Infection Severity of <i>Arceuthobium oxycedri</i> (Viscaceae) in Two Protected Areas in the Alborz Mountains, Iran: A Case Study

ABSTRACT This study investigated the severity of mistletoe infection on junipers and the relationship between tree characteristics, physiographic factors and mistletoe infection. The study was conducted in two protected areas: Parvar and Miankouh Tash in Semnan Province, Iran. A dwarf mistletoe‐infested area of 400 ha was sampled by placing a 250 × 250 m grid over a map of each study area. A 20 × 20 m (400 m 2 ) plot was placed at the intersection of grid lines. This resulted in a total of 110 forested plots: 45 plots in Parvar and 65 plots in Miankouh Tash. In each plot, all live junipers were measured for total height, crown area, collar diameter and crown diameter. Mistletoe severity was estimated using the 6‐class dwarf mistletoe rating system (DMR). Dead trees were also recorded and examined for evidence of past mistletoe infection. The Boruta algorithm was used for feature selection, and the selected features were utilised for predicting DMR using a random forest (RF) model. Over 40% of junipers in the sample plots were infected and 7.5% were dead. Mortality was greater in severely infested plots compared to uninfested and lightly infested plots. The results of univariate analysis demonstrated that the largest trees across plots were more severely infected. Infection severity was highest on steep terrains (slopes more than 60%) where the largest junipers were growing. Based on Boruta feature selection algorithm, collar diameter, tree height and crown area of junipers have been detected as the most important variables for the explanation of infection severity. Crown diameter, geographical aspect, density of juniper trees, altitude and juniper canopy cover were classified as tentative variables in this model. Slope percentage beside two variables related to stand characteristics including all woody species (basal area and tree density) were the rejected variables. The correlation coefficient ( r ) and root mean square error (RMSE) between the predictions by RF and the measured DMR were 0.83 and 0.53 respectively. These results indicate that the selected environmental variables can be used for spatial prediction of DMR in both study areas. Because increased mortality was observed for severely infected areas, more efficient management methods are needed to control the dwarf mistletoe populations in the Alborz Mountains.