Weiwei Zhan

Weiwei Zhan

Assistant Professor at University of Central Florida

Blog

2022

Geomorphological Precursors for Landslide Early Detection

Early detection of rockslides at high-elevation and well-vegetated slopes remains challenging. This study used satellite and unmanned aerial vehicle (UAV) optical remote-sensing (ORS) images to track evidence of slope deformation and examine potential geomorphological precursors of five large rockslides in China. All the surveyed landslides had cracks or scarps and rockfalls within the landslide source area before the onset of rapid sliding. These precursors can be identified in ORS images taken several years or decades before the rapid slope failure, which provides sufficient time for the landslide early detection in practice. Local topography affects the spatial locations of cracks or scarps. Rockfalls within the landslide source area tend to locate at “key blocks” where slope mass provides forces resisting sliding. The rockfall area ratio, defined as the accumulated area of rockfalls over the landslide source area, ranged from 0.33 to 0.92 before rapid slope failure. The landslides developed on anti-dip and igneous rock slopes show a more significant rise of rockfall area ratio before the slope failure than the landslides on dip slopes. Given the broad availability of ORS data, this study could shed light on the ORS- based landslide early detection and landslide kinematics study.

Open-source Geospatial Liquefaction Database and Associated Tools

OpenLIQ is a geospatial database of global seismic liquefaction hazards. The OpenLIQ database currently includes ~290,000 samples (~63% of them are liquefaction points) and 19 geospatial predictors for 51 worldwide earthquakes. In addition, the OpenLIQ project includes a semi-automatic GIS workflow that can automatically conduct geospatial sampling of liquefaction and non- liquefaction points from liquefaction inventories (georeferenced liquefaction points and/or polygons), extract candidate geospatial predictors, and generate maps for quality check and visualization purposes. This workflow only requires limited manual inputs (e.g., incomplete or complete liquefaction inventory, and raster data for geospatial predictors). The GIS workflow will be used to process new liquefaction inventories, and the OpenLIQ database will be used to advance applications of different machine learning and uncertainty quantification methods in global geospatial liquefaction modeling.

Evaluating the Accelerogram-based Liquefaction Detection Method Using Continuous Earthquake Recordings

I further evaluate the performance of the new method using continuous earthquake recordings at four seismic stations. The results suggest the method performs better at the soft-soil sites than at the stiff-soil sites and tends to make false negative predictions for very strong earthquakes. This study highlights the applicability of the new accelerogram-based method to regions with continuous earthquake monitoring.

Site Response Temporal Change Evaluation

To examine if the large interevent variability of seismic site responses at liquefiable sites is reducible, I investigated the temporal variability of seismic site responses at a liquefiable site using the sliding window spectral ratio method. I used stacked images of surface-to-borehole spectral ratios to track changes in seismic site response characteristics at temporal scales varying from seconds to decades. I observed that resonance frequencies dropped by ~70% during the mainshock of a liquefaction-triggering strong earthquake (indicating significant soil softening behaviors) and had an ephemeral increase by ~40% due to the strong soil dilatant behavior. In addition, I observed that the resonance frequency recovered after a strong earthquake at two phases with different recovery speeds and fluctuated along the seasonal changes of hydrologic conditions in the shallow subsurface.

Site Response Nonlinearity Assessment

Site response nonlinearity describes the change of site responses from weak to strong events, shedding light on nonlinear soil dynamic behaviors. I studied the site response nonlinearity at four liquefiable vertical array sites (with liquefaction surface manifestation during at least one earthquake; the site has at least two ground motion sensors, one at the surface and one at the depth). I processed thousands of ground motion waveforms and computed surface-to-borehole spectral ratios (frequency-wise site amplification curves). I used shape parameters to quantify site response nonlinearity and developed site-specific regression models to evaluate the site vulnerability to nonlinear site responses. I found that the liquefaction could enhance the amplification of low-frequency and very-high-frequency ground motions, and these liquefaction effects vary from site to site.

Geospatial Surrogate Models for Site Response Complexity Assessment

The term site response complexity describes the discrepancy between observed site response (i.e., empirical transfer function ETF) and theoretical site response (i.e., theoretical transfer function TTF) predicted using the widely-used one-dimensional site response methods (i.e., SH1D). I derived TTF for ~700 vertical array sites using SH1D simulations and compared them with ETF derived from a large ground motion dataset. I observed noticeable trends between site response complexity and topography: Good-matching sites are mostly located in flat sedimentary basins (where the SH1D assumptions are most valid), and poor-matching sites are often located near mountain/basin edges, and high-interevent-variability sites are located within mountainous areas.

ML-based Ground-Motion Modeling

Ground-motion models (GMM) estimate the ground shaking intensity measures given a set of the earthquake source, path, and site variables, which provides the basis for any seismic risk assessment. In several ongoing works, I improved the accuracy of GMMs and discovered new site variables by integrating nonparametric machine learning and feature selection techniques. I observed that geospatial environmental data contributed to ground-motion modeling.

