Canal Seepage Reduction by Soil Compaction - ITRC · Canal Seepage Reduction by Soil Compaction ... The Standard Proctor test (ASTM D698) was developed in the 1920s when static rollers

Canal Seepage Reduction by Soil Compaction http://www.itrc.org/papers/canalseepage.htm ITRC Paper No. P 09-001

Canal Seepage Reduction by Soil Compaction

Charles M. Burt1; Sierra Orvis2; Nadya Alexander3 Abstract: Simple in-situ vibratory soil compaction of earth lined canals was tested to determine the impact on seepage losses. Commercial equipment was used for vibratory compaction of long sections of five irrigation district earthen canals. Ponding tests were conducted before and after compaction. When the sides and bottoms of the canals were compacted, seepage reductions of about 90% were obtained; reductions of 16 – 31% were obtained when only sides were compacted. DOI: CE Database subject headings: canal seepage; in-situ compaction; soil; irrigation; sheepsfoot.

INTRODUCTION Irrigation districts that rely upon long, open canals share a common problem: canal seepage. Canal seepage can lead to: • Reduced water deliveries to farmers • Increased pumping costs if the water in the canals is lifted by pumps • Increased drainage problems, possibly causing crop yield and health problems • Loss of water supply in a basin if the seepage goes to a salty aquifer or into the ocean • Increased diversion from rivers, resulting in decreased in-stream flows The two most common solutions for reducing seepage are lining canals or replacing them with pipes. These options bring along with them additional benefits, such as stabilization of banks (with canal lining) or reduced need for access and fewer drownings (with pipelines). However, these solutions are expensive. A typical piping cost in California for an irrigation district is around $400US-$1000US per meter for pipe sizes in the 1.2-1.5 meter diameter range (flows in the 0.6-0.9 cubic meters per second [CMS] range). Canal lining costs are often around $700,00US per kilometer for a canal with about 3-5 CMS capacity, which is prohibitive for most irrigation districts. A medium-sized district may have more than 150 kilometers of unlined canals. The general concepts of soil compaction for seepage reduction and soil consolidation are well-documented in civil engineering, under the category of “soil mechanics”. New canal bases are often compacted using standard engineering techniques. However, most canals were built crudely, and the authors were unable to locate any published results from

1 Professor, Dept. of BioResource and Agricultural Engr., and Chair, Irrigation Training and Research Center, California Polytechnic State Univ., San Luis Obispo, CA 93407-0730. E-mail: [email protected] 2 Graduate Student and Engineering Assistant, Irrigation Training and Research Center, California Polytechnic State Univ., San Luis Obispo, CA 93407-0730. E-mail: [email protected] 3 Student, Dept. of BioResource and Agricultural Engr., California Polytechnic State Univ., San Luis Obispo, CA 93407-0730. E-mail: [email protected]


