利用多时相Sentinel-1 SAR数据反演农田地表土壤水分

土壤水分是陆面生态系统水分和能量循环的重要变量,在农田干旱监测、作物长势监测和作物估产等应用研究中具有重要的作用。该文结合基于变化检测的Alpha近似模型,利用Sentinel-1卫星获取的多时相C波段合成孔径雷达(synthetic aperture radar,SAR)数据,实现了农田地表土壤水分的反演。该文首先利用微波辐射传输模型验证了Alpha近似模型在土壤水分反演中的合理性。研究发现,对于土壤散射占主导的区域,Alpha近似模型对辐射传输模型有较好的近似,能够有效地消除地表粗糙度和植被对雷达后向散射系数的影响。在此基础上,结合怀来研究区多时相Sentinel-1 SAR数据,利用Alpha近似模型构建了土壤水分观测方程组,通过求解方程组得到了农田地表土壤水分。地面验证结果表明,土壤水分反演的均方根误差(root mean square error,RMSE)为0.06 cm3/cm3,平均偏差为0.01 cm3/cm3,精度较好。该文研究为利用高重访周期、多时相的Sentinel-1 SAR数据获取农田地表土壤水分提供了参考。     

基于Sentinel-1A与Landsat 8数据的北黑高速沿线地表土壤水分遥感反演方法研究

土壤含水量是影响水文和气候变化的基本参数,研究土壤含水量分布,对气候变化、水资源分布、农作物估产等有着重要的现实意义和科学价值。本文以2015年6月21日的Sentinel-1A(哨兵1号)双极化合成孔径雷达影像为基础,结合同时段辅助光学影像Landsat 8,对北安—黑河高速沿线地区不同植被覆盖程度下复杂地表土壤含水量进行反演研究,探讨不同极化组合方式在不同土地利用方式下的土壤水分含量反演结果。结果表明:VH极化及VH与辅助变量NDVI(Normalized Difference Vegetation Index)组合反演精度分别为52.1%和53.6%,整体效果并不理想。VV极化(VV Polarization)图像和双极化VV/VH(VH Polarization)组合在裸露和低植被地区反演更具有优势,其精度分别为75.4%和59.5%,而在高植被覆盖度地区并不适用。VH极化反演结果中耕地土壤含水量比实际值低9.37%,VV极化在低植被区域土壤含水量比实际值低10.45%,在灌木及耕地地区VV/VH反演结果精度比单极化及其组合反演结果低,最高精度模型的反演是VV结合NDVI。VV与辅助变量NDVI结合能综合反映复杂地表环境下土壤含水量,其精度达84%,标准误差RMSE为2.07,比VV极化反演精度提高8.8%,RMSE比VV极化降低2.704。VV与辅助变量NDVI组合方式在中等植被覆盖地区土壤含水量反演更具有优势,并能够更好地发挥哨兵1号C波段合成孔径雷达在土壤水分研究中的潜力与有效性。     

Multiscaling analysis of soil roughness variability

Soil surface roughness (SSR) is a parameter highly suited for the study of soil susceptibility to wind and water erosion. The development of a methodology for quantifying SSR has typically been based on field techniques to obtain data, rather than on the indexes used for interpreting soil roughness variability. One of the most used indexes to evaluate SSR is the random roughness (RR), easily calculated from the heights obtained with a pin meter. The RR index was obtained from soil elevation measurements collected at the intersections of a 2 × 2-cm² grid in a 100 × 400-cm² plot from three different types of soil. SSR values for all soil types were obtained after passing three different tillage tools (chisel, tiller, and roller) through three types of soils at field conditions. The RR index was calculated using the standard deviation (SD) of the lines parallel to the direction of tillage. Lines were 20 mm apart.Since RR assumes vertical random roughness without correlation, the variability of SSR was assessed using structure function (SF) to complement the study. Therefore, the main objective of this analysis was to better illustrate the variability of SSR in relation to spatial distribution. The SF was highly sensitive to soil roughness variability and depended on the tillage tool treatments and soil types, thereby illustrating the origin of the soil roughness variability, either from the soil itself or from the tillage tool used. We also demonstrate that the concept of a generalised Hurst exponent derived from the SF improves our ability to differentiate among the cases.     

