When Measuring Biodiversity, Do Individuals Matter?

Post provided by Samuel RP-J Ross

Close up of a black-capped babbler (Pellorneum capistratum), one of the species included in our study.

Close up of a black-capped babbler (Pellorneum capistratum), one of the species in our study.

Our newly-developed method simulates intraspecific trait variation when measuring biodiversity. This gives us an understanding of how individual variation affects ecosystem processes and functioning. We were able to show that accounting for within-species variation when measuring functional diversity can reveal details about ecological communities which would otherwise remain unseen. Namely, we found a negative impact of selective-logging on birds in Borneo when accounting for intraspecific variation which we could not detect when ignoring intraspecific variation.

Why Biodiversity Matters

Biodiversity is important for many reasons. One of the main reasons is its contribution to the range of goods and services provided by ecosystems (i.e. ecosystem services) that we can take advantage of, such as natural food resources or climatic regulation. It’s generally believed that biodiversity contributes to these services by increasing and maintaining ‘ecosystem functioning’ – often defined as the rate at which ecosystems are turning input energy (e.g. sunlight) into outputs (e.g. plant biomass).

Measuring Biodiversity

As an approximation of ecosystem functioning, ecologists are increasingly measuring the biodiversity of a given community, assemblage or ecosystem by looking at the range of functions performed by species in an ecosystem. This is a great alternative to simply counting the number of species present. So, we’re measuring the functional diversity of an ecosystem, not just the taxonomic (how many species) or phylogenetic (evolutionary relatedness) components of biodiversity.

Functional diversity is based on the principle that we can get some indication of what functional roles species perform in an ecosystem based on their behavioural, morphological or physiological traits. For example, the size and shape of a bird’s beak likely tells us a little bit about what the species eats and how it forages. So, we usually measure functional diversity by getting data on the functional traits of different species in a community and using various functional diversity indices. These are usually calculated by computing convex hulls or by clustering species into a tree based on differences between species in these functional traits. Once we have these data, we can estimate how productive an ecosystem is, based on the range of functions being performed by all species in the community.

This is a widely-accepted way of measuring biodiversity, but it has a few caveats. Ecosystems are hugely complex and so all mathematical models that try to describe the way these ecosystems work naturally must oversimplify them. But one simplification when measuring functional diversity is that functional diversity studies do not account for intraspecific variation – the variation that occurs between individuals of the same species. This within-species variation is often important to a range of ecological and evolutionary processes, and individual differences can have large impacts in contexts from classical Darwinian evolutionary theory to decision making and collective behaviour.

So, we might be missing something by not accounting for the variation within species when we measure the functional diversity of ecosystems, especially since we know that functional differences between species can have large impacts on ecosystem functioning, and that often intraspecific variation can be as large as the variation that occurs between species.

Including Intraspecific Variation in Functional Diversity

Primary forest habitat, Sabah, Malaysian Borneo

Primary forest habitat, Sabah, Malaysian Borneo

One of the main reasons most studies don’t include intraspecific variation when measuring functional diversity is that we usually don’t have enough information about the amount variation that occurs within species. To get around this, we decided to make use of a global trait database which has data on several important functional traits for most bird species on earth. These kinds of databases are growing in popularity where empirical data on functional traits for many species are missing or cannot be measured directly in the field.

We used data on the real bird community composition of two rainforest bird communities in Borneo, based on past work by Felicity Edwards, David Edwards, Keith Hamer and others. The data from one of our communities came from natural ‘primary’ rainforest which had been largely undisturbed by human activity, whereas the other came from rainforest that had been selectively-logged with the largest trees removed for timber extraction. Our data included species names and abundances for all the birds found in each community.

Selectively-logged forest, Sabah, Malaysian Borneo. Note the open canopy level, dense understorey vegetation, and the clearance for a road.

Selectively-logged forest, Sabah, Malaysian Borneo. Note the open canopy level, dense understorey vegetation, and the clearance for a road.

By using the data from the Wilman et al. (2014) global trait database for birds, and data on trophic position of these species from past work in this ecosystem, we collected functional trait data for all the species in our study.

We chose 14 functional traits, including traits relating to foraging and resource requirements, such as clutch size, body mass, trophic position and foraging strategy. After accounting for correlations between these functional traits, we then calculated functional diversity for both communities based on average values for each species (although we had to slightly change the way we format our input data for the five commonly-used indices we used to calculate functional diversity).