within the US to reduce seepage by running compaction equipment over the canal surfaces (without doing the standard, costly over-excavation of soil and careful compaction of replaced soil in layers). Therefore, the Irrigation Training and Research Center (ITRC), with support from the USBR and the California State University Agricultural Research Initiative, researched the effectiveness of in-situ compaction of canal banks and canal bottoms for reducing canal seepage. Concepts of Soil Compaction Two of the major dams in California (Oroville Dam and San Luis Dam) are earth-filled dams rather than concrete structures. The success of these dams certainly proves that proper soil compaction can reduce seepage. Soil laboratory tests for compaction (Proctor and Modified Proctor) have specified procedures by the American Society for Testing and Materials (ASTM). Soil samples are compacted in specified layer thicknesses, with specified weights dropped a specified number of times from a specified height. In a compaction test, this is typically done with a number of samples, each having a different soil moisture content. A graph is developed, illustrating what the moisture content should be during construction. Generally, the purpose of this compaction is to build a solid soil foundation that will not settle over time or deflect when subjected to transient loads. Seepage reduction is not a typical goal of compaction, except for dams. Some soils compact better than others and as the level of compaction increases, the soil hydraulic conductivity (seepage rate) decreases. The increase in bulk density will also depend upon the moisture content during compaction; if the soil is too moist or too dry, it will not achieve the “optimum bulk density”. The details of compaction are more complex than a simple graph might suggest, however. The optimum moisture content is not only dependent on soil type/gradation; it also depends upon the amount of energy applied to the soil for compaction purposes. Therefore, different laboratory compaction test procedures yield different results. The Standard Proctor Optimum Moisture Content (SPOM) varies considerably from the Modified Proctor Optimum Moisture Content (MPOM) for the same soil. Further, energy levels (hammer mass, drop height, and number of drops) applied in determining SPOM and MPOM are arbitrary in nature. The Standard Proctor test (ASTM D698) was developed in the 1920s when static rollers were state-of-the-art. The Modified Proctor test (ASTM D1557) was developed in the 1950s and reflected the increased compaction capabilities of vibratory compactors. Today, technology has advanced even further in terms of energy applied by at least a few compactors, including vibration speeds, pressure, and configuration of the pads/rollers. In general, the higher applied energy, the lower the optimum moisture content will be. The optimum moisture content for high bulk density cannot necessarily be converted into the optimum moisture content for reduced seepage. Some engineers believe that a


slightly-moister-than-“optimum” soil in the field provides the best seepage reduction for non-clay soils. There are challenges here for in-situ compaction—in-place compaction of existing canal banks and bottoms—for seepage reduction: 1. Normally, one only has a small window of time during which the canals are “dry”.

Because clay soils can liquefy during compaction, if they are too wet, clay soils should probably be compacted at moisture contents slightly below the “optimum” moisture content – if the objective is seepage reduction. It is extremely difficult to dry a clay soil to below the “optimum” moisture range; therefore, it is extremely difficult to properly compact clay soils. On the positive side, usually the clay soils have low seepage rates without compaction. However, what small amounts of water do seep out often show up on the surface of the soil next to the canal and create operational headaches for farmers and district personnel.

2. If one applies a laboratory compaction test (at various moisture contents) to a soil, it is fairly easy to determine differences in bulk density. It is entirely different in determining the effect on hydraulic conductivity. Over a wide range of compaction conditions, the laboratory tests will yield very similar results – a drastic reduction of hydraulic conductivity. To look at the lab results, one could conclude that as long as there is some compaction, the seepage would almost be eliminated. However, this level of effectiveness has not been seen in the field.

3. A canal is removed from service for up to a few weeks prior to compaction. That means at each depth from the surface, there is a different moisture content.

Compaction Equipment The best type of compaction equipment will vary, depending upon the soil type. For example, according to Bader (2001), a vibrating sheepsfoot roller may be excellent for compacting clay, but ineffective for gravel or sand. In gravel or sand, a vibrating plate roller would have much better results. For the compaction research, ITRC purchased a Universal Vibratory Wheel (UVW) Roller from MBW, Inc. This is a sheepsfoot in-drum configuration. The operator can select the mode of vibration for a particular condition. It can be operated in static mode (clays), vibratory mode (little host machine down pressure, for sands and gravels), and in a mode that combines variable host machine down pressure to a maximum of roughly 130 kN (at the end of an excavator arm) with full vibration (clays, mixed soils, most native soils). Soil Compaction for Sealing Canals The best source found for information on earth lining of canals is a publication by ANCID (Australian National Committee on Irrigation and Drainage, 2001) entitled “Open Channel Seepage and Control”. This publication focuses on bringing soil material to the site one layer at a time and properly compacting each layer. The publication does mention “in-situ” compaction. The senior author has talked to engineers from dozens of


irrigation districts in California about this, and has not encountered anyone who has tried in-situ compaction before this research. After a preliminary presentation of the research results (Burt et al., 2008) one engineer from Colorado (Zimbelman, 2009) did comment that compaction had produced good seepage reduction in the Northern Colorado Water Conservancy district near Loveland.

