Evaluation of Bioengineering Soil Erosion Control Techniques in Standard USLE Plots

M.S.M. Amin, H.L. Yong and M. Rashidi Bakar

Department of Biological and Agricultural Engineering Faculty of Engineering Universiti Putra Malaysia 43400 Serdang, Selangor DE

Abstract An erosion control study on Serdang series soil was conducted in standard USLE plots at DBAE Field Station, UPM. The bioengineering erosion control techniques include vetiver (Vetiveria zizanioides), legume (Arachis pintoi), spot turfing and close turfing with cowgrass (Axonopus compressus), hydroseeding and few combinations of hydroseeding with biomats. A plot was left bare as a control. Close turfing gave better soil protection than the other grass species, reducing soil loss by 99% compared to the bare plot. The addition of "fibromat" to the hydroseeding plot resulted in significantly lower soil loss. All hydroseeding plots overlaid with biomats gave better protection, resulting in C factor lower than 0.004. Close turfing produced C factor of 0.004, compared to 0.017 for spot turfing, 0.021 for hydroseeding only, 0.122 for vetiver and 0.213 for legume. From statistical correlation results, soil loss from the bare plot was better correlated with KE>25 than raindepth, EI30 and AIm. 1.0 Introduction Looking briefly into the history of land use, it seems that human interference by clearing of natural vegetation covers result in serious soil erosion. Excessive runoff generated from logging activities, golf courses and highway constructions usually moves directly from drainage structures into waterways and cause considerable sedimentation in nearby streams and lakes. Tropical countries like Malaysia has a climate which is abetted by monsoon. Without taking proper mitigation, high intensity rainfall strike on denuded slope causing a spate of landslides in the country. Traditional methods
1

have been devised to combat erosion such as retaining wall, sheet piles and concrete embankments. However, such solutions may not be acceptable mainly due to the cost implications. An alternative approach is bioengineering, a method using life plants alone or combined with dead or inorganic materials to arrest and prevent slope failures and erosion (Franti, 1996). Advantages of bioengineering solutions are 1) less expensive and lower maintenance than structural measures; 2) environmental compatibility with landscape and limited access sites; 3) strengthen the soil by binding action of vegetation roots; 4) environment friendly of wildlife habitat, water quality improvement and aesthetics; 5) use of natural by-products such as rice straw, jute, coconut fibres etc. 2.0 Bioengineering Techniques Bioengineering control measures have been applied to highway construction to improve slope stability and minimise slope erosion. However, most estimates of soil erosion emphasised on agricultural land. Soil loss equations have been developed using data from studies conducted on cropland. Little information on bioengineering characteristics and performances has been obtained. There are numerous studies where runoff and soil losses under natural and artificial rainfall are measured. Sulaiman (1989) documented soil loss from isolated land use in Peninsular Malaysia and soil loss was much greater in urban development area. He also pointed out an alarming increase in the rate of soil loss following a greater intensity of the land use. Evidence shows that tropical soils erode more

