The Impact Of Climate Change On Shallow Landslides In The Loess Plateau

The surface of the Loess Plateau is fragmented and ravines are vertical and horizontal,which is the prone area of shallow landslide geological disaster.The landslide disaster in the area is mainly induced by rainfall.Under the background of global climate change,the increase of extreme climate and weather phenomenon,the change of rainfall event frequency and structure will have an important impact on the temporal and spatial distribution of landslide disaster.The study of the influence of climate change on the temporal and spatial distribution of landslides in the future can provide a reference for the work of disaster prevention and mitigation in the region,and has important practical significance.In this paper,taking Baota region of Yan'an as a case study,under three GCM?Global Climate Model?and two typical concentration scenarios of RCP?Rrepresentative Concentration Pathway?,the research area rainfall downscaling prediction data is used to drive the landslide prediction model TRIGRS?Transient rainfall infiltration and grid-based regional slope-stability model?,combined with Rosenblueth point estimation method to solve the uncertainty of soil parameters,the distribution of landslide in Baota region from1979 to 2100 was simulated.Based on the analysis of the simulation results of 13188landslides,the change trend of future rainfall events under different climate models and typical concentration scenarios and the response law of temporal and spatial distribution of shallow landslides to climate change in the future study area are explored.The main conclusions are as follows:?1?The future precipitation in the study area shows different trends under three global climate models,and the corresponding extracted rainfall event frequency and structure that may induce landslides are also different.GFDL-ESM2G model and MIROC5 model showed an upward trend on the annual scale of rainfall,while IPSL-CM5A-MR model showed a slight downward trend.In the whole time series,the annual rainfall of GFDL-ESM2G model in the same period is the largest of the three models,and the trend of increasing rainfall of MIROC5 model is the most obvious.Taking the mean value of the three models,the future precipitation in the study area will increase generally,and the annual precipitation shows a high degree of dispersion.The corresponding extracted rainfall events show the same trend as the annual rainfall,and the frequency,cumulative rainfall and duration of rainfall events that may induce landslides increase with time.In the future,the frequency of rainfall events in the range of cumulative rainfall E?90 mm<E<120 mm?,duration D?120 h<D<180 h?is significantly higher In historical period H?1979-2018?.?2?The future landslides in the study area also show different trends.Under MIROC5model,the change trend of cumulative landslide area is the largest.Under RCP8.5 scenario,the annual average cumulative landslide area in H period rises from 7.10km² to 10.45km² in F₂ period,an increase of 47.17%.Under RCP4.5 scenario,the annual average single landslide area in H period rises from 6.82 km² to 9.77 km² in F₂ period,an increase of43.16%.The increasing trend of shallow landslide area of GFDL-ESM2G model is slightly smaller than MIROC5,and the annual average cumulative landslide area of F₂ period increases by 23.1%and 31.14%respectively under the scenarios of RCP4.5 and RCP8.5.The landslide area of IPSL-CM5A-MR model basically has no upward trend,but has a slight downward trend under the scenario of RCP8.5.Taking the mean of the results of the three models,the annual average cumulative landslides of F₂ period increased by 33.12%and 27.4%compared with H period under the scenarios of RCP4.5 and RCP8.5.In the case of increased rainfall,the change proportion of future accumulated landslide area is larger than that of annual rainfall,while in the case of increased rainfall,the change proportion of future accumulated landslide area is smaller than that of annual rainfall.With the decrease of rainfall,the number of events that can reach the threshold of rainfall induced landslides decreases more obviously,which shows that the future change of landslide area in the study area is more sensitive to the decrease of rainfall.?3?In the future,the dominant factor of the cumulative area growth of landslides in the study area is the increasing frequency of landslides.Compared with the increase of the area of single landslide,the increase of the frequency of future landslide has more obvious influence on the increase of cumulative landslide area.On the time scale of every ten years,there is a high correlation between the number of landslides and the change of the annual average cumulative landslide area.The correlation coefficients of the two groups of data under the same RCP scenario of the same model are more than 0.9,while the change trend of the single field landslide area is not obvious.The median of the landslide area is in a relatively stable state,and the area of the largest landslide event shows a slight upward trend This is due to the climate change leading to the increase of the extreme value of the accumulated rainfall in the study area.?4?The months with more landslides in the study area will be more obvious in the future.Under the RCP4.5 scenario,the strongest response period of landslides to climate change is the middle of this century,and the strongest response period under the RCP8.5scenario is the end of this century.In the future,no matter in the scenario of RCP4.5 or RCP8.5,the three models all show the phenomenon of monthly distribution backward shift within a certain year.