Collaborating with scientists in Argonne National Laboratory, I introduced an innovative approach to wind modeling that simultaneously models wind speed and wind direction. The key motivation behind this work was to address algorithmic bias arising from the dominance of wind speed information while providing much-needed attention to wind direction, which has been historically overlooked due to its circular nature. By developing this novel framework, we achieved a more comprehensive understanding of wind behavior, enabling us to characterize both the wind direction and wind speed distribution accurately. Our work has far-reaching implications in, for example, air pollution and wind energy modelings, and underscores the importance of addressing algorithmic bias in scientific research.
The 45th annual meeting of the SIAM Southeastern Atlantic Section will be hosted at Virginia Tech. The meeting will take place on March 25-26, 2023, and will contain the usual combination of plenary talks, minisymposia sessions, and contributed talks and posters.
MS description: As the amount of available computing resources have increased over the past decades, so too has the need for new methodologies for processing ever-larger amounts of data and simulating physical models with ever-greater complexity. In this minisymposium, we will discuss novel mathematical and statistical techniques for tackling these challenges in a variety of environmental and earth sciences applications. The methods we discuss will touch upon areas such as uncertainty quantification, model correction, and extreme-scale numerical methods.
11:30-12:00 Qiuyi Wu, University of Rochester, A conditional approach for joint estimation of wind speed and direction under future climates