quickly when disturbed and that the impact of erosion is greater than in temperate counterparts (Edwards, 1983; El-Swaify, 1977). Ahmad (1990) highlighted the problems of soil erosion on the North-South Expressway. Unprotected and improperly installed measures on cut slopes exposed the soil surface to rills and gullies erosion. Effective erosion control depended on fast initial vegetation growth and cover. Biodegradable mulch has been used for erosion for many years and there are extensive literature confirming its effectiveness (Lal, 1977a; Jenning and Jarrett, 1985; Cazzuffi, 1994; Mostaghimi et al., 1994). Mulch encourage plant growth by protecting seeds and seedlings, slow down runoff, protect the soil from raindrop impact and increasing soil moisture. The effect of surface cover types, their combinations and percentage ground cover on soil loss were studied by Grace et al. (1998). They found that erosion mat treatment with grass seedling gave better percentage of cover and hence, effective in mitigating erosion losses with a 98% and 88% reduction in cut slope and fill slope sediment yield respectively. Biodegradable mat and hydroseeding (a combination of seeds, fertiliser, tackifier and mulch) are commonly used on construction sites. However, hydroseeding only did not give good protection on bare slope. Non-germinated seeds are transported down the slope with the runoff. Low percentage of grass cover encourages concentrated flow and gullies erosion (Ahmad, 1990; Mostaghimi et al., 1994). Vegetation intercepts the kinetic energy of the raindrops and inhibits sediment detachment and soil erosion. Elwell and Stocking (1976) suggested that about 60% of vegetation cover are sufficient to cope with erosive forces. In a study on Mediterranean shrub cover in Valencia, Spain, Andrew and associates (1998) found that percentage vegetation covers play important role in soil loss and effective with covers higher than 30%. This is in agreement with the study conducted by Thornes (1990). Vetiver (Vetiveria zizanioides) has been applied for erosion control and slope stabilisation on highway projects in Malaysia. It is planted as a hedge across the slope, which acts as a natural barrier that slows down the runoff and allow sediments to be deposited behind the barrier. As a result, natural terraces built up behind the hedge, which further reduces water velocity and soil and water losses. Koon and Lim (1991)
2

showed that vetiver was able to reduce runoff and soil loss to 73% and 98%, respectively compared to the bare soil. The literature clearly indicates that careful planning and implementation of slope construction minimise soil erosion. Additional information is needed to properly select appropriate vegetation measures from currently available alternatives. The selection varies in cost and erosion control efficiency. The objective of this study was to quantify the effect of commonly used bioengineering slope erosion control techniques. The effect of biodegradable mat on vegetation growth and development was examined. The initial costs of the erosion control techniques were also considered to provide a comparison of cost-benefit. 3.0 Materials and Methods Experimental Design The site selected for the study is located at the Biological and Agricultural Engineering Department Field Station, UPM. The site consists of ten standard Universal Soil Loss Equation (USLE) plots, which measure 1.8 m wide by 22 m long, on 9 % slope. The soil was classified as Serdang Series (sandy-clay). These plots were provided with 25 cm deep, 10 cm wide reinforced concrete partition to form the perimeter on three sides, with 10 cm depth extended into the ground. At the downslope end was placed a series of metal roof covered with lids to prevent the direct entry of rainfall. The metal roof acts as a divisor, which divided the runoff into equal portions and passed one part or one-fifteenth through the central slot of the metal roof, into a calibrated, covered divisor tank, while the remaining 14/15 flowing to waste. The excessive runoff from the divisor tank was then subdivided further where onefourth of the flow was collected in a second calibrated, covered tank. The weight of soil in both tanks was adjusted in accordance with the proportion of the total runoff passing into the tanks. Thus, the total soil loss from each plot was fifteen times the weight of soil in the divisor tank plus sixty times the weight of soil in the second tank. Both tanks were carefully emptied and cleaned after each measurement.

The plots were given the following ten treatments: 1. Vetiver (Vetiveria zizanioides), the leaf was trimmed monthly to about 40 cm height; 2. Legume (Arachis pintoi); 3. Bare (control); 4. Hydroseeding after laying coco-rice straw mat ("coco-fibromat" + hydroseeding); 5. Hydroseeding before laying rice straw mat (hydroseeding + "fibromat"); 6. Hydroseeding after laying rice straw mat ("fibromat" + hydroseeding); 7. Hydroseeding; 8. Hydroseeding after laying geojute ("geojute" + hydroseeding); 9. Spot turfing with cowgrass (Axonopus compressus); 10. Close turfing with cowgrass. Vetiver, legume and cowgrass were planted on 1/1/1998. Vetiver and legume were planted 15 cm and 25 cm apart between the clumps respectively. The vertical interval between rows for both grasses was one meter. In close turfing, 23 cm square sod was laid close together and compacted to an even thickness on the surface of the soil. While in spot turfing, the sods were pegged down about 20 cm apart side by side on the soil. The hydroseeding was done by a local contractor on 21/4/1998 and applied at recommended rates with a high pressure sprayer. The hydroseeding contained a mixture of 50 kg of limestone, 50 kg of 15-15-15 NPK, 80 kg of paper mulch, 50 kg rock phosphate, 50 kg of aerolite, 25 kg of Japanese millet (Echinochloa crusgalli) seed, 50 kg of Ruzi grass (Brachiaria ruziziensis) seed, and enough water to fill 3/4 volume of the container. It was applied at a rate of 27 kg/m2. Sample Analysis The total surface runoff was measured and stirred after each runoff-producing storm. The volume of runoff collected was calculated from the knowledge of depth of runoff in the tanks. 500 ml of water sample was taken by grabsampling technique. Gravimetric analysis was carried out to determine the sediment concentration on a storm or daily basis. Soil loss was calculated by the product of the sediment concentration (g/l) and volume of
3

