Quantifying systematic drifts in UCNτ data using Gaussian processes

Measurements of the neutron lifetime provide a useful way to probe the Standard Model by helping to determine V_ud. The UCNτ experiment has provided the most precise measurement of the neutron lifetime. Located at the Los Alamos Neutron Science Center, the experiment uses a magneto-gravitational bottle to store ultracold neutrons before counting them in-situ with a Zinc Sulfide alpha detector coated with ¹⁰B. During the running of the experiment in 2021, there were some large, unexpected drifts in the data due to water being introduced into the neutron trap as part of a systematics test. These drifts could not be properly normalized by the established analysis method, and so this data set was not used in the determination of the neutron lifetime. I will present an alternative analysis method using Gaussian processes that better accounts for the drifts in the data and describes the systematic uncertainty introduced by these drifts. This method better quantifies this uncertainty and predicts a larger systematic uncertainty associated with the data taken when water was in the trap compared to data taken during normal running conditions.