IPB University Logo

SCIENTIFIC REPOSITORY

IPB University Scientific Repository collects, disseminates, and provides persistent and reliable access to the research and scholarship of faculty, staff, and students at IPB University

AI Repository
 
Building and Categories


      View Item 
      •   IPB Repository
      • Final Assignments
      • Master Final Assignments
      • MF - Forestry
      • View Item
      •   IPB Repository
      • Final Assignments
      • Master Final Assignments
      • MF - Forestry
      • View Item
      JavaScript is disabled for your browser. Some features of this site may not work without it.

      Pemodelan Prediktif Biomassa Atas Permukaan pada Areal Regenerasi Alami di Hutan Tanaman Lahan Gambut Menggunakan Data Sentinel-2

      Thumbnail
      View/Open
      Cover (464.5Kb)
      Fulltext (1.347Mb)
      Date
      2026
      Jenis/Type
      Tesis
      Subtype
      Theses
      Author
      Cahyaningtyas, Anggita Utami
      Setiawan, Yudi
      Putra, Erianto Indra
      Metadata
      Show full item record
      Abstract
      Biomassa atas permukaan merupakan komponen penting dalam inventarisasi hutan dan estimasi cadangan karbon, tetapi pendugaannya melalui pengukuran lapangan memerlukan waktu, tenaga, dan biaya yang besar. Penelitian ini dilakukan pada areal hutan tanaman industri di lahan gambut tropis Sumatra yang telah mengalami pemanenan dan selanjutnya berkembang melalui regenerasi alami. Proses regenerasi tersebut menghasilkan kondisi tegakan yang heterogen dengan variasi struktur dan karakteristik vegetasi, sehingga hubungan antara variabel spektral dan biomassa atas permukaan menjadi lebih kompleks. Penelitian ini bertujuan mengembangkan model prediksi biomassa atas permukaan melalui integrasi hasil inventarisasi data lapangan dan Sentinel-2 serta membandingkan performa beberapa pendekatan pemodelan pada kondisi tegakan tersebut. Penelitian menggunakan 145 plot sampel dan empat algoritma, yaitu linear model, generalized additive model, random forest, dan support vector regression. Tinggi pohon yang tidak diukur diprediksi menggunakan model Weibull sebelum perhitungan biomassa atas permukaan. Hasil penelitian menunjukkan RF memberikan performa prediksi terbaik dengan RMSE sebesar 142,918 Mg ha?¹, MAE sebesar 103,758 Mg ha?¹, dan R² sebesar 0,177. Model random forest menggunakan sebelas variabel prediktor terpilih dan menghasilkan peta distribusi biomassa atas permukaan secara spasial. Meskipun random forest merupakan model dengan performa terbaik, nilai R² yang relatif rendah menunjukkan kemampuan variabel spektral Sentinel-2 dalam menjelaskan variasi biomassa masih terbatas pada kondisi tegakan yang heterogen. Keterbatasan tersebut menunjukkan variasi struktur tegakan dan karakteristik vegetasi pada areal regenerasi alami belum sepenuhnya dapat direpresentasikan oleh data spektral yang digunakan.
       
      Aboveground biomass is an important component of forest inventories and carbon stock estimation, but its assessment through field measurements requires considerable time, labor, and cost. This study was conducted in an industrial plantation forest area on tropical peatland in Sumatra that had undergone harvesting and subsequently developed through natural regeneration. This regeneration process resulted in heterogeneous stand conditions with variations in stand structure and vegetation characteristics, making the relationship between spectral variables and aboveground biomass more complex. This study aimed to develop an aboveground biomass prediction model by integrating field inventory data and Sentinel-2 data and to compare the performance of several modeling approaches under these stand conditions. The study used 145 sample plots and four algorithms, namely linear model, generalized additive model, random forest, and support vector regression. Tree heights that were not measured in the field were predicted using a Weibull model prior to aboveground biomass estimation. The results showed that RF provided the best predictive performance, with an RMSE of 142.918 Mg ha?¹, MAE of 103.758 Mg ha?¹, and R² of 0.177. The random forest model used eleven selected predictor variables and produced a spatial distribution map of aboveground biomass. Although random forest showed the best performance among the evaluated models, its relatively low R² indicates that the ability of Sentinel-2 spectral variables to explain biomass variability remains limited under heterogeneous stand conditions. This limitation suggests that variations in stand structure and vegetation characteristics in naturally regenerating areas were not fully represented by the spectral data used in this study.
       
      URI
      http://repository.ipb.ac.id/handle/123456789/179501
      Collections
      • MF - Forestry [1607]

      Copyright © 2020 Library of IPB University
      All rights reserved
      Contact Us | Send Feedback
      Indonesia DSpace Group 
      IPB University Scientific Repository
      UIN Syarif Hidayatullah Institutional Repository
      Universitas Jember Digital Repository
        

       

      Browse

      All of IPB RepositoryCollectionsBy Issue DateAuthorsTitlesSubjectsThis CollectionBy Issue DateAuthorsTitlesSubjects

      My Account

      Login

      Application

      google store

      Copyright © 2020 Library of IPB University
      All rights reserved
      Contact Us | Send Feedback
      Indonesia DSpace Group 
      IPB University Scientific Repository
      UIN Syarif Hidayatullah Institutional Repository
      Universitas Jember Digital Repository