|
|
| Bildiri Özetleri | |
| Ana Sayfaya Dönüş |
| ISD Ana Sayfası |
|
G. Erpul 1, D. L. Norton 1, D. Gabriels 2 1 USDA-ARS National Soil Erosion Lab., 1196 SOIL Bldg, Purdue Univ., West Lafayette, IN, 47907-1196. 2 Department of Soil Management and Soil Care, University Ghent, Coupure Links 653, B 9000 Ghent, Belgium. ABSTRACT A series of tests to evaluate the rain-splash detachment and transport under windless and wind-driven rains were conducted in rainfall simulation facility of a wind tunnel. Windless rains and rains driven by the horizontal wind velocities of 6, 10, and 12 ms-1 were applied on a soil packed in 20 by 55 cm pan placed on both windward and leeward slopes of 7, 15, and 20%. It is found that the rain-splash detachment and transport was highly influenced by rain kinetic energy rate and the angle of raindrop incidence, both being the function of horizontal wind velocity. Mass distribution curves along the prevailing wind direction were determined, and total splash erosion based on the first moment of mass system was calculated. Unlike windless rains, wind-driven rains led to net rain-splash transport in the wind direction, and soil particles moved more than 10 m in the wind direction, especially in the windward slopes where the angle of raindrop incidence were larger than those of the leeward slopes. Results of the study depict that, in addition to its role in the detachment process, the wind plays very important role in the transport process. INTRODUCTION Soil transport by rain-splash has been most widely neglected in recent erosion models because its contribution, when compared to that of transport by surface flow, is very small (Kinnell, 1991); although soil detachment by rain-splash has been widely accepted as a main process that initiates soil erosion (Ellison, 1947). Our study hypothesizes that soil detachment and transport by rain-splash under wind-driven rains would differ from that under windless rain, and the soil transported by rain-splash could be significant process to the extent that it may not be negligible process. Soil detachment tends to increase due to the increased kinetic energy of drops (Pedersen and Hasholt, 1995) and a change in angle of raindrop incidence (Van Heerden, 1967) in wind-driven rains. Wind is also a possible factor capable of transporting detached particles by raindrop impact as well as slope and overland flow (Moss and Green, 1983). The splash process can cause net transportation in one direction under the influence of slope and wind direction (Moeyersons, 1983). De Lima (1989) demonstrates the significance of wind as mainly affecting raindrop splash anisotropy which determines the direction and extent of splash erosion. Most of the studies on modeling spatial distribution of soil particles by rain-splash are carried out with windless rains, and some mathematical models are developed regardless of wind effects (Poesen 1986). In those studies, the rain obliquity on sloping surfaces is introduced to obtain an action of some directionally inclined affect and accordingly to achieve net soil transportation by rain-splash itself. Whereas, Moeyersons (1983) and Moss and Green (1983) reported that wind direction and velocity is the prime factor determining the rain obliquity and hence the extent of particle detachment and transport, and the rain obliquity without wind and slope is only rare or random. The objective of this study is to present experimental results performed on detachment and transport by rain-splash in a wind tunnel facility of rainfall simulator and to indicate the magnitude and the extent of splash erosion under wind-driven rain. MATERIALS AND METHODS Three agricultural soils, Kemmel1 silt loam (28.9% sand, 58.6% silt, and 12.5% clay), Kemmel2 loam (37.8% sand, 44.4% silt, and 17.7% clay,) and Nukerke silt loam (22.2% sand, 60.1% silt, and 17.8% clay) from near Gent-Belgium were used in this study. The soil samples were collected from the Ap horizon and air-dried prior to the experiment. Soil was sieved into three aggregate fractions, 1.00 - 2.75 mm, 2.75 - 4.80 mm, and 4.80 - 8.00 mm, and the weighting factors assigned to each fraction were 28%, 32% and 40%, respectively. A 5-kg soil sample was then packed loosely into a 55-cm-long and 20 cm-wide pans after three fractions of aggregates were evenly mixed. The study was conducted in a wind tunnel rainfall simulator facility of Gent University, Belgium (Gabriels et. al., 1997). A continuous spray system of downward oriented nozzles delivered a median drop size of 1.00, 1.61, 1.54, and 1.54 mm for 0.0, 6.0, 10.0, and 12.0 ms-1 wind velocities, respectively at 1.5 bar operating pressure (Erpul et al., 1998). Simulated rains without wind and driven by horizontal wind velocities of 6, 10, and 12 ms-1 were applied to the soil pans under freely drained conditions for a 45 min-duration. The slope gradients were 7, 