Directory UMM :Data Elmu:jurnal:S:Soil & Tillage Research:Vol52.Issue1-2.Sept1999:

(1)

Runoff and soil erosion as affected by plastic

mulch in a Hawaiian pineapple ®eld

Y. Wan

*

, S.A. El-Swaify

Department of Agronomy and Soil Science, University of Hawaii at Manoa 1910 East-West Road, Honolulu, HI 96822, USA

Received 11 January 1999; received in revised form 26 April 1999; accepted 7 May 1999

Abstract

The effects of plastic mulch in Hawaiian pineapple culture on runoff generation and soil erosion are poorly understood and not covered in the literature. Field rainfall simulation experiments were conducted on an Oxisol commonly used in Hawaii for pineapple production. Four treatments were no cover (hereafter called bare), plastic mulch as the sole cover (plastic), pineapple crowns as the sole cover (crown), and both plastic and crowns as covers (plastic±crown). The average slope of these plots was 4.2% and all were shaped into ridge-and-furrow con®gurations. For each treatment, three successive storms were imposed: a `dry run' with an intensity of 35 mm hÿ1on the initially dry soil; a `wet run' with the same intensity on the following day; and a `very wet run' with an intensity of 62 mm hÿ1that immediately followed the wet run. Runoff was measured and sampled every 5±10 min until a steady state was reached. Results indicated that plastic mulch itself substantially accelerated runoff generation and soil erosion due to its impervious nature. The simultaneous presence of plastic mulch and pineapple crowns, however, tended to retard runoff generation and reduce soil erosion. For all storms, runoff and erosion rates in the plastic±crown plot were smaller than those of the bare and the crown plots (ca. 30 50% less) due to the formed micro-basins in plastic±crown system that enhanced in®ltration. Plots without plastic mulch displayed soil surface sealing that impeded in®ltration. The practical implication of the research is that the plastic mulch used in Hawaiian pineapple plantations may not necessarily increase runoff and soil erosion.#1999 Elsevier Science B.V. All rights reserved.

Keywords:Runoff; Soil erosion; Plastic mulch; Pineapple

1. Introduction

Many intensive agricultural systems have the poten-tial to induce accelerated soil erosion on site and water quality deterioration off site. The ridge-and-furrow system, commonly practiced for cultivation of row crops, is a typical example of these practices. In such

system, runoff in the ridge (interrill) area is delivered laterally into the furrow (rill), producing concentrated overland ¯ow (rill ¯ow). This concentrated ¯ow pos-sesses higher erosive power than unconcentrated ¯ow under the same land and rainfall conditions, and serves as the primary vector for sediment transport. For a given ridge-and-furrow system, soil erosion in the furrow is largely determined by sediment and runoff delivery from the ridge area, which, in turn, depends on rainfall intensity, antecedent soil water content, and the plantation practices applied in the system.

*Corresponding author. Tel.: 956-8708; fax: +1-808-956-6539

E-mail address:ywan@hawaii.edu (Y. Wan)

(2)

In Hawaii, the ridge-and-furrow system is typically

used for pineapple (Ananas comosus(L.) Merr.)

cul-ture, and plastic mulching over the ridge area has been a common practice. The plastic mulch is convention-ally applied in 0.6-m wide strips over the bed. The edges are covered by soil forming a furrow between two beds. The primary purpose of using plastic mulch is to control nematode following soil fumigation. Other bene®ts include increasing soil temperature, retaining soil moisture, and controlling weed growth. Early research indicated that soil loss from Hawai-ian pineapple land ranged up to 35 Mg haÿ1yearÿ1 and was higher than that from sugarcane ®elds (El-Swaify and Cooley, 1980). This has been attributed to the unpaved ®eld access roads and exposed ®eld during tillage and early growth stage (El-Swaify et al., 1993). The plastic mulch effects on the dynamics of runoff generation and soil erosion are, however, poorly understood and not documented. Available information in the literature has well demonstrated that crop residue mulch enhances in®ltration and reduces soil erosion (e.g., Gregory, 1984; Alberts and Neibling, 1994). Non-erodible materials such as rock are also found to be able to reduce soil erosion by serving as a protective soil cover (e.g., Simanton et al., 1984; Benkobi et al., 1993). This concept has been adapted for soil erosion prediction models such as the Revised Universal Soil Loss Equation (Renard et al., 1991). In a plastic mulched pineapple ®eld, the plastic mulch can be considered as a non-erodible material, and the overall soil surface area susceptible to soil erosion is less than without plastic mulch. However, plastic mulch is also an impervious surface capable of inducing runoff and increasing soil erosion. It is clear that a quantitative understanding of the dual opposite roles of plastic mulch requires empirical experimenta-tion. The objective of this study was to assess plastic mulch effects on runoff generation and soil erosion under varying antecedent soil water contents and rainfall intensities.

