细沟侵蚀阶段玉米季坡耕地氮素流失特征

研究紫色土区坡耕地玉米全生育期细沟侵蚀阶段水土及氮素流失规律,以期为研究区氮素流失有效防控提供科学依据。采用人工模拟降雨与野外径流小区相结合的方法,开展降雨强度为1.5mm/min条件下玉米全生育期细沟侵蚀阶段地表径流、壤中流和侵蚀泥沙中氮素流失特征的研究。结果表明:细沟侵蚀阶段,玉米各生育期地表径流量、壤中流量和侵蚀产沙量总体表现为随降雨时间延长呈先增加后平稳的变化趋势。地表径流中总氮、可溶性总氮、硝态氮和侵蚀泥沙中总氮流失量总体呈现先增加后平稳的趋势,而地表径流中铵态氮流失量变化趋势在降雨前期呈现波动性变化,降雨后期逐渐平稳。壤中流中总氮、可溶性总氮、硝态氮、铵态氮流失量则随着降雨时间延长呈现平稳的变化趋势。细沟侵蚀阶段地表径流中氮素流失总量在玉米苗期最大,为628.77mg/m2;壤中流中氮素流失总量在拔节期和抽雄期最大;侵蚀泥沙中氮素流失总量在苗期最大,为144.95mg/m2。壤中流为氮素流失主要途径,硝态氮为氮素流失主要形态。     

Rill formation and evolution caused by upslope inflow and sediment deposition on freshly tilled loose surfaces

Rill erosion easily occurs on tilled surfaces because the soil shear resistance is less than the runoff shear stress. However, rill erosion formation and evolution on tilled surfaces under upslope inflow conditions remain unclear. The objective of this study was to investigate the rill formation process and rill erosion amount on tilled surfaces via flow scouring under different inflow rates (4, 6, and 8 L min⁻¹) at a 15° slope. Close-range photogrammetric technology was applied to measure the rill morphology during the experiments. The results suggested that the rill formation process due to inflow could be divided into two distinct stages, namely, before and after runoff reached the downslope end, i.e., stages I and II, respectively. At stage I, runoff washed soil particles along the downslope direction. In this process, due to limited transport capacity of runoff, washed soil particles were deposited at the runoff head and formed a soil mound, which blocked the flow path, after which the runoff direction changed within the 6.7°− 50.6° range with an average moving distance of 0.62 m. As a result, a curved rill and a series of soil mounds were left on the surface. The inflow rate affected rill morphology by influencing the runoff direction change angle and runoff moving distance between the mounds on the surface. Herein, the rill formation process is referred to as downslope-trending erosion by inflow (DTEI). At stage II, runoff reached the downslope end, and a rill channel was formed throughout the slope. Thereafter, DTEI was largely reduced, and headward erosion was strengthened. As a result, the rill morphology quickly changed, such as the rill depth and rill width, which gradually increased with ongoing headward erosion. During DTEI, the rill paths were curved due to sediment deposition, and the sediment deposition conditions varied under the different inflow rates. Therefore, the rill curvature (RC) differed (1.039 ± 0.014). The RC decreased with headward erosion progression at stage II. The total sediment yield (TSY) increased with increasing inflow rate. Under inflow rates ranging from 4 to 6 L min⁻¹ and 6–8 L min⁻¹, the TSY increased 2.1–2.4 times. Consequently, DTEI on tilled surfaces significantly affects the initial rill morphology and evolution at the later stage. Hence, on slopes, its role should be considered in rill erosion assessment.     

Long-term erosional responses after fire in the Central Spanish Pyrenees: 1. Water and sediment yield

This paper reports the results of a study on how fire effects water and sediment losses in the Central Spanish Pyrenees where land abandonment results in an increase of the scrubland and fire-affected surfaces. In 1991, two plots (control and burnt) were installed to collect runoff, suspended sediments and solutes. One of the plots was burnt (burnt-1) and the other was covered by shrubs and herbs (control). During 1993, another burnt plot was installed (burnt-2). Data was collected from the three plots until 1999 and is used as the basis for assessing the soil erosion changes after the fire. Runoff, solute and suspended sediment concentration and erosion rates were slightly greater during the first 2 years following fire on burnt-1 compared to the control plot. The average runoff coefficients during the 8-year study period were 4.2% and 7%, the total sediment concentrations were 189 and 181 mg l⁻¹, and the erosion rates were 89 and 143 kg ha⁻¹ year⁻¹ for the control and burnt-1 plots, respectively. In contrast, burnt-2 plot responded differently than burnt-1 plot to fire, although these two plots were designed to be replicated treatments. The average runoff coefficient for burnt-2 plot was two times greater than that for the control plot. Likewise, total sediment concentration was 649 mg l⁻¹ and erosion rate was 566 kg ha⁻¹ year⁻¹, which are 3.6 and 4.5 times greater than the control plot. Differences between burnt and control plots were analysed at hourly, daily, monthly and yearly temporal scale. Both burned plots began to recover after the fire. However, burnt-1 plot had a recovery time of 2 years compared to burnt-2 plot; while 7 years after the fire, the runoff sediment concentration and erosion rates from burnt-2 were higher than those for the control plot. Water and sediment losses from the two burnt plots were markedly different.     

