Hi Todd I agree, Not sure why you have -ve numbers coming into the calculation since the formula squares the proportions and when summed up will be less than 1. Hence sum(pi*pi) > 0 and also 1-sum(pi*pi) > 0.
Take a look at this website for useful information with examples from the UK: http://www.countrysideinfo.co.uk/simpsons.htm Also recently (2012) there is an interesting article regarding weighted Gini-Simpson Index. SeeGuiasu and Guiasu (2012). They show the new weighted Gini-Simpson index behaves well when the number of species is large. Reference: Guiasu, R.C., and Guiasu, S. (2012). “The Weighted Gini-Simpson Index: Revitalizing an Old Index of Biodiversity,” International Journal of Ecology, vol. 2012, Article ID 478728, 10 pages, 2012. doi:10.1155/2012/478728 (available online). Best, Ling Ling Huang Sacramento City College --- On Fri, 8/10/12, Nicolas PERU <[email protected]> wrote: From: Nicolas PERU <[email protected]> Subject: Re: [ECOLOG-L] Simpson's Index of Diversity_aquatic plants To: [email protected] Date: Friday, August 10, 2012, 4:27 AM Dear Todd, I don't really understand how you can obtain a negative Simpson's Index value. Simpson's index formula is : H = 1-sum(pi²) where pi is the relative abundance of the ith species. So, H is always greater or equal to 0. You can always use Simpson's index as far as you have a set of relative abundances (ie. sum to one). The kind of abundance measure has no importance for the index. Of course it change the type of diversity you want to estimate. You must know that Simspson's Index give a very high importance to abundant species and don't care about rare one. So if one species have a very high weight in comparaison to the other ones, your index will be low even if you have many species. Consequently, you must ask yourself if it make sense for you. Regards Nicolas Le Thu, 09 Aug 2012 01:38:51 +0200, Shuskey, Todd <[email protected]> a écrit: > I am studying how the biodiversity within the weedbeds of a freshwater lake > has changed and one parameter I would like to investigate is species > diversity using Simpson's Index. The data I have available is the wet > weights and dry weights for each species from quadrat sampling. I do not > have available the actual number of plants within each quadrat. For another > comparison, I am using the Proportional Index of Community Similarity. > Can either of these data sets be used in the Simpson's Index of Diversity? > My problem is that many of the weights are less than one, which gives me a > negative value when trying to calculate Simpson's Index. Is there a proper > transformation that I can do with the data? > > Thank you in advance, > Todd Shuskey >
