Temporal correlations among demographic parameters are ubiquitous but highly variable across species

Abstract

Temporal correlations among demographic parameters can strongly influence population dynamics. Our empirical knowledge, however, is very limited regarding the direction and the magnitude of these correlations and how they vary among demographic parameters and species' life histories. Here, we use long-term demographic data from 15 bird and mammal species with contrasting pace of life to quantify correlation patterns among five key demographic parameters: juvenile and adult survival, reproductive probability, reproductive success and productivity. Correlations among demographic parameters were ubiquitous, more frequently positive than negative, but strongly differed across species. Correlations did not markedly change along the slow-fast continuum of life histories, suggesting that they were more strongly driven by ecological than evolutionary factors. As positive temporal demographic correlations decrease the mean of the long-run population growth rate, the common practice of ignoring temporal correlations in population models could lead to the underestimation of extinction risks in most species.

Keywords: capture-recapture; demographic correlation; demography; environmental stochasticity; slow-fast continuum; stochastic population dynamics; temporal covariation.

FIGURE 1

Estimation of two types of correlations between survival and reproduction. In practice, reproductive parameters are often estimated within a year, while survival probabilities between two consecutive years. This means that two types of correlations between survival and reproduction could be estimated, that is, the correlation between reproduction in a given year and either the probability of survival from the previous breeding season to the current one (i.e. survival from time t − 1 to t, type 1) or survival to the next breeding season (i.e. survival from time t to t + 1, type 2). This distinction is critical because these two types of correlations are likely to be influenced by different processes such as carryover effects of the previous non‐breeding season (causing positive correlation) versus reproductive cost carried over subsequent survival (causing negative correlation). Consequently, the direction and magnitude of the correlation likely depend on the period over which the correlation is estimated

FIGURE 2

Temporal correlations estimated for three pairs of demographic parameters including juvenile survival (${\mathrm{\Phi }}_j$), adult survival (${\mathrm{\Phi }}_{\mathit{ad}}$) and reproductive success ($\pi$). We estimated correlations between reproductive success and both survival from the previous and to the next reproductive season, leading to two correlation estimates (green and blue, respectively). Species names: AF = Antarctic fulmar, BBA = black‐browed albatross, BS = bighorn sheep, BT = blue tit, EO = Eurasian oystercatcher, GS = golden‐mantled ground squirrel, GT = great tit, HS = house sparrow, Ki = kittiwake, MG = mountain goat, RD = roe deer, SP = snow petrel, SS = Savannah sparrow, WD = white‐throated dipper, WS = Weddell seal. For notation, ‘SD’ indicates standard deviation, ‘CRI’ indicates credible interval and ‘P’ indicates the proportion of the posterior distribution that has the same sign as the posterior mean

FIGURE 3

Temporal correlations estimated between productivity ($\mathrm{\Omega }$) and four demographic parameters, including juvenile survival (${\mathrm{\Phi }}_j$), adult survival (${\mathrm{\Phi }}_{\mathit{ad}}$), breeding probability ($\psi$) and reproductive success ($\pi$). We estimated correlation between productivity and both survival from the previous and to the next reproductive season, leading to two correlation estimates (green and blue, respectively). Species names: BT = blue tit, EO = Eurasian oystercatcher, GS = golden‐mantled ground squirrel, GT = great tit, HS = house sparrow, Ki = kittiwake, RD = roe deer, SS = Savannah sparrow, WD = white‐throated dipper. For notation, ‘SD’ indicates standard deviation, ‘CRI’ indicates credible interval and ‘P’ indicates the proportion of the posterior distribution that has the same sign as the posterior mean

FIGURE 4

Temporal correlations estimated for three pairs of demographic parameters including juvenile survival (${\mathrm{\Phi }}_j$), adult survival (${\mathrm{\Phi }}_{\mathit{ad}}$), reproductive probability ($\psi$) and reproductive success ($\pi$). We estimated correlations between reproductive performances and both survival from the previous and to the next reproductive season, leading to two correlation estimates (green and blue, respectively). Species names: AF = Antarctic fulmar, BBA = black‐browed albatross, BS = bighorn sheep, GS = golden‐mantled ground squirrel, Ki = kittiwake, MG = mountain goat, Oy = Eurasian oystercatcher, SP = snow petrel, WS = Weddell seal. For notation, ‘SD’ indicates standard deviation, ‘CRI’ indicates credible interval and ‘P’ indicates the proportion of the posterior distribution that has the same sign as the posterior mean

FIGURE 5

Temporal correlation between reproductive success and subsequent juvenile survival as a function of generation time. ‘P’ indicates the proportion of the posterior distribution that has the same sign as the posterior mean