conventional placement in horizontal lifts. Figure 3 Soil-cement plating oriented parallel to the slope. The soil-cement on the earlier constructed sites was inspected by a team of engineers in 1985 and all of the sites were inspected in 1991, 1996, and 2005. This paper documents the findings of the teams that inspected the soil-cement. The information in this paper should be useful for designers when designing soil-cement structures, selecting the materials, specifying the method of compaction, and specifying curing methods for soil-cement used for wave protection and protection against erosion from localized storm water runoff. Exposure Conditions All of the sites within the study area are located within 100 miles and east of Amarillo, Texas. A detailed study of freeze-thaw potential in the area was beyond the scope of this study; however, the average monthly precipitation and temperature data for the period between 1979 and 2000 are given in Table 1. The potential for freeze-thaw damage in this relatively arid climate should be low compared to freeze-thaw damage potential in humid climates. 4
Table 1 Temperature and Precipitation Data for Amarillo, Texas from 1971 to 2000 2 Mix Proportioning and Construction of Soil-Cement Plating Soils and Mix Proportioning The soils used for making the soil-cement were obtained onsite and generally had a Unified Soil Classification of SP, SM, SC, or SC-SM. Soils in the northernmost watershed were generally of the SC-SM variety and required more cement to make desirable soil-cement. Soils in the southernmost watershed were classified as SP and required less cement to attain the desired properties. Cement content of soil-cement is commonly expressed as a percentage of the total oven-dry weight of soil used in the soil-cement mixture. Cement contents for soil-cement installed in the study area were determined in the laboratory on the basis of compressive strength, freeze-thaw durability, and wet-dry durability as set forth in the Soil-cement Laboratory Handbook, Portland Cement Association, 1959. The 1992 edition of this publication contains the same criteria as that of the 1959 edition. 3 The cement content of the soil-cement installed in the study area ranged from 10 percent for soils with fewer fines to 12 percent for the soils having more plastic-fines. Batching and Mixing The soil-cement was mixed by a pug mill on all but one site where the soil-cement placed in the auxiliary spillway was mixed in place by a roto-tiller. The pug mill produced a well proportioned uniform mix. The in-place mixing produced weaker soil-cement, especially at the edge of placement lanes (i.e. construction joints) where it has deteriorated in several places (Photo 3). 5
Photo 3 Deterioration of construction joints occurred where soil-cement was mixed in place Soil-cement was mixed in a pug mill on one site and transported to another site. This soil-cement, having been transported from one site to the other, was noticeably soft and seemed to deteriorate more than soil-cement placed on the sites where the mix was produced. Apparently, the increase in time elapsed from mixing to placement and compaction adversely affected the quality of the soil-cement. 6
Placement The soil-cement plating for wave protection was placed in two 6-inch lifts or one 8-inch lift on 3.5:1 (horizontal:vertical) slopes. The soil cement placed in chutes was generally placed in two 6-inch lifts. Soil-cement installed in auxiliary spillways was placed in one 8-inch layer. Soil-cement plating for wave protection placed on some of the earlier constructed sites was placed across the slope in lanes parallel with the centerline of the dam (Figure 4). This method proved difficult as the compaction equipment tended to slide off line as it traveled across the slope. Placement and compaction was easier to control when lanes were oriented up and down the slope, perpendicular to the centerline of the dam (Figure 5). Later applications were oriented up and down the slope. Where two 6-inch lifts were placed, the edges of the lanes in the top lift were generally offset from the edges of the lanes in the bottom lift. Figure 4 - Lanes oriented across the slope. Figure 5 - Lanes oriented up and down slope. 7
Compaction Soil-cement was compacted by a steel wheel vibratory compactor, a pneumatic tire roller, or a track type front-end loader. Density was specified to be 100 percent of the maximum dry density as determined by ASTM D 558, Standard Test Methods for Moisture- Density (Unit Weight) Relations of Soil-Cement Mixtures. The moisture content at time of compaction was specified to be from 1 percent below optimum to 2 percent above optimum. Construction documentation points to difficulty meeting the requirement for 100 percent maximum density; actual density was likely in the 95 to 97 percent range. Requiring greater than 95 percent compaction on 3.5:1 (horizontal:vertical) slopes is not realistic or practical. Curing Specifications required the soil-cement to be moist-cured for a period of seven days after placement. A moist soil blanket was used for curing and in most cases it was not removed after curing. Performance Overall the soil-cement has performed well even though there has been some apparent freeze-thaw damage and erosion from localized storm water runoff. Where two 6-inch lifts were installed, the top 6 inches has completely deteriorated on a few of the sites, but the remaining 6 inches of soil-cement on these sites is likely to provide another 25 years of service. Soil-cement on some of the sites has suffered no apparent damage. Mechanism of Deterioration The assumed mechanisms of deterioration were saturation followed by freezing which resulted in expansive forces that damaged the saturated soil-cement and erosion from water flowing over the surface. Delamination was caused by freezing water that expanded in vertical cracks and popped chunks of the soil-cement loose. Where water flowed over the surface of the damaged soil-cement, the materials that crumbled from freeze-thaw damage were eroded away. Larger chunks remain in-place. Thickness of the damaged layers ranged from 1 to 2 inches as evidenced by the remaining portion of damaged layers (Photo 4). The two-inch delamination of layers occurred in soil-cement having lower strengths, and possibly higher permeability. Only one inch thick delaminated layers occurred in soil-cement of a higher quality. 8
Photo 4 Remaining portion of layers indicates delamination of layers that are 1 to 2 inches thick. Where soil-cement was placed in lanes oriented up and down the slope, the joints were oriented with the flow. Localized storm water runoff tended to erode these joints (Figure 6). Figure 6 Joints oriented with the flow tended to erode. Where the joints were oriented perpendicular to the flow, freeze damage occurred just down slope of the joint, apparently due to the expansive forces of water trapped in the joint (Figure 7). 9