surface runoff (l), and expressed in kilogram per hectare (kg/ha). The total soil loss was obtained by summing up the stormwise losses for each of the runoff plots. Vegetation cover was quantified each month during the study using a visual assessment method. A rod with 20 fixed observation points was placed at 20 random locations in hydroseeding plots. Each observation point was classified as either covered or bare. Rainfall Erosivity Index Rainfall was measured with tipping bucket pluviometer, which was located at the head of the plots. A laptop computer was used with data loggers to record and download the precipitation depth and intensity data. The data obtained from this record for kinetic energy and rainfall erosivity computations as EI30, KE>25 and AIm. The EI30 is the product of the kinetic energy of the storm and its maximum 30-minute intensity (I30). The SI metric-unit version of the energyintensity equation is (Foster et al, 1981):
E = 0.119 + 0.0873log10I E = 0.283 I 76 mm/h [1] [2]

I>76 mm/h

where E is expressed in MJ.mm/ha.h and I is in mm/h. AIm (Lal, 1976) is defined as the product of the amount of rainfall per storm, A (cm) and its maximum 7.5 minutes intensity, Im (cm/h). The index has the unit cm2/h. Kinetic energy (KE) computed by Hudson (1971):
KE = 29.8 [3]

127 .5 I

I >25 mm/h

where KE is in J/m2, and I is the rainfall intensity in mm/h. The erosivity index is the total KE caused by the storm at intensities above 25 mm per hour. Cover Management Factor (C) Cover management factor, C in the USLE was evaluated by summing soil loss and rainfall erosivity index (Mutchler et al., 1994) from the following equation:

C=

A RKLSP

[4]

where, C = cover management factor. A = soil loss (t/ha). R = rainfall erosivity index (MJ.mm/ha.h). K = soil erodibility factor Mg.ha.h/ha.MJ.mm) LS = slope length and steepness factor. P = support practise factor. The standard plots were managed so P equalled 1 and topographic factor LS is unity for each plot. The K value computed from the same field was 0.02 Mg.ha.h/ha.MJ.mm (Yong, 1998). Upon substitution of these values, C in the USLE was computed by the ratio below:
C = 0.05
[5]

