Abstract
Studies of the global mean surface temperature trend are typically conducted at a single (usually annual or decadal) time scale. The used scale does not necessarily correspond to the intrinsic scales of the natural temperature variability. This scale mismatch complicates the separation of externally forced temperature trends from natural temperature fluctuations. The hiatus of global warming since 1999 has been claimed to show that human activities play only a minor role in global warming. Most likely this claim is wrong due to the inadequate consideration of the scale-dependency in the global surface temperature (GST) evolution. Here we show that the variability and trend of the global mean surface temperature anomalies (GSTA) from January 1850 to December 2013, which incorporate both land and sea surface data, is scale-dependent and that the recent hiatus of global warming is mainly related to natural long-term oscillations. These results provide a possible explanation of the recent hiatus of global warming and suggest that the hiatus is only temporary.
Highlights
The downward part of a long-term natural oscillation can give us a picture of the global warming pause. Is this the cause of the recent hiatus? In addition, temperature change rates, commonly used in global warming studies, are sensitive to outliers in temperature time series resulting from small-scale variability or events. Is this small-scale variability one of the reasons for the hiatus of global warming? Here we attempt to answer these questions with the Multi-Resolution Analysis (MRA) and the wavelet power spectrum analysis of global surface temperature anomaly (GSTA) time series
The DWT based MRA decomposes the original series into a low frequency component and a high frequency component and this process can be repeated for the approximation part at various levels, producing a series of overall pictures of the geophysical time series at various scales and detail parts corresponding to various decomposing levels or scales as well
MRA undoubtedly increases the chance of scale match between intrinsic scales of a natural process or pattern and the data analysis scale in that many scales rather than a single scale are used in multi-resolution analysis
Summary
We use the monthly mean GSTA (relative to the 1961–1990 mean) time series, which incorporates both land surface air temperature and sea surface temperature data and spans the time period from January 1850 to December 2013, to study the global mean surface temperature trend. The continuous Wavelet transform, which decomposes a time series into time-frequency space, can be used to determine both the dominant modes of variability and how those modes vary in time by means of Wavelet power spectrums[11] These two functions of wavelet analysis are useful in this study in that they can help us gauge if the recent hiatus affects the overall global warming trend when observed at bigger temporal scales and if the hiatus is related to the existence of some small-scale (high frequency) natural oscillations (e.g. ENSO with a return period of 2–7 years) in the last two decades. The period of the oscillation determined by the Mexican hat wavelet at the scale s is about 4 s, which is used to estimate the duration of oscillation determined by the wavelet power spectrum using the Mexican hat wavelet in our study
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
