For the 15 years before rotavirus vaccine introduction in 2006, annual rotavirus activity in the United States showed a distinct spatiotemporal pattern, peaking first in the Southwest and last in the Northeast. We modeled spatiotemporal trends in rotavirus laboratory detections from 2009 to 2021. Laboratories reporting to the National Respiratory and Enteric Virus Surveillance System were eligible for inclusion in a given surveillance year (July to June) if ≥1 polymerase chain reaction or enzyme immunoassay rotavirus test per week was reported during ≥26 weeks and totaling ≥100 annual tests. For each laboratory, the season peak was the week with the highest 7-week moving average of the number of rotavirus positive tests during the national season, defined as the period with a 3-week moving average of >10% rotavirus positivity lasting ≥2 consecutive weeks. We input peak week as a continuous variable and the geospatial coordinates of each laboratory into a spherical variogram model for Kriging spatial interpolation. We also created a state-level bivariate choropleth map using tertiles of the 2010-2019 average birth rates and rotavirus vaccine coverage. Following the established biennial trend, the 2010-2011, 2012-2013, 2014-2015, 2016-2017, and 2018-2019 surveillance years had >10% rotavirus positivity for ≥2 weeks and were included in the geospatial analysis. During all 5 seasons included in the geospatial analysis, the earliest peak week occurred in Oklahoma, Arkansas, and the western Gulf coast, a pattern markedly different from prevaccine seasons. These states also had the average lowest rotavirus vaccine coverage and highest birth rate, suggesting that more rapid accumulation of susceptible children drives annual rotavirus season activity. Increasing vaccine coverage remains a key tool in reducing rotavirus burden.