We were able to establish baseline “norms” in the form of Shewhart charts for indicators that could be broadly categorised as indicators of change in:

(i) **population demography** because of altered recruitment, survival and mortality. Measures were:

- proportion females;
- proportion of first-year birds (having hatched in that calendar year or in the preceding calendar years);
- proportion deaths from starvation or disease.

(ii) **nutritional status** measures were:

- body weight;
- fat score;
- condition index.

(iii) **physiological stress** (fluctuating asymmetry)

Descriptions of how these indices are calculated can be found in the supporting PBMS report.

Where the health index reports as a proportion of the population (e.g. percentage of females) then for statistical reasons the index is presented as the Arcsine of the square root of the proportion. This value can theoretically range between 0 and 1.57, representing 0% and 100% of the population, respectively.

### Proportions

The proportion of female and juvenile birds in each year's sample. Years where the value falls outside of the 95% prediction intervals may indicate change in the demographic structure of the species population.

Health index value

Mean or fitted regression for 2000-2016

Mean or regression after 2016

95% prediction intervals for 2000-2016

95% prediction intervals after 2016

### Body Weight

Median post-mortem body weight of birds each year for different sexes and ages of birds. The sample size in any single year was at least 5 birds. Years where the value falls outside of the 95% prediction intervals may indicate change in the physical condition birds in that particular demographic which may indicate, in the case of years below the lower 95% prediction interval, that food availability may be restricted.

Health index value

Mean or fitted regression for 2000-2016

Mean or regression after 2016

95% prediction intervals for 2000-2016

95% prediction intervals after 2016

### Putative cause of death

The proportion of birds that died of starvation or disease in each year. Years that fall below the lower 95% prediction interval indicates that there are fewer starved birds in the population (or demographic group), conversely an increase above the upper 95% prediction interval indicates there are more starved birds.

Health index value

Mean or fitted regression for 2000-2016

Mean or regression after 2016

95% prediction intervals for 2000-2016

95% prediction intervals after 2016

### Fat score

The proportion of birds categorised as starved each year, based on their fat score.Years that fall below the lower 95% prediction interval indicates that there are fewer starved birds in the population (or demographic group), conversely an increase above the upper 95% prediction interval indicates there are more starved birds.

Health index value

Mean or fitted regression for 2000-2016

Mean or regression after 2016

95% prediction intervals for 2000-2016

95% predition intervals after 2016

### Condition

In the graphs below, years for which the average condition index (CI) is below the lower 95% prediction interval indicates that there are fewer birds with low fat reserves in the population (or demographic group), conversely an increase above the upper 95% prediction interval indicates there are more birds with low fat reserves. Low fat reserves means a bird is less likely to survive prolonged periods where they are unable to hunt successfully, or that the bird has starved. The condition index is used as a continuous measure of body condition, rather than categorical such as fat score. The mean condition of birds per year was calculated for different ages and sexes. The Condition Index is calculated as follows:

CI = (BW - GW) / SD

where:

- CI = Condition Index
- BW = whole post-mortem body weight (g)
- GW = Wet weight of gizzard contents (g)
- SD = Sternum diagonal, which is the distance between the posterior point of the sternum keel plate and the distal point of the clavical (mm)

Health index value

Mean or fitted regression for 2000-2016

Mean or regression after 2016

95% prediction intervals for 2000-2016

95% prediction intervals after 2016

### Fluctuating asymmetry

Higher levels of fluctuating asymmetry in feathers may indicated that birds in the population were subject to physiological stress during the period that those feathers were grown. The mean fluctuating asymmetry index per year in the weight of the 10th Primary feather is calculated for different ages and sexes of birds. The mean fluctuating asymmetry index is calculated as follows:

FA10w = 0.798 * SQRT( | FWl - FWr | - ME )

where:

- FA10w = fluctuating asymetry index of 10th primary feather weight
- FWl = weight of the left 10th primary feather (g)
- FWr = weight of the right 10th primary feather (g)
- ME = measurement error estimated from 2 repeat measures of the trait

Health index value

Mean or fitted regression for 2000-2016

Mean or regression after 2016

95% prediction intervals for 2000-2016

95% prediction intervals after 2016