Oktaviana, Pratnya Paramitha (2026) Pengembangan Estimasi Risiko Banjir Berbasis Spatio-Temporal Extreme Value untuk Perhitungan Premi Asuransi Bencana Lingkungan. Doctoral thesis, Institut Teknologi Sepuluh Nopember.
![[thumbnail of 7014231001-Doctoral.pdf]](http://repository.its.ac.id/style/images/fileicons/text.png) |
Text
7014231001-Doctoral.pdf - Accepted Version Restricted to Repository staff only Download (16MB) | Request a copy |
Abstract
Perubahan iklim di Indonesia yang sering terjadi salah satunya adalah meningkatnya curah hujan. Bencana alam yang sering terjadi akibat curah hujan yang tinggi adalah bencana banjir. Jakarta merupakan salah satu daerah di Indonesia yang paling sering mengalami bencana banjir. Tingginya frekuensi dan dampak banjir menyebabkan meningkatnya kebutuhan akan mekanisme perlindungan finansial, salah satunya melalui asuransi banjir. Dalam penerapan asuransi banjir, perhitungan premi yang akurat menjadi sangat penting karena premi harus mencerminkan tingkat risiko banjir yang dihadapi oleh suatu wilayah sehingga dapat memberikan perlindungan yang adil bagi masyarakat dan berkelanjutan bagi penyedia asuransi. Pada penelitian ini, dilakukan analisis risiko bencana banjir dengan cara menggabungkan aspek spatial kondisi daerah dan temporal historis curah hujan di Jakarta, khususnya Jakarta Pusat dan Jakarta Utara. Penelitian yang dilakukan ini mempunyai tiga poin penting, antara lain: yang pertama, perhitungan probabilitas terjadinya banjir di Jakarta akibat curah hujan ekstrem berdasarkan perhitungan probabilitas temporal menggunakan data curah hujan ekstrem dengan analisis extreme value dan probabilitas spasial menggunakan analisis spasial data raster (grid) kerentanan banjir InaRisk; kedua, estimasi risiko kerugian akibat banjir di Jakarta Pusat dan Utara, hasil digunakan sebagai komponen perhitungan premi asuransi banjir; ketiga, perhitungan premi asuransi banjir di Jakarta Pusat dan Utara berdasarkan perhitungan premi asuransi banjir. Penelitian ini tidak hanya berfokus pada estimasi peluang terjadinya banjir akibat curah hujan ekstrem, tetapi juga mengintegrasikan analisis risiko, estimasi kerugian fisik, dan perhitungan premi asuransi dalam satu kerangka penilaian risiko banjir. Analisis temporal dilakukan menggunakan Extreme Value Theory (EVT) dengan metode Peak Over Threshold (POT) untuk memodelkan probabilitas kejadian curah hujan ekstrem. Analisis spasial dilakukan melalui pemodelan kerentanan banjir berbasis Geographic Information System (GIS) dan analisis raster. Kerentanan spasial diperoleh melalui integrasi faktor sosial, ekonomi, fisik, dan lingkungan yang direpresentasikan dalam bentuk raster kerentanan. Pendekatan ini memungkinkan identifikasi variasi tingkat risiko banjir antarwilayah secara lebih detail dan representatif. Estimasi kerugian fisik dilakukan menggunakan flood inundation model berbasis raster dengan