ENNR 429: Final Year Project

Final Draft Report

Soil erosion control on Banks Peninsula: A Bioengineering approach

Project Supervisor: Dr Tom Cochrane Dominic Fletcher Student Number: 34909188 Third Professional Year Natural Resources Engineering

Submitted: 4th of October

Executive Summary Human induced topsoil erosion is a global problem that causes significant pollution of waterways and loss of soil fertility. Although erosion is a natural process, human activities have greatly increased its occurrence and effects. The main causes of accelerated topsoil erosion are deforestation and inappropriate land management practices. On the Port Hills, which are part of Banks Peninsula, accelerated topsoil erosion is a concern. Historic deforestation and current pastoral farming have resulted in loss of workable land and sediment pollution of the local streams. Urban development has also contributed to accelerated topsoil erosion, as the soil is often left exposed during construction. The most common types of erosion process that occur in the Port Hills are gullying and mass movement, although sheet and stream bed/bank erosion are also significant. The soil type most affected by this erosion is Port Hill’s loess, due to its high erodibility and prevalence. This project compares the effectiveness of two erosion control methods at reducing hydraulic surface erosion on Port Hills’ loess in the Port Hills. The two methods tested were planting native vegetation with compost, and applying polyacrylamide. These two methods were tested for a variety of slopes in the lab and rainfall intensities at a field site in Bowenvale Reserve, and the runoff properties were measured. It was found that polyacrylamide was very ineffective at reducing hydraulic surface erosion on Port Hill’s loess at both the field site and for the lab slopes tested. Compost combined with the native vegetation proved to be a very effective erosion control method in all instances. It was recommended that future work on hydraulic surface erosion in the Port Hills include the use of GIS and WEPP. Both programs can be used (combined or separately) for predicting the amount of erosion expected if compost and native vegetation are used as an erosion control in selected areas.

i

Table of Contents Executive Summary

i

1

Introduction

1

1.1 Background

1

1.2 Objectives

1

2

3

4

5

6

7

8

Hydraulic surface erosion mechanics

3

2.1 Introduction to hydraulic surface erosion

3

2.2 Splash erosion

3

2.3 Sheet erosion

4

2.4 Rill erosion

4

2.5 Gully erosion

5

History of erosion on Banks Peninsula

6

3.1 History and description of Banks Peninsula

6

3.2 Types of erosion on the Port Hills

8

Bioengineering soil erosion control methods

11

4.1 Soil protection techniques

11

4.2 Vegetation as an effective soil protection technique

13

4.3 The advantages of native vegetation

15

Soil amendments for erosion control

16

5.1 Polyacrylamide

16

5.2 Compost

17

Field testing

19

6.1 Purpose of field testing

19

6.2 Field site description

19

6.3 Field apparatus

21

6.4 Field data collection procedure

22

Lab testing

23

7.1 Purpose of lab testing

23

7.2 Lab apparatus

23

7.3 Rainfall simulator calibration

25

7.4 Lab procedure

28

Results

29

8.1 Field Results

29

8.2 Lab Results

30

8.3 Comparison of results

31

ii

9

Discussion

32

9.1 Effectiveness and applicability of each erosion control measure

32

9.2 Cost estimate of compost and native vegetation method

33

10

Difficulties encountered

34

11

Conclusions

35

12

Recommendations

36

13

Acknowledgements

37

14

References

38

Appendix A: Elevation view of Bowenvale Reserve and the field site.

41

Appendix B: Hill Laboratories Soil Analysis Form

42

Appendix C: Schematic of field site runoff diversion equipment

44

Appendix D: Lab apparatus schematic

45

Appendix E: Rainfall simulator calibration data and rainfall distribution graphs

48

Appendix F: Field site experimental data and calculations

52

Appendix G: Laboratory experimental data and calculations

54

List of Figures Figure 2.5.1: The process of tunnel-gully formation. ........................................................5 Figure 3.1.1: Banks Peninsula in proximity to Christchurch. (Reproduced from http://artworks.freeparking.co.nz/newssite/countryglen/news.php?&pg=3) ...............6 Figure 3.1.2: New Zealand Fundamental soils layer of Banks Peninsula showing the distribution of soils.....................................................................................................7 Figure 3.2.1: Collapsed tunnel gullying at Summervale Reserve in the Port Hills. (Reproduced from Trangmar, 2003)..........................................................................9 Figure 3.2.2: Tunnel gullies in Westmorland, Port Hills 1977. (Reproduced from Trangmar, 2003) .......................................................................................................9 Figure 3.2.3: Worsley Spur 1977. Sheet, rill and gully erosion causing significant impacts to the Port Hills landscape. (Reproduced from Trangmar, 2003) ............................10 Figure 4.1.1: Erosion control nets made from jute fibre. (Reproduced http://www.beltonindustries.com/erosion_short.html) ..............................................12 Figure 4.1.2: Erosion control net made from coir fibre. (Reproduced from http://www.cocopeat.com.au/products/erosionControlNets.asp) .............................12 Figure 4.1.3: Diagram showing a secured live brush mat. (Reproduced from MNSPN, 2006) .......................................................................................................................13 Figure 5.2.1: Compost windrow at Timaru recycling centre. The compost was created from domestic green waste. ....................................................................................17 Figure 5.2.2: A blower truck used by Rural Supply Technologies to apply compost (Reproduced from Coulson, 2005) ..........................................................................18 Figure 6.2.1: Bowenvale Reserve with respect to other facilities on the Port Hills. ........19 Figure 6.2.2: Photo of the field site prior to construction of testing apparatus. ...............20 iii

Figure 6.3.1: Field site schematic showing experimental set up.....................................21 Figure 6.3.2: Two photos showing the constructed field site at Bowenvale Reserve. ....22 Figure 7.2.1: Schematic of lab apparatus and set up. ....................................................23 Figure 7.2.2: Photos of the finished lab apparatus .........................................................24 Figure 7.2.3: Photo showing the runoff collection tubs and the method for covering them. ................................................................................................................................25 Figure 7.3.1: Rainfall simulator as set up over lab apparatus for calibration and testing 25 Figure 7.3.2: The control box for the rainfall simulator ...................................................26 Figure 7.3.3: Simulated rainfall variation on horizontal surface at 0.67 m. See Appendix E for more details. ...................................................................................................27 Figure 7.3.4: Variation in simulated rainfall for 290 slope over the three test plots. ........27 Figure A 1: Altered aerial photo of Bowenvale Reserve on the Port Hills showing field site. (Reproduced and altered from http://www.ccc.govt.nz/ Parks/Natural Areas/ port_hills_recreation.asp, (2006))............................................................................41 Figure E 1: Simulated rainfall variation under 1 sec sweep delay for each of the 4 rows of catch cans. ..........................................................................................................48 Figure E 2: Simulated rainfall under 2 secs of sweep delay ...........................................49 Figure E 3: Variation in simulated rainfall over the length of the simulator for 1 sec of sweep delay. ...........................................................................................................49 Figure E 4: Simulated rainfall variation over simulator length for 2 sec sweep delay. ....50

List of Tables Table 3.1.1: Typical nutrient loss if 1cm of topsoil is eroded (i.e. 100 tonnes/hectare eroded) (CRC, undated)............................................................................................7 Table 6.2.1: Soil data as obtained from Hill Laboratories (Reproduced from Analysis Form, attached as Appendix B)...............................................................................20 Table 8.1.1: Summary of the two observed storms and their impacts on the field plots .29 Table 8.2.1: Averaged results from the three lab experiments for each of the three slopes (4.80 (Notch 1), 15.40 (Notch 2), and 28.60 (Notch 3)) after 26 mm of simulated rainfall. ....................................................................................................30 Table 8.2.2: Schedule of lab testing ...............................................................................31 Table 8.3.1: The effectiveness of the two erosion control methods when compared with current soil coverage conditions..............................................................................31 Table 9.2.1: Cost of a 20 mm compost blanket in Christchurch. ....................................33 Table 9.2.2: Cost of purchasing PB5 size Carex flagellifera...........................................33 Table A 1: Soil properties obtained from the NZ fundamental soils layer for the Bowenvale Reserve site. This data proves that the soil tested is Port Hill’s loess. .41 Table E 1: Distribution and numbers of catch cans for the widthwise and lengthwise variation in simulated rainfall ...................................................................................48 Table E 2: Rainfall depth calibration data for Notch 3. ...................................................51 Table F 1: Data and calculations from Storm Event 1 ....................................................52 Table F 2: Data and calculations from Storm Event 2 ....................................................53 iv

Table G 1: Data from the Notch 1 (4.80) experiments. The saturated data values were used for all reported results.....................................................................................54 Table G 2: Data from Notch 2 (15.40) experiments. Saturated values reported. ............55 Table G 3: Data from Notch 3 (28.60) experiments. Saturated values reported. ............56

v

1 1.1

Introduction Background

Soil erosion is a significant problem facing not only New Zealand, but also the rest of the world. Although erosion is a natural process, the scale and impacts of it have been greatly increased due to anthropogenic causes. Of the different erosion processes, hydraulic erosion has the greatest impact, affecting 1094 x 106 hectares (Lal, 1994) of global land area, which is approximately 8.4 % of the total land. It is often the nutrient rich topsoil that is eroded, leaving the land less fertile and in extreme cases unable to support vegetation. Eroded sediments often end up in waterways, polluting them and damaging aquatic ecosystems. A significant cause of increased erosion is change in land use, especially pertaining to deforestation of previously stable hill country for arable farming. Removing vegetation from hill slopes destabilises the topsoil, as the roots that previously bound the soil decompose and exposes the soil layer to surface erosion by wind and water. Replanting the slopes with agricultural grass varieties, for livestock grazing, often fails to prevent further surface erosion, because the grasses fail to provide adequate protection and stabilisation to the soil matrix. In addition to this, grazing livestock can add to erosion by damaging vegetation and shearing the soil surface with their hooves (Bechmann and Stålnacke, 2005). On Banks Peninsula in Canterbury, hill slopes have been historically cleared and converted for agricultural grazing. Current urban development involves stripping vegetation cover, leaving the soil surface bare for significant periods of time. These practices have caused accelerated erosion of the highly erodible Port Hill’s loess soils and increased sediment input into the local streams and rivers. These soils are eroded mostly by hydraulic surface erosion processes such as rilling and gullying (Trangmar, 2003). This has been identified as a large enough concern, such that soil conservation work is now being done by the Christchurch City Council and Environment Canterbury to reduce the amount of soil eroding and especially entering local waterways. Sediment adversely affects the water quality of waterways by increasing the suspended sediment concentration, turbidity and nutrient levels of the water. These increased levels can result in the physical smothering of aquatic biota, destruction of suitable habitat and potentially eutrophication of the waterway (Painter, 2005). 1.2

Objectives

The expected outcome of this project was to determine the most effective method for reducing topsoil erosion of Port Hill’s loess. This control method can then be applied within urban developments, such as new housing subdivisions and roads, to reduce the amount of soil eroded and entering the region’s waterways. The aim of this project was to compare the effectiveness of two different soil erosion control methods in resisting hydraulic surface erosion of Port Hill’s loess soil, for specific conditions on an area of Banks Peninsula called the Port Hills. The two erosion controls Dominic Fletcher

34909188

Final Year Project

11/01/2007

1

methods are the planting of native vegetation with compost, and the application of polyacrylamide (PAM). These two methods will be compared with bare soil, for a variety of slopes and rainfall intensities, and the applicability of each for specific situations discussed. The effectiveness was evaluated by measuring the runoff volume and quality, from both field and lab plots containing the two erosion controls, and comparing these with the runoff from bare soil and existing vegetation cover plots. The objective of this project was to determine the most effective erosion control method from the two methods tested, for a slope range of 8.4% to 48% and varying rainfall intensities.

