I am not sure about how to interpret different values of the Simpson's index of diversity. For example: if I have two communities where
1-D (community 1) = 0.92
1-D (community 2) = 0.89
the first community is about 3% more diverse than the second one. Does this mean that the two communities are very similar or that they substantially differ?
I would suggest you to use a statistical approach to make your inference more robust. If you have replicates for each community, you can test for mean differences using a t-test, for example. If you only have a single sample for each community, but the sample is large enough to be representative of the universe of each community, you can use bootstrap techniques and generate statistical distributions of Simpson values for each community. Then, you may want to contrast both distributions to infer about significant statistical differences.
I hope that helps.
I advise you to do the analysis in the past 3.18. It allows you to establish the confidence intervals for different bootstrap diversity indexes in addition to calculating the statistical differences for the different diversity indices between two samples (Diversity permutation test) and (Diversity t test). This could be an inconvenience if you have to make multiple comparisons. Attached the manual of Past. 3.18 for that analysis. It is important to have several transects per location and samples must work to make your results representative.
Diversity permutation test
Expects two columns of abundance data with taxa down the rows. This module computes a number of diversity indices for two samples, and then compares the diversities using random permutations 9999 random matrices with two columns (samples) are generated, each with the same row and column totals as in the original data matrix.
Diversity t test
Comparison of the Shannon and Simpson diversities in two samples.
I agree with Dr. Enrique.... You can use Past software which is freely available online.
Emanuele,
Seems to me that you are trying to answer not one but rather two distinct questions: 1) which community is more diverse? (already addressed above), and 2) how similar are the two communities/assemblages?
Something to keep in mind is that even if the diversity (Simpson's or other) between the two communities is similar, or even the same, you will not be able to tell how different they are from each other in terms of their species (generic, etc.) composition.
You can address your second question by using ANalysis Of SIMilarity (ANOSIM), (there are other approaches as well). This non-parametric approach tests whether you can reject the null hypothesis that the similarity between communities is greater than or equal to the similarity within the communities of interest. The test is available in several statistical software packages including the Vegan package in R, PRIMER, and the aforementioned PAST.
Hope this helps. Shout if you have additional questions.
Kal
Step 1. Please prepare the sample plot wise value
Step 2. Test the normality using Shapiro–Wilk, Kolmogorov–Smirnov
then according to characteristics of the data
a. If the data are normal and we compare index values between 2 sites apply t-test
& for more than two sites we can apply the One-Way ANOVA
but
ii. If the data are not normal and we compare index values between 2 sites apply Mann Whitney U test
& for more than two sites we can apply the Kruskal Wallis test
Simpson's (and Shannon's) have an important drawback as the values are not directly comparable. Firstly, because they are unitless, it makes the comparison challenging. Secondly, and more importantly, if the true diversity of one community is twice as that in another community, the values returned by either index don't. Thus, the 3% difference in the values you show not necessarily means that true diversity of one community is 3% greater than the other. A better approach is to use Hill numbers, which are related to these popular diversity measures. For more information you can read Host and/or Chao for details. Hope this helps!
In the article below, we used ANOVA to test for differences in Shannon Diversity, Morisita's Similarity, and Jaccard's Similarity Indices, using an arc-sin transformation on index numbers to meet ANOVA assumptions. We initially also looked at Simpson's Diversity but the results were so similar to Shannon's we omitted them from our final analyses. But testing for differences in Simpson's should be similar.
Article Effects of 3 forest management systems on herpetofaunal dive...
Article My Village Biodiversity: Documentation of Western Ghats Biod...
Article Geospatial analysis of forest fragmentation in Uttara Kannad...
- Badges
- Science method
- Similar topics
- Biological Science
- Ecology