A probabilistic approach to river network detection in digital elevation models

River networks are often derived from digital terrain models and are affected by uncertainty and errors of the corresponding elevation data. The analysis of the spatial distribution of the errors provides information on the confidence level of the derived networks. However indications on the most probable river network as a whole are missing. This study proposes a method to indicate which is the river network maximising the sum of the probability values along the network, given a map reporting the likelihood that a cell belongs to the network itself. The method is considering the inverse of the channel probability map as pseudo-DEM from which drainage networks are derived. A reference network is used to assess the spatial match of the extracted river networks using the Euclidean distance as simple comparison parameter. The network extracted from the inverse of the channel probability map is the closest to the reference. The use of a probabilistic approach to error modelling significantly increases the values of channel probability for extracted river networks and the spatial match with a ground reference dataset.     

Assessment of gully erosion susceptibility using different DEM-derived topographic factors in the black soil region of Northeast China

As a primary sediment source, gully erosion leads to severe land degradation and poses a threat to food and ecological security. Therefore, identification of susceptible areas is critical to the prevention and control of gully erosion. This study aimed to identify areas prone to gully erosion using four machine learning methods with derived topographic attributes. Eight topographic attributes (elevation, slope aspect, slope degree, catchment area, plan curvature, profile curvature, stream power index, and topographic wetness index) were derived as feature variables controlling gully occurrence from digital elevation models with four different pixel sizes (5.0 m, 12.5 m, 20.0 m, and 30.0 m). A gully inventory map of a small agricultural catchment in Heilongjiang, China, was prepared through a combination of field surveys and satellite imagery. Each topographic attribute dataset was randomly divided into two portions of 70% and 30% for calibrating and validating four machine learning methods, namely random forest (RF), support vector machines (SVM), artificial neural network (ANN), and generalized linear models (GLM). Accuracy (ACC), area under the receiver operating characteristic curve (AUC), root mean square error (RMSE), and mean absolute error (MAE) were calculated to assess the performance of the four machine learning methods in predicting spatial distribution of gully erosion susceptibility (GES). The results suggested that the selected topographic attributes were capable of predicting GES in the study catchment area. A pixel size of 20.0 m was optimal for all four machine learning methods. The RF method described the spatial relationship between the feature variables and gully occurrence with the greatest accuracy, as it returned the highest values of ACC (0.917) and AUC (0.905) at a 20.0 m resolution. The RF was also the least sensitive to resolutions, followed by SVM (ACC = 0.781–0.891, AUC = 0.724–0.861) and ANN (ACC = 0.744–0.808, AUC = 0.649–0.847). GLM performed poorly in this study (ACC = 0.693–0.757, AUC = 0.608–0.703). Based on the spatial distribution of GES determined using the optimal method (RF + pixel size of 20.0 m), 16% of the study area has very high level susceptibility classes, whereas areas with high, moderate, and low levels of susceptibility make up approximately 24%, 30%, and 31% of the study area, respectively. Our results demonstrate that GES assessment with machine learning methods can successfully identify areas prone to gully erosion, providing reference information for future soil conservation plans and land management. In addition, pixel size (resolution) is the key consideration when preparing suitable datasets of feature variables for GES assessment.     

A comparison of LiDAR-based DEMs and USGS-sourced DEMs in terrain analysis for knowledge-based digital soil mapping