Dominic Fletcher

34909188

Final Year Project

11/01/2007

2

2 2.1

Hydraulic surface erosion mechanics Introduction to hydraulic surface erosion

Hydraulic surface erosion is the movement of soil particles by water, which in natural systems comes from precipitation, especially rainfall. There are four distinct categories of hydraulic surface erosion: splash, sheet, rill and gully. The occurrence of each of these processes depends on the rainfall, soil, topographic and vegetation characteristics of the area being eroded (ARC, 1999). The process can also depend upon the erosion process that preceded it. For example, sheet erosion can lead into rill erosion, which can lead into gullying, although this is not a strict relationship and is dependent on the area characteristics mentioned previously. 2.2

Splash erosion

Splash erosion is where soil particles are displaced by the impact force of water droplets. This type of erosion is mainly dependent on the rainfall, vegetation cover and soil characteristics. The rainfall characteristics most important to splash erosion are the droplet size, intensity and droplet kinetic energy (KE). The droplet size determines the droplet potential energy; the rainfall intensity determines the droplet size and density pattern; and the kinetic energy is a measure of the impact force. These characteristics determine the force a droplet can apply to a soil particle to move it (Spon, 1995). Vegetation can intercept rainfall, as determined by the basal area (area sheltered by leaves), and either store this on the leaves allowing it to evaporate, or temporarily detain it until it runs down the plant stalk and onto the ground (Temporarily Intercepted Flow, TIF). Hence, vegetation can alter the erosive energy of the water droplets, depending on the vegetation properties. This does, however, sometimes mean that the water is collected into larger droplets, thereby increasing the splash erosion (Spon, 1995). For example, broadleaf vegetation can concentrate the raindrops due to a high basal area and high canopy vegetation can increase the potential energy of these droplets increasing their erosive energy. An important soil property is cohesion. This is a measure of the soil’s ability to stick together and resist movement. The more cohesive a soil, the better able it is to resist impact forces due to raindrops. This cohesion however can depend on the moisture content of the soil, which in turn is determined by vegetation and climate. For example, a dry soil may have less cohesion than when it is wet and so is more susceptible to splash erosion in dry conditions.

Dominic Fletcher

34909188

Final Year Project

11/01/2007

3

2.3

Sheet erosion

Sheet erosion occurs on soil surfaces exposed to rainfall, and is where soil particles, either dislodged by splash erosion or by turbulent overland water flow, are moved down slope by a ‘sheet’ of water. This sheet of water further scours soil surfaces down a slope, eroding more soil particles. This form of erosion depends upon the occurrence of runoff, as dictated by the rainfall intensity and the infiltration capacity of the soil. The infiltration capacity is strongly influenced by the moisture content, porosity and stratification of a soil. Vegetation can increase the infiltration capacity of a soil, because it alters the soil moisture content and can increase the porosity due to root action (Spon, 1995). The long-term presence of vegetation can also increase the organic matter in the soil, which is significant for runoff and erosion, as organic matter improves soil structure, and increases permeability and water holding capacity (ARC, 1999), increasing the infiltration capacity of the soil. Soil fertility is also improved by organic matter allowing further colonisation of plant species and hence more soil stabilisation. The topography of an area strongly influences the extent of sheet erosion. Trangmar (2003) states that the intensity of sheet erosion increases with the slope, because this increases the runoff velocities thereby increasing the erosive power of the water. Long continuous slopes also allow runoff to build up velocity and concentrate flow, increasing the erosiveness of the water (ARC, 1999). Vegetation also has an influence on the erosiveness of sheet erosion. Suitable vegetation can slow the runoff velocity, hold the soil in place and even filter suspended sediments out off the runoff. The sediment filtering ability of vegetation is determined by its density, shape and resilience (Rickson and Morgan, 1995) 2.4

Rill erosion

Rill erosion occurs in localised regions where sheet erosion has been more severe (than in other regions) and has incised a small channel into the hill slope. Because these channels are lower than the surrounding topography, water flows into them and hence the channel is further eroded (Trangmar, 2003). These small channels are termed rills and by definition should be less than 300 mm deep (ARC, 1999). Rills larger than 300 mm deep are defined as gullies. The occurrence of rill erosion depends on the soil characteristics and often, spatial variations in the soil, vegetation and topography. Of special concern to this project is the formation of preferential flow paths around the base of vegetation. Under certain conditions, the density of vegetation can cause runoff to concentrate in particular flow paths and hence create rills. Special consideration needs to be given to plant spacing, in order to avoid the formation of rills.

Dominic Fletcher

34909188

Final Year Project

11/01/2007

4

2.5

Gully erosion

There are two types of gully erosion: Deep Rill and Tunnel. Both types of gullying are influenced by the soil properties, such as depth of soil and stratification, and the topography. Often vegetation does not have any affect on gullying, because the extent of erosion is such that vegetation can not hold the soil together. Deep Rill gullying is where rills have become sufficiently deep (Section 2.4). The channelled water further increases the erosion within the gullies, eventually causing collapse of the gully banks and forming a larger gully. Once this process has started, it is very hard to stop. Tunnel gullies are a specific type of gullying, where instead of eroding the soil surface, the water seeps into the soil and forms underground cavities, which eventually collapse into open gullies (Figure 2.5.1). Tunnel-gullies can be identified by the presence of: water and sediment discharges from surface cracks or small holes (tunnel outlets), at changes in slope angle or cut faces; collapse holes aligned down slope; and open gullies up to 2 m deep filled with collapsed debris (Trangmar, 2003). Tunnel-gullies often occur where a soil is dry and desiccation cracks form on the surface, as runoff can flow into these cracks and erode the subsoil (Trangmar, 2003). It follows that the moisture content of the soil influences the formation of tunnel gullies. It is therefore also important to note that vegetation can increase the likelihood of tunnel gullies occurring, by reducing the soil moisture content until cracks occur (ARC, 1999).

Figure 2.5.1: The process of tunnel-gully formation. [Reproduced from Trangmar (2003)] Dominic Fletcher

34909188

Final Year Project

11/01/2007

5

3 3.1

History of erosion on Banks Peninsula History and description of Banks Peninsula

Banks Peninsula was initially an island formed by volcanic activity, which was eventually connected to the mainland (South Island) by the spreading alluvial plains created by the Waimakariri River. Banks Peninsula is located just south of Christchurch City, with some suburbs located on or at the base of a part of the Peninsula, the Port Hills, which overlook the Canterbury plains and Lyttleton Harbour (Figure 3.1.1)

Christchurch Port Hills

Figure 3.1.1: Banks Peninsula in proximity to Christchurch. (Reproduced from http://artworks.freeparking.co.nz/newssite/countryglen/news.php?&pg=3) A significant amount of the topsoil found on Banks Peninsula is Port Hill's loess (Figure 3.1.2); a sandy loam formed from wind blown volcanic soil particles. This soil is typically highly dispersive (structurally unstable), has poorly aggregated subsoils and has a high fine sand content and exchangeable sodium percentage, in addition to a low organic matter content (Trangmar, 2003). Erosion and redeposition of loess topsoil has been occurring in the Port Hills since the onset of loess deposition in the Pleistocene period (Trangmar, 2003). However, this erosion has been greatly increased since the first 50 years of European settlement, when most of the native forest on the Port Hills was felled or burned for timber or grazing land (DoC, 2006). This topsoil erosion has decreased the fertility of the land, reducing the sustainable crop yield that can be taken from it. Table 3.1 illustrates the typical nutrient loss for soil erosion in Canterbury, from I cm of topsoil. However, these values are an average and will most likely be less for loess, due to its low organic matter component.

Dominic Fletcher

34909188

Final Year Project

11/01/2007

6

Table 3.1.1: Typical nutrient loss if 1cm of topsoil is eroded (i.e. 100 tonnes/hectare eroded) (CRC, undated) Nutrient Total N Total P Total K Total S Total Mg Total Ca

Amount lost (kg/ha) 350 90 1000 60 650 1050

Figure 3.1.2: New Zealand Fundamental soils layer of Banks Peninsula showing the distribution of soils. Dominic Fletcher

34909188

Final Year Project

11/01/2007

7

3.2

Types of erosion on the Port Hills

Trangmar (2003) states that the main types of erosion, as determined by the most aerially extensive, that are currently occurring in the Port Hills are tunnel-gully and rapid mass movements or soil slips. Although less extensive, soil creep, rockfall, sheet wash and stream bed/bank erosion are also present in the Port Hills. However, this project is concerned with reducing erosion due to land use change, such as during urban development, which is the erosion due to surface hydraulic processes. Hence only tunnel-gully and sheet wash erosion will be discussed in detail below. Tunnel-gullies are believed to be a major contributor of sediment to Port Hill’s streams, influencing the sedimentation, water quality and stream ecology in the Heathcote River and the Avon-Heathcote Estuary. Stream bed/bank erosion is also considered a potentially significant contributor to sediment pollution in the local waterways. Tunnel-gully erosion on the Port Hills mainly occurs in the loessial soils, due to their high structural instability as discussed previously (Trangmar, 2003). The significant content of expandable clay particles also enables soil shrinkage in summer, leading to the formation of desiccation cracks and increasing the potential for tunnel-gullies as discussed in Section 2.5. Tunnel-gullies are more commonly found on slopes ranging between 10 and 28 degrees, and are the dominant erosion type on slopes with a sunny aspect (i.e. NE, N, NW and W), due to greater seasonal changes in soil moisture content (leading to desiccation cracks and hence tunnel-gullies). Conversely, tunnel-gullies are less common on shaded, moister slopes (i.e. E, SE and S), due to more constant soil moisture contents (Trangmar, 2003). Trangmar (2003) states that the following land management practices can increase the occurrence of tunnel gullying in the Port Hills. ‰ ‰ ‰ ‰ ‰

Uncontrolled stormwater discharge onto loessial soils (i.e. road runoff from the Summit Road or Rapaki Track). Depletion of covering vegetation due to excessive stock grazing (i.e. sheep eating the grass down to the stubble) Disturbance of the subsoil from engineering activities (i.e. earthworks, foundation works, overburden) Use of unstabilised recompacted loess as a fill material. Cut faces creating tunnel outlets and hence improving the subsurface tunnel flow.

An impact of tunnel-gullying is the disruption of the ground surface, which decreases the potential carrying capacity of pastoral land and creates a hazard to stock in the form of collapse holes (Figure 3.2.1 and Figure 3.2.2). Tunnel-gullies can also route water flow onto/into soil zones prone to soil creep and/or slipping, further increasing the erosion. The sediment created from this erosion can block pipes, drains and gutters during storm events, resulting in localised flooding and property damage. These off-site effects of tunnel-gully erosion all increase the costs of maintaining the local waterways (Trangmar, 2003). Dominic Fletcher

34909188

Final Year Project

11/01/2007

8

Figure 3.2.1: Collapsed tunnel gullying at Summervale Reserve in the Port Hills. (Reproduced from Trangmar, 2003)

Figure 3.2.2: Tunnel gullies in Westmorland, Port Hills 1977. (Reproduced from Trangmar, 2003) Sheet wash or sheet erosion is common on unsealed tracks and paths, and especially bare excavated soil surfaces, within the Banks Peninsula area. However, the dominant land use on the Port Hills is pastoral and hence sheet erosion is not prevalent, with the notable exception of construction and excavation sites (e.g. roads, residential developments). As mentioned in Section 2.1, sheet erosion can lead to rill and gully erosion. This happened in the late 1970’s during the development of the Westmorland subdivision, and at Worsley Spur (Figure 3.2.3), resulting in construction of settlement ponds to reduce the sediment input into the Heathcote River (Trangmar, 2003). This illustrates that sheet erosion has been a significant process on the Port Hills in recent history, during significant land use changes (i.e. in this case urban development), and so should be considered when attempting to reduce the impacts of erosion. Dominic Fletcher

34909188

Final Year Project

11/01/2007

9

Figure 3.2.3: Worsley Spur 1977. Sheet, rill and gully erosion causing significant impacts to the Port Hills landscape. (Reproduced from Trangmar, 2003)