After this, we wanted to see how intraspecific variation altered the values of functional diversity as well as the relationship between our two bird communities for our five chosen functional diversity indices. We generated a range of ecological limits for each trait based on literature values and mathematical formulae, then randomly assigned functional trait values to individuals for all their functional traits based on a normal distribution of trait values within these limits.

In this way, we generated ‘virtual’ individuals which were randomly generated numbers along 14 trait axes (one for each functional trait), with species-specific value ranges. We could then use the standard methods to position these individuals in multidimensional space (here, in 14 dimensions) to calculate the diversity of each community. We then repeated this procedure 1000 times for each community to make sure we avoided bias based on the values of rare or extreme species.

When Do Individuals Matter in Functional Diversity?

Once we had our values of functional diversity for all 1000 virtual communities, we could look for an effect of including intraspecific variation on overall values of functional diversity for either community. We discovered that including intraspecific variation lowers the values of functional richness and divergence (two components of functional diversity) based on our indices, but increases functional evenness (the third component of functional diversity). This led us to wonder why our values of functional diversity actually decreased most of the time when we increased the diversity of functional traits in an ecosystem. Our conclusion was that it was probably due to the mathematical properties of these functional diversity indices.

View of the selectively-logged forest from canopy level.

View of the selectively-logged forest from canopy level.

Most of these values were calculated as a proportion of the total diversity across the regional species pool, that is, the total diversity of all individuals and species in both the primary and logged forest combined. An increase in the diversity of one community when including intraspecific variation can never proportionally match an increase in the diversity across both communities, because these communities include different species. On the other hand, functional evenness increased because at the community level, clustering is reduced when displacing individuals from their species-level mean trait values. This results in a more even community than when individuals cluster on identical trait values within species.

Next, we wanted to look at the difference in functional diversity between the primary and logged forest communities when accounting for within-species variation. Traditionally, when only calculating functional diversity based on mean trait values, we get only one value of functional diversity, and so we can easily compare the functional diversity of multiple communities. However, when we produce 1000 possible values of functional diversity for each community because of our bootstrapping methods to consider intraspecific variation, it became more difficult to compare these values.

So, we devised a statistical method that allowed us to calculate the proportion of times each community has higher functional diversity. We applied this to our data for each of our five functional diversity indices. It showed that only two indices were higher in a given direction a significant proportion of the time. These were our two closely-related functional divergence indices: functional dispersion and Rao’s quadratic entropy.

What did we find out? Well, past work by Felicity Edwards and others has shown that when looking at functional diversity based on average values of functional traits for all species, there’s little difference in the functional diversity of birds in the natural primary forest habitat and the selectively-logged forest habitat. However, when we looked at the functional diversity of these bird communities using simulated values of within-species variation, we saw that this is not the case.

We actually found that the logged forest community had lower functional divergence than the primary forest community when considering intraspecific variation. This means that perhaps selective-logging has a greater impact on biodiversity in this ecosystem than we thought based on functional diversity calculated using only on species-level average functional trait values. This could have implications for our understanding of functional diversity and the impacts of logging in Borneo and beyond. Where possible studies should consider the important role than individual variation plays in ecosystems.

To find out more, read our Methods in Ecology and Evolution article ‘Incorporating intraspecific trait variation into functional diversity: Impacts of selective logging on birds in Borneo’.

You may also find past work in this system interesting:

Edwards DP, Woodcock P, Newton RJ, Edwards FA, Andrews DJ, Docherty TD, Mitchell SL, Ota T, Benedick S, Bottrell SH, Hamer KC. (2013a) Trophic flexibility and the persistence of understory birds in intensively logged rainforestConservation biology 27, 1079-1086.

Edwards FA, Edwards DP, Hamer KC, Davies RG. (2013b) Impacts of logging and conversion of rainforest to oil palm on the functional diversity of birds in SundalandIbis 155, 313-326.

Advertisements

3 thoughts on “When Measuring Biodiversity, Do Individuals Matter?

  1. Terrific work. I’ve been thinking about functional guild/life history/trait diversity in the context of ecological integrity for quite some time. This approach looks like it might solve the persistent issue of potential circularity in assessments of integrity, i.e., “these species are important because they are sensitive to anthropogenic disturbance.” Better to simply document, matter-of-factly, which traits win and which traits lose along gradients of disturbance.

  2. Pingback: Functional diversity assessment – traits! | The Waterthrush Blog

  3. Pingback: Back on solid ground (for now) | The Infrequent Musings of an Early-Career Ecologist

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s