November 20th, 2018
Examining spatiotemporal patterns of bird movements using EBP observations and ringing recoveries
There are several sources of information on bird movements, including networks of weather radars , ringing data, and tracking data, that complement the information obtained from observational data gathered by EBP. Such information is particularly necessary if we are to obtain a good understanding of migratory connectivity and population-specific migration routes. LIFE EBP Action 12 started to explore how these different datasets can be combined to provide a more comprehensive picture of bird movements. This is an important approach for the development of the EBP project and has the potential to create powerful synergies among the existing research networks. Such analyses have the potential to provide crucial information for the conservation and management of migratory bird populations, which may be affected by environmental conditions at each of the different stages of their annual cycles.
In this post we explore possible ways in which data on marked individuals from the EURING Databank (EDB) and EBP observational data can be combined to describe the spatiotemporal nature of avian movement patterns. This may also provide useful ideas for the Eurasian African Bird Migration Atlas that is to be developed by EURING on behalf of CMS.
Here we use data on the Eurasian Curlew Numenius arquata and the Common Redstart Phoenicurus phoenicurus to exemplify how ringing data can highlight spatial connectivity patterns between different populations which cannot be detected from observational data alone.
The maps
To easily visualize how the information provided by each data source can complement each other, we prepared week by week animated maps at a 30x30 km square resolution combining EBP occurrence data and EURING databank recovery information.
The recovery of each individual bird was assigned to a “region of origin”, according to the area of Europe in which it was first encountered (Figure 1). Moreover, we added a three-week buffer to all ringing recovery encounters to overcome the sparseness of the ringing data compared to the EBP observational data. This means that an individual ringing recovery record is added to the map for both the week preceding and the week following the focal week, in addition to being mapped in the focal week itself.
Figure 1. Regions of origin as defined for the bird recoveries.
Results for Eurasian Curlew
EBP data shows that the Eurasian Curlew winters primarily along the coasts of the UK, France, Belgium, the Netherlands, Denmark, and Norway, with fewer observations along coastal Iberia and Italy (Figure 2). Spring migration shows birds moving into inland regions of the UK, central Europe, and northern Europe for the duration of the breeding season. In mid to late summer, Eurasian Curlew move out of inland and northern regions and observations become more concentrated in coastal areas by late September/early October. EBP data show substantial overlap in wintering, migration and breeding range, particularly in the UK and continental countries bordering the North Sea.
Figure 2. Eurasian Curlew weekly animated occurrence map based on EBP data.
Ringing data show that the majority of individuals originate from West, Central and Northern regions (Figure 3). Patterns are difficult to grasp in the animated map combining all recoveries (Figure 4) but those focussed on each of the three regions with the most data (western, central and northern regions of Europe) reveal some more interesting patterns that could not be discerned from the observational EBP data alone (Figures 5-7).
Figure 3. Locations of first encounter of Eurasian Curlew recoveries (the colour of the recoveries refers to their region of origin; cf. Figure 1).
Figure 4. Eurasian Curlew weekly animated map combining occurence EBP data (yellow dots) and EURING Databank recoveries (the colour of the recoveries refers to their region of origin; cf. Figure 1).
Ringing recovery data demonstrate that despite largely overlapping wintering, migration and breeding ranges, there is substantial migratory connectivity between particular Eurasian Curlew populations. Eurasian Curlew wintering in UK, France, Belgium and the Netherlands comprise both local breeders, but also birds which breed in Finland and Russia (Figure 5).
Figure 5. Eurasian Curlew weekly animated map combining occurence EBP data (yellow dots) and EURING Databank recoveries whose region of origin is western Europe (cf. Figure 1).
While Figure 5 shows movement between wintering Eurasian Curlew in the UK and France to northern Europe, the reverse can be seen in Figure 6. Eurasian Curlew first ringed in northern Europe, primarily originating in Finland as breeding birds, spend the winter in coastal area of the East Atlantic, primarily in the UK and France with some in Belgium and the Netherlands. Ringing data also clearly demonstrate that in July and August, Denmark is a staging site on autumn migration for birds originating from northern Europe.
Figure 6. Eurasian Curlew weekly animated map combining occurence EBP data (yellow dots) and EURING Databank recoveries whose region of origin is northen Europe (cf. Figure 1).
