Connections!

Could you image making 190 of these plots? A correlation matrix does just that in 1 pretty figure!

Figure 1: Could you image making 190 of these plots? A correlation matrix does just that in 1 pretty figure!

One of my favorite packages in R is Corrplot. It allows you to create really cool visual correlation matrices so you can quickly see what parameters are significantly correlated to each other.

Okay, let me translate that! In our Plot of the Week, I have a 19 different time series and I want to know which are correlated to each other (Table 1).  Your first thought might be to plot each parameter against each other (such as Figure 1 showing Frost Days versus Growing Season Length)…but I would have to make 190 plots in order to test all 19 time series against each other! What a pain!

Table 1: The 19 time series used in our Plot of the Week.

Table 1: The 19 time series used in our Plot of the Week.

Instead, a correlation matrix calculates the R (Pearson correlation coefficient) of each time series combination and arranges it in a visual graphic (Figure 2).

So this plot of the month is showing 190 correlations…but in a much more helpful (and colorful) way!

Plot of the Week: A Corrplot of the Annual Indices!

Our plot of the week is a correlation matrix investigating the correlations between the annual climate extreme indices calculated for the near-shore Chesapeake Bay region and four teleconnection indices (NAO, AMO, PDO, and MEI).

I have my R script set up to only plot correlations which are significant…so if a colorful circle is plotted, there is a statistically significant linear trend. The bigger the circle, the strong the trend! And, blue is a negative trend while red is a positive trend!

Figure 2: The Plot of the Week, a correlation matrix!

Figure 2: The Plot of the Week, a correlation matrix!

So what do you see?

Not counting the time series when plotted against each other (which is a perfect linear correlation!), there were 72 significant relationships. So 48% of the total correlations made in this plot had a relationship!

Since I don’t want to overwhelm you with 72 different correlations, I will briefly a few that I personally find interesting! Hopefully you can take a look at this correlation matrix and find a relationship or two that you find neat!

Years with a strong AMO are likely to have fewer Frost Days and more Summer Days and Tropical Nights.

Years with a lot of Icing Days are likely to also have many Frost Days, longer colds spells, and a shorter growing season. That makes sense!

Years with a high quantity of total precipitation are likely to have fewer Frost Days and dry spells and more likely to have many days with at least 20 mm of precipitation.

These correlations matrices are fun to make and really helpful when investigating trends! You will probably see more of these! (Hint, Hint)

 

 

Kari Pohl

About Kari Pohl

I am a post-doctoral researcher at NOAA and the University of Maryland (Center for Environmental Science at Horn Point Laboratory). My work investigates how climate variability and extremes affect the diverse ecosystems in Chesapeake Bay. I received a Ph.D. in oceanography from the University of Rhode Island (2014) and received a B.S. in Environmental Science and a B.A. in Chemistry from Roger Williams University (2009). When I am not busy being a scientist, my hobbies include running, watching (and often yelling at) the Boston Bruins, and taking photos of my cat.
This entry was posted in Uncategorized. Bookmark the permalink.

Leave a Reply

Your email address will not be published. Required fields are marked *