phenomena was more severe in March. The 1998 monthly rainfall distribution showed that the highest rainfall occurred in August, indicating the necessity for provision of good vegetation covers, while the lowest rainfall was in October. The average monthly 30-minutes intensity were generally high (>12.72 mm/h) (Soong et al., 1980), except for July, October and November, indicated by lower erosivity indices. Runoff and Soil Loss Figures 2 and 3 show total runoff depth and soil loss of each plot. The data were analysed separately according to the study period. Over the entire 1-year study period, statistical analysis showed that there were no significant differences on soil loss among the treatments (Table 1). The bare plot had significantly greater soil loss and runoff than all of the other treatments. However, there were no significant differences in runoff between the plots with legume and vetiver, vetiver and spot turfing and finally spot turfing and close turfing. Close turfing had 99.4% and 90.0% less soil loss and runoff respectively compared to bare plot (Fig. 2). Spot turfing was the next most effective treatment with a 97.3% and 76.9% less soil loss and runoff with respect to the bare plot. Both plots used cowgrass which covered well the soil surface. The grass intercepted raindrops and decreased the drops impact pressure. The raindrop energy was dissipated before it struck the ground, causing less erosion. Runoff was also greatly reduced by infiltration into the root systems. Vetiver gave 81.2% and 61.8% less soil loss and runoff. Vetiver planted as hedges across the plot slowed down the runoff and sediments deposited behind the hedges. As a result, it reduces water velocity, soil and water losses. The legume was least effective with 67.0% and 41.4% reduction in soil loss and runoff, respectively to the bare plot. This may be due to poor propagation of the legume with less fertiliser input. The bare plot produced 170.3 t/ha/y of soil loss during the study period, which was far greater than the acceptable limit, 13 t/ha/y (Morgan, 1979). Plots treated with legume and vetiver also gave soil loss above the permissible value with 56.1 t/ha/y and 32.0 t/ha/y, respectively. Soil loss from the spot and close turfing, which were the better protection
4

A R

Statistical Analysis Two statistical tests were used to analyse erosion data, correlation analysis and ANOVA. Simple correlation were computed to determine the relation between soil loss (kg/ha) from control plot for each storm with various erosivity indices including raindepth, EI30, KE>25 and AIm. The runoff and soil loss means of individual treatment were compared for significant differences using Duncan's multiplerange test. 4.0 Results and Discussion Rainfall Characteristics Table 1 gives the monthly rainfall distribution for the study site in 1998. Average 30-minutes intensity, I30 and three erosivity indices such as EI30, KE>25 and AIm are also included. The 10year average monthly rainfall distribution from UPM station over the years 1988 to 1997 was compared to the monthly rainfall distribution for 1998 as shown in Fig. 1. The total annual rainfall for 1998 was 1587.5 mm, which can be considered as a dry year because it was less than the 10-year average annual rainfall in UPM, i n 2400 mm. The El N ~o Southern Oscillation occurred at the beginning of the study year reduces the rainfall and the

among the treatments, produced 4.5 t/ha/y and

1.0 t/ha/y, respectively, were lower than the acceptable limit.

Table 1: Rainfall characteristics, runoff and soil loss of different treatments for the study period.
January February March April May Jun July August September October November December Total Mean*

Rainfall (mm) Intensity I30 (mm/h) Erosivity Indices

65.0 14.4 EI30 (MJ.mm/ha.h) KE>25 (J/m2) AIm (cm2/h) 326 1305 39

143.0 21.6 1503 3748 82

88.0 16.6 549 1600 56

162.5 21.9 1310 3257 95

199.5 22.5 2249 4056 148

123.5 22.4 1367 2261 96

134.0 11.8 613 1616 51

240.0 14.9 1190 3117 102

149.0 22.2 2317 2974 109

36.8 4.6 44 32 3

75.8 7.1 178 499 14

170.4 15.0 1533 2621 94

1587.5 13178 27086 891

Runoff (mm)

