“…Classical approaches consider the relation between climate elements (e.g., temperature or precipitation) and a tree-ring proxy, for example, tree-ring width (TRW) or maximum latewood density (MXD), by scaling or building linear regression models (Briffa et al, 1992;Cook et al, 2019;Cook & Kairiukstis, 1990;Esper et al, 2005Esper et al, , 2012Gurskaya et al, 2012;Lara et al, 2020;Li et al, 2012;Wilson & Luckman, 2003). ML algorithms are tested as transfer functions for this relationship by training artificial neural networks, random forests, or boosted regression trees (Gu et al, 2019;Jevšenak et al, 2018;Jevšenak & Skudnik, 2021;Salehnia & Ahn, 2022).…”