Evolution of overland flow connectivity in bare agricultural plots

Soil surface roughness not only delays overland flow generation but also strongly affects the spatial distribution and concentration of overland flow. Previous studies generally aimed at predicting the delay in overland flow generation by means of a single parameter characterizing soil roughness. However, little work has been done to find a link between soil roughness and overland flow dynamics. This is made difficult because soil roughness and hence overland flow characteristics evolve differently depending on whether diffuse or concentrated erosion dominates. The present study examined whether the concept of connectivity can be used to link roughness characteristics to overland flow dynamics. For this purpose, soil roughness of three 30‐m² tilled plots exposed to natural rainfall was monitored for two years. Soil micro‐topography was characterized by means of photogrammetry on a monthly basis. Soil roughness was characterized by the variogram, the surface stream network was characterized by network‐based indices and overland flow connectivity was characterized by Relative Surface Connection function (RSCf) functional connectivity indicator. Overland flow hydrographs were generated by means of a physically‐based overland flow model based on 1‐cm resolution digital elevation models. The development of eroded flow paths at the soil surface not only reduced the delay in overland flow generation but also resulted in a higher continuity of high flow velocity paths, an increase in erosive energy and a higher rate of increase of the overland flow hydrograph. Overland flow dynamics were found to be highly correlated to the RSCf characteristic points. By providing information regarding overland flow dynamics, the RSCf may thus serve as a quantitative link between soil roughness and overland flow generation in order to improve the overland flow hydrograph prediction. Copyright © 2016 John Wiley & Sons, Ltd.     

Spatial and temporal variations of overland flow during rainfall events and in relation to catchment conditions

Improved knowledge on overland flow (OF) generation and its dynamics (i.e. spatial and temporal variations) is essential to understand catchment hydrology, a prerequisite for better water resources and soil management. In this study, our main objective was to quantify the dynamics of OF during rainfall events and to assess its main factors of control. The research study was undertaken in an agricultural 23‐ha catchment of a communal pasture in KwaZulu‐Natal (South Africa) experiencing Mediterranean climate and with variations of soil, topography and vegetation conditions. The dynamics of OF was evaluated during three rainfall seasons (2007 to 2010) by using 1 × 1‐m² microplots (n = 15) located at five landscape positions. At each location, a microplot was equipped with an automatic tipping bucket linked to a logger to estimate the delay between the start of the rain and the start of OF [i.e. the time to runoff initiation (TRI)]. Multivariate analysis was applied to the total OF and TRI data and the information on selected environmental factors (rainfall characteristics; soil type; soil clay content, Clay; proportion of the soil surface covered by vegetation, Cov; proportion of the soil surface covered by crusting, Crust; mean slope gradient, S; soil bulk density, ρ_b; soil water tension at different depths, SWT). The average OF rate over the 3‐year study period varied 2.3‐fold across the catchment (from 15% footslope to 35% backslope), whereas the average TRI varied by a 10.6‐fold factor (between 0.6 min at bottomland and 6.4 min at footslope). TRI temporal variations correlated the most with event duration (r = 0.8) and cumulative amount of rainfall since the onset of the rainy season (r = ?0.47), whereas TRI spatial variations were controlled the most by Crust (?0.97 < r < ?0.77). Ultimately, TRI spatial variations were modelled and mapped in an attempt to model OF dynamics over the entire microcatchment. Copyright © 2012 John Wiley & Sons, Ltd.     

Terrestrial laser scanning soil surfaces: a field methodology to examine soil surface roughness and overland flow hydraulics

