Successfully eradicated invasions are ideal opportunities for understanding the factors governing biological invasions and developing robust management strategies should the same, or similar, organisms again invade. We used geospatial analyses and habitat suitability modeling to reconstruct the spatiotemporal dynamics of an invasive vineyard pest, the European grapevine moth (Lobesia botrana [Denis & Schiffermüller]), in northern California. L. botrana detections were most strongly autocorrelated at local spatial scales (≤250 m) and remained clustered up to ~10 km. Generalized linear model, boosted regression tree, and random forest modeling methods performed well in predicting habitat suitability for L. botrana; annual mean temperature, elevation, and distance to the nearest road were identified as important predictors. Hotspots in L. botrana occurrence were spatiotemporally dynamic, yet habitat suitability was less important than purely spatial effects in explaining hotspot persistence. Our results indicate that local regulatory response to novel L. botrana detections was appropriate; 500 m treatment zones around detections are sufficient given the apparent propensity for very local movement by L. botrana. Our results also confirm the role of anthropogenic effects in L. botrana spread and support the establishment of quarantine procedures to limit human-mediated dispersal. Lastly, ensemble predictions provide a fine-scale measure of relative risk for a portion of northern California in the event of future L. botrana introductions.