Study On Suspended Sediment Sources In Typical Collapsing Hill Watershed In Granitic Region In The South

To analyse suspended sediment sources in a typical collapsing hill watershed in the granitic region,a collapsing hill watershed in the Longmen town,Anxi County,was selected as the research object.85 soil samples from four kinds of sediment source(erosive woodland,tea plantations,the cultivated land and the erosive area of collapsing hill area)were collected,in the meantime,erosion suspended sediment in the sampler in the channel was also collected.34 kinds of fingerprint factors of the samples were analyzed,then the best combination of fingerprint factors was screened,and the migration regularities of the sediments derived from different sources were figured out.Main results were as followed:(1)The nonparametric Kruskal-Wallis test was used to test all the factors,and the significant difference(P<0.05)factors were identified.The corresponding fingerprint factors in each sediment collection domain are as follows:There were 5 factors(Li,Co,Ni,Sb and Ca)in the sediment collection domain called A.There were 19 factors(Li,Ti,Cr,Mn,Co,Ni,Cu,Zn,Sr,Cd,Sn,Pb,Ca,Al,V,As,P,quartz and kaolin)in B.There were 26 factors(Li,B,Ti,Cr,Mn,Co,Ni,Zn,Sr,Cd,Sn,Ba,Tl,Pb,Bi,Ca,Mg,V,Ga,As,P,K,quartz,kaolin,pyroxene and plagioclase)in C.And the D had 28 factors(Li,Ti,Cr,Mn,Co,Ni,Cu,Zn,Sr,Cd,Sn,Ba,Tl,Pb,Bi,Fe,Ca,Mg,Al,V,Ga,As,P,K,quartz,kaolin,pyroxene and plagioclase).(2)The multivariate regression analysis was used to obtain the best combination of fingerprint factors.So the best fingerprint combination of A was Li,Sb and Ca.There were 4 factors(Ti,Li,Sn and Ca)in the sediment collection domain called B.The C was Ca,Li,Sn and Mn.And sediment collection domain D was Ca,Li,Sn,K and Ba 5 kinds of fingerprint factors.The cumulative contribution rate of each sediment collection area was more than 90%,which met the requirements of the analysis.The correct classification rates from different sources of sediment were above 80%.It showed the reliability of the established discriminant model and was planed for the next step.(3)Using a multivariate mixed model quantitatively could obtain uspended sediments contribution of different sediment sources in different rainfall conditions.At the same time,it was concluded that the percentage contribution of suspended sediment by erosive woodland was approximately 55.26%and about 20%by tea plantations and cultivated land in the A.The relative sediment contribution rates were tea garden(21.51%),cultivated land(20.07%)and gully erosion area(58.42%)in the B.There was the erosion of forest land(17.08%),tea(42.18%),(19.31%)land and collapsing erosion area(21.43%)in the C.There was the erosion of forest land(21.76%),tea(24.55%),(29.99%)land and collapsing erosion area(23.70%)in the D.The goodness of fit showed that all were greater than 0.8.So it proved once again that the hybrid model results could be accepted.(4)The coupling relationship between the suspended sediment sediment source contribution rate and rainfall:The contribution rate of tea plantation and farmland in the sub basin A increased with the increase of rainfall data 130.The contribution percentage of tea plantation in B was positively correlated with rainfall rainfall and average rainfall intensity,and the contribution of cultivated land increased with the increase of rainfall intensity.The contribution of suspended sediment in the gully erosion area was positively correlated with rainfall rainfall,duration and average intensity,and it was positive correlation.In the sub basin C,the percentage of the contribution of the forest land was positively correlated with the average rainfall intensity,and the contribution of the tea plantation was positively correlated with the rainfall rainfall,and the contribution value of the erosion area was only positively correlated with the rainfall duration.There was a significant positive correlation between the sediment contribution value of tea plantations and rainfall duration in the sub basin D.And so was rainfall.