MEMBERSHIP IN SCIENTIFIC & PROFESSIONAL ASSOCIATIONS
(1) Member of American Geophysical Union (AGU).
(2) Member of Society of Exploration Geophysicists (SEG).
(3) Member of American Statistician Association (ASA).
(4) Member of Environmental and Engineering Geophysical Society (EEGS).
(5) Member of New York Academy of Science (NYAS).
RESEARCH EXPERIENCE
(1) 01/2005 ─ 12/2006, University of California, Berkeley.
NSF project for hydrogeological inversion ─ Vadose Zone Characterization and Monitoring. This project is to develop a general vadose zone hydrological inversion framework using geophysical methods, particularly the tomographic Ground-Penetrating-Radar (GPR) method.
I investigated the applicability and limitations of two types of ray-tracing models, the curved-ray and straight-ray methods, through forward and inverse modeling studies, and provide guidelines for applications. In the forward modeling study, GPR traveltime calculations are compared between the two ray-tracing methods under various flow conditions, and study the sensitivity of the differences in the first-arrival traveltime calculations to factors including geometry, heterogeneity and data acquisition parameters. In the inverse modeling study, the effects of the differences are compared between these two methods on the quality of the inversion within a MCMC Bayesian inversion framework. Summaries are made regarding to the favorable situations for the straight-ray model where it may work as an alternative of the curved-ray model for shallow subsurface characterization.
(2) 01/2004 ─ 12/2004, University of California, Berkeley.
RPSEA (Research Partnership to Secure Energy for America) ─ Reservoir Characterization in North Sea. This project is to develop approaches for efficient exploration, production, and transportation of natural gas resources below seafloor.
I developed a joint inversion approach for estimating reservoir fluid saturations and porosity, by coupling seismic amplitude versus angle (AVA) and marine controlled-source electromagnetic (CSEM) forward models into a stochastic framework. The approach is demonstrated using data from Troll field site in North Sea. A paper is recently published in Geophysics. This study demonstrates how we can improve subsurface characterization by combining different yet complementary information to better constrain our geophysical or hydrological models; it provides rational evaluation of parameter uncertainty and its consequences and can be used to model data acquisition requirements needed to meet desired resolution of parameter estimates, or whether it is possible at all.
(3) 08/2001 ─ 12/2003, University of California, Berkeley.
Near Surface Water Content Estimation and Monitoring: Investigations within California Vineyards (Mondavi Winery in Northern California), Funded by USDA NRI Soils & Water Programs and NSF. A comprehensive stochastic approach for forward modeling of the state variables in the vadose zone, as well as for inverse modeling of the hydraulic parameters, is developed and applied to a field site in Napa Valley, California. The approach enables us to improve estimates of soil properties such as porosity and hydraulic conductivity, which gives more accurate prediction of the soil moisture variation and the water budget in the soil so that we can provide better guidance for agricultural activities. A paper was published in 2005 in Water Resour. Res.
(4) 09/2000 ─ 07/2001, Princeton University.
Use of NEXRAD Data to Estimate Rainfall Rate and Rainfall Kinetic Energy Flux - This project supported by NSF is to study how to make use of remote sensing radar data to provide rainfall rate and rainfall kinetic energy concurrently, while covering a large area with high spatial and temporal resolution. I analyzed and summarized mechanisms and patterns of the storms in the Goodwin Creek area using TITAN (Thunderstorm Identification, Tracking, Analysis, and Nowcasting System). Meanwhile, I used a GIS imagery processing software ─ ERDAS, and completed the Middlesex County Land Cover Classification Project.
(5) 10/1997 ─ 07/2000, Peking University.
National Key Project 95-11 ─ Influences of Volcanic Activities on Global Warming. This project is to develop exploratory data analysis methods and analyze the influences of different external forcing factors (volcanic eruptions, solar periods, human activities) on global warming.
I generalized the expression of the Natural Orthogonal Polynomial, which can be used for signal processing as well as nonlinear system analyses. I also constructed the tight B-spline wavelet analysis method for signal processing or time series analysis, and helped develop a data analyses software package ─ RCEOF (Rotated Complex Empirical Orthogonal Function Analysis). Papers were published in Acta Scientiarum Naturalium Universitatis Pekinensis and Acta Meteorologica Sinica.
ONGOING PROJECTS AND RESEARCH PROPOSALS
Theoretical development of subsurface characterization approaches
Geophysical survey provides a key tool for hydrologic characterization in the vadose zone (and in some cases deep subsurface as well), yet a characterization approach that can handle the nonlinearities and complexities inherent in vadose zone flow and transport processes and geophysical phenomena, is always a challenge. One difficulty in joint hydrogeological-geophysical site characterization is the question of how to incorporate geophysical information into an inversion approach, which combines both hydrological and geophysical data. The previous practices are often in fact a sequential or iterative process, where errors associated with geophysical inversion or data processing are not quantified or incorporated in the hydrogeological parameter estimation. Another difficulty in hydrogeological inversion is that convergence to an optimum usually cannot be guaranteed. Parameter space in the earth sciences is often characterized by multiple minima. Moreover, it is always a fundamental issue concerning the GPR forward models: a simplified and computationally efficient model probably cannot represent the real world very well; however, an accurate yet complicated model may be practically inapplicable. It is very important to identify applicability and limitations of different geophysical forward models, which usually requires demanding computation.
My research team will develop, optimize, and generalize various vadose zone characterization approaches. These include not only inverse techniques/framework such as nonlinear regression, Maximum Likelihood, MAP, Bayesian, and Hidden Markov Chain methods, and so on, but also geophysical models, for example, the ray-tracing methods against the full-waveform method.
Environmental problems
The application of geophysical techniques and the inversion approaches I developed can also be used for environmental remediation studies for the vadose zone and/or aquifers. However, in addition to geophysical models and flow and transport models, approaches are needed to elucidate reactions associated with coupled biological, chemical, and hydrologic processes. Flow and transport models such as GMS, TOUGH2, and Hydrus2D, will be adopt and improved to have reactive transport models, so that we can accurately represent those above processes and predict their roles in contaminant remediation over multiple scales and in heterogeneous environments. In certain cases, the society and government pay more attention to larger scale environmental problems. For example, when studying the salinization problem which is encountered in most areas of the world, people may use local scale flow and transport models to study the fate of salt discharged by industrial, agricultural, or municipal human activities, but it is also important to develop a regional model (e.g., for a hydrologic region such as a river basin) to study the salt budget within a region and evaluate the relative importance and social effects of different salinity sources. This type of study will need to bring together people with different background and organizations.
My research team will collaborate with colleagues from multidisciplinary programs, including geophysics, geochemistry, hydrogeology, and geomicrobiology to develop such tools and approaches. Through integration of experimental, numerical and theoretical studies, our research team will develop approaches for detecting and monitoring changes in physical, chemical and biological properties as a function of space and time and for evaluating the effects of such changes at different spatial scales.