Spatiotemporal variations of soil surface roughness from in-situ laser scanning

Microtopography and roughness are highly dynamic properties of the soil surface and important factors governing surface runoff and erosion processes. While various remote sensing technologies were successfully applied for topography measurements at different spatial scales, there is a lack of field studies that collected systematically microtopography data over long observation periods. In this paper an approach to measure and quantify surface roughness in the field based on laser scanning technologies is presented. Between June 2004 and November 2005 97 in-situ measurements were conducted in a test site with two different sandy substrates in vegetation-free conditions. Two-dimensional high-resolution (1 mm) datasets where generated for eight micro erosion plots of 0.25 to 2.9 m² in size. Dynamics and pattern formation were quantified for surface roughness and surface height changes. Roughness patterns at different scales were analyzed by local roughness indices using sliding windows of 3 to 55 mm in size. Results show strong spatial and temporal dynamics in surface roughness as well as substrate-specific variations. Temporal roughness variations could be detected and were linked to precipitation patterns. The methods presented in this paper are considered suitable to generate high-resolution datasets on spatiotemporal and multi-scale microtopography patterns and to advance the understanding of surface processes at small scales in natural environments.     

The stone and boulder content of Swedish forest soils

Stone and boulder content has a strong influence on many soil processes and is a crucial factor to take into account when estimating element pools. In this paper we present the results of a large scale inventory of the stone and boulder content of forest soils, which employed a surface penetration method (Viro's method) in 1943 plots across Sweden. The volume of stones and boulders was 43.4%, on average, although it was 47.6% for the most common soil type on glacial till parent material. The variation was spatially correlated and there were regional patterns, which were related to bedrock composition and Quaternary development. A stoniness function was developed for the upper 30 cm of the mineral soil and this was used to translate the penetration depths into stone and boulder content. The function was based on a reference data set of 200 soil pits. A relationship was found between stone and boulder content and surface boulder frequency, but the strength of the relationship varied regionally. In some regions surface boulders may provide an indicator of stone and boulder content in the soil.     

SAR遥感技术在农业土地利用遥感调查的应用

成像雷达遥感全天候、全天时工作和穿透一些地物的特点，以及独特的成像原理，使其得以广泛的应用。该研究针对热点雷达数据ERS-2，以地处云量较多的杭州湾海涂围垦区为研究样区，根据地物目标的时域散射特性，采用分区分类策略，对预处理后的多时相雷达数据进行农业土地利用类型分类。研究结果表明全天时、全天候的雷达遥感数据能够替代多光谱遥感数据进行农业土地利用遥感调查，并显示出巨大的应用潜力。     

Assessing the impact of urban sprawl on soil resources of Nanjing city using satellite images and digital soil databases

The Yangtse delta area is one of the most rapidly developing areas in China. There are mega-cities like Shanghai and Nanjing and the surrounding urban areas of different sizes including those along the lower reach of the Yangtse river from Shanghai to Nanjing. In combination with their satellite counties and towns, they form one of the most densely distributed urban areas in China. This is a case study conducted in Nanjing city to evaluate the impact of urban sprawl on soil resources using satellite images and digital soil database maps. The extent of the developed land in the study area and the impact of development on soil resources at a scale of 1:200,000 are estimated. The soil types occupied by the urbanization process are determined by overlaying the soil map on the satellite images (Landsat TM) of the study area at different times (1984, 1995, 2000 and 2003). This study uses a geographic information system (GIS) to combine urban land use maps of different times derived from satellite images with data on soil characteristics contained in soil databases. The results document the rapid expansion of urbanization in Nanjing city, as well as the soil types occupied by the urbanization process, and their quality. The urban area has increased 43,544 ha, 2 times more than in 1984. The urban area expanded at an annual rate of 6.9%. Thirty of the total 32 soil types (soil families) within the city were utilized by the urbanization process among which Loamy typic-Fe-leachic-stagnic anthrosol ranked the highest (12,007 ha). The loss of surface land to urban use in Nanjing city has ranged from 4.8% in 1984 to 11.8% in 2003. Soils of the first class (5349 ha) and second class (20,781 ha) were 61.5% of the total occupied soil area. Results for Nanjing show that residential, commercial, and industrial development, known as “urban sprawl,” appear to follow soil resources, with the better agricultural soils being the most affected. Several soil types appear to be on the verge of being replaced by urban sprawl. Growing urbanization may threaten food security, soil diversity and sustainability. The extent and geographic distribution of soil quality and the pedodiversity for land presently under urbanization in the study area may be determined through modeling.     

