The urban area of Linyi City is endowed with the basic conditions for the development of karst collapse. In this area, Ordovician limestone with karst development is widely distributed, and the limestone is covered by a thin Quaternary overburden with a binary phase structure. From the 1980s to the early 21^st century, the rapid development of Linyi City, the increase of groundwater exploitation and the decline of the overall regional water level led to the frequent occurrence of geological disasters of karst collapses, causing serious economic losses and social impact.

In this study, the karst collapse in Linyi City is taken as the research object. The distribution law, geological conditions, and inducing factors of karst collapses are studied through statistical analysis. According to the analysis of the genesis mechanism of karst collapse in Linyi City, an early warning model based on the occurrence index of karst collapse is proposed. Taking the water level as the main monitoring factor, this model is constructed to comprehensively judge the early warning by the combination of multi-factors, and is verified by historical data.

During the high incidence period from 1993 to 2012, a total of 17 karst collapses occurred in the study area. These karst collapses were caused by geological conditions such as topography, geological structure, and formation lithology, as well as inducing factors such as groundwater exploitation, human engineering activities, and precipitation. The distribution law of karst collapse reflects the consistency with its related influencing factors. Karst collapses are mainly distributed in the caprock area of double-layer structure with shallow karst development and small overburden thickness. They are also distributed near the surface water body and fault zone. Generally, karst collapses take place in the funnel areas and in the seasons with a large variation of water levels, They also occur in the influence range of human engineering activities. With a comprehensive index method, an early warning model based on the judgment made through multi-factors is established. The model comprehensively considers geological conditions and inducing factors and selects 10 influencing factors, including karst development degree, caprock thickness, caprock structure, distance from structure, distance from surface water, distance between water level and limestone roof, water level amplitude, distance from the center of depression cone, precipitation, and human engineering activities. Different weights of influencing factors and the range of influencing factors of each sub-condition are given and the early warning level is divided into four grades according to the occurrence index value.

The model is verified by using historical data of the collapse when it occurred. For example, the data of 2003 was used for simulation and early warning validation. The karst collapse occurred in the east of the No.32 building in Lanshan community on May 8 and June 22 in 2003, and all the collapse points were located in the area at a high early warning level, and also fell into the time period of high early warning level. The results show that the model has high reliability in the verification of early warning in key monitoring areas of Linyi City, which can provide a reference for monitoring and early warning of karst collapse in other areas.