Wildland fire smoke exposures present an increasingly severe threat to public health, and thus there is a growing need for studying the effects of protective behaviors on improving health. Smoke Sense, a citizen science project, provides an interactive platform for participants to engage with a smartphone app that records air quality, health symptoms, and behaviors taken to reduce smoke exposures. We propose a new, doubly robust estimator of the structural nested mean model that accounts for spatially- and time-varying effects via a local estimating equation approach with geographical kernel weighting. Moreover, our analytical framework is flexible enough to handle informative missingness by inverse probability weighting of estimating functions. We evaluate the new method using extensive simulation studies and apply it to Smoke Sense survey data collected from smartphones for a better understanding of the relationship between smoke preventive measures and health effects.