Limited flexibility in departure timing of migratory passerines at the East-Mediterranean flyway

Abstract

The rapid pace of current global warming lead to the advancement of spring migration in the majority of long-distance migratory bird species. While data on arrival timing to breeding grounds in Europe is plentiful, information from the African departure sites are scarce. Here we analysed changes in arrival timing at a stopover site in Israel and any links to Enhanced Vegetation Index (EVI) on the species-specific African non-breeding range in three migratory passerines between 2000–2017. Differences in wing length between early and late arriving individuals were also examined as a proxy for migration distance. We found that male redstart, but not females, advanced arrival to stopover site, but interestingly, not as a response to EVI phenology. Blackcap and barred warbler did not shift arrival timing significantly, although the arrival of blackcap was dependent on EVI. Barred warbler from the early arrival phase had longer wings, suggesting different populations. Our study further supports the existence species-specific migration decisions and inter-sexual differences, which may be triggered by both exogenous (local vegetation condition) and endogenous cues. Given rapid rate of changes in environmental conditions at higher latitudes, some migrants may experience difficulty in the race to match global changes to ensure their survival.

A Comparison Between Quantile Regression and Linear Regression on Empirical Quantiles for Phenological Analysis in Migratory Response to Climate Change

“…This means that overall, species have advanced their spring arrival. This is consistent with many previously published studies of bird phenology (Jonzén et al, 2006;Tøttrup et al, 2006;Lehikoinen et al, 2019;Usui et al, 2017;Aharon-Rotman et al, 2021), here with added details of how this effect changes across the arrival time distribution, both within species and generally across species. As a closing observations we notice that for most species i, βmeq it (τ ) and βqr it (τ ) follow each other closely.…”

“…, x jp ), consisting of an intercept term 1, the year t j that individual j was recorded, and p additional covariates. Typical covariates are the binary age (juvenile or adult) and sex (female or male), but continuous covariates such as wing length may also be used (Aharon-Rotman et al, 2021). We assume that t j takes values in T = {1, .…”

“…We will use year as the covariate of primary interest, as the interpretation of the associated coefficient will tell us what change has occured over time. Other covariates may also be used, such as the age and sex of birds or other biometrics (Aharon-Rotman et al, 2021). The empirical arrival distribution of any particular year can be visualized as a step function with variable step sizes.…”

“…The purpose of this article is to compare linear models for empirical quantiles with linear quantile models without (Koenker, 2005) or with (Geraci and Bottai, 2014) random effects. The latter model, which essentially is a mixed effects model for quantile regression, has recently been employed for phenological analyses (Aharon-Rotman et al, 2021). The remainder of our paper is structured as follows.…”

“…Thus, we have highly reliable data, and it is possible to incorporate the covariates age and sex in the single species models. The Eurasian Blackcap is quite a common choice of species for phenological analysis, and an interesting modeling approach can be found inAharon-Rotman et al (2021) for this species, using lqmm with year as block factor and fixed effect. The distribution of birds across age and sex levels are shown in Table1, whereas the number of birds recorded each year is illustrated in Figure2.…”