Conventional roughness–resistance relationships developed for pipe and open‐channel flows cannot accurately describe shallow overland flows over natural rough surfaces. This paper develops a new field methodology combining terrestrial laser scanning (TLS) and overland flow simulation to provide a high‐resolution dataset of surface roughness and overland flow hydraulics as simulated on natural bare soil surfaces. This method permits a close examination of the factors controlling flow velocity and a re‐evaluation of the relationship between surface roughness and flow resistance. The aggregate effect of flow dynamics, infiltration and depression storage on retarding the passage of water over a surface is important where runoff‐generating areas are distant from well‐defined channels. Experiments to separate these effects show that this ‘effective resistance’ is dominated by surface roughness. Eight measurements of surface roughness are found to be related to flow resistance: standard deviation of elevations, inundation ratio, pit density (measured both perpendicular and parallel to the flow direction), slope, median depth, skewness of the depth distribution and frontal area. Hillslope position is found to affect the significant roughness measures. In contrast, infiltration rate has little effect on the velocity of water fronts advancing over the soil surfaces examined here and the effect of depression storage is limited. Overland flow resistance is depth dependent where complex microtopographic structures are progressively inundated. Copyright © 2010 John Wiley & Sons, Ltd.     

The importance of surface controls on overland flow connectivity in semi‐arid environments: results from a numerical experimental approach

In semi‐arid environments, the characteristics of the land surface determine how rainfall is transformed into surface runoff and influences how this runoff moves from the hillslopes into river channels. Whether or not water reaches the river channel is determined by the hydrological connectivity. This paper uses a numerical experiment‐based approach to systematically assess the effects of slope length, gradient, flow path convergence, infiltration rates and vegetation patterns on the generation and connectivity of runoff. The experiments were performed with the Connectivity of Runoff Model, 2D version distributed, physically based, hydrological model. The experiments presented are set within a semi‐arid environment, characteristic of south‐eastern Spain, which is subject to low frequency high rainfall intensity storm events. As a result, the dominant hydrological processes are infiltration excess runoff generation and surface flow dynamics. The results from the modelling experiments demonstrate that three surface factors are important in determining the form of the discharge hydrograph: the slope length, the slope gradient and the infiltration characteristics at the hillslope‐channel connection. These factors are all related to the time required for generated runoff to reach an efficient flow channel, because once in this channel, the transmission losses significantly decrease. Because these factors are distributed across the landscape, they have a fundamental role in controlling the landscape hydrological response to storm events. Copyright © 2013 John Wiley & Sons, Ltd.     

Effect of saltating sediment on flow resistance and bed roughness in overland flow

The acceleration of saltating grains by overland flow causes momentum to be transferred from the flow to the grains, thereby increasing flow resistance and bed roughness. To assess the impact of saltating sediment on overland flow hydraulics, velocity profiles in transitional and turbulent flows on a fixed sand-covered bed were measured using hot-film anemometry. Five discharges were studied. At each discharge, three flows were measured: one free of sediment, one with a relatively low sediment load, and one with a relatively high sediment load. In these flows from 83 to 90 per cent of the sediment was travelling by saltation. As a result, in the sediment-laden flows the near-bed velocities were smaller and the velocity profiles steeper than those in the equivalent sediment-free flows. Sediment loads ranged up to 87·0 per cent of transport capacity and accounted for as much as 20·8 per cent of flow resistance (measured by the friction factor) and 89·7 per cent of bed roughness (measured by the ratio of the roughness length to median grain diameter). It is concluded that saltating sediment has a considerable impact on overland flow hydraulics, at least on fixed granular beds. Saltation is likely to have a relatively smaller effect on overland flow on natural hillslopes and agricultural fields where form and wave resistance dominate. Still, saltation is generally of greater significance in overland flow than in river flow, and for this reason its effect on overland flow hydraulics is deserving of further study. © 1998 John Wiley & Sons, Ltd.     

Changes in overland flow and infiltration after a rangeland fire in a Mediterranean scrubland

Changes in overland flow and infiltration after a wildfire (summer 1989) in a typical Mediterranean scrubland were measured during the winters of 1990, 1991, 1992 and 1995 by means of simulated rainfall. Infiltration increases gradually from 1990 (sixth months after the forest fire) to 1995 (five and a half years after the forest fire). Overland flow decreases from 45% of rainfall after the forest fire to less than 6% five and a half years later. The reduction in overland flow was greatest in the first two years after the fire because of the quick recovery of vegetation. The steady-state infiltration capacity increased every year after the fire. Runoff and infiltration changes are mainly determined by the gradual recovery of vegetation. © 1998 John Wiley & Sons, Ltd.     