Dominic Fletcher

34909188

Final Year Project

11/01/2007

10

4 4.1

Bioengineering soil erosion control methods Soil protection techniques

Schiechtl and Stern (1996) state that “Soil protection techniques rapidly protect soil, by means of their covering action, from surface erosion and degradation. They improve water capacity and promote biological soil activity.” It is important to note that this refers only to surface erosion, such as sheet and rill, and involves the promotion of biological activity, such as by vegetation and insects. Although this project only considers the effectiveness of two such soil protection techniques, there are many other techniques that may also be applicable for use in the Port Hills. Some of these techniques, such as turfing, grass-seeding, direct seeding, erosion nets, seed mats, brush mats, are discussed briefly below. Turfing is where mats of grass (or turf), that are cut to include the root zone, are placed on a slope and staked down to secure them. The turf mats then grow, attaching themselves to the slope with their roots and thereby covering and stabilising the topsoil. This technique essentially involves manually vegetating a slope, thereby avoiding delays in growing time and providing almost instantaneous protection. However, decomposition and desiccation are two concerns with turfing, and so applied turf needs to be carefully maintained initially, so that it does not dry out or rot. Turfing is also an expensive, labour intensive method and so is not always suitable (ARC, 1999). There are two types of turf, Natural and “Rolled”. Natural turf is thicker making it stronger and more resilient to adverse conditions (e.g. drought…). This means that natural turf is more versatile and can be used for a greater range of slopes and conditions. Because “Rolled” turf is generally thin, it is only applicable for low stress locations on wellprepared moderate slopes (Schiechtl and Stern, 1996). Grass seeding involves spreading seed out on a slope and maintaining it (e.g. watering, fertilisation, and keeping birds away) until the seed grows and covering the slope and stabilising the topsoil. There are several types of commonly used seed mix, such as hayseed and standard seed, which can be applied to a slope by a number of methods, such as hydroseeding, dry seeding and mulch seeding. Grass seeding provides little or no initial protection against topsoil erosion. Standard seeding is simple and cheap but requires the addition of fertile topsoil for most applications in order for it to grow. Hayseed is used together with other seeds and is more effective in sub-alpine/alpine zones. However, there are limited quantities available and this method adds 10-30% to the costs of other seeding types, such as standard seeding (Schiechtl and Stern, 1996). Hydroseeding uses water, combined with the seeds, fertiliser, soil improver and a binding agent, to apply the seeds to a slope. This method is useful for steep, rocky and inaccessible slopes but requires adequate site access for machinery. Applying seed by this method also provides limited resistance to mechanical erosion (e.g. splash, rill…).

Dominic Fletcher

34909188

Final Year Project

11/01/2007

11

Dry seeding is very similar to hydroseeding, the only difference being that dry seeding uses compressed air to apply the seeds as opposed to water. This has the benefit of not requiring a water source to be used. Mulch seeding is the application of seeds embedded with a thick layer of (usually organic) mulch or compost. This method provides immediate, effective protection to the topsoil from mechanical erosion, and is especially useful in winter when there is no plant growth (ARC, 1999). However, this method is only a short term solution and is only feasible for highly accessible sites, greater than 1 hectare, due to economic factors (i.e. economies of scale) (Schiechtl and Stern, 1996). Direct seeding of trees and shrubs is a low cost, simple method most suited for difficult sites. This method does not provide immediate erosion protection, but is very good for accelerating natural succession, such that the slope will be permanently protected when the plants have established an ecosystem. This method does however require species pre-adapted to the site-specific conditions. For example, Dicksonia fibrosia (a type of native tree fern) will not survive on highly exposed faces of the Port Hills that are subjected to hot dry nor-westerly winds, as it is pre-adapted to live in deep rainforest. Erosion control nets are used in conjunction with other methods (such as those mentioned previously) to reinforce a soil surface and prevent sediment movement. These nets are commonly made from jute fibre (Figure 4.1.1), synthetic materials and metal wire, although other materials such as coir fibre (from coconuts) are also used (Figure 4.1.2).

Figure 4.1.1: Erosion control nets made from jute fibre. (Reproduced http://www.beltonindustries.com/erosion_short.html)

Figure 4.1.2: Erosion control net made from coir fibre. (Reproduced from http://www.cocopeat.com.au/products/erosionControlNets.asp) Dominic Fletcher

34909188

Final Year Project

11/01/2007

12

Seed mats protect against mechanical erosion, provided no water flows beneath them. These mats are made from geotextile fabric filled with seeds, fertiliser and soil, and cost more than seeding methods. Although seed mats are durable, they are only suitable for even ground and are predominantly used in shallow waterways. Live brush mats (or brush mattresses) are an immediate and effective erosion protection measure, and are made from live stakes, branch cuttings and root bulbs wired together and staked to a slope (MNSPN, 2006) (Figure 4.1.3). This method is very labour intensive in the construction phase and moderately expensive (Schiechtl and Stern, 1996). There are also concerns with the common usage of willows, as they are not preadapted to all sites and they are susceptible to disease.

Figure 4.1.3: Diagram showing a secured live brush mat. (Reproduced from MNSPN, 2006) There are a number of supporting practices that can increase the effectiveness of these erosion control techniques. The addition of fertile topsoil as a growth medium for plants, to areas where the subsoil will not support adequate growth is a common practice. Surface roughening is another support practice that assists sediment entrapment by reducing the runoff velocities and capturing mobile sediment, providing the roughness furrows are perpendicular to the slope (i.e. follow the natural slope contours) (ARC, 1999). 4.2

Vegetation as an effective soil protection technique

“Vegetation is without question the most effective long term form of erosion control for protecting surfaces that have been disturbed.” (ARC, 1999). As removal of site adapted vegetation is often the cause of excess erosion, it follows that replanting a slope with appropriate vegetation is the most obvious solution. Gray and Sotir (1996) state that soil loss due to rainfall erosion can be decreased a hundred fold by maintaining a dense cover of vegetation. In New Zealand, vegetation is often used effectively in the stabilisation and rehabilitation of eroded lands and riparian zones (Phillips citing van Kraayenoord & Hathaway, 1986).

Dominic Fletcher

34909188

Final Year Project

11/01/2007

13

Schietchtl and Stern (1996) state that the reason vegetation is an effective erosion control is because “The presence of large numbers of plants seeds or plant parts per unit surface area protects the soil from the deleterious effects of mechanical forces. Moreover vegetation covers improve soil climate and water capacity which in turn lead to better growing conditions for succeeding plant life”. Vegetation protects the soil by either shielding it from rainfall and runoff, or reinforcing the soil matrix. Morgan and Rickman (1995) explain the reinforcement of soil as follows, “The roots and rhizomes of the vegetation interact with the soil to produce a composite material in which the roots are fibres of relatively high tensile strength and adhesion embedded in the matrix of lower tensile strength. The shear strength of the soil is therefore enhanced by the root matrix”. In addition to this, vegetation is also known to act as a filter or barrier to sediment-laden runoff (Phillips citing Greenway, 1987), and reduces wind erosion by exerting drag on the airflow and thereby reducing the shear velocity of wind (Morgan and Rickman, 1995). Once vegetation is established, it will indefinitely protect the slope from excess erosion, because it creates a resilient, self-sustaining ecosystem that can buffer climatic change and variation (Phillips, 2005). This is an example of an ecological service provided to society. This service has an economic benefit because once a slope is planted with established vegetation, there will be none or little maintenance required for the vegetation to continue to provide effective surface erosion protection. There are however several factors that need to be considered when replanting an eroded slope with certain chosen vegetation. These factors can be divided into the following sub-categories: suitability of vegetation to site, effectiveness, ease of use, and economics. The chosen plant species must be pre-adapted to the site conditions, non-invasive and resistant to local pests and disease. The vegetation must also effectively prevent erosion in the accepted and required time period. The ease of use category refers to the availability of species, their ease of propagation and establishment, and their growth habits and rates (Phillips, 2005). The time of planting affects the cost of acquiring the desired plants. Often is it best to order the plants a long time before they are needed, ensuring that they are available and at their least expensive. The aesthetic qualities (e.g. smell, sight, intrinsic value to indigenous cultures) of particular vegetation should also be considered, especially for urban areas (Gary and Sotir, 1996). It should be made clear that the erosion mentioned here applies firstly to excess erosion, as defined by erosion greater than the ‘natural’ or non-human influenced levels, and that erosion is a natural process and should be expected to occur on some small scale. Secondly, the effects of vegetation mentioned here are only for surface erosion. Mass movement, such as rockfall, slips and fault block movement, can occur regardless of the vegetation cover and in some cases vegetation can even cause mass movement erosion (e.g. root action).

Dominic Fletcher

34909188

Final Year Project

11/01/2007

14

4.3

The advantages of native vegetation

As mentioned in Section 4.2, vegetation pre-adapted to a specific site is the most effective in controlling surface erosion for an indefinite period. This has implications for the use of native plant species for sustained erosion control, as they are pre-adapted for their indigenous environments. In New Zealand there is now an approach to using vegetation for erosion control that considers soil type and location (based on DoC’s Ecological Regions) and matches these to specific native plant species (Dodd, Douglas, et. al., 2006). This approach is based on the assumption that native vegetation has evolved along side the soil at a location, and so is best suited to growing in it. However, vegetation for erosion control is also chosen by its growth form, height and initial maintenance requirements and so even if a native plant is adapted to a certain site’s conditions it may not be appropriate given its other properties (Phillips, 2005). It should also be noted that exposed sites cannot be immediately planted with native forest species that are pre-adapted to the region and soil, because they require certain conditions during the seedling and intermediate growth stages, such as adequate moisture and shade. In order for slopes to be reforested, natural succession needs to be seeded by first planting second order succession species, such as tussocks. Only once these plants are present, can certain other plants (e.g. Pittosporum sp.) begin to grow and eventually replace them. The more established an ecosystem becomes the more resilient it is and therefore more capable at sustained erosion control. It is important that the process of natural succession also be considered when planting native vegetation for erosion control. The role of exotic species as nursery plants, or plants that create adequate microclimate conditions such that other plants can grow beneath them, is only recently being explored. For example, gorse hedge was the focus of an expensive and mildly unsuccessful operation to exterminate invasive exotic species, on the premise that they made land unusable for grazing and prevented the growth of native species. However, it is now known that gorse provides an excellent form of shelter for native seedlings, that when fully grown out compete the gorse and kill it. The use of nursery plants for erosion control has at present not been thoroughly considered. Native seeds are currently being used in hydroseeding techniques (Coulson, 2005b). Historically native hydroseeding has produced highly variable results, often disappointing with poor establishment and survival rates, due largely to the poor quality of seeds. However the results from hydroseeding are improving, due to dormancy testing of native seeds. A concern with native hydroseeding is the long establishing time, typically 2-3 years, which limits its use to areas where an immediate visual impact is not required, such as quarries and mines (Coulson, 2005b). More research into improving the survival rates of native seeds is also required before native hydroseeding can be more feasible. At present there is very little quantitative data the rates of groundcover and canopy spread of native plant species, or on the ability of native plant species to filter sediment from runoff (Phillips, 2005). More research is required before the benefits of using native vegetation for surface erosion control can be evaluated and discussed. Dominic Fletcher

34909188

Final Year Project

11/01/2007

15

5 5.1

Soil amendments for erosion control Polyacrylamide

Anionic polyacrylamide (PAM) is a soil flocculent used to bind soil particles together and hence prevent them from eroding. This compound also promotes infiltration in some soils, thereby increasing the effective infiltration capacity and hence reducing runoff. Polyacrylamide is created from the polymerisation of accylmide and other related monomers (Partington and Mehuys citing Barvenik, 1994). It is only effective as a soil amendment if it is negatively charged (anionic), with a charge density between 18 and 30%, has a molecular weight of 12-15 x103 kg/mole and contains an active ingredient content of at least 80% (Partington and Mehuys citing Sojka and Lentz, 1999, Green et al., 2000). Polyacrylamide has been used successfully for short-term soil erosion control in many agricultural soils and situations, especially in furrowed land (Ajwa and Trout, 2001). However, the effectiveness of PAM is often dependent on the particular site conditions and has been shown to be ineffective in certain situations (Ajwa and Trout, 2001 and Mehuys and Partington, 2005). The most relevant site conditions appear to be soil type, rainfall intensity and volume, and land use practices (Choi et. al., 2001). Slope does not appear to significantly affect the effectiveness of PAM. The concentration of applied PAM also influences its effectiveness as an erosion control. Because the required concentration depends on the soil properties, the most effective concentration differs for individual soils and hence is very site specific. Guideline ranges give some indication of suitable concentrations but these can be inappropriate for some soil types. The method of PAM application influences its effectiveness. Flanagan et. al. (2001) states that liquid application of PAM and subsequent drying reduces erosion and runoff considerably more than dry granule application of PAM. Increasing the application concentration of PAM has also been shown to reduce the volume of erosion and runoff for some soils (Alberts et. al., 2001), but appears to have negligible effects other soils, such as sandy loams (Ajwa and Trout, 2001). The number of applications over a season also seems to influence the effectiveness of PAM. Aase and Bjorneberg (2000) concluded that whilst both single and multiple applications of PAM reduced runoff and soil loss for their specific situation, multiple applications resulted in a greater reduction and hence effectiveness. However, the United States Practice Standard for Anionic Polyacrylamide (2001) states, in the case of agricultural (or horticultural) irrigation of PAM, that application should only occur at the start of the season to prevent over-application and waterlogging of the soil. This implies that too much PAM will increase the infiltration into a soil to the point of saturation, as runoff is inhibited and hence more water enters the soil. However, waterlogging is still a function of the soil properties and rainfall intensity/volume, and hence the amount of PAM that will cause waterlogging depends of these two factors.

