DEVELOPMENT AND CALIBRATION OF A RAINFALL SIMULATOR

R. Pall, W. T. Dickinson, D. Reals, and R. McGirr

School of Engineering, UniversityofGuelph, Guelph, Ontario NIG 2W1

Received 2 July 1982, accepted 25 February 1983

Pall, R., W. T. Dickinson, D. Reals, and R. McGirr.Can. Agric. Eng. 25: 181-187.

1983. Development and calibration of a rainfall simulator.

A rainfall simulator involving a large-capacity wide-angle spray nozzle and a spray interception device has been developed for the soil erosion research program at Guelph. A rotating disk with multiple variable aperture openings hasbeen used for spray interception. Calibration tests show that the simulated rainfall intensity and the uniformity ofapplication are affected by the aperture angle, nozzle pressure, disk angular velocity and the interaction of nozzlepressure and aperture angle. Aperture angle has the greatest effect on intensity. Nozzle pressure demonstrates the mostsignificant effect on uniformity of simulated rainfall. The uniformity of distribution for a small plot is also affected bythe size of the collector units considered in the determination of the uniformity coefficient. For selected combinationsof nozzle pressure, aperture angle, and disk angular velocity, the simulated rainfall intensity and the uniformity ofdistribution can be represented by a linear model involving the plot dimensions of length and width.

INTRODUCTION

One very effective approach to thestudy of soil erosion processes and theanalysis of possible remedial strategies involves the use of rainfall simulation. Al

though to date it has proven virtually impossible to reproduce artificially all thephysical characteristics of natural rainfall,rainfall simulation offers researchers con

trol of precipitation conditions such as intensity, frequency and duration. Simulation, therefore, has become a powerfulresearch tool for the development andevaluation of soil conservation programs.

Although the use of rainfall simulationfor soil erosion research is not new, literature regarding the development andparticularly the calibration of such simulators appears sparse. Therefore the purpose of this paper is to describe thedevelopment of a simulator system, calibration methodology, and some calibration results.

TYPES OF RAINFALL

SIMULATORS

The artificial reproduction of rainfall tostudy different hydrologic processes hasled to the development of different typesof rainfall simulators (Mutchler andHermsmeier 1965). On the basis of thedrop-forming mechanism used, simulators can be grouped into two major categories: (i) dripping rainfall simulatorsform drops at the tip of a material and initiate the fall of drops with zero velocity,and (ii) nozzle rainfall simulators forcewater drops from a nozzle at significantvelocities.

In dripping simulators the commonpractice has been to form drops at the tipof a material by some suitable device (i.e.a drop former) until the weight of the dropovercomes the surface tension force of the

drop former and the drop falls with an initial velocity of zero. In the early stage ofdevelopment of such simulators, hangingyarns (Barnes and Costel 1957) and glasscapillary tubes (Adams et al. 1957) wereused as drop formers. Recent designs include metal tubes (Mutchler and Molden-hauer 1963; and Gabriels and DeBoodt1975), hypodermic needles (Romkens etal. 1975; Walker et al. 1977), and polyethylene and plastic tubing (Chow andHarbough 1965; Chow and Yen 1974;Kinnell 1974).

To create a wider drop size distribution,many different types of nozzles have beenused for rainfall simulation (Meyer andMcCune 1958; Bubenzer and Meyer 1965;Morin et al. 1967; Amerman et al. 1970;Cluff and Boyer 1971; Rawitz et al. 1972;Marston 1978; Meyer and Harmon 1979).In nozzle simulators, the spray pattern,median drop size and drop size distribution are governed by the shape characteristics and discharge of the nozzle. Available nozzles can produce drop and energycharacteristics comparable to natural rainfall but relatively high capacity limits theiruse in rainfall simulation. The problemsof high flow rates have been resolved bythe following methods.

1. Covering a large area by spraying thewater upward (Simulators using F typeand sprinkler irrigation nozzles belong tothis group).

2. Physically moving the nozzle backand forth across the plot (e.g. the rainu-lator developed by Meyer and McCune(1958)).