Control Legume Vetiver Spot Turfing Close Turfing CMH HM MH HB# GH

32.5 25.5 23.5 15.0 5.0 -

72.9 53.9 39.9 16.0 7.6 -

64.8 27.7 23.5 14.5 2.7 -

74.6 55.7 42.8 22.3 7.1 -

92.4 43.1 19.5 29.5 11.4 22.9 20.8 20.9 26.9 23.4

66.3 26.4 14.8 14.3 5.4 9.3 6.5 6.7 14.2 15.6

44.5 22.9 33.2 14.0 8.7 8.3 10.7 12.0 11.9 12.6

66.2 51.1 44.6 15.4 9.9 13.6 9.8 15.2 28.1 17.4

58.5 44.7 8.8 7.4 3.6 7.6 5.5 8.1 25.1 22.0

9.1 3.6 1.0 1.0 0.9 1.1 1.1 1.1 1.9 2.2

7.9 2.8 0.7 1.1 0.8 0.7 0.9 0.8 1.6 0.7

78.0 33.9 2.9 3.7 3.4 4.1 3.3 3.1 3.0 4.6

666.7 391.3 255.2 154.2 66.5 67.6 58.6 67.9 112.7 98.5

55.6A 32.6B 21.3BC 12.9CD 5.5D 8.5a 7.3a 8.5a 14.1a 12.3a

Soil Loss (kg/ha)

Control Legume Vetiver Spot Turfing Close Turfing CMH HM MH HB# GH

8750 4875 4595 1885 945 -

29587 18699 7187 409 25 -

13833 6878 3320 278 6 -

31254 17729 13470 1435 13 -

57184 3657 914 343 8 109 47 54 1046 349

8697 1925 1190 141 2 30 8 10 937 207

3872 523 593 16 2 10 5 5 519 30

3002 1097 701 10 3 5 3 3 764 42

4984 538 38 11 6 6 3 3 712 81

512 17 2 1 0 0 0 0 3 2

395 10 1 1 1 1 1 1 1 1

8235 190 3 0 0 2 2 2 2 2

170305 56138 32014 4530 1011 163 69 78 3984 713

14092A 4678B 2668B 378B 84B 20b 6b 7b 332a 59b

"Coco-fibromat" + hydroseeding CMH = cocoMH = Hydroseeding + "fibromat" "Fibromat" + hydroseeding # Hydroseeding only "geojute" + hydroseeding * Means with same letter are not significantly different ( = 0.05)

No significant differences were observed during the 8-months study among the hydroseeding plots with biomats. Hydroseeding alone had significantly greater soil loss than all other treatments. However, the runoff depths among all treatments were not significantly different. Both hydroseeding + "fibromat" and "fibromat" + hydroseeding plots were the most effective treatments, represented 98.3% and 98.0% reduction in soil loss, respectively as compared to "control" plot (hydroseeding only) (Fig. 3).
6

Runoff were reduced in both treatments by 48.0% and 39.8%, respectively. Both results showed no significant difference either hydroseeding was done before or after laying the "fibromat". "Coco-fibromat" + hydroseeding reduced soil loss and runoff by 95.9% and 40.0%, respectively. The increased in soil loss in this treatment may be due to poor germination of seeds at the downslope, as "coco-fibromat" was not placed in close contact with the ground. "Geojute" + hydroseeding gave 82.1% and 12.6% less soil loss and runoff, respectively, compared to the control plot. It was least

effective compared to other "fibromat" treatments, possibly due to the 25 mm woven

1988-97 300 Rainfall (mm) 250 200 150 100 50 0
Ju ne h M ay

open mesh of "geojute" allows high intensity storms to wash away a part of the seeds.
1998 I30 25.0 20.0 15.0 10.0 5.0 0.0
t Se pt em be r O ct ob er No ve m be r De ce m be r

Ja nu ar y Fe br ua ry

M ar c

Ap ril

Ju ly

Month

Figure 1: Monthly rainfall distribution at UPM

Soil Loss 200000 Soil Loss (kg/ha) 150000 100000 50000 0 Bare LegumeVetiver

Runof f 800

Au gu s

Soil Loss 5000 Soil Loss (kg/ha)

Runoff (mm)

Runoff 120 100 80 60 40 20 0

400 200 0 S. C. Turf ing Turf ing

3000 2000 1000 0 CMH HM MH HB GH Tre atm e nt

Tre atm e nt

Figure 2: Total soil loss and runoff for each treatment over one-year study period. All biomat treatments had greater amount of vegetation cover compared to the control plot without biomat. This would result in greater interception, decreased raindrop energy, and decreased runoff due to increased canopy. As shown in Fig. 3, hydroseeding alone did not give good protection on soil erosion although its soil loss, 4.0 t/ha, was much lower than the acceptable limit. Figure 4 shows the soil loss ratio for each treatment during the one-year study period. The data were plotted with a best-fit polynomial
7

