Manfaat Prediksi Cuaca Jangka Pendek Berdasarkan Data Radiosonde dan Numerical Weather Prediction (NWP) untuk Pertanian Daerah

Indonesia as a continent- maritime has a climate characteristics in different regions. Information on fast and accurate weather forecasts to be one important needed an effective method of forecasting the weather. The study was based on radiosonde data and Numerical Weather Prediction (NWP) using observational data at 00 Universal Time Coordinates (UTC) and 12 UTC. Radiosonde data analysis performed on the surface variables and K-index that are used to study the stability of the atmosphere and the potential of a thunderstorm. The results of the analysis of radiosonde data showed fluctuations in the surface variables during the observation and the highest K-index in the rainy season. NWP method using 4 variables predictant and 9 predictor variables. The analysis was done by reducing the dimensions of a predictor variable on the horizontal dimension (grid) and higher dimensions (levels) by using the Principal Component Analysis (PCA) and multivariate regression analysis. Weather prediction shows a fairly relevant results between observational data and the data shown by the allegations that the correlation value by 67%