Clustering improvements for the Run-2 and Run-3 precession frequency analysis in the Muon g-2 experiment

The Fermilab Muon g-2 experiment, which measures the muon anomalous magnetic moment a_μ, published analysis of Run-1 data in April 2021 to a precision of 460 ppb.  This result confirmed the value of a_μ measured at the earlier Brookhaven experiment, and the averaged experimental value stands in tension with the Standard Model prediction to 4.2 σ.  Analysis of the Run-2 and Run-3 data is now underway.  As the statistical uncertainty decreases, uncertainties on systematic effects are being reduced to reach the ultimate systematic uncertainty goal of 70 ppb.  When a muon decays, the resulting positron is detected by a calorimeter, where its time and energy is reconstructed. One important source of uncertainty for the Run-1 precession frequency (ω_a) analysis was pileup, which occurs when two decay positron hits are reconstructed as a single positron.  In this talk, I will present an improvement to the algorithm that clusters calorimeter hits in order to reconstruct the positron information.  By incorporating detector time resolution information into the algorithm, pileup is reduced by a factor of four.  I will present preliminary blinded ω_a results and uncertainties for Run-2 and Run-3 using this improved reconstruction.