Dominic Fletcher

34909188

Final Year Project

11/01/2007

16

5.2

Compost

Compost can be an effective soil amendment for reducing surface erosion. It reduces erosion by covering the exposed soil, preventing particle movement, and by absorbing water, such that runoff is less severe and takes longer to occur. Compost also continues to protect the soil under wet/saturated conditions, with a high structural stability over its useful life (Legacy and McCoy, 2003). The use of compost over time can also increase the infiltration capacity of the topsoil, as the organic matter content is increased. Compost also encourages faster vegetation re-growth further stabilising soils (Coulson, 2005 and Sherman, 2003). Although use of compost as an erosion control is relatively new in New Zealand, its use has been increasing, following international trends over the last 5 years. This increase in use is due to a greater understanding of its effectiveness, applicability and availability. Compost is considered by the US EPA (Environmental Protection Agency) as a best management practice for storm water management, as it prevents soil erosion, reduces the volume of runoff and hence pollution of stormwater (Goldstein, 2006). In an urban environment, compost can be considered to be an aesthetically pleasing erosion control method (Coulson, 2005), increasing its acceptability and uptake. Compost is also more available in New Zealand now due to increased production from waste recycling initiatives (Coulson, 2005). Figure 5.2.1 shows compost created from domestic green waste in Timaru.

Figure 5.2.1: Compost windrow at Timaru recycling centre. The compost was created from domestic green waste. There are many different methods of using compost to control erosion, such as blankets, filter socks and berms (Coulson, 2005). Compost blankets that cover entire slopes are typically used to control hydraulic surface erosion. However, compost blankets do not control erosion due to concentrated flow (Coulson, 2005) and measures should be taken to prevent concentrated run on from up-slope entering the protected slope (ECAN, 2006). It is also recommended that compost blankets be regularly inspected and maintained, especially after high rainfall events (ECAN, 2006). Dominic Fletcher

34909188

Final Year Project

11/01/2007

17

The factors that influence the performance of compost blankets are the depth of application, and compost properties, such as average particle size (coarseness). Dixon et al. (2004) states that whilst coarser composts can reduce inter-rill erosion and runoff rates, too large a particle distribution results in inadequate surface sealing and hence protection of the soil from splash erosion. The greater the depth of application, the more rainfall can be absorbed; however, past a certain depth the increase in soil protection is negligible (Coulson, 2005). The economics of applying more compost mean that an optimum depth is often decided upon based on a cost/benefit ratio. There are some issues associated with the use of compost. The standardisation of compost quality is required (Coulson, 2005), in order to ensure a predictable performance, and to prevent the addition of high concentrations of pollutants, such as heavy metals and pathogens (Sherman, 2003). This standardisation requires a standard compost method, and compost constituent properties, to be specified and ensured. A New Zealand Compost Standard was created and published in December 2005 as NZS4454:2005 for Composts, Soil Conditioners and Mulches (UC Composting, 2006) for the reasons mentioned above, and is in the process of being implemented throughout New Zealand. Different blends of composts are also necessary. For example, it has been shown to be effective to blend 50% wood chips with compost to keep the compost from blowing away in high winds (Sherman, 2003). As such, different types of compost should be made available on a commercial basis for use as an erosion control, in order to ensure predictable and site adapted use of compost. The New Zealand Compost Standard provides for different compost mixtures to be standardised and hence made commercially available. Another issue with the use of compost for erosion control is that of application method. Until recently, compost could only be applied to certain location, due to the fact that it was spread manually by a loader or person. However the advent of pneumatically operated trucks, called blower trucks (Figure 5.2.2), has increased the number of situations in which it is economical and possible to apply compost (Coulson, 2005).

Figure 5.2.2: A blower truck used by Rural Supply Technologies to apply compost (Reproduced from Coulson, 2005) Dominic Fletcher

34909188

Final Year Project

11/01/2007

18

6 6.1

Field testing Purpose of field testing

The purpose of having field test site is to obtain data from the two tested erosion controls for comparison with the lab results. This is important because the lab results are obtained for very specific controlled conditions, which are not imposed under natural field conditions. The results from the field tests determine the effectiveness of each erosion control method under a variety of rainfall intensities and other climatic conditions. It is important that the effectiveness is measured under natural conditions, as this determines the likely performance of the erosion controls in the uncontrolled environment they are to be used in. 6.2

Field site description

Bowenvale Reserve is a Christchurch City Council (CCC) owned recreational reserve in the Port Hills. It is 237 hectares in area with a 12 km perimeter and is surrounded by Victoria Park and Mt Vernon Park (Figure 6.2.1 below). Although part of the reserve is currently in use as pastoral land, Bowenvale Reserve is valued by the CCC as a popular recreational facility for mountain bikers and walkers/runners. This is reflected in The Cashmere Spur And Bowenvale Valley Reserves Management Plan (July 1991), which was designed to recognise and protect the reserve’s “natural values” (CCC, 2006). Bowenvale Reserve can be accessed from Bowenvale Avenue at the base of the hills in Cashmere (CCC, 2006), or from Summit Road along the hill ridge.

Figure 6.2.1: Bowenvale Reserve with respect to other facilities on the Port Hills. (Reproduced from http://www.ccc.govt.nz/Parks/NaturalAreas/port_hills_recreation_victo riapark.asp (2006)) The field site is located at the top of Bowenvale Reserve (Elevation 391.7 m above mean sea level, NZ Map Grid co-ordinates: N: 5733915, E: 2482391 (as obtained from GPS survey data)). Figure A1 in Appendix A shows the location of the field site on an aerial photo of Bowenvale Reserve. The field site faces due west, has an average slope Dominic Fletcher

34909188

Final Year Project

11/01/2007

19

of 12.70, and the local vegetation cover consists of pastoral grasses and sporadic clumps of native tussocks (such as Poa Cita). A photo of the site before construction can be seen as Figure 6.2.2.

Figure 6.2.2: Photo of the field site prior to construction of testing apparatus. Soil samples were taken from the field site and sent to Hill Laboratories for analysis of the soil’s chemical properties. The following soil properties were found as shown in Table 6.2.1 below. As expected, the soil was found to have low phosphorus and organic matter contents, a medium dry density of 910 kg/m3, and a medium Cation Exchange Capacity (CEC). When compared with data from the New Zealand Fundamental Soils Layer (NZFSL), this data appears to confirm that the soil sample is Port Hills Loess (see Table A 1, Appendix A). Table 6.2.1: Soil data as obtained from Hill Laboratories (Reproduced from Analysis Form, attached as Appendix B)

Dominic Fletcher

34909188

Final Year Project

11/01/2007

20

6.3

Field apparatus

The field site plots were set up with the assistance of Peter McGuigan, the Natural Resources Engineering Group Technician. The field apparatus consists of four plots, divided by aluminium sheets to prevent water running on to them from up slope and contamination from the surrounding plots. The four plots are a grassed control (representative of natural current site conditions), bare soil, bare soil with polyacrylamide, and a compost blanket with native vegetation. Figure 6.3.1 shows the field site schematic including all dimensions and labels. A lu m in u m s h eet meta l d iv id ers to p rev en t ru n -o n fro m u p s lo p e an d fro m o th er p lo ts .

600 m m

Co n tro l

Bare s o il

PA M

Com pos t and veg etation

11 0 0 m m

Ru n o ff

20 L Plas tic b u ckets fo r ru n o ff co llectio n

Figure 6.3.1: Field site schematic showing experimental set up. Before the field site facilities could be set up, the funnels and dividers had to be cut and folded from a 5 mm aluminium sheet. This was done under the guidance and assistance of Kevin Wines, Fluids Lab Technician. A design sketch showing all dimensions can be seen in Appendix A, Figure A 2. Aluminium was used as this material resists corrosion and primarily because there was a scrap sheet available from the Fluids Lab. Once all the materials were available and manufactured to specification, the exact location of the plots were marked out using string line. The aluminium dividers were driven into the ground using a metal hammer, the holes for the buckets were dug, and the funnels were driven and dug into place. 4 plastic rainfall gauges were also inserted into the soil, one in each corner of the field site, for the estimation of rainfall events. After this, the turf layer (~30 mm) was removed from the bare soil and PAM plots, and 5.25 L of PAM (concentration 200 mg.L-1, which is equivalent to an application rate of 20 kg.ha-1) was carefully poured onto the PAM plot. 15 Carex flagellifera, with root ball diameter 60 mm and depth 70 mm, were placed into 60 mm diameter holes, spaced 250 mm apart and 50 mm from the plot borders. The foliage was cut to a standard height of 120 mm. After the vegetation was planted the plot was covered in a uniform 20 mm layer of compost. Dominic Fletcher

34909188

Final Year Project

11/01/2007

21

It was deemed necessary to deter members of the public from tampering with the field site. This was done by erecting a low square net wire fence and attaching University of Canterbury branded signage, that informed observers of the purpose for the plots and provided contact details for further enquires. These measures can be seen in Figure 6.3.2, which shows the finished field site plots.

Figure 6.3.2: Two photos showing the constructed field site at Bowenvale Reserve. Shortly after construction was completed however, it was noticed that groundwater was filling the bucket holes and causing the buckets to float out of the holes, lifting the funnels out of place. To remedy this situation, relief trenches were dug into the downhill side of the bucket holes to an approximate depth of 200 mm. 6.4

Field data collection procedure

The following procedure was observed for the collection of field site data. After each rainfall event, the rain gauges were read, the rainfall depth recorded and an average rainfall depth calculated. The gauges were then emptied and returned to the ground. The runoff buckets were lifted out of their holes and the height of runoff measured and recorded. The volume of runoff was then calculated from this height and the area of each bucket. After this the buckets were stirred to ensure a uniform sediment concentration, and a 250 mL runoff sample was collected in a plastic bottle for each plot. The buckets were then emptied, rinsed with water to flush out any remaining sediment, and placed back in the holes. The runoff samples were taken to the Engineering Environmental Lab at the University of Canterbury for further analysis. The suspended sediment was measured as per standard procedure, by weighing a 0.22 micron filter, passing a 100 mL runoff volume through it (aided by ~ 10 MPa suction), and drying the filter in a ~1050C oven for at least 1 hr. The dry sediment laden filters were taken out of the oven and re-weighed to get the mass of sediment and hence an approximate sediment concentration in the runoff.

