The tropics are a critical part of climate system, modulating global temperature and precipitation patterns on a variety of timescales. Yet, natural climate variability across much of the tropics remains poorly characterized due to the dearth of instrumental observations prior to 1970s. Massive corals are now routinely used to reconstruct past tropical climate variability, however the accuracy of these records have been questioned as corals growing on the same reef can produce very different geochemical signals. Diagenetic alteration, prevalent in both modern and fossil corals, is another factor that significantly compromises the accuracy of coral records. To quantify intercolony reproducibility at Palmyra Atoll (6˚N,162˚W), we compare oxygen isotope (δ18O) and Sr/Ca records across in five overlapping modern corals. Our results demonstrate that Palmyra corals reconstruct similar monthly to interannual climate variability, however absolute δ18O and Sr/Ca are systematically offset among different corals. These intercolony offsets have a negligible impact on climate records derived from modern corals, however reconstructions based on fossil corals may be unable to resolve mean climate changes < ~1.4˚C. We present an alternative approach to using fossil corals, where overlapping cores are used to constrain climate-driven differences in mean δ18O and Sr/Ca between two time periods, providing more reliable estimates of mean climate change. Lastly, we explore the potential of using secondary ion-probe mass spectrometry (SIMS) to extract reliable climate records from altered fossil corals. By targeting only a few pristine spots within the coral skeleton, SIMS analyses reproduce bulk Sr/Ca estimates and yield credible Sr/Ca-based SST records from three modern coral samples. Applying this technique to an heavily altered fossil coral where bulk estimates yield improbably cool temperatures, SIMS allows us to bypass diagenesis and extract a more reliable Sr/Ca-based SST record. Overall, these results demonstrate that SIMS is a powerful tool that can be used to verify the accuracy of reconstructions from fossil corals.