prilagođeno pretraživanje po punom tekstu

ŠUMARSKI LIST 7-8/2020 str. 24     <-- 24 -->        PDF

following standard dendrochronological procedures (Stokes and Smiley, 1968). The cores were then digitized using ATRICS system1 (Levanič, 2007) and annual radial growth was measured to the nearest 0.01 mm using WinDENDRO software. Each tree ring series was then visually and statistically crossdated using PAST-4. We used the COFECHA program (Holmes, 1983, Holmes, 1994) to check for errors in dating. Tree growth almost universally shows a non-climatic, age-related biological growth trend that must be removed before any dendroclimatological analyses. Individual tree-ring width series were therefore standardised to remove long-term trends (Cook, 1985) using a smoothing loess function. To calculate index chronologies each year’s ring width was divided by that year’s fitted value to give a dimensionless index with a mean of 1. This procedure removed non-climatic trends due to tree age, size and the effects of stand dynamics (Cook, 1985). Index values were also pre-whitened using an autoregressive model selected on the basis of the minimum