Dan Hodyss

Background

About Dan Hodyss

I design and build traditional statistical and ML/AI modelling systems for large dimensional problems in the geosciences. This includes both state estimation as well as predictive/forecast models from both a deterministic as well as probabilistic perspective using both classical methods as well as ML/AI. Presently, my focus is on diffusion models for prediction and data assimilation. Please see the following selected articles for further information:Hodyss, D, 2012: Accounting for skewness in ensemble data assimilation. Mon. Wea. Rev, 140, 2346-2358.Hodyss, D et al, 2014: Inducing tropical cyclones to undergo Brownian motion: A comparison between Itô and Stratonovich in a numerical weather prediction model. Mon. Wea. Rev. 142, 1982-1996.Hodyss, D, E. Satterfield, et al, 2016: Inaccuracies with multimodel postprocessing methods involving weighted, regression-corrected forecasts. Mon. Wea. Rev, 144, 1649-1668.Hodyss, D, W.F. Campbell, J. Whitaker, 2016: Observation-dependent posterior inflation for the ensemble Kalman filter. Mon. Wea. Rev, 144, 2667-2684. Hodyss, D, J. Anderson, N. Collins, W. Campbell, A. Reinecke, 2017: Quadratic Polynomial Regression using Serial Observation Processing: Implementation within DART. Mon. Wea. Rev, DOI:/10.1175/MWR-D-17-0089.1Hodyss, D. and M. Morzfeld, 2025: Using diffusion models to do data assimilation. arXiv:2506.02249. https://doi.org/10.48550/arXiv.2506.02249