2. Materials and methods

2.1. Soil and treatments

Field rainfall simulation experiments were con-ducted on the Wahiawa silty clay soil (a Rhodic

Eutrustox) at the Poamoho Experimental Station on the island of Oahu, Hawaii. Mean annual rainfall in the area is ca. 1100 mm with 1-h 10-year storm of ca. 60± 70 mm. The soil is widely used for pineapple culture and diversi®ed agriculture and has been shown to undergo signi®cant erosion that is linked to non-point source water pollution problems (El-Swaify and Cooley, 1980; El-Swaify et al., 1993). Selected soil properties are shown in Table 1.

Prior to rainfall simulation, the site was disked and smoothed with a rototiller to simulate conventional seed-bed conditions and to maximize surface unifor-mity. Each plot was prepared for rainfall simulation so as to consist of two 8.5-m-long and 0.6-m-wide ridges with a 0.5-m-wide and 0.15-m deep triangular furrow in between. A plastic border was inserted in the center of each ridge as the drainage divide, resulting in a plot width of 1.2 m. The average slope for the furrow bed was 4.2% and that for the furrow side was ca. 30%. This layout was prepared to mimic the agricultural ridge-and-furrow system used for pineapple culture in Hawaii.

A total of four plots representing four different treatments were prepared. These plots were a control plot with the bare soil (hereafter called bare), a plot with plastic mulch as the sole ridge cover (plastic), a plot with the pineapple crowns as the sole ridge cover (crown), and a plot with plastic mulch and crowns as ridge covers (plastic±crown). For plots with plastic mulch, a standard plastic sheet (60 cm wide), which is commercially available for pine-apple culture, was laid over the bed and the edges were covered with soil, providing ca. 45% coverage for the ridge area. The crowns were planted, as done by growers, by insertion through the plastic according to the marks in the sheet. The areal co-verage of the pineapple crowns in the plot was ca. 20%. The plastic plot and the plastic±crown plot represented actual stages of ®eld preparation for pine-apple culture involving plastic mulch. The other two plots were tested to aid understanding the processes involved.

2.2. Rainfall simulation

(3)

nozzles from low line pressure (6.4104Pa). The median drop diameter of simulated rain was 2.2 mm. Rain fell from a height of 3 m above the soil

surface resulting in kinetic energy of ca.

20 J mÿ2mmÿ1 of rainfall. Domestic water supply with an average water temperature of 25₈C was used. Wind effects were minimized by use of a wind-shield around the rainfall simulator. Three storms were simulated on each plot. The ®rst storm (dry run), applied at the soil water condition prevailing at the time (ca. 0.20 kg kgÿ1), had a rainfall intensity of 35 mm hÿ1. The second storm (wet run) with the same rainfall intensity was imposed on the following day to simulate high antecedent soil water conditions (ca.

0.40 kg kgÿ1), and to determine the in¯uence of

storm-induced surface sealing or compaction on run-off rate and soil erosion. The third storm (very wet run) was applied immediately following the wet run and

had a rainfall intensity of 62 mm hÿ1. Simulated

storms were terminated when a steady runoff rate (steady state) was deemed present. For the dry run, this took ca. 3±5 h. For the wet and very wet runs, steady state was reached within 30 min and rainfall was extended for 1 h.

2.3. Sampling and measurements

For each simulated storm, time to runoff initiation was recorded. Runoff was volumetrically measured and sampled every 5±10 min. Visual observations of ponding and soil surface conditions were made and recorded both during and after simulated storms. After steady state was reached, the dimension and velocity of the rill ¯ow at 1, 2, and 3 m upslope from the sampling outlet were measured. A ¯uorescent dye was used for ¯ow-velocity measurement (Gilley et al., 1990) and a mm-scale ruler was used to determine ¯ow dimension. Flow stream power was calculated as follows:

!gSQ=w (1)

where!stream power (W mÿ2),water density

at a given temperature (kg mÿ3),gacceleration of gravity (9.81 m sÿ2), Sslope (m mÿ1), Q¯ow rate (m3sÿ1), andw¯ow width (m).

