From Swimming with Seagrasses to Statistics


Somewhere down the road ecology evolved from this:

To something like this:



As a member of the Seagrass Ecosystem Research lab here at Florida International University I am, by default, a researcher in two long-term monitoring projects.  The first is the FCE-LTER seagrass monitoring project and the other is a seagrass monitoring project in the Florida Keys National Marine Sanctuary.  Both datasets have resulted in mounds and mounds of data throughout the years.  As a side project to my dissertation I arrogantly decided to conquer said data mounds and turn out some ground-breaking revelations about benthic community structure in South Florida.  There's just one problem:  I am buried in MOUNDS of data!  To analyze these data, it turned out, would be much more complicated than just getting the R code (R is a stats program that hates me, but it's free so I put up with the abuse!) correct so the program will tell me what I want to know.  Here is the first figure I generated from this dataset:
Figure 1:  R-plot, or as my friend Rob put it, "It looks like a unicorn threw up on your computer."  Thanks Rob!
Somehow, R did not tell me what I wanted to know, nor did it write up a nice paper for me to publish.  My purpose in generating the first plot was to see if there was any relationships in community structure (species densities) at different regions of the FKNMS (denoted by the 8 different symbols) over time (different colored symbols).  The answer was simple:  No.  While pretty, this figure shows me that there isn't much of a difference among years or across regions for this set of years (I have data for every year beginning in 1996).  Where are my ground-breaking results?!

This lack of a relationship is good for the seagrass and macroalgae.  However, it has forced me to use my brain in ways in which I was not accustomed.  It has spurred me in a different direction with this dataset, though....AND, it's exciting!  I'm still in the exploratory stages of the analysis but I'm finding some interesting results.  Herein lies the beauty of both long-term datasets and statistics.  I am, by no means, a stats whiz - I really struggle with the topic sometimes and throughout my master's degree I threatened the life of my computer, R, and all South Texas seagrasses at one point or another.  Though I haven't found anything ground-breaking yet, the sheer amount of data I have at my fingertips allows me to investigate new hypotheses.  The benefits are two-fold:  1) I will never be the person who can read a stats book and apply the analysis years down the road.  I need the practice - like riding a bike - and with this dataset I get that much needed practice.  2)  I get to train my brain to think outside the box and be creative...with MATH!  That I enjoy the mathematical brain massage has shocked me.

Comments

  1. I am running into a similar problem Jenn, so I know how you are feeling! I am trying to see if there has been any significant changes in diatom community structure over the years that could be related to changes in rainfall/water management. However, the answer is no for a lot of the sites! But I have come to realize that this makes sense because not much change has occurred in terms of hydrologic restoration. So, I too am going to have to come up with creative ways to analyze the data and find out the important story underlying the unicorn throw up mess!

    ReplyDelete
  2. This comment has been removed by the author.

    ReplyDelete

Post a Comment

Popular posts from this blog

Diatom of the month - February 2016: Cyclotella meneghiniana

Diatom of the month - January 2016: Gomphonema parvulum