University of Guilan Caspian Journal of Environmental Sciences 1735-3033 2025 09 04 Geostatistical analysis of density indices in traditionally managed oak forests 1 17 9028 10.22124/cjes.2025.9028 EN Loghman Ghahramany Department of Forestry, Faculty of Natural Resources, University of Kurdistan; Dr. Hedayat Ghazanfari Center for Research and Development of Northern Zagros Forestry, University of Kurdistan, Sanandaj, Iran Mahtab Pir Bavaghar Department of Forestry, Faculty of Natural Resources, University of Kurdistan; Dr. Hedayat Ghazanfari Center for Research and Development of Northern Zagros Forestry, University of Kurdistan, Sanandaj, Iran Journal Article 2025 09 04 The spatial variation in tree density and basal area provides crucial information for forest health assessment, sustainable forest management, ecosystem monitoring, carbon storage assessment, and climate change mitigation. Literature reviews confirm that geostatistical methods have been effectively applied across various forest management applications. However, their use in traditionally managed oak forests (silvopastoral systems) remains underexplored. The primary challenge in applying these methods lies in the clumped spatial distribution of trees within pollarded oak forests under silvopastoral management.This research applies geostatistical techniques to study the spatial distribution of tree density and basal area in the pollarded oak forests (9,178 ha) of the Northern Zagros region, Northwest Iran, managed under a silvopastoral system. Field measurements were taken in 2019 using a random-systematic sampling design across 117 georeferenced circular plots (0.1 ha each). The mean tree density was 291 stem ha^-1 (CV = 65.6%), while basal area averaged 14.12 m² ha^-1 (CV = 49.9%). Variogram analysis showed isotropic behavior and high spatial dependence (SDD = 88% for tree density and 80% for basal area). An exponential model explained 72% and 68% of variability in tree density and basal area, respectively. Small nugget effect values (0.0384 for density, and 0.0476 for basal area) indicated the reliability of the models. Ordinary kriging produced the best predictions, with relative errors of 33.3% (MAEr) and 44.4% (rRMSE) for tree density as well as 30.63% (MAEr) and 41.8% (rRMSE) for basal area. Although higher rRMSE values reflected local deviations, t-test results revealed no significant differences between measured and estimated values. This study underscores the suitability of kriging methods for mapping spatial variations in tree density and basal area, offering valuable approach for forest health assessment and management. https://cjes.guilan.ac.ir/article_9028_55335e46225c6b0a14952e5e5109979a.pdf