After rainfall simulation, plastic mulch in the plas-tic±crown and plastic plots was removed and undis-turbed surface soil samples were taken at ®ve locations for bulk density measurements. Similar

Table 1

Selected soil properties of the Wahiawa soil

Soil property Depth (m) Replicates Mean (standard deviation)

pH (soil water ratio1 : 1) 0±0.1 3 5.80(0.04)

Organic carbon (%) 0±0.1 5 1.52(0.03)

Field bulk density (Mg mÿ3) 0±0.1 5 1.01(0.05)

0.3±0.5 5 1.29(0.04)

Dispersed particle size distribution (%)

sand (>53mm) 0±0.1 3 1.33(0.12)

silt (2±53mm) 0±0.1 3 9.59(0.50)

clay (<2mm) 0±0.1 3 89.08(0.47)

Wet-sieved aggregate size distribution (%)

4±10 mm 0±0.1 5 9.69(0.32)

2±4 mm 0±0.1 5 15.11(0.77)

1±2 mm 0±0.1 5 21.49(1.66)

0.5±1 mm 0±0.1 5 17.39(2.37)

0.25±0.5 mm 0±0.1 5 14.92(2.30)

0.125±0.25 mm 0±0.1 5 11.30(1.79)

0.063±0.125 mm 0±0.1 5 6.32(1.03)

<0.063 mm 0±0.1 5 3.78(0.78)

Steady-state infiltration rate (mm hÿ1₎a _19.2±25.4

ULSE Erodibility (Ton/acre/EI)a _0.14

(4)

measurements were also performed in surface soil that was exposed to simulated rainfall. Sediment concen-trations of the collected runoff samples were deter-mined in the laboratory. Soil loss per unit land area (g mÿ2hÿ1) was computed using sediment concentra-tion and measured runoff volumes. Sediment delivery from ridge areas into furrow bed was estimated using an interrill erosion model obtained by Wan (1996) for the soil:

EKiIqs1:15 (2)

where Einterrill sediment delivery rate

(kg mÿ2hÿ1), Kiinterill erodibility (kg h mÿ4),

which has a value of 630 for the particular soil studied, Irainfall intensity (m hÿ1),qrunoff rate (m hÿ1), andSinterrill slope (m mÿ1).

3. Results and discussion

3.1. Runoff generation and runoff rate

The plastic mulch signi®cantly in¯uenced runoff generation under dry soil conditions. Time to runoff initiation for the dry run was in the order: plastic < bare < crown < plastic±crown with 50, 75, 120, and 210 min after rainfall started, respectively (Fig. 1). Visual observations on the bare control plot indicated that the ridge area generated a thin layer of overland ¯ow that laterally drained into the furrow bed. This thin sheet ¯ow was an essential component of runoff generated in the plot. For the plastic mulch plot, runoff

was generated earliest due to the impermeable nature of the plastic mulch. Intercepted rain formed ponds on small depressions in the plastic mulch. However, under continuous raindrop impact, these ponds col-lapsed and water ¯owed into the furrow. Unlike the continuous sheet overland ¯ow on the bare plot, runoff from the interrill area was channeled into the furrow through a system of small rills. For the crown plot, runoff was delayed until 120 min after rainfall started, possibly because of rainfall interception by crowns that slowed sheet ¯ow into the furrow and reduced storm-induced soil surface sealing. For the plastic± crown plot that is characteristic of factual conditions in recently planted ®eld, runoff was generated later than with all other treatments. Visual observations indicated that each crown generated a tiny basin and that ponded water in the plastic sheet was fun-nelled into the hole where the crowns were planted. For the wet run and very wet run, time to runoff generation was similar and short for all the treatments (1±5 min), and thus the difference between treatments was limited.

The effect of plastic mulch on runoff generation has signi®cant implications from a management perspec-tive because low intensity events occur more fre-quently than high intensity events. When crowns are planted, rainfall in®ltrates through the plastic± crown interfaces, is stored in the soil pro®le, and is thus available for post-rain plant growth. The time required for reaching steady state after runoff initia-tion was ca. 40 min for the plastic±crown plot, while that for the other treatments ranged from 1 to 2 h (Fig. 1). This re¯ected the effect of substantial water in®ltration as shown by a very delayed runoff initia-tion time on this plot.

