Comparative analysis of deadwood abundance and characteristics in protected versus managed forests in Northeastern Turkey

Abstract Deadwood is essential for maintaining forest ecosystem health and biodiversity. This study investigates and compares the abundance and characteristics of deadwood in Eastern spruce ( Picea orientalis Link. At Carr) stands located within the Velikoy Forest Management Unit (FMU) and Karagol-Sahara National Park (NP) in northeastern Turkey. A total of 476 randomly selected sampling plots were assessed —236 in Velikoy FMU and 240 in Karagol-Sahara NP—based on stand type, canopy cover, and age class. Deadwood was classified into standing deadwood, stumps, and downed deadwood, with volumes measured for each category. To model deadwood volume, we applied both fixed and mixed non-linear models alongside a Generalized Linear Mixed Model (GLMM), incorporating stand parameters, physiographic variables, and geographical factors. Model performance was evaluated using adjusted R 2 ( R 2 _adj ), root mean square error (RMSE), Akaike’s Information Criterion (AIC), and Bayesian Information Criterion (BIC). The results showed that in Velikoy FMU, the composition of deadwood volume was 29.5% standing, 37.8% stumps, and 32.7% downed. In contrast, Karagol-Sahara NP exhibited 47.8% standing, 22.7% stumps, and 29.4% downed deadwood. Statistical analyses revealed significant positive correlations between total deadwood volume and factors such as basal area, stand density, stand volume, tree count, site index, stand age, and distance to settlements. However, physiographic variables like slope and aspect showed no clear association. The fixed-effects model yielded R 2 _adj = 0.714, RMSE = 10.376 m 3 ha −1 , AIC = 1584.9, and BIC = 1590.4. The mixed-effects model improved performance significantly with R 2 _adj = 0.787, RMSE = 5.815 m 3 ha −1 , AIC = 1262.2, and BIC = 1267.4. The GLMM identified basal area, stand density, and distance to settlements as the most influential predictors of deadwood volume. The GLMM model further reduced -2LogL, AIC, and BIC to 737.7, 741.7, and 749.5, respectively, underscoring the effectiveness of incorporating random effects in modeling. These findings emphasize the value of deadwood in forest management and its relationship with structural stand attributes. The integrated modeling framework offers robust tools for informing biodiversity-oriented forest management and conservation planning. Enhancing basal area and stand density, particularly in areas distant from human settlements, can foster deadwood accumulation and support biodiversity.