Status and trends of common and widespread bird species

This measure relates to indicator 1.5.1 – Species composition and diversity.

Background

The Department of Conservation’s (DOC’s) National Biodiversity Monitoring Programme has been designed to assess status and trend in common and widespread terrestrial species across all public conservation land (PCL) in New Zealand. These species are of particular importance because they have considerable influence on the structure, biomass and function of most ecosystems. Declines in these species may be difficult to detect over relatively short timeframes, particularly in the case of bird populations. However, even relatively small declines can reflect significant reductions in the number of individuals, which can have subsequent effects on ecosystem structure, function and services. Such declines have been well documented for some of New Zealand’s rarer forest-dwelling birds, but there is also increasing evidence that significant declines have occurred for what were previously considered common and widespread species.

Grey warbler, tomtit, silvereye and bellbird are the most commonly observed birds on PCL.

What did we measure?

DOC has developed a national monitoring programme to assess status and trend of biodiversity across of all of the land it manages. The programme collects data on indicators and measures of ecological integrity outlined in the Department’s Biodiversity Outcome Monitoring Framework (PDF, 1.07 MB).

A terrestrial monitoring programme has been established at approximately 1400 plot locations spaced evenly across PCL. Each year, 5-minute bird counts are obtained from between 230 and 270 plots with every plot being measured once over a 5-year rotation cycle. See (Table 4) for the sample sizes in different ecosystems during the eight seasons from 2012/13-2019/20.

5-minute bird count observations are obtained from five stations at each plot location. Occupancy (presence adjusted for detection probability) and relative abundance are estimated for the most common species, using a statistical model that accounts for under-detection. Species richness (number of species per plot) is estimated for indigenous and introduced species separately.

What did we find?

Figures

Figure 1: Estimated average indigenous species richness at each reserve or conservation area over the seasons 2012/13 to 2019/20.

Figure 2: Estimated average occupancy for indigenous species across public conservation land in woody and non-woody ecosystems over the seasons 2012/13 to 2019/20.

Figure 3: Estimated occupancy for introduced species across public conservation land in woody and non-woody ecosystems.

Figure 4: Estimated occupancy (5-minute bird counts, 5MBC) of the seven most common bird species across public conservation land. Values are shown for different ecosystem types (i.e. woody, non-woody) and conservation statuses (i.e. national parks, non-national parks).

Figure 5: Estimated relative abundance (5-minute bird counts, 5MBC) of the seven most common bird species across public conservation land. Values are shown for different ecosystem types (woody, non-woody) and conservation statuses (national parks, non-national parks).

Tables

Table 1: Species richness on public conservation land broken down by ecosystem and indigenous/introduced species and ecosystem type. The values are derived from the posterior predicted site richness estimates shown in Figure 1.
Species group Ecosystem Mean richness Standard error 95% range across sites
all all 8.83 0.09 (0.5, 15.4)
all non-woody 5.53 0.19 (0.37, 15.64)
all woody 10.28 0.07 (5.4, 15.29)
indigenous all 6.31 0.07 (0.28, 11.67)
indigenous non-woody 3.16 0.12 (0.23, 10.26)
indigenous woody 7.70 0.06 (3.39, 11.95)
introduced all 2.52 0.04 (0.13, 7.16)
introduced non-woody 2.37 0.10 (0.11, 8.14)
introduced woody 2.58 0.04 (0.22, 6.25)

 

Table 2: Estimated occupancy model coefficients (standard error) for common bird species on PCL.
Species Intercept Time (Years) Woody Time*woody National park
Bellbird -1.71 (0.2) 0.01 (0.04) 2.88 (0.24) 0.03 (0.05) 0.92 (0.16)
Blackbird -1.5 (0.21) -0.03 (0.04) 2.39 (0.24) 0.02 (0.05) -0.7 (0.15)
Chaffinch -1 (0.19) 0.04 (0.04) 2.34 (0.23) 0.01 (0.05) -0.46 (0.15)
Fantail -3.06 (0.25) 0.06 (0.05) 3.07 (0.27) 0.06 (0.06) -0.44 (0.15)
Grey Warbler -1.65 (0.21) 0.01 (0.04) 4.42 (0.28) 0.01 (0.06) 0.02 (0.19)
Silvereye -0.9 (0.2) 0 (0.04) 2.8 (0.24) -0.06 (0.05) -0.72 (0.15)
Tomtit -2.42 (0.22) 0.01 (0.04) 4.04 (0.26) 0.03 (0.05) 1.23 (0.18)

 

Table 3: Estimated relative abundance model coefficients (standard error) for common bird species on PCL.
Species Intercept Time (Years) Woody Time*woody National park
Bellbird -0.03 (0.09) -0.03 (0.02) 0.33 (0.1) 0.02 (0.02) 0.06 (0.03)
Blackbird -0.51 (0.1) -0.04 (0.03) 0.13 (0.11) 0.02 (0.03) -0.27 (0.05)
Chaffinch 0.24 (0.09) 0 (0.02) -0.01 (0.1) 0.02 (0.02) -0.28 (0.04)
Fantail -1.05 (0.16) 0.03 (0.03) 0.23 (0.17) 0.02 (0.03) -0.23 (0.06)
Grey Warbler -0.48 (0.1) -0.03 (0.02) 0.82 (0.1) 0.04 (0.02) -0.25 (0.03)
Silvereye 0.32 (0.09) -0.01 (0.02) 0.25 (0.1) -0.02 (0.02) -0.19 (0.04)
Tomtit -0.44 (0.1) 0 (0.02) 0.58 (0.11) 0.01 (0.03) 0.09 (0.03)