基于合成孔径雷达的农作物后向散射特性及纹理信息分析——以吉林省农安县为例

及时掌握农作物类型、时空分布和结构信息,是合理调整农业结构的重要科学依据。针对光学遥感依赖于太阳辐射,在农作物生长周期内常受制于云雾的影响而无法获取到光学遥感数据的问题,本文采用全天时全天候、不受云雾等天气影响的合成孔径雷达（synthetic aperture radar,SAR）影像,探讨典型农作物的后向散射特性和纹理特征,为采用合成孔径雷达影像实现高精度农作物大面积监测提供科学依据。以吉林省农安县为例,利用12景Sentinel-1B双极化SAR影像数据,经影像预处理和统计分析,研究不同极化的农作物后向散射特性和纹理信息。结果表明：3种作物（大豆、玉米、水稻1和水稻2）同极化（VV）的SAR后向散射系数在生长周期内均高于交叉极化（VH）,农作物植株形态改变极化方式的能力为-25~-15 dB。3种作物在整个生长期内,后向散射系数呈现较大波动,各阶段后向散射特征差异明显。在生长初期,土壤对后向散射特征起主要作用,在SAR图像上表现为暗色调;随着作物生长,冠层散射叠加土壤散射作用占据主要位置,散射值随作物生长呈现逐渐增加的趋势,在SAR图像上表现为亮色调;拔节（分枝）后（7月10日后）作物的后向散射信号除冠层散射作用外,还主要受到土壤含水量及其与作物相互作用的影响,因此拔节后两种水稻后向散射系数下降幅度较大。水稻对雷达波的吸收强于玉米和大豆,整体上后向散射系数第2种水稻<第1种水稻<玉米<大豆,尤其在VH极化方式下表现更明显。对作物SAR纹理信息的研究表明纹理信息的均值、方差和协同性对于农作物的SAR识别更有效,最佳纹理信息为VH极化均值,有利于识别3种作物;VV极化方差和VV协同性有助于区分两种水稻;SAR影像识别作物的最佳时相为5月23日至7月10日。     

Effects of soil rock fragment content on the USLE-K factor estimating and its influencing factors

Rock fragments are an important component of soil, and their presence has a significant impact on soil erosion and sediment yield. In this paper, the effects of rock fragments in the topsoil profile (RFP) and rock fragments on the soil surface (RFS) on the soil erodibility factor (K) were assessed at a global scale. The spatial pattern of the relationship between stoniness and erodibility (R_S-K) and its predominant factors were explored through correlation analysis, pattern analysis, and random forest model analysis. The results were as followings: (1) The existence of RFP increased K by 2.84%. The RFS of the mountain land and desert/Gobi reduced K by 18.7%; therefore, once the RFP and RFS were taken into account in the calculation, K was 6.98% lower. (2) The predominant factors of the effect of RFS and the joint effect of RFP and RFS were elevation and slope gradient. The predominant factors of the effect of RFP were annual average precipitation and annual average temperature. (3) In assessing and mapping soil erosion in large regions, special attention should be given to areas with large rock fragment contents, a relatively high altitude, and the presence of steep slope. If rock fragments were not taken into consideration, the mapping results of soil erosion may be biased. This article made the calculation of K more complete and accurate, thereby improving the accuracy of regional soil erosion estimation. This research was of significance for the investigation of global hydrological effects and simulation of the global soil carbon budget.     

沙区旱垄作对油菜生长环境的影响

垄作是沙区旱作农田常用的集水防风耕作技术。本文通过野外试验观测和土壤样品分析,对不同结构的垄作与平作下油菜生长环境进行了研究。结果表明,垄作相对于平作增加了生长季内土壤耕作层水分含量,降低了其波动幅度,改变了土壤水分在垂直剖面上的分布,使湿润锋位于耕作层,有利于油菜对土壤水分的吸收,对缓解干旱对作物生长的胁迫有积极作用。垄作下土壤易蚀性颗粒含量降低,地表粗糙度和垂直风速梯度增大,有效降低了土壤可蚀性和近地表风速,对防治土壤风蚀和保护作物幼苗有重要作用。不同结构的垄作比较,垄高15 cm、垄沟比1/12和垄高25 cm、垄沟比1/24的两种垄作在油菜生长季内集水效果较优,而垄高25 cm、垄沟比1/6的垄作在农田休闲期内防治风蚀效果较好。故结构合理的垄作是沙区旱作农田微观土地利用结构调整的有效措施。     

Quantifying surface and subsurface cast production by earthworms under controlled laboratory conditions

