Sampling Design for Trees Outside Forest in Indonesia

Abstract

The existing National Forest Inventory (NFI) system of Indonesia provides forest resource data on state forested lands only. Meanwhile, the Government Regulation No. 23/2021 provided a mandate to the Ministry of Environment and Forestry (MoEF) to conduct NFI not only at state forests, but also at customary forests and private/community forests. This new mandate requires a new sampling design for conducting NFI at customary and community forests, which are collectively named as Trees Outside Forest (TOF) lands. This study was therefore aimed at proposing an appropriate sampling design for TOF lands. This study was conducted by first discussing relevant criteria for defining TOF lands with the MoEF and other related parties. The main sampling design components, i.e., cluster-plot design and sample size (number of cluster-plot), were determined based on available community forest inventory data in Java. The TOF cluster-plot was designed to consist of three sample plots, in which each sample plot consists of three nested-plots that would provide fairly balanced contributions of small, medium, and large trees when estimating stand attributes. Appropriate number of cluster-plot was determined based on variation coefficients of stand attributes and targeted precision (i.e., allowable sampling error). In the final step, the proposed sampling design was tested at TOF lands in West Kalimantan and West Java. Based on discussions with relevant parties, this study defined a TOF land as "a land classified as pure dry agriculture, mixed dry agriculture, dry shrub, or wet shrub according to the Indonesian land cover system with minimum area of 0.25 ha with trees higher than 5 m and canopy cover more than 10%”. Appropriate cluster-plot for TOF land inventory consisted of three plots with a distance of 50 m. Each plot consisted of three nested subplots with a radius of 5 m for small trees with 5≤DBH<20 cm, a radius of 15 m for medium trees with 20≤DBH<35 cm, and a radius of 25 m for large trees with DBH≥35 cm. The pilot testing in West Kalimantan and Java confirmed that such cluster-plot design enable the field measurement of a cluster-plot within one day. To conduct NFI at TOF lands with a sampling error of 15%, this study recommended 70 cluster-plots for each island in Indonesia. The MoEF, however, might then adjust the number of cluster-plots by considering available budget and targeted precision for TOF lands inventory.