The influence of low-frequency variability on tree-rings based climate reconstruction: a case study from central Italy (Roman coast)

Abstract

Tree rings are among the best sources of proxy data for reconstructing past climatic records. In this study we explore for the first time what type of climatic signals can be reconstructed from stone pine ( Pinus pinea L.) based on tree-rings from central Italy (Roman coast). Samples from 112 stone pine trees from stands with different age classes were collected at two locations, Castel Fusano and Castelporziano. In determining the particular target variable for climate reconstruction we explored a wide range of climatic signals (from monthly to multiple year scale) for correlations with tree ring chronologies produced using a variety of detrending methods. We reconstructed short term (autumn-early winter) and long term (3 years precipitation) signals during the 150 years of available data using the “classical” detrending method but also methods preserving their low frequency variability (ABD and RCS) within the chronologies. By setting the best multiple year precipitation drivers at an annual scale and applying a simple percentile threshold approach, we identified the wettest and driest climatic events. The best accuracy in identifying the climatic thresholds was obtained with the ABD method, which also showed the best cross spectral correlation with a long precipitation record. Our reconstruction underpins that since ca. 1850 the Roman coast has experienced its driest conditions in terms of 2-3 year rainfall sums during the last 50 years of the 20 th Century. This finding may be used in the context of identifying the long-term natural variability of the region in relation to climate change as it is expected to affect the Mediterranean.