UV-spectral luminescence scanning: technical updates and calibration developments

Abstract

Spectral luminescence scanning (SLS) is a novel technique that uses a UV light source and line-scan camera to generate photoluminescence images of carbonate materials, such as corals. The camera in the Avaatech XRF core scanner records luminescence signals in three spectral domains of visual light, providing Red, Green and Blue (RGB) luminescence intensity data. Spectral luminescence Green/Blue ratios (G/B) of coral skeletons have previously been employed as a proxy to reconstruct river runoff. Prior G/B reconstructions have been formulated based on indirect G/B-runoff relationships (e.g. modelled discharge), as coral cores were drilled from regions where reliable long-term instrumental data were lacking, i.e. Madagascar. Here, we provide additional evidence that G/B is directly related to runoff by comparing instrumental data with four coral cores from the Keppel Islands, Australia; a region where instrumental data are both reliable and plentiful. A four coral core G/B-composite record was found to correlate significantly with precipitation, stream height level and stream discharge rate over a 53 year period. The strongest G/B relationship observed was with stream discharge rate, which explained 37 % of the total interannual variance of G/B. Modifications to the Avaatech XRF core scanner are ongoing. Here, we describe the use of a new commercially available light cut off filter (Schott GG 455 nm long pass filter) to block reflected Blue light from the UV light source, and compare it with the previously employed 450 nm filter. Conversion of the 450 nm filtered data to 455 nm filtered data was carried out by a linear correction function based on major axis regression, providing statistically similar G/B data for monthly resolved coral records, as well as offering greater insights into the nature and cause of skeletal luminescence. In addition to modifications to the scanner, developments in the sample preparation are described here. We show that when treating coral cores with NaOCl to remove organic contaminants, soaking once or twice for 24 h can have different effects on absolute G/B values. Corals must therefore be treated consistently to ensure accurate cross core comparisons. A single 24 h treatment is sufficient in most cases; however, when resistant contaminants remain a second 24 h treatment improves the signal. Absolute values can therefore not be compared when cores are cleaned using different treatment methods.