Revegetation techniques on clay soil slopes in northern Italy

Several standard techniques currently used in the reclamation and revegetation of former quarry land in Italy's northern Apennine region were compared. The trial site, a 40-degree slope with clay soil, high pH and a little organic matter, was arranged in a randomized block design. The techniques assayed were hydro-seeding+tillage, hydro-seeding on sod, straw mulch+tar, jute mats+tillage+hydro-seeding, jute mats+tillage and a control. Each plot was enriched with equal amounts of chopped straw (120 g m⁻²), worm compost (140 g m⁻²) and the fertilizer Biosol^® (140 g m⁻²); the binder Terravest^® (36 g m⁻²) was also used in the three hydro-seeded plots. Ground cover, species, runoff and erosion were compared in each treatment. The overall 3-year data indicate significant differences in the treatments tested. There is a clear distinction between the treatments that promoted good grass cover growth and checked erosion (straw and tar, jute on tilled soil, and hydro-seeding on jute and tilled soil) on the one hand and the hydro-seeded treatments (on sod and on tilled soil), which showed no differences from control, on the other. © 1997 John Wiley & Sons, Ltd.     

Losses of nitrate-nitrogen in water draining from under autumn-sown crops established by direct drilling or mouldboard ploughing

The leaching of nitrate-N under autumn-sown arable crops was measured using hydro-logically isolated plots, about 0.24 ha in area, from 1984–1988. Fluxes of water and nitrate moving over the soil surface (surface runoff), at the interface between topsoil and subsoil (interflow), and in the subsoil (drainflow) were monitored in plots with mole-and-pipe drain systems (drained plots); surface runoff and interflow only were monitored in ‘undrained’ plots. Half the drained and undrained plots were direct-drilled, and on the other half seedbeds were prepared by tillage to 200 mm. Tillage increased the total leaching loss of nitrate by 21 % compared with direct drilling in drained plots. About 95% or the nitrate moving from the soil was present in the water intercepted by the subsoil drains in these plots. In undrained plots less water and nitrate were collected in total; more of the nitrate was present in interflow on ploughed plots and in surface runoff in direct-drilled land. Losses of nitrate for the whole experiment from 1978-1988 were analysed. This showed that, between the harvest of one crop and the spring application of fertilizer to the next, loss of nitrate-N from ploughed land (L_p) was approximated by L_p=22+Fkg N ha^?1, where F was the autumn fertilizer-N applied. After fertilizer was applied in spring, loss of nitrate-N depended on rainfall such that for 100 mm rainfall about 30% of the fertilizer-N was lost by leaching. About 18% more nitrate-N was lost from direct-drilled land than from ploughed land in spring, but the total loss was generally small compared to that over winter. The apparent net mineralization of organic-N was measured in 1988. In autumn and winter there was little effect of tillage treatment (26 and 31 kg N ha^?1 on direct drilled and tilled plots respectively). However, over the year 83 kg N ha^?1 were mineralized in tilled plots, and 67 kg N ha^?1 in direct-drilled plots. Five factors governing the leaching of nitrate are assessed and this identified that fertilizer nitrogen application to the seedbed of winter sown crops and the mineralization of nitrogen from the residues of the previous crop are the most significant factors for nitrogen leaching in the UK.     

A comparison of runoff and erosion rates on bare and grassed loamy sand soils

Abstract. Runoff and erosion rates on loamy sand soils were measured at the Hilton experimental site, Shropshire. Little runoff and no erosion occurred on a steep runoff plot inserted into permanent grassland. Most precipitation infiltrated into the soil and was transmitted down slope as interflow within the Ah horizon. In contrast, considerable amounts of runoff and erosion occurred on the bare arable plots. These results suggest that incorporation of strips of permanent grassland into arable systems could assist soil and water conservation. Considering the high erosion rates on bare steep plots, conversion of slopes > 12° to permanent grassland seems appropriate.     

