Three Gorges Reservoir lies in the upstream of Yangtze River, massive landslides
hazard is frequent and widespread because of the rough terrain and susceptible
geological environment, which is seriously threatened the local settlement and their
development. As a result the convenient hazard assessment and prediction in this area
is critical for disaster prevention and reduction. This research is exampled in Wanzhou
county, Chongqing city, and the assessment and prediction model’re established
according to the probability method. Landslide hazard assessment is the
environmental conditions of the landslides occurrences, combining with precipitation
and rainfall probability prediction of Meteorological Department as the triggering
factor, the prediction model are established, which can be used in both regional scale
and site-specific rainfall-induced landslide prediction in the study area.
This paper consists of nine chapters: (1) Introduction; (2) Background of the
study area; (3) Distribution and mechanism of landslides in the study area; (4)
Objective weight of factors in hazard assessment model ; (5) Zonation method of the
regional-scale landslide hazard degree; (6)The site-specific landslide hazard degree
assessment method; (7) Study on relation between the rainfall-induced landslides and
precipitation ; (8) Study on the prediction method of rainfall-induced landslide ; (9)
Summarization and prospect.
This paper consists of five main achievements as below:
(1) The regular pattern of distribution and mechanism of landslides in the study
area’re deeply Analyzed, temporal distribution is analyzed separately in interannual
scale, annual scale and day scale; statistics of spatial distribution is analyzed by the
influences of landslides and the environmental factors including lithology, stratum, terrain and slope shape and so on. Mechanism of landslides is dissertated by both
environmental factors and triggering factors.
(2) The model of regional scale landslide hazard zonation is established by the
landslide occurrence probability and information entropy to get objective weights of 6
bottom factors, which is applied in wanzhou county. The result is verified by 3
different classification methods and validity is proved for the consistence of different
classification methods.
(3) By combining the 8 factors of landslide influence and surface phenomena and
their information entropy weights, the site-specific landslide hazard assessment model
is established and applied in 4 typical landslides in Wanzhou County. The validity of
results is proved by the deep displacement monitoring data and the literatures.

(4) analyzing the relation between rainfall and landslide is carried out by
statistics of precipitation and landslides data in 2 decades, precipitation of 24 hours on
the day landslide occurring and precipitation 10 days before landslide occurring are
analyzed, and rain types of rainstorm and Continuous is also taking into account.
Then the expression of probability of rainfall-induced landslide is established after
calculation of 4 days rainfall related by SPSS with the Binary logistic method.
(5) Prediction model of rainfall-induced landslide is built by taking precipitation
and rainfall probability offered by Meteorological Department. The result of
prediction is classified into four grades which include green expressing with the first
grade, yellow with the second grade and orange red with the third grade, red with the
fourth grade. Prediction examples in study area is simulated by both regional scale
and site-specific scale by the model and prediction system platform is established by
software engineering and GIS development.