While LiDAR-based digital elevation models (DEM) are more accurate and precise than the USGS-sourced DEM that are widely used in soil mapping in the US, their high cost and other problems prohibit an easy decision of adopting them in service-oriented soil mapping conducted by a government agency like USDA-NRCS. This study compares the performances of LiDAR-based DEM and the USGS-sourced DEM in calculating slope gradient as an input for knowledge-based digital soil mapping (KBDSM), aiming to provide scientific evidence and more importantly, propose a scientific approach to evaluating the two types of DEM for KBDSM. We conducted the comparison by evaluating how closely the DEM-based slope gradient values match the field-measured values. For a small watershed in northern Vermont, US, we prepared three DEM, including a 10-m DEM interpolated from the 7.5-minute USGS topographic map, a 1-m DEM based on LiDAR points, and a 5-m DEM resampled from the 1-m DEM. When calculating slope gradient, we applied two neighborhood sizes (10 m and 30 m), two neighborhood shapes (square and circular), and three slope gradient algorithms (Evans–Young, Horn, and modified Zevenbergen–Thorne) to the three DEM. We then compared the calculated slope gradient values with the values measured by soil scientists at 159 sample locations in the study area. Statistics show that across all the tested settings, the LiDAR-based DEM perform significantly better than the USGS-sourced DEM. We conclude that LiDAR-based DEM may considerably improve the quality of inputs for KBDSM. We also find that the results from the 1-m LiDAR-based DEM and the resampled 5-m DEM do not show considerable and consistent differences.     

Centimetre-scale digital representations of terrain and impacts on depression storage and runoff

The microscale topography of a 20 × 10 m hillslope in a burned portion of Kootenay National Park, British Columbia, Canada was determined through laser scanning at a sub-centimetre scale. Digital terrain models (DTM) were developed for this hillslope at the 0.75 cm, 10 cm, 25 cm, 50 cm, 75 cm and 1 m scales. Random Roughness, Tortuosity and Mean Upslope Depression were computed for bare and vegetated surfaces created at each resolution to determine mean depression storage as a function of resolution. Results for these various roughness parameters and associated depression storage values are not found to change appreciably between the vegetated and bare surface DTMs, and moreover do not change notably with increasing grid cell resolution. Depression storage was also computed using a GIS-based method, the values of which are considered to be, at the highest resolution of 0.75 cm, the most realistic for comparative purposes. Mean depression storage using this latter approach decreases significantly with an increase in grid cell size, in contrast to the values based on Random Roughness which remain stable with increasing grid cell size. The generation of Hortonian overland flow was modelled at different scales using the GIS-derived DTMs. The values of depression storage for the various scales of DTM are demonstrated to have significant influences on overland flow generation. The scale of the DTM affects the changes over time between the proportion of water going into depression storage and that contributing to overland flow, with more water being retained as depression storage at a particular value of rainfall excess after infiltration for the smaller DTM scalings. When overland flow develops from a rising groundwater table in a simulation exercise, water depths are initially lower than the primary roughness elements on the land surface and water is relatively disconnected. As water depths increase, more integrated connections form.     

基于DEM的废弃矿山小流域地形特征分析

以北京市某废弃采石场的小流域地形为研究对象,对分辨率为0.3～7 m的栅格DEM分别进行小流域参数和地形参数的提取和分析.结果发现,利用不同分辨率DEM提取的河网形态结构有显著差异,网格尺寸7,4和3m的河网结构与2,1,0.7,0.5,0.4和0.3m的河网结构有显著差异,二者主河道位置变化明显,但流域面积、河道长度及比降、平均坡度和高程参数相差不大,而河道总长度和河网密度参数相差较大.根据分析结果确定DEM的比较适宜的分辨率为0.5m,在此基础上利用主成分分析法得出高程和坡度为主成分.对研究区设置20°,25°,30°和35°的4个不同坡度的标准径流小区,观测日降雨量、坡度、径流量和土壤侵蚀量并分析和计算它们之间关系式,计算出整个流域在不同坡度分级及不同雨量级的平均径流量和土壤侵蚀量.研究成果对于废弃矿山修复和生态重建具有参考价值.     

小流域数字地面模型(DTM)的建立及其在水土保持遥感动态监测中的应用

以陕北安塞纸坊沟流域为研究区,采用航测精密立体测法建立了全流域1975年和1987年两期20m×20m格网数字高程模型和6条5m×5m断面的数字高程模型,利用Foxpro数据库系统建立了相应的DTM数据库,基于DPP测绘软件包和ARC/INFO、APSIS地理信息系统软件平台,分别绘制了多幅两期全流域和三条典型小沟的监测专题图件。通过对产生的图件和DTM数据进行定量和定性分析,对沟谷的扩张、下切及溯源侵蚀进行了较精确的动态监测。初步建立了地面三维数据采集→DTM数据库→监测专题图件→动态监测先进而又实用的科研技术系统,为较高精度和信息化小流域动态监测,提供了比较理想的技术方法和工作基础。     