In contrast, Eurasian Curlew ringed in central Europe comprise multiple different populations (Figure 7). Birds originally ringed in central Europe appear to be either migrants or local breeders. Some winter in Eastern Atlantic coastal areas and breed in central Europe, primarily Germany. Other individuals are likely captured on their migration to breeding locations in Russia. Some Eurasian Curlew are captured in south-central Europe only in March, and are thus likely migrants wintering on the North African Mediterranean coast and breed either in central Europe or are captured on their migration to breeding sites in northern Europe.
Figure 7. Eurasian Curlew weekly animated map combining occurence EBP data (yellow dots) and EURING Databank recoveries whose region of origin is central Europe (cf. Figure 1).
Results for Common Redstart
EBP data show that Common Redstart are a typical Afro-Palearctic migrant, and do not start appearing in southern Europe until early March (Figure 8). By early April, birds have reached the UK, the Netherlands and northern Germany, and by late April and early May occur as far north as northern Scandinavia. Birds begin disappearing from northern Europe throughout July and August on autumn migration, and there are few records of Common Redstart anywhere in continental Europe after mid October.
Figure 8. Common Redstart weekly animated occurrence map based on EBP data.
The majority of individuals were ringed in West, Central and Northern regions (Figure 9). As with Eurasian Curlew patterns are difficult to grasp in the animated map combining all recoveries (Figure 10) but those focussed on each of the three regions with the most data (western, central, south-central and northern regions of Europe) reveal some more interesting patterns of movement that could not be discerned from the observational EBP data alone (Figures 11-14). We do not discuss migration through the eastern Mediterranean here, as there are too few EBP and ringing data from this region.
Figure 9. Locations of first encounter of Common Redstart recoveries (the colour of the recoveries refers to their region of origin; cf. Figure 1).
Figure 10. Common Redstart weekly animated map combining occurence EBP data (yellow dots) and EURING Databank recoveries (the colour of the recoveries refers to their region of origin; cf. Figure 1).
While the EBP data show Common Redstart migrating in spring and autumn through both Iberia and Italy, ringing recovery data demonstrate that these birds originate from different populations. Common Redstart first encountered in northwestern countries in Europe (UK, France, Belgium and the Netherlands) migrate almost exclusively through Iberia, with very few ringing recoveries in Italy on either northwards or southwards migration (Figure 11).
Figure 11. Common Redstart weekly animated map combining occurence EBP data (yellow dots) and EURING Databank recoveries whose region of origin is western Europe (cf. Figure 1).
In contrast, a few individuals first encountered in northern or central European countries migrate through Italy, though the majority from both regions appear to migrate through Iberia in both spring and autumn, and move on to return to breed in either northern or central Europe. A few individuals first encountered in central Europe were clearly migrants which breed in northern Europe (Figures 12-13).
Figure 12. Common Redstart weekly animated map combining occurence EBP data (yellow dots) and EURING Databank recoveries whose region of origin is northen Europe (cf. Figure 1).
Figure 13. Common Redstart weekly animated map combining occurence EBP data (yellow dots) and EURING Databank recoveries whose region of origin is central Europe (cf. Figure 1).
On the other hand, the majority of Common Redstart observed in Italy appear to be birds which both migrate through and breed in this region, though a few individuals encountered here do migrate to central or northern Europe (Figure 14).
Figure 14. Common Redstart weekly animated map combining occurence EBP data (yellow dots) and EURING Databank recoveries whose region of origin is south-central Europe (cf. Figure 1).
Future opportunities
Here we have adopted a very simple approach to explore how EBP data on year round variation in occurrence might be combined with data on marked birds to show the year round movement patterns of different populations. A more comprehensive analysis would need to take account of a number of additional variables including the time of year when birds were marked, spatial variation in recovery probabilities and the potential for differences in movement patterns between age and sex categories. Despite these limitations the results are encouraging, demonstrating clear differences in population-specific movement patterns that can be linked to overall variation in occurrence.
The Eurasian African Bird Migration Atlas will provide a first opportunity to develop some of these ideas further. In the medium term we look forward to large-scale quantitative models that will provide robust assessments of population-specific movement patterns by combining large-scale occurrence data of the type gathered by EBP with data on individual and population-specific movements from ringing, tracking and other data sources.
A contribution by: Samantha Franks & Stephen Baillie (BTO)