beberapa skenario parameter hidrologis, meliputi curah hujan ekstrem, koefisien limpasan (runoff coefficient), dan drainage loss. Hasil simulasi genangan kemudian diintegrasikan dengan data bangunan untuk menghitung potensi kerugian fisik berdasarkan luas bangunan terdampak dan kedalaman genangan banjir. Perhitungan premi asuransi dilakukan menggunakan pendekatan risk-based premium berdasarkan model Picard, di mana premi ditentukan melalui integrasi probabilitas kejadian banjir dan estimasi kerugian fisik. Hasil yang didapatkan adalah: dengan pendekatan temporal dihasilkan probabilitas terjadinya satu atau lebih kejadian banjir akibat curah hujan ekstrem melebihi 122 mm/hari di Jakarta Pusat adalah 32,10% dalam 5 tahun, dan 97,90% dalam 50 tahun. Di Jakarta Utara, probabilitas terjadinya satu atau lebih kejadian banjir akibat curah hujan ekstrem melebihi 126 mm/hari adalah 21,80% dalam waktu 5 tahun dan 91,50% dalam waktu 50 tahun. Dengan pendekatan spasial dihasilkan bahwa Jakarta Pusat memiliki kelas kerentanan banjir tinggi dengan probabilitas kejadian banjir sebesar 0,067 per tahun dan 0,888 dalam 50 tahun. Jakarta Utara juga tergolong berisiko tinggi, dengan probabilitas banjir sebesar 0,045 per tahun dan 0,852 dalam 50 tahun. Pemodelan kerugian fisik menunjukkan bahwa kedalaman genangan, luas bangunan terdampak, serta variabel hidrologis seperti run-off coefficient dan drainage loss merupakan faktor paling dominan dalam menentukan besarnya kerugian. Kerugian meningkat secara signifikan seiring bertambahnya kedalaman banjir, di mana klaster banjir ekstrem (>100 cm) menghasilkan kerugian rata-rata tertinggi meskipun jumlah bangunan terdampak relatif sedikit, sedangkan banjir dangkal (≤30 cm) berdampak pada jumlah bangunan terbesar dengan tingkat kerugian yang relatif rendah. Jakarta Utara menunjukkan kerugian fisik yang lebih besar dibanding Jakarta Pusat pada setiap tingkat kedalaman genangan, dipengaruhi oleh kondisi topografi rendah, banjir rob, dan konsentrasi bangunan bernilai ekonomi tinggi. Pendekatan risk-based pricing menghasilkan premi asuransi yang proporsional terhadap tingkat risiko banjir, sehingga peningkatan probabilitas kejadian banjir dan estimasi kerugian fisik akan diikuti oleh peningkatan besaran premi yang harus dibayarkan. Secara komparatif, premi di Jakarta Pusat sedikit lebih tinggi dibanding Jakarta Utara pada klaster risiko tertinggi akibat nilai kerusakan bangunan yang lebih besar. Penyesuaian coverage ratio menghasilkan premi asuransi banjir yang lebih realistis karena mencerminkan praktik underwriting, di mana proporsi pertanggungan tidak selalu meningkat seiring kenaikan risiko fisik. Secara keseluruhan, hasil skenario menunjukkan bahwa Skenario 3 yang melibatkan run-off-coefficient dan drainage loss merupakan dasar paling tepat untuk penetapan premi asuransi banjir berbasis risiko karena paling mendekati kondisi aktual.