DEM空间尺度对可照时数模拟结果的影响 ——以浙江省仙居县为例

以浙江省仙居县为实验区,基于4种不同空间分辨率DEM提取相关地形因子,结合可照时数分布式模型,利用数字地形分析和空间数据叠置分析等方法,模拟1月和7月4种空间分辨率下的可照时数,并定量分析DEM尺度效应对模拟结果的影响。结果表明：（1）模拟结果的空间异质性随分辨率减小而减小,其平均值逐渐增加,且1月增幅大于7月,最大值随分辨率的变化不大,而最小值差异较大,标准差逐渐减小。（2）受海拔和地形遮蔽影响,平地和山脊处可照时数最多,海拔200-400m区间最少,400-1100m区域可照时数随海拔增加而增加。以10m分辨率结果为参照,30m、90m和900m分辨率下的差值随分辨率减小而增加,海拔<100m处差值最小,700-900m区域差值最大,1月日均多算0.7、1.4和2.9h,7月日均多算0.5、0.9和2.3h。（3）坡度0-55°范围内,可照时数随坡度增加而减少,30m、90m和900m分辨率与10m分辨率的差值随分辨率降低而增加,且最小差值均在<5°区域,最大差值在不同坡度等级,1月日均多算2.1、1.8和1.7h,而7月日均多算0.3、0.6和1.2h。（4）受太阳高度角和方位角影响,可照时数在南坡-北坡间的差异较大,东南-西南坡、东坡-西坡以及东北-西北坡之间差异较小,30m、90m和900m分辨率与10m分辨率的差值随分辨率降低而增加,在偏北坡,1月差值大于7月,最大差值在1月的北坡上,日均多算1.4、2.5和4.8h,在偏南坡上,1月差值小于7月,最大差值在7月的南坡或西南坡上,日均多算0.5、0.9和2.1h。     

The effectiveness of airborne LiDAR data in the recognition of channel-bed morphology

High-resolution topographic data have the potential to differentiate the main morphological features of a landscape. This paper analyses the capability of airborne LiDAR-derived data in the recognition of channel-bed morphology. For the purpose of this study, 0.5 m and 1 m resolution Digital Terrain Models (DTMs) were derived from the last pulse LiDAR data obtained by filtering the vegetation points. The analysis was carried out both at 1-D scale, i.e. along the longitudinal channel profile, and at 2-D scale, taking into account the whole extent of the channel bed. The 1-D approach analyzed the residuals of elevations orthogonal to the regression line drawn along the channel profile and the standard deviation of local slope. The 2-D analysis was based on two roughness indexes, consisting on the local variability of the elevation and slope of the channel bed. The study was conducted in a headwater catchment located in the Eastern Italian Alps. The results suggested a good capability of LiDAR data in the recognition of river morphology giving the potential to distinguish the riffle-pool and step-pool reaches.     

Regional Patterns of Streamwater Acidity and Catchment Afforestation in Galloway,SW Scotland

This paper investigates the regional relationships between highflow water chemistry and conifer forest cover in 95 streamsspread over an area of approximately 2000 km² in Galloway,south-west Scotland, paying particular attention to the use ofdetailed catchment area data. Catchment forest data areextracted from a 30 m resolution tree height map derived fromsatellite imagery; geology data from a digital geology map; andaltitude and slope data from a 50 m resolution Digital TerrainModel (DTM). The results show that over the entire region pHlevels are lower with increasing catchment afforestation whilstaluminium concentrations are higher. Concentrations of sulphate arealso higher with increasing afforestation, which suggests thatconifers play a primary role in increasing the acidity levels ofstreams by exacerbating aerial acid deposition in the entireregion.     

