Case study of a landslide continuous probability rainfall threshold analysis based on the prediction interval principle

Abstract

Bazhong City is located on stratum dominated by red-bed conditions. This type of weak geological condition with sand and mudstone interbedding is very easily affected by climatic conditions to produce rainfall-type landslides. On the basis of landslide data statistics collected in Bazhong City from 2011 to 2019, this paper uses ERA5-Land rainfall data to statistically analyze the correlation between rainfall and landslide events in Bazhong City. The landslide events in Bazhong City are greatly affected by rainfall events lasting for 10 days. Considering the influence of rainfall seepage and other processes, an effective cumulative rainfall-duration threshold curve for Bazhong City is obtained via median nonlinear fitting. Then, on the basis of the prediction interval, the rainfall threshold formula under different landslide occurrence probabilities is obtained and the critical threshold curve with a non-exceeding probability of 1% in Bazhong City is calculated and verified. Subsequently, a continuous probability distribution fitting function of landslide occurrence is established and a continuous probability distribution surface with a good fitting effect in Bazhong City is obtained. This allows a definite probability of whether future rainfall events will induce landslides to be obtained, providing an important basis for engineering disaster prevention and mitigation.

Landslide risk reduction for sustainable land use direction in Tawangmangu District, Karanganyar Regency year 2023

“…Rainwater infiltration reduces the shear strength of soil masses, increases the weight of rocks and soil masses, increases shear forces, and decreases the stability coefficient [13]. Landslides are one of the most IOP Publishing doi:10.1088/1755-1315/1314/1/012026 2 widespread, impactful, and destructive geological disasters worldwide, especially in mountainous areas [14] Changes in land use and land cover in mountainous areas can increase vulnerability to landslides due to modifications in topography, vegetation, and material characteristics [15]. Karanganyar is an area vulnerable to natural disasters such as landslides, typhoons, floods, fires, land movements, collapsed houses, and accidents [16].…”

Enhancing rainfall estimation for water resource management through multiple linear regression approach in Khordha District, Odisha, India

“…The multiple linear regression (MLR) stands out among empirical methods due to its widespread use, as it effectively illustrates the linkage between several contributing factors, providing an accurate prediction of rainfall (Amiri et al 2015;Andrews 1974;Krzywinski and Altman 2015;Raval et al 2021). Additionally, multiple linear regression aids in identifying the more influential variables compared to others (Tuleya et al 2007;Yu et al 2015;Zhang et al 2023).…”

Research on machine learning forecasting and early warning model for rainfall-induced landslides in Yunnan province

“…Currently, the prediction of RIL can be categorized into two main types: critical rainfall models and probabilistic forecasting models. (1) Critical rainfall models directly link landslide occurrence to rainfall by identifying empirical relationships between various rainfall indicators and landslide occurrences to determine critical rainfall conditions 5 , 6 . There are various empirical methods available for spatiotemporal prediction of RIL , which have been applied in disaster assessment or warning systems worldwide 7 .…”