Fadil, Aufa Dreihazi Muhammad and Ir. Aditya Nugraha Putra, S.P., M.P. (2024) Pemetaan dan Klasifikasi Hutan Rakyat di Kawasan Daerah Aliran Sungai Pegunungan Vulkanik Berbasis Google Earth Engine, Random Forest, dan Mega Composite Indeks Vegetasi. Sarjana thesis, Universitas Brawijaya.
Abstract
Daerah aliran sungai (DAS) adalah wilayah yang dibatasi oleh punggung gunung atau bukit, mengalirkan dan mengumpulkan air hujan ke dalam sistem sungai. Tutupan lahan di Sub DAS Sumber Brantas dan Kali Konto terdiri dari berbagai tipe, termasuk hutan alami, hutan rakyat, hutan produksi, lahan pertanian, semak belukar, padang rumput, dan lahan kosong. Hutan rakyat, dikelola oleh masyarakat di luar kawasan lindung, memberikan manfaat ekonomi dan sumber daya bagi masyarakat. Karena penguasaan lahan dan pengelolaan hutan rakyat adalah di tangan masyarakat, maka muncul suatu permasalahan dimana hutan rakyat ini terancam beralih fungsi ke penggunaan lahan lainnya. Penelitian ini bertujuan untuk menganalisis keberhasilan kombinasi Google Earth Engine, mega composite indeks vegetasi, dan algoritma Random Forest dalam memantau keberadaan hutan rakyat, serta mengevaluasi keberadaan hutan rakyat untuk mitigasi bencana hidrometeorologi di Sub DAS (Sub DAS Sumber Brantas dan Kali Konto). Analisis dilakukan dengan memanfaatkan citra Sentinel 2A perekaman Oktober 2022 hingga Oktober 2023 yang kemudian dilakukan klasifikasi terbimbing (supervised classification) berdasarkan penggunaan lahan yang dilakukan melalui platform Google Earth Engine dan memanfaatkan algoritma Random Forest. Penentuan titik pengamatan didasarkan pada area yang diduga hutan rakyat dari hasil klasifikasi dan berada di luar kawasan Perhutani. Hasil pengamatan di lapangan yang kemudian dijalankan klasifikasi menggunakan algoritma Random Forest dengan menggunakan mega composite indeks vegetasi sebagai base. Titik pengamatan yang berjumlah 453 digunakan sebagai training point untuk menjalankan algoritma Random Forest dan 30 titik pengamatan digunakan sebagai titik uji validasi. Hasil analisis Random Forest menunjukan 69 tutupan lahan yang berbeda di setiap wilayah Sub DAS, dengan luasan yang berbeda beda di setiap desa. Keberadaan hutan rakyat berpengaruh dalam memitigasi bencana hidrometeorologi. Bencana tanah longsor dapat ditekan karena keberadaan hutan rakyat yang didominasi oleh vegetasi pepohonan yang berperan penting dalam menstabilkan agregat tanah. Bencana lainnya seperti banjir dapat ditekan dengan keberadaan hutan rakyat karena aliran air hujan dapat cepat diserap oleh tanah, dan efek kanopi yang dimiliki pepohonan dapat meminimalisir terjadinya limpasan permukaan. Pemanfaatan kombinasi teknologi Google Earth Engine, mega composite indeks vegetasi, dan algoritma Random Forest dapat memantau keberadaan dan persebaran hutan rakyat di kawasan daerah aliran sungai pegunungan vulkanik (Sub DAS Sumber Brantas dan Kali Konto) dengan hasil tingkat akurasi sebesar 86%.
English Abstract
The Watershed (DAS) is an area bounded by mountain ridges or hills, which channels and collects rainfall into the river system. Land cover in the Sub Watershed of Sumber Brantas and Kali Konto consists of various types, including natural forests, community forests, production forests, agricultural land, shrubs, grasslands, and vacant land. Community forests, managed by communities outside protected areas, provide economic benefits and resources for the local population. However, due to the control and management of community forests lying in the hands of the community, there arises a problem where these forests are at risk of conversion to other land uses. This research aims to analyze the success of the combination of Google Earth Engine, mega composite vegetation indices, and the Random Forest algorithm in monitoring the presence of community forests. It also aims to evaluate the existence of community forests for the mitigation of hydrometeorological disasters in the Sub Watershed of Sumber Brantas and Kali Konto. The analysis was conducted utilizing Sentinel 2A imagery recorded from October 2022 to October 2023. Supervised classification was performed based on land use through the Google Earth Engine platform, utilizing the Random Forest algorithm. Observation points were determined based on areas suspected to be community forests from the classification results and located outside the Perhutani forest area. Field observations were then used to classify the areas using the Random Forest algorithm, with the mega composite vegetation index as the base. The observation points, totaling 453, were used as training points to execute the Random Forest algorithm, while 30 observation points were utilized as validation points The Random Forest analysis showed 69 different land covers in each sub-watershed area, with varying sizes in each village. The presence of community forests plays a significant role in mitigating hydrometeorological disasters. Landslides can be reduced due to the presence of community forests dominated by tree vegetation, which plays a crucial role in stabilizing soil aggregates. Other disasters such as floods can also be mitigated by the presence of community forests because rainwater can be quickly absorbed by the soil, and the canopy effect of trees can minimize surface runoff. The utilization of the combination of Google Earth Engine technology, mega composite vegetation index, and the Random Forest algorithm can monitor the presence and distribution of community forests in the volcanic mountain watershed area (Sub Watershed of Sumber Brantas and Kali Konto) with an accuracy level of 86%.
Item Type: | Thesis (Sarjana) |
---|---|
Identification Number: | 0524040190 |
Divisions: | Fakultas Pertanian > Agroekoteknologi |
Depositing User: | Unnamed user with username nova |
Date Deposited: | 14 May 2024 06:37 |
Last Modified: | 14 May 2024 06:37 |
URI: | http://repository.ub.ac.id/id/eprint/218897 |
![]() |
Text (DALAM MASA EMBARGO)
Aufa Dreihazi.pdf Restricted to Registered users only Download (8MB) |
Actions (login required)
![]() |
View Item |