What is the purpose of this project?

Historically, ecological community assessment focusses on plants which is insufficient because high condition plant communities can be missing key fauna, and conversely, degraded plant communities can be important habitat for fauna. The bird community metrics developed in this project will allow for bird community condition to be explicitly measured, helping to provide the missing elements in biodiversity assessments, improve cost-effectiveness and comparability, and enable an holistic view of biodiversity that includes both plant and animal communities. The metrics developed in this project will allow land managers and conservation agencies to evaluate bird community condition and how this responds to pressures, essential for recovering Australia’s biodiversity.

What does community condition mean?

‘Condition’ of an ecological community is not an objectively measurable attribute, but represents a value judgement, usually involving comparison of an instance of a community with what that community would be expected to look like under ‘reference’ conditions. In the Australian context, reference condition is usually derived from ‘best on offer’ sites, representing expert-judged examples of the highest condition. In this project, we are using expert judgements of community condition to identify and understand what ‘best on offer’ looks like for each community. Throughout the survey you will be asked to rate the condition of the bird community on a scale of 0-100, where 100 represents the best version of the community you are aware of.

Should I consider historical bird assemblages and community condition when I think about a community that rates 100 on condition?

No. For the purpose of this project, we are focussing community condition relative to a ‘best on offer’ benchmark, rather than a historic or theoretical pristine community.

Why isn’t there information about the vegetation or other site features?

This project aims to develop measures of bird community condition based purely on the bird community at a site, so we have intentionally avoided providing any site-level details about the vegetation. The bird records we are using for this project are also obtained from BirdLife Australia’s Birdata platform, which does not include systematic information about the vegetation at all bird survey sites.

Why isn’t there information about which season the survey was conducted in?

The relationship between season and the bird assemblage, including which species occur and how abundant they are, is really complicated. We would need to get you to assess the condition of a much larger number of bird lists from a range of seasons to be able to explicitly incorporate season into our condition models. Instead, we would like you to assume that all the surveys you see were conducted in the most suitable survey season. At first, our condition metrics will only be suitable for this season. Later in the project we will examine seasonal changes and hope to do some combination of: understanding which communities have metrics that are seasonally changeable, developing season-specific condition metrics, developing guidelines about what season bird surveys should be conducted in to be compatible with our condition metrics.

Note that it is possible that some of the bird lists you see will include species that are not typically present in the optimal survey season. In this circumstance we ask that you make a note of your concern and proceed as if that species had been anomalously detected in the optimal survey season.

What kind of expertise do I need to have to take part in this project?

We are primarily looking for people who have significant experience observing birds in the relevant bird communities. For example, if the survey is about the Inland Woodland and Mallee community, you might have done bird surveys in one region of the community over a couple of seasons, or you might have done bird surveys across a large span of the region over many seasons.

We are also interested in hearing from people who might have no, or very limited on-ground experience in these communities if they have engaged significantly with data from the relevant communities.

How does the survey work?

We will ask you a series of questions in an online survey. For each bird community you agree to evaluate, this survey will show you a series of species lists from 25 2-ha 20-minute bird surveys out of BirdLife Australia’s Birdata platform. For each survey, we will ask you to estimate the condition of the bird community that the survey represents on a scale of 0-100 based on your personal experience. When doing this we ask that you assume that all bird surveys were conducted in the last 10 years, by experienced birders, in non-urban environments, at an appropriate time of day and in appropriate weather conditions. At the end of the survey, you will be asked about the attributes that you considered when generating your condition scores. Each community type should take around half an hour to complete.

How do I make sure my responses are saved?

Every time you move through the survey using the backwards and forwards arrow buttons at the bottom of the survey your progress will be saved. NOTE, using the browser ‘back’ button will lose any unsaved data you have entered.

Can I see the other assessments I have made?

At the bottom of your survey there is a back arrow button. You can use this to navigate back to earlier sections of the survey including your previous assessments. Navigating with this back arrow will also save your progress. If you attempt to navigate with the browser button you will lose any unsaved progress. Unfortunately, the survey tool we are using (Qualtrics) doesn’t allow us to provide easier access to past assessments. 

Do I need to assess condition for all the communities?

Definitely not. Each survey only includes condition assessments for a single bird community. We plan to have between 20 and 30 surveys asking experts to assess the condition of different bird communities. We would love it if you could spare the time to assess as many communities as you are familiar with but there is no obligation on your behalf.

Why are you asking me to assess community condition based on a single survey?

We recognize that a single 2-ha 20-minute bird survey is a very incomplete sample of any community. Once we have developed a first-take metric of community condition based on experts’ condition estimates from single surveys, we will experiment with how best to combine information from multiple surveys to provide the most reliable estimates of community condition.

On that basis, we ask you to imagine that you have to make a preliminary assessment of community condition based on information from a single 2 ha 20 minute bird survey, and assuming that the survey you see is quite typical of surveys done at that site.

What if something other than community condition is driving the species found at that site (e.g. observer skills, weather conditions, time of day)

For the purposes of this exercise, we need you to assume that the survey was conducted in favourable conditions by expert bird watchers. We are going to use the information from your assessments to model the predictors of community condition. This model will ultimately be used to judge the condition of the bird community as surveyed by experts according to a sensible survey protocol. 

Why are you specifically asking for estimates for 2-ha 20-minute surveys and what does that mean for my condition estimates?

Over the course of the project, we hope to calibrate our metrics to a range of other survey methods but for now the focus is firmly on the 2-ha 20-minute approach. It is important that we use data from one survey methodology to develop the preliminary condition estimates – that way the survey effort and detectability of species is comparable between surveys in the same broad habitat type. The 2-ha 20-minute survey is one of the most commonly conducted bird survey methods and it occurs over a reasonably small area, reducing the chance of capturing multiple bird communities in a single survey.

Data from 2-ha 20-minute bird surveys only includes species occurring within the 2-ha area during the 20-minute survey time. That means that birds heard or sighted outside the survey perimeter or before or after the survey time period should not be included in the recorded data.

Not all of the sites/birds in this community seem right to me, how did you come up with the communities?

We filtered data from more than 1.5 million individual bird surveys shared by BirdLife Australia’s BirdData portal. The filtering approach allowed us to focus on analysing data for terrestrial birds surveyed using a uniform 2-ha 20-minute approach. A total of 75,000 bird lists were acquired, which we then further sampled to focus on balanced bioregion and Major Vegetation Group, resulting in a set of 57,488 survey points.

We used hierarchical clustering and expert consultation to identify potential clusters. We presented these clusters to experts and asked them to assess whether: the computationally derived communities represent bird communities that are recognisable in nature, whether any of the communities need to be subdivided or grouped together, and whether there were bird communities missing from the set.

The expert consultation and refinement resulted in some clusters being categorised as degraded communities and subsequently excluded from the bird community typology, three communities further refined by splitting them into two smaller communities based on the hierarchical clustering results. This resulted in a total of 32 communities which form our Australian Bird Community Typology (currently we are not eliciting condition estimates for all of these).

What should I do if a particular bird list looks unfamiliar or wrong to me?

If you are primarily familiar with a particular region of the community and there are surveys that include species from a different region in the community, feel free to rate the condition of the community as you would one with analogous species (e.g. switch an unfamiliar robin species for one you are familiar with).

If you are uncomfortable with that or the survey feels more unfamiliar or wrong, please write us a note and skip that bird list in the online survey.

How many communities are there?

We have identified 32 bird communities at this stage but recognise that we are missing some significant communities including but not limited to lowland tropical rainforests, coastal, marine, and aquatic birds. Over the course of the project we hope to extend our scope to include these missing communities.