Climate change in south-west Australia: Using tree rings to reconstruct the past and future

A project undertaken at the University of Western Australia and supervised by Dr Louise Cullen

Understanding past climates is the key to predicting climate change and its impacts. In Australia, however, the lack of climate records longer than ~100 years hampers our ability to understand natural climate cycles and develop scenarios for the future. One means of extending instrumental records is to use natural archives of climate, such as those available from ice cores, speleothems and lake sediments. The annual growth rings of trees, in particular, provide unparalled records of climate (Figure 1). This is because tree rings are absolutely dated, annually resolved and since trees can live for several centuries (and occasionally millennia), tree rings can provide relatively long climate records. Although tree rings are frequently used elsewhere in the world as a climate proxy, they have received relatively little attention in mainland Australia. Nevertheless, preliminary research in the 1970s and 80s pointed to the potential of the relatively long-lived (~250 to 300 years) native conifer Callitris (Cupressaceae) for dendroclimatological studies (using tree rings as climate proxies). New studies, using modern dendroclimatological techniques, are now needed to further investigate the potential of Callitris to expand Australia's network of climate data.

Traditionally, tree-ring widths have been used as the primary source of climate information. However, in the last two decades there has been a growing use of stable isotopes, particularly of carbon d13C), in tree rings as a climate proxy. d13C in tree rings of Callitris could provide new opportunities to develop more accurate records of climate: Callitris typically grows in semi-arid regions, where relative humidity and rainfall should strongly control stomatal conductance and, ultimately, tree-ring d13C. Nevertheless, the often assumed contention that stable isotopes in tree rings provide the best means of reconstructing past climate has not been fully tested. Therefore, the aim of our study was to investigate and compare the dendroclimatic potential of tree-ring widths and d13C of Callitris. We focused our study on three broadly co-occurring species of Callitris in the semi-arid south-west Western Australia (SWWA), C. columellaris, C. canescens and C. preissii (Figure 2).

Tree-ring widths of Callitris canescens and C. columellaris have significant dendroclimatic potential, particularly as a proxy for rainfall. The ring-width chronologies (year-to-year record of growth) of C. canescens and C. columellaris had relatively strong (r > 0.6) and positive correlations with autumn-winter (March-September) rainfall: years with below average autumn and winter rainfall reduce growth rates and, hence, the width of tree rings. The autumn-winter period is when SWWA receives much of its rainfall and it is unsurprising that rainfall during this time is a significant influence on tree growth in a semi-arid environment. On the other hand, the correlations between the d13C chronologies and autumn-winter rainfall were weaker than those found for ring widths (r < 0.50), despite rainfall being the dominant climatic signal in both proxy types. In addition, neither ring-widths nor d13C of C. preissii were strongly correlated with rainfall or temperature (all r < 0.36). We conclude that tree-ring widths of C. canescens and C. columellaris appear to have the most potential to provide records of past rainfall in south-west Western Australia. On the other hand, d13C in tree rings, and C. preissii are currently less useful as climate proxies.

We used the tree-ring width chronology of Callitris columellaris to reconstruct and investigate natural variability in rainfall in south-west Western Australia. The chronology developed from C. columellaris was not only the longest available from the three species (350 years compared with 130 years for C. canescens and 70 years for C. preissii), but also had the strongest correlation with autumn-winter rainfall. Using the C. columellaris chronology, autumn-winter rainfall was reconstructed back to 1654 AD, revealing considerable multi-decadal variability in rainfall (Figure 3). Rainfall exhibits fluctuations from dry periods that often last 20 to 30 years to periods of above average rainfall that tend to persist for only 15 years or so. Our data support conclusions made previously from modelling studies that rainfall in south-west Western Australia can exhibit significant natural variability. Our study also demonstrates the potential of C. columellaris to provide long records of rainfall for the region.


Multi-decadal scale variability in autumn-winter rainfall in south-western Australia since 1655 AD as reconstructed from tree rings of Callitris columellaris. Cullen, L E & P F Grierson (2008). Clim Dyn DOI 10.1007/s00382-008-0457-8


Figure 1. Tree rings showing variation in growth from year-to-year, a function of changes in climate.

Figure 2. Callitris in south-west Australia.

Figure 3. Reconstructed autumn-winter (March-September) rainfall. Annual values are plotted in gray and a 20-year smoothing spline in black.