Rainfall variability over northeastern Australia derived from coral luminescence data

Twenty coral cores from the Australian Institute of Marine Science's archive of long Porites spp. coral cores from the Great Barrier Reef were chosen for this study. The cores selected were from sites extending along much of the length of the Great Barrier Reef (between 11°-23°S); were located in nearshore waters, which would be affected by freshwater flood plumes (between 0 and 30km from the mainland) and previous analyses of annual coral growth records for these cores indicated that records started prior to 20th century.

The cores were collected from reefs offshore from Red Wallis Is, Pascoe River, Jeannie River, Conical Rocks, Normanby Island, Dunk Island, Coombe Island, Brooke Island, Great Palm Island, Pandora Reef, Havannah Island, Archeron Island, Magnetic Island, Hook Island, Stonehaven Beach, North Molle Island, Cid Harbour, South Molle Island, Lupton Island and Humpy Island.

Luminescence (excitation k (nm)/emission k (nm) = 390/490) was measured on slices from each core at 0.25 mm intervals with a 2 mm diameter beam using optical luminometry along a major growth axis of each coral core slice.

Monthly instrumental rainfall for northeast Queensland including the catchments of rivers draining into the Great Barrier Reef was sourced from the Australian Bureau of Meteorology for the period 1900-2008. These data were used to calculate annual summer rainfall (October-March) for 1901 to 2008.

17 of the annual coral luminescence series were significantly correlated with the instrumental rainfall record. These 17 series were used to develop a suite of reconstructions of Queensland rainfall (using PCA and linear regression analyses) - using between 17 and a single coral core. The reconstructions differ in the start year (depending on available coral records). All reconstructions were significantly calibrated and verified against the instrumental climate record.

The longest reconstruction based on 3 coral records starts in 1685 and triples the length of the instrumental record. The single core reconstruction starts in 1639 but the 3-coral reconstruction is considered more robust.