| Accurately monitoring the rill geometry and rill fow characteristics is important to improve soil erosion prediction accuracy and the performance of erosion prevention practices.Quantitative researches on dominated rill erosion processes at different rill development stages provide theoretical basis for erosion predictions on hillslopes.Based on the deficiencies of current studies on rill erosion and its monitoring methodology,simulated rainfall and upslope inflow experiments using 3D laser scanning and photogrammetry techniques were conducted to establish an automatic rill erosion monitoring system and data post processing method.Rill and rill network development processes were studies.Headcut advance,bed incision,sidewall expansion and their key processes and impacting factors were quantified.Rill erosion prediction equations based on different dominated rill erosion processes were fitted.Main results were as follows:(1)Rill geometry and rill flow chracteristics automatic monitoring system and data post processing methods were established based on photogrammetry.Rill width and rill flow width could be measured directly from perpendicularly shot photographs after proportional scale calibration.Based on interpolation principal,dynamic variations of rill depth and rill flow depth could be measured.Rill depths and rill flow depths obtained by manual measurements with a steel ruler were 3.3%-5.1% and 91.0%-178.5% higher than those obtained by photogrammetry.Technology was developed to capture 5-cm-spaced cross-sections along a soil flume at 3 s time intervals.Two off-the-shelf ordinary digital cameras were positioned 3 m above the soil bed.They were controlled by a program to trigger simultaneously and download images to the laptop.Methods depending on color difference of images and elevation difference of DEMs were applied to detect discontinuities between rill wall and bed.Rill widths were calculated by differentiating the coordinates of discontinuities.Volumetric method was used to calculate flow velocity with measurements of flow depths using ultrasonic wave generators.Sediment concentration was determined by manual sampling.The results showed that different rill width calculation methods exhibited comparable outcomes and achieved satisfactory accuracy.Sediment discharge showed a significant positive linear correlation with rill widening rate while exhibited a 5-25 s time lag compared to rill widening rate peaking.Total sediment discharge calculated by photogrammetry was 3.1% lower than that calculated by manual sampling.Flow velocity decreased with time progress and showed a significant negative power correlation with rill width.The advantages of the current methodology include high spatial and temporal monitoring resolution,100% automated monitoring fashion,comprehensive data post-processing,and the potential to be generalized to large scale river/reservoir bank failure monitoring.(2)Rill and rill network development processes on loessial hillslope were studied.Rill networks evolved in a converging way and reached maturity at the 60-80 min rainfall duration.Main rill length and rill width,depth,and degree of contour line departure increased with increased rains,while rill width/depth ratio showed the opposite trend.Secondary rill length and rill density increased at the 0-40 min rainfall duration,and then both decreased at the 40-80 min rainfall duration.The maximum values of rill erosion rate and hillslope erosion rate occurred at the bed incision dominated stage,while the minimum values of them occurred at the headcut advancing dominated stage.Hydrodynamic parameters of overland flow and rill flow showed different variation trends with the increase of rainfall duration.Shear stress,stream power and unit stream power were increased with fluctuations of unit slope length under two rainfall intensities.It showed strong positive linear correlation between rill erosion per unit width and three hydrodynamic parameters.The critical values of shear stress,stream power and unit stream power reached the maximum values at initial phase of rill development.Scour effect of lateral interfluve flow and meander cutoffs of rill flow were two sub-processes of rill piracy.Rill length and density decreased due to rill piracy specific in merging of secondary rills into main rills.Plow pan and secondary headcuts played key roles in main rill bed incision and sidewall expansion processes,while both had little impact on secondary rills.(3)Headcut advancing process was quantified.Sediment delivery and headcut advancing rate increased while initial headcut height and secondary headcut number did not strictly increase with the increase of inflow rate and slope gradient.Linear equation which included unit flow rate and slope gradient was used to predict the rill length time series.Soil loss on the hillslope dominated by rill head advance was determined by headcut advancing rate,headcut height and the number of secondary headcuts developed below the initial headcut.Soil loss increased with the increase of headcut advancing rate or headcut height in a power function while showed a linear correlation with the number of secondary headcuts.Soil loss can be modeled with a non-linear regression equation with a determination coefficient of 0.932.(4)Bed incison process was quantified.Sediment delivery,rill bed incision rate and average rill depth increased with inflow rate and bed slope.In a well-developed rill channel,rill bed incision could be divided into three phases: pre-headcut formation(dominated by rill flow shear stress),headcut incision(dominated by headcut advancing)and post-headcut incision(dominated by rill flow shear stress).Headcut incision phase,which only accounted for <15% of total experimental time,produced >65% of rill bed sediment.Initial headcut advancing rate could be predicted by a non-linear function based upon soil characteristics,rill flow shear stress and headcut height.Sediment delivery showed a linear function with the product of inflow rate and squared bed slope.(5)Sidewall expansion process was quantified.Sediment delivery and rill width increased with the increase of inflow rate,bed slope and the decrease of initial rill width.Toe scour,crack development,sidewall failure and block detachment and transport,in sequence,were the four main processes of rill widening.Time lag occurred between sediment peak and soil block failure.Basal scour arc length,tension crack length and width decreased with initial rill width and increased with time,flow discharge and bed slope.Basal scour arcs were divided into three patterns according to different shapes in comparison to the failure arcs.Exponential equations were used to predict the rill width time series.Sediment delivery equations based on the disaggregation of concentrated flow entrainment and mass failure were also fitted.(6)Rill erosion prediction models based on different dominated rill development processes were established.The headcut advancing,interaction of headcut advancing * sidewall expansion and interaction of headcut advancing * bed incision * sidewall expansion showed significant effects on unit rill erosion amount.As far as a single rill on loessial hillslope,percentage of headcut advancing on rill erosion was 45.9%-87.2%,which exhibited the highest,and followed by bed incision and sidewall expansion.Rill erosion prediction equations based on headcut advancing,bed incison and sidewall expansion were fitted and showed satisfactory validation accuracy.Rill headcut advancing rate was the best indicator for rill erosion prediction. |
PDF Full Text Request