Earthworm casts, formed when organic substrates and soil minerals pass through the digestive tract, may protect soil organic matter from biological degradation if they persist in the soil. Yet, the stability of casts is affected by their location in the soil profile because surface casts are exposed to more disruptive forces (wetting-drying, freezing-thawing) than subsurface casts. It is not known whether environmental conditions affect the proportions of surface and subsurface casts produced by earthworms. This study investigated how surface and subsurface cast production by juveniles of Aporrectodea spp. and Lumbricus spp. was affected by temperature. Two juveniles of Aporrectodea spp. or Lumbricus spp. were added to plexiglass chambers filled with soil, and five replicate chambers were incubated in the dark at 5°C, 10°C, 15°C or 20°C for 1 week. Most of the casts produced by Aporrectodea spp. and Lumbricus spp. were surface casts, with <10% of casts deposited below the soil surface. The earthworms studied produced more casts, and a greater proportion of surface casts, as the soil temperature increased. These results can be used to estimate the quantity of surface and subsurface casts produced by earthworm populations under field conditions and determine the susceptibility of cast-associated organic matter to decomposition in the medium- to long-term.     

Impact of Collembola and Enchytraeidae on soil surface roughness and properties

The aim of the study was to quantify the alteration of soil surface roughness caused by the casting activity of mesofauna. Undisturbed soil monoliths with a surface area of 5000mm² were taken from the upper 4cm of the Ap-horizon in agricultural land. Two tillage systems were studied: conventional tillage (CT) and conservation tillage (CS). The sampling plots were mechanically compacted by wheeling with graded loads. Sampling occurred in spring after compaction and before seeding. The soil monoliths were defaunated before inoculating one half of the monoliths with 150 individuals of Collembola (Folsomia candida) and the other half with 100 individuals of Enchytraeidae (90% Enchytraeus minutus and 10% E. lacteus). Soil surface roughness was measured using a noncontact laser scanner: before inoculation and 6 months later. Photographs show the soil surface covered with casts. Results from laser scanning show that in most cases the surface roughness increased due to mesofaunal activity. However, roughness decreased when cracks were filled with casts. The casting activity and surface roughness changes are highest in uncompacted soil. A high degree of soil compaction significantly reduced the activity at the surface. In most cases the soil surface is more altered in CT than in CS. When the content of carbon and nitrogen were determined, both elements were accumulated in the casts but more by Collembola than Enchytraeidae. The results are discussed in the context of the hemiedaphic (F. candida) and the euedaphic (Enchytraeus spp.) mode of living. Received: 6 December 1996     

Influence of soil surface features and vegetation on runoff and erosion in the Western Sierra Madre (Durango,Northwest Mexico)

In mountainous areas, runoff and soil erosion are closely linked to soil surface features, particularly stoniness. Depending on the size of rock fragments (gravel, pebbles, stones and/or blocks) and especially the way they are integrated into the soil matrix, they may facilitate or hinder infiltration and promote soil losses. The present study examines the role of different soil surface features and their influence on runoff formation and on soil erosion in an area seriously affected by overgrazing.Based on measurements made on hillslopes for 2 years at the plot scale, the results show that grass cover, pebbles and sand content increase runoff and erosion. Inversely, slope value, tree cover percentage, structural stability and organic matter content are negatively correlated with runoff and soil losses.It is shown that the correlations can be explained by the major role played by the surface features on hydrologic behaviour of the hillslopes. Two main surface features were identified and hydraulically characterised, namely: (i) crusted surfaces with embedded gravel widespread on gentle slopes which induce high runoff and erosion rates; and (ii) stony surfaces, where free pebbles and blocks protect the top soil against raindrops and overland flow kinetic energy and lead to reduce runoff and soil losses.     

Gully and sheet erosion on subtropical mountain slopes: Their respective roles and the scale effect

As most mountains in tropical and subtropical zones, the Western Sierra Madre suffers active present erosion, which may create some constraints to the social and economic development in the area.The objectives of this study of soil degradation in the Western Sierra Madre, are to determine the respective roles of gully and sheet erosion. This research is based on field observations, field measurements of runoff and, soil losses at the plot, as well as the watershed scales as an analysis of an exhaustive census of the few gullies located in an experimental area.Measured soil losses in the Western Sierra Madre are high although there are few gullies. Most of the sediment yield seems to originate in widespread degraded areas where stoniness is the main evidence of a previous stage of erosion. Previously overgrazing and deforestation were determined as the factors of the appearance of new soil surface characteristics which explain the high runoff and sediment productions. The soil compacted by cattle trampling reduces infiltration. The decrease of the vegetation cover triggers a rise in the splash effect and thus, a soil sealing.These processes induce an increase in runoff and soil losses. The main erosion type has been described as sheet erosion: it is characterised by the removal of fine soil particles and the remains of gravels, pebbles and blocks, which constitute a pavement on the soil. Gullies generally appear on the bottom of wide valleys and depressions, where soils are thick. It is shown that sheet erosion is two orders of magnitude higher than gully erosion at the hillslope scale.Due to the spatial distribution of land use and the geological context such as the heavily degraded areas close to the main rivers, the reduction of runoff and soil loss rates within the extension of a considered area, commonly observed in hydrology, only applies up to the elementary catchments scale (1 to 50 km²). Above this area, runoff coefficient and soil loss rates increase.     

