KAUST Ph.D. student Ghulam Qadir received a best poster award from the from the Italian Environmetrics Society and the International Environmetrics Society in July. Photo courtesy of Ghulam Qadir.
By David Murphy, KAUST News
Ghulam Qadir, a third-year Ph.D. student in KAUST Associate Professor Ying Sun's Environmental Statistics research group, recently received a best poster award from the Italian Environmetrics Society (GRASPA) and the International Environmetrics Society (TIES) at the GRASPA 2019 conference held from July 15 to 16 in Pescara, Italy.
Qadir's winning poster—titled "Estimation of Spatial Deformation for Non-stationary Processes via Variogram Alignment"—was based on the student's first Ph.D. research project at KAUST. It is co-authored by his Ph.D. supervisor Sun and Associate Professor Sebastian Kurtek, who is based in the Department of Statistics at the Ohio State University, U.S.
"I won the award due to the scientific quality, structure and organization of the poster presentation," Qadir stated.
At KAUST, Qadir's main research interest centers on developing covariance models for multivariate nonstationary random fields with applications to environmental data. His work involves developing models for univariate as well as multivariate spatial data.
KAUST Associate Professor Ying Sun (pictured here) supervises Ph.D. student Ghulam Qadir, who is part of her Environmental Statistics research group. File photo.
"This convenient working assumption is often unrealistic for statistical modeling of environmental processes. Therefore, in this work, we have developed a method to avoid the stationarity assumption by transforming the spatial domain of the process by using the tools from Elastic Functional Data Analysis. The proposed method allows us to model the nonstationary spatial covariance functions of the univariate spatial data," Qadir explained.
In the future, Qadir hopes to develop statistical models that can uncover many subtle features, such as nonstationarity, periodicity, asymmetric cross-spatial dependence and non-trivial cross-spectral features of any spatially referenced data.
"These models will have a wide range of applications, such as, for instance, in modeling environmental and atmospheric processes for inference as well as predictions at unobserved locations," he noted.