EXPERIMENTAL VERIFICATION OF SEDIMENT EROSION / YIELD AND ITS MODELLING

Abstract

Extensive studies are being conducted globally to address soil erosion/yield, water pollution, and urban flooding. Though many primary factors, including rain and soil parameters, are responsible eventually, rainfall granulometry and intensity (I) of falling rain play a vital role in soil erosion. A comprehensive understanding of natural rainfall events is impossible as it varies spatially and temporally. Since the spatial and temporal distribution characteristics under natural rain are beyond control, there is always a scope for research in this area. The natural rainfall reflected a strong correlation between the various rain parameters, such as raindrop size, kinetic energy (KE), and rainfall intensity (I). Natural precipitation characterized as unpredictable, inconsistent, and inaccessible when needed; however can, be controlled to some extent in experimentation, and therefore, the principles of physical modeling form to be an alternative to studying soil erosion/yield process of in situ detachment of the sediments by adopting a simulation technique. The rainfall simulator is an advanced means for characterizing and correlating the raindrop parameters, which is essential in studying soil erosivity potential. The technique of physically-based modeling through empirical and conceptual relationships helps correlate these rain parameters. However, their estimation requires precise instrumentation, which is seldom available. The accuracy and usefulness of the simulator improve with the similarity between natural and simulated rainfalls. The considerations for simulator design depend on the information required to fulfill the research purpose. The literature study indicated considerable variations in the methodology adopted and the developed empirical relationships between various rainfall parameters in different localities. Multiple attributes of the raindrop generation mechanism under certain simulation conditions with the empirical relationships obtained for simulated and natural rainfall are discussed in the present study to shed light on further research work by the readers. The present study engaged six nozzles of different capacities and a Laser Precipitation Monitor (LPM) in obtaining the empirical relationships between different erosivity parameters. The simulator was calibrated to simulate natural rainfall conditions in the laboratory based on rain granulometry, drop size distribution, drop terminal velocity, and kinetic energy of raindrops. Different linear and non-linear regression relationships were developed and tested statistically to correlate the pressure, median volume/drop diameters (D50) of rain, the kinetic energy of raindrop per unit area per unit time (KEtime), kinetic energy expended per unit rain quantity (KEvol), with the rain intensity (I). The estimated KEtime and KEvol ranged from 10.384 to 572.273Jm−2h−1 and 0.57 to 17.51Jm−2mm−1, respectively, comparable to the natural rain at specified rain intensities. The present study also developed a generalized exponential equation to correlate D50–I and a power-law-based equation for erosivity and rainfall depth. The adequacy of the developed relationships was verified with the residual analysis and MAE, MSE, and RMSE, indicating the significance of the relationship. Developed correlations shall be helpful in the estimation of various rainfall parameters with simple measurements of the most common parameters like rain intensity and depth. The results of the present study shall enable the researchers to develop event-scale physically-based models on soil erosion/yield. The technique of physically-based modeling through empirical and conceptual relationships helps to correlate the rain parameters contributing to soil erosion that differ by locality due to changes in the microphysics of rainfall. In the present study, six pressurized nozzles (Spraying System Co.) of different capacities were engaged with Laser Precipitation Monitor (LPM) to generate various rain intensities (21.14 to 78.93 mm/hr) to register rain granulometry. A soil erosion flume of 2.50×1.25×0.56m with an adjustable longitudinal slope was used to investigate the sediment transportation induced by rainfall and runoff. Experiments were performed at flat, 5, 10, and 15% slopes of the erosion flume to record the runoff and sediment yield. Various rain characteristics were estimated by conducting 32 experiments for each slope category. The simulator produces the median drop sizes of 0.38 to 2.11mm at generated rain intensity coincides with natural rain. Various linear and non-linear empirical relationships were developed through multiple regressions to correlate sediment yield and rain intensity with the optimization of parameters. The developed relationships yielded the best fit (R2= 0.75 to 0.93; P<0.001) with observed and estimated parameters. The metrics used to test the developed regression equations accurately with low MAE, MSE, and RMSE values. The rain intensity and slope were found to be the primary contributing factors to sediment transport. The results indicated that the simulator helps understand the complex task of soil erosion with hydrologic and geomorphic processes in laboratory experimentation with sufficient accuracy for measuring sediment transport. The present study developed various relationships to correlate the rain and sediment yield parameters. The adequacy of the developed relationships was validated using data observed from three different LULC (high dense, medium dense, and fallow) conditions and slope gradients (8, 12 & 16%) in natural rain conditions. Multiple regression and ANOVA analyses obtained high retrieval accuracies for non-linear (logarithmic and power) models developed for sediment yield. The models proposed in this study will help estimate the quantum of soil erosion/yield by measuring the most commonly used parameters.