Wavelet Analysis of Average Monthly Temperature New Delhi 1931- 2021 and Forecast until 2110

Abstract

The identification method in the CurveExpert-1.40 software environment revealed asymmetric wavelets of changes in the average monthly temperature of New Delhi from 1931 to 2021. The maximum increment for 80 years of the average monthly temperature of 5.1°C was in March 2010. An analysis of the wave patterns of the dynamics of the average monthly temperature up to 2110 was carried out. For forecasting, formulas were adopted containing four components, among which the second component is the critical heat wave of India. The first component is the Mandelbrot law (in physics). It shows the natural trend of decreasing temperature. The second component increases according to the critical law. The third component with a correlation coefficient of 0.9522 has an annual fluctuation cycle. The fourth component with a semi-annual cycle shows the influence of vegetation cover. The warming level of 2010 will repeat again in 2035-2040. From 2040 the temperature will rise steadily. June is the hottest month. At the same time, the maximum temperature of 35.1°C in 2010 in June will again reach by 2076. But according to the second component of the heat wave, the temperature will rise from 0.54°C to 16.29°C. The annual and semi-annual cycles had an insignificant effect on the June temperature dynamics. Thus, the identification method on the example of meteorological observations in New Delhi made it possible to obtain summary models containing a different number of components. The temperature at a height of 2 m is insufficient. On the surface, according to space measurements, the temperature reaches 55°C. As a result, in order to identify more accurate asymmetric wavelets for forecasting, the results of satellite measurements of the surface temperature of India at various geographical locations of meteorological stations are additionally required.