FIELD EXPERIMENTS IN CALIFORNIA ITRC contacted four irrigation districts in the San Joaquin Valley of California that were experiencing seepage problems: • Panoche WD • Chowchilla WD • San Luis Canal Co. (also known as Henry Miller Reclamation District) • James ID Seepage tests were conducted on two Panoche WD canals, and it was determined that the seepage rates were very low. Plus, the soil was a heavy silty clay loam and it would be impossible to dry the soil enough for compaction. Therefore, further efforts at Panoche WD were halted. The other three districts had sandier soil, so compaction trials were conducted there. The results of one canal compaction effort in Chowchilla WD cannot be reported because the well that would have supplied the water for post-compaction seepage tests failed and was not repaired. One site in James ID had very low pre-compaction seepage rates, so the site was not compacted. All the compaction work was performed “in-situ”, meaning that there was no addition of soil, and no over-excavation and replacement of compacted soil layers. The compaction was performed on the soil surface “as-is,” with the exception of some smoothing of canal banks. Two of the three districts have conducted subsequent compaction efforts to reduce seepage in their canals. Seepage Tests Prior to and after compaction, ponding tests were conducted to determine the seepage rates (shown in Fig. 1).


Figure 1. Pre-compaction seepage test – James ID Main Canal

The ponding tests involved the following: • The entire canal pool that was to be compacted was filled with water to the normal

operating depth. • The ends of the pool were sealed to prevent water from entering or leaving the pool. • Weather data was recorded from the nearest California Irrigation Management

Information System (CIMIS) station, to estimate evaporation losses. The estimated volume of evaporation was estimated as:

Evaporation volume = ETo × 0.95 × (Surface area of the pond), where - ETo (grass reference evapotranspiration) is from the CIMIS station

as the sum of hourly data, for the duration of the test - 0.95 is the assumed “crop coefficient” for a still, deep body of

water several feet below the canal banks - The area is the area of the ponded water surface

• Redundant water level sensors were installed to measure the change in water depth versus time (see Fig. 2). The sensors used were 5 psi DRUCK PDCR 1830 pressure transducers with a ±0.06% accuracy. Water levels were recorded once per 5 minutes using a Telog PR-31 12-bit datalogger. Differences in readings between the sensors contributed to an error of 1-4% of total seepage, depending upon on the location. The transducers were firmly secured to different stakes in the ponds. Manual readings were also taken of relative water depths, but only to confirm pressure transducer data reasonableness.


Figure 2. Pressure transducer (“Druck” brand) data from ponding test – Henry Miller RD • The water was replenished as needed with a metered supply to maintain a fairly

constant water level. The flow meter used during the seepage tests was a SeaMetrics magnetic meter AG2000-400.

• There was no attempt made to examine the effect of silt deposition during the ponding tests. Both the pre- and post- compaction tests used the same canal water supplies.

• Water temperatures were measured. It was assumed that seepage rates are directly proportional to hydraulic conductivity, which is directly proportional to kinematic viscosity. Therefore, the water temperatures were used to provide relative hydraulic conductivities in pre- and post-compaction tests. The seepage rate for the post-compaction test was adjusted to the same water temperature as the pre-compaction test.

• Measurements began after water had been standing in the pool for several days, and continued for 1-3 days.

• The wetted areas of the pools were determined with survey data of 5 cross sections of each pool.

