While the COVID-19 pandemic upended many aspects of life as we knew it, its effects on U.S. public transit were especially dramatic. Many former transit commuters began to work from home or switched to traveling via private vehicles. But for those who continued to work outside the home and could not drive—who were more likely low-income and Black or Hispanic—transit remained an important means of mobility. However, most transit agencies reduced service during the first year of the pandemic, reflecting reduced ridership demand, increasing costs, and uncertain budgets. To analyze the effects of the pandemic on transit systems and their users, we examine bus ridership changes by neighborhood in Boston, Houston, and Los Angeles from 2019 to 2020. Combining aggregated stop-level boarding data, passenger surveys, and census data, we identify associations between shifting travel patterns and neighborhoods. We find that early in the pandemic, neighborhoods with more poor and non-white households lost proportionally fewer riders; however, this gap between high- and low-ridership-loss neighborhoods shrank as the pandemic wore on. We also model ridership change controlling for multiple factors. Ridership in Houston and LA generally outperformed Boston, with built environment and demographic factors accounting for some of the observed differences. Neighborhoods with high shares of Hispanic and African American residents retained more riders in the pandemic, while those with higher levels of auto access and with more workers able to work from home lost more riders, all else equal. We conclude that transit’s social service role elevated during the pandemic, and that serving travelers in disadvantaged neighborhoods will likely remain paramount emerging from it.

据我们所知，虽然 COVID-19 大流行颠覆了生活的许多方面，但它对美国公共交通的影响尤其严重。许多以前的通勤者开始在家工作或改用私家车出行。但对于那些继续外出工作且无法开车的人（他们更有可能是低收入的黑人或西班牙裔）来说，公共交通仍然是一种重要的出行方式。然而，大多数交通机构在大流行的第一年减少了服务，反映出乘客需求减少、成本增加和预算不确定。为了分析这一流行病对交通系统及其用户的影响，我们研究了 2019 年至 2020 年波士顿、休斯顿和洛杉矶各社区的公交车乘客量变化。结合汇总的车站登机数据、乘客调查和人口普查数据，我们确定不断变化的旅行模式和社区之间的关联。我们发现，在大流行初期，贫困和非白人家庭较多的社区失去的乘客比例相对较少；然而，随着大流行的持续，高客流量损失社区和低客流量损失社区之间的差距缩小了。我们还对控制多个因素的客流量变化进行建模。休斯顿和洛杉矶的乘客量总体上优于波士顿，其中建筑环境和人口因素是造成观察到的一些差异的原因。在其他条件相同的情况下，拉美裔和非裔美国人居民比例较高的社区在疫情中保留了更多的乘客，而那些拥有较高汽车通行率和更多能够在家工作的工人的社区则失去了更多的乘客。我们的结论是，在大流行期间，交通的社会服务作用得到提升，而为弱势社区的旅行者提供服务可能仍然是最重要的。