Population demography of lark buntings: post-fledging survival, fecundity, and breeding decisions

Abstract

Grassland bird species have experienced the largest population declines of any terrestrial avian group in the past four decades in North America. Lark buntings (Calamospiza melanocorys), endemic to the grasslands of North America, have declined by 2.5% per year in Colorado and 2.1% per year in the High Plains physiographic region between 1966 and 2003. Although reasons for these declines have yet to be elucidated, factors influencing reproductive productivity and juvenile survival on the breeding grounds are among the possibilities. To assess population declines, direct and reliable estimates of survival, fecundity, and dispersal are required. Consequently, I estimated demographic parameters (i.e., post-fledging survival and fecundity) for lark buntings in a relatively extensive (62% grassland remaining within a 21,600 km2 area) native shortgrass landscape on the Pawnee National Grassland, Colorado in 2001-2003. In chapter 1 ,1 applied an information-theoretic approach to evaluate factors affecting post-fledging survival of lark buntings. I estimated daily and 22-day post-fledging survival (n = 206, 82 broods) using radio-telemetry and color bands to track fledglings. For 2001-2002 data, I employed the joint model in program MARK to examine the effects of drought condition, time in season, age, nestling condition (rank, condition index, or brood size), mark type (radio-marked versus band-only), and sex of attending parent on post-fledging survival. Estimated daily survival probabilities (± SE) were higher under normal precipitation (2001: 0.933 ± 0.010) and mild drought conditions (2003: 0.933 ± 0.013) than during a severe drought (2002: 0.908 ± 0.011). Post-fledging daily survival probabilities in 2001 and 2002 were best explained by models that incorporated drought condition, time in season (quadratic trend), ages ≤ 3, and rank x drought interaction. Daily survival probabilities were lower under severe drought conditions than in a normal year; the model-averaged coefficient for the additive effect of drought on survival of fledglings was ^βdrought = -3.99 (95% CI = -7.94, -0.05). Models also revealed greater survival in mid-season (^βT2 = -0.002, 95% CI = -0.003, -0.001). Survival was lower for recently-fledged young (ages ≤ 3) than older fledglings. Rank was an important predictor of fledgling survival only during the severe drought of 2002. Both mark type and sex of attending parent had no effect on survival. Survival estimates that account for age, condition of young, ecological conditions and other factors are important for parameterization of realistic population models. My results suggest that age-specific estimates can help identify critical time periods within the species' life-cycle, and that species-specific estimates of post-fledging survival are more informative than generalized estimates used in population growth models. In chapter 2, I evaluated the stability of a breeding population of lark buntings using population-specific values for two demographic parameters, fecundity and post-fledging survival, and quantified breeding decisions of females. I followed 67 radio-marked female lark buntings captured on their initial nests to determine annual fecundity and evaluate factors affecting it, such as nest survival and breeding decisions (renesting, ceased breeding, dispersal). Collectively, the 67 females built 112 nests (1.67 ± 0.07 nests per female per season (mean ± 1 SE); range 1-3) of which 45 were subsequent nests (34 second nests and 11 third nests). Daily nest survival estimates were similar for initial and subsequent nests with overall nest survival (DSP19) of 30.7% and 31.7%, respectively. Nest predation was the most common cause of failure (92%). Capture and radio-marking of females did not affect nest survival. Lark bunting dispersal probabilities increased among females that (1) fledged young from initial nests, and (2) lost their initial nest late in the season. Conservative and liberal estimates of mean annual fecundity were 0.96 ± 0.11 and 1.24 ± 0.09 female offspring female-1. Given these fecundity estimates along with estimates of juvenile survival obtained for this species, adult survival values of 70-76% are necessary to achieve a stable population. Adult survival estimates for prairie passerines range between 55-65%, indicating that this study area is unable to maintain a stable population in the absence of immigration. My results suggest that population declines of prairie birds are caused, at least in part, by breeding ground phenomena. Further consideration of the quality of breeding areas in the context of the predator communities is clearly needed.