Supplementary Online Material: Additional Details on the Application of IUCN Criteria

Supplementary online material: additional details on the application of IUCN criteria to GB birds

The rationale behind projecting trends forward into the future (Criteria A3 and A4)

The IUCN Red List Guidelines (IUCN 2016) recommend that species should, whenever possible, be assessed against all criteria, including exploring future trends under criteria A3 and A4.

• Criterion A3 deals with population reductions projected or suspected to be met in the future ten years or three generations (whichever is longer, but up to a maximum of 100 years) (IUCN 2016), although no species or subspecies were assessed as qualifying under solely A3 in the GB assessments.
• Criterion A4 deals with reductions observed, estimated, inferred, projected or suspected over any ten year or three generation time period (up to a maximum of 100 years into the future), where the time period must include both the past and the future, and where the reduction or its causes may not have ceased OR may not be understood OR may not be reversible (IUCN 2016).

The Guidelines present different examples of how trends can be projected forward into the future, but leave it to assessors to decide what method to select. For the GB assessments, we used a constant proportional (exponential) rate of decline, based on a known rate between two points (three generations or less apart) in the past. This annual rate was then in turn projected forward by the required number of years.

Annual indices were available for many GB species; these showed that for some species, simply projecting the trend from the previous three generations forward might not be appropriate, as the trajectory of the more recent element of the trend differed from that in the past. Some of these deviations could be considered to be fluctuations, while others appeared to be obvious changes in a species’ fortune that should be reflected in our estimation of what the future trend may be. We adapted the approach we used to deal with these different scenarios.

Method 1: Species that show a relatively consistent decline

For category A2, which explores observed, estimated, inferred or suspected declines, we took two data points three generations apart to calculate an annual rate of decline. This was then projected forward up to three generations (fig. 1). Generally if a taxon qualified under criterion A2, it was also listed under criteria A3 and A4 as the same rate of change was used, over the same length of time.

Fig. 1. Grey Partridge Perdix perdix, a species showing a relatively consistent trend. The red line shows the exponential rate of decline between two data points three generations apart, which was then projected forward to allow A3 and A4 trends (the same as the A2 trend) to be derived.

Method 2: Species that show what could be considered to be natural fluctuations

The IUCN Red List Guidelines specify that the downward phase of a fluctuation should not normally count as a reduction when assessing against criterion A. While carrying out the assessments, we noted that fluctuations also occurred while species were declining. In order to produce robust assessments, which were not necessarily the worst case scenarios resulting from minor or major fluctuations, we used the underlying three generation trend (calculated the same as Method 1) for these species. This was then projected forward up to three generations and the species qualified under criteria A2, 3 & 4 (fig. 2).

Fig. 2. Non-breeding Velvet Scoter Melanitta fusca is shown as an example of a species with fluctuations over the last three generations. Although there was a rapid fall in numbers in recent years, we did not feel it sensible to project that forward (not least because the decline appears to have ceased), so the underlying three generation trend was used.

Method 3: Species that show a substantial deviation away from the underlying three generation trend, but for which we do not expect the trend to continue at the same deviated rate in the future. In such cases, we still considered it unreasonable to default to the underlying three generation trend

The IUCN Red List Guidelines recommend using a ‘moving window’ method to assess species against criterion A4. To do this, a time series of past and future trends should be created. Again we estimated future trends by projecting forward the trend over the previous three generations (see Method 1). The moving window method then calculates three generation trends for all time frames that include at least one past year and at least one future year (fig. 3). The maximum of these reductions was used to determine the assessment under criterion A4 (see IUCN (2016) for additional information).

Fig. 3. Common Kestrel Falco tinnunculus as an example of a species showing a substantial deviation away from the underlying three generation trend; however, we don’t expect the recent decline to continue at the same rate. The method compares the difference between the black and red lines always over a three generation time window. The blue line represents the steepest decline.

4: Species that show inflection points at some stage over the last three generations, and expect the recent trend to continue at the same rate

Some species have shown clear changes in their trends within the last three generations, and for some of these it seemed reasonable to expect the more recent trend to continue into the future. To project this forward by the required number of years we used the Lowess smoothing procedure on the species index (log scale) in the statistical package R 3.1.0 (R Development Core Team 2015). The decline over three generations, which included include both the past and the future, was then used to determine the assessment under criterion A4 (fig. 4).

Fig. 4. BBS trend for breeding Greenfinch Chloris chloris showing a clear change in fortunes from 2005 onwards. We used a Lowess smoothing to project the recent trend forward to 2018 to cover three generations from 2005 to estimate the decline over this period.

Information on which method was used to project trend forward for the various assessments are shown in the supplementary online results spreadsheet.

Supplementary species notes

Evaluation of Black Grouse trend in abundance 2005–16, and over three generations 1997–2016

Nicholas Aebischer, GWCT

Aim: Use the trends available from a range of annual Black Grouse Tetrao tetrix surveys across Britain to produce a 2016 estimate of male Black Grouse numbers that is comparable with the last national Black Grouse survey in 2005 (Sim et al. 2008). Combine this with the change from the first national survey in 1995/96 to evaluate change over the three-generation period 1997–2016.

1. Scotland 2005–16

Since 2005, roughly two-thirds of Scottish Black Grouse have been counted mostly annually by regional groups, of which there can be several within each of the four broad regions used in the 2005 survey (Sim et al. 2008). I used these counts to calculate average annual change per local region from the slope of the regression of log (count) versus year. Using the average annual change, I projected backwards from the last count to 2005, apportioned the 2005 survey estimate for the corresponding broad region pro rata, then projected forward to 2016. The sum of these 2016 projections gave me a total Scottish estimate for 2016 of 2,871 males, compared with the 3,333 estimated for 2005.

