Although locally common across most of northern U.S. and southern Canada, the Bobolink, like most species of open grassy fields and tallgrass prairie habitats, was poorly represented in the 1992-2006 MAPS database: 460 adult individuals were banded and only 17 between-year recaptures were recorded at 12 stations located within 5 Bird Conservation Regions (BCRs).  As shown on the temporal display of results, however, 447 of the 460 individual adults were banded during the final 5 years of the 15-year study period, with 55% of those banded during the final 2 years, 2005 and 2006.  Moreover, as shown on the spatial display of results, 418 (91%) of the 460 individual adults were banded in a single BCR, Central Mixed-grass Prairie (BCR 19).  As a result, usable year-specific estimates for lambda, adult apparent survival, and productivity were only available for 4, 3, and 5 years, respectively, while analogous usable BCR-specific estimates were only available for 5, 2, and 2 BCRs, respectively

Temporal and spatial analyses of 1992-2006 MAPS data produced remarkably different mean indices of adult population density for Bobolink (5.4 adults per station from the selected time-dependent temporal model and 17.6 adults per station from the selected BCR-specific model).  This difference was caused by the very high index of adult population density in BCR 19, within which 97% of the adults were banded.  These results suggest that the overall index of population density for Bobolink was high, about 45% higher than the mean index for all sparrow (Emberizidae) species, but that locally the species can sometimes occurs at densities that are as much as 3 times higher than its overall mean density.  As perhaps expected, both annual and spatial variabilities in the index of adult population density (38.9% and 53.3%, respectively) were high, about 50% and 25%, respectively, higher than the analogous mean variabilities for all sparrow species.  The linear time model for the index of adult population density produced a significantly positive Beta (0.160 adults per station per year), suggesting an increasing population trend for Bobolink, but this result was likely biased to some extent by the establishment of several new stations, at which large numbers of Bobolinks were captured, during the final few years of the 15-year study period.

The weighted geometric means of the 1992-2006 model-averaged annual and BCR-specific lambda estimates from MAPS (1.289 and 1.276, respectively, both of which were significantly different from 1.0) also suggested a significant increasing population for Bobolink.  Moreover, the linear time model for lambda indicated a significantly positive Beta (0.045), suggesting that the positive population trend itself tended to increase, a result that also may have been biased to some extent by the establishment of new stations during the final few years of the 15-year study period.  These MAPS results were in sharp contrast to the 1992-2006 survey-wide population trend from the North American Breeding Bird Survey (BBS; a lambda of 0.990, which was significantly different from 1.0) which indicated a significantly declining Bobolink population.  Perhaps surprisingly, both annual and spatial variabilities in lambda for Bobolink (24.6%, and 6.1%, respectively) were not much greater than the analogous mean variabilities for all sparrow species (22.7% and 5.2%, respectively).

Mean model-averaged estimates of adult apparent survival from temporal (0.695) and spatial (0.694) analyses for Bobolink were almost identical to each other and were very high, among the very highest estimates relative to their body masses of all 158 species included on the website.  Interestingly, however, Bobolink had lower mean temporal and spatial recapture probabilities (0.133 and 0.131, respectively) than any sparrow species, including such open grassland-inhabiting species such as Savannah and Grasshopper sparrows which had mean temporal or spatial recapture probabilities ranging from 0.289 to 0.418.  This perhaps suggests a poor fit for capture-mark-recapture models for Bobolink.  Perhaps unexpectedly, however, both annual and spatial variabilities in adult apparent survival for Bobolink (13.5% and 5.9%, respectively) were quite low, only about 60% and 50%, respectively, as high as the analogous mean variabilities for all sparrow species.

Temporal and spatial analyses produced mean productivity indices for Bobolink (0.109 and 0.066, respectively, from the selected models) that were extremely low, substantially lower than the analogous productivity indices for each sparrow species and nearly as low as the analogous indices for Dickcissel.  It seems likely that these very low reproductive indices may have been an important driver of the species’ population decline.  Annual variability in productivity (68.9%) was about 40% higher than the mean annual variability for all sparrow species, while spatial variability in productivity (28.4%) was about 40% lower than the analogous spatial variability for all sparrow species.

