Abstract. Maternal characteristics, social dynamics, and environmental factors can all influence reproduction and survival as well as shape trade-offs that might arise between these components of fitness. Short-lived mammals like the golden-mantled ground squirrel (GMGS; Callospermophilus lateralis) tend to maximize effort towards current reproduction at the expense of survival but may be complicated by other aspects of the species’ life history and environment.. Here, we use 25 years of data (1995-2020) collected from a population of GMGS at the Rocky Mountain Biological Research Laboratory in Gothic, Colorado, to test the effect of several maternal characteristics (e.g., age, experience, timing of litter emergence), social context (e.g., litter sex ratio, kin density), and environmental context (e.g., date of bare ground, length of vegetative growing season) on survival of reproductive female GMGS using Cox proportional hazard models. Our results indicated that social dynamics (i.e., density) and environmental conditions (i.e., standardized first day of permanent snow cover and length of growing season) explained significant variation in annual maternal survival, while maternal characteristics did not. A higher density of related breeding females and the total number of females (both related and unrelated to the focal mother) were associated with an increase in the mortality hazard. A later standardized date of the first day of permanent snow cover and a shorter growing season both reduced the maternal mortality hazard. Together, our results suggest that factors extrinsic to the squirrels affect maternal survival, and thus may also influence local population growth and dynamics in GMGS and other short-lived, territorial mammal species.

Study Site & Data Collection

The Rocky Mountain Biological Research Laboratory (RMBL) is located in the East River Valley of Gunnison, Colorado (38°58'N, 106°59'W). The 13-ha study site is situated at an elevation of 2,900 m above sea level. A complete census of the GMGS population at RMBL began in 1990, with detailed pedigrees constructed for all adult females beginning in 1995.  We used data collected from 1995 to 2020 on 141 resident females for which we had 249 annual observations and 131 mortality events. Females entered the study the year of their first reproduction, which occurred at age one or two for nearly all individuals. Because we only included reproductive females, any instance where an individual disappeared from the study site was classified as a mortality event as post-breeding dispersal is extremely rare (<1% of females, Van Vuren, unpublished).

To develop a census database for analysis, field methods involved monthly trapping events during the active season and daily observations. Once trapped, individuals were identified by ear tag (Monel 1005-1) and by unique dye mark (Nyanzol-D) applied to the fur, weighed (g) with a Pesola scale, and assessed for reproductive status based on nipple development. Individuals were classified as ‘alive’ if they were trapped or seen during daily observations, but adult females who failed to return to the system were classified as ‘dead.’ For immigrant individuals that were not born in the study site but entered and established a territory later, we used timing of entrance and mass at capture as indicators of age. Immigrants who were captured late in summer and had a mass consistent with other young of the year were classified as juveniles and assigned a known age; by contrast, immigrants who were captured in spring could be yearlings or older and mass could not be used as a categorical difference, making their age unknown. 

Additional details regarding the methods and analyses are provided in the manuscript.