Figure 3: Total soil loss and runoff for each treatment over eight-months study period. equation. The results indicate that the soil erosion - time relationship were curvilinear. The soil losses were high in the initial study period and decreased with time. In Fig. 4, close and spot turfing gave better protection after one and two months, respectively. Cowgrass is a fast growing stoloniferous perennial and quickly forms a dense mass and ground-hugging turf. Vetiver required 8 months to form dense and tight hedges, which blocked the movement of soil. Least effective cover was found from the legume which took 10 months due to poor propagation as was mentioned earlier.

Runoff (mm)

600

4000

I30 (mm/h)

"Fibromats" and "geojute" treatments gave good protection immediately after installation.
Legume
Soil Loss Ratio (%) 100 80 60 40 20 0
Soil Loss Ratio (%)

Vetiv er
100 80 60 40 20 0

2 y = 0.3103x - 9.654x + 72.827 2 R = 0.7384

y =0.2585x 2 - 7.2142x +49.498 R2 =0.6446

Month

M onth

Spot Turfing
Soil Loss Ratio (%)

Close Turfing
Soil Loss Ratio (%) 100 80 60 40 20 0

100 80 60 40 20 0

y =0.3389x2 - 5.7891x +23.055 R2 =0.8381

y = 0.147x 2 - 2.3282x + 8.11 R2 = 0.5065

Month

Month

Figure 4: Monthly distribution of soil loss ratio for each treatment.

Soil Loss (kg/ha) per EI (%) 30

100 80 60 40 20 0 0

y = 0.0004x 3 - 0.0845x 2 + 4.4322x + 11.832 R2 = 0.9824

20

40

60

80

100

% Cover Hydroseeding Poly. (Hydroseeding)

Figure 5: Effect of percentage of cover on soil loss per EI30 on hydroseeding only. The biomats offered nearly 100% coverage and protected the soil surface from raindrop impact, hence reduced soil detachment. The Millet and Ruzi grasses from plots overlaid with biomats appeared to establish more quickly compared to the hydroseeding only. The high water retention capacities provided by the biomats encourage the growth of the grasses. Interception was increased by grass development and detachment was reduced as the root network develops.
8

Hydroseeding alone only gave good protection in October, after 6 months in which Ruzi grass covered about 90% of the plot (Fig. 5). The increase of soil loss (kg/ha) per EI30 from 0% 30% cover was due to the formation of rill erosion which diverted the concentrated flow to the centre of the divisor. This demonstrates that hydroseeding alone allows the formation of rill, hence the occurrence of rill erosion.

Correlation coefficients, r relating soil loss from bare plot to various erosivity indices including the raindepth are show in Table 2. There was no strong correlation between the erosivity indices and soil loss. EI30 had a lower correlation when studies conducted in the tropics (Lal, 1977: Balasubramanian and Sivanappan, 1981). Hudson (1971) observed that erosion is entirely caused by rainfall at intensities above a threshold level, 25 mm/h. At intensities lower than this level, soil erosion is negligible. By leaving out the energy of the non-erosive rain, a better correlation was obtained between the KE>25 and soil loss from bare plot. Table 2: Correlation of soil loss from bare plot with erosivity indices. Erosivity index Raindepth EI30 KE>25 AIm Costs Table 3 shows the cost per square meter of each treatment, including costs of labour, fertiliser, seed application and mulching. Vetiver and Arachis pintoi are more expensive than turfing, but the latter gave better protection. The cost of hydroseeding depends on the landscape. Often, overlaying with "fibromat" and "geojute", are more costly but they are very effective erosion control measures. Cover Management Factor Cover management factor, C for various treatments are given in Table 4. Hydroseeding overlaid with "fibromat" was found to give the best protection with C value lower than 0.001 (0.0004) as compared to 0.004 treated with "geojute" and 0.021 treated with hydroseeding only. Meanwhile, close turfing gave good protection with C value 0.004, followed by 0.017 for spot turfing, 0.122 for vetiver and 0.213 for the legume. Two assumptions were made in the computation of the C factor. First, the slope length and steepness factor, LS is
9

