Sandra,

You raised a very relevant issue.

Biodiversity assessments are being implemented around the globe mostly using surrogate taxa to extrapolate to all animal and plant kingdoms. The decision about which group to use as surrogate is very controversial as is deals with the apropriateness of each group to be used as surrogate. Which other groups does it represent? Is it easy and economically viable to sample? Does it have threatened species of concern that may be affected by the sampling? and so on...

Another approach is the use of vegetation indexes or environmental characteristics as a proxy to biodiversity. 

Some sources for discussion are:

http://www.madingleymodel.org/index.html

http://biodiversitymapping.org/index.htm

Ponce-Hernandez et al. 2004. Assessing carbon stocks and modelling win-win scenarios of carbon sequestration through land-use change. FAO.

http://www.fao.org/docrep/007/y5490e/y5490e00.htm#Contents

GLOBIO. Modelling human impacts on biodiversity.

http://www.globio.info

Any development in this area will be valuable for many fields of environmental activity, including biodiversity research, environmental licencing, reserve selection and environmental policy implementation.

Best,

Onildo

This is going to depend on what aspect of biodiversity you want to measure and at what scale.  What scale are you interested?

Hi

I agree with Mr. Chris. Obviously, it is depend on the scale which you desire to study.

Thanks for your answers, of course all indicators are scale dependent, but the issue is to provide insights into the relevant scale for implementation. If we want to target planning purposes then the questions raised apply from local to global. In particular national level monitoring schemes like the one in the UK are looking for answers of these questions; perhaps concrete examples in other countries can help to move towards implementation.

Apart the scale, what do you mean for "measure" biodiversity? Biodiversity by itself cannot be measured! You could measure, for instance, species richness and use it as a proxy for biodiversity, but you would have no information on both genetic and community/ecosystem diversity.

Hi Sandra,

Have you checked the IPB which is used in forestry? or some complexity indexes that were developed long ago by Cancela?

Best,

Guy

Sandra,

You raised a very relevant issue.

Biodiversity assessments are being implemented around the globe mostly using surrogate taxa to extrapolate to all animal and plant kingdoms. The decision about which group to use as surrogate is very controversial as is deals with the apropriateness of each group to be used as surrogate. Which other groups does it represent? Is it easy and economically viable to sample? Does it have threatened species of concern that may be affected by the sampling? and so on...

Another approach is the use of vegetation indexes or environmental characteristics as a proxy to biodiversity. 

Some sources for discussion are:

http://www.madingleymodel.org/index.html

http://biodiversitymapping.org/index.htm

Ponce-Hernandez et al. 2004. Assessing carbon stocks and modelling win-win scenarios of carbon sequestration through land-use change. FAO.

http://www.fao.org/docrep/007/y5490e/y5490e00.htm#Contents

GLOBIO. Modelling human impacts on biodiversity.

http://www.globio.info

Any development in this area will be valuable for many fields of environmental activity, including biodiversity research, environmental licencing, reserve selection and environmental policy implementation.

Best,

Onildo

Dear Onildo, thank you for your answer that is very relevant in adition to the links you provided, I would like to add another link from the UK biodiversity indicators program http://jncc.defra.gov.uk/page-1824  The UK has information about its biodiversity, collected across a broad spread of species and habitats both by professionals and by expert amateurs. Then the other issue is to be able to rely on expert amateur’s data as many others countries do…but then we encountered the issues I raised in terms of effectiveness, implementation, accuracy…..

Not sure if this answers your question, but you may be interested in looking at the Biodiversity Index developed by my team at the University of Northampton.  It's intended for use by non-specialists in an urban setting:

http://www.biodiversityindex.org/

Please message me if you want more details.  

Regards,

Jeff

PS - I've uploaded a short article about the Index:

https://www.researchgate.net/publication/259574796_Ollerton_J._%282013%29_The_Biodiversity_Index__a_tool_for_facilities_management._Essential_FM_Report

Article Ollerton, J. (2013) The Biodiversity Index – a tool for faci...

you may want to check the following paper on EBVs( essential biodiverisity variables)

Bruford, M. W., Brummitt, N., Butchart, S. H. M., Cardoso, A. C., Coops, N. C., & Dulloo, E. (2013). Essential Biodiversity Variables, 339(January), 277–278.

The most straightforward approach would be to identify your measure(s) of biodiversity, such as species richness or a measure including also frequency values, or a measure of morphological phylogenetic diversity (or all of them) and establish a model between parameters (presence, abundance, spatial distribution) of the indicator species and these measures. The accuracy and performance of the indicator can then be tested by looking at the model fit. This should be repeated over time to see whether the predictive value of the indicator is consistent.

Thank you Robert, the difficult is usually to test and find the best model fitness, do you have any paper you can suggest for this discussion. Many thanks in advance

You can either do a classic model fit using R2 (for example with least squares approximation). This would be either non-linear or linear, depending on how the data behave and how you can transform them. Or you can use the AIC criterion. For the first, any good statistical software would serve, or else you can use R, and for the second there are also R packages available. The most important aspect of this is not necessarily the approach towards the model fit, but the quality and quantity of the data.

This may be difficult to generalize but for specific system, I would agree with Robert Lucking approach. One modification would be to model biodiversity measure (e.g., phylogenetic diversity) as function of the abundance of each species in your community with the goal to identify key indicator (of higher diversity) species. Alternatively you could regress this with a set of environmental variables to identify key indicator variable(s). Instead of trying to identify the best fit model, one option would be to do model averaging. Here are a few links on model averaging:

Basically, I agree with Robert's approach; however, I do not understand why if you use measures of SPECIES RICHNESS or MORPHOLOGICAL PHYLOGENETIC DIVERSITY you should call them BIODIVERSITY... Biodiversity is a much wider concept that cannot be measured directly. So, my thought is that if you measure, for instance, species richness, then you have just to call it SPECIES RICHNESS. After that, you can make considerations on biodiversity, taking into account that other of its levels were not considered (genetic and ecosystem levels, or functional levels).

”difficult is usually to test and find the best model fitness”

not necessarely...

how many taxonomic groups do You intend to include into the test?

Thank you for feeding this exchange that I found quite inspiring. Still I think that species richness is not the best indicator for biodiversity (many papers exist already on the literature). Biodiversity indicators need to be able to capture also the quality of the site and when possible consider dynamic changes as well. Another important consideration is that useful indicators target to help conservation planning needs to be spatial in nature. One way to capture environmental conditions in tandem with species information is to develop ecological-niche based spatial models. Thus this modelling approach generally define a function that links the fitness of individuals to their environment (Hirzel and Le Lay, 2008; Rotenberry et al., 2006). Theoretically, if we know precisely the habitat characteristics suitable for a species, it is possible to rebuild its ecological niche from the environmental variables describing its habitat (see BIOMOD, etc). I do believe this is the best way to derive spatial models as a surrogate for biodiversity assessments. In all, an integrated approach is needed if we aim to target operational planning options to improve conservation. Any other opinion is welcome!

We tried to assess this problem developing a biodiversity relevant environmental indicator set at the landscape level especially suitable as surrogate for land use related anthropogenic influence on biodiversity: see Rüdisser et al. Ecological Indicators 15 (2012) 208–216