The presence of plastic mulch also in¯uenced run-off rate as shown by the steady-state runrun-off rates and runoff coef®cients (Table 2). The runoff coef®cient (ROC), de®ned as the percentage of rainfall converted into overland ¯ow during an event, was consistently highest for the plastic mulch plot, with mean steady-state runoff rates ranging from ca. 19 mm hÿ1for the dry run to 47 mm hÿ1for the very wet run. The bare plot and the crown plot had similar runoff rates and ROC values, indicating that the limited crown cover (20%) was not effective in increasing in®ltration or preventing soil surface sealing. The plastic±crown plot had consistently lower runoff rates and ROC values

(5)

than other treatments for all the runs. This was likely because surface soil covered by plastic mulch was protected from raindrop impact, and thus maintained a higher in®ltration capacity than the bare or crown plots. Visual observation after rainfall simulation indicated that the soil portion under the mulch con-sisted of large aggregates and crumbs, similar to the condition before rainfall simulation. The average bulk density of the mulched soil was 1.01 Mg mÿ3, sig-ni®cantly smaller (p< 0.01 with at-test) than that of

surface soil without mulch (1.17 Mg mÿ3). This

showed that surface compaction and sealing were induced when soil was exposed to rainfall impact. A similar conclusion was also reached in early studies on the soil (El-Swaify, 1980; Sutherland et al., 1996). The compaction and sealing effect may also be par-tially responsible for the slightly higher runoff rates for the wet run than for the dry run as indicated in Table 2, thus adding into the effect of reduced soil water saturation de®cit (increased soil water content) on in®ltration rate.

3.2. Soil loss and sediment concentration

Table 3 shows sediment concentrations and soil loss rates in the steady state. For all runs, the plastic mulch plot consistently produced the highest sediment con-centration and soil loss rate while the plastic±crown plot always displayed lowest sediment concentration and soil loss rate. This difference was largely induced by the treatment effect on runoff rate, which in turn

determines the stream power of the ¯ow when slope is constant (Eq. (1)). Fig. 2 shows the strong correlation between steady-state soil loss rate and ¯ow stream power under the experimental conditions.

It is well known that detached sediment is entrained more easily by ¯ow than undetached soil due to its lower interparticle cohesion (Hairsine and Rose, 1992). Under our experimental conditions, interrill erosion in the ridge area delivered sediment into the furrow bottom for further transport by rill ¯ow. Thus, the plastic mulch acted as a non-erodible soil cover for interrill erosion. The amount of sediment delivered into the furrow for plots with plastic mulch should be correspondingly less than the plots without plastic mulch given similar ¯ow discharge conditions. The

Table 2

Treatment effects on steady-state runoff rates and runoff coeffi-cients

Treatment Dry run Wet run Very wet run

Steady runoff rate (mm hÿ1)a

Bare 11.45(0.19) 13.56(0.40) 37.24(0.45) Plastic 18.43(0.56) 19.89(0.21) 47.12(0.15) Crown 9.382(0.25) 13.99(0.34) 37.37(0.26) Plastic±crown 8.271(0.34) 11.65(0.23) 31.67(0.43)

Runoff coefficient (%)

Bare 15.40 38.26 57.32

Plastic 23.95 48.83 74.10

Crown 9.534 39.11 57.50

Plastic±crown 5.432 26.96 48.35

a_{Mean of the final three measurements with standard deviation}

in parentheses.

Table 3

Treatment effects on steady-state sediment concentrations and soil loss ratesa

Treatment Dry run Wet run Very wet run

Sediment concentration (g lÿ1)

Soil loss rate (g mÿ2hÿ1)

a_{Mean of the final three measurements with standard deviation}

in parentheses.

(6)

estimated sediment delivery rates using Eq. (2) for the bare plot were ca. 63, 74, and 364 g mÿ2hÿ1for the dry, wet, very wet runs, respectively. According to the coverage provided by plastic mulch and crowns, the interrill sediment delivery rates were at most 50% of these values for the plastic±crown plot, and 80% for the crown plot. Comparing these values with soil loss rates in Table 3, it is apparent that the amount of sediment delivered by interrill erosion exceeded rill ¯ow detachment capacity during the dry and wet runs but was below ¯ow detachment capacity during the very wet run. This assumption can be con®rmed by the data in Fig. 3 in which sediment concentration is plotted against time. For all plots, sediment concen-tration during the dry and wet runs ranged from ca. 1.5 to 5.0 g lÿ1, and the temporal variation during these runs was limited. However, for the very wet runs sediment concentration peaked within 20 min and then tended to decline with time. This trend during the very wet run probably resulted from an initial transport of deposited sediment from previous runs and a subsequent detachment of in-situ soil after the deposited sediment was depleted. This analysis further illustrates that soil loss in pineapple ®eld is more

related to runoff rate or runoff erosive power than to interrill sediment delivery into the furrow.