Soil surface roughness measurement—methods, applicability, and surface representation

In a laboratory rainfall simulator study soil surface roughness was measured using contact (roller chain, pin meter) and noncontact devices (laser scanner, photogrammetry). Soil surfaces with two initial roughness conditions (aggregates < 20 mm and < 63 mm) were investigated before and after 90 mm of simulated rainfall. Measured plot area was 50 by 55 cm. A comparison of soil roughness measurement techniques was undertaken with regard to data acquisition and computation efforts, resolution, precision and capability to represent soil surface features.As for the contact methods, resolution (cm range) and precision (mm range) is limited which constrains their application to calculation of simple surface parameters. Resolution and precision in the sub millimeter range could be obtained with the laser scanner, while for the photogrammetric method the measurement uncertainty was approximately 1–2 mm. Measurement time was highest (90 min) for the pin meter technique, though data were ready to use for analyses. Laser scanner measurements took 34 min. Several steps of data post-processing required 30 more minutes. Data acquisition was fastest for photogrammetry (5 min), but expert knowledge as well as special hard- and software were necessary for time-consuming photo analyses, taking about 120 min.The chain and pin meter were compared using a profile index. Profile lengths matched well for smooth surfaces; on rough surfaces, the chain meter gave shorter profile lines. Between profile index from chain measurements and the random roughness derived from pin meter data a polynomial regression could be found. Parameters of distributions of elevations as well as inclinations and depressions were used to compare the laser scanner and the photogrammetric technique. Generally the laser scanner was able to reproduce small aggregates as well as voids in between them, while DEM from stereophotos was smoothed between major aggregates. This led to skewed distributions of elevations and inclinations, as well as to a lower surface area (up to 39%), and a lower depression volume (up to 68%). Shapes of depressions were significantly different as well. The used photogrammetric technique is supposed to be successful in producing adequate DEMs for already smoothed surfaces, e.g. after rainfall events.The study revealed different fields of application and limitations of the compared devices. Using a non-adequate technique for certain situations will definitely have implications on further analyses concerning connectivity of runoff pathways, surface protection from raindrop impact or runoff detachment and sediment transport.     

Probability density function: A tool for simultaneous monitoring of pore/solid roughness and moisture content

We present a novel approach to multiscale fractal image analysis for monitoring the dynamics of the soil pore/solid network roughness due to moisture content changes. Roughness of the gray-level probability density function (PDF) of subsequent images of the drying soil was expressed in terms of its Hurst exponent (H_PDF), which correlated significantly with the soil dielectric permittivity and gravimetric water content, as well as with micro-horizon depth and the temporal progress of drying. We documented an intermittent character of the water content dynamics, correlating with oscillations of the pore/solid interface roughness. Our technique of PDF roughness analysis of digital images, exemplified here for soil moisture monitoring, can be applied for monitoring the other complex systems, for instance plant growth in greenhouse.     

基于地基激光雷达监测不同地表覆盖条件的土壤侵蚀试验研究

土壤侵蚀量是土壤退化风险评估的重要指标,地基激光雷达(TLS)为土壤侵蚀量动态监测提供新技术手段,但野外不同地表覆盖条件对监测精度的影响尚不明确。基于2019年11月和2021年7月在鹰潭红壤生态实验站获取的地表裸露、石块、稀疏象草、马尾松、马尾松+石块,马尾松+稀疏象草等不同地表覆盖条件的6个径流小区两期TLS数据,运用移动曲面拟合滤波和反距离加权插值方法估算土壤侵蚀量,结合实测侵蚀量数据评价TLS精度,并探讨不同地表覆盖条件下TLS的最小变化识别度(minLOD)。研究表明,TLS监测精度与土壤侵蚀量呈正相关关系,更适用于中等强度以上或长时间发生明显侵蚀的土壤侵蚀区监测。不同地表覆盖条件下minLOD为4～60 mm,TLS监测相对误差(RE)依次为：地表裸露(RE=–9.4%)<马尾松+石块(RE=12.6%)<石块(RE=15.5%)<稀疏象草(RE=–18.9%)<马尾松(RE=23.4%)<马尾松+稀疏象草(RE=–25.2%);地表覆盖条件不仅影响土壤侵蚀强度、地表粗糙度,也产生点云滤波及空间插值误差,进而影响TLS识别minLOD准确度和土...     

