Improving the Navys Numerical Atmospheric Predictions by Advancing Novel Transformative Low Cost Forecast Ensemble Creation Methods
Abstract
ABSTRACT # APPROVED FOR PUBLIC RELEASEThe Navy#s tactics must account for atmospheric conditions, particularly visibility conditions. I seek to amplify the Navy#s military edge by improving numerical atmospheric predictions (NAPs). Transformative low-cost methodsto construct large ensembles of high-resolution NAPs will improve NAPs through boosting the ensemble-mediated conversion of observations into NAP corrections. My team and I will build on my past promising work to investigate a transformative class of low-cost ensemble construction methods: Probit-space Ensemble Size Expansion (PESE; pronounced #peace#). PESE efficiently and scalably converts small NAP ensembles into large NAP ensembles through incorporating the users# expertise on forecast ensemble statistics. This incorporation is flexible, novel, and transformative. Furthermore, the resulting large NAP ensembles will have features consistent with the predicted weather conditions (i.e., flow-dependency). Finally, PESE avoidsthe pitfalls of alternative methods that construct largeNAP ensembles from small NAP ensembles. In this project, we will investigate (Objective 1) the realism and statistics of PESE-created NAP ensembles, (Objective 2) PESE#s impacts on numerical aerosol prediction, and (Objective 3) PESE#s potential impacts on numerical weather prediction. A series of three increasingly sophisticated PESE methods will be tested in Objective 1. The most sophisticated PESE method is an assumption-free method that incorporates not only the users# expertise on univariate forecast distributions, but also their expertise on spatial forecast decorrelation lengths. In Objectives 2 and 3, we will first explore the statistical distributions and the complicated nonlinear multivariate statistical relationships embedded in large ensembles of high-resolution numerical weather and aerosol models. We will then perform data assimilation experiments with PESE methods to assess PESE#s impacts on numerical aerosol prediction and numerical weather prediction. Our project will not only improve the Navy#s NAPs, but also greatly expand our understanding of multivariate forecast statistics.Our project will be executed using case studies over a Navy-identified strategically important maritime trade chokepoint that is experiencing increased security threat: the Malacca Strait (NAVPLAN 2022). We will use high-resolution NAP models and ensemble data from the Navy Global Environmental Model (NAVGEM). Furthermore, we will produce a PESE-containing Python package (PyPESE) that will be transferred into the NAP workflows of the Naval Research Laboratory (NRL) and the Fleet Numerical Meteorology and Oceanography Center (FNMOC). Finally, because our work is highly relevant to the Navy#s NAPs,three NRL scientists (Dr. Elizabeth Satterfield, Dr. Edward Hyer and Dr. Juli Rubin) have agreed to collaborate with us. Our collaboration will enable my team to acquire the necessary NAVGEM data and to smoothly transfer PESE into the Navy#s NAPs.
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- Nov 26, 2024
- Source ID
- N000142512023
Entities
People
- Man-yau Joseph Chan
Organizations
- Office of Naval Research
- Ohio State University
- United States Navy