Pemetaan Banjir Rob Wilayah Medan Utara Menggunakan Artificial Neural Network dan GIS

Abstract

Large cities in Indonesia, such as Jakarta, Surabaya, Semarang and Medan are vulnerable to tidal flooding, especially in the northern region which borders directly with sea waters. Thousands of residents in the northern area of Medan experienced tidal flooding due to high tides that submerged their settlements. This study was carried out to meet the need for tidal flood zone mapping based on the factors causing tidal flooding in the North Medan area as a basis for stakeholders in the context of handling tidal flood mitigation. Vulnerability factors to tidal flooding include data on rainfall, drainage density, land use (land use), distance to the river, soil type, elevation, slope, aspect, distance to the estuary. Data analysis using Geographic Information System (GIS) and Artificial Neural Network (ANN). The research location covers the sub-districts of Medan Belawan, Marelan and Medan Labuhan. The results of the ANN modeling with the MAPE formula showed the accuracy of the experimental train 1 data was 64.57% and the experimental test da2ta 1 was 75.75%. Meanwhile, the experimental train data 2 was 70.36% and the experimental test data 2 was 75.54%. The experimental train data 3 is 71.65% and the experimental test data 3 is 50.88%. The experimental train data 4 was 70.54% and the experimental test data 4 was 78.55%. The resulting flood susceptibility map can be used for tidal flood mitigation planning in the three sub-districts.

Kota besar di Indonesia, seperti Jakarta, Surabaya, Semarang dan Medan terancam terhadap banjir rob, khususnya di wilayah utara yang berbatasan langsung dengan perairan laut. Ribuan warga di kawasan utara Medan mengalami banjir rob akibat pasang air laut yang merendam permukiman mereka. Studi ini dilaksanakan untuk memenuhi kebutuhan akan pemetaan zona banjir rob berlandaskan faktor-faktor penyebab banjir rob di wilayah Medan Utara sebagai dasar bagi pemangku kepentingan dalam rangka penanganan untuk mitigasi banjir rob. Faktor kerawanan terhadap banjir rob mencakup data curah hujan, drainage density, land use (tata guna lahan), jarak ke sungai, jenis tanah, elevasi, kemiringan, aspek, jarak ke muara. Analisis data menggunakan Geographic Information System (GIS) dan Artificial Neural Network (ANN). Lokasi penelitian mencakup kecamatan Medan Belawan, Marelan dan Medan Labuhan. Hasil pemodelan ANN dengan rumus MAPE menunjukkan akurasi data train percobaan 1 sebesar 64,57% dan data tes percobaan 1 sebesar 75,75%. Sementara data train percobaan 2 sebesar 70,36% dan data tes percobaan 2 sebesar 75,54%. Data train percobaan 3 sebesar 71,65% dan data tes percobaan 3 sebesar 50,88%. Data train percobaan 4 sebesar 70,54% dan data tes percobaan 4 sebesar 78,55%. Peta kerawanan banjir yang dihasilkan dapat digunakan untuk perencanaan mitigasi banjir rob di tiga kecamatan tersebut.