Two preprocessing algorithms for climate time series

Area of Science:

• Environmental science
• Data science

Background:

• Climate time series data often contain outliers that can affect analysis.
• Existing preprocessing methods may not be optimal for all types of outliers.

Purpose of the Study:

• To propose and evaluate two novel algorithms for preprocessing climate time series data.
• To compare the performance of the proposed algorithms against existing methods.

Main Methods:

• Autoregressive cost update mechanism for outlier detection.
• Wavelet transform for outlier detection, a pattern recognition technique.
• Benchmarking using synthetic datasets and real-world temperature measurements.

Main Results:

• The autoregressive method is effective for solitary outliers.
• The wavelet-based method is superior for multiple consecutive outliers.
• Combined application of both methods provides a robust signal preprocessing tool.

Conclusions:

• The proposed algorithms offer effective solutions for climate time series preprocessing.
• The choice of algorithm depends on the nature of the outliers (solitary vs. consecutive).
• These methods enhance the reliability of climate data analysis.