Conservation of soil moisture by different stone covers on alpine talus slopes (Lassen, California)

This study reports on the influence of stone covers with different clast sizes on the soil moisture of alpine talus slopes in Lassen (California). Fifteen four-plot sets were sampled in the dry season (July 1990) in sandy areas and in talus covered with pebbles, cobbles, or blocks between 2740 and 2775 m. Three depths (0–5, 5–10, 10–15 cm) were sampled. Field moisture content increased gradually with depth in all soil profiles, and also in plots covered by increasingly larger rocks. Surface soils in sand areas were very dry, but under rocks had water contents 6 to 14 times greater. Differences among plots decreased with depth, but subsoil samples in sand were still drier than those beneath any stone cover at similar depths. Blocks were most effective in conserving moisture; water content below them was higher than even in deep (10–15 cm) sand soils. Soil temperatures were recorded in sand and under blocks for an 11-day period. Minima were not significantly different, but average maxima were 5.6°C lower under blocks than in sand, which reached highs 4.4°C lower than the air. Differences in soil moisture among talus types are ascribed to lower evaporation losses under stones, due to both disruption of capillarity by the coarse particles, which prevented water flow to the talus surface, and to their efficient reduction of maximum temperatures. An irrigation experiment was conducted at 2110 m on a steep talus on the Chaos Crags from July 18 to Aug. 2, 1993. Four 100×75 cm plots with the same surface types than at Lassen received 22.5 mm water; moisture content was then periodically sampled. Watering produced similar water distributions among soil depths and talus types to those in Lassen. Evaporation occurred quickly in bare soils due to high air and soil temperatures. The sand surface was already dry 2 days after watering, but stone-covered plots remained moist until day 15, when soils under blocks still retained 77–97% of the water content (percent by weight) at the start of the test.     

Characterization of soil profiles in a landscape affected by long-term tillage

Soil movement by tillage redistributes soil within the profile and throughout the landscape, resulting in soil removal from convex slope positions and soil accumulation in concave slope positions. Previous investigations of the spatial variability in surface soil properties and crop yield in a glacial till landscape in west central Minnesota indicated that wheat (Triticum aestivum) yields were decreased in upper hillslope positions affected by high soil erosion loss. In the present study, soil cores were collected and characterized to indicate the effects of long-term intensive tillage on soil properties as a function of depth and tillage erosion. This study provides quantitative measures of the chemical and physical properties of soil profiles in a landscape subject to prolonged tillage erosion, and compares the properties of soil profiles in areas of differing rates of tillage erosion and an uncultivated hillslope. These comparisons emphasize the influence of soil translocation within the landscape by tillage on soil profile characteristics. Soil profiles in areas subject to soil loss by tillage erosion >20 Mg ha⁻¹ year⁻¹ were characterized by truncated profiles, a shallow depth to the C horizon (mean upper boundary 75 cm from the soil surface), a calcic subsoil and a tilled layer containing 19 g kg⁻¹ of inorganic carbon. In contrast, profiles in areas of soil accumulation by tillage >10 Mg ha⁻¹ year⁻¹ exhibited thick sola with low inorganic carbon content (mean 3 g kg⁻¹) and a large depth to the C horizon (usually >1.5 m below the soil surface). When compared to areas of soil accumulation, organic carbon, total nitrogen and Olsen-extractable phosphorus contents measured lower, whereas inorganic carbon content, pH and soil strength measured higher throughout the profile in eroded landscape positions because of the reduced soil organic matter content and the influence of calcic subsoil material. The mean surface soil organic carbon and total nitrogen contents in cultivated areas (regardless of erosion status) were less than half that measured in an uncultivated area, indicating that intensive tillage and cropping has significantly depleted the surface soil organic matter in this landscape. Prolonged intensive tillage and cropping at this site has effectively removed at least 20 cm of soil from the upper hillslope positions.