2021

Parameter Sensitivity Analysis of the New Accelerogram-based Method

The effectiveness of accelerogram-based liquefaction triggering procedure is relying on the two frequency-related parameters of surface ground motions (RL and MIFr). We selected eight factors to represent source, path, and site effects, and applied the one-way Analysis of Variance (ANOVA) method to rank the correlation between these eight factors and RL and MIFr separately. The results suggest that liquefaction generally has a dominant effect on RL and MIFr. However, earthquake magnitude and epicentral distance could have similar effects on RL or MIFr, respectively, as liquefaction does, which needs to be accounted for in applications and improvements of the accelerogram-based liquefaction triggering procedure.

Numerical Modeling of Landslide Dynamic Processes

To further investigate the extraordinary mobility of rock avalanches, I employed the discrete element method to simulate the entire process of an earthquake-induced landslide (landslide initiation, propagation, and deposition). Through the case study, I observed that the progressive rock fragmentations and violent interparticle collisions significantly contributed to the extraordinary high mobility.

Geotechnical Site Characterization of Deep Sediment Soils

The dynamic penetration test (DPT) and the Menard pressuremeter test (PMT) have been widely used in geotechnical survey of deep soils for megadam foundations in western China. The DPT measures are not well utilized due to the lack of correction factors and of empirical relationships for deep soils. I developed a method for DPT correction and empirical relationships between the corrected DPT blow counts with deformation modulus and bearing capacity of deep soils.

Accelerogram-based Real-time Liquefaction Hazard Detection

I developed a new accelerogram-based method and a worldwide database with paired liquefaction observations and ground motion records. The new method consisted of a logistic regression model and an automated signal processing procedure that extracts two physics- informed features for liquefaction detection from the raw accelerograms using different signal processing techniques. I also evaluated the performance and sensitivity of the accelerogram-based liquefaction detection method to different earthquake source, path, and site conditions.

2018

Geomorphological Surveys of Successive Landslide Dams

Earthquakes can not only cause immediate slope failure but also affect slope stability and landscape evolution in the short and long terms. Through the extensive field surveys in the earthquake-impacted area, I found an interesting near-fault slope unit where at least three landslide events occurred and dammed the river in the past. From the hazard assessment and prevention perspective, it is important to explore landslide hazard management at such successive sliding and dam sites. I collected multi-temporal remote sensing images and terrain data through unmanned aerial vehicle (UAV) surveys and conducted geomorphological and geological surveys. I concluded that this site is a good example of multiple reactivations of a prehistorical giant landslide under different triggering conditions: a reactivation of an older landslide during a strong earthquake, and a second reactivation during a rainfall event several years later. Meanwhile, I proposed a generalized framework for assessing landslide dam cascading hazards, including landslide volume estimation, empirical landslide dam stability evaluation, GIS-based barrier lake volume estimation, and dam-breach flooding parameter estimation.

Geological and Geotechnical Surveys of a Deep-Seated Rockslide

The 2008 Ms 8.0 Wenchuan earthquake triggered more than 200 thousand landslides, including hundreds of catastrophic deep-seated rockslides. The Among these catastrophic rockslides, the Daguangbao (DGB) landslide is the largest landslide triggered by the Wenchuan earthquake (28°38′16.93′′ E, 104°06′02.35′′ N). The DGB landslide moved the slope mass with volume of 1.2 billion cubic meters to covered an area of 7.8 km2, and formed a landslide dam nearly 600 m high. We had done detailed geological, geophysical, and geotechnical surveys to study the failure mechanism of this giant landslide. I observed that a 3-meter-thick fault gouge was well extended along the sliding bed and with a maximum burial depth of ~400 m. Geotechnical laboratory experiment results suggested that the weakest sublayer within the fault gouge consists of silt soils with strain hardening and liquefaction behaviors. I concluded that slippery fault gouge and strong earthquake-induced shaking led to the giant landslide.

Landslide Inventory Development and Spatial Pattern Analysis

Landslide inventory plays an important role in regional landslide susceptibility assessment and uncovering landslide formation mechanisms. I studied the remote-sensing-based landslide mapping and landslide inventory development for earthquake-impacted areas following strong earthquakes. I conducted statistical analyses to investigate the spatial distribution pattern of coseismic landslides with respect to different seismic, terrain, and geological factors. In this setting, I found high landslide density at lower slope sections where the landform evolves from upper, broad valleys to lower, deep-cut gorges. The spatial distribution of the coseismic landslides did not seem correlated to the location of any known active faults. In contrast, I revealed that a previously unknown blind fault segment is the plausible seismogenic fault given the characteristic landslide density break across that fault line.

2017

Empirical Modeling of Landslide Mobility

Landslide-impacted area is another crucial element for individual landslide hazard assessment for slope units close to infrastructure and community. Rock avalanche is a high-mobility type of landslide that can travel a few to tens of kilometers and move a mass of a million to billion cubic meters.