Deriving river networks and catchments at the European scale from medium resolution digital elevation data

The extraction of drainage networks and catchment boundaries from digital elevation models (DEMs) has received considerable attention in recent years and is recognized as a viable alternative to traditional surveys and the manual evaluation of topographic maps. However, most studies have covered limited areas due to the lack of detailed information and/or the lack of highly efficient algorithms. In this paper we present an application that delineates river networks and catchment boundaries across the European continent from a medium resolution (250 m) DEM. We exploit novel algorithms based on the concepts of mathematical morphology and implement a landscape stratification for drainage density.A flow direction grid is computed using an efficient algorithm for the removal of spurious pits. River networks are then derived by imposing a variable threshold for the minimum contributing area needed to form and maintain a channel. This is achieved through a landscape stratification that reflects the ability of the terrain to develop different drainage densities. It is shown that the analysis of environmental characteristics coupled with the analysis of local slope versus contributing area enables river network mapping with a spatially varying drainage density. The result has been validated by comparing the derived data to digital river and catchment data from other sources and with varying scales of observation.     

黄河上游内蒙古河段塑槽输沙需水量分析

为了解冲积性河流塑槽输沙需水量,合理调配水资源,以内蒙古河段为例,依据挟沙水流能量平衡方程,研究了冲积性河流挟沙水流的能量耗散原理及水流塑槽和输沙能量的分配模式,给出了反应冲积性河流冲淤状态的塑槽输沙需水量计算方法,由平滩流量、河道来沙量和冲淤量三者组成的关系确定塑槽输沙需水量。利用内蒙古河段1960-2013年实测资料,计算得到内蒙古河段塑槽输沙水量,汛期来沙量0.7～1.1亿t,塑造2 000～3 000 m~3/s的中水河槽、淤积水平控制30%以下,需要的汛期塑槽输沙水量为94.6～141.2亿m~3;分析了不同条件下汛期塑槽输沙水量变化规律:来沙量一定,同样的中水河槽规模,控制淤积水平越低,需要的塑槽输沙水量越大;来沙量一定,同样的淤积水平,塑造中水河槽规模越大,需要的塑槽输沙水量越大;控制同样淤积水平,塑造同样中水河槽规模,汛期来沙量越大,需要的塑槽输沙水量越大。内蒙古河段塑槽输沙需水量计算结果和结果变化规律可为黄河上游内蒙古河段水资源配置提供参考。     

A mid–late Holocene flood record from the alluvial reach of the Mahi River, Western India

Flooding of rivers in India is linked with the peak monsoons. Investigating the linkage between monsoonal patterns and flood history of various rivers is therefore of fundamental importance in the Indian context. In the present study, the slackwater deposits in the alluvial reaches of the Mahi river basin, western India have been documented. These occur in the ravines incised during the early Holocene on an alluvial surface comprising sediments of Late Pleistocene age. The slackwater deposits occur at elevations up to 20 m from the present river level and extend to about 500 m inland. The carbonate rich sediments forming the ravine cliffs have provided bank stability and the dissections in the ravines have helped in the accumulation of slackwater deposits due to backflooding of the floodwater from the main channel. Recent gullies have incised the sediments and exposed deposits related to major flood events. The best exposures of slackwater deposits have been observed at Dodka. The sediment succession of the slackwater deposits is dominated by bedsets and laminasets of silt and sand separated by colluvial sediments. Four events of flood deposition occurred during the mid to late Holocene. Two units of slackwater deposits, SWD 2 and 4 have been dated by IRSL at 4.6 ± 1 ka and 1.7 ± 0.5 ka. The stratigraphy of these deposits indicates that the first two slackwater units (SWD1 and SWD2) have resulted due to flooding in a regime of intense monsoon. The other two units, however, represent extreme high magnitude floods in a period of low average precipitation.     