3. Intercepting a major portion of thespray by introducing a physical obstruction (as done by Morin et al. (1967) byplacing a rotating disk with an aperture inthe nozzle spray path).

CANADIAN AGRICULTURAL ENGINEERING, VOL. 25, NO. 2, WINTER 1983

SELECTION OF A SIMULATOR

The first step in the development of arainfall simulator is the establishment of

selection criteria depending upon the rainfall characteristics required and objectivesof the research program. For erosion research some of the most important characteristics for rainfall simulation have

been outlined by Bubenzer (1979) to be:1. Drop size distribution similar to nat

ural rainfall.

2. Drop impact velocity approximatingterminal velocity of natural raindrops.

3. Rainfall intensity in the range of therequirements of the research program.

4. Uniform rainfall and random dropsize distribution.

5. Total energy corresponding to natural rainfall.

6. Accurate reproduction of givenstorms.

Most researchers using rainfall simulators (Borst and Woodburn 1940; Meyerand McCune 1958; Bertrand and Parr1961; Chow and Harbough 1965; Nassifand Wilson 1975; Munn and Huntington1976; Shriner et al. 1977) agree with theaforementioned criteria, and Meyer (1965,1979) added the following characteristics:

7. Rainfall intensity nearly continuousthroughout the study area.

8. Angle of impact nearly vertical formost drops.

9. Sufficient area of coverage.10. Satisfactory characteristics under

varying climates.11. Complete portability from site to

site.

We believe that, in addition to theabove criteria, a reliable rainfall simulatormust have efficient and simple controls.

The developers of the various rainfallsimulators have claimed excellent per-

181

formance for their units, but certainstrengths and limitations can be identifiedto be characteristic of the two main types.The dripping simulators are simple in design and operation. They produce dropsof nearly uniform size (2.2-5.5 mm) depending upon the type and dimensions ofthe drop former. The application ratesover the covered area are fairly uniform.The main advantage of this type of simulator is that it can produce large-size dropsat a low rate of application. The drop sizedistribution is generally narrow but can beimproved at the cost of uniformity of distribution. The randomness of drop sizedistribution is always a big problem, although in some units the drop former and/or plot are oscillated to achieve this.Forced air has also been used to resolve

this problem. As drops start from zero velocity, the drop former must be at a sufficient height from the impact surface toattain reasonable rainfall energy levels.Intensity in these simulators is controlledby changing the pressure head and the sizeof the drop former. In the light of thesecharacteristics, the dripping simulators arebest suitable for infiltration and runoff

studies. They are also useful in splash erosion research where uniform drop size isan advantage.

Nozzle simulators produce storms ofwider drop size distribution. The type Fnozzle (Parsons 1943) was specially developed to produce a high-energy spray atlow application rates. The drop size produced was large but the impact velocity ofupward sprayed drops was small and resulted in a reduced-energy spray. The velocity was controlled by the height atwhich the vertical component of the velocity became zero. From a drop velocitypoint of view, the nozzle simulators facingupwards are similar to the dripping rainfallsimulators. The rainulator developed byMeyer and McCune (1958) uses theSpraying Systems 80100 Veejet nozzle.This nozzle produces a long and narrowspray pattern with intensity decreasingwith distance from the spray center. At apressure of 41.4 kPa the drop size distribution is similar to natural rainfall. The

median drop size of 2.1 mm at an intensityof 50 mm/h is about 15% lower than that

of similar intensity natural rain. The kinetic energy produced by operating thenozzle at a height of 2.44 m is about 80%of the similar intensity natural rainfall.Swanson (1965), and Bubenzer and Meyer(1965), have used the same nozzle on arotating boom and oscillating laboratorysimulators, respectively. Recently, Meyerand Harman (1979) have used 80100 and80150 Veejet nozzles and the concept of

182

oscillation. The 80150 nozzle producesdrop characteristics slightly better than the80100 nozzle. Holland (1969) and Lusby(1977) have tried Rainjet 78C nozzles inlarge size simulators but the kinetic energydeveloped has been much lower than thatof Veejet nozzles.

