Skip to main content

New insights into the patterns and drivers of avian altitudinal migration from a growing crowdsourcing data source

Cite this dataset

Tsai, Pei-Yu et al. (2020). New insights into the patterns and drivers of avian altitudinal migration from a growing crowdsourcing data source [Dataset]. Dryad.


Altitudinal migration is a common and important but understudied behavior in birds. Difficulty in characterizing avian altitudinal migration has prevented a comprehensive understanding of this behavior. To address this, we investigated the altitudinal migration patterns and explored potential drivers for a major proportion (~70%) of the entire resident bird community along an almost 4,000 m elevational gradient on the main island of Taiwan. Based on the occurrence records collected by citizen scientists, we examined the seasonal shifts in the center and the upper and lower boundaries of elevational distributions for 104 individual species. We then built phylogeny-controlled regression models to investigate the associations between the birds’ seasonal distribution shifts and seven of their traits, and examined whether the observed shifts can be explained by three main hypotheses on potential drivers. Results showed that at least 60 species (58%) seasonally changed their distributions along elevations. While most of them (42 species) tended to move downhill in winter, a considerable number of species (14) tended to move uphill. While the species breeding at high or low elevations tended to move downhill in winter, those breeding at medium-low elevations tended to move or extend their distributions to higher elevations. Our regression models suggested that seasonal variations in climates and food availability could be major drivers of the behavior. However, the three hypotheses can only partially explain the observed downhill migration patterns and none of them can well explain the uphill patterns, indicating an important knowledge gap. This study investigated avian altitudinal migration from a new perspective with a novel and generalizable approach, and revealed interesting patterns that could be difficult to identify with conventional approaches. It demonstrated the power of citizen science data to provide new insights into this behavior by characterizing the general patterns and mechanisms across a large number of species.

Usage notes

The zip file contains one R script and one "data" folder. The "data" folder contains the eBird data (ebd_TW_relFeb-2019_ebirdp.txt), phylogenetic trees (phylogenytree.nex), trait data (Taiwan_Breeding_Bird_Trait.csv), a shapefile of the study area (Taiwan_diss_without_island), and dem data (in the folder of "twdtm_asterV2_30m). All results and major figures reported in the associated paper can be obtained by running the R script (almig_tw_202009.R).


National Science and Technology Council, Award: 170-2621-B-001-001-MY2

Academia Sinica, Award: Biodiversity Research Center