I am using CPCe with random point counts in a 1m x 1m quadrate to determine the abundance of sponge, algae, corals, and other species on the reef. I noticed that the Shannon-Weaver index was calculated upon processing the .cpc files, but some of my values have either really low negative numbers (-500) or really high numbers (9000). I know that this index typically should be between 0 - 1, so why am I receiving such odd numbers? Is this a problem with how I am processing the initial data after identifying all of my points? Or is the algorithm behind this calculation still buggy? If anyone has had any luck with this, please let me know how you fixed this situation.
In many cases I have got the value more than 1.000 -- which I developed through C language
Abhijit Mitra Thank you very much for your input. What exactly does a value of more than 1 mean? Also, what does a value of less than 0 mean? Again, I just had no idea that you could get either more than one or less than 0 for this index.
The value in excess of 1.000 is common and easy to get if you know a bit of C language
Tridip Boruah and Abhijit Mitra , have you also experienced receiving a low, negative value for this index?
Yes, I did receive negative data once but could not really find a way to explain the results.
I think the CPCe's Shannon-Weaver Index is processing incorrectly, you should check your formulations or calculations, perhaps there are some incorrect process
The Shannon-Wiener Index (or just "Shannon" but not "Shannon-Weaver" -- there's a complicated story behind the authorship but it was Norbert Wiener, not Warren Weaver who contributed to Claude Shannon's advance) cannot take negative values but it very commonly takes positive values greater than one. In theory, it could have very large positive values but, in practice, values as high as 5 are unusual.
[The exponent of the Shannon index is a count of species in a community equivalent to the one sampled, except for having perfect evenness (i.e. all species equally abundant), which means that the value of "exponential Shannon" always falls far short of the number of species in the real (always highly uneven) community actually sampled. e^5 is nearly 150 and few studies of diversity include hundreds of different species, so Shannon values of as much as 5, although theoretically possible, are rarely seen.]
If the software you are using yields negative outputs or large positive ones, there is something seriously wrong.
The Shannon Index is not much better than meaningless but, if you want to use it, the now-conventional approximation (using ni/N as an alternative to the unknowable pi) is easy to calculate. My recommendation would be to put your data into an Excel spreadsheet and crunch the numbers within that. Stick close to the data and to the approximation Shannon's equation, without the risk of someone else leaving a bug in their software. (Though I'd also recommend that you consider calculating Brillouin's H and approaching the Shannon value through Zahl's jackknife adaptation of Pielou's "pooled quadrat" approach. And, if you do not understand why that extra complexity may be needed, I'd recommend that you avoid invoking any index before understanding what it represents!)
There are n number of programming to calculate the index. However, some gives negative results and some can't read the decimals. Those must be avoided
- Badges
- Science method
- Similar topics
- Biological Science
- Ecology