always unity. The second assumption was that the soil erodibility factor, K was constant during the study period whereas in reality the K factor can vary because of the effect of soil loss. For instance, the bare plot was exposed for too long until the topsoil deteriorated from excessive erosion and this significantly changed the K value. As a result, we obtained a value of 0.646 for bare soil instead of the theoretical value of 1.000. Table 3: Contract cost of each treatment. Treatment Vetiver Legume (Arachis pintoi) Spot turfing Close turfing Hydroseeding Geojute Fibromat RM/m2 10.50 8.00 2.00 4.00 2.00 - 3.00 1.60 2.50

r 0.41 0.56 0.70 0.64

Table 4: C factor of each treatment. Treatment Bare Legume Vetiver Spot turfing Close turfing Hydroseeding Geojute + hydroseeding Coco-fibromat + hydroseeding Fibromat + hydroseeding Hydroseeding + fibromat C factor 0.646 0.213 0.122 0.017 0.004 0.021 0.004 0.001 <0.001 <0.001

5.0 Conclusion Runoff and soil loss from different treatments using standard USLE plots were measured

during a one-year study period. Bioengineering soil erosion control techniques were found to have significant effect on reducing soil loss. The following conclusions can be drawn from this study: 1. In one-year study period, close turfing with cowgrass was the best treatment, reducing soil loss by 99% compared to the bare plot. 2. Hydroseeding + "fibromat" gives better protection among the plots treated with hydroseeding. This technique reduced soil loss by a factor of fifty-seven compared to hydroseeding only. Hydroseeding overlaid with "fibromat" gave the best protection with a C factor of 0.0004. 3. The "fibromat" can be considered to be the most reliable erosion control technique since it provides a more secure cover to protect the soil surface from raindrop impact and enhance the growth and development of vegetation. 4. Without biomat, hydroseeding alone required 6 months to form about 90% cover in order to have effective protection. 5. The KE>25 can be considered as a better erosivity index than any other commonly used indices. 6. A combination of control measures usually improve protection from erosion Acknowledgements This work is a part of IRPA Project 51350. The financial support of MPKSN is acknowledged. The authors are also indebted to all staff of the DBAE Field Station. Their help in this project is greatly appreciated. References Ahmad, A. K. Z. (1990). Erosion of Slopes. In: C. H. Chin, M. J. D. Dobie and R. A. Nicholls (Eds.), Seminar on Geotechnical Aspects of the North-South Expressway, PLUS Projek Leburraya Utara-Selatan Bhd., Malaysia, pp 283-292. Andreu, V., J. L. Rubio and R. Cerni (1998). Effects of Mediterranean Shrub Cover on Water Erosion (Valencia, Spain). J. Soil and Water Cons. 53(2): 112-120.
10

