Silver fir tree ring width: a proxy for winter temperature variability in the Carpathians?

A composite Silver fir tree ring width chronology from seven mountain sites in the Eastern Carpathian Mountains (Europe) was established for AD 1588-2021, with SSS>0.8 and EPS>0.85 for AD 1750-2012. The bootstrap correlation analysis of the tree ring index with monthly climate parameters (temperature and precipitation) shows a positive and relatively time constant response to mean winter-spring temperature (November to March). The correlation between Silver fir tree ring proxy and winter-spring temperature is high and statistically significant (0.556 at p<0.05). The reconstruction statistics (R2, RE, CE and DW) indicate a good skill of the regression model between proxy data and winter temperature back to 1901. RE and CE statistics range between 0.32 and 0.39, and DW has values between 2.05 and 2.18. These results show good reliability of the model, and for the entire period, the reconstruction explains ~ 30% of winter temperature variability. The temperature reconstruction from AD 1750 shows inter-decadal fluctuation induced by low frequencies sinusoids (waves). The reconstructed mean winter temperatures for the 1750-2012 period was -2.93°C with -0.31°C colder than the 1961-2009 reference period. The longest period with high frequencies of years with low temperatures was recorded in 1740-1800, coinciding with the end of the Little Ice Age. After this coldest winter period, a six-year period with extreme warm winters was identified. The warming trend was more distinguishable science AD 1880 to the present, especially through the high frequency of mild winters. The coldest reconstructed winters for entire period were find in 2003 (anomalies= -1,56), 2012 (anomalies = -1,32) and 1965 (anomalies = -1,24). The warmest winters were recorded in 2001 (anomalies = +1,71), 1998 (anomalies = +1,48) and 2007 (anomalies = +1,37). The pattern of spatial correlation between proxy data and winter-spring temperature releases a wide extend of high correlation (>0.5), covering the North-Western Carpathians, continues with the Eastern chain of the Carpathian Mountains and finishes with the extreme South-East of Romania. Correlation with the Central Europe gridded temperatures is significant (>0.4), and with the Alpine Arc grid, temperatures are quite low (0.3). This result provides a regional scale of the winter-spring temperature reconstruction, suggesting a possible west-east gradient across Europe, potentially influenced by the interplay between the eastward expansion of Atlantic influence and the westward expansion of the West Asian influence.