This study compares the performance of three approaches to estimate background seismicity rates, i.e., Reasenberg declustering, Gamma distribution fit, and the Nearest Neighbor Distance. We analyze the respective performance of the methods in volcanically active Hawaii and tectonically active Southern California. Background rate estimation is essential for probabilistic seismic hazard assessment, yet remains challenging due to catalog, instrumentation, and methodological uncertainties. Using both observed and simulated ETAS-based catalogs, the study explores each method’s sensitivity to temporal rate changes. Our results show that Nearest Neighbor and Gamma resolve an increase in Hawaii's background rate following the 2018 Kilauea eruption, whereas Reasenberg is less responsive to such changes. Transient increases in background rates that persist for several years should be considered in time-dependent seismic hazard models for volcanically active regions. We also find that volcano seismicity appears highly clustered in space and time during the 2018 Kilauea eruption, which is notably different from tectonic seismicity.