15, and 20% (4, 8.53, and 11.310) facing to windward and leeward. Intensity was measured with 5 small collectors placed next to the soil pan and with the same slope gradient and aspect as soil pan during the simulated rains. In this way, intensity measurements were made truly representative of each run without any need for correction (Sharon, 1980; Lima 1990). Upslope and downslope splash were collected by troughs of 18 ´ 120 cm and for each run, side splash was also collected by splashboards. Side splash was obtained only in windless rains and the rains driven by 6 ms-1 and evaluated in total splash regardless of the distance, and it disappeared in the rains driven by 10 and 12 ms-1. The detachment and the spatial redistribution by rain-splash inside the soil pan were also included in total splash calculations (Van Heerden, 1967). On each soil and each slope aspect with two replicates, 24 runs, in total 144 runs, were performed. The particles trapped in the troughs were washed into beakers, oven-dried, and weighted. Stepwise polynomial regression analysis was performed to determine the particle trajectories and mass distribution curves, using SAS Proc-reg procedure (SAS Institute, 1990). The polynomial equations obtained from the stepwise regression procedure were integrated with 150 predicted values at every 4-cm over the length where rain-splash was observed. The distance, over which particles traveled varied with wind velocity, slope gradient and slope aspect.  where e was splash in grams. The calculation of total splash erosion was based on the first moment of mass system, which is the center of gravity of mass distribution curves, xc  ê is the total splash erosion and is the mass of soil splashed over the distance xi. The kinetic energy of simulated rainfalls was measured by splash cup method (Ellison, 1947). The exponential relationship exists between the applied wind velocities (ms-1) and the kinetic energies of rains (Jm-2mm-1) in the tunnel (Erpul, et.al., 1999). Since intensities varied depending on rain inclination from vertical, slope gradient and aspect, the rate of expenditure of rainfall kinetic energy ERR , of which units are Jm-2 s-1 (or Wm-2 ) was calculated:  where ERA is the amount of rainfall kinetic energy expended per unit quantity of rain which has a units of Jm-2mm-1 . I is rainfall intensity . Angle of raindrop incidence were calculated from the angle of rain inclination from the vertical (a) and slope gradient (Ø) with respect to the slope aspect:  The procedure for computation of total splash erosion rate ( in ) from was based on the starting time of overland flow because the study involved in not only evaluating rain-splash transport but also investigating overland flow transport under wind-driven rains. It is likely expected that rain-splash continued until appreciable flow depth reached during 45-min simulated rains. However, no other reference point to partition the contribution of these two processes was available except overland flow start. In this case, we assume that rain-splash transport falls by a large extent through the flow depth range of 0-2 mm, and beyond it becomes negligible (Moss and Green, 1983). Finally, the power equations were developed for every soil to examine the importance of ERR and sin Ø in total splash erosion rate, E in gcm-1 s-1 , and for this SAS Proc-reg procedure (SAS Institute, 1990) was used with the log-linear (power) regression model of: log ( E ) = log ( a ) + b log ( ERR ) + c log sin Ø DISCUSSIONS The rain intensity highly changed with the angle of raindrop incidence, which was function of rain inclination and slope gradient and aspect. It was greatest when rain was vertical (windless rain), and it decreased with decreasing angle of raindrop incidence (Sharon, 1980) in the wind-driven rains. It had a value as low as 29.12 mmh-1 in leeward slopes where angle of raindrop incidence was 11.330 (Table 1). In the leeward slope of 11.310, rains driven by either 10 ms-1 or 12 ms-1 fell almost parallel to the soil surface with only angle of incidences 12.250 and 11.330, respectively. With the angle of incidence higher than 300, the change in the intensity was not as low and noticeable as those lower 300. This led to windward rain kinetic energy rates, ERR , differ from those of leeward, and Fig.1a depicts ERR as influenced by wind velocities on both windward and leeward slopes.  