Introduction
My global warming research continues to surprise me in ways I never expected. In this article, I share insights on the timing of global warming development.
I recently wanted to quantify how many years of data are needed for my global warming simulator to accurately predict the temperature change patterns I detected in 2017. Would 10, 20, 30, or more years of data be needed to quantify these patterns of change, I wondered?
In previous unpublished work, I investigated the sensitivity of the starting year of my global warming simulations. I compared starting the global warming models in 1960, 1970, and 1980. The work results indicated that I could use any starting point, but 1980 gave the most consistent temperature change patterns. For this reason, I now use 1980 as the starting point of my global warming models.
Background
I conducted this research to answer a simple question. I wondered if I could have detected the patterns of temperature change if I had started this work in 1995, 2006, or even 2010. Since my work saw the patterns of temperature changes using up to six decades of data, I wanted to identify the minimum number of years of data needed for my simulator to predict the temperature change patterns accurately. As previously mentioned, I now start the global warming models in 1980.
The Analysis
To complete this work, I wrote a new Alteryx workflow. This workflow allowed me to run the global warming simulator between arbitrary starting and ending dates. For example, I could simulate temperature changes from 1980 to 2000 (a 20-year simulation), or I could run from 1980 to 2022 (a 42-year simulation). This flexibility has helped me test my simulation approach in ways that were cumbersome to accomplish before I wrote the workflow.
This workflow took over 12 hours to create nearly 42 million global warming models, spanning from 20 to 42 years (Figure 1). Without the superb computational speed of Alteryx, this would have been a very long process to complete.
Processing the Output
I could make this a very long article. I choose not to, so I will keep this simple.
The mathematical approaches I have built to simulate future temperatures need nearly 30 years of data before stable temperature predictions are achieved (Figure 2). This result means that if I had daily temperature data starting in 1980, it would have taken until about 2010 before the patterns of temperature change I have been writing about would have been detectable in the configurations I have shown.
I guess I was lucky that my work evolved when it did. It took me years to realize its significance. Now I have a new technique that will help me to understand where global warming patterns are stable and where there might be accelerations or decelerations in the time rate of temperature change. With this new simulation capability, I will attempt to detect areas that exhibit significant temperature accelerations (getting hotter or cooler faster) or decelerations (the temperature rate of change is slowing).
Final Thoughts
I doubt that many people have read this article. If you have, thank you! Global warming impacts all of us and is something we need to understand better. For example, if you live in Germany or France, Figure 3 shows how significantly daily maximum temperatures change per decade in April compared to what is happening in the US and Canada. The technique I used in this work gives me even more confidence in my research.