A network theory approach for a better understanding of overland flow connectivity

Hydrological connectivity describes the physical coupling (linkages) of different elements within a landscape regarding (sub‐) surface flows. A firm understanding of hydrological connectivity is important for catchment management applications, for example, habitat and species protection, and for flood resistance and resilience improvement. Thinking about (geomorphological) systems as networks can lead to new insights, which has also been recognized within the scientific community, seeing the recent increase in the use of network (graph) theory within the geosciences. Network theory supports the analysis and understanding of complex systems by providing data structures for modelling objects and their linkages, and a versatile toolbox to quantitatively appraise network structure and properties. The objective of this study was to characterize and quantify overland flow connectivity dynamics on hillslopes in a humid sub‐Mediterranean environment by using a combination of high‐resolution digital‐terrain models, overland flow sensors and a network approach. Results showed that there are significant differences between overland flow connectivity on agricultural areas and semi‐natural shrubs areas. Significant positive correlations between connectivity and precipitation characteristics were found. Significant negative correlations between connectivity and soil moisture were found, most likely because of soil water repellency and/or soil surface crusting. The combination of structural networks and dynamic networks for determining potential connectivity and actual connectivity proved a powerful tool for analysing overland flow connectivity. Copyright © 2016 John Wiley & Sons, Ltd.     

Do not only connect: a model of infiltration‐excess overland flow based on simulation

The paper focusses on connectivity in the context of infiltration‐excess overland flow and its integrated response as slope‐base overland flow hydrographs. Overland flow is simulated on a sloping surface with some minor topographic expression and spatially differing infiltration rates. In each cell of a 128 × 128 grid, water from upslope is combined with incident rainfall to generate local overland flow, which is stochastically routed downslope, partitioning the flow between downslope neighbours. Simulations show the evolution of connectivity during simple storms. As a first approximation, total storm runoff is similar everywhere, discharge increasing proportionally with drainage area. Moderate differences in plan topography appear to have only a second‐order impact on hydrograph form and runoff amount. Total storm response is expressed as total runoff, runoff coefficient or total volume infiltrated; each plotted against total storm rainfall, and allowing variations in average gradient, overland flow roughness, infiltration rate and storm duration. A one‐parameter algebraic expression is proposed that fits simulation results for total runoff, has appropriate asymptotic behaviour and responds rationally to the variables tested. Slope length is seen to influence connectivity, expressed as a scale distance that increases with storm magnitude and can be explicitly incorporated into the expression to indicate runoff response to simple events as a function of storm size, storm duration, slope length and gradient. The model has also been applied to a 10‐year rainfall record, using both hourly and daily time steps, and the implications explored for coarser scale models. Initial trails incorporating erosion continuously update topography and suggest that successive storms produce an initial increase in erosion as rilling develops, while runoff totals are only slightly modified. Other factors not yet considered include the dynamics of soil crusting and vegetation growth. Copyright © 2014 John Wiley & Sons, Ltd.     

Assessing the impact of the hydraulic properties of a crusted soil on overland flow modelling at the field scale

Soil surface crusts are widely reported to favour Hortonian runoff, but are not explicitly represented in most rainfall‐runoff models. The aim of this paper is to assess the impact of soil surface crusts on infiltration and runoff modelling at two spatial scales, i.e. the local scale and the plot scale. At the local scale, two separate single ring infiltration experiments are undertaken. The first is performed on the undisturbed soil, whereas the second is done after removal of the soil surface crust. The HYDRUS 2D two‐dimensional vertical infiltration model is then used in an inverse modelling approach, first to estimate the soil hydraulic properties of the crust and the subsoil, and then the effective hydraulic properties of the soil represented as a single uniform layer. The results show that the crust hydraulic conductivity is 10 times lower than that of the subsoil, thus illustrating the limiting role the crust has on infiltration. Moving up to the plot scale, a rainfall‐runoff model coupling the Richards equation to a transfer function is used to simulate Hortonian overland flow hydrographs. The previously calculated hydraulic properties are used, and a comparison is undertaken between a single‐layer and a double‐layer representation of the crusted soil. The results of the rainfall‐runoff model show that the soil hydraulic properties calculated at the local scale give acceptable results when used to model runoff at the plot scale directly, without any numerical calibration. Also, at the plot scale, no clear improvement of the results can be seen when using a double‐layer representation of the soil in comparison with a single homogeneous layer. This is due to the hydrological characteristics of Hortonian runoff, which is triggered by a rainfall intensity exceeding the saturated hydraulic conductivity of the soil surface. Consequently, the rainfall‐runoff model is more sensitive to rainfall than to the subsoil's hydrodynamic properties. Therefore, the use of a double‐layer soil model to represent runoff on a crusted soil does not seem necessary, as the increase of precision in the soil discretization is not justified by a better performance of the model. Copyright © 2005 John Wiley & Sons, Ltd.     

