| Researches on climate change and human activities impacts on runoff and sediment load in the watershed scale can not only contribute to early warning of watershed flood and sediment disaster,but also provide scientific basis to evaluate environmental effects of ecological construction.Based on the research weekness and gap of the impacts of extreme precipitation change and land use spatial pattern on watershed runoff and sediment load,this paper analyzed the dynamic changes characteristics of extreme precipitation,extreme temperature,runoff and sediment load in Xiaolihe watershed located in loess hilly-gully region in the middle reaches of the Yellow River from 1959 to 2013,by using moving average method,anomaly analysis,Sen slope estimation,Mann-Kendall non-parametric test and mutation point test.This paper also discussed the land use spatial and temporary variation processes,spatial autocorrelation pattern,landscape pattern spatial and temporary changes and their spatial differentiation characteristics affected by topography changes,by using GIS technology,land use transfer matrix,land use dynamic degree,spatial autocorrelation pattern analysis and landscape pattern analysis.Based on the mutation point inspection on runoff and sediment load,this paper selected the optimal extreme precipitation index affecting watershed runoff and sediment load.This paper quantified the contributions of extreme precipitation change and human activities to watershed runoff and sediment load based on the selected optimal precipitation index by using double mass curve method and regression analysis.Based on the model calibration and validation of SWAT model,this paper discussed the response mechanism of watershed runoff and sediment load to land use landscape pattern change,and analyzed the response characteristics of watershed runoff and sediment load to climate change,land use change and other human activities.The main research results were as follows:(1)During 55 years from 1959 to 2013,the annual precipitation,flood season precipitation,spring precipitation,summer precipitation and autumn precipitation showed insignificant decreasing trends,while the winter precipitation showed an insignificant increasing trend.In 17 extreme precipitation indices(R10,R20,R50,R10 p,R20p,R50 p,R1d,R5 d,SDII,R90 p,R90t,R95 p,R95t,R99 p,R99t,CDD and CWD),CDD and SDII showed insignificant increasing trends,while other 15 indices showed insignificant decreasing trends.The annual and seasonal average temperature showed increasing trends,among which annual average temperature,spring average temperature and winter average temperature showed significant increasing trends.In 16 extreme temperature indices,cold indices(FD,ID,TN10 p,TX10p,TNn and CSDI)showed decreasing trends,among which ID,TX10 p and TN10 p showed significant decreasing trends;warm indices(TR,HD,SU,TN90 p,TX90p,TNx,TXn,TXx,WSDI and DTR)showed increasing trends,among which SU,DTR,HD and TN90 p showed significant increasing trends.This result showed that watershed tended to be dry and warm obviously.(2)All land use types of the watershed showed positive global spatial autocorrelation,but this correlation gradually decreased with the increase of distance.Land use types had different local spatial cluster or abnormal features,and the local cluster areas showed spatial differentiation with elevation,slope gradient and slope aspect.From 1990 to 2009,landscape fragmentation degree gradually increased,landscape shape became more complex,landscape types become more diversification,and landscape heterogeneity increased.Landscape pattern metrics were significantly correlated with elevation,slope gradient and slope aspect.(3)The annual runoff,flood season runoff,annual sediment load and flood season sediment load showed highly significant decreasing trends,respectively,and their decrements per year were 29.13×104 m3,18.6×104 m3,6.33×104 t and 6.49×104 t respectively during past 55 years.The abrupt change years of runoff and sediment load all appeared in 1970.Thus,the whole research period was divided into two periods: 1959-1970 as a baseline period and 1971-2013 as a controlled period.Compared with baseline conditions,the annual and flood season runoff deceased by nearly 40% and 50%,respectively,and the annual and flood season sediment load decreased by nearly 72% in controlled period.Correlation coefficients between watershed runoff,sediment load and annual precipitation,flood season participation,16 extreme precipitation indices varied between 0.28~0.9 in baseline period,while these correlation coefficients varied between 0.18~0.62 in controlled period.This result showed that runoff and sediment load were mainly affected by precipitation in baseline period,while both were affected by the combination of precipitation and human activities in controlled period.During the baseline period,controlled period and the whole research period,R90p(annual total precipitation when daily precipitation >90th percentile in the 1961-1990 period)had the best correlation with runoff and sediment load.(4)R90p was selected as the best extreme precipitation index to represent extreme precipitation conditions.The contribution rates of extreme precipitation change and human activities on the runoff and sediment load changes were quantitatively evaluated.The results showed that,the impacts of human activities on the runoff and sediment load changes were dominated in the 1970 s,1980s and in 1999-2013,the contribution rates of human activities to runoff reduction were 57.3%,62.1% and 75.4%,respectively,the contribution rates of human activities to sediment load reduction were 91.8%,92.9% and 90.3%,respectively,so the influence of human activities on sediment load was much greater than that on runoff.While the impact of extreme precipitation change on the runoff and sediment load changes were dominated in 1990-1998,the contribution rates of R90 p to runoff and sediment load reduction were 78.2% and 57.9%,respectively,so the influence of extreme precipitation on runoff was greater than that on sediment load in this period.(5)SWAT model parameter database that fitted in the Xiaolihe watershed was built.Through the model parameter sensitivity analysis,model calibration and model precision verification,the simulation effect evaluation indices of R2,NSE and PBAIS varied between 0.84~0.89,0.80~0.86 and-7.6%~2.3%,respectively,which indicated the model simulated precision was accepted.Based on the calibrated SWAT model and designing different scenarios of land use and climate,the impact of land use landscape pattern variation on runoff and sediment load was analyzed,and the contribution rates of climate change,land use change and other human activities to runoff and sediment load change were evaluated.The results exibited that,the landscape pattern of land use significantly affected the runoff and sediment load in the Xiaolihe watershed.At the landscape level,annual runoff and sediment load were significantly positively correlated with NP,PD,LSI and SHDI,while annual runoff and sediment load were significantly negatively correlated with LPI,COHESION,CONTAG and SHEI.Setting runoff and sediment load in1971-1998 as a reference,contribution rates of climate change and land use change to runoff reduction were 29.8% and 70.2%,respectively,contribution rates to sediment load reduction were 13% and 87%,respectively in 1999-2013.Meanwhile,contribution rates of other human activities except land use change to runoff and sediment load reduction(such as engineering construction,soil and water conservation measures,etc.)were 5.2% and 3.4% respectively in 1999-2013. |
PDF Full Text Request