2. England 2005–16

Two surveys of Black Grouse in England have been carried out since 2005, in 2006 and 2014 (Warren et al. 2015). I used the counts from the two English surveys to calculate average annual change from 2006 to 2014. I applied the average annual change to the 2005 survey estimate for England (Sim et al. 2008) to obtain an English estimate for 2016. This gave 2,510 males in 2016, compared with the 1,521 estimated for 2005.

3. Wales 2005–16

Black Grouse are monitored annually in Wales, with counts of 195 (estimated from graph) in 2005 and 361 in 2016 (Anon. 2016). This compares with the 2005 estimate of 213 from Sim et al. (2008), so for comparability the count of 361 was increased pro-rata to 394.

4. Combined GB total 2005–16

Combining the above estimates from Scotland, England and Wales gives an estimated total for 2016 of 5,775 males, compared with the 2005 total of 5,067 estimated by Sim et al. (2008). This gives an overall increase of 14.0% between 2005 and 2016, equivalent to an annual increase of 1.2% per annum.

5. Evaluation of GB trend over three generations (1997–2016)

Sim et al. (2008) provided estimates for the number of male Black Grouse in Scotland, England and Wales in the 1995/96 national Black Grouse survey. Assuming that these estimates held for 1996, and with the matching estimates for 2005, I calculated the average annual change between 1996 and 2005. I used the annual change to estimate the numbers in 1997, totalled the numbers across countries, then calculated the overall change from 1997 to 2016 (a span of 19 years, i.e. three generation lengths). The value for the overall change over three generations, ending in 2016 is -9.2%.

Additional considerations

With regard to the magnitude of change in Scotland, it is worth noting that the Scottish regional group counts are not based on random sampling, but are carried out in self-selected areas. By their nature (and the tendency for areas that lose grouse to be dropped from coverage, or at least not to be covered regularly), these tend to be areas that hold (relatively) healthy populations. This might be expected to produce a bias towards a more positive trend. Conversely, coverage of areas known to be occupied will not pick up any range expansion, so may also be subject to a negative bias. On balance, we believe that the overall impact errs towards a slight positive bias, implying that the 2016 Scottish estimate may be slightly too high, and the GB 1997–2016 change slightly lower than calculated. In addition, the error attached to the calculated estimate of -9.2% is considerable because a power of 11 applied to the annual change greatly amplifies the error on the estimate of change. It seems reasonable to take the view that -9.2% is not distinguishable from -10% based on the data available.

6. Assessment of GB population size

The 2016 estimate of population size obtained above was specifically derived so as to be comparable with the 2005 national survey estimate, so the evaluation of the 1997–2016 change is sound (although as explained above, the Scottish estimate may be slightly too high owing to the nature of the regional counts). When focusing on population size instead of population change, however, the English estimate of 2,510 is unrealistic relative to the 2014 survey result of 1,437. The discrepancy arises because the 2005 national survey estimate of 1,521 was already unrealistically large relative to the English 2006 survey result of 1,029 (although the 2005 95% confidence interval of 798–2,364 was so wide that it included the 2006 result). I believe therefore that using a 2005 baseline of 1,521 for England is too high for predicting total numbers in England in 2016. Instead it is more sensible to use the 2014 England survey count of 1,437 males and project that forward to 2016 with an annual change of 4.3%, which yields 1,562 males. This gives a revised GB population estimate of 4,827 males in 2016. Assuming a 1:1 sex ratio, this equates to 9,654 mature individuals.

Evaluation of the current status of breeding Woodlarks in Great Britain

The Woodlark Lullula arborea was assessed as being Vulnerable in Great Britain, because the breeding population was estimated to be less than 10,000 mature individuals and evidence suggests the population has undergone a decline of at least 10% over three generations (Category C1). The last national survey in 2006 (Conway et al. 2009) put the population at 3,100 (2,500–3,700) pairs. This represented an 88% increase since 1997; however, there is evidence to suggest that the increase has stalled, with declines in several key areas since then. To further explore this, we collated local/regional data to see if it was possible to determine any underlying trend (see below). Although in the end it was not possible to determine a robust population trend over three generations, the authors agreed, on balance, that Woodlark had likely undergone a decline of at least 10% over the last three generations (a period of 11 years). Note: RBBP data is incomplete for this species, and Woodlark was dropped from the species covered by RBBP after 2012.

References

Anon. 2016. Black Grouse Annual Lek Monitoring, Wales 2016 Newsletter. RSPB Cymru, Natural Resources Wales, Denbighshire County Council & COFNOD.

Conway, G., Wotton, S., Henderson, I., Eaton, M., Drewitt, A., & Spencer, J. 2009. The status of breeding Woodlarks Lullula arborea in Britain in 2006. Bird Study 56(3): 310–325.

IUCN Standards and Petitions Subcommittee. 2016. Guidelines for Using the IUCN Red List Categories and Criteria. Version 12. Prepared by the Standards and Petitions Subcommittee.

Sim, I. M., Eaton, M. A., Setchfield, R. P., Warren, P. K., & Lindley, P. 2008. Abundance of male Black Grouse Tetrao tetrix in Britain in 2005, and change since 1995–96. Bird Study 55(3): 304–313.

Warren, P., Atterton, F., Baines, D., Viel, M., Deal, Z., Richardson, M., & Newborn, D. 2015. Numbers and distribution of Black Grouse Tetrao tetrix males in England: results from the fourth survey in 2014. Bird Study 62(2): 202–207.