Dominic Fletcher

34909188

Final Year Project

11/01/2007

22

7

Lab testing

7.1

Purpose of lab testing

The purpose of the lab testing is to determine effectiveness of the two erosion controls under fixed conditions, such as rainfall intensity, duration, slope, and droplet size. To test the effects of varying slope on the sediment concentration and runoff volume, it was necessary to construct an apparatus to hold the test soil plots at a variety of slopes. 7.2

Lab apparatus

Construction of this lab apparatus was an expensive and time-consuming process, due to the fact that the frame had to be constructed out of steel in order to support the 182 kg weight of soil used (0.2 m3 of density 910 kg/m3). The steel frame supports a plastic tub containing three test plots: bare soil with PAM, compost planted with Carex flagellifera, and a grassed control. Each plot has dimensions 600 mm by 1100 mm, which are the same dimensions as the field site plots. Figure 7.2.1 shows a schematic of the apparatus set up, and Figure 7.2.2 shows the completed steel frame holding the three plots. A design schematic of the apparatus can be seen in Appendix D.

PAM

Compost

Control (Natural conditions)

1000 L tub with pump Apparatus Ladder frame (bases in bold, top dashed)

Rainfall simulator

Gate

Figure 7.2.1: Schematic of lab apparatus and set up.

Dominic Fletcher

34909188

Final Year Project

11/01/2007

23

Figure 7.2.2: Photos of the finished lab apparatus After the steel frame was assembled (welded and bolted together) and hot dip galvanised, with the invaluable and considerable assistance of Ian Sheppard and Kevin Wines, Fluids Lab Technicians, a wooden base and plastic tub were attached (Figure 7.2.2). The completed apparatus was then transported up to the Bowenvale field site and the plastic tub carefully filled with soil, with the assistance of Peter McGuigan. The soil was dug out of the ground in squares of approximate spade width (~200mm), after which 50 mm of turf was removed and the soil squares were levelled to 100 mm depths (all measured form the soil surface). These squares were carefully arranged, in the order that they appeared in the ground, into the plastic tub sections. One section was filled with turfed soil for the simulation of current soil coverage conditions in the lab. The total volume of soil extracted for use was 0.2 m3. Once the filled apparatus was returned to the lab, the soil was carefully levelled and any gaps filled in. 15 Carex flagellifera, of standard foliage height 120 mm, were planted into the middle plastic tub section, at approximate square spacing of 250 mm, with 50 mm from the plot borders. 20 mm of compost was then spread to cover the bare soil surface. 5.25 L of PAM (200 mg/L concentration) was applied to the left plot, and this was allowed to dry for at least one day. The grassed control plot was allowed to grow. Six slots were cut into the tub at the plot soil surface and six plastic funnels inserted into the slots to collect the runoff. A thin plastic or metal strip was left in the middle of each plot runoff slot (hence six slots) to prevent bowing of the plastic sidewall. A 12 mm diameter hole was also drilled into each section at the collection end to allow some drainage of the soil water, however during the course of simulations it proved more effective to block these holes with rubber bungs to allow saturation to occur quicker. The runoff was collected via the plastic funnels in three 40 L plastic tubs. In order to prevent rainfall from the rainfall simulator directly entering the buckets, a black polyethylene sheet was attached to a Dexion steel and timber frame, which was bolted to the sides of the plastic tub and protruded 400 mm over the hinge end of the apparatus. Figure 7.2.3 shows the final arrangement as described.

Dominic Fletcher

34909188

Final Year Project

11/01/2007

24

Figure 7.2.3: Photo showing the runoff collection tubs and the method for covering them. 7.3

Rainfall simulator calibration

The rainfall simulator used was the new University of Canterbury constructed simulator. As such it was untested and required an optimal assembly method to be determined. The simulator is made from a modified aluminium ladder, and operates by a stepper motor turning a PVC pipe with two nozzles inserted in it. The rotation arc and speed of the PVC pipe is controlled by the stepper motor, as operated from an electronic controller (Figure 7.3.2) which sets the sweep delay time, or the time waited between each turn of the PVC pipe. Directly below the PVC pipe is an aperture that allows water to fall beneath the simulator onto the test plots. On either side of the aperture, water collection spouting routes the water accumulated during the sweep waiting periods back into the 1000 L plastic tub that the water is pumped from. The photo of the simulator assembled and set up to operate can be seen as Figure 7.3.1

Figure 7.3.1: Rainfall simulator as set up over lab apparatus for calibration and testing Dominic Fletcher

34909188

Final Year Project

11/01/2007

25

As the water is pumped from 1000 L tub, via a pressure hose, through the PVC pipe at a constant rate, it is only the sweep delay that influences the amount of simulated rainfall that leaves the aperture. However the rainfall volume that reaches the test plot depends not only on the sweep delay, but also the height of the plot from the aperture and to some extent the wind intensity and direction. Because of these influences it is necessary to calibrate the rainfall simulator for a given plot height and sweep delay to determine the rainfall volume and distribution pattern. The rainfall simulator is designed to be operated at a pressure of 6 PSI and an approximate height of 2.5 m, so that the simulated rainfall has similar properties to natural rainfall. This pressure ensures a droplet size distribution similar to that of natural rainfall. The height is designed to allow the rain drops to reach terminal velocity before they hit the soil surface, as occurs with natural rainfall.

Figure 7.3.2: The control box for the rainfall simulator The rainfall simulator was calibrated twice for this series of experiments, once at the beginning of testing and once after testing. This was necessary because the first calibration was done inside on a horizontal collection plot, whereas the experiments were conducted outside on varying slopes. The final calibration was done outside, on the apparatus at the maximum slope of 290, in order to compare the greatest rainfall difference in the experiments and to determine the accuracy of the assumed rainfall volume. Both calibrations were performed by covering the plot area of interest with catch cans (square plastic buckets used to collect rainfall), setting the sweep delay on the control box (Figure 7.3.2), and running the simulator at a specific pressure (6 PSI for all experiments) for 10 mins. Each catch can was weighed to determine the mass of water, and hence the volume, and this was divided by the can mouth area to get a depth of rainfall. Figure 7.3.3 and Figure 7.3.4 show the rainfall intensity distribution patterns for the horizontal and 290 slopes respectively. Appendix E contains all the calibration data, results and graphs showing the distribution pattern. The rainfall simulator appears to produce a peak in volume around the two nozzles and lower volumes elsewhere. However, the variation in volume is slight (Figure 7.3.4) and hence the average rainfall depth was assumed to be accurate. Dominic Fletcher

34909188

Final Year Project

11/01/2007

26

Catchcan distribution 70.0

Rainfall intensity (mm/hr)

60.0

50.0

40.0

30.0

20.0

10.0

0.0

1 sec W idthwise

1 sec Lengthwise

2 sec W idthwise

2 sec Lengthwise

Standard deviation

13.5

7.1

7.1

7.2

Average intensity

44.6

55.9

26.3

32.2

Figure 7.3.3: Simulated rainfall variation on horizontal surface at 0.67 m. See Appendix E for more details.

Variation in Rainfall for Notch 3 30.0

Average intensity mm/hr

25.0

20.0

Standard deviation (mm/hr)

15.0

Average intensity (mm/hr)

10.0

5.0

0.0

PAM

Compost

Standard deviation (mm/hr)

4.0

3.1

Control 4.7

Average intensity (mm/hr)

23.4

24.2

22.6

Plot

Figure 7.3.4: Variation in simulated rainfall for 290 slope over the three test plots.

Dominic Fletcher

34909188

Final Year Project

11/01/2007

27

7.4

Lab procedure

As there was not enough available space in the Civil Fluids Laboratory, the testing was done outside the Concrete Laboratory, and because of the worth of some of the equipment, especially the pump and rainfall simulator, it was deemed necessary to provide some form of security measure. As such a security fence was wired around the assembled apparatus and a padlock placed on the gate (Figures 7.2.1 and 7.3.1). This was still not considered secure enough to leave the most expensive equipment outside, and hence the pump was stored in the Fluids Lab when the simulator was not in operation. Another issue associated with the outdoor facility was that of weather damage to the electronics. This meant that all portable electronics were stored in the Fluids Lab when not in use, and all not portable electronics covered with black polythene sheet and sealed tight with tape. For these two reasons, the first step in the testing procedure was to take the pump outside, place it in the 1000 L tub and connect the pressure hoses to the rainfall simulator. Then the control box and associated cables were brought out and connected to simulator also. Special care was taken to ensure that all power cords were connected to a RCD supply, to prevent short-circuiting due to water. When the equipment was ready, the runoff funnels and collection tubs were cleaned and covered with the black polyethylene sheet. The pump was switched on and the adjusted for constant pressure at 6 PSI. The sweep delay dial was set to 2, which for test conditions is equal to 26 mm/hr, or approximately the 50 year Annual Return Interval storm for the field site (obtained from NIWA’s HIRDS software program). The simulator was switched on (on switch, Figure 7.2.2), and the control type switch was flipped to manual, starting the rainfall simulation. A stopwatch was used to time each 1 hr simulation. Once a simulation was running, the funnels were checked regularly to determine when runoff began for each plot and this was recorded. When the simulation was complete, the control type switch was turned to comms control, stopping the rainfall. Then all electronic equipment was turned off at the power source. The volume of runoff was determined for each plot, by measuring the height of runoff in the collection tubs, and multiplying this by the horizontal area. Approximately 100 mL of water was used to wash sediment from the funnels into the tubs. After this, the runoff was stirred rigorously to ensure a consistent sediment concentration, and a 120 mL runoff sample taken in a plastic water sample jar. The suspended sediment concentration was then determined as per the method used for the field samples (Section 6.3). The runoff tubs were then emptied and cleaned, ready for the next simulation. Normally three simulations were completed in a day.

Dominic Fletcher

34909188

Final Year Project

11/01/2007

28

8 8.1

Results Field Results

Due to time constraints and scheduling issues, data was only collected for two storms on the Bowenvale field site. It was difficult to schedule time to collect data immediately after a storm event, due to the irregular nature of storms. If a dry period occurs after a storm and samples are not collected in reasonable time, the rainfall in the rainfall gauges can evaporate rendering any runoff results inaccurate. However, the above aside, two large storms (112 mm and 42 mm respectively) were observed and their effects on the erosion controls measured. Table 8.1.1 summarises the storm characteristics and the results obtained from the four field plots. It should be noted that the runoff from Storm 2 exceeded the bucket capacity for all plots but the compost, and hence the runoff volume is not known. Table 8.1.1: Summary of the two storms and their impacts on the field plots Ev e n t # 1 A ve ra g e ra in fa ll

C o n t ro l 42

B a re s o il 42

PAM 42

Com pos t 42

V o lu m e o f ra in fa ll R u n o ff vo lu m e % R u n o ff TS S Ev e n t # 2 A ve ra g e ra in fa ll

27.7 7.3 26.2 43 C o n t ro l 112

27.7 5.9 21.4 391 B a re s o il 112

27.7 7.9 28.6 481 PAM 112

27.7 5.2 18.8 85 Com pos t 112

L L % m g/L

V o lu m e o f ra in fa ll R u n o ff vo lu m e % R u n o ff TS S

74 > 25.8 > 34.8 207

74 > 25.8 > 34.8 933

74 > 25.8 > 34.8 915

74 2.8 3.8 56

L L % m g/L

mm

mm

This data appears to suggest that of the two erosion control methods (PAM, Natives and Compost), the native vegetation and compost method is much more effective. When compared with the bare soil, the percentage reductions in suspended sediment from the native/compost plot, for Storm 1 and Storm 2 respectively, are 78.3% and 94%. When compared with the grassed control (current natural conditions) the results are less conclusive, with a 97.7% increase (Storm 1) and a 72.9% reduction (Storm 2) in suspended sediment. Compost also appears to reduce the volume of runoff significantly as expected. It would appear that polyacrylamide has no or even a negative effect on surface soil erosion, with percentage increases in suspended sediment of 1019% (Storm 1), 342% (Storm 2) when compared with the control, and 23% (Storm 1), -1.9% (slight reduction for Storm 2) when compared with the bare soil. The increase in sediment may be in part due to the wash off of PAM. The volume of runoff also appears to increase if PAM is used, with increases of 9.1% compared with the control and 33% compared with bare soil observed. There is however great variability in these results and limited conclusions can be drawn from the data set. The raw data obtained from the field site and the percentage reduction calculations can be seen in Appendix F. Dominic Fletcher

