Search

• Calculation unit division induced soil spatial aggregation in SWAT model. • Effect of soil spatial aggregation on simulation was separated and quantified.Soil spatial aggregation caused a significant increase in annual runoff and sediment load. • The effect of soil spatial aggregation on sediment load was greater than on runoff. • Optimal sub-watershed division scheme was discussed. Distributed hydrological model divides watersheds into distinct calculation units to simulate hydrological processes. Within the model, different calculation unit division schemes induce input elements, such as soil and land use, to certain degrees of generalization, thus affecting the simulation result. Analyzing the effect of the calculation unit division on the simulation result is an important methodology to reduce uncertainty and improve the simulation accuracy of the model. Previous researchers have focused on the comprehensive effect of underlying surface generalization (such as soil, land use, and terrain generalization) caused by the calculation unit division on the simulation results, while it did not separate the effect of the generalization of the individual input factor on the simulation result. In this paper, the Dongjiang headwater region in southern China was selected as the study area. The single type of land use and slope were input into the Soil and Water Assessment Tool (SWAT) model to fix the effects of land use and terrain caused by calculation unit division on the simulation results. Eleven calculation unit division schemes were used to analyze the impact of soil spatial aggregation on runoff and sediment load simulations. The calculation unit division schemes included 11 sub-watershed division schemes and their hydrological response unit (HRU) thresholds were set at 8 %. The drainage area thresholds of the 11 sub-watershed division schemes were 30 ha, 60 ha, 120 ha, 240 ha, 480 ha, 1000 ha, 1600 ha, 2200 ha, 2800 ha, 3400 ha and 4000 ha. The results demonstrated that the soil spatial distribution showed an obvious aggregation with an increase in the drainage area threshold in the SWAT model. Of these, the Krasnozem soil type which covers the largest area in the watershed, increased significantly with an increase in the drainage area threshold. However, the areas of all the other soil types decreased significantly. The Lioth soil type covers the smallest area in the watershed decreased and even disappeared when the threshold in the SWAT model was greater than 1000 ha. Furthermore, soil spatial aggregation caused a significant increase in the annual runoff and sediment load (P < 0.05), with the sediment load (Re = 29.78 %) increasing at a greater rate compared to runoff (Re = 1.74 %). Increased soil aggregation significantly increased the sediment loads in the maximum 1-day, the continuous maximum 5-day, the continuous maximum 7-day, and the sediment load in flood processes (P < 0.001). Refining the calculation unit division scheme will cause a significant increase in the number of calculation units. Due to the computer processing power limitations, we cannot further refine the calculation unit division scheme. Based on our current operational calculation unit division scheme, we did not identify the optimal division scheme in the study region. These results can provide reference information for future research into hydrological model uncertainty and can improve the SWAT model simulation accuracy. [ABSTRACT FROM AUTHOR]

Summary Land cover and rainfall regime are two important factors that affect soil erosion. In this paper, three land cover types – grass cover, litter cover and bare land – were employed to analyze surface runoff, subsurface flow and sediment loss processes in relation to the rainfall regimes in the red soil region of China. Five rainfall regimes were classified according to 393 rainfall events via a k-means clustering method based on the rainfall depth, duration and maximum 30-min intensity. The highest surface runoff coefficient and erosion amount were found on bare land in all five rainfall regimes, and the lowest were found on grass cover. The litter cover generated the highest subsurface flow rate, followed by the grass cover; the lowest was on bare land. For grass cover and litter cover plots, rainfall events of rainfall regime IV which had the longest duration, greatest depth and lowest intensity had the highest surface runoff coefficient, soil erosion amount and subsurface flow rate. For bare land, storm rainfall events of rainfall regime V had the highest intensity, lowest depth and duration, had the highest surface runoff coefficient and soil erosion amount, but the lowest subsurface flow rate. The highest subsurface flow rate of bare land happened in rainfall regime IV. Surface cover was urgently needed to reduce soil erosion. When the lands under dense surface cover, more attention should be paid to rainfall events that of long duration, high depth but low in intensity which commonly occurred in spring. The interactions of surface–subsurface flow and its effects on soil erosion and nutrient loss were worth considering in the red soil region. [ABSTRACT FROM AUTHOR]

Abstract: Vegetation recolonization has often been used to control roadside slope erosion, and in this paper, four restoration models – Natural Restoration, Grass, Grass & Shrub, Sodded Strip – were chosen to recolonize the plants on a newly built unpaved roadside slope in the Three Gorges Reservoir Area. After eight months growth, eight rainfall simulations (intensity of 90mmh−1 for 60min) and in-situ soil shear strength test were then carried out to identify the impacts of vegetation on roadside slope erosion and soil shear strength. The erosion on cutslopes was higher than that on fillslopes. The runoff coefficient and soil detachment rate were significantly lower on the Grass & Shrub model (4.3% and 1.99gm−2 min−1, respectively) compared with the other three, which had the highest surface cover (91.4%), aboveground biomass (1.44kgm−2) and root weight density (3.94kgm−3). The runoff coefficient and soil detachment rate on roadside slopes showed a logarithmic decrease with the root weight density, root length density and aboveground biomass. The soil shear strength measured before and after the rainfall was higher on Grass & Shrub (59.29 and 53.73kPa) and decreased on Grass (46.93 and 40.48kPa), Sodded Strip (31.20 and 18.87kPa) and Natural Restoration (25.31 and 9.36kPa). Negative linear correlations were found between the soil shear strength reduction and aboveground biomass, root weight density and root length density. The variation of soil shear strength reduction was closely related to the roadside slope erosion, a positive linear correlation was found between runoff coefficient and soil shear strength reduction, and a power function was shown between soil detachment rate and soil shear strength reduction. This study demonstrated that Grass and Grass & Shrub were more suitable and highly cost-effective in controlling initial period erosion of newly built low-volume unpaved road. [Copyright &y& Elsevier]

Summary: Annual runoff and annual suspended sediment loads of hydrological gauging stations along the mainstream of the Yangtze River basin (Pingshan station, Yichang station, Hankou station and Datong station) and main tributaries (Beipei station in Jialingjiang River, Wulong station in Wujiang River and Huangzhuang station in Hanjiang River) were analyzed with the help of Mann–Kendall trend analysis and linear regression analysis. Research results indicate that (1) changing patterns of runoff and sediment loads are different in different parts of the Yangtze River basin. No significant trend is detected for annual runoff at all stations at >95% confidence level. Changes of sediment loads, however, demonstrate different pictures in the Yangtze River basin. The sediment loads are in increasing trend in Pingshan station- the most upstream station on the Yangtze River basin (this increasing trend is significant at >95% confidence level after about 1990), but are in decreasing trend at other stations (including stations in the tributaries studied in this paper). This decreasing trend becomes more obvious from Yichang station to Datong station. (2) Water reservoirs exerted more influences on changes of sediment loads than on runoff, which is the main reason for the decreasing trend of sediment loads found in most stations. (3) Influences of water reservoirs on sediment loads are more obvious in the tributaries than in the mainstream of the Yangtze River basin, while in the mainstream the variation patterns of sediment loads are determined by multiple factors. [Copyright &y& Elsevier]