Balasubramanian, G. and R. K. Sivanappan (1981). Effect of Degree of Slope and Rainfall Erosivity on Soil Erosion and the Influence of Mulching on Runoff and Soil Loss. Proceedings of the South-East Asian Regional Symposium on Problems of Soil Erosion and Sedimentation, January 27-29, 1981, Asian Institute of Technology, Bangkok, pp 29-36. Cazzuffi, D., F. Monferino, R. Monti and P. Rimoldi (1994). Experimental Evaluation of the Erosion on Bare and Geosynthetically Protected Slopes. In: R. J. Rickson (Ed.), Conserving Soil Resources, Cranfield, pp 413-421. Edwards, K (1983). Analysis of Runoff and Soil Loss from Long-term Plots. In: S. A. El-Swaify, W. C. Moldenhauer and A. Lo (Eds.), Soil Erosion and Conservation of America, Ankeny, pp 472-479. Elwell, H. A. and M. A. Stocking (1976). Vegetal Cover To Estimate Soil Erosion Hazard in Rhodesia. Geoderma 15: 61-70. El-Swaify, S. A. (1977). Susceptibilities of Certain Tropical Soils to Erosion by Water. In: D. J. Greenland and R. Lal (Eds.), Soil Conservation and Management in the Humid Tropics, John Wiley & Sons, London, pp 71-77. Foster, G. R., D. K. MCcool, K. G. Renard and W. C. Moldenhauer (1981). Conversion of the Universal Soil Loss Equation To SI Metric Units. J. Soil and Water Cons. 36(6): 355-359. Franti, Thomas G. (1996). Bioengineering for Hillslope, Streambank and Lakeshore Erosion Control. Cooperative Extension, Institution of Agriculture and Natural Resources, University of Nebraska-Lincole. Grace, J. M. III, B. Rummer, T. A. Dillaha and R. A. Cooke (1994). Evaluation of Erosion Control Techniques on Forest Roads. Trans. of the ASAE 41(2): 383-391. Hudson, N. W. (1971). Soil Conservation. Bt Batsford Ltd., London. Jenning, G. D. and A. R. Jarrett (1985). Laboratory Evaluation of Mulches in Reducing Erosion. Trans. of the ASAE 28(5): 1466-1470.

Koon, P. K. and F. W. Lim (1991). Vetiver Research in Malaysia - Some Preliminary Results on Soil Loss, Runoff and Yield. In: Vetiver Newsletter no. 5, March 1991. Vetiver Information Network, ASTG, World Bank, Washington D.C. Lal, R. (1976). Soil Erosion on Alfisols in Western Nigeria, III. Effect of Rainfall Characteristics. Geoderma 16: 389-401. Lal, R. (1977a). Soil - Conserving Versus Soil Degrading Crops and Soil Management for Erosion Control. In: D. J. Greenland and R. Lal (Eds.), Soil Conservation and Management in the Humid Tropics, John Wiley & Sons, London, pp 81-86. Lal, R. (1977b). Analysis of Factors Affecting Rainfall Erosivity and Soil Erodibility. In: D. J. Greenland and R. Lal (Eds.), Soil Conservation and Management in the Humid Tropics, John Wiley & Sons, London, pp 49-56. Morgan, R. P. C. (1979). Soil Erosion, Longman, London. Mostaghimi, S., T. M. Gidley, T. A. Dillaha and R. A. Cooke (1994). Effectiveness of Different Approaches for Controlling Sediment and Nutrient Losses from Eroded Land. J. Soil and Water Cons. 49(6): 615 - 620. Mutchler, C. K., C. E. Murphree and K. C. McGregor (1994). Laboratory and Field Plots for Erosion Research. In: R. Lal (Ed.), Soil Erosion Research Methods, 2ed., SWSS-ISSS, Ankeny, pp 11-37. Soong, N. K., G. Haridas, C. S. Yeoh and P. H. Tan (1980). Soil Erosion and Conservation in Peninsular Malaysia. A Monograph, Rubber Res. Inst. Malaysia. Sulaiman, W. W. H. (1989). Fighting Erosion in the Humid Tropics. In: O. Van Cleemput (Ed.), Soil for Development. ITC - Ghent, Publication series no. 1, Belgium, pp 85-107. Thornes, J. B. (1990). The Interaction of Erosional and Vegetational Dynamics in Land Degradation: Spatial Outcomes. In: J. B.

Thornes (Ed.), Vegetation and Erosion, John Wiley & Sons, Chichester, pp 41-55. Yong, H. L. (1998). Comparison of Four Agronomic Erosion Control Measures. Bachelor Degree Project, UPM, Serdang.

11