34909188

Final Year Project

11/01/2007

29

8.2

Lab Results

A large good quality data set was obtained from the lab experiments due to the controlled nature of the testing. Three simulated storms of 26 mm/hr intensity (2 year ARI) were performed for each of the slopes: 8.4% (4.80, Notch 1), 26.5% (15.40, Notch 2), and 47.9% (28.60, Notch 3). Each storm was run for 1 hr at steady state runoff. The results obtained can be seen as Table 8.2.1 below. The raw data obtained from these test experiments can be seen in Appendix G. Table 8.2.1: Averaged results from the three lab experiments for each of the three slopes (4.80 (Notch 1), 15.40 (Notch 2), and 28.60 (Notch 3)) after 26 mm of simulated rainfall. Slope Runoff

Notch 1 (8.4%) Notch 2 (26.5%) Notch 3 (47.9%)

PAM Average 19.3 16.7 17.5

Compost Control Std dev Average Std dev Average 0.4 1.5 0.0 12.5 0.1 3.7 0.4 9.2 0.7 4.0 0.3 10.0

Bare soil Std dev Average 0.1 0.8 0.6 -

% Runoff

Notch 1 (8.4%) Notch 2 (26.5%) Notch 3 (47.9%)

92.9 83.5 97.2

3.0 0.7 3.9

8.2 21.3 25.6

0.4 2.4 1.7

60.2 46.1 55.6

3.9 4.2 3.1

-

% % %

TSS

Notch 1 (8.4%) Notch 2 (26.5%) Notch 3 (47.9%)

488 1362 1403

3.5 31.8 536.0

353 203 277

30.4 29.0 60.8

636 394 395

123.7 3.5 40.3

-

mg/L mg/L mg/L

L L L

The results in Table 8.2.1 suggest that compost and native vegetation is more effective at retaining sediment than the grassed control and the polyacrylamide, for all of the slopes tested. The data also indicates that as the slope increases the volume of runoff from the compost and native vegetation plot increases, but the total suspended sediment (TSS) concentration decreases. The results for Notch 1 show the greatest total suspended sediment concentration for the compost and control plots. This is likely to be due to the soil and compost being disturbed when they where placed into the tub. Table 8.2.2 shows the schedule of testing including some observations. The suspended sediment concentration from the control plot (grassed natural field conditions) is less for the two greatest slopes. This may be the result of grass re-growth that occurred after the Notch 1 experiments (Table 8.2.2). However, the TSS concentration did stabilise at ~400 mgL-1 for the other two slopes, suggesting that a good pastoral grass cover on Port Hill’s Loess will produce a constant sediment loss over the tested slope range. The volume of runoff also decreased with increasing slope, possibly due to grass re-growth also. While the volume of runoff from the PAM plot decreased with increasing slope, the TSS concentration increased significantly for Notch 2 and 3. For the Notch 2 results this may be due to the PAM being washed away during Notch 1 simulations. However, as PAM was reapplied (1.3 L of 200 mgL-1) and allowed to dry after Notch 2 and before Notch 3 simulations, this does not explain the increased TSS from Notch 3. It is possible that the increase in TSS was due to the increase in slope. Dominic Fletcher

34909188

Final Year Project

11/01/2007

30

Table 8.2.2: Schedule of lab testing Ex p e r i m e n t n u m b e r

D a te

S l o p e (D e g )

Tria l

28/07/06

4.8

1 2

01/08/06 01/08/06

4.8 4.8

3

01/08/06

4.8

4 5 6

28/08/06 28/08/06 28/08/06

15.4 15.4 15.4

7 8 9

29/08/06 29/08/06 29/08/06

28.6 28.6 28.6

8.3

R e m a rks P A M ap p lie d b efo re , n o t re po rt e d in re s ult s D is t u rb e d s o ils a n d c om pos t R ain fall s im u lat o r b re a k d o w n S ig n ific a n t g ra s s g ro w t h o n c on t ro l plo t P A M rea p p lie d be fo re e x p e rim e n t -

Comparison of results

Both the field and lab results indicate that polyacrylamide is ineffective at controlling surface water soil erosion. Table 8.3.1 below shows that when compared with the grassed control plot, PAM applied to bare soil results in increased soil erosion, similar to that observed on the bare soil field plot. This result is similar to that found by Ajwa and Trout (2001), who determined that polyacrylamide is ineffective at controlling soil erosion on sandy loam soils, as observed here on Port Hill’s Loess (a sandy loam). Compost and native vegetation however, appears to be a highly effective method of controlling water surface soil erosion, with reductions in sediment concentration observed for all experiments, except Storm 1 on the field site. The Storm 1 result may be due to other weather conditions or simply the freshness of the compost. The freshness of the compost is considered to affect the sediment concentration, because the smaller compost particles on the surface have not been eroded yet. The “settling in period”, or the time it takes for these smaller particles to initially erode is hence important for the suspended sediment concentration in the runoff. Effective comparisons between Notch 2 (15.40) and the field site (12.70) were not possible due to the differences in rainfall intensity, which affects the amount of erosion that occurs. Table 8.3.1: The effectiveness of the two erosion control methods when compared with current soil coverage conditions. % D iffe re n ce w h e n co m p a re d to co n tro l R u n o ff

TS S

Dominic Fletcher

S lope N ot c h 1 (8.4% ) N otc h 2 (26. 5% ) N otc h 3 (47. 9% ) F ield 1 (22.5% ) F ield 2 (22.5% )

P AM 54 81 43 9.1 -

C o m p o st -88 -60 -61 -28.2 -

B a re so il -18.2 -

% % % % %

N ot c h 1 (8.4% ) N otc h 2 (26. 5% ) N otc h 3 (47. 9% ) F ield 1 (22.5% ) F ield 2 (22.5% )

-23 246 72 1018.6 342.0

-45 -49 -30 97.7 -72.9

809.3 350.7

% % % % %

34909188

Final Year Project

11/01/2007

31

9 9.1

Discussion Effectiveness and applicability of each erosion control measure

The results of these experiments show that under the test conditions, compost and native vegetation is an effective topsoil erosion control measure. However, it is not known what effect the native vegetation had on the results. It is likely that the erosion control was provided solely from the compost blanket, with the native vegetation providing very minor shielding from splash erosion, and intercepting a negligible proportion of the sheet wash flow. However, the purpose of the vegetation is to provide future protection, as it will grow large enough to eventually adequately shield the soil. The results of this experiment show that juvenile Carex flagellifera, planted through a compost blanket, do not have a negative impact on the soil protection properties of the compost blanket. This means that erosion protection of Port Hill’s Loess by a compost blanket is not significantly impacted by planting juvenile Carex flagellifera, enabling planting for future protection as the compost degrades. More studies are required however to determine how the performance of the compost and native vegetation system changes over several years. The influence this method has on encouraging natural succession also needs to be determined. The results also showed that the compost requires time to settle and consolidate after it is applied. The finer particles are expected to erode initially, after which the compost performance at reducing suspended sediment improves. This was observed in these experiments. All of the observed erosion is considered to have come from the compost, as the soil surface is covered and protected. Although compost was successful for these given test conditions, it is known that compost blankets do not control erosion caused by a concentrated flow (Coulson, 2005). Coulson (2005) suggests that it is possible to inject polyacrylamide into the compost blanket to add extra strength and flocculate the soil, hence increasing the effectiveness against concentrated flow. The degradation rate of PAM and its applicability for specific composts need to be considered if this approach is taken. The Erosion and sediment control guidelines (2006) produced by the local regional council, Environment Canterbury, outline best management practices for using compost blankets to control soil erosion. To overcome the problems caused by concentrated flow, the guidelines recommend that all flow up-slope from the compost blanket be routed away by interception trenches. Another option is to use compost socks at the top of compost blanket to absorb run-on flow. Polyacrylamide proved to be ineffective at influencing topsoil erosion for these experiments. It is likely that PAM will not work on Port Hill’s Loess due to its ionic properties, and hence PAM is not recommended as a soil erosion control method for the Port Hills. However, these experiments compare PAM on bare soil against a grassed control. In hindsight, a better way to compare these two situations would be to apply PAM to a grassed plot. The comparison of PAM on bare soil to the bare soil plot on the field site still however proved that PAM is ineffective on Port Hill’s Loess.

Dominic Fletcher

34909188

Final Year Project

11/01/2007

32

9.2

Cost estimate of compost and native vegetation method

The economic cost of applying compost and native vegetation as a soil erosion control needs to be investigated further; however some estimates of cost for the Christchurch area are discussed below. The wholesale price of compost in Christchurch is approximately $16/tonne, as compared to $40/tonne for Tauranga, Napier, and Auckland (Clough, 1999). The price of compost in Christchurch is cheaper due to a competitive compost producing market. Using the 20 mm depth of compost application used in these experiments, and assuming a compost density of 590 kg/m3 (MfE, 2005), the cost of applying a compost blanket in Christchurch is $1,888 per hectare (Table 9.2.1). The maintenance cost for this blanket was approximated to be $14.54 per hectare, assuming the compost loss is equal to 10x the TSS concentration obtained from Notch 1 experiments. This compost loss accounts for biodegradation and erosion. The total cost for a 1 year design life compost blanket was estimated to be $1,900 per hectare. Table 9.2.1: Cost of a 20 mm compost blanket in Christchurch. Price incl. GST $/tonne $ 16.00 Life span years 1

Density tonnes/m3

Depth of application mm

Volume of compost Cost m3/ha $/ha 0.59 20 200 $ 1,888.00 Erosion rate Maintenance requirements Cost Total cost kg/ha.year m3/ha.year $/ha $/ha 909 1.54 $ 14.54 $ 1,902.54

As acquiring and applying compost is becoming increasingly cheaper, due to new technologies and greater composting resource availability (As discussed in Section 5.2), the cost of purchasing and planting the native plants becomes the limiting factor. The wholesale price of Carex flagellifera for the PB5 size used in these experiments is ~$2.50 per plant. Using a plant spacing of 300 mm, the cost of purchasing Carex flagellifera was calculated to be $279,000 per hectare planted (see Table 9.2.2). The seasonal nature of vegetation availability constrains when this technique can be applied. Using native seeds instead of seedlings could reduce this cost, but it is not recommended due to the specific germination requirements, insect predation and high mortality rate of native seeds. Non-native species can also be used with this method depending on their suitability to a site. Table 9.2.2: Cost of purchasing PB5 size Carex flagellifera. Plant price $/plant (PB5 size) $2.50

Recommended spacing mm/apart 300

Number of plants number/ha 111556

Cost of vegetation $/ha $278,890.00

The total raw material cost (purchasing vegetation and compost) is estimated to be $282,700 per hectare. Of this approximately 99% of the cost is in the vegetation. The actual cost of applying this technique will depend upon the method of compost application (i.e. by hand or blower truck), and the planting labour required.