Temporal analyses of Bobolink vital rates showed that lambda was very strongly but not significantly positively correlated with both adult apparent survival and post-breeding effects, and rather weakly and non-significantly positively correlated with productivity.  The lack of significance of these temporal correlations was likely because they were based on only 3 or 4 years of data.  These results suggest that temporal variation in lambda was driven by temporal variation in all three important vital rates, but primarily by adult apparent survival and post-breeding effects, and to a much lesser extent by productivity.  Although these results may suggest that mortality of both adults and young birds on the non-breeding grounds could be the most important driver of temporal variation in lambda, it is likely that site-fidelity (lack of emigration) and recruitment of young on the breeding grounds, driven primarily by breeding habitat quality and stability, could also be important drivers of temporal variation in lambda.  In addition, adult apparent survival was strongly but non-significantly positively correlated with post-breeding effects and very weakly and non-significantly positively correlated with productivity.  Post-breeding effects were also rather strongly but non-significantly negatively correlated with productivity.

Temporal analyses of Bobolink vital rates also showed that the index of adult population density was very strongly negatively correlated with both lambda and adult apparent survival, rather strongly negatively correlated with post-breeding effects, and essentially not correlated at all with productivity.  Again, because of the few years of data involved in these correlations, none of them were significant.  These results suggest that population regulation in Bobolink was very strongly driven in a density-dependent manner, primarily through adult apparent survival and secondarily through post-breeding effects, but not all through productivity.  Thus, the order of strength of the positive correlations between lambda and the three other vital rates was the same as the order of strength of the negative correlations between these three vital rates and the index of adult population density.  This suggests that the temporal variation in lambda was driven virtually entirely in a density-dependent manner, a rather unusual situation among Neotropical-wintering migratory species.  Moreover, the positive correlation between adult apparent survival and post-breeding effects further suggests that temporal variation in lambda was likely primarily positively driven on the breeding grounds through temporal variation in site fidelity (lack of emigration) and recruitment of young, caused in turn by temporal variation in habitat quality and stability, rather than by temporal variation in survival of adults and young on the non-breeding grounds.  This is because, if survival of adults and young was driven in a density-dependent manner on the non-breeding grounds, we would likely expect a negative relationship between adult apparent survival and post-breeding effects, rather than the positive relationship that we found.

Because usable BCR-specific adult apparent survival estimates and productivity indices were available for Bobolink from only 2 BCRs, essentially all spatial correlations were based on data from only those 2 BCRs and thus provided virtually no useful information.  Nevertheless, it might be of interest that the signs of all three spatial correlations between lambda and the other three vital rates, as well as the signs of the spatial correlations between post-breeding effects and both adult apparent survival and productivity, were the same as the signs of the analogous temporal correlations.  These results suggest that, in general, the drivers of spatial variation in lambda for Bobolink may have been similar to the analogous drivers of temporal variation in lambda.

Summary of research and management hypotheses – We preface these remarks by noting that Bobolink was relatively poorly represented in the 1992-2006 MAPS database and, despite producing apparently consistent and informative results, much more demographic data on Bobolink will be required in order to develop definitive management and conservation plans.  Nevertheless, we suggest that research and management efforts to reverse declining populations (or maintain possible stable or increasing populations) of Bobolink should concentrate first on determining and then maintaining or enhancing habitat conditions on the breeding grounds that promote high rates of site fidelity of adult birds and high rates of recruitment of young birds, especially during years when, and in regions where, low rates of site-fidelity and recruitment appear to lead to population declines.  Such habitat conditions will likely involve high year-to-year habitat stability as well as high innate habitat quality.  Second, because MAPS data suggests that overall productivity for Bobolink may be low and deficient, research and management efforts should be directed toward determining habitat conditions on the breeding grounds that promote overall high productivity, especially in years when low productivity may be associated with population declines.  Although Bobolink population dynamics appear to be regulated primarily in a density-dependent manner, considerations of weather and climate change should be included in these research and management efforts.  Finally, although adult apparent survival seems to be high and not deficient, research and management should be directed toward determining and maintaining or enhancing habitat conditions on the non-breeding grounds that promote high survival of both adult and young birds.  As