Many developers have used the conceptof spray interception by using a slottedrotating disk. Spraying Engineering Company's 7LA nozzles were used by Amer-man et al. (1970), and Rawitz et al. (1972)in a modified version of sprinkler infiltro-meters. No attempt was made to measuredrop size. Spraying Systems Fulljet nozzles (1.5H30 and 1HH12) used by Morinet al. (1967), Cluff and Boyer (1971), andMarston (1978) were preferred for erosionresearch. The spray characteristics of the1.5H30 Fulljet nozzle at a pressure of 60.8kPa were similar to natural rainfall. The

median drop size of 2.6 mm comparedwell with 2.5 mm for 50 mm/h natural

rainfall intensity. The kinetic energy ofthe simulated rainfall was better than sim

ulators using Veejet nozzles. These simulators were portable and stable up to awind speed of 25 km/h.

From available literature regardingtested rainfall simulators, only the rotatingdisk simulators using Fulljet nozzles comevery close to satisfying all the aforementioned criteria. Rainulator and drippingsimulators fall second and third, respectively, in the rating. Particularly on the

"DRAIN

basis of better performance, drop characteristics and energy levels, a rotating diskrainfall simulator was selected for devel

opment for the erosion research programat Guelph.

DESCRIPTION OF GULEPH

SIMULATOR

The development and construction ofthe rainfall simulator has addressed the

essential characteristics mentioned ear

lier. Special attention has been given tothe cost, portability and operational features of the unit. A schematic diagram ofthe unit is presented in Fig. 1. The interception concept of Morin et al. (1967)used in a rotating disk simulator wasadopted. The major deviation from theoriginal design is that the Guelph unit hasonly one disk with variable apertures. Theshape of the aperture is similar to a sectorof a frustrum of a cone. Two equallyspaced apertures have been used, so thatthe simulator can be operated at lower diskangular velocities than the original designof Morin et al. (1967).

The nozzles used to date include Fulljet1.5H30 and 1HH12, manufactured bySpraying Systems Co., Bellwood, 111. Thedigits following H indicate the dischargeof the nozzles in US gallons per minute ata pressure of 48 kPa. Both nozzles havea wide spray pattern with a nominal sprayangle of about 90°.

The modification in the original design

TELESCOPIC

^ STAND

TEST PLOT

Figure 1. Schematic diagram of the rainfall simulator.

CANADIAN AGRICULTURAL ENGINEERING, VOL. 25, NO. 2, WINTER 1983

of application, and rainfall kinetic energy.The nozzle has been tested for operatingpressures of 30.4, 40.5, 60.8, 81.1 and101.3 kPa, disk aperture angles of 5, 15,20, and 40°, and disk angular velocities of15, 40, 70 and 100 rpm.

The intensity of simulated rain has beendetermined by collecting the volume ofwater during the known run period (approximately 30 min) in closely placed

106-mm-diameter and 178-mm-high cylindrical cans. From the volume of water

collected, the intensity of application withrespect to length and width of plot hasbeen computed. The uniformity of application has been determined according tothe procedure of Christiansen (1942).Each run was replicated twice. In caseswhen the two observations did not match,a third replication was performed.

is that the variable aperture is achieved bya double disk mechanism. In the disk assembly, one disk with two equally spacedapertures of 40° overlaps a second disk ofsimilar shape and size with three 40° apertures located 90° from center to center.

Sketches of the disks are presented inFigs. 2 and 3. The upper disk can be rotated over the lower disk to vary the aperture opening. Two apertures can beachieved by using alternate openings onboth disks. An option has also been provided for a single aperture angle. This canbe achieved by matching the intermediateopening on the lower disk with any opening on the upper disk. With this arrangement, one or two apertures from 0° to 40°can be obtained.

The disk is powered by a 0.124-kWD.C. electric motor. A high-quality SCR(Boston Gear Ratiopack RP1) motor control and single action 90° gear reducer areused to adjust the disk speed up to 100rpm.

