iiinova — where your digital self keeps evolving

Stephen Vickers

@stephen8vickers.bsky.social

557 followers 143 following 26 posts

Ecological Statistician at BioSS. ‘S’ permit bird ringer & RIN member 2022-26.

Posts Replies Media

Stephen Vickers @stephen8vickers.bsky.social · 07/09/2026

Our GPS tracking work has found Iceland Whoopers Swans spend less than a third of their time in protected areas! Read all about it in our latest paper published fully open access today in Biological Conservation 🦢 www.sciencedirect.com/science/arti...

1157

Stephen Vickers @stephen8vickers.bsky.social · 01/11/2025

My apps that visualise trends in the BTO annual ringing and NRS reports have been updated to include the 2024 data, which was released yesterday. Change over time plots: stephenvickers.shinyapps.io/ringing_tota... Spatial trends maps (ringing only): stephenvickers.shinyapps.io/ringing_map/

0153

Stephen Vickers @stephen8vickers.bsky.social · 02/03/2025

The Whoopers are starting to move and will soon be back on their Icelandic breeding grounds! Since @waterbirdcm.bsky.social started ringing Whooper Swans in 2023, we've tagged >600 inds. and gathered >750 sightings. Please keep an eye out and continue to report any sightings to us as they leave!

Map of whooper movements between sightings of colour rings

Flock of colour marked Whooper swans in flight

0123

Stephen Vickers @stephen8vickers.bsky.social · 14/01/2025

Flocking was associated with significantly larger non-breeding range COA shift rates, particularly in mixed-age flocks. However, flocking had no significant effect on breeding range COA shifts. 8/10

A Estimated shifts in non-breeding centres of abundance 1970–2019 for 81 analysed species in Model B. Arrows are coloured by migratory flocking behaviour – age-separated flocks (light-blue), mixed-age flocks (dark-blue) and solo (red) and transparency is set according to reciprocal error in estimated shift, such that more transparent arrows indicate lower confidence in estimates. B Phylogenetic tree of the 81 species analysed for shifts in Christmas Bird Count annual centre of abundance and incorporated into the Phylogenetic Generalised Least Squares (PGLS) model. Nodes are coloured by migratory flocking behaviour – age-separated flocks (light-blue), mixed-age flocks (dark-blue) and solo (red). C Estimated marginal mean annual shifts in Christmas Bird Count centre of abundance 1970–2019 for migratory flocking behaviour based upon 81 species of North American migratory birds in a Phylogenetic Generalised Least Squares (PGLS) model incorporating uncertainty in shift rates and controlling for biological and demographic traits. Error bars indicate standard error

130

Stephen Vickers @stephen8vickers.bsky.social · 14/01/2025

The data goes into PGLS models weighted by uncertainty in COA shift rate. One model for each season for a binary flocking var. for 122 species. Another model for each season where we categorise flocks as either age-separated or mixed-age, for 81 species. 7/10

Phylogenetic Generalised Least Squares (PGLS) model coefficients for predictors of annual shift rates (metres per year) of Breeding Bird Survey (left) and Christmas Bird Count (right) centres of abundance between 1970 and 2019. Model A assessed migratory flocking as a binary flocking vs. solo migrants for 122 species. Model B refined the assessment of migratory flocking differentiating between species that flock in age-separated and mixed-age flocks, for a reduced pool of 81 species. Red points indicate significant results inferred from credible 95th percentiles (error bars) that exclude zero

130

Stephen Vickers @stephen8vickers.bsky.social · 14/01/2025

We collated information on migratory flocking behaviour of North American migratory birds, building upon the Beauchamp et al. 2011 dataset with species accounts from Birds of the World Online and age-cohort timing calculated from USGS banding data. 5/10

Examples of the methodology for assessing overlap in timing of migration between age cohorts using banding data, showing examples of a species with high cohort overlap and thus concurrent migration timing of age classes (Vesper Sparrow Pooecetes gramineus, A–C) and a species with relatively low overlap and thus non-concurrent migratory timing (Sharp-shinned Hawk Accipiter striatus, D–F). After initially fitting GAMs to latitudes of banding events across the year for hatch-year and after hatch-year age cohorts of a given species across all USGS banding events 1960–2019 (A and D), we restricted the dataset to the autumn migration period, assessed as the temporal region where latitude shows a clear negative trend and refit the GAM models (southward migration; B and E). We then normalised the GAM-predicted mean latitudes for each age class to a 0–1 scale and calculated the overlap in area under the curve as an index of cohort temporal overlap during migration (C and F)

130

Stephen Vickers @stephen8vickers.bsky.social · 14/01/2025

We hypothesised in species that migrate solo, inter-generational change may need change in genetics of migration distance/direction. However, flocking migrants can benefit from social learning, which might act faster – but inter-generational learning can only happen in mixed-age flocks! 3/10

Pathways of seasonal range colonisation in migratory species. Migratory innovation can occur when individuals or groups make novel movements (F1), which can arise from (a) exogenous mechanisms (e.g., drift by winds or geomagnetic anomalies), exploratory dispersal within a seasonal stage, or abmigration (following other species), as well as (b) through the emergence of novel endogenous navigation programmes. Successful colonisation depends on their offspring (F2) or other conspecifics repeating the novel migration route in subsequent years. In solitary migrants (c, d), innovations are unlikely to be transferred if they originate from stochastic drift or dispersal, as F2 individuals will continue to follow inherited programmes to the original range (c). Novel endogenous navigation programmes can similarly only be transferred to F2 in solitary migrants if they are genetically heritable (d). In social migrants (e, f), by contrast, route innovations can be transferred between generations if F2 individuals follow returning F1 innovators on return migrations, regardless of whether the initial innovation mechanism was stochastic e or endogenous f

141

Stephen Vickers @stephen8vickers.bsky.social · 18/10/2024

The BTO Ringing and NRS report 2023 data has now been added to my visualisation apps: Explore graphs of numbers ringed etc. and NRS records over time: stephenvickers.shinyapps.io/ringing_tota... Explore a map of numbers ringed, etc. by region: stephenvickers.shinyapps.io/ringing_map/

030

Stephen Vickers @stephen8vickers.bsky.social · 25/09/2024

Between periods 1 (Oct ’21-Feb’22) and 3 (Aug ’22-Jan ‘23) we also found expansion of species groups associated with outbreaks. Landbirds and seabirds became more strongly associated with outbreaks, reflecting expansion in wild species impacted by HPAI. 6/8

100

Stephen Vickers @stephen8vickers.bsky.social · 25/09/2024

Narrower species groups show our method picks up associations in groups we already thought were important, like wildfowl (ducks, geese, etc.), amongst some change across periods, like non-native gamebirds post-release. Unfortunately, not many seabird species could be tested. 5/8

100

Stephen Vickers @stephen8vickers.bsky.social · 25/09/2024

But focussing on single-species is risky because patterns may match by chance when testing so many, so we look at consistency across groups of species. Period 2 is ~spring/summer ’22 (not many poultry outbreaks), 1 & 3 are the autumn/winters either side (lots of outbreaks). 4/8

100

Stephen Vickers @stephen8vickers.bsky.social · 25/09/2024

This gives us lots of species- and period-specific associations with HPAI outbreaks, most of which were positive (only positives shown here) 3/8

100

Stephen Vickers @stephen8vickers.bsky.social · 25/09/2024

Using #eBird generated abundance distributions for 152 wild bird species, we can look at maps of each species’ abundance, together with farm density, poultry numbers, and where the outbreaks were, to find species abundance patterns that best match poultry outbreaks. 2/8

110