 

Table 4: Number of plots measured annually for the different ecosystem class and land status combinations.
Ecosystem Conservation land status 2012 2013 2014 2015 2016 2017 2018 2019 Total
non-woody national park 21 21 19 22 17 25 22 19 147
non-woody non-national park 42 56 38 56 58 59 59 40 368
woody national park 8 51 54 73 60 64 52 53 362
woody non-national park 23 138 112 115 119 124 140 113 771
total total 94 266 223 266 254 272 273 225 1648

Map

Figure 6: Bird species richness and estimated indigenous dominance. Indigenous dominance is the proportion of species present that are indigenous.

Data quality

This measure is classified as a partial measure and complies with the data quality guidelines used in the Environmental Reporting framework.

Glossary of terms

5-minute bird counts (5MBC) are observer-based counts conducted at a fixed location. These counts require that all birds seen or from the point irrespective of distance are recorded over a 5-minute period. The 5MBC method is an index of relative abundance. Rather than measuring the actual density or abundance of birds directly, it records some of the individuals present and uses this measure to infer relative abundance without attempting to adjust for detection probability.

Credible intervals (CI) indicate that the true mean is likely to lie within the interval with a 95% probability given the posterior probability distribution.

Indigenous species are those native species inhabiting a defined region as a result of natural processes with no human intervention.

Introduced species are species that occupy an area as a result of non-natural human processes, either deliberate or accidental.

Occupancy refers to the probability that a site is inhabited by a given species whilst also accounting for imperfect detection.

Richness is a count of the number of different species that occupy a given area.

Methodology

This factsheet focusses on bird species composition (richness), occupancy and relative abundance. Communities of diurnal (daytime) birds are surveyed at each plot location using a cluster of up to 5 different bird count stations (at least 150 m apart) between 1 hour after official sunrise and 1pm hours using the 5-minute bird count (5MBC) method. Only terrestrial birds were considered for analysis (n = 51). Aquatic birds such as gulls or ducks were removed. The occasional diurnal record of morepork (ruru) was also removed as there are alternative acoustic detection and analysis methods for nocturnal (night time) species.

The data were modelled using a Bayesian hierarchical multispecies occupancy model with richness equal to the sum of species occupying a site. Occupancy \((z_{i,j})\) was species \(i\) specific and depended on cover (woody, non-woody). Detection \(x_{i,j,k}\) at a listening station \(k\) was also species specific. A site \(s[j]\) level random effect was included for both detection and occupancy. Index j refers to a site-season (a site in a specific season, as some sites had data from two seasons, e.g. 2013/14 and 2018/19) while s[j] refers to the site corresponding to site-season j.
The model was specified as follows \[ \begin{array}{rcl} x_{i,j,k}|z_{i,j} & \sim & bernoulli(p_{i,j} \times z_{i,j}) \\ z_{i,j} & \sim & bernoulli(\psi_{i,j})\\ logit(p_{i,j}) & = & \alpha_{i} + u_s[j] \\ logit(\psi_{i,j}) & = & \beta_{0,i} + \beta_{1,i}woody_{j} + v_s[j] \\ \end{array} \] with priors \[ \begin{array}{rcl} \alpha & \sim & \mathcal{N}(\mu_{\alpha}, \sigma_{\alpha}) \\ \beta & \sim & \mathcal{N}(\mu_{\beta}, \sigma_{\beta}) \\ v_j & \sim & \mathcal{N}(0, \sigma_1) \\ u_j & \sim & \mathcal{N}(0, \sigma_2) \\ \end{array} \] and hyper priors \[ \begin{array}{rcl} \mu_k & \sim & \mathcal{N}(0, 10) \\ \sigma_k & \sim & \Gamma(0.01, 0.01). \\ \end{array} \]

For the most common species on PCL the model was augmented with a negative binomial term for the total count \(y\) per species/site/season combination with a species-dependent overdispersion parameter \(r\), conditional on occupancy and detection: \[ \begin{array}{rcl} x_{i,j,k}|z_{i,j} & \sim & bernoulli(p_{i,j} \times z_{i,j}) \\ y_{i,j,k}|z_{i,j} & \sim & negbin(\lambda_{i,j} \times x_{i,j},r_i) \\ z_{i,j} & \sim & bernoulli(\psi_{i,j})\\ logit(p_{i,j}) & = & \alpha_{0,i} + \alpha_{1,i}woody_{j} + u_s[j] \\ logit(\psi_{i,j}) & = & \beta_{0,i} + \beta_{1,i}woody_{j} + v_s[j] \\ \end{array} \]

Additional resources

Walker, Susan and Monks, Adrian and Innes, John: Status and change in native forest birds on New Zealand’s mainland, 1969-1979 to 1999-2004. Landcare Research, 2017.

Gaston, Kevin J: Common ecology. Bioscience, 2011.

Elliott, Graeme P and Wilson, Peter R and Taylor, Rowley H and Beggs, Jacqueline R: Declines in common, widespread native birds in a mature temperate forest. Biological Conservation, 2010.

Dorazio, RM and Royle, JA: Estimating size and composition of biological communities by modeling the occurrence of species. Journal of the American Statistical Association, 2005.