State of NZ Garden Birds 2020 | Te Āhua o ngā Manu o te Kāri i Aotearoa
What are our birds telling us?
Birds act as backyard barometers – telling us about the health of the environment we live in. They are signalling significant changes in our environment over the last 10 years. We should be listening.
Using cutting−edge techniques, Manaaki Whenua – Landcare Research has distilled a substantial information base – bird counts gathered by New Zealanders from over 39,000 garden surveys since 2010 – into simple but powerful metrics.
Key findings from the 2020 survey
Good news for four native species:
- Kererū counts have moderately increased (by 79%) over the past 10 years – a bigger increase than the 2019 results showed (43%).
- Fantail (pīwakawaka) counts also show a slightly larger increase over the long-term.
- Tūī (kōkō) counts still show a shallow increase over the long-term.
- Silvereye (tauhou) counts are no longer moderately decline over the long-term – they now show only a shallow decline over the past 10 years. And, for the first time, silvereye counts show a 5-year increase – shallowly increasing nationwide and moderately increasing in three regions (Manawatū-Whanganui, Marlborough, and Taranaki).
- Myna counts have been increasing in some regions over the past 5 and 10 years, but have been rapidly increasing in Wellington.
- Starling counts continue to show a moderate decline nationally over the long term.
- The numbers of song thrush, goldfinch and dunnock are no longer showing national declines over either the short or long term.
How we calculate NZ Garden Bird Survey results
The general idea behind the survey is to convert the individual bird counts people contribute into meaningful estimates of wider population changes over time. Counts from every garden surveyed are linked to their location within a neighbourhood, suburb, district and region to calculate how bird counts change over time at each of these spatial scales.
We expect a species whose population is increasing over time to show an increase in counts and vice versa. While this is simple in principle, because there are different numbers of gardens in different regions, and different people decide to participate in some years, it might appear as though bird count numbers have changed, when it is just the proportion or location of survey returns that have changed.
The type of garden surveyed (rural or urban), whether people feed birds, and the total number of gardens surveyed in a region also add complexity. Statisticians refer to this as ‘noisy data’ because there are so many variables to take into account.
To reduce this noise, we use cutting edge statistical techniques to account for these variables. Following bootstrap analysis and bias-correction of the modelled data, estimated trends in bird populations over the past 5- and 10-year periods are summarised nationally and regionally according to their direction (decline or increase) and size (rapid to shallow).
These long-term trends are called ‘signals’ and show a persistent increase or decrease in the abundance of a species. We categorise these signals in the following way:
When using samples to estimate wider populations, we need some way of measuring whether the sample actually reflects the wider population. We used 80% confidence intervals for the average percentage change in each species’ counts to evaluate confidence in our estimates. We generally have more confidence at the regional level than district or suburb.
The easiest way to improve this confidence at the district or suburb level is to increase the number of bird counts done. The more people who participate, the greater the strength of our evidence for what’s happening to garden birds at the local scale.