CD Skripsi
Pemodelan Peramalan Curah Hujan Menggunakan Statistical Downscaling Dan Principal Component Regression (Pcr) Di Kabupaten Siak
Indonesia is a country with a tropical climate that has rainy and dry seasons, so it requires efforts to cope with climate change against the handling of agricultural crops, especially palm coconut. Weather forecast modeling could use data from the Global Circulation Model (GCM). Statistical downscaling is a method of adapting global circulation models to variable data on a local scale. The problem of multicolinearity is a common one in statistical downscaling modeling. A method that can be used in dealing with this multicolinearity problem is the use of Principal Component Regression (PCR) by eliminating unstable structures in the model and reducing the variance of the regression coefficient. This research aims to obtain rainfall prediction modeling using statistical downscaling and PCR, which can provide an understanding of weather change preparedness. The observation data used is rainfall data located at the location of Libe Estate of PT. SMART Tbk Division SMART Research Institute in 2013–2022, as a response variable. Predictor variables use GCM output data with CMIP6 simulation. The study shows that PCR modeling with RMSE 97.06–131.69 has a R^2 value of around 14.25%-20.49%. PCR modeling with dummy variables can improve the performance of models with a RMSE value of only 24.46 - 35.83 with R^2 value of 89.02% - 90.24%.
Keywords: Statistical downscaling, principal component regression, global circulation model (GCM), multicolinearity, rainfall.
Tidak tersedia versi lain