Temporal autocorrelation in environmental conditions influences population dynamics through its effects on vital rates. However, a comprehensive understanding of how and to what extent temporal autocorrelation shapes population dynamics is still lacking because most empirical studies have unrealistically assumed that environmental conditions are temporally independent. Mast seeding is a biological event characterized by highly fluctuating and synchronized seed production at the tree population scale, as well as a marked negative temporal autocorrelation. In the current context of global change, mast seeding events are expected to become more frequent, leading to strengthened negative temporal autocorrelations and thereby amplified cyclicality in mast seeding dynamics with cycles of length 2 years. Theory predicts that population growth rates are maximized when the environmental cyclicality of consumer resources and their generation times are closely matched. To test this prediction, we took advantage of the long-term monitoring of a wild boar population, a widespread seed consumer species characterized by a short generation time (ca. 2 years). As expected, simulations indicated that its stochastic population growth rate increased as mast seeding dynamics became more negatively autocorrelated. Our findings demonstrate that accounting for temporal autocorrelations in environmental conditions relative to generation time of the focal population is required, especially under global warming where the cyclicality in resource dynamics is likely to change.

From 1983 to 2016, a capture-mark-recapture-recovery (CMRR) program has been running on a wild boar population located in the North of France (i.e. 11,000 ha forest of Châteauvillain-Arc-en-Barrois). Demographic data were collected annually during two specific periods: the capture period, occurring from March to September, and the hunting period, occurring from October to February. Thus, it allowed us to obtain accurate estimates of all vital rates of the studied wild boar population. 

More specifically, alive females were trapped from March to September, individually marked (or identified for females already marked) and weighed (in kg) before being released. Additionally, information on the reproductive status of females shot during the hunting season was recorded. To do so, hunting bags were analyzed from October to February. Body mass (in kg), reproductive status, which is defined as reproductive or not based on the presence of fetuses in the uterus or of Graafian follicles and/or corpora lutea in the ovaries), as well as litter size were recorded on all shot females. Last but not least, stomach contents of all shot individuals, regardless of their sex, were analyzed to indirectly assess acorn production. 

All vital rates were estimated from the demographic data collected during (re)captures and hunting bag analyses. Three stages corresponding to the following female body mass classes were considered: small (< 30 kg), medium (between 30 and 50 kg), and large (> 50 kg) females, and that for 2 environmental states corresponding to good and poor years of acorn production.

The annual survival probability of females of body class i S_i was estimated as the probability of not dying from natural causes (1 – Mn_i) multiplied by the probability of not being shot (1 – h_i) (Eq. 1). 

(1) S_i =  (1 - Mn_i) x (1 - h_i)   

Surviving females either remained in the same body mass class until the next year (pSS for the probability of small females to remain small and (1 – pML) for the probability of medium females to remain medium) or moved upward the following class (pSM for small females becoming medium, pSL for small females becoming large and pML for medium females becoming large). Large females remain in that state from one year to the next. Females from all body mass classes produced offspring. The fecundity R_i was estimated as the product of breeding proportion (BP_i), litter size (LS_i), postnatal survival (Spn), and sex ratio (assumed to be balanced, see Servanty et al. 2007 for supporting empirical evidence) (Eq. 2). 

(2) R_i =  BP x LS_i x 0.5 x Spn

Offspring produced either remained in the small body mass class (piOs) or moved upward the following class (1-piOs). We refer the readers to Touzot et al. (2020) for a more thorough description of the analytical steps for estimating vital rates. The only vital rate responding to acorn production was the probability for a female to participate in reproduction (BP_i). Thus, the probability for a female to breed was higher during good than poor years of acorn production for both medium and large-sized females.