Global-scale patterns of avian survival with latitude
We found support for the oft-toted latitudinal survival gradient, but this depended on both the geographic region and taxa being considered. Specifically, we demonstrate that the previously noted inverse relationship between latitude and survival is only weakly borne out across northern hemisphere avifauna overall, but that this effect is strengthened when considering only passerines or species inhabiting the mainland. In contrast, the relationship was only evident in the southern hemisphere for survival estimates from New World birds, the vast majority of which were passerines. When considered independently, there was no indication that nonpasserines had higher survival with decreasing latitude in either hemisphere. Overall, our meta-analysis reveals that while some tropical birds may be longer lived than their temperate counterparts, the shape of the latitude-survival response is likely to differ among species and between hemispheres.
Our synthesis is the first to assess global-scale patterns in avian survival rates; previous studies have either been limited geographically (Karr et al. 1990; Peach et al. 2001; Lloyd et al.2014), or have focused on a narrower range of species, such as raptors (Newton et al. 2016) or shorebirds (Méndez et al. 2018). To date, the most extensive analysis of avian survival and latitude comes from a study of 12 locations spanning 60° across the Americas (Muñoz et al. 2018). Our global-scale analysis compliments that of Muñoz et al. (2018), who reported a linear decrease in survival of roughly 2.1% for every 10° increase in latitude for passerine birds from Alaska to Peru, similar to what we observed for northern hemisphere species worldwide. Granted both our studies used a meta-analytical approach, Muñoz et al. (2018) conducted their analysis using a Bayesian mode of inference and considered only forest-dwelling passerines, while our study includes survival estimates of both passerines and nonpasserines from a variety of habitats, which we investigated using a maximum-likelihood approach. We also fit regression lines for latitude both north and south of the equator rather than testing the relationship between survival and absolute latitude. This latter point is particularly important, given that one general explanation for spatial patterns in life-history traits is that they arise from natural selection imposed by latitudinal gradients in environmental conditions (Cardillo 2002), which differ between hemispheres (Chown et al. 2004). Despite our use of different methods, the fact that we obtained some common results lends increased support to the overall relationship. Moreover, with our analysis, we provide a strong mechanistic basis for understanding variation in survival rates, as it better reflects the climatic variables that underlie latitude in the northern and southern hemispheres.
Hemispheric asymmetries in other patterns of avian life-history traits, such as timing of reproduction (Covas et al. 1999), clutch size (Moreau 1944; Martin et al. 2006; Lloyd et al. 2014), and parental care (Russell et al. 2004; Llambías et al. 2015), are well documented. The global patterns we identified are also congruent with the idea of a differential response of life-histories between hemispheres ― we detected an inverse relationship between survival and latitude in the northern hemisphere but found little indication that this association was mirrored by southern hemisphere species overall. Only when we analyzed biogeographic realms in the southern hemisphere separately did we find that New World birds showed higher survival with decreasing latitude. This pattern is deceptive, however, since southern hemisphere nonpasserines account for little more than 1% of the effect sizes analyzed in the New World data subset. We therefore interpret this result as evidence of the latitudinal survival gradient in South American passerines. This means that for Old World birds, tropical species had similar survival rates to birds from the austral zone, and this was likely to be true regardless of whether they were passerines or nonpasserines. Survival estimates from Australasia and Oceania, biogeographic realms not traditionally included in the New / Old world classification, also reflected this same pattern and showed no evidence of a negative relationship with latitude.
Such differences may be explained, in part, by the historical geography and latitudinal positions of the continents. For the last 15 million years, South America has extended roughly 20° further into the southern hemisphere than continental landmasses in the Old World. Thus, one reason we may have detected a negative trend in survival for southern hemisphere birds, but only in the New World, could simply be due to the greater range of latitudes and climatic conditions available to landbirds from South America with which to adapt. For example, latitudes greater than 35° S are characterized by higher seasonality and mean annual temperatures ≤0°C (Chown et al. 2004); thus, this result may be indicative of a threshold response of avian survival to freezing temperatures and / or a more seasonal environment. Supporting this idea, mean survival of South American passerines that occurred at latitudes higher than 35° S (survival rate = 0.38, n = 8) was lower on average than those from the highest latitudes occupied by birds in Africa (Old World survival at 34° S = 0.69, n = 19). Only one other study has addressed the question of a latitudinal survival gradient in the southern hemisphere; Lloyd et al. (2014) found no indication of higher survival for birds living in tropical Malawi compared to austral South Africa. Our results are congruent with those findings and suggest that higher survival of tropical birds may be a pattern localized primarily to passerines from the northern hemisphere and in South America, where factors such as a more seasonal environment may limit resource availability and constrain species survival.