Dominic Fletcher

34909188

Final Year Project

11/01/2007

33

10 Difficulties encountered This project encountered many set backs and many technical issues were overcome during the course of experiments. The first major difficulty was constructing the lab apparatus. The apparatus design (reproduced in Appendix C) looked feasible on paper, but during construction the design had to be further defined and altered, to best use the available materials and for the ease of construction and operation. The technical assistance provided by Kevin Wines and Ian Sheppard proved invaluable during this process, as often a theoretically sound idea was difficult or less efficient to construct and use. Discussing and explaining the ideas and requirements clearly always resulted in the best option being determined, saving time, money, and labour. The biggest issue associated with the use of technical assistance was the need to schedule personnel in advance. Due to the other work commitments of the technical staff, tasks often took considerably longer than anticipated. Starting tasks as early as possible and allowing excess time helps to prevent delays in the project schedule. A problem associated with the field site was the lifting of the collection buckets out of their holes by groundwater. This problem resulted in inaccurate runoff collection for Storm 1 and as such the results may not be entirely representative of the actual situation. This situation was remedied by digging a relief trench for each bucket hole to allow water to drain down slope. Another problem encountered with the buckets was the fact that they were too small to collect the runoff from Storm 2, resulting in an unknown volume of runoff for that event and possible affecting the TSS results. This problem would best be resolved by checking the buckets after 24 hours of rainfall, and not waiting for the end of a storm, as this may be several days over which a large volume of runoff can be produced. An unavoidable set back occurred during lab rainfall simulations for Notch 2. The step magnets attached to the drive shaft in the stepper motor fell off, resulting in the simulator not working. This meant that no simulations could be run for the two weeks it took to fix the motor. Whilst this situation was unavoidable, it did not result in significant time lost, as during this period other project related tasks were done, ensuring some progress was still made. That fact that extra time had been allowed for such a situation also meant that the stoppage had minimum impact. Significant difficulty was encountered when trying to increase the slope of the apparatus lid. This was because the hydraulic jack used to lift the lid did not have enough travel for the entire distance, and the wheels attached to the frame’s base enabled the apparatus to move during lifting, causing a potential safety hazard. It required two people to change the lid slope, and even then with some difficulty. This issue was not resolved. Another issue associated with the lab apparatus was that of leaking around the runoff collection funnels. Whilst this was kept to a minimum, it still induced some inaccuracy into the results. This problem was temporarily resolved by taping the edges of the funnels with black duct tape, preventing water movement under the funnel lips. However, this was a short-term solution and hence a more durable sealing method needs to be applied to prevent leakage during future apparatus use. Dominic Fletcher

34909188

Final Year Project

11/01/2007

34

11 Conclusions The purpose of this project was to determine the effectiveness of two methods at controlling water induced soil surface erosion on the Port Hills in Canterbury. The two methods tested were polyacrylamide, a soil flocculent, and native sedge species (Carex flagellifera) planted with compost blankets. These two methods were compared with a current site vegetation cover (pastoral grass) control plot, for a variety of slopes, in both the lab and at a field site in Bowenvale Reserve in the Port Hills. In addition to this, the two methods were also compared against a bare soil control at the field site. The comparison was made by collecting the runoff from each plot and comparing the volume and the suspended sediment concentration. The results of both the lab and field experiments show that polyacrylamide is ineffective at controlling water induced soil surface erosion in Port Hill’s loess. This was determined from comparisons to current Bowenvale Reserve site conditions and bare Port Hill’s loess soil. It was hence recommended that polyacrylamide not be used as an erosion control method in the Port Hills. The results also showed that the native vegetation with compost was effective at controlling soil erosion, for slopes ranging from 50 – 300 and for the Port Hills’ 50 year annual return interval storm of 26 mm/hr. The vegetation and compost method significantly reduced the volume of runoff (minimum of 28% reduction observed) and the concentration of suspended sediment (minimum of 30% reduction observed).

Dominic Fletcher

34909188

Final Year Project

11/01/2007

35

12 Recommendations It is recommended that these experiments be followed up with a series of more rigorous field experiments, for a variety of locations on the Port Hills. The effectiveness of the compost and native vegetation method needs to be determined over a time scale of 2-3 years, and the extent to which that method encourages natural succession studied. The cost of the native vegetation and compost system needs to be accurately determined and compared in terms of cost benefit ratios with other soil erosion control methods. This should further define the applicability and feasibility of this system. The lab apparatus used for these experiments can be greatly improved, especially to facilitate changing the slope. This apparatus was constructed out of durable materials with the intention that it could be reused for future experiments in this field; however the difficulties with increasing the slope need to be remedied before it can be safely operated. It is recommended that future users of the apparatus attach a hydraulic lifting arm between the base strut and the lid, enabling safe and easy slope changes. It is also recommended that the wheels be chocked to prevent movement during testing and lifting. Before new soil is placed in the plastic tub, it is recommended that the plastic funnels be bolted to the tub and a silicon sealant applied around them to prevent runoff leakage during collection. It is also recommended that all design calculations be doublechecked by another person, to ensure that minor mathematical errors do not result in large construction problems and set backs. There is great potential for future work in this field, especially incorporating the use of Geographic Information Systems (GIS) and the Water Erosion Prediction Project software model (WEPP). GIS can be used to identify potential erosion sites were the native vegetation and compost method may be suitable, what type of native vegetation is best suited to the site conditions, and the predicted sediment erosion of an area if that method is applied. WEPP can similarly be used (alone or in conjunction with GIS) to predict the erosion from a slope after this control method has been applied.

Dominic Fletcher

34909188

Final Year Project

11/01/2007

36

13 Acknowledgements I would like to acknowledge the follow persons who have been invaluable in assisting with this project. ¾ Kevin Wines ¾ Ian Sheppard ¾ Peter McGuigan ¾ Tom Cochrane ¾ Mike Weavers ¾ Stuart Toase ¾ Christchurch City Council

Dominic Fletcher

34909188

Final Year Project

11/01/2007

37

14 References Aase, J. K. and Bjorneberg, D. L. (2000) Multiple Polyacrylamide Applications for Controlling Sprinkler Irrigation Runoff and Erosion. Applied Engineering in Agriculture, 16(5), page 501. Alberts, E. E., Ghidey, F. and Thompson, A.L. (2001) Effect of Polyacrylamide on Erosion for Various Rainfall Kinetic Energies. Proc. ASAE Annual International Meeting (Sacramento, CA), July 30 - Aug 1, pages 3-4. Ajwa, H. and Trout, T. J. (2001) Polyacrylamide Effects on Infiltration in San Joaquin Valley Sandy Loam Soils. Proc. ASAE Annual International Meeting (Sacramento, CA), July 30-Aug 1, pages 1-11. Auckland Regional Council (ARC) (1999) Erosion & sediment control guidelines for land disturbing activities. Technical publication No. 90, Auckland Regional Council, Auckland. Bechmann, M. and Stålnacke, P. (2005) Effect of policy-induced measures on suspended sediments and total phosphorus concentrations from three Norwegian agricultural catchments. Science of the Total Environment, 344(1-3), pages 129-142. Canterbury Regional Council (CRC) (undated) Establishing Shelter for Soil Conservation in Canterbury. Canterbury Regional Council, Christchurch. Choi J. D., Kim, K. S. and Kweon, K. S. (2001) Effect of PAM Application on Soil Erosion of a Sloping Field with a Chinese Cabbage Crop. Proc. ASAE Annual International Meeting (Sacramento, CA), July 30-Aug 1, page 1-8. Christchurch City Council (CCC) (2006) Parks: Natural areas. [Online] Available: http://www.ccc.govt.nz/Parks/NaturalAreas/ [Accessed: 2006, Sept 19th] Clough, P. and Dolan, L. (1999) The economics of waste management and recycling in New Zealand. Proc. WasteMINZ 11th Annual Conference (Queenstown), November. Page 10. Coulson, R. (2005) The Use Of Composts In Erosion Control. Erosion control Seminar, 11th –13th September. Rural Supply Technologies Ltd. Coulson, R. (2005b) Establishing Indigenous Species with Hydroseeding. Erosion control Seminar, 11th –13th September. Rural Supply Technologies Ltd. Department of Conservation (DoC) (2006) Native plants natural to Banks Peninsula [Online] Available: http://www.doc.govt.nz/RegionalInfo/010~Canterbury/004~Conservation/Motukarara-Nursery/102~Native-Plants-BanksPeninsula.asp [Accessed: 2006, June 6th] Dixon, P. M., Glanville, T. D., Laflen J. M., Persyn, R. A. and Richard, T. L. (2004) Environmental Effects Of Applying Composted Organics To New Highway Dominic Fletcher

34909188

Final Year Project

11/01/2007

38

Embankments: Part 1. Interrill Runoff And Erosion. American Society of Agricultural Engineers, 47(2): 463−469. Dodd, M., Douglas, G., Fung, L., et. al. (undated) Designing sustainable farms - plant solutions for eroding hill country. Landcare Research, HortResearch and AgResearch [Online] Available: http://www.landcareresearch.co.nz/research/biodiversity/greentoolbox/#Articles [Accessed: 2006, April 28th ] ECAN (2006) Erosion and sediment control guidelines. Environment Canterbury Report no: R06/23, Environment Canterbury, Christchurch. Flanagan, D.C., Peterson, J.R., and Tishmack J.K. (2001) Effects of PAM Application Method and Electrolyte Source on Runoff and Erosion. Proc. International Symposium of Soil Erosion Research for the 21st Century (Honolulu), Jan 3-5, pages 179-182. Goldstein, N. (2006) It’s Official! Compost Tools on EPA List as Storm Water BMPS. Biocycle, February 2006, page 21. Gray, D.H., and Sotir, R.B. (1996) Biotechnical and Soil Bioengineering. Slope Stabilization. A Practical Guide for Erosion Control. John Wiley & Sons Inc., New York Lal, R. (Ed.) (1994) Soil Erosion Research Methods, 2nd Edition. Soil and Water Conservation Society. St Lucie Press, Florida, USA. (Pages 2-4) Legacy, D. and McCoy, S. (2003) A vegetation success: Contractor counts on compost for slope stabilization. Biocycle, July 2003, pages 29-30. Massachusetts Nonpoint Source Pollution Management (MNSPM) (undated) Brush mattresses [Online] Available: http://projects.geosyntec.com/NPSManual/Fact%20Sheets/Brush%20Mattresses.pdf. [Accessed: 2006, June 8th] Mehuys, G. and Partington, M. (2005) Effectiveness of Polyacrylamide in Reducing Soil Erosion on Steep Slopes. Proc. ASAE Annual International Meeting (Tampa, FL), July 17 – 20, pages 1-8. MfE (2005) Greenwaste and Organic Disposal Options For the West Coast. Ministry for the Environment. Page 12-13. Available: www.wcrc.govt.nz/.../F8BD353D-DEFF-406ABE04F0B67D94BFA4/30725/GreenwasteandOrganicDisposalfinal.pdf [Accessed: 2 Oct, 2006] Painter, D. (2005) Defining Storm Water. Lecture notes for ENNR 305: Ecological Engineering 1, University of Canterbury. Pages 1-6. Phillips, C. (2005) Erosion and Sediment Control Using New Zealand Native Plants What Do We Know? Erosion control Seminar, 11th –13th September. Landcare Research Dominic Fletcher

34909188

Final Year Project

11/01/2007

39

Practice Standard for Anionic Polyacrylamide (PAM) Erosion Control (2001). Conservation Code 450. Natural Resources Conservation Service, Pages 1-3. Rickson, R.J., and Morgan R.P.C. (Eds.) (1995) Slope Stabilization and Erosion Control: A Bioengineering Approach. E&FN Spon, London, UK. (Pages 7-35) Schiechtl, H.M., and Stern, R. (1996) Ground Bioengineering Techniques for Slope Protection and Erosion Control. Barker, D. H. (Ed.) Blackwell Science LTD Sherman, R. (2003) America’s Largest Compost Market: Texas Transportation Department Accelerates Highway Use of Compost. Biocycle, July 2003, pages 24-28 Trangmar, B. (2003) Soil Conservation Guidelines for the Port Hills. Landcare Research Contract Report: LC0203/111, Landcare Research. UC Composting. [Online]. Available: http://www.civil.canterbury.ac.nz/Compost/ (Accessed: 2006, September 16)

Dominic Fletcher

34909188

Final Year Project

11/01/2007

40

Appendix A: Elevation view of Bowenvale Reserve and the field site.