DEM栅格分辨率对丘陵山地区定量土壤-景观模型的影响

基于数字高程模型（Digital Elevation Model,DEM）的定量土壤-景观模型的精度依赖于DEM栅格分辨率,而DEM栅格分辨率如何影响土壤-景观模型及其预测精度目前研究较少。以西南丘陵山地区一典型汇水盆地为研究对象,以该区2.5、5、10、20和30 m DEM为基础,利用逐步线性回归方法建立起研究区不同分辨率下的定量土壤-景观模型,并应用这些模型预测研究区内土壤表层碱解氮含量的空间分布,进而比较DEM不同分辨率下土壤-景观模型及其预测精度。结果表明,随着DEM栅格分辨率的降低,比汇水面积、地形湿度指数的均值逐渐增加;平均坡度逐渐降低;曲率变化的范围逐渐减小。地形指数的这一变化规律对土壤-景观模型及其预测结果产生显著影响,模型的校正决定系数、平均绝对误差和均方根误差都以5 m栅格分辨率为转折点,分辨率低于5 m,模型的校正决定系数显著减小,平均绝对误差和均方根误差显著增加。     

基于地形因子的土壤有机碳最优估算模型

基于数字地面模型(Digital Terrain Model,DTM),同时考虑因子组合和分辨率构建土壤有机碳(SOC)最优估算模型。在7 100 km2范围内,选取了71个分辨率和22个地形因子中不多于5个因子的所有可能组合,构造了2 514 820个模型。采样点随机分为两组,6 362个训练样点构造数据挖掘模型,其他2 208个为验证样点。根据模型相关系数r值大小从中选取了不同个数因子组合以及相应分辨率的最优模型,并根据这些模型生成对应的土壤有机碳图。结果表明:单个地形因子模型和栅格大小之间的关系表现出多样化,并不是分辨率越高模型结果越好。单因子模型r值的大小并不能决定其在因子组合模型中的重要性。不同的因子及其组合有其特定的最适分辨率,最佳分辨率范围约为60~150 m。综合数据的存储空间和计算量、模型复杂度、预测精度以及空间表达能力,该地区最优模型由相对坡位、高程、归一化高程及多尺度山谷平坦指数等4个变量组成,对应分辨率为121.6 m。同时与多种克里格空间插值方法生成的土壤有机碳空间分布图进行了对比分析,发现无论几个变量的组合,其空间预测能力均较克里格空间插值方法更能表达SOC的空间变化,预测精度也较高。     

Relationship between suspended sediment load, channel geometry and land area increment in the Yellow River Delta

The annual river discharges and suspended sediment loads into the Yellow River Delta show a declining tendency with some distinct fluctuations over the last 50 years. The decrease of river discharge and suspended sediment load and the change in the river channel must influence the evolution of the Yellow River Delta. During this period several new river mouths formed via channel switch at the river delta, and the old watercourse was gradually abandoned. Recently, in years with very low annual suspended load, erosion of land area has been recorded for the delta. The aims of this work are (1) to determine what is the critical suspended sediment load needed in order to maintain the land balance of the Yellow River Delta for two periods before and after 1976 (the last time the channel shifted), and (2) to examine the variation in the channel geometry and gradient in response to changes in suspended sediment load and delta area at the river mouth. In order to estimate these critical values, we used statistical method to analyze the relationships between land area increment, and suspended sediment load and channel geometry. In order to examine the variation in the channel geometry, the channel cross-sections of the Q1 and Q6 were compared. The results show that to maintain the land area balance between 1953 and 1973, when the river mouth was the Diaokouhe, the critical annual suspended sediment load entering the delta was 4.21  10⁸ tonnes/a. After the main channel switched to Qingshuigou in 1976, the critical value to maintain the Qingshuigou mouth between 1976 and 1997 was 1.51  10⁸ tonnes/a. To maintain the land area balance for the entire Yellow River Delta between 1976 and 1997 the critical suspended sediment load was 3.18  10⁸ tonnes/a. The annual mean channel thalweg elevation and channel gradient at the river delta increase with increasing land area increment at the Qingshuigou mouth. The critical channel gradient at the channel reach between cross sections Q1 and Q6 is 0.000095. The channel has narrowed during the time period from 1976 to 1997. Also, lateral channel migration has decreased remarkably, resulting in enhanced hydraulic efficiency of the deltaic channel and artificial levees. This channel geometry evolution was influenced by river adjustment and human activities. These results are of importance for the management of the lower Yellow River channel and the delta. Future water diversion or river damming should consider the balance between suspended sediment delivery and delta growth.     

