Exploring the Mt. St. Helens Succession Study

Studies of Mt. St. Helens dramatically altered how ecologists understand succession. For this project, I used data from the Mount Saint Helens succession study conducted by Roger del Moral.  This is the longest continuous record of vegetation recovery on a new volcanic surface.

Multivariable Comparisons: Time-Series vs Scatter Plot

Understanding the species diversity of an environment is a fundamental concept in ecology. The two main measures of this are richness and evenness. Richness is the number of species in a given area, and evenness accounts for both the number of species in a given area and the distribution or abundance of each species. To account for the difference between richness and evenness, ecologists calculate a diversity index called the Shannon-Wiener H’:

where  is the proportion of individuals in species i in an environment.

Suppose that I would like to compare the average rainfall and the percent growth of H’ per year in a given plot. Below, I compare the advantages and disadvantages of two graphical types: time-series charts and scatter plots.

Time-Series

The first set of graphics use time-series charts to compare average rainfall and H'. The first display is a dual axis time-series chart filtered by plot name.  The problem is that this graphic limits our ability to compare the time-series for multiple plots.  If I were to filter for only Abraham Plain and Lahar, the paths for Average H’ and average rainfall would be adjusted to reflect only those two plots.

In order to compare between each plot more effectively, I chose to use small-multiples indexed by plot name. In small-multiples, a series of graphics is indexed by changes in a given variable. Edward Tufte was particularly fond of this design, since it is inherently multivariate and comparative. The advantage of a small-multiple display in this situation is clear. They allow us to directly compare the average rainfall and H’ of different plots.

As you can see, one of the design difficulties I experienced in developing this was that the final display is organized in a single column. If I were to continue designing this graphic, I would have liked to display the small multiples in more than a single column. Since there are twelve different plots, I would have preferred a 3x4 grid.

Is a time-series the best way to compare these two variables?  One flaw of comparing H’ and average rainfall using a time-series is that this display limits our ability to compare and identify relations between the variables. According to Edward Tufte, “The problem with time-series is that the simple passage of time is not a good explanatory variable.” I should also point out that line charts give the impression that our “Year” variable is continuous, when it is actually discrete. To show the significance of this, I’ve included displays showing the time-series with a discrete “Year” variable.

Let’s look at the same data using a chart style that is more geared towards identifying relationships and correlations. 

Scatter Plot

The second set of graphics use scatter plots to compare average rainfall and H’. Edward Tufte considered the scatter plot to be the greatest of all graphical designs. It links multiple variables and allows the viewer to assess the possible causal relationship between the plotted variables. It confronts causal theories with empirical evidence.

The first scatter plot is filtered by year. When I tried filtering by plot, I struggled to set up trend lines that took year into consideration.  I included this scatter plot in the storybook to show this problem. Additionally, this shows that the “Year” variable is discrete. Similar to the time-series, I  have also included a small-multiple display. 

Average Growth Percent of Nitrogen Fixers

In addition to the above displays, I have also included a display comparing the average plant growth percent of nitrogen fixing plants and other plants included in the study. Lupines Lepidus is a nitrogen fixing plant which was a key player in succession on Mt. St. Helens.

Part of this display is a map with the labeled sites and colored by Impact Type. This allows the viewer to see the geographic location of individual sites in this study, and explore correlations between impact type and the plant growth of nitrogen fixing plants.

Bibliography

Chang, C. (n.d.). Causes of Plant Diversity. Retrieved from The Chang Lab: http://cynthiacchang.weebly.com/research.html

del Moral, R. (2001). The Vegetation Recovery of Mount St. Helens: A personal view of successional processes. Retrieved from University of Washington: http://faculty.washington.edu/moral/pdf/Folio%2001.pdf

del Moral, R. (n.d.). Mount St. Helens Memoir. Retrieved from University of Washington: http://faculty.washington.edu/moral/MountStHelensMemoirs.html

del Moral, R., & Rozzell, L. (2005). Long-term effects of Lupinus Lepidus on vegetation dynamics at Mount St. Helens. Retrieved from University of Washington: http://faculty.washington.edu/moral/pdf/2005LupinePatch.pdf

Tufte, E. R. (2013). The Visual Display of Quantitative Information (Second ed.). Cheshire, Connecticut: Graphics Press LLC.