Rain inclination from vertical, slope gradient, and slope aspect influenced the angle of raindrop incidence. The angles of rain inclination are 52°, 66°, and 67° for the rains driven by 6.0, 10.0, and 12.0 ms-1, respectively (Gabriels et.al., 1997). It was clear that increasing horizontal wind velocities resulted in decreasing angle of raindrop incidences, and this was more pronounced in the leeward slopes. Wind-driven rains mostly had the angles of incidences > 30° in the windward slopes, whence mostly associating with those < 30° in the leeward slopes (Table 1, Fig.1b). Not only lower ERR but also smaller ø resulted in decreased total splash rate (Fig.1c) in the leeward slopes. The data obtained on ERR and ø shows a quite continuous range within the limit of the study when the data of either aspects are combined. The following power equations for the total rain-splash rate were developed by regression analysis for Nukerke silt loam, Kemmel1 silt loam and Kemmel2 loam, respectively. The analysis of variance indicated both rain kinetic energy rate and angle of raindrop incidence were significant at the 0.0001 level of significance for three soils. E = 6.8290 x 109 ERR 2.2593 (sin ø)1.7973 (r2 = 0.9177) (9) E = 8.9842 x 1010 ERR 2.5411 (sin ø)1.6280 (r2 = 0.9311) (10) E = 2.0223 x 1010 ERR 2.3872 (sin ø)1.7714 (r2 = 0.9502) (11) E = 2.2116 x 1010 ERR 2.3896 (sin ø)1.7470 (r2 = 0.9170) (12) Equation (12) was developed for the all data obtained in the present study. E is expressed in gcm-1 s-1 , and ERR in Wcm-2 . Measured and predicted values for Nukerke silt loam are shown in Fig.2.  Furthermore, center of gravity of mass distribution curves could give us a significant insight about the extent and magnitude of rain-splash under the wind-driven rains. Center of gravity increases with increasing horizontal wind velocities, depending on ERR and ø . Increase was much more discernible in the windward slopes than in the leeward slopes since rain kinetic energy rates and the angle of drop incidences were relatively higher in those slopes. For example, Fig. 3 illustrates the first moment of mass system in windward slope is notably greater than that of leeward slope at all applied wind velocities for Nukerke silt loam in the runs with slope gradient of 40.  In all cases of wind-driven rains, more rain-splash occurred in longer distances in the windward slopes due to both higher rain kinetic energy rates and higher angle of raindrop incidences compared to those of leeward slopes. In windless rain, change in the center of gravity with respect to slope aspect was insignificant. In the present study, wind determined the rain-splash trajectories, directing the lifted soil particles, and this process caused net transportation in prevailing wind direction. Theoretically, we could consider that detachment and transport of rain-splash is constant over the entire uniform slope.  Fig.4 presents this theoretical case at which rain-splash attains its maximum value at a given distance and remains the same from that point on (Moeyerson, 1983; Van Heerden, 1967). The extent and magnitude of this net migration process will vary with the rain kinetic energy rate and the angle of raindrop incidence. REFERENCES De Lima, J. L. M. P. (1989). Raindrop splash anisotropy: Slope, wind, and overland flow velocity effects. Soil Technol., 2: 71-78. De Lima, J. L. M. P. (1990). The effect of oblique rain on inclined surfaces: A nomograph for the rain-gauge correction factor. Journal of Hydrology 115: 407-412. Ellison, W. D. 1947 (7 parts). Soil erosion studies. Agric. Eng. 28: 145-146; 197-201; 245-248; 297-300; 349-351; 407-408; 447-450. Erpul, G. D. Gabriels, D. Janssens. (1998). Assessing the drop size distribution of simulated rainfall in a wind tunnel. Soil and Tillage Research 45: 455-463. Erpul, G. D. Gabriels, D. Janssens. (1999). The effect of wind on size and energy of small simulated raindrops: a wind tunnel study. Geophysics (in press). Gabriels, D., W. Cornelis, I. Pollet, T. Van Coillie, and M. Quessar. (1997). The I.C.E. wind tunnel for wind and water erosion studies. Soil Technology 10: 1-8. Gabriels, D., K. Tack, G. Erpul, W.M. Cornelis, D.L. Norton and J. Biesemans. (1997). Effect of wind-driven rain on splash detachment and transport of a silt loam soil: a short slope wind tunnel experiment. Workshop on Wind and Water Erosion, November 17-18, 1997, Ghent-Belgium. Kinnell, P. I. A. (1991). The effect of flow depth on sediment transport induced by raindrops impacting shallow flows. Trans. ASAE 34: 161-168. Moeyersons, J. (1983). Measurements of splash-saltation fluxes under oblique rain. Catena suppl. 4: 19-31. Moss, A. J. and P. Green. 1983. Movement of solids in air and water by raindrop impact. Effects of drop-size and water-depth variations. Aust. J. Soil Res. 21(3): 373-382. Pedersen, H. S. and B. Hasholt. (1995). Influence of wind speed on rainsplash erosion. Catena 24: 39-54. Poesen, J. 1986. Field measurements of splash erosion to validate a splash transport model. Z. Geomorph. N. F., Suppl. -bd 58: 81-91. SAS Institute. (1990). SAS/STAT guide for personal computers. SAS Inst., Cary, NC. Sharon, D. (1980). The distribution of hydrologically effective rainfall incident on sloping ground. Journal of Hydrology 46: 165-188. Van Heerden, W. M. (1967). An analysis of soil transportation by raindrop splash. Trans. ASAE 10: 166-169. |