Baylor Statistics Seminar Series

Fall 2017 will bring many exciting and notable speakers to Baylor’s Department of Statistical Science.  A full list can be viewed here.  Be sure to mark your calendar for Thursdays at 3:30 in Marrs McLean so that you don’t miss any talks.


ENVR Early Investigator Award

The American Statistical Association awarded five individuals with the Early Investigator Award at the Joint Statistical Meetings in August 2017. They are Matthias Katzfuss of Texas A&M; Ephraim Hanks of Penn State; Elizabeth Mannshardt of the EPA; Ying Sun of KAUST, and myself. Note that in the picture below, Matthias Katzfuss is not present, and Montse Fuentes (far left) received the Distinguished Achievement Award.

Radiosonde Outlier Detection

A paper with Ying Sun of KAUST and former Master’s student Joshua Browning was recently accepted for publication in Environmetrics in which a new method for identifying outliers in skewed, heavy-tailed, multivariate distributions is developed.  The method is applied to wind observations from historical radiosonde launches in the National Center for Atmospheric Research archive.  The global historical radiosonde archives date back to the 1920’s and contain the only directly observed measurements of temperature, wind, and moisture in the upper atmosphere, but they contain many random errors. Most of the focus on cleaning these large datasets has been on temperatures, but winds are important inputs to climate models and in studies of wind climatology.



NSF PFI:BIC Award Granted

A multidisciplinary team, led by Professor Tzahi Cath of Colorado School of Mines, has received a $1 million award from the National Science Foundation to develop an innovative monitoring and control system for small wastewater treatment facilities.  Professor Hering is a co-PI on this award and will be working to develop quality control methods that will learn from the massive amount of data produced from the Mines decentralized wastewater treatment testbed in a smart and efficient way.  A pilot method is already in place in which various dimension reduction methods were tested both in simulation and for real scenarios, see Kazor et al. (2016)



Mines Newsroom Announcement




Land Cover Fusion

A paper has just been accepted to Environmetrics with first author and PhD student Nicolas Rodriguez-Jeangros receiving an Honorable Mention for the work by the American Statistical Association’s Section on Statistics and the Environment.  In this work, we fuse multiple land cover (LC) products together to produce a single “super-product.”



Short Description:  The type of land cover (LC) present on the Earth, such as open water, forest, or grasslands, drives many environmental processes and is thus important to track over time.  In recent decades, multiple maps depicting LC have been produced by classifying images obtained from satellites orbiting the Earth. In these maps, each pixel is assigned a specific LC category, and the size of the pixel is referred to as the spatial resolution of the map.  Despite vast improvements in satellite technologies, there is no LC map that has the spatial resolution and the frequency in time required for long-term environmental studies.  Consequently, in this paper, the authors propose the Spatiotemporal Categorical Map Fusion (SCaMF) methodology to fuse multiple existing LC maps into a single map with any spatial and temporal resolution chosen by the end user. SCaMF is illustrated by fusing six LC maps over the Rocky Mountains.  The fused map is generally able to reproduce known LC features even better than some of the individual LC maps, and as a byproduct of the method, the uncertainty in the fused map is also produced.  Given the enormous size of the LC maps, the method is computationally intense, so experiments that justify some computing short-cuts are performed.