Size characteristics of sediment in interrill overland flow on a semiarid hillslope,Southern Arizona

This study examines the size characteristics of sediment removed from a semiarid hillslope by interrill overland flow. Rainfall simulation experiments were conducted on a runoff plot 18 m wide and 35 m long established on a piedmont hillslope in southern Arizona. The top of the plot coincided with the hillslope divide, and its outlet was located within a shallow rill. Samples of runoff were obtained from two cross-sections located in the interrill portion of the plot upslope of the rill and from a calibrated flume through which was directed interrill overland flow reaching the bottom of the plot. Analyses of sediment contained in these samples showed that sediment in interrill flow is finer than the matrix soil. The fineness of the interrill sediment compared to the matrix soil appears to be due to the inability of interrill overland flow to transport the coarser fraction of the sediment supplied to it by raindrop detachment. This finding implies that the rate of soil erosion in interrill areas is not. as is commonly supposed, limited by the rate at which raindrops can detach sediment but by the rate at which they detach sediment of a size that the overland flow is competent to transport. The relative fineness of sediment eroded from this hillslope is consistent with other evidence for the recent evolution of shrub-covered hillslopes in southern Arizona.     

Soil erosion and overland flow in Japanese cypress plantations: spatio-temporal variations and a sampling strategy

ABSTRACT

Many studies have focused on soil erosion in unmanaged Japanese cypress plantations because the sparse understory vegetation and litter covering the forest ground enhance soil erosion. In this study, soil erosion, litter, and overland flow measurements were conducted over 14 months to identify the spatio-temporal variation and examine the optimal sample size. Fifteen traps (each 0.25 m wide) were installed in line along the bottom of a 15-m-wide slope. Soil erosion and overland flow had large spatial variations as compared to litter. The temporal coefficient of variation of soil erosion and overland flow was highest during dry seasons, while smaller during wet seasons. The random sampling analysis showed that the rate of decrease in spatio-temporal variation became moderate as the sample size increased beyond six. This result indicated that the optimal sample size was five, the total width of which was equivalent to about 8% of the monitored slope width.     

An experimental study on the evolution law of roll wave parameters in overland flow

Roll waves commonly occur in overland flow and have an important influence on the progress of soil erosion on slopes. This study aimed to explore the evolution and mechanism of roll waves on steep slopes. The potential effects of flow rate, rainfall intensity and bed roughness on the laws controlling roll wave parameters were investigated. The flow rates, rainfall intensities and bed roughness varied from 5 to 30 L/min, 0 to 150 mm/h, and 0.061 to 1.700 mm, respectively. The results indicate that roll waves polymerize significantly along the propagation path, and bed roughness and rainfall affect the generation and evolution of roll waves. The wave velocity, length and height decreased with bed roughness, whereas the wave frequency increased with increasing bed roughness under fixed flow rate and rainfall intensity conditions. Rainfall increased the wave velocity and wavelength and decreased the wave frequency. The wave velocity, height and wavelength tended to increase with an increasing flow rate. Rainfall promoted the generation of roll waves, whereas bed roughness had the opposite effect. The generation of roll waves is closely related to the Froude number (Fr) and flow resistance. In this experiment, the range of the Reynolds number for the roll waves generated in the laminar region was 142–416, and the range of the flow resistance coefficient was 0.64–4.85. The critical value of the Fr for flow instability in the laminar region was approximately 0.57. Exploring the generation and evolution law of roll waves is necessary for understanding the processes and dynamic mechanisms of slope soil erosion.     