Seeding and mulching + seeding effects on post‐fire runoff,soil erosion and species diversity in Galicia (NW Spain)

The effects of two different soil rehabilitation treatments on runoff, infiltration, erosion and species diversity were evaluated in a shrubland area in Galicia (NW Spain) after an experimental fire by means of rainfall simulations. The treatments compared were: seeding, seeding + mulching and control (untreated). Rainfall simulations were conducted 9 months after fire and the application of soil rehabilitation treatments. A rainfall rate of 67 mm h⁻¹ was applied for 30 min to each runoff plot. Seeding significantly increased plant species richness in the treated plots relative to the control plots, although it had no effect on diversity or evenness. Rehabilitation treatments did not significantly increase soil cover or affect runoff and infiltration. Soil losses were low in all cases, varying from 75·6 kg ha⁻¹ in the seeded + mulched plots to 212·1 kg ha⁻¹ in the untreated plots. However, there were no significant differences in sediment yields between treatments. The percentage of bare soil appeared to be a critical variable in controlling runoff and erosion. Copyright © 2010 John Wiley & Sons, Ltd.     

The effect of simulated runoff on the erosion of gypsiferous soils

The effect of runoff generation on the erosion of gypsiferous soils under different experimental conditions was investigated by applying simulated rainfalls of 48 and 58 mm h⁻¹ to experimental plots exhibiting representative characteristics of these soils. The 45 experiments indicate differing erosive behaviour as test conditions varied. Suspended gypsum yields ranged from 0-3.9 gm⁻² h⁻¹ the higher values occurring with the steeper slopes and dry soil conditions. The differences among the soils tested were mainly due to variations in underlying rock outcrop, whilst within each soil type, slope was the most important variable. Transport of suspended gypsum by runoff is inversely related to rock outcrop, and positively related to slope. The highest erosion rates are registered in Gypsic Calcisols, followed by Gypsic Calcisols (lithic phase), while production of suspended gypsum is not verified for Dystric Leptosols. By transporting gypsum particles down-slope, runoff is probably responsible for the dynamic evolution of gypsiferous hillslopes as well as the genetic evolution of the studied soils.     

Surface runoff and soil erosion estimation using the SWAT model in the Keleta Watershed,Ethiopia

This study evaluates surface runoff generation and soil erosion rates for a small watershed (the Keleta Watershed) in the Awash River basin of Ethiopia by using the Soil and Water Assessment Tool (SWAT) model. Calibration and validation of the model was performed on monthly basis, and it could simulate surface runoff and soil erosion to a good level of accuracy. The simulated surface runoff closely matched with observed data (derived by hydrograph separation). Surface runoff generation was generally high in parts of the watershed characterized by heavy clay soils with low infiltration capacity, agricultural land use and slope gradients of over 25 per cent. The estimated soil loss rates were also realistic compared to what can be observed in the field and results from previous studies. The long‐term average soil loss was estimated at 4·3 t ha⁻¹ y⁻¹; most of the area of the watershed (∼80 per cent) was predicted to suffer from a low or moderate erosion risk (<8 t ha⁻¹ y⁻¹), and only in ∼1·2 per cent of the watershed was soil erosion estimated to exceed 12 t ha⁻¹ y⁻¹. Expectedly, estimated soil loss was significantly correlated with measured rainfall and simulated surface runoff. Based on the estimated soil loss rates, the watershed was divided into four priority categories for conservation intervention. The study demonstrates that the SWAT model provides a useful tool for soil erosion assessment from watersheds and facilitates planning for a sustainable land management in Ethiopia. Copyright © 2010 John Wiley & Sons, Ltd.     

