Indonesia is a country with a population that still uses the agricultural sector as the highest livelihood. The Central Statistics Agency said 38.61% were in this sector. However, the issue of climate change caused by an increase in global temperature has begun to impact the agricultural sector in several regions. Therefore, this study aims to understand the effect of increasing surface temperature anomaly on the Oldeman climate pattern in Dumai City. The data used are in-situ observation rainfall data for 1981 -2010, rainfall reanalysis data for 1991 -2020, surface temperature anomaly data for 1991 -2020. The data validation methods used are spearman correlation and Kendall-tau methods. The approach to explore changes in the Oldeman climate pattern uses time series analysis techniques such as ACF, PACF, and moving average of 1 year lag-time with a period of 10 years. The rainfall data validation test result show that the average spatial correlation value is 0.86. It can be used to explore time-series climate data in 3 locations in Dumai. As a result, time series analysis found a positive trend in rainfall data significantly along with the temperature anomaly but the Oldeman's climate pattern analysis in Dumai is monitored as stable in type B1. Overall, it is concluded that the change in the surface temperature anomaly variability and trend as a manifestation of climate change in Dumai has not impacted in the Oldeman agricultural climate patterns.

The limitations of in-situ observations of the upper air are one of the obstacles in analyzing the weather. The use of data models can be one solution. The purpose of this study was to determine the accuracy of the data model in providing upper air information using RAOB as a visualization tool for aerological diagrams and sounding information analyzers. The data used are radiosonde observation data from the Cengkareng meteorological station and 1000 – 100 mb ECMWF pressure level models at the same location as the in-situ observations. The time chosen is when the haze and mist occur at the observation time 00 UTC for 5 events each. The method used is pearson correlation and simple visual verification. The results obtained that the correlation of the significant point plot data diagram when the mist occurs is 0.76 and when the haze occurs is 0.67 and visually as a whole show that the model data is quite close to the observation data. Correlation of 59 sounding information as a whole produces a value of 0.85 – 0.99 when the Mist occurs and a value of 0.89 – 0.99 when the Haze occurs. It is hoped that these results can be used as a consideration for the use of data models in filling in the gaps in radiosonde observation data. Keywords: Sounding Information, RAOB, Radiosonde, ECMWF Models