In the early design, the disk was shapedto a shallow cone from a 0.5-mm stainless

steel sheet, but the maintenance of closecontact between the disks was a problem.In a subsequent model, the disk has beenmachined from an aluminum sheet to a

shallow cone with an approximate sideslope of 7.5%. The tail edges of the diskaperture openings have been knife-shaped.The lower disk has four equally spaced 6-mm-diameter holes on a 65-mm-radium

circle (Fig. 2) while the upper disk hasfour 6-mm-wide, 40° arched slots at a radius of 60 mm located 90° from center to

center (Fig. 3). The holes in the lower diskcan be lined to any position with the slotin the upper disk. Socket head screwsthrough the slots into the holes are usedto adjust the aperture angle. Six holes of3-mm diameter are drilled along the periphery of the lower disk. A nut, bolt andwasher system through these holes maintains close contact between the two disks

near the periphery.The disk unit is placed in a horizontal

plane about 20 mm below the nozzle insuch a way that the center of the apertureangle and nozzle orifice are in a verticalline. A panlike collector is mounted underthe disk to collect intercepted water spunfrom the outer periphery of the disk. Thebottom of the pan is shaped convex upward with a circular opening of 200-mmdiameter below the nozzle. A 38-mm hose

is used to drain the pan.

CALIBRATION PROCEDURE AND

PERFORMANCE RESULTS

The developed rainfall simulator hasbeen calibrated for intensity, uniformity Figure 2. Detailed description of the lower disk.

CANADIAN AGRICULTURAL ENGINEERING, VOL. 25, NO. 2, WINTER 1983 183

Figure 3. Detailed description of the upper disk.

velocity of the disk. Theoretically, angular velocity should not have any effect onthe intensity of application, but the results(Fig. 4) indicate an irregular variation ofintensity with disk, angular velocity at allcombinations of nozzle pressure and aperture angle. Application of the Duncanmultiple range test indicates that only theangular velocity of 15 rpm has a significant effect on the intensity at the 5% level.The change in intensity and nozzle discharge with pressure shows a similartrend. At all pressures, the increase in intensity with aperture size is approximatelylinear (Fig. 7).

Further analysis has revealed that thenozzle pressure, aperture angle, disk angular velocity, and the combination ofnozzle pressure and aperture angle has asignificant effect on the uniformity coefficient. However, as exemplified in Fig.8, the uniformity coefficient is influencedprimarily by nozzle pressure. An increasein pressure increases the nominal sprayangle, resulting in an improvement in theuniformity of similated rainfall. Application of the Duncan multiple range testindicates that, except for the conditionsinvolving 15 rpm or 5° or both, the uniformity coefficient is not significantly affected by disk angular velocity or apertureangle.

In all cases, the uniformity coefficientsare noted to slightly lower than those reported by Morin et al. (1967). The difference in experimental procedure may beresponsible for these discrepancies. Morinet al. (1967) did not mention the size ofthe container used to determine the uni-

At 15, 40, 70, and 100 rpm, twoequally spaced apertures provide precipitation with application intervals of 2,0.75, 0.43 and 0.3 sec, respectively.These application intervals are considerably smaller than the 10 sec recommendedby Sloneker and Moldenhauer (1974), andYoung (1979) to avoid delay in surfacesealing.

The average rainfall intensity and coefficient of uniformity of distribution on aplot 848 mm x 742 mm at various combinations of nozzle pressure, angular velocity of the disk and aperture size areshown in Figs. 4 and 5, respectively.Analysis of the data has shown that theaperture angle, nozzle pressure, angularvelocity and the interaction of pressureand aperture angle have a significant effect on the intensity reproduction at the1% level. The influence of aperture sizeon rainfall intensity (Fig. 6) is greater thanthe effects of nozzle pressure and angular

184

E

£ '3

CO

z

y-Z

15

9 -

2z

<

IOO

in

<

90-

80-

70 -

15

15

A = 5°

APERTURE ANGLE,°NOZZLE PRESSURE,kPa

40 70 100

A =20

80-

A= 15°

70-

60

50

15 40

180

160-

-•30.4P '50 -

40 70 IOO

140-

15 40

ANGULAR VELOCITY OF DISK (rpm)

70

70

Figure 4. Simulated rainfall intensity as a function of system variables.