Erosion rates and nutrient losses affected by composted cattle manure application in vineyard soils of NE Spain

Land preparation for mechanisation in vineyards of the Anoia–Alt Penedès region, NE Spain, has required major soil movements, which has enormous environmental implications not only due to changes in the landscape morphology but also due to soil degradation. The resulting cultivated soils are very poor in organic matter and highly susceptible to erosion, which reduces the possibilities of water intake as most of the rain is lost as runoff. In order to improve soil conditions, the application of organic wastes has been generalised in the area, not only before plantation but also every 3–4 years at rates of 30–50 Mg ha^− 1 mixed in the upper 30 cm.These organic materials are important sources of nutrients (N and P) and other elements, which could reduce further fertilisation cost. However, due to the high susceptibility to sealing of these soils, erosion rates are relatively high, so a higher nutrient concentration on the soil surface increases non-point pollution sources due to runoff.The aim of this study is to analyse the influence of applied composted cattle manure on infiltration, runoff and soil losses and on nutrients transported by runoff in vineyards of the Alt Penedès–Anoia region, NE Spain. In the two plots selected for the analysis, composted cattle manure had been applied in alternate rows 1 year previous to the study. In each plot soil surface samples (0–25 cm) were taken and compared to those of plots without manure application. The study was carried out at laboratory scale using simulated rainfall. Infiltration rates were calculated from the difference between rainfall intensity and runoff rates, and the sediment and total nitrogen and phosphorus were measured for each simulation. In addition, the influence of compost was investigated in the field under natural rainfall conditions by analysing the nutrient concentration in runoff samples collected in the field (in the same plots) after seven rainfall events, which amount different total precipitation and had different erosive character.Compost application increases infiltration rates by up to 26% and also increases the time when runoff starts. Sediment concentration in runoff was lower in treated (13.4 on average mg L^− 1) than in untreated soils (ranging from 16.8 to 23.4 mg L^− 1). However, the higher nutrient concentration in soils produces a higher mobilisation of N (7–17 mg L^− 1 in untreated soils and 20–26 mg L^− 1 in treated soils) and P (6–7 mg L^− 1 in untreated soils and 13–19 mg L^− 1 in treated soils). A major part of the P mobilised was attached to soil particles (about 90% on average) and only 10% was dissolved. Under natural conditions, higher nutrient concentrations were always recorded in treated vs. untreated soils in both plots, and the total amount of N and P mobilised by runoff was higher in treated soils, although without significant differences. Nutrient concentrations in runoff depend on rainfall erosivity but the average value in treated soils was twice that in untreated soils for both plots.     

Effects of field reorganisation on the spatial variability of runoff and erosion rates in vineyards of Northeastern Spain

This study analyses the spatial variability of runoff and erosion rates in vineyards due to mechanisation works. Runoff samples were collected at three positions in two plots after 33 erosive events in three years (2001, 2003, 2004) with different rainfall patterns. Three replications were considered at each position. Soil properties were evaluated in order to analyse its relationship with runoff and erosion rates. Runoff and erosion rates were, on average, higher in the levelled plot (HD), ranging between 8·4 and 34·3 per cent, than in the non‐levelled plot (LD) ranging between 8·2 and 24·1 per cent. Mean sediment concentration in runoff ranged between 6 and 8 g L⁻¹ in the HD plot and about 4·6 g L⁻¹ in the LD plot, but with high differences within the plot. In the HD plot, runoff‐rainfall rates were significantly higher (at 95 per cent level) in the upper part of the slope and decreased along the slope, while in the LD plot, differences in runoff rates were not significant and similar to those observed in the less disturbed areas of the HD plot. The higher susceptibility to soil sealing in areas where the original topsoil was removed conditioned runoff rates. In the lower part of the HD plot runoff rates were, on average, 20 per cent lower than in the upper part of the slope. In those positions runoff rates up to 79 per cent were recorded. Organic matter content and water retention capacity at different potentials are the soil characteristics related to the differences on runoff and erosion rates in the resulting soils. Copyright © 2009 John Wiley & Sons, Ltd.     

Impact of small mammal disturbances on sediment yield from grassland and shrubland ecosystems in the Chihuahuan Desert

This study examines water and sediment movement on Summerford bajada in the Jornada Basin, N.M. Forty-five rainfall simulation experiments were conducted on 1- and 2-m² runoff plots in grassland, degraded grassland, and shrubland communities. Within the shrubland community separate experiments were conducted in shrub and intershrub environments. Regression analyses indicate that for a 30-min rainfall at approximately 130 mm h⁻¹, water yields on these environments are negatively related to the percentage of ground covered by vegetation and/or litter. In the degraded grassland and intershrub environments, sediment concentration is positively correlated with the average diameter of small mammal disturbances, suggesting that animal digging is an important factor controlling rates of erosion in these environments. Sediment concentration is not correlated with any surface property in the grassland or shrub environments. An analysis of water yields and sediment concentrations at 5-min intervals during the 30-min simulated rainfall experiments reveals that the influence of the above-mentioned factors on runoff and erosion is established during events as short as 10–15 min.     

