Background As a result of low content of acid insoluble material in bedrock and the slow weathering rate of rock in Karst area, the rate of soil formation is very little, and extremely serious environmental and social problems will occur while soil is eroded. Moreover, it is difficult to monitor soil erosion accurately because of the complex structures above ground and underground at Karst area. Here we are trying to find appropriate method to monitor the amount, variation pattern, and formation mechanism of soil erosion during the natural restoration of a typical clear cutting in Karst area in Puding county of Guizhou province.

Methods Drilling method is appropriate to trace soil erosion in situ according to previous studies, it was combined with runoff plot method to be used in this study. Test bras had been set by uniform distribution of a interval of 1 meter and the distance from top of bras to soil surface was measured each time to obtain the erosion depth between ever two measurement. on the other hand, runoff plot was built around the clear cutting plot and sediment was collected and weighed after every heavy rainfall. The result of these two methods was compared.

Results 1) During the 4.5 years of restoration after clear cutting, the result by drilling method showed an overall trend of erosion, with a cumulative erosion depth of 1.46 cm (Equivalent erosion modulus 2 066.7 t/(km²·a)), which had a large fluctuation at the first one year of the natural recovery, then the fluctuation, decreased (except the test of April 2015(about 3 years of the natural recovery)), and tended to be stable after 3.5 years. 2) The erosion modulus obtained by the runoff plot method was 0.78 t/(km²·a), and the sediment can only be collected after one rainfall of 67.6 mm, which rainfall intensity reached 42.4 mm/h within two months after clear cutting. 3) There was a huge difference between the erosion modulus obtained from the drilling method and runoff plot method (about 2 065.92 t/(km²·a)). This might be resulted from the special surface rock and soil distribution pattern of Karst area and the different observation emphases and properties between the two monitoring methods. The results by the two methods could not be compared generally. The drilling method was more representative at this situation on account of it could monitor the special move form of Karst (soil creeping and underground soil loss) by situ monitor. 4) During the early natural restoration stage, rainfall played the most important role in influencing the soil erosion due to soil exposure, but coaction of litter, soil organic matter and other factors might reduce soil erosion after the restoration of vegetation.

Conclusions The amount and variation of soil erosion at the clear cutting plot are understood at certain level in this research, but the feasible major influencing factors, which have a very important effects on soil and water conservation, are only speculated. Further research is needed.