Spatial and temporal variation in abundance and recruitment of burrowing ghost shrimp Trypaea australiensis was examined across 3 south-eastern Australian estuaries using a hierarchical sampling design, over a 2 yr period. We tested the hypothesis that abundances of shrimp were different between plots (10s to 100s of metres apart), sites within estuaries (kilometres apart), estuaries (100s of kilometres apart) and through time. More frequent sampling at 1 site also examined temporal variation at scales of months, seasons and years. Another aim was to investigate the reliability of using counts of burrow openings to indirectly measure the relative abundance of T, australiensis. Significant and interactive variability was detected at the scale of plots for the mean numbers of shrimp and recruits. Components of variation, however, suggested there was patchiness in abundance at all spatial scales within estuaries, particularly between replicates separated by metres. Indeed, between-replicate variance was greater than for any of the temporal scales examined. Despite this small-scale patchiness, patterns of abundance and recruitment were generally consistent across broad geographic areas. Numbers of shrimp generally increased throughout spring and summer periods, and recruitment also occurred during these times. The patterns of variation observed in this study highlight the importance of including appropriate scales of sampling in future monitoring studies of T. australiensis and in experiments concerned with detecting the effects of bait harvesting on populations of shrimp. We also conclude that counting burrow openings is not a reliable predictor of the relative abundance of T. australiensis.