Reducing water erosion in a gypsic soil by combined use of organic amendment and shrub revegetation

Degraded gypsic soils in the centre of Spain can be rehabilitated with organic amendment and shrub revegetation. Erosion has been measured on plots of 2×0·5 m² under simulated rainfall of 70 mm h⁻¹ and a kinetic energy of 18 J mm⁻¹ m⁻². Samples of water runoff and sediments were studied in the summer of the years 2002 and 2003. The presence of shrub Atriplex halimus (Chenonodiaceae) significantly reduces runoff from 16·9 to 6·7 ml m⁻² min⁻¹ and sediments from 0·16 to 0·02 g m⁻² min⁻¹. When sewage sludge is applied the differences among plots with and without bushes disappear. Although both treatments independently applied are efficient as erosion control measures, the combined use of revegetation and organic amendment allows a reduced dose of sewage sludge with the same effect on erosion. A low dose of sludge is desirable in view of the accumulation of toxic chemicals. Copyright © 2005 John Wiley & Sons, Ltd.     

百喜草覆盖和敷盖对中国南方红壤坡地径流和产沙的影响

Rainfall,runoff (surface runoff,interflow and groundwater runoff) and soil loss from 5 m × 15 m plots were recorded for 5 years (2001-2005) in an experiment with three treatments (cover,mulch and bare ground) on sloping red soil in southern China.Surface runoff and erosion from the Bahia grass (Paspalum notatum Flugge) cover plot (A) and mulch plot (B) during the 5 years were low,despite the occurrence of potentially erosive rains.In contrast,the bare plot (C) had both the highest surface runoff coefficient and the highest sediment yield.There were significant differences in interflow and surface runoff and no significant difference in groundwater runoff among plots.The runoff coefficients and duration of interflow and groundwater runoff were in the order plot B > plot A > plot C.Effects of Bahia grass cover were excellent,indicating that the use of Bahia grass cover can be a simple and feasible practice for soil and water conservation on sloping red soil in the region.     

Soil mineral nitrogen concentration within cycles of flood irrigation: Effect of rice stubble and fertilization management

Two field experiments using macro (3.5 × 12m) plots and ¹⁵N-labelled fertilizer on micro (155 mm internal diameter) plots were undertaken to measure the effects of rice (Oryza sativa L.) stubble management and nitrogen fertilizer strategies on N transformations within a series of intermittent flood irrigations. Nitrate concentration fell by 90% during each flooding, and analysis of ¹⁵N micro plots showed the loss was due to denitrification rather than leaching. Over 52% of the ¹⁵N was lost. Apparent loss over four irrigations from macro plots receiving 60 kg urea-N ha⁻¹ was 37 kg N ha⁻¹ while unfertilized plots lost 19 kg N ha⁻. Stubble incorporation reduced nitrate accumulation rate and increased immobilization, thereby reducing denitrification losses by 23%. Nitrate concentration in the 0–100 mm soil layer increased after the soil water content fell below field capacity during the drying portion of each cycle, but the net nitrification rate fell with increasing number of cycles.Ammonium content in the top 100mm of soil fell from 35kgNha⁻¹ to 3kgNha⁻¹ over four irrigations. This fall was ascribed to the combined effects of nitrification and immobilization. Immobilization was greatest on plots where large quantities of rice stubble had been incorporated, and over 50% of the applied ¹⁵N was retained in the soil on these plots compared with 40% (SED = 3.5%) on plots where stubble had been burnt. We conclude that the poor response of rice to fertilization at sowing is due to a combination of denitrification and immobilization of applied nitrogen during cycles of wetting and drying prior to permanent flood.     

Simulated rainfall evaluation of revegetation/mulch erosion control in the Lake Tahoe Basin—1: method assessment

