As I continue to work with global air temperature data, I am learning that global warming has several characteristics that I did not expect. The usage of Alteryx and Tableau allow me to demonstrate what I mean by that statement.
In this article, I focus on maximum air temperature changes that have happened in the US and Canada. I provide a video in which I visualize decade-by-decade changes in state-based average monthly maximum temperatures. The patterns of warming and cooling that I have visualized are both spatially and temporally variable, and surprising in many ways. This analysis indicates that we need to think critically about what the term “global warming” really means.
Here are the links to the previous three articles on this topic:
Part 2 – Temp changes over the past 50 years (the precursor to this article)
The Tableau dashboards shown in this article are a result of processing a lot of data this past weekend. I went back and downloaded the entire contents of ghcnd (2.98 Gb) and re-processed the data from over 5.3K monitoring stations. I did this so that I could include the entire year of 2016 in this analysis. The amount of data included is summarized:
- Data from 5,259 world-wide monitoring stations were extracted. These stations have Tmax data from 1960 to now (see the final figure of this article);
- Data from 3,221 monitoring stations in the US and Canada were included. These were processed to complete the work included in this article;
- Over 58 million daily Tmax readings were processed from the US/Canada stations from 1960 to now. These are the readings that have no data quality problems.
Several Alteryx workflows were used to crunch the data and Tableau was used to visualize the results shown in this article. The Alteryx workflows I developed and used are shown here. I realize that I can no longer even think about working without Alteryx and Tableau. My brain has never experienced recreational drugs. However, it knows something better than drugs, and that is the endorphine release created by the explosive combination of Alteryx+Tableau.
What I Have Learned About Max Air Temp Changes
I have now learned that not all areas in the US and Canada have undergone heating over the past 50 years. There are many examples of large sections of the US where cooling has occurred over the 50-year time scale of analysis. These zones are best visualized in the monthly-based dashboards shown below.
As I continue to think about this work, it is clear that there are many factors that have had an impact on maximum air temperatures. It is clear that in the general sense, many places have heated up. However, there are many zones where maximum monthly air temperatures have fallen. The atmosphere is not in a state of thermodynamic equilibrium. I think it is important to realize that the atmosphere is not uniformly heating up.
In the video shown below, I discuss some of these findings. Of particular importance is the pervasive and ubiquitous heating demonstrated in the months of March and September. Upon further review, December and January can also be included in that category.
On the opposite side of the equation, are October and November, both of which have exhibited significant cooling periods over time. Of 10 potential decades of measured changed, these two months have been in cooling conditions for 8 of the decades (see the third table below).
Tabular and Graphical Results
I’m not going to spend a lot of time discussing these tabular results. The table titles are self-explanatory. The dashboards shown below summarize the work that was done, on a month-by-month basis. I included these tables to help readers realize how much data was used to complete this analysis.
Number of Daily Tmax Readings
Number of Tmax Monitoring Stations
Average Daily Tmax Values by Month and Decade
Average Daily Tmax Change by Month and Decade
Note that the last column shown in this table, for the 2010’s decade, is the degree of cooling or warming that has occurred since the 1960’s. You will find these numbers in each of the 12 dashboards shown below, as the reference line in the lower right corner scatter plots. Thank goodness for Tableau level of detail calculations.
Dashboards of Monthly Tmax Change From the 1960’s to the 2010’s
Over 1.5 million daily readings of Max daily temperatures were used to compute each of these dashboards. I got really good computing on my son’s abacus. In total, over 58 million measurements were used to compute the intermediate results for the 1970’s, 1980’s, 1990’s, and 2000’s.
Between 2,900 and 3,200 monitoring stations were used to create the spatial coverages. The data is superb and the trends are what really happened. There can be no arguing about the validity of these calculations and results.
To see the initial work I did that was based on analysis of the actual monitoring station data, please refer to this article. It was this initial analysis that indicated to me that this approach would produce very nice insights into the warming and cooling trends shown below.
The clustering of warming and cooling zones was very pronounced even in the point (i.e., monitoring station) data. The presences of those large-scale clusters were totally unexpected, they blew me away, and they confirmed the validity of this data and of this approach.
January (2010’s – 1960’s)
February (2010’s – 1960’s)
March (2010’s – 1960’s)
April (2010’s – 1960’s)
May (2010’s – 1960’s)
June (2010’s – 1960’s)
July (2010’s – 1960’s)
August (2010’s – 1960’s)
September (2010’s – 1960’s)
October (2010’s – 1960’s)
November (2010’s – 1960’s)
December (2010’s – 1960’s)
If anyone wants the 645 Mb file that was used to create this work, you can get it here. The zip file (88 Mb) contains the monthly aggregated Tmax/Tmin records for the 5,259 world-wide monitoring stations. The entire time series history of each station is included in this 4.66 million record file, with all data quality issues removed.
My Tableau workbook that was used to produce the graphical output for this work will be placed on my Tableau public website soon. There are some fancy techniques buried in that workbook for those looking for some clever techniques.
I think that the Makeover Monday members should take this data file and go to work. There are about 1 million findings to be had in this data. Since I did all the hard work making this data file, they can reap the benefits of contributing to the knowledge base of global warming/cooling. It would be bad-to-the-bone if that could happen. @mix_pix are you game?