Soil Preparation During the first compaction work at the H Lateral in James ID, it was learned that if the canal banks were smoothed off first, the compactor could operate much more quickly. Subsequent locations were therefore lightly smoothed off. There was no opportunity to


obtain the “optimum” moisture content for compaction. Field conditions and availability required that the compactor begin work as soon as the canal had dried down enough to use the equipment without making mud. Moisture contents were different at various depths in the banks and bottom. Laboratory Tests Texture samples (about 20 per canal section) were taken at various depths. Texture samples were either sent to a commercial laboratory for analysis, or were tested by ITRC using equipment and procedures specified by ASTM D421 and D422. Laboratory preparation of compacted soil samples was done using two techniques. With both techniques, the soil used had been collected in the field immediately prior to compaction and was sealed in a plastic bag to retain the original moisture content. The techniques were:

1. Proctor Method (ASTM D698) using commercial equipment including a 2.5 kg (5.5 lb) drop hammer.

2. Water compaction. Approximately 2 cm of soil at a time was placed in the permeameter cylinder. Water was poured into the cylinder until it was about 2-3 cm above the soil and then allowed to drain. This was repeated until the entire cylinder was full of water-compacted soil.

Undisturbed soil samples for bulk density measurements were taken with an AMS® 3.5 cm × 30.5 cm (1.375” × 12”) split core sampler and signature series slide hammer. The soil above the sample depth was excavated by hand to reveal the top of the sample zone before the sampler was used. Sample locations were selected to cover a variety of locations along each measured cross section (see Fig. 3).

Figure 3. James ID H Lateral example cross section showing location of soil sampling as

dots Hydraulic conductivity was tested using a constant head permeameter (Humboldt Mfg. Compaction Permeameter (H-4145)). Equipment and Costs


The side soil was compacted using a 45-thousand pound (20 metric ton) Kobelco excavator with an MBW, Inc. UVW 36-inch (91-cm) roller attached to the end of the boom (Figs. 4 and 5). Installed immediately between the UVW-36 roller and the end of the excavator boom was a UV-10K exciter. This exciter is a hydraulically driven vibration mechanism. Since the vibratory exciter was hydraulically driven, the excavator operator could engage and disengage the exciter when he felt it was necessary. For large canal bottoms, a ride-on unit was used (Fig. 6).

Figure 4. Side compaction with Kobelco excavator – James ID Main Canal

Figure 5. 36” vibratory roller (MBW, Inc.) attached to the end of an excavator arm


Figure 6. Ride-on vibratory compactor for bottom of canals – Chowchilla WD Main

Canal The compaction accessories cost about $25,000US (not including the cost of the excavator) installed. An experienced operator was able to compact the sides of 1.6 kilometers of canal (both sides, meaning 3.2 kilometers total) in about 8 days. The cost for the operator, transport of the excavator, and the excavator rental was about $3.90US per meter of canal length, with about 3 meters of compaction on each side of the canal.

RESULTS Undisturbed soil core samples were collected at various depths, locations along cross sections, and at different cross sections of each pool in an attempt to characterize the range of bulk densities in the canal. Samples were taken before compaction (Pre-Compact) and after compaction (Post-Compact). Results are summarized in Tables 1 and 2.

Table 1. Undisturbed bulk densities pre- and post-compaction

Canal Value Depth Range, cm.

0 - 10 10 - 20 20 - 30 30 - 41 41 - 51 51 - 61 All James ID

Main Pre-Compact Avg. BD, gm/cm3 1.33 1.31 1.51 1.54 1.55 1.62 1.46

n 7 8 7 5 5 5 37 cv 0.14 0.19 0.03 0.05 0.05 0.03 0.13

Post-Compact Avg. BD 1.53 1.62 1.57 1.54 1.62 1.63 1.59 n 6 7 8 7 8 7 43

cv 0.09 0.09 0.08 0.19 0.08 0.04 0.10 James ID H Lateral

Pre-Compact Avg. 1.44 1.54 1.67 1.59 1.77 1.79 1.58 n 19 19 18 7 6 7 76

cv 0.19 0.12 0.07 0.15 0.02 0.03 0.14 Post-Compact Avg. 1.45 1.79 1.81 2.03 1.85 1.83 1.79

n 2 3 3 2 2 2 14 cv 0.14 0.02 0.02 0.18 0.01 0.07 0.12

Chowchilla # 2 (Main)

