The main limiting factor in obtaining precise and accurate uranium-series (U-series) ages of corals that lived during the last few hundred years is the ability to constrain and correct for initial thorium-230 ( 230Th 0), which is proportionally much higher in younger samples. This is becoming particularly important in palaeoecological research where accurate chronologies, based on the 230Th chronometer, are required to pinpoint changes in coral community structure and the timing of mortality events in recent time (e.g. since European settlement of northern Australia in the 1850s). In this study, thermal ionisation mass spectrometry (TIMS) U-series dating of 43 samples of known ages collected from living Porites spp. from the far northern, central and southern inshore regions of the Great Barrier Reef (GBR) was performed to spatially constrain initial 230Th/ 232Th ( 230Th/ 232Th 0) variability. In these living Porites corals, the majority of 230Th/ 232Th 0 values fell within error of the conservative bulk Earth 230Th/ 232Th atomic value of 4.3±4.3×10 -6 (2σ) generally assumed for 230Th 0 corrections where the primary source is terrestrially derived. However, the results of this study demonstrate that the accuracy of 230Th ages can be further improved by using locally determined 230Th/ 232Th 0 values for correction, supporting the conclusion made by Shen et al. (2008) for the Western Pacific. Despite samples being taken from regions adjacent to contrasting levels of land modification, no significant differences were found in 230Th/ 232Th 0 between regions exposed to varying levels of sediment during river runoff events. Overall, 39 of the total 43 230Th/ 232Th 0 atomic values measured in samples from inshore reefs across the entire region show a normal distribution ranging from 3.5±1.1 to 8.1±1.1×10 -6, with a weighted mean of 5.76±0.34×10 -6 (2σ, MSWD=8.1). Considering the scatter of the data, the weighted mean value with a more conservative assigned error of 25% (i.e. 5.8±1.4×10 -6) that encompasses the full variation of the 39 230Th/ 232Th 0 measurements is recommended as a more appropriate value for initial 230Th corrections for U-series dating of most Porites samples from inshore regions of the GBR. This will result in significant improvement in both the precision and accuracy of the corrected 230Th ages related to those based on the assumed bulk Earth 230Th/ 232Th 0 value of 4.3±4.3×10 -6. However, several anomalously high 230Th/ 232Th 0 values reaching up to 28.0±1.6×10 -6 occasionally found in some coral annual bands coinciding with El Niño years imply high 230Th/ 232Th 0 sources and highlight the complexities of understanding 230Th/ 232Th 0 variability. For U-series dating of young coral samples from such sites where anomalous 230Th/ 232Th 0 values occur, we suggest replicate dating of multiple growth bands with known age difference to verify age accuracy. © 2011 Elsevier Ltd.