Page 61 - 全球气候变化及其影响Global Climate Change and Its Impacts-185×260

Chapter II Evidence for Global Climate Change

cipitation that might affect trend judgment.
The applicability of the moving average method is reflected in its smoothing effect on
data. It can effectively filter out short-term noise and abnormal fluctuations, making long-
term trends more prominent, and is particularly effective for analyzing climate data with
pronounced seasonality or significant short-term variability. When analyzing monthly aver-
age temperature data, the moving average can eliminate random fluctuations in temperature
between months, better revealing the long-term temperature change trends over time. More-
over, the moving average method is relatively simple, does not require complex mathemati-
cal models or assumptions, and has low computational costs. However, the moving average
is not without flaws. The choice of window length significantly impacts the results. If the
window is too short, it may fail to adequately smooth the data, leaving long-term trends still
obscured by short-term fluctuations. If the window is too long, it may over-smooth the data,
losing some important short-term variation information and responding sluggishly to trend
changes. Additionally, the moving average merely smooths the data without conducting an
in-depth analysis of the trend’s underlying nature. It cannot provide explicit quantitative met-
rics like the rate of trend change, as linear regression does.
In addition to linear regression and moving averages, other methods can be used for
identifying and analyzing long-term climate change trends, such as wavelet analysis and
empirical mode decomposition. Wavelet analysis can decompose data at different time scales
while preserving both temporal and frequency information, offering advantages for analyzing
climate data with complex periodicities and non-stationary characteristics. Empirical mode
decomposition candecompose time series data into multiple intrinsic mode functions (IMFs),
each representing fluctuation components at different temporal scales, facilitating deeper un-
derstanding of the multi-scale characteristics of climate change. In practical applications, it
is often necessary to integrate multiple methods according to specific data characteristics and
research objectives, cross-validate and complement each other, to more accurately identify
and analyze long-term climate change trends.

II. Division and Basis of Climate Change Characteristics in Different
Stages

Climate change is an extremely complex and dynamic process that encompasses the
interactions and feedback mechanisms among multiple spheres within the Earth system,
including the atmosphere, oceans, and land. To gain deeper insights into the inherent laws
underlying this process, it is of paramount significance to classify its phases based on a se-
ries of critical factors such as the rate, magnitude, and impact intensity of climate change.
This classification approach not only helps to clearly present the complete picture of climate
change across different temporal scales, but also provides a systematic and comprehensive
perspective for in-depth research on climate change and the formulation of effective response
strategies.

• 53 •