Pre-Compact Avg. 1.38 1.47 1.51 1.45 1.48 1.55 1.47 n 35 35 35 15 16 15 151


n 13 13 12 13 14 13 78


cv 0.11 0.13 0.08 0.08 0.12 0.12 0.11 Chowchilla

# 1 (Ash Bypass)

Pre-Compact Avg. 1.48 1.52 1.50 1.62 1.61 1.55 1.50 n 20 19 20 6 7 6 78

cv 0.09 0.04 0.04 0.08 0.07 0.06 0.07 Post-Compact Avg.

No data; no samples were collected n cv

HMRD Swamp

Pre-Compact Avg. 1.40 1.48 1.54 1.51 1.65 1.70 1.55 n 7 7 8 8 7 7 44


n 8 6 8 8 8 7 45 cv 0.07 0.05 0.10 0.11 0.09 0.07 0.09

HMRD Delta

Pre-Compact Avg. 1.17 1.28 1.34 1.26 1.22 1.17 1.23 n 4 3 2 3 4 3 19


n 2 2 2 2 2 2 12 cv 0.21 0.11 0.02 0.01 0.10 0.06 0.13

Table 2. Relative bulk densities (post-compaction/pre-compaction) of undisturbed soil

core samples at a various depths Depth Range, cm

Canal 0-10 10-20 20-30 30-41 41-51 51-61 All James ID Main 1.15 1.24 1.04 1.00 1.04 1.01 1.09

James ID H Lateral 1.01 1.16 1.09 1.27 1.04 1.02 1.13 Chowchilla #2 (Main) 1.02 1.04 0.99 1.06 1.05 0.96 1.02

HMRD Swamp 1.06 1.00 1.01 1.04 0.97 0.96 1.01 HMRD Delta 0.98 1.02 1.05 1.12 1.21 1.39 1.13

Laboratory measurements of saturated hydraulic conductivity and corresponding bulk densities were performed on disturbed soil samples collected from four canal pools. Results are found in Table 3.

Table 3. Laboratory results of hydraulic conductivity and bulk density

Location Value

Water Compacted Proctor Compacted

K (cm/sec)

Bulk Density

(gm/cm3) K (cm/sec)

Bulk Density

(gm/cm3) Chowchilla #1 (Ash Bypass)

Avg 0.09 1.41 0.008 1.75 n 8 8 8 7

cv 0.77 0.04 2.70 0.05 Chowchilla #2

(Main) Avg 0.24 1.40 0.097 1.63

n 14 14 14 14 cv 0.28 0.09 0.56 0.07

James ID H Lateral

Avg 0.10 1.46 0.052 1.74 n 10 10 9 10

cv 0.60 0.08 0.90 0.09 James ID Main Avg 0.17 1.38 0.003 1.68

n 4 4 4 4 cv 0.94 0.07 1.61 0.14

Figure 7 illustrates the relationship between percentage silt and seepage reduction at various sites. The silt content ranged from about 20 to 50%.


Figure 7. Percent reduction in seepage rate compared to percent silt by site

a Values were adjusted to reflect post-compaction seepage through compacted areas only Table 4 shows the field results of the in-situ compaction. Seepage reduction varied from 16% to 90%. The three sites at which the canal bottom was compacted had much better results than the two sites at which only the sides were compacted.