======================================================================================================================================
Climate change in Indonesia is frequently manifested through increasing rainfall intensity. One of the most common natural disasters associated with high rainfall is flooding. Jakarta is among the regions in Indonesia that experience flooding most frequently. The high frequency and significant impacts of flooding have increased the need for financial protection mechanisms, one of which is flood insurance. In the implementation of flood insurance, accurate premium calculation is essential because premiums should reflect the level of flood risk faced by a particular area, ensuring fair protection for policyholders while maintaining the sustainability of insurance providers. This study analyzes flood risk by integrating the spatial characteristics of the study area with historical rainfall patterns in Jakarta, particularly Central Jakarta and North Jakarta. The study has three main objectives. First, it estimates the probability of flooding caused by extreme rainfall by combining temporal probability analysis, using extreme rainfall data and Extreme Value Theory, with spatial probability analysis based on the InaRisk flood vulnerability raster dataset. Second, it estimates flood-related losses in Central and North Jakarta, which are subsequently used as inputs for flood insurance premium calculations. Third, it calculates risk-based flood insurance premiums for both study areas. This study does not only focus on estimating the probability of flood occurrence due to extreme rainfall but also integrates flood risk analysis, physical loss estimation, and insurance premium calculation within a single flood risk assessment framework. The temporal analysis was conducted using Extreme Value Theory (EVT) with the Peak Over Threshold (POT) method to model the probability of extreme rainfall events. The spatial analysis was performed through flood vulnerability modeling using Geographic Information Systems (GIS) and raster-based analysis. Spatial vulnerability was derived by integrating social, economic, physical, and environmental factors represented as vulnerability raster layers. This approach enables a more detailed and representative identification of variations in flood risk across different areas. Physical losses were estimated using a raster-based flood inundation model under several hydrological scenarios, including extreme rainfall, runoff coefficient, and drainage loss parameters. The simulated flood inundation results were then integrated with building footprint data to quantify potential physical losses based on the extent of affected buildings and flood depth. Insurance premiums were calculated using a risk-based premium approach based on the Picard model, in which premiums are determined through the integration of flood occurrence probability and estimated physical losses. This approach allows premium values to reflect the level of flood risk and the potential economic impacts associated with flood events. The results obtained indicate that, under the temporal approach, the probability of at least one flood event caused by extreme rainfall exceeding 122 mm/day in Central Jakarta is 32.10% over a five-year period and 97.90% over a 50 year period. In North Jakarta, the probability of at least one flood event caused by extreme rainfall exceeding 126 mm/day is 21.80% over five years and 91.50% over 50 years. Under the spatial approach, Central Jakarta is classified as having high flood vulnerability, with an estimated annual flood probability of 0.067 and a 50 year probability of 0.888. North Jakarta is also classified as high risk, with an annual flood probability of 0.045 and a 50-year probability of 0.852. Physical loss modeling indicates that inundation depth, the affected building area, and hydrological variables such as the runoff coefficient and drainage loss are the most influential factors in determining total losses. Losses increase substantially with increasing flood depth. The extreme-flood cluster, with inundation depths exceeding 100 cm, produces the highest average losses despite affecting a relatively small number of buildings. In contrast, shallow flooding of 30 cm or less affects the largest number of buildings but produces comparatively lower losses. North Jakarta exhibits greater physical losses than Central Jakarta at each inundation-depth level, largely because of its low-lying topography, tidal flooding, and the concentration of high-value buildings. The risk-based pricing approach results in insurance premiums that are directly proportional to flood risk, whereby higher probabilities of flood occurrence and greater estimated physical losses are associated with increased premium levels. Comparatively, premiums in Central Jakarta are slightly higher than those in North Jakarta for the highest-risk cluster because of the greater value of building damage. Applying a coverage ratio produces more realistic flood insurance premiums because it reflects underwriting practices, in which the proportion of insured coverage does not necessarily increase in line with physical risk. Overall, the scenario analysis indicates that Scenario 3, which incorporates both the runoff coefficient and drainage loss, provides the most appropriate basis for risk-based flood insurance pricing because it most closely represents actual conditions and is consistent with regulatory principles.
| Item Type: | Thesis (Doctoral) |
|---|---|
| Uncontrolled Keywords: | Banjir, Curah Hujan, Extreme Value, Kerugian, Probabilitas, Premi Asuransi, Flooding, Rainfall, Extreme Value, Loss, Probability, Insurance Premiums |
| Subjects: | G Geography. Anthropology. Recreation > GE Environmental Sciences > GE300 Environmental management |
| Divisions: | Faculty of Civil, Planning, and Geo Engineering (CIVPLAN) > Environmental Engineering > 25001-(S3) PhD Thesis |
| Depositing User: | Pratnya Paramitha Oktaviana |
| Date Deposited: | 22 Jul 2026 02:26 |
| Last Modified: | 22 Jul 2026 02:26 |
| URI: | http://repository.its.ac.id/id/eprint/136410 |
Actions (login required)
 |
View Item |