30.4 P

100

IOO

CANADIAN AGRICULTURAL ENGINEERING, VOL. 25, NO. 2, WINTER 1983

A = APERTURE ANGLE , °P = NOZZLE PRESSURE, kPa

ANGULAR VELOCITY OF DISK {rpm)

IOO

Figure 5. Uniformity of application of simulated rainfall as a function of system variables.

formity coefficient. In the field evaluationof sprinkler nozzles a change in size of thecontainer has been shown to have a verysmall effect on the uniformity coefficient,because the area covered by the containeris small in comparison with the total areacovered (the diameter of the containerwith respect to the dimensions of the covered area could be represented by a point).The same condition does not exist for lab

oratory rainfall simulators. The area covered by a single-nozzle rainfall simulatoris generally small (1-2 m2). A 50%change in container dimensions can resultin a significant variation in the uniformitycoefficient. Therefore, this aspect must betaken into consideration when uniformitycoefficients are determined and com

pared.Comparison of the uniformity of distri

bution for the nozzle with and without the

disk (Fig. 9) has revealed that inclusion ofthe disk improved the uniformity coefficient for all aperture angles tested. Although, as stated earlier, only a 5° apertureyielded uniformity coefficient significantly greater than the other openings, alldisk situations examined resulted in a significantly better uniformity than no diskcondition.

The variation of rainfall intensity anduniformity coefficient with plot size isshown in Fig. 10 for a selected combination of nozzle pressure, aperture angle,

P= NOZZLE PRESSURE, kPaS = ANGULAR VELOCITY OF DISK , rpmA =APERTURE ANGLE OF DISK , degrees

160

140

^ 120

EE

— IOO

in

zId

£ 80

60

<

a

40

20

P 20 40 60 80 IOO

s 20 40 60 80 IOO

A iO 20 30 40 50

180

160

2

ID

5

120

100

80

60

40

20

Figure 6. Effect of nozzle pressure, aperture angle and disk angular Figure 7.velocity on the intensity of simulated rainfall.

CANADIAN AGRICULTURAL ENGINEERING, VOL. 25, NO. 2, WINTER 1983

P=N0ZZLE PRESSURE

10 15 20 25 30 40

APERTURE ANGLE (Degrees )

Simulated rainfall intensity as influenced by aperture angleand nozzle pressure.

185

P = NOZZLE PRESSURE , kPaS = ANGULAR VELOCITY OF DISK, rpmA = APERTURE ANGLE OF DISK, degrees

z

89r

88-

87-

86-

85-

4

84- «^

83-

82-

81-

80-

79-

78-

77-

UJ

o

2<r

ou.

z3

30

_L-

40 50 60 70 80

NOZZLE PRESSURE( kPa)

90

NO DISK

100

20

^20"10

40

40~

20

^060~

30

80

_eo'40

100

\\ocT60 Figure 9. Comparison of the uniformity of distribution of simulated

rainfall with and without disk.Figure 8. Effect of nozzle pressure, aperture angle and disk angular

velocity on the uniformity of distribution of simulated rainfall.

and disk angular velocity, i.e. P = 81.1kPa, A = 40°, and 5=100 rpm. The results of this study indicate a linear modelof the form,

a + bL + cW (1)

where Z is rainfall intensity or uniformitycoefficient, L is length of the plot, W iswidth of the plot, and a, b and c areregression coefficients. The coefficientsfor the situation illustrated in Fig. 10 are169, -0.146, and -0.161 for the rainfallintensity model, and 100, -0.085, and-0.045 for the uniformity coefficientmodel. The respective determination coefficients were computed to be 0.989 and0.977.

The experimental results for simulatedrainfall using the procedure outlined byBeals et al. (1983) revealed that the velocity of fall for a particular size of dropis comparable to the terminal velocity ofsimilar drops reported by Laws (1941).The average deviation for the velocity offall from the Laws data was 9%. Thiscould be due to experimental procedure.In the experimental procedure of Laws(1941) the drop size was measured independently from the drop velocity, whilein this study the drop size and the fall velocity were measured on the same photograph.

