Statistical Analysis of Long-Term Climate Data – Case Study on the Western Part of Romania between 1961 and 2020

Abstract

Climate change is a reality, manifesting on a global scale, and with variable effects on the environment, human health, performance and socio-economic balances. This study analyzed climate data (temperature, precipitation) over a period of 60 years, between 1961 and 2020. The climate data were recorded in the Western area of ​​Romania, at the Arad Meteorological Station and the Sannicolau Mare Meteorological Station. The multiannual calculated mean value over the study period for temperature (T, °C) was = 10.92±0.11 °C, and the multiannual average value calculated for precipitation (P, mm) was = 564.54±14.17 mm. In the case of temperature, the annual mean values, compared to the multiannual mean value, showed differences between DT = -1.44 °C in the case of 1980 year, and DT = 1.76 °C in the case of 2019 year. In the case of precipitation, the annual values, compared to the multiannual mean value, showed differences between DP = -289.04 mm in the case of 2000 year, and DP = 234.56 mm in the case of 2010 year. In the first half of the interval (1061 – 1990) only three years (1961, 1966 and 1990) presented the annual mean temperature above the multiannual mean, within statistical safety conditions (p<0.001, ***). In the second half of the interval (1991 – 2020) themperature values ​​above the multiannual mean were recorded in the case of 16 years, in statistical safety conditions. In the decade 1991 – 2000 three years presented temperature values ​​above the multiannual mean (1992, 1994, and 2000), in the decade 2001 – 2010 four years (2002, 2007, 2008 and 2009) presented temperature values ​​above the multiannual mean, and in the decade 2011 – 2020 nine consecutive years (2012 – 2020) presented temperature values ​​above the multiannual mean. Multivariate analysis (PCA) explained the positioning, mode and action intensity of the monthly periods in the study interval according to the principal components, based on the the values ​​of climatic factors (T, °C; P, mm). Cluster analysis grouped the years in the study interval based on similarity in relation to the values ​​of climatic parameters. Sinusoidal models described the variation of temperature (p = 0.00026) and precipitation (p = 0.5955) over the entire study interval. Ranking analysis ranked the years in the study interval in relation to temperature values ​​and quartile threshold.