Received: 23 Jan 2019 – Discussion started: 12 Feb 2019
Abstract. Studies about landslide modeling and monitoring are becoming more diverse. Data assimilation is an approach to combine mechanism models and observations. In this study, an improved particle filtering algorithm is used to assimilate the transient rainfall infiltration and grid-based regional slope-stability analysis (TRIGRS) model and landslide surface deformation monitoring data observed with GPS and InSAR. After assimilation calculation, FS has been effectively corrected, rather than continuously decreasing as the background model output. The root mean square difference (RMSD) tends to decrease from a maximum of 0.084 to a minimum of 0.026 in the process of assimilation, which means the assimilation process makes the model output FS closer to the actual observations. The friction angle (φ), which is an investigated parameter, can be updated and fed back in each step of assimilation. The value of the investigated parameter makes the model output closer to the observation. The groundwater pressure head is output as an assimilation result simultaneously.
How to cite. Xue, C., Nie, G., Dong, J., Wu, S., Wang, J., Li, X., and Zhang, X.: Landslides Data Assimilation Using TRIGRS Based on Particle Filtering, Nat. Hazards Earth Syst. Sci. Discuss. [preprint], https://doi.org/10.5194/nhess-2019-16, 2019.
This paper provides an approach to apply data assimilation method to stability analysis and parameter update and feedback in a landslide. The experiment is implemented by particle filter algorithm. The result FS sequence of TRIGRS output decreases continuously with time and the assimilation can effectively correct the FS of the model output. The RMSD of FS indicates the assimilation results can correct the estimation of TRIGRS output close to actual observations.
This paper provides an approach to apply data assimilation method to stability analysis and...