186

CONCLUSIONS

The rainfall simulator developed withlarge-capacity wide-angle spray nozzleand a rotating disk with multiple variableaperture openings has provided an experimental facility which meets a variety ofselection criteria established for a soil ero

sion research program. The simulatoryields a wide range of useful simulatedrainfall intensities and exhibits excellentspatial uniformity characteristics. Thedrop size and velocity of fall of drops ofsimulated rain are comparable to published data on terminal velocities of sim-

I = INTENSITY, mm/hUC= UNIFORMITY COEFFICIENT, PERCENT

o-I

a.

u_

o

xt-Q

10 20 30 40 50 60 70 80 90 IOO 110 120 130 140 150

LENGTH OF PLOT , cm

Figure 10. Simulated rainfall intensity and its uniformity of distribution as influenced by plotdimensions.

CANADIAN AGRICULTURAL ENGINEERING, VOL. 25, NO. 2, WINTER 1983

ilar drops. The device is portable, economical to build, and simple to adjust andoperate. It provides a versatile and usefulfacility for the soil erosion laboratory.

REFERENCES

ADAMS, J. E., D. KIRKHAM, and D. R.NIELSON. 1957. A portable rainfall simulator infiltrometer and physical measurement of soil in place. Soil Sci. Soc. Am.Proc. 21:473-477.

AMERMAN, C. R., D. I. HILLEL, and A. E.PETERSON. 1970. A variable-intensitysprinkling infiltrometer. Soil Sci. Soc. Am.Proc. 34: 830-832.

BARNES, O. F. and G. COSTEL. 1957. Amobile infiltrometer. Agron. J. 49:105-107.

BEALS, D., W. T. DICKINSON, D. S.HELSDON, and R. PALL. 1983. An approach for characterizing falling water drop-plets. Can. Agric. Eng. 25: 189-192.

BERTRAND, A. R. and J. F. PARR. 1961.Design and operation of Purdue sprinklerinfiltrometer. Purdue University Res. Bull.No. 723: 1-16.

BORST, H. L. and R. WOODBURN. 1940.Rainfall simulation studies of the effect of

slope erosion and runoff. Soil ConservationService, SCS-TP-36: 1-36.

BUBENZER, G. D. 1979. Rainfall characteristics important for simulation. Proc. Rainfall Simulation Workship Tuscon, Arizona.U. S. Dep. Agric, Sci. Educ. Adm., Agric.Review Manual ARM-W-10. pp. 22-34.

BUBENZER, G. D. andL. D. MEYER. 1965.Simulation of rainfall and soils for labora

tory research. Trans. ASAE (Am. Soc.Agric. Eng.) 8: 73-75.

CHOW, V. T. and T. E. HARBOUGH. 1965.Raindrop production for laboratory watershed experimentation. J. Geophys. Un.Res. 70:6111-6119.

CHOW, V. T. and B. C. YEN. 1974. A laboratory watershed experimental system.Hydr. Eng. Series No. 27, Civ. Eng. Dep.,Univ. of Illinois, Urbana, 111. pp. 1-196.

CHRISTIANSEN, J. E. 1942. Irrigation bysprinklers. Univ. of California Agric. Exp.Sta. Bull. 670.

CLUFF, C. B. and D. G. BOYER. 1971. Theuse of a realistic rainfall simulator to deter

mine relative infiltration rates of contribut

ing watersheds to the lower gila belowpainted rock dam. Water Resour. CentreArizona.

GABRIELS, D. and M. DeBOODT. 1975. Arainfall simulator for soil erosion studies in

the laboratory. Pedologie XXV(2): 80-86.HOLLAND, M. E. 1969. Design and testing

of rainfall system. Col. State Univ. Exp.Sta. CER 69-70, MEH21, Fort Collins,Colorado.

KINNELL, P. I. A. 1974. Splash erosion:Some observations with a splash cup technique. Soil Sci. Soc. Am. Proc. 38:657-660.

LAWS, J. O. 1941. Measurement of fall velocity of water drops and raindrips. Am.Soc. Goephys. Trans. 22: 709-721.