Relating nitrogen export patterns from a mixed land use catchment in NW Spain with rainfall and streamflow

The temporal variability in nitrogen (N) transport in the Corbeira agroforestry catchment (NW Spain) was analysed from October 2004 to September 2008. Nitrate (NO₃–N) and total Kjeldahl nitrogen (TKN) loads and concentrations were determined at various timescales (annual, seasonal and event). The results revealed a strong intra‐annual and inter‐annual variability in N transport influenced by weather patterns and consequently by the hydrological regime. Mean annual export of total N in the catchment was 5.5 kg ha^?1 year^?1, with NO₃–N being the dominant form. Runoff events comprised 10% of the study period but contributed 40 and 61% of the total NO₃–N and TKN loads, respectively. The NO₃–N and TKN concentrations were higher during runoff events than under baseflow conditions, pointing to diffuse sources of N. The mobilization of TKN during runoff events was attributed to surface runoff, while NO₃–N might be related to subsurface and groundwater flow. Runoff events were characterized by high variability in N loads and concentrations. Higher variability was observed in N loads than in N concentrations, indicating that event magnitude plays an important role in N transport in this catchment; event magnitude explained approximately 96% of the NO₃–N load. However, a combination of variables related to runoff event intensity (rainfall, discharge increase and kinetic energy) explained only 66% of the TKN load. Copyright © 2014 John Wiley & Sons, Ltd.     

Study on effect of surface roughness on overland flow from different geometric surfaces through numerical simulation

Effect of variability in surface roughness on overland flow from different geometric surfaces is investigated using numerical solution of diffusion wave equation. Three geometric surfaces rectangular plane, converging and diverging plane at slopes 1 to 3% are used. Overland flow is generated by applying rainfall at constant intensity of 10 mm/h for period 30 min and 100 min. Three scenarios of spatial roughness conditions viz. roughness increasing in downstream direction, roughness decreasing in downstream direction and roughness distributed at random are considered. Effect of variability of roughness on overland flow in terms of depth, velocity of flow and discharge along the distance from upstream to downstream for different geometric surfaces are discussed in detail. Results from the study indicate that roughness distribution has significant effect on peak, time to peak and overall shape of the overland flow hydrograph. The peak occurs earlier for the scenario when roughness increases in downstream direction as compared to scenario when roughness is decreasing in downstream for all three geometric surfaces due to very low friction factor and more velocity at the top of the domain. The converging plane attains equilibrium state early as compared to rectangular and diverging plane. Different set of random values result in different time to peak and shape of hydrograph for rectangular and diverging plane. However, in case of converging plane, the shape of computed hydrographs remains almost similar for different sets of random roughness values indicating stronger influence of converging geometry than effect due to variation of roughness sequence on computed runoff hydrograph. Hierarchically, the influence of geometry on overland flow is stronger than the influence of slope and the influence of slope is stronger than the influence of roughness. Copyright © 2013 John Wiley & Sons, Ltd.     

Assessing plot-scale impacts of land use on overland flow generation in Central Panama

Land use in Panama has changed dramatically with ongoing deforestation and conversion to cropland and cattle pastures, potentially altering the soil properties that drive the hydrological processes of infiltration and overland flow. We compared plot-scale overland flow generation between hillslopes in forested and actively cattle-grazed watersheds in Central Panama. Soil physical and hydraulic properties, soil moisture and overland flow data were measured along hillslopes of each land-use type. Soil characteristics and rainfall data were input into a simple, 1-D representative model, HYDRUS-1D, to simulate overland flow that we used to make inferences about overland flow response at forest and pasture sites. Runoff ratios (overland flow/rainfall) were generally higher at the pasture site, although no overall trends were observed between rainfall characteristics and runoff ratios across the two land uses at the plot scale. Saturated hydraulic conductivity (K_s) and bulk density were different between the forest and pasture sites (p < 10⁻⁴). Simulating overland flow in HYDRUS-1D produced more outputs similar to the overland flow recorded at the pasture site than the forest site. Results from our study indicate that, at the plot scale, Hortonian overland flow is the main driver for overland flow generation at the pasture site during storms with high-rainfall totals. We infer that the combination of a leaf litter layer and the activation of shallow preferential flow paths resulting in shallow saturation-excess overland flow are likely the main drivers for plot scale overland flow generation at the forest site. Results from this study contribute to the broader understanding of the delivery of freshwater to streams, which will become increasingly important in the tropics considering freshwater resource scarcity and changing storm intensities.     