Sedimentation-induced eutrophication in large river floodplains – An obstacle to restoration?

Conservation and restoration of floodplain ecosystems and their specific functions often suffer from a deficient knowledge on basic processes and dynamics such as sedimentation and related nutrient inputs. In this study, we present the results of a joint analysis of site conditions and vegetation of alluvial meadows along a gradient of increasing distance to the main channel of the river Rhine. In particular we scrutinized whether the flooding frequency mediated by the ground elevation or the distance to the river channel determines the degree of eutrophication by river water and sediments.We found a remarkably strong transversal gradient in the functional floodplain. Extremely high concentrations of CAL-soluble phosphorous (more than 150 mg kg⁻¹) were measured in close proximity to the main channel. These sites were characterized by low vascular plant species richness and a significantly higher proportion of species with high nutrient demands. At a distance of 300 m from the main channel P concentrations in the soil dropped to a quarter of this figure, whereas the average species richness per plot strongly increased. Contrary to our expectations, the influence of the flooding duration was of minor importance for the nutrient status. We can specify that strong eutrophication effects induced by sediment deposits are confined to the close proximity to the main channel within a distance of 200–300 m. At more distant sites, target vegetation of flood meadows was obviously not affected by recent nutrient input in the course of flooding. Concerns by conservationists that the restoration of active floodplains by the removal of dykes will generally lead to unwanted eutrophication by river water and sediments require revision and thorough specification.     

Holocene alluvial morphopedosedimentary record and environmental changes in the Bardenas Reales Natural Park (NE Spain)

A Holocene alluvial morphostratigraphy, based on geomorphological mapping, sedimentological analysis, soil development and radiocarbon chronology is presented for the semiarid Blanca basin in the Bardenas Reales Natural Park (NE, Spain). Four morphopedosedimentary alluvial units (MU1, MU2, MU3 and MU4) were differentiated, which may correspond to two palaeogeographical stages. The first one includes the morphopedosedimentary unit MU1 and is the most extensive in the area. It is made up of a complex aggradation (fill up) sequence comprising a braided channel system, a sand–mud flat (9200–7900 cal BP) and a playa-lake environment related to a nearly closed depression. An intensive fluvial entrenchment period is recognized, leading the opening of the Blanca basin, before the second stage that includes the units MU2, MU3 and MU4. This set of units forms a sequence of cut-and-fill and nested fill terraces recording the alluvial activity in the Blanca basin during the last 6 ka. The MU2 unit, 5900–3300 cal BP in age, represents a period of high geomorphic instability with rapid alluvial sedimentation/channel incision oscillating changes. On the other hand, the MU3 unit has an age of 1050–500 cal BP and appears as a thin extensive alluvial cover overlying previous units. Finally, the MU4 unit comprises superposition of numerous flood depositional sequences, centimetric in scale, dated from 450–10 cal BP. The established soil chronosequence supports the numerical chronology of the four units. These accumulation periods (MU1, MU2, MU3, MU4) are associated with dry phases and/or high hydrological variability (floods and droughts) at a regional scale. Changes on the North Atlantic atmospheric circulation mode over Southern Europe may provide the key to explain the nature of alluvial activity during Holocene in Bardenas Reales.     

Rangeland hillslope lengths: A case study at the Walnut Gulch Experimental Watershed,southeastern Arizona

