Unsupervised detection of climate anomalies with Python

A step-by-step tutorial using the Google Earth Engine Python API and pycaret library

The increasing frequency of climate anomalies is an expected consequence of global warming. In this context, analytical tools that can identify outliers in time series are very useful, as they allow more accurate inferences about the impact of extreme weather events.

Unsupervised machine learning methods, of which Isolation Forest is widely used, are valuable tools for this purpose.

This tutorial uses a monthly time series of Climate Water Deficit values extracted from the TerraClimate dataset via Google Earth Engine’s Python API. A valid Google Earth Engine account is required for this step.

That’s it! Based on this analysis, we can see that for the period from 2000 to 2021, the anomalies in the climatic water deficit time series for the region under consideration have become relatively more frequent since December 2018.

Note: This tutorial is intended for didactic purposes only. It’s only about exploring the given techniques and tools.

For more information, see:

Related posts:

This is very nice project and his quality is satisfy. I believed that this project is successful very quickly and great for us. i hope, it will be success for future plan. Project is very good . I am…

Quick fact to start off the article-Bitcoin routinely drops around 85% of it’s value every few years. It then goes on to make new all time highs around the time of the next upcoming halving. The next…

Was the best at KFDA so far.. “Le Cercle” is published by Eylül F. Akıncı in choreodramaturgy.