Modelling light-induced discoloration of 40 wood species used in artistic objects: a colorimetric and chemometric approach

Lignocellulosic surfaces in heritage artifacts are susceptible to irreversible light-induced degradation. This study quantifies UV-B-induced discoloration across 40 wood species and assesses whether baseline CIELAB parameters predict photostability. After 240 h of UV-B exposure, color changes were monitored by colorimetry. Photo-yellowing was dominant and strongly correlated with Δ E * ( r = 0.811, p < 0.001). A regression model explained 48.5% of Δ E * variance ( R 2 = 0.485, RMSE = 2.52); higher initial yellowness was associated with less discoloration, although the underlying cause requires further investigation of the wood’s chemical structure. Stratified models showed greatest explanatory power for below-median b * ( R 2 = 0.687, RMSE = 1.98) and above-median C * ( R 2 = 0.660, RMSE = 2.07). PCA identified coupled yellowing–reddening and lightness-shift patterns. K -means clustering identified four photostability cohorts, enabling a matrix linking color compatibility and photostability. Findings support considering photostability in wood selection for conservation and artworks, while highlighting the need to incorporate chemical properties in future models.