For Space/Time Geostatistics
In Exposure, Disease and Risk Mapping
Room 14, Rosenau Hall
Department of Environmental Sciences and Engineering
School of Public Health
University of North Carolina at Chapel Hill
BMElab director: Marc Serre
The BMElab at the University of North Carolina
The BMElab is a research group of dynamic and innovative researchers and students at the University of North Carolina at Chapel Hill dedicated to the development of geostatistics, spatial regression and risk assessment to (i) map environmental and heath processes, (ii) find drivers of environmental pollutions and adverse health outcomes, and (iii) solve environmental and public heath injustices.
The BMElab has made substantial contributions to (and gets its name after) the Bayesian Maximum Entropy (BME) method of modern spatiotemporal geostatistics. The BMElab has published concepts, mathematical tools and computer programs for the BME methods, including the well-known BMElib package of spatiotemporal Geostatistics used worldwide by scientists and researchers in over 40 different countries.
Facilities and Equipment
The BMElab is currently located in room 14 in the basement of Rosenau Hall. The BMElab features a large computational lab with state of the art computational facilities for many BMElab researchers. The BMElab has both local workstations with space for several researchers. These computational facilities include 10 local PC workstations with up to 4 processors (type Intel Xeon, 3.07 GHz) and 12 GB RAM on a single workstation, 3 printers, and advanced spatial modeling software, including arcGIS and MATLAB, installed locally on each computer to increase computational efficiency. The BMElab is equipped with high speed Ethernet network sockets able to provide network access to up to 16 computers, making it easy for visiting researchers to connect their personal laptops to the internet.
In addition the BMElab has campus wide access to high performance research computing (ITS Research Computing), which provide a world-class computing infrastructure as well as other technology tools and capabilities to support interdisciplinary research in high end computational science. Amongst the high performance computers available are a Linux cluster (KillDevil) with more than 9500 computing cores across 774 servers and large memory ranging from 48 GB to 1 TB per core, and Linux cluster (Kure) with more than 1840 computing cores across 230 blade servers. The facilities also include a large variety of scientific software titles allowing to perform specialized tasks in Science, Geographic Information System (GIS), Mathematics and Statistics and advanced Visualization.