4. Conclusions

In a dry soil condition, plastic mulch alone sub-stantially accelerated runoff generation due to its impervious nature. The simultaneous presence of plastic mulch and pineapple crowns, however, tended to retard runoff generation. For all the simulated storms, the plastic plot displayed the highest runoff and soil erosion rates while the plastic±crown plot had the least runoff and soil erosion. The primary mechan-ism responsible for the delayed runoff generation and reduced soil loss in the plastic±crown treatment was that the plastic±crown system enhanced in®ltration through the formed micro-basins.

The results obtained in this study have practical implications for soil and water conservation in pine-apple culture. Firstly, it is important to plant pinepine-apple crowns as soon as possible after the plastic mulch is laid in the ®eld. Secondly, suf®cient amount of soil should cover the plastic mulch edges to guide water through the holes where crowns are planted. Because this study was conducted on a relatively mild slope, the bene®ts of the combined plastic±crown treatment can only be assured in this slope range. For land with steeper slopes, furrow and plastic mulch layouts must be installed at contours and mild slopes to assure these bene®ts.

Acknowledgements

CTAHR Journal Series No. 4410. Financial support from USDA Section 406, T-STAR Grant Agreement No. 91-34135-6177, and ®eld assistance from T. Burke, A. Mirza, S. Oshita, and J. Zhang are greatly acknowledged. Thanks are also due to National Soil Erosion Research Laboratory, USDA-ARS, for pro-viding the rainfall simulator.

References

Alberts, E.E., Neibling, W.H., 1994. Influence of crop residues on water erosion. In: Unger, P.W. (Ed.), Managing agricultural residues. Lewis Publishers, Boca Raton, FL, pp. 19±39. Fig. 3. Change of sediment concentration with time during dry run,

(7)

Benkobi, L., Trlica, M.J., Smith, J.L., 1993. Soil loss as affected by different combination of surface litter and rock. J. Environ. Qual. 22, 657±661.

Dangler, E.W., El-Swaify, S.A., Ahuja, L.R., Barnett, A.P., 1976. Erodibility of selected Hawaii soils by rainfall simulation. USDA-ARS, Western Region, Oakland, CA.

El-Swaify, S.A., 1980. Physical and mechanical properties of Oxisols. In: Theng, B.K.G. (Ed.), Soils with variable charge. New Zealand Soc. Soil Sci., Wellington, pp. 303±324. El-Swaify, S.A., Cooley, K.R., 1980. Sediment losses from small

agricultural watersheds in Hawaii (1972±1977). Sci. and Education Admin., Agric. Reviews and Manuals, ARM-W-17, Oakland, CA.

El-Swaify, S.A., Zhang, J., Palis, R., Rose, C.W., Ciesiolka, C.A.A., 1993. Erosion problems and conservation needs of pineapple culture. Acta Horticulturae 334, 227±239. Foster, G.R., Neibling, W.H., Nattermann, R.A., 1982. A

programmable rainfall simulator. ASAE Paper No. 82-2570, Chicago, IL.

Gilley, J.E., Kottwiz, E.R., Simanton, J.R., 1990. Hydraulic characteristics of rills. Trans. ASAE 33, 1900±1906.

Gregory, J.M., 1984. Prediction of soil loss by water and wind for various factions of cover. Trans. ASAE 27, 1345±1350. Hairsine, P.B., Rose, C.W., 1992. Modeling water erosion due to

overland flow using physical principles I, Sheet flow. Water Resources Res. 28, 237±243.

Renard, K.G., Foster, G.R., Weesies, G.A., McColl, K.D., Yoder, D.C., (Coordinators), 1991. Predicting soil erosion by waterÐ A guide to conservation planning with the revised universal soil loss equation (RUSLE). USDA, Agricultural Handbook No. 703.

Simanton, J.R., Rawitz, E., Shirly, E.D., 1984. Effects of rock fragments on erosion of semiarid rangeland soils. In: Soil erosion and productivity of soil containing rock fragments. SSSA Spec. Publ. 13. SSSA, Madison, WI, pp. 65±72. Sutherland, R.A., Wan, Y., Ziegler, A.D., Lee, C.-T., El-Swaify,

S.A., 1996. Splash and wash dynamics: an experimental investigation using an Oxisol. Geoderma 69, 85±103. Wan, Y., 1996. Soil erosion processes and sediment enrichment