Species-Area curves

Species-area curves are useful for answering the question "How many plots do we have to survey to include most of the species in an ecosystem?"

The following figures each show three curves.  In red is the Species-area curve for both transects, based on the average of five runs through all 101 randomly-ordered plots.  In blue is the Species-area curve for the Uphill transect, calculated starting at plot 1 and going straight through to plot 51.  In green is the Species-area curve for the Contour transect, calculated starting at plot 52 and running straight across to plot 101.

1993 Species-Area Curves
The curves show that in the Contour transect, almost all (73%) species present have been accounted for after just 20 plots.  On the other hand, in both the Uphill transect, and for both transects overall, new species are sill being added after 45 plots (yields 30 species).  To account for most of the species in the system as a whole, it appears that about 60 plots are needed (though a few final species could still be added by surveying all 101 plots).

1998 Species-Area Curves
The patterns are a bit different from 1993.  This can be partially explained by the fact that we found fewer species in 1998 than the class found in 1993.  About 35 (rather than 20 in 1993) plots are needed to find most of the species present in the Contour transect.  The Uphill transect has far fewer species in 1998 than in 1993: and new species are being added right up to the end.  Overall, 60 plots are needed to get to 30 species, compared with 45 plots in 1993.   Species are added very slowly the last 50 plots.


Could we have saved ourselves some time?

We wondered whether we could have surveyed smaller plots (3 m x 30 m) and got just as much information as from our larger (6 m x 30 m) plots.  Since we kept track of the two halves (3 m x 30 m) of each 6 m x 30 m plot separately (designated "A" and "B" sides), we were able to try to answer this question.

The following Species-area curve shows that although more small plots are needed to get a given number of species, we don't necessarily need to survey twice as many plots.  It depends on how many species we want to get.  The saving of surveying more, smaller plots is significant only when a comparatively few (15) species are desired.  To get 25 species, we need to survey 27 big plots, or as many as 51 small plots (close to twice as many).  To get 30 species, we need to survey 62 big plots, or as many as 101 small plots.   The savings are less significant the more species we want to get.
 
All plots
6 m x 30 m
"A" Plots
3 m x 30 m
"B" Plots
3 m x 30 m
To get 15 species...
7 plots
10 plots
9 plots
To get 20 species...
12 plots
24 plots
19 plots
To get 25 species...
27 plots
51 plots
30 plots
To get 30 species...
62 plots
101 plots
95 plots
 


Diversity Measures
 
One reason to look at species diversity is to see whether the few species that are most abundant dominate the system, or whether there are many species that are all more or less equally abundant, with none dominant.  One way of conducting such analysis is to determine the abundance of each species (for this analysis we used species density, in terms of stems / ha), and then rank this list in order of decreasing abundance.
 
As the following ranked abundance graphs indicate (only the 10 most abundant species are shown, since species 11-35 are of such low abundance), similar patterns were found in 1993 and 1998.  In both years, all transects were pretty much dominated (more than half the total abundance accounted for) by two or fewer species.  The Contour transect has the least diversity, since most total abundance is accounted for by the fewest number of species: sugar maple is completely dominant.  In the Uphill transect, and for the ecosystem as a whole, both sugar maple and hemlock are the dominant species.  Neither transect can be considered very diverse.

1993 Ranked Abundance

1998 Ranked Abundance

Another way of presenting the same information is to plot the cumulative abundance against the number of species.  In all cases (all plots, uphill plots, and contour plots), 2 species account for 50% of the cumulative abundance, and 10 species account for about 80% of the cumulative abundance.  The remaining 25 species are responsible for very little of the ecosystem's total abundance.

1993 Cumulative Abundance

1998 Cumulative Abundance

 

Diversity Indices

There are a number of ways to measure species diversity in an ecosystem.   A simple count of the number of species found is one way: this gives equal weight to all species present, regardless of their abundance.  More complicated formulas have been derived to yield diversity indices, and one common formula is (where pi is the proportion of total abundance accounted for by species i, and Ib is the index of diversity):

Ib = (1 - S pib+1 ) / b, -1 =< b

When B = -1, this formula gives the species count (s - 1); when B = 0, the formula gives Shannon's index of diversity; when B = 1, the formula gives Simpson's index.  Simpson's index tends to give more weight to species that are more abundant, and less weight to those that are less abundant. Larger values of Ib indicate more diversity.

As the following table shows, in both cases (1993 and 1998), and regardless of the measure of divesity used, there is more diversity on the uphill transect than on the contour transect.  We would expect this, since there is a greater variety of habitats on the uphill transect.  The diversity of the uphill transect is comparable to that of the two transects combined.

 
Species Count
Shannon's Index
Simpson's Index
1993
All Plots
33
2.38
0.61
Uphill (Xeric)
32
2.47
0.63
Contour (Mesic)
24
2.03
0.56
1998
All Plots
33
2.30
0.88
Uphill (Xeric)
29
2.34
0.84
Contour (Mesic)
27
2.02
0.73
 
Whether or not these indices actually tell us anything about what we have in the field, or whether areas of similar diversity are in fact in any way comparable, is another question entirely.