Background Accurately identifying local information in wind speed time series is crucial for revealing the intrinsic patterns of wind speed variations. Hami city, a typical region for wind speed research, offers valuable insights into local wind speed trends. Previous studies have primarily focused on large-scale wind speed changes at national or regional levels, lacking precise guidance on small-scale variations. Additionally, these studies often relied on wavelet analysis method, resulting in high costs and low efficiency. Building on prior research, this paper introduces new methods to explore small-scale wind speed variations, providing essential data support and a methodological foundation for soil and water conservation efforts.

Methods The daily average wind speed data from 1960 to 2019 in Hami city were selected, and the final data was obtained through outlier processing and interpolation. The overall trend of the average wind speed in Hami over the years was analyzed using the linear trend method, and the Mann-Kendall method was used to test for abrupt changes in the average wind speed sequence. Empirical mode decomposition was used to identify the inherent change characteristics of the wind speed sequence and reconstruct it.

Results 1) The average wind speed in Hami city has shown a decreasing trend in the past 60 years, with a decrease rate of 0.0301 m/(s·a), which is consistent with the trend of surface wind speed changes in China and even globally. However, there has been a slight increase since 2000, and the long-term changes in wind speed have a certain degree of abrupt change. The annual average wind speed experienced abrupt changes in 1981 and 1986. 2) There are two typical inherent time scales for wind speed series: interannual scale and decadal scale. Among them, decadal oscillation dominates the wind speed variation, with 30 years as the main variation period. 3) The VMD method is effective in identifying the nonlinear trend of the average wind speed in Hami over the years. The decomposed and reconstructed wind speed sequence can accurately represent the long-term trend and short-term waveform characteristics of Hami wind speed, providing a new method for wind speed research.

Conclusions By analyzing the changes in average wind speed in Hami city, this study validates the effectiveness of the VMD method in capturing nonlinear trends and multi-scale oscillations in wind speed data. It provides a new approach for wind speed research, enhances our comprehensive understanding of regional wind speed variations, and holds significant importance for improving the ecological environment of Hami city.