Effect of climate change on overland flow generation: a case study in central Germany

The impact of global climate change on runoff components, especially on the type of overland flow, is of utmost significance. High‐resolution temporal rainfall plays an important role in determining the hydrological response of quick runoff components. However, hydrological climate change scenario analyses with high temporal resolution are rare. This study investigates the impact of climate change on discharge peak events generated by rainfall, snowmelt, and soil‐frost induced runoff using high‐resolution hydrological modelling. The study area is Schäfertal catchment (1.44 km²) in the lower Harz Mountains in central Germany. The WaSiM‐ETH hydrological model is used to investigate the rainfall response of runoff components under near future (2021–2050) and far‐distant future (2071–2100) climatic conditions. Disaggregated daily climate variables of WETTREG2010 SRES scenario A1B are used on a temporal resolution of 10 min. Hydrological model parameter optimization and uncertainty analysis was conducted using the Differential Evolution Adaptive Metropolis (DREAM_(ZS)) uncertainty tool. The scenario results show that total runoff and interflow will increase by 3.8% and 3.5% in the near future and decrease by 32.85% and 31% in the far‐distant future compared to the baseline scenario. In contrast, overland flow and the number and size of peak runoff will decrease moderately for the near future and drastically for the far‐distant future compared to the baseline scenario. We found the strongest decrease for soil‐frost induced discharge peaks at 79.6% in the near future and at 98.2% in the far‐distant future scenario. It can be concluded that high‐resolution hydrological modelling can provide detailed predictions of future hydrological regimes and discharge peak events of the catchment. Copyright © 2014 John Wiley & Sons, Ltd.     

Estimation of irrigation flow by hydrograph analysis in a complex agricultural catchment in subtropical China

Estimating the amount of irrigation water is challenging at the catchment scale because of the difficulties in direct measurement and interactions between the flow components. The objectives of the study were to characterize the catchment flows in an agricultural catchment with an irrigation system in subtropical China and to estimate catchment irrigation flow using hydrograph analysis methods. A weighting model and multiple regression models were established to estimate catchment irrigation outflow according to the hydrographs of the inflows and outflows of the catchment. The multiple regression models took into consideration the drainage time of base flow, resulting in better estimation on an event and annual basis. Using the MR‐6d method, the estimated irrigation outflows amounted to 3700 mm, 2600 mm and 2760 mm during 2001, 2002 and 2003 respectively, which covered 70%, 60% and 64% respectively of the total catchment outflows in the corresponding years. Copyright © 2007 John Wiley & Sons, Ltd.     

The performance of grass filter strips in controlling high‐concentration suspended sediment from overland flow under rainfall/non‐rainfall conditions

There is little information on the performance of vegetative filter strips (VFS) in filtering high‐concentration sediment from subcritical overland flow. Flume experiments on simulated grass strips were conducted using combinations of three slope gradients (3°, 9° and 15°), five 1‐m‐wide slope positions (from upslope to downslope), two flow rates (60 and 20 L min^‐1 m^‐1) and sediment concentrations of 100–300 kg m^‐3 under simulated rainfall and non‐rainfall conditions. The results showed that sediment deposition efficiency increased with VFS width as a power function. Rainfall significantly reduced sediment deposited within VFS. Higher sediment concentration corresponded to a larger sediment deposition load but reduced deposition efficiency. Flow rate had a negative effect on deposition efficiency but no effect on deposition load. Sediments were more easily deposited at the upper slope position than downslope, and the upper slope position had a higher percentage of coarse sediments. The deposited sediment had significantly greater median diameters (D₅₀) than the inflow sediment. A greater proportion of coarse sediments larger than 25 µm in diameter were deposited, and particles smaller than 1 µm and of 10–25 µm had a better deposition performance than particles of 1–10 µm. Rainfall reduced the deposited sediment D₅₀ at a slope gradient of 3° and had no significant influence on it at 9° or 15°. A higher sediment concentration led to a smaller D₅₀ of the deposited sediment. Rainfall had no significant effect on overland flow velocity. Both the deposited sediment load and D₅₀ decreased with increasing flow velocity, and flow velocity was the most sensitive factor impacting sediment deposition. The results from this study should be useful to control sediment flowing into rivers in areas with serious soil erosion. Copyright © 2013 John Wiley & Sons, Ltd.     