Figure 7 Joints oriented perpendicular to the flow were subject to freeze damage. The potential for freeze damage would be reduced with watertight joints. In an attempt to improve the water tightness of joints, the edge of a placement lane should be dense and clean when soil-cement mix is placed against it. During construction it would be best to maintain a live joint so that the soil-cement on each side of the joint could be compacted concurrently without damage to the previously placed soil-cement. This would effectively eliminate the construction joint and provide a more watertight continuous slab. Soil Dependent Quality The quality of soil-cement varied from site to site. Soil-cement made with SP and SM soils produced a very durable product. Soils containing plastic fines (e.g. SC and SM-SC soils) produced soft soil-cement that could easily be gouged with a knife. Construction records noted that there were clay balls in the mix on one site. Some of these have since eroded leaving small holes or pock marks in the soil-cement surface. The soft soil-cement deteriorated much more than soil-cement made with soils containing fewer plastic fines (e.g. SP and SM materials) (Photo 5). Soil-cement made with SP and SM soils could not be gouged with a knife. Equipment tracks are still visible in the soil-cement surface on some of the sites where SP materials were used, indicating very little deterioration of the more durable soil-cement (Photo 6). 10
Photo 5 - Deteriorated soft soil-cement is common where SM-SC soils were used to make soil-cement. Photo 6 - Equipment tracks after 25 years indicate little deterioration of soil-cement made with SP soils. Photos 5 & 6 - Obvious difference in the quality of soil-cement made with different soils Curing A soil blanket was used for curing and in most cases it was not removed after curing. Over time, some of the curing soil eroded away while the remaining soil channeled and concentrated storm water runoff into specific areas. This concentration of runoff and the protection afforded by the curing materials left in place, resulted in uneven deterioration of the soil-cement (Photo 7). A moist blanket of soil is an excellent method of curing; however, it would have been better to remove the curing materials after the curing period to promote uniform wear of the surface. Photo 7 Uneven deterioration caused by curing materials left in place. 11
Woody Vegetation Woody vegetation was a problem on some sites. The soil-cement plating apparently functioned like a mulch to help retain ground moisture. This promoted tree growth on sites where there were otherwise few trees. Trees sprouted in open cracks and at the edge of the soil-cement plating. There was more woody vegetation in the cracks where one lift was installed than where two lifts were installed. This was likely due to the fact that, where only one lift was installed, soil would have been present at the bottom of the crack to germinate and support plant growth. Conversely, where two lifts were installed, the bottom layer of soil-cement between the crack and the ground deterred the germination of woody vegetation. Trees that grew to a diameter larger than two inches damaged the soil-cement. In a few cases the larger trees caused the soil-cement to heave and breakup around the tree with the damage radiating several feet from the tree trunk. A more voracious maintenance program to prevent trees from getting larger than two inches in diameter would have been prudent (Photos 8 & 9). Photo 8 - Soil-cement heaved and broke up around trees larger than approximately 6 inches. Photo 9 - Small diameter trees caused little damage. Some were killed with herbicide but not removed. Photos 8 & 9 Large trees damaged soil-cement, smaller trees did little damage. Damage from Waves The damage from wave action was relatively minimal on the site where the soil-cement was subjected to a permanent pool (Photo10). Other sites were subjected to wave action during flood stage only; the soil-cement installed for protection against wave damage on these sites showed no signs of damage from wave action. 12
Photo 10 Minimal wave damage on a site with a permanent pool Conclusion Thin soil-cement plating for wave protection and protection against erosion from localized storm water runoff is a viable alternative to the conventional horizontal lift configuration. The quality of the soil-cement, the joint orientation and quality, and exposure conditions are all factors to be considered. Unfortunately, there is no formula for predicting how a specific soil-cement would perform in a specific situation; however, some recommendations can be formulated from observations made during this study: Use cleaner sands with less plastic fines for best soil-cement durability, Mix with a pug mill or drum mixer. Mixing with a tiller in-place produces a poor mix, especially at the construction joints. Locate the mixer on or very near the site to limit the time from mixing to placement. Mixing on a remote site some distance from the placement area results in poor quality soil-cement. Placing soil-cement in two lifts with offset joints will deter the growth of woody vegetation at the joints. 13
Ensure the edge of placement lanes are dense and clean whenever soil-cement is placed against them and strive to maintain a live joint so soil-cement on each side of the joint may be compacted concurrently without damage to the previously placed soilcement. Remove trees at an early age before growth to a diameter greater than 2 inches. Place and compact soil-cement in lanes oriented up and down the slope, as opposed to being oriented across the slope. When plating a slope, specify compaction of soil-cement to 95% of the maximum dry density as determined by ASTM D 558. Remove soils used for curing, after the curing period, as opposed to leaving them on the surface to erode at random. In a more demanding service and where environmental conditions are more severe, consider using the more conventional horizontal lift configuration. 14
References 1. TR 69 - Riprap For Slope Protection Against Wave Action, 5/24/1982, U.S. Department of Agriculture, Natural Resources Conservation Service, http://www.info.usda.gov/ced/ftp/ced/tr69.pdf 2. National Weather Service Amarillo Weather Forecast Office, 1900 English Road, Amarillo, Texas 79108, http://www.srh.noaa.gov/ama/climate/panhandleprecip.htm 3. Soil-Cement Laboratory Handbook, 1992, Portland Cement Association, 5420 Old Orchard Road, Skokie, Illinois 60077-1083, www.cement.org 15