Simple Search:

Comparative assessment of pre-felling sampling designs in lowland forest of Peninsular Malaysia


Citation

Abd Hamid, Nurul Atiqah (2016) Comparative assessment of pre-felling sampling designs in lowland forest of Peninsular Malaysia. Masters thesis, Universiti Putra Malaysia.

Abstract / Synopsis

In Peninsular Malaysia, forest sampling during pre-felling (pre-f) inventory is to gather information on tree stocking, species, forest condition, distribution of big trees, bamboo and rattan before commercial felling of trees ≥ 45 cm dbh. Currently, the pre-f inventory is carried out using 20 m x 50 m plot at 10% sampling intensity with 95% confidence level and ±20% sampling error. Even though the sampling design used in pre-f inventory is systematic sampling, it is not known whether this type of design is the most accurate one. Four tested sampling designs in this study are (i) random sampling, (ii) cluster sampling, (iii) systematic sampling, and (iv) transect line sampling. The objectives of this study were to compare the four sampling designs in pre-f of mixed tropical forest. This was done by comparing which sampling design is closer to population data set of 30 ha for trees ≥ 30 cm dbh in estimating mean basal area of trees per ha and volume per ha by species groups of Dipterocarp, Non-Dipterocarp and All Species. This study was carried out using a 30 ha (1000 m x 300 m) forest area within the 50ha Demography Project established by the Forest Research Institute Malaysia (FRIM) in Pasoh Forest Reserve, Negeri Sembilan. The sample estimates were done at 95% confident level and replicated three times based on 20 m x 50 m plot information. The following results were obtained when basal area per ha (m²/ha) of All Species for trees ≥ 30 cm dbh was inventoried: (i) Systematic sampling underestimated by 0.94% for All Species; (ii) Transect line sampling overestimated by 2.18% for Dipterocarp species; (iii) Cluster sampling overestimated by 1.03% for Non-dipterocarp species. Similarly, to estimate mean volume per ha (m³/ha) of All Species for trees ≥ 30 cm dbh, the following results were best obtained: (i) Systematic sampling underestimated by 1.16% for All Species; (ii) Transect line sampling overestimated by 1.06% for Dipterocarp species; (iii) Cluster sampling overestimated by 3.77% for Non-dipterocarp species. Meanwhile, in estimating dbh (cm) and height (m), it was found that random sampling would be a preferred sampling design to inventory Dipterocarp and Non-dipterocarp species. The difference means of dbh (cm) and height (m) to the actual population is overestimated 0.39% and underestimates 0.31% for Dipterocarp species. For Non-dipterocarp species, the difference is underestimating 0.07% and overestimates 0.19% respectively. However, systematic sampling would be the most accurate sampling design to inventory dbh (cm) and height (m) of All Species which is only recorded overestimate of 0.13% and 0.18% respectively. Overall, systematic sampling was best to estimate basal area (m²/ha), volume (m³/ha), dbh (cm), and height (m) as it gave the closest, albeit underestimated or overestimated to the actual. Therefore, it is proven that systematic sampling is the most accurate sampling design for pre-f inventory.


Download File

[img]
Preview
PDF
FH 2016 16 IR.pdf

Download (2MB) | Preview

Additional Metadata

Item Type: Thesis (Masters)
Subject: Forests and forestry - Statistical methods
Subject: Sampling (Statistics)
Call Number: FH 2016 16
Chairman Supervisor: Associate Professor Kamziah Abd Kudus, PhD
Divisions: Faculty of Forestry
Depositing User: Mr. Sazali Mohamad
Date Deposited: 22 Aug 2019 12:07
Last Modified: 22 Aug 2019 12:07
URI: http://psasir.upm.edu.my/id/eprint/70120
Statistic Details: View Download Statistic

Actions (login required)

View Item View Item