Pemodelan Statistical Downscaling Dengan Projection Pursuit Regression Untuk Meramalkan Curah Hujan Bulanan Di Sentra Produksi Padi Jawa Timur

Asyeifa, Vella Rochmana (2017) Pemodelan Statistical Downscaling Dengan Projection Pursuit Regression Untuk Meramalkan Curah Hujan Bulanan Di Sentra Produksi Padi Jawa Timur. Undergraduate thesis, Institut Teknologi Sepuluh Nopember.

[img]
Preview
Text
1313100073-Undergraduate_Theses.pdf - Published Version

Download (3MB) | Preview

Abstract

Peramalan curah hujan di wilayah yang memiliki iklim tropis, seperti Indonesia, memiliki permasalahan yang kompleks. Topografi dan interaksi antara lautan, daratan dan atmosfer yang kompleks mempersulit prediksi curah hujan di Indonesia. Oleh karena itu, diperlukan model peramalan curah hujan yang akurat pada skala lokal dengan mempertimbangkan informasi tentang sirkulasi atmosfer global yang didapatkan dari luaran General Circulation Model (GCM). GCM bisa digunakan untuk memperoleh informasi dalam skala lokal dengan melakukan statistical downscaling. Statistical downscaling merupakan model basis regresi untuk menentukan hubungan fungsional antara variabel respon dan variabel prediktor. Variabel respon yang digunakan dalam penelitian ini adalah observasi curah hujan, sedangkan variabel prediktor adalah iklim global luaran dari GCM. Penelitian ini dilakukan di sentra produksi padi Jawa Timur, yaitu Kabupaten Ngawi, Lamongan, Bojonegoro, Jember dan Banyuwangi. Metode statistical downscaling yang digunakan adalah Projection Pursuit Regression (PPR). Dalam pemodelan PPR optimalisasi dilakukan sebanyak lima kali dengan simulasi banyak fungsi 1 sampai 5. Validasi model dilakukan dengan kriteria Root Mean Square Error Prediction (RMSEP) terkecil. Pola antara hasil ramalan dan observasi menunjukkan bahwa hasil ramalan curah hujan di lima kabupaten mendekati data observasi. Sehingga model yang diperoleh merupakan model yang baik untuk meramalkan curah hujan di lima kabupaten. ============================================================ The prediction of rainfall in the area with the climate tropical, such as Indonesia, is hard to do. The Topography and interaction complex between sea, mainland, and atmosphere makes the prediction of rainfall in Indonesia complicated. Therefore, the prediction models precipitation accurate on the local scale consider the information about the circulation global atmosphere established from General Circulation Model (GCM) are required. GCM can be used to acquire the information in the scale of the local or regional by doing statistical downscaling. Statistical downscaling is a regression based model to determine the functional relationship between variable response and variable predictors. In this research variable response uses data of rainfall observation, meanwhile variable predictor uses climate scale global from GCM. The research of rainfall take place in rice production center in East Java, included Ngawi, Lamongan, Bojonegoro, jember and Banyuwangi regency. The exact statistical downscaling method in this research is Projection Pursuit Regression (PPR). PPR modeling have five times optimizing with simulation function 1 until 5. Validation model criteria use Root Mean Square Error Prediction (RMSEP). The best model choose based on how many function which has the smallest RMSEP value. The pattern between forecasting and observation shows the result of rainfall forecasting in five regency are close to the data observation. So the optained model is a good model to forecast rainfall in five regency.

Item Type: Thesis (Undergraduate)
Additional Information: RSSt 519.536 Asy p
Uncontrolled Keywords: GCM, PPR, rainfall forecasting, statistical downscaling
Subjects: H Social Sciences > HB Economic Theory > Economic forecasting--Mathematical models.
Divisions: Faculty of Mathematics and Science > Statistics > (S1) Undergraduate Theses
Depositing User: Vella Rochmana Asyeifa
Date Deposited: 13 Feb 2018 07:29
Last Modified: 05 Mar 2019 08:39
URI: http://repository.its.ac.id/id/eprint/47825

Actions (login required)

View Item View Item