One‐ and two‐dimensional modelling of overland flow in semiarid shrubland,Jornada basin,New Mexico

Two distributed parameter models, a one‐dimensional (1D) model and a two‐dimensional (2D) model, are developed to simulate overland flow in two small semiarid shrubland watersheds in the Jornada basin, southern New Mexico. The models are event‐based and represent each watershed by an array of 1‐m² cells, in which the cell size is approximately equal to the average area of the shrubs. Each model uses only six parameters, for which values are obtained from field surveys and rainfall simulation experiments. In the 1D model, flow volumes through a fixed network are computed by a simple finite‐difference solution to the 1D kinematic wave equation. In the 2D model, flow directions and volumes are computed by a second‐order predictor–corrector finite‐difference solution to the 2D kinematic wave equation, in which flow routing is implicit and may vary in response to flow conditions. The models are compared in terms of the runoff hydrograph and the spatial distribution of runoff. The simulation results suggest that both the 1D and the 2D models have much to offer as tools for the large‐scale study of overland flow. Because it is based on a fixed flow network, the 1D model is better suited to the study of runoff due to individual rainfall events, whereas the 2D model may, with further development, be used to study both runoff and erosion during multiple rainfall events in which the dynamic nature of the terrain becomes an important consideration. Copyright © 2006 John Wiley & Sons, Ltd.     

Assessment of overland flow variation and blue water production in a farmed semi-arid water harvesting catchment

Upgrading agriculture in semi-arid areas and ensuring its sustainability require an optimal management of rainfall partition between blue and green waters in the farmed water harvesting catchment. The main objective of this study is to analyze the influence of heterogeneous land use on the spatial and temporal variation of rainfall partitioning and blue water production within a typical farmed catchment located in north-eastern Tunisia. The catchment has an area of 2.6 km² and comprises at its outlet a dam, which retains the runoff water in a reservoir. Overland flow and soil water balance components were monitored during two cropping seasons (2000/2001 and 2001/2002) on a network of eleven plots of 2 m² each with different land use and soil characteristics. The hydrological balances of both the catchment and reservoir have been monitored since 1994.Observed data showed a very large temporal and spatial variability of overland flow within the catchment reflecting the great importance of total rainfall as well as land use. During the 2001/2002 season the results showed a large variation of the number of observed runoff events, from 27 to 39, and of the annual overland flow depths, from 8 mm (under vineyard on calcaric cambisols) up to 43 mm (under shrubs-pasture on haplic regosols), between the plots. The annual runoff amounts were moderate; they always corresponded to less than 15% of the annual rainfall amount whatever the observation scale. It was also observed that changes in land use in years with similar rainfall could lead to significant differences in blue water flow. An attempt for predicting the overland flow by the general linear regression approach showed an r² of 31%, the predictors used are the class of soil infiltration capacity, the initial moisture saturation ratio of the soil surface layer and the total rainfall amounts.These experimental results indicate that the variation in land use in a semi-arid catchment is a main factor of variation in soil surface conditions and explain the major role played by the former on hydrological behavior of the upstream area and on rainfall partition between overland flow and infiltration. Therefore, to predict the water harvesting capacities in terms of blue water production of a farmed catchment in semi-arid areas it seems essential to consider precisely its land use and its temporal evolution related to management practices.