Athaghaly, Alief (2025) Prediksi Curah Hujan Ekstrem Provinsi Aceh Menggunakan Spatial Extreme Value Dengan Pendekatan Max Stable Processes. Other thesis, Institut Teknologi Sepuluh Nopember.
![]() |
Text
06311940000019-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (2MB) | Request a copy |
Abstract
Data Informasi Bencana Indonesia (DIBI) dari Badan Nasional Penanggulangan Bencana (BNPB) menyatakan dalam 25 tahun kebelakang provinsi Aceh mengalami 1.032 kejadian bencana banjir serta 115 kejadian bencana longsor. Kejadian ini telah memberikan dampak kerusakan yang sangat besar bagi pemerintah dan masyarakat Aceh. Curah hujan ekstrem merupakan salah satu faktor utama yang menyebabkan bencana alam banjir dan longsor terjadi. Penelitian ini menerapkan metode Spatial Extreme Value (SEV) dalam pemodelan curah hujan ekstrem dengan pendekatan Max Stable Processes (MSP) model Brown-Resnick. Data yang digunakan pada penelitian ini berupa curah hujan harian dari 6 titik pengamatan yang tersebar di Kecamatan/Kota di Provinsi Aceh (Baiturrahman, Muara Dua, Johan Pahlawan, Tapak Tuan, Lut tawar, dan Kuala Simpang) dengan rentang waktu 1 Desember 2004 – 30 November 2024. Pengambilan data ekstrem dilakukan dengan metode Block Maxima menggunakan R Largest Order dimana nilai r yang digunakan sebesar 2. Pembagian data ekstrem menjadi data training dan testing dengan proporsi 80% dan 20%. Data ekstrem kemudian ditransformasi ke unit margin Frechet Z. Selanjutnya dihitung koefisien ekstremal yang menghasilkan nilai diantara 1.12 – 2 yang menunjukkan adanya hubungan di beberapa pasangan lokasi. Dibentuk model trend surface dari kombinasi koordinat longitude latitude sehingga ditentukan model terbaik berdasarkan TIC terkecil. Didapatkan model trand surface terbaik adalah model yang melibatkan koordinat longitude pada parameter bentuk dan parameter skala. Menggunakan model terbaik tersebut dilakukan estimasi parameter secara spasial mengikuti model Brown-Resnick. Selanjutnya diukur tingkat akurasi model yang dihasilkan dari estimasi parameter spasial berdasarkan perbandingan return value dan data testing. Nilai MAPE yang didapatkan adalah 36.9% dimana nilai ini berada pada kategori layak untuk digunakan. Langkah terakhir yaitu menghitung prediksi curah hujan ekstrem untuk periode 10, 15, dan 20 tahun ke depan. Hasil prediksi untuk periode 10 tahun kedepan pada lokasi Baiturrahman, Muara Dua, Johan Pahlawan, Tapak Tuan, Lut tawar, dan Kuala Simpang secara berturut-turut adalah 217.9449 mm/hari, 108.9558 mm/hari, 232.449 mm/hari, 131.3434 mm/hari, 137.1291 mm/hari, 114.06 mm/hari. nilai prediksi ini terus meningkat seiring bertambahnya waktu untuk seluruh titik pengamatan. Seluruh hasil prediksi tersebut masuk dalam kategori sangat lebat. Hasil ini dapat digunakan oleh pihak terkait seperti Badan Nasional Penanggulangan Bencana (BNPB) atau Badan Penanggunalangan Bencana Aceh (BPBA) untuk melakukan upaya mitigasi bencana alam.
===========================================================================================================================================
The Indonesia Disaster Information Database (DIBI) from the National Disaster Management Agency (BNPB) reported that over the past 25 years, Aceh Province has experienced 1,032 flood incidents and 115 landslide events. These disasters have caused significant damage to both the government and the people of Aceh. Extreme rainfall is one of the main factors contributing to the occurrence of these natural disasters. This study applies the Spatial Extreme Value (SEV) method for modeling extreme rainfall using the Max Stable Processes (MSP) approach with the Brown-Resnick model. The data used in this research consist of daily rainfall observations from six monitoring points distributed across districts/cities in Aceh Province (Baiturrahman, Muara Dua, Johan Pahlawan, Tapak Tuan, Lut Tawar, and Kuala Simpang) covering the period from December 1, 2004, to November 30, 2024. Extreme data extraction was conducted using the Block Maxima method with the R Largest Order approach, where the value of rrr used is 2. The extreme data were divided into training and testing datasets with a proportion of 80% and 20%, respectively. The extreme data were then transformed into Frechet Z unit margins. Subsequently, the extremal coefficient was calculated, resulting in values between 1.12 and 2, indicating relationships between certain location pairs. A trend surface model was formed based on combinations of longitude and latitude coordinates, and the best model was determined based on the smallest TIC value. The best trend surface model involved the longitude coordinate in both the shape and scale parameters. Using this optimal model, spatial parameter estimation was performed following the Brown-Resnick model. The accuracy of the model resulting from the spatial parameter estimation was evaluated by comparing the return values with the testing data. The obtained MAPE value was 36.9%, which falls within the acceptable category for practical use. The final step was to predict extreme rainfall for the next 10, 15, and 20 years. The prediction results for the next 10 years at the locations of Baiturrahman, Muara Dua, Johan Pahlawan, Tapak Tuan, Lut Tawar, and Kuala Simpang are, respectively, 217.9449 mm/day, 108.9558 mm/day, 232.449 mm/day, 131.3434 mm/day, 137.1291 mm/day, and 114.06 mm/day. These predicted values show an increasing trend over time at all monitoring points and are categorized as very heavy rainfall. These findings can be utilized by relevant agencies such as the National Disaster Management Agency (BNPB) or the Aceh Disaster Management Agency (BPBA) to implement disaster mitigation efforts effectively.
Item Type: | Thesis (Other) |
---|---|
Uncontrolled Keywords: | Curah Hujan, Spatial Extreme Value, Max Stable Processes, Brown-Resnick, r-Largest Order, Return Level,Rainfall, Spatial Extreme Value, Max Stable Processes, Brown-Resnick, r-Largest Order, Return Level |
Subjects: | Q Science > QC Physics > QC866.5 Climatology--Forecasting. |
Divisions: | Faculty of Mathematics, Computation, and Data Science > Actuaria > 94203-(S1) Undergraduate Thesis |
Depositing User: | Alief Athaghaly |
Date Deposited: | 01 Aug 2025 06:40 |
Last Modified: | 01 Aug 2025 06:40 |
URI: | http://repository.its.ac.id/id/eprint/125430 |
Actions (login required)
![]() |
View Item |