Field site

Bowenvale Reserve

Figure A 1: Altered aerial photo of Bowenvale Reserve on the Port Hills showing field site. (Reproduced and altered from http://www.ccc.govt.nz/ Parks/Natural Areas/ port_hills_recreation.asp, (2006)) Table A 1: Soil properties obtained from the NZ fundamental soils layer for the Bowenvale Reserve site. This data proves that the soil tested is Port Hill’s loess. Bowenvale Reserve Soil type Steepland soil

77a

Dominic Fletcher

NZSC BMA-BFA

34909188

pH Max

Min 6.4

Salinity Max Min 5.8 0.04

Final Year Project

CEC Max 0

Min 24.9

11/01/2007

Carbon Max Min 12 3.9

41

2

Appendix B: Hill Laboratories Soil Analysis Form

Dominic Fletcher

34909188

Final Year Project

11/01/2007

42

Dominic Fletcher

34909188

Final Year Project

11/01/2007

43

Appendix C: Schematic of field site runoff diversion equipment

Funnel dimensions 600 mm

50 mm

50 mm Fold lines

300 mm

100 mm

Sheet metal depth 120 mm

1100 mm or 2400 mm

Dominic Fletcher

34909188

Final Year Project

11/01/2007

44

Appendix D: Lab apparatus schematic

Steel Base

Angle iron

900 mm

5 mm steel plate for wheels to attach to

2000 mm

Front

150 mm

Bolt holes

Back

700 mm

Dominic Fletcher

34909188

Final Year Project

11/01/2007

45

Side view of base

1100 mm

0

29

150 mm

700 mm

900 mm

Li d

2000 mm

550 mm

1100 mm

670 mm

Dominic Fletcher

34909188

Final Year Project

11/01/2007

46

Plastic tub from above

610 mm

605 mm

610 mm 1230 mm

1095 mm

1990 mm

100 mm

Soil surface Plastic Tub Angle iron Fiberboard base

1100 mm

Dominic Fletcher

34909188

Final Year Project

11/01/2007

47

Appendix E: Rainfall simulator calibration data and rainfall distribution graphs The following graphs show the variation in simulated rainfall at a vertical distance of 1530 mm from the nozzles, pressure 6 PSI, and sweep delays of 1 and 2 seconds. Table E 1 below shows the catch cans distribution and numbers. The widthwise distribution consists of four rows of 6 cans, spaced 300 mm apart, and Figures E 1 and 2 show the variation in rainfall depth for each row. The lengthwise distribution shows two rows of 12 catch cans positioned directly beneath the PVC rotating pipe. The cans were placed snugly next to each other. Table E 1: Distribution and numbers of catch cans for the widthwise and lengthwise variation in simulated rainfall Can number and distribution 19 20 21 18 17 16 7 8 9 6 5 4 13 1

14 2

15 3

22 15 10 3 16 4

23 14 11 2 17 5

18 6

24 13 12 1 19 7

20 8

21 9

22 10

23 11

24 12

Widthwise 1 sec 12.0

10.0

Rainfall in mm

8.0 Front row Second row

6.0

Third row Back row

4.0

2.0

0.0 1

2

3

4

5

6

Catch can

Figure E 1: Simulated rainfall variation under 1 sec sweep delay for each of the 4 rows of catch cans.

Dominic Fletcher

34909188

Final Year Project

11/01/2007

48

Rainfall variation for 2 sec delay 7.0

6.0

Rainfall in mm

5.0

Front row

4.0

Second front Third row 3.0

Back row

2.0

1.0

0.0 1

2

3

4

5

6

Catch can

Figure E 2: Simulated rainfall under 2 secs of sweep delay

Lengthwise Rainfall variation for 1 sec delay 12.0

10.0

Rainfall in mm

8.0

First row

6.0

Second row

4.0

2.0

0.0 1

2

3

4

5

6

7

8

9

10

11

12

Catch can

Figure E 3: Variation in simulated rainfall over the length of the simulator for 1 sec of sweep delay. Dominic Fletcher

34909188

Final Year Project

11/01/2007

49

Rainfall variation lengthwise for 2 sec delay 12.0

10.0

Rainfall in mm

8.0

Front row

6.0

Second row

4.0

2.0

0.0 1

2

3

4

5

6

7

8

9

10

11

12

Catch can

Figure E 4: Simulated rainfall variation over simulator length for 2 sec sweep delay. It can be noted from all of these graphs that the greatest simulated rainfall volume occurs directly under the constant pressure nozzles, in both the length and width directions. However, the rainfall volume does not differ significantly over the 2.15 m by 1.23 m test area for these two sweep delay times. This data was obtained after 10 mins of simulated rainfall, and for a sweep delay of 2 sec, the average intensity is 26 mm/hr. This value is approximately equal to the 2 year ARI (Annual Return Interval) storm for Bowenvale Reserve in the Port Hills. Table E 2 over, shows the data obtained from the final simulator calibration, performed on the lab apparatus at a slope of 280, for a sweep delay of 2 sec. It was found that the average rainfall intensity over the plastic tub area was 23.4 mm/hour, which is not too different from the 26 mm/hr assumed and had little effect on the results for the 1 hour simulation period.

Dominic Fletcher

34909188

Final Year Project

11/01/2007

50

Table E 2: Rainfall depth calibration data for Notch 3. Total average rainfall intensity Sweep delay Time of operation Dry bucket weight Dimensions

23.4 mm 2 sec 10 min 163.6 g 180 mm

PAM Bucket weight Water weight Water volume g g m3 332.5 168.9 0.000169 296.5 132.9 0.000133 296.7 133.1 0.000133 270.1 106.5 0.000107 245 81.4 0.000081 274.7 111.1 0.000111 325.2 161.6 0.000162 299.6 136 0.000136 286.5 122.9 0.000123 308.1 144.5 0.000145 296.9 133.3 0.000133 292.1 128.5 0.000129 302.8 139.2 0.000139 284 120.4 0.000121 294 130.4 0.000131 261.9 98.3 0.000098 272.9 109.3 0.000109 276.9 113.3 0.000113 Average Average Intensity mm/hr

mm 5.2 4.1 4.1 3.3 2.5 3.4 5.0 4.2 3.8 4.5 4.1 4.0 4.3 3.7 4.0 3.0 3.4 3.5 3.9 23.4

Compost Bucket weight Water weight Water volume g g m3 297.8 134.2 0.000134334 264.8 101.2 0.000101301 317.1 153.5 0.000153654 293.2 129.6 0.00012973 312.6 149 0.000149149 308.5 144.9 0.000145045 301.1 137.5 0.000137638 300.7 137.1 0.000137237 294.1 130.5 0.000130631 297.3 133.7 0.000133834 269.1 105.5 0.000105606 276 112.4 0.000112513 Average Average Intensity

mm 4.1 3.1 4.7 4.0 4.6 4.5 4.2 4.2 4.0 4.1 3.3 3.5 4.0 24.2

Dominic Fletcher

34909188

Control Bucket weight Water weight Water volume g g m3 265.6 102 0.000102102 288 124.4 0.000124525 286.3 122.7 0.000122823 280.3 116.7 0.000116817 321.9 158.3 0.000158458 319.2 155.6 0.000155756 293.8 130.2 0.00013033 314.6 151 0.000151151 313 149.4 0.00014955 284.3 120.7 0.000120821 309.6 146 0.000146146 293.6 130 0.00013013 269 105.4 0.000105506 286.4 122.8 0.000122923 285 121.4 0.000121522 237.1 73.5 7.35736E-05 251.2 87.6 8.76877E-05 240.2 76.6 7.66767E-05 Average Average Intensity mm/hr

mm 3.2 3.8 3.8 3.6 4.9 4.8 4.0 4.7 4.6 3.7 4.5 4.0 3.3 3.8 3.8 2.3 2.7 2.4 3.8 22.6

mm/hr

Final Year Project

11/01/2007

51

Appendix F: Field site experimental data and calculations

Table F 1: Data and calculations from Storm Event 1 Event 1 Rainfall gauge Rainfall depth collected Average rainfall Volume of rainfall per plot Runoff volume Depth in bucket Volume % Runoff

Filter weight

Before After Change per 100 mL sample TSS

1

2

40 40 40 40 + 2 mm evaporation

348.0

438.0

-348.0

Dominic Fletcher

90.0

34909188

41 mm 42 mm

Control Bare soil PAM Compost 110 90 120 79 mm 7.3 5.9 7.9 5.2 L 26.2 21.4 28.6 18.8 %

0.1997 0.204 0.0043

0.2069 0.246 0.0391

0.2062 0.2543 0.0481

43

391

481

0.1986 g 0.2071 g 0.0085 g 85 mg/L

Bare soil PAM Compost -18.2 9.1 -28.2

42.0 % Difference

Compared to bare soil Control PAM Compost 1.3 2.0 -0.7 % Difference Difference

Difference

4

27.72 L

Compared to control Bare soil PAM Compost -1.3 0.7 -2.0 % Difference Difference

Difference

3

809.3

1018.6

97.7

Control PAM Compost 22.2 33.3 -12.2

-306.0 % Difference

Final Year Project

-89.0

23.0

-78.3

11/01/2007

52

Table F 2: Data and calculations from Storm Event 2 Event 2 Rainfall gauge Rainfall depth collected Average rainfall Volume of rainfall per plot Runoff volume Depth in bucket Volume % Runoff

1

2

3

108 112

114

112

4 114 mm 112 mm 73.92 L

Control Bare soil PAM Compost 390 390 390 43 mm 25.8 25.8 25.8 2.8 L 34.8 34.8 34.8 3.8 %

Filter weight

Before After Change per 100 mL sample TSS

0.2015 0.2222 0.0207

0.1986 0.2919 0.0933

0.1986 0.2901 0.0915

207

933

915

0.2 g 0.2056 g 0.0056 g 56 mg/L

Compared to control Runoff

Difference % Difference

TSS

Difference % Difference

Bare soil PAM 726.0 350.7

708.0 342.0

Compost -151.0 -72.9

Compared to bare soil Runoff

Difference % Difference

TSS

Difference % Difference

Dominic Fletcher

Control -726.0 -77.8

34909188

PAM -18.0 -1.9

Compost -877.0 -94.0

Final Year Project

11/01/2007

53

Appendix G: Laboratory experimental data and calculations

Table G 1: Data from the Notch 1 (4.80) experiments. The saturated data values were used for all reported results. Notch 1

Average

Average

Average

16.5 16.9 19.6 19.0 18.0 18.5 19.3

82.4 84.4 90.8 95.0 88.2 90.1 92.9

TSS Concentration 100ml sample mg/L (ppm) 290 460 485 490 431 478 488

Compost Trial 1 Run 1 Run 2 Run 3 With trial Without trial Saturated

0 0.3 1.5 1.5 0.8 1.1 1.5

0 1.7 7.9 8.5 4.6 6.1 8.2

109 549 331 374 341 418 353

Control Trial 1 Run 1 Run 2 Run 3 With trial Without trial Saturated

5.3 6.5 12.4 12.6 9.2 10.5 12.5

26.5 32.5 57.5 63.0 44.9 51.0 60.2

1123 876 723 548 818 716 636

PAM Trial 1 Run 1 Run 2 Run 3 With trial Without trial Saturated

Dominic Fletcher

Runoff Volume L

% Runoff %

34909188

Final Year Project

11/01/2007

54

Table G 2: Data from Notch 2 (15.40) experiments. Saturated values reported.

Average

Average

Average

PAM Run 1 Run 2 Run 3 All Saturated

Runoff TSS Concentration Volume % Runoff 100ml sample L % mg/L (ppm) 16.6 83.0 1486 16.6 83.0 1384 16.8 84.0 1339 16.7 83.3 1403 16.7 83.5 1362

Compost Run 1 Run 2 Run 3 All Saturated

2.4 3.4 3.9 3.2 3.7

14.1 19.6 23.0 18.9 21.3

393 223 182 266 203

Control Run 1 Run 2 Run 3 All Saturated

9.2 8.6 9.8 9.2 9.2

46.1 43.1 49.1 46.1 46.1

450 391 396 412 394

Dominic Fletcher

34909188

Final Year Project

11/01/2007

55

Table G 3: Data from Notch 3 (28.60) experiments. Saturated values reported.

Average

Average

Average

Runoff TSS Concentration Volume % Runoff 100ml sample L % mg/L (ppm) PAM Run 1 15.0 83.3 1724 Run 2 17.0 94.4 1024 Run 3 18.0 100.0 1782 All 16.7 92.6 1510 Saturated 17.5 97.2 1403

Compost Run 1 Run 2 Run 3 All Saturated

3.8 3.8 4.1 3.9 4.0

24.4 24.4 26.8 25.2 25.6

227 234 320 260 277

Control Run 1 Run 2 Run 3 All Saturated

9.2 9.6 10.4 9.8 10.0

51.2 53.4 57.8 54.1 55.6

460 423 366 416 395

Dominic Fletcher

34909188

Final Year Project

11/01/2007

56