Total: 1
Current pulsar timing arrays could detect gravitational wave memory, which would provide a unique test of general relativity. In order to optimize our searches for gravitational wave memory in pulsar timing arrays, we integrate the optimal frequentist statistics for GW memory into the \texttt{enterprise} code suite. We additionally test a noise marginalized version of this statistic, and compare it's effectiveness to the simpler maximum posterior method. We perform these tests on simulated pulsar timing array with red and white noise content, as well as the NANOGrav $15$-year data set. We find no significant variations in our detection statistics between the noise-marginalized and the maximum-posterior methods for the simulated datasets, and only small variations in the NANOGrav $15$-year comparisons. We conclude that for realistic PTAs the noise-marginalization method could be useful for assigning a false alarm probability to Bayesian results, though the maximum posterior method is sufficient for identifying potentially detectable signals.