LUSBY, G. C. 1977. Determination of runoffand sediments by rainfall simulation. InErosion research techniques, erodibility andsediment delivery. Geo. Abstracts Ltd.Norwich, England.

MARSTON, D. 1978. The use of simulatedrainfall in assessing the erodibility of variousstubble management practices. Conf. onAgric. Eng. pp. 74-77.

MEYER, L. D. 1965. Simulation of rainfallfor soil erosion research. Trans. ASAE

(Am. Soc. Agric. Eng.) 8: 63-65.MEYER, L. D. 1979. Methods of attaining

desired characteristics in rainfall simulation.

Proc. Rainfall Simulation Workshop Tuscon, Arizona. U. S. Dep. Agric, Sci. Educ.Admin., Agric Review Manual ARM-W-10 pp. 35^4.

MEYER, L. D. and W. C. HARMON. 1979.Multiple intensity rainfall simulator for erosion research on row sideslopes. Trans.ASAE (Am. Soc Agric Eng. 22: 100-103.

MEYER, L. D. and D. L. McCUNE. 1958.Rainfall simulation for runoff plots. J.Agric Eng. 39: 644-648.

MORIN, J., D. GOLDBERG, and I. SEGI-NER. 1967. A rainfall simulator with a ro

tating disk. Trans. ASAE (Am. Soc. AgricEng.) 10: 74-79.

MUNN, J. R., Jr. and G. L. HUNTINGTON.1976. A portable rainfall simulator for ero

CANADIAN AGRICULTURAL ENGINEERING, VOL. 25, NO. 2, WINTER 1983

dibility and infiltration measurements onrugged terrain. Soil Sci. Soc. Am. J. 40:622-624.

MUTCHLER, C. K. and W. C. MOLDEN-HAUER. 1963. Applicator for laboratoryrainfall simulator. Trans. ASAE (Am. SocAgric. Eng.) 6: 220-222.

MUTCHLER, C. K. and L. F. HERMS-MEIER. 1965. A review of rainfall simu

lators. Trans. ASAE (Am. Soc Agric.Eng.) 8: 67-68.

NASSIF, S. H. and E. H. WILSON. 1975.The influence of slope and rain intensity onrunoff and infiltration. Hydrol. Sci. Bull.XX: 4-12.

PARSONS, D. A. 1943. Discussions on theapplication and measurement of artificialrainfall as type FA and F infiltrometer. Am.Soc. Geophys. Un. Trans. 29: 485-487.

RAWITZ, E., M. MARGOLIN, and D. HILLEL. 1972. An improved variable-intensitysprinkling infiltrometer. Soil Sci. Soc. Am.Proc. 36: 533-535.

ROMKENS, M. J. M., L. F. GLENN, D. W.NELSON, and C. B. ROTH. 1975. A laboratory rainfall similator for infiltration andsoil detachment studies. Soil Sci. Soc Am.

Proc. 39:158-160.

SHRINER, D. S., G. H. ABNER, and L. K.MANN. 1977. Rainfall simulation for en

vironmental application. Oak Ridge National Laboratory, Environ. Sci. Div. Publ.No. 1067.

SLONEKER, L. L. and W. C. MOLDEN-HAUER. 1974. Effect of varying the on offtime of rainfall similator nozzles on surface

sealing and intake rate. Soil Sci. Soc. Am.Proc 38: 157-159.

SWANSON, N. P. 1965. Rotating boom rainfall simulator. Trans. ASAE (Am. SocAgric Eng.) 8: 71-72.

YOUNG, R. A. 1979. Interpretation of rainfallsimulation data. Proc. Rainfall Simulation

Workshop Tuscon, Arizona. U. S. Dep.Agric, Sci. Educ. Admin., Agric. ReviewManual ARM-W-10. pp. 108-112.

WALKER, P. H., J. HUTKA, A. J. MOSS,and P. I. A. KINNELL. 1977. Use of ver

satile experimental system for soil erosionstudy. Soil Sci. Soc Am. J. 41: 610-612.

187