Revegetation of road cuts and fills is intended to stabilize those drastically disturbed areas so that sediment is not transported to adjacent waterways. Sediment has resulted in water quality degradation, an extremely critical issue in the Lake Tahoe Basin. Many revegetation efforts in this semiarid, subalpine environment have resulted in low levels of plant cover, thus failing to meet project goals. Further, no adequate physical method of assessing project effectiveness has been developed, relative to runoff or sediment movement. This paper describes the use of a portable rainfall simulator (RS) to conduct a preliminary assessment of the effectiveness of a variety of erosion‐control treatments and treatment effects on hydrologic parameters and erosion. The particular goal of this paper is to determine whether the RS method can measure revegetation treatment effects on infiltration and erosion. The RS‐plot studies were used to determine slope, cover (mulch and vegetation) and surface roughness effects on infiltration, runoff and erosion rates at several roadcuts across the basin. A rainfall rate of ≈60 mm h⁻¹, approximating the 100‐yr, 15‐min design storm, was applied over replicated 0·64 m² plots in each treatment type and over bare‐soil plots for comparison. Simulated rainfall had a mean drop size of ≈2·1 mm and approximately 70% of ‘natural’ kinetic energy. Measured parameters included time to runoff, infiltration, runoff/infiltration rate, sediment discharge rate and average sediment concentration as well as analysis of total Kjeldahl nitrogen (TKN) and dissolved phosphorus (TDP) from filtered (0·45 μm) runoff samples. Runoff rates, sediment concentrations and yields were greater from volcanic soils as compared to that from granitic soils for nearly all cover conditions. For example, bare soil sediment yields from volcanic soils ranged from 2–12 as compared to 0·3–3 g m⁻² mm⁻¹ for granitic soils. Pine‐needle mulch cover treatments substantially reduced sediment yields from all plots. Plot microtopography or roughness and cross‐slope had no effect on sediment concentrations in runoff or sediment yield. RS measurements showed discernible differences in runoff, infiltration, and sediment yields between treatments. Runoff nutrient concentrations were not distinguishable from that in the rainwater used. Copyright © 2004 John Wiley & Sons, Ltd.     

Effect of land management on runoff and soil losses from two small watersheds in St Lucia

The influence of land use on runoff and soil loss was assessed on two small watersheds in the Eastern Caribbean island of St Lucia, under contrasting land management regimes. The data generated from these watersheds revealed that the soil losses from an intensively cultivated agricultural watershed were 20‐times higher in magnitude than that of a forested watershed both for peak rainfall event and for total duration of analysis. This was due to higher surface runoff rates and exposure of soil to direct raindrop impact within cultivated areas. Whereas the forest canopy cover in combination with higher infiltration capacities of the forested land reduced the erosive runoff from the forest watershed and thus the soil loss. Moreover, the energy intensities of large storms in excess of 40 mm were estimated and found to range between 400 MJ mm ha⁻¹ h⁻¹ and 1834 MJ mm ha⁻¹ h⁻¹. 1

1 Megajoules‐millimeters per hectare‐hour.

Soil loss from the agricultural watershed was strongly correlated (R² = 0·85) to storm energy‐intensity (EI₃₀). However, the correlation of soil loss with the EI₃₀ (R² = 0·71) was poor for the forest watershed due to the effect of canopy vegetation, which significantly reduced the energy of raindrop impact. Over the study period, cumulative soil losses were 10·0 t ha⁻¹ for the agricultural site and 0·5 t ha⁻¹ for the forest site. 2

2 Metric tons per hectare.

The largest storm observed during the study period resulted in erosion losses of 3·78 t ha⁻¹ and 0·2 t ha⁻¹ from the agricultural and forest sites respectively. The regression models were developed using the measured data for prediction of runoff and soil loss over the watersheds of St Lucia under similar conditions. This study contributed towards efficient watershed management planning and implementation of suitable water conservation measures in St Lucia. Copyright © 2005 John Wiley & Sons, Ltd.     

Using Cs‐137 fingerprinting technique to estimate sediment deposition and erosion rates from Yongkang depression in the karst region of Southwest China

Understanding the erosion and deposition rates is very important for designing soil and water conservation measures. However, existing methods of assessing the rates of soil loss present many limitations and are difficult to apply to in karst areas, and there is still very little data in this areas. Karst depressions comprise geomorphologically important sources and sinks for sediments and can provide the long‐term history records of environmental changes. But there have been few similar studies focused on its sediments in the world. In this paper, the Cs‐137 technique was employed to estimate the sediment deposition rate of karst depression to assess the surface erosion. The results indicate that the average deposition rate, deposition amount and specific deposit yield for the Yongkang catchments since 1963 were estimated to be 4·32 mm y⁻¹, 3·16 t y⁻¹and 20·53 t km⁻² y⁻¹, respectively. The results obtained were consistent with the actual monitoring data of local runoff plots, and confirm the validity of the overall approach. So it was suggested that the mean specific sediment yields of 20 t km⁻² y⁻¹ can be representative of the soil loss rates in the regions. Copyright © 2010 John Wiley & Sons, Ltd.