Biography
Russell Limber "Russ" is a senior data scientist in Knoxville, Tennessee. He was previously a PhD student in Data Science Engineering with the University of Tennessee; Bredesen Center. Originally from Ithaca N.Y., Russ attended the State University of New York, College of Environmental Science and Forestry (SUNY ESF) where he received his B.S. in Environmental Science while minoring in Math, Physics and Applied Statistics. He worked for two years with Dr. Mitchel Soderberg in the Syracuse University High Energy Physics Department as a research assistant, collecting circuitry data for the FermiLab's Deep Underground Neutrino Experiment (DUNE). While at SUNY ESF, he also worked for Dr. Russell Briggs to help identify the optimal tree species for Christmas tree farmers in Upstate NY to grow in order to manage blight and changing seasonal weather patterns. He went on to receive his M.S. in Data Analytics through Western Governors University. He currently works for Dr. Jitendra Kumar at Oak Ridge National Laboratory where he uses remote sensing and geospatial data alongside machine learning and statistical analysis, to analyse Arctic river ice breakup, forest structural changes in the Tropics and fire danger assessment in California. His dissertation revolves around his work on Arctic river ice breakup and is titled "Modeling River Ice Breakup Throughout Interior Alaska." His work primarily entails modeling river ice breakup timing through the use of a Long Short Term Memory Model (LSTM), a type of sequential deep learning algorithm, using meteorological inputs. He has also applied this methodology to meteorological inputs provided by various CMIP6 model scenarios in an effort to forecast how river ice breakup patterns are likely to change in Alaska USA under different climate change scenarios. The analysis also includes methods for identifying causality, such as an analysis of the interaction of precipitation and temperature over time on influencing river ice breakup. In addition, Shapley value analysis was used to identify influence among variables over time, after applying the LSTM. The motivation for this work stems from a desire to help undeserved and overlooked native communities. Better understanding future river ice breakup patterns will help native communities of the region (as well as more recent inhabitants) better prepare for a changing landscape, that will inevitably alter their transportation, food supply, culture and overall way of life.
