An optimal approach for balancing health and economic challenges has been the epicenter since the pandemic, while differentiating the impacts of various levels of intervention on homogeneous units is challenging. The risk-rating system China developed to categorize COVID-19 risk for localities and match them with corresponding tactics can serve as an ideal alternative. Apart from this, human mobility is a key mechanism through which activities emerge and viruses spread, bringing both merits and hurdles to urban economies. For these reasons, this paper examines the causal effects of medium vis-à-vis high COVID-19 risk designated to urban sections on intra-city mobility during the Zero-COVID period, using a dynamic difference-in-differences approach. It then extends to assess local travel trajectories of 368 Chinese cities during the pandemic and post-pandemic periods within a non-linear time-varying latent factor framework. Utilizing the latest Baidu Mobility data and national risk-level data at daily and weekly frequencies, we find that interventions devised for high risk tend to have an unexpectedly far-reaching impact particularly on recreational travel beyond affected urban sections. People reacted to the emergence of high risk eight days in advance, and the striking impact lasted for around four weeks after the intervention starts. We further provide stylized facts on stratified mobility dynamics, suggesting that a) western cities are less resilient, and b) regional disparities tend to widen during the Zero-COVID period but shrink alongside the reopening.