| Rock desertification and soil erosion are prominent ecological problems in the southwest karst region,threatening the safety of terrestrial ecosystems and constraining socio-economic development.Since the implementation of the rock desertification control project,the karst ecological environment has improved significantly,and the temporal and spatial distribution of runoff and sediment has also undergone significant changes.However,the spatial and temporal variation characteristics of runoff sediment in karst watersheds and its main drivers are not quite clear at present.On the other hand,SWAT(Soil and Water Assessment Tools)model,as an important tool for watershed hydrological simulation,neglects the rapid recharge process of runoff through sinkholes to underground aquifers when simulating the hydrological process in karst watersheds,which somewhat restricts the application of SWAT model in karst basins.In view of this,the present study takes a typical karst watershed,the Dabang River Basin,as the study area,and combined the hydrogeological characteristics of the depression-sinkhole system,explored the proportional relationship between the rapid and retarded flow of the depression-sinkhole system through artificial simulation tests,and an optimization of the SWAT model was attempted based on the experimental results,and at the same time,the soil and water loss situation of the basin from 2008 to 2020 was calculated by combining the long series high-frequency measured runoff and sediment data of the Dabang River basin.Further,the spatial and temporal distribution characteristics,evolution trends and distribution of cold and hot zones of runoff sediment in the basin were analyzed using the methods of Mann-Kendall test,Getis-Ord G_i*analysis,etc.And the partial least squares regression(PLSR)model was used to decouple the main drivers affecting the variation of runoff and sediment yield in karst basin.The main results are as follows:(1)The proportional relationship between rapid and retarded flow was investigated by indoor artificial simulation tests.The results showed that under different rainfall intensities and bedrock exposure rates,the time required for rapid flow out of the surface ranges from 1.51 to 2.32s,with an average of 1.82s,and the variation pattern is that with the increase of rainfall intensity,the velocity of rapid flow accelerates,while the increase of bedrock exposure will reduce the velocity of rapid flow.The time required for retarded flow out of the surface ranged from 64.23to 165.32s,with an average of 112.14s.The variation pattern showed that the velocity of retarded flow was accelerated with the increase of rainfall intensity and bedrock exposure rate.And under this experimental condition,the average ratio relationship between rapid and retarded flow rates was about 1:62.(2)Based on the DEM,land use,soil type and meteorological data of the Dabang River basin,the SWAT model of the basin was constructed,and monthly runoff measured data from Gaoche and Huangguoshu hydrological stations were used for the calibration and validation,with2008-2014 as the calibration period and 2015-2020 as the validation period.The applicability of the model in the present study area was assessed in terms of the coefficient of determination(R~2)and Nash efficiency coefficient(NSE),and the results showed that in the calibration period,R~2=0.84,NSE=0.83 at Huangguoshu hydrological station and R~2=0.71,NSE=0.68 at Gaoche hydrological station;in the validation period,R~2=0.79,NSE=0.55 at Huangguoshu hydrological station and R~2=0.75,NSE=0.55 at Gaoche hydrological station,indicating that the SWAT model is basically applicable to karst basin.(3)Some optimization of the SWAT model was attempted according to the average scaling relationship between the rapid and retarded flow velocity.The optimization method is to add the pond module and simulate the depression-sinkhole system with the pond module,and modify the formula for calculating the daily recharge amount of aquifer in karst area,so as to portray the rapid recharge process of surface water through the depression-sinkhole system to underground aquifer.The results showed that the R~2 and NSE of both hydrological stations increased to different degrees in both periods,except for the NSE of Gache hydrological station in the validation period,which did not change before and after optimization.On the other hand,the water yield of the No.4 and No.19 sub basins where the ponds are located showed a decreasing trend after model optimization,indicating that the surface runoff decreased and baseflow increased after SWAT model optimization,i.e.,the water yield was redistributed,thus better characterized the rapid recharge of runoff through the depression-sinkhole system to groundwater.Therefore,the optimized SWAT model is more applicable than the original model in portraying the hydrological processes in karst basins.(4)The statistical characteristics of monthly runoff and sediment yield in the study area from 2008 to 2020 showed that the minimum value of runoff is 5.96 mm,the maximum value is 329.20 mm,and the average value is 60.14 mm;the minimum value of sediment modulus is 0.01 t/ha,the maximum value is 0.99 t/ha,and the average value is 0.17 t/ha.The variation characteristics of runoff and sediment yield and rainfall are consistent,with high values mainly occurring in summer and autumn,mostly concentrated in May to October,and low values mainly occurring in winter and spring.In terms of the annual variation trend,the runoff sediment process shows a trend of increasing and then decreasing.In addition,the peak of sediment yield was earlier than the peak of runoff yield in 2008,2009,2010,2012,2014 and 2020.In terms of spatial distribution,it can be found that the high and low value zones of the runoff and sediment yield are similarly distributed in the two stages,with the high value zones being distributed in the northwestern,northern and central-eastern parts of the study area,and also in the northeastern and southern parts in the later stages;while the low value zones are mainly distributed in the central-western,northeastern and southern parts of the study area.(5)Through Mann-Kendall time series trend analysis and mutation test,it was found that the runoff and sediment yield in the study area showed an overall increasing trend in terms of temporal variation,and mutations occurred during the study period;while spatial variation showed non-equilibrium,with increasing and decreasing zones co-existing.The results of the Getis-Ord G_i*analysis showed that the distribution of the hot zones of runoff and sediment production(major prevention and control area of soil erosion)is similar,with the most concentrated in the northwestern,northern,northeastern,eastern and southeastern parts of the study area;while the distribution of the cold zones(slight area of soil erosion)is different,with the cold zones of runoff production mainly concentrated in the western and southwestern parts of the basin,the sediment production cold zones are relatively small and mainly concentrated in the west of the center of the basin.(6)A PLSR model between runoff sediment and 12 drivers was constructed,and the four main factors influencing runoff production were identified by the variable importance for projection(VIP),namely R25,BEP,R50 and FVC,which together explained 63.07%of the variation in runoff;and five main factors influencing the variation of sediment production were also identified,namely R25,R50,FVC,BEP and A,and which together explained a total of 59.59%of the variation in sediment in the study area.In addition,the results of the study indicated that the specific reasons for the increase in runoff and sediment yield are the increase in extreme rainfall R25 and the decrease in bedrock exposure rate,while the spatial decrease was due to the decrease in R50 and the increase in vegetation coverage.The present study provides deeper insights into the spatial and temporal variation characteristics and their driving mechanisms of runoff and sediment production in heterogeneous karst watersheds,which can help karst regions to develop prospective soil and water conservation and watershed management programs in the context of regional realities.Meanwhile,based on the simulation tests,it provides a reference for further deeper optimization of the SWAT model in karst basins in the future.In addition,the present study also has important reference significance for the planning and rational utilization of water resources in karst basins. |
PDF Full Text Request