Rangeland hillslopes provide much of the sediment supplied to channel systems and their lengths exert a fundamental constraint on hillslope diffusive processes. However, information regarding lengths of rangeland hillslopes, and how best to estimate them, is limited. In this study, three groups of watersheds (10 in total) were selected from the Walnut Gulch Experimental Watershed according to their geology, soil and vegetation characteristics. Group 1 watersheds were at lower elevations dominated by shrubs, Group 3 were at high elevations dominated by grass, and Group 2 were mixed shrub and grass. Their hillslope lengths were calculated from 1 m-resolution DEMs using three methods: a flow routing algorithm, slope-area relationships, and inverted relationship with drainage density. Parameters that characterize the current watersheds, including Hack's exponent and coefficient, watershed shape coefficient, channel concavity and steepness, and surface roughness, were quantified and related to hillslope lengths. Results shows: (1) estimated hillslope lengths were different for the three methods and between the three groups of watersheds; (2) hillslope lengths that measured from the flow routing algorithm for the ten selected watersheds primarily ranged from 30 to 100 m, with a median value of 63.0 m, which was 20%–50% greater than those derived from slope-area plots or drainage densities; (3) hillslope lengths estimated from the flow routing method were greater in Group 3 watersheds than in Group 2 and then in Group 1 watersheds. We attributed these differences in hillslope lengths to the historic epeirogenic pulses, watershed and drainage network morphology, and differences in vegetation characteristics; (4) measured hillslope lengths from the flow routing algorithm were best correlated with hillslope relief, then surface roughness, channel steepness and concavity. These results would benefit the applications of hydrological and erosion models in rangelands.     

Stability of an upland gravel-bed stream, Swinhope Burn, Northern England

Some upland gravel-bed streams often undergo frequent channel change in response to floods and changes in sediment supply. However, in others, where local conditions permit, long-term channel planform stability can be observed. This paper uses evidence from Swinhope Burn in upper Weardale, Northern England, to demonstrate relative channel stability over a 180-year period.Channel planform change is identified using seven historical maps and air photographs dating from 1815 to 1991. The 1.4-km study reach has retained a stable meandering pattern over a period of 180 years, with a temporary but dramatic change to a straight, low sinuosity, partly divided channel, at some point between 1815 and 1856 (identifiable on the 1844 Tithe Map). Channel planform stability, observed over the historic period, is related in part to the low channel gradient upstream of a cross-valley moraine situated at the lower end of the study reach. This moraine has led to partial closure of the valley system and the development of a small upstream floodplain or ‘sedimentation zone’ (average width of 150 m). The resulting low channel gradient (0.012) inhibits coarse bedload transport and encourages overbank sedimentation. Cohesive banks promote lateral channel stability and the wide floodplain reduces potential coupling between the channel and valley-side slope sediment sources.The probable cause of the observed channel planform change is the combination of a succession of four major floods in the River Wear catchment during the 1820s, and episodic inputs of sediment generated by upstream metal mining between 1823 and 1846. Large floods in the past 40 years have produced very little evidence of lasting channel change. However, locally a channel avulsion has been documented following a flood in February 1997. This represents a useful modern analogue for previous mechanisms of channel change.     

Suitability of S factor algorithms for soil loss estimation at gently sloped landscapes

The reliability and the estimation technique of the S factor limit the use of the Universal Soil Loss Equation (USLE) at a regional scale. The accuracy of the S factor principally depends on the digital elevation model (DEM) accuracy and the precision of algorithms. Few studies were conducted on the accuracy of the S factor for a gently rolling landscape at a regional scale. We select a rectangular study site of 1.3 km² in northern Germany as a case study to explore the effects of algorithm, horizontal resolution and vertical precision of DEM as well as terrain on the S factor. Horizontal resolution includes 1 m, 5 m, 10 m, 25 m, 50 m and 100 m. We use the following eight algorithms: maximum slope gradient, maximum downhill slope gradient, second-order finite difference, third-order finite difference, third-order finite difference weighted by reciprocal of squared distance, third-order finite difference weighted by reciprocal of distance, frame finite difference and simple difference. The results are as follows: (1) the algorithms affect the accuracy of the S factor at any given resolution of DEM, in particular more obvious for the lower DEM resolutions; (2) the accuracy of the S factor decreases as horizontal resolution decreases; (3) the most suitable algorithm is maximum downhill slope gradient for 1 m, and maximum slope gradient for the 5 m to 100 m resolution; (4) vertical precision also affects the accuracy of the S factor. However, vertical precision almost does not change the order of rank of the S factor with eight different algorithms at any given resolution of DEM; (5) the more complex the terrain is, the worse the accuracy of the S factor is represented. Among all the affective factors, horizontal resolution is the most important for the accuracy of the S factor. Selecting the most suitable algorithm is the best way for reducing the errors of the S factor when applying the USLE.