Table 4. Compaction results comparison of sites Irrigation District Chowchilla WD James ID James ID Henry Miller RD Henry Miller RD

Location Main Canal,

between roads 11-12

Lateral H, from Main

Canal to TO N Main Canal

Swamp 1 Ditch, between Turner

Island & Deep Well East Delta Canal

Compaction Sides Ya Y Y Y Y Bottom N Y N Y Y – with ride-onb

Cost, $US 4,845 3,240 15,800 1,945 3,100 Canal Length, m. 1,292 308 3121 527 921 Canal Width, m. 8.2 4.6 17.7 8.2 5.8

Texture Sandy Loam Sandy Loam Sandy Clay Loam Clay Loam Silt Loam

Pre-Comp. Seepage LPS/m2 x 103 0.800 3.28 0.261 2.41 0.646

Post-Comp. Seepage LPS/m2 x 103 0.674 0.458 0.179 0.259 0.0646

% Seepage Reduction 16 86 31 89 90 Notes: a The Chowchilla WD Main Canal site had sections of rip-rap along the canal banks that could not be compacted, resulting in a lower % seepage reduction b Ride-on indicates a large self-propelled unit


Discussion The three sites that had the sides and bottom compacted (HMRD Swamp, HMRD East, and James H) all had seepage rate reductions greater than 85%. The site with only the sides compacted (James Main) had only 31% reduction (72% when adjusted for compacted areas only), and the site with only parts of the sides compacted (Chowchilla Main) had only 16% reduction (33% when adjusted for compacted areas only). Small-scale testing of undisturbed and disturbed soil samples provided large variations in individual values, as indicated by the coefficients of variation (cv) in Tables 1 and 2. It would have been difficult or impossible to predict field-level reduction in seepage based on the laboratory test results. Table 3 shows that while bulk density values are relatively tightly clustered for a given testing procedure and sampling location, the opposite is true for hydraulic conductivity measurements – as evidenced by the cv values. The undisturbed bulk densities of before and after compaction shown in Tables 1 and 2 were not greatly different, but in all cases there was an increase in bulk density from compaction, as expected. It is hypothesized that as long as only one layer of soil is sufficiently compacted, the hydraulic conductivity will be significantly reduced. It may also be the case that, because within any single canal pool the seepage rate is not uniform, the compaction may have a very significant effect on a relatively small area that might not have been sampled. The cost to sample the soils and conduct laboratory tests of bulk density and hydraulic conductivity is much greater than the cost to compact the canals. Because of the difficulty in comparing laboratory results to field performance, in addition to the large laboratory costs, it is recommended to forgo laboratory tests. Instead, if there appears to be a large amount of seepage and the soils can dry out sufficiently well to allow physical compaction, it is recommended to move straight to compacting the canals.

CONCLUSION For all tested soil types, in-situ compaction with a vibratory roller reduced seepage. The seepage reduction was significant (86 – 90%) when both the sides and bottom were compacted. Since silt contents in the 20% range and the 50% range both had high percent reductions, it appears that within the range observed (20-50%) silt content of the soil does not significantly affect the seepage rate reduction due to compaction. Incomplete compaction, however, appears to greatly affect the effectiveness of the compaction.

ACKNOWLEDGEMENTS

Funding for this project was supplied by the Fresno office of the Mid-Pacific Region of the US Bureau of Reclamation, and by the California State University Agricultural Research Initiative (CSU/ARI).


REFERENCES

ANCID. (2001). “Open Channel Seepage and Control, Vol. 2.1.” Australian National

Committee on Irrigation and Drainage, March 2001. Bader, C. D. (2001). “Soil compaction: Modern solutions to an age-old need.”

Forester Media, Inc. < http://www.gradingandexcavation.com/january-february-2001/soil-compaction-solutions.aspx>, (May 27, 2009).

Burt, C.M., M. Cardenas, and M. Grissa. (2008). Canal Seepage Reduction by Soil

Compaction. Preliminary results, presented at the 2008 International Irrigation Association Technical Conference. Anaheim, CA

Zimbelman, D. (2009). Personal communication. [email protected] (303)

663-0215.

Documents

Canal Seepage Reduction by Soil Compaction - ITRC · Canal Seepage Reduction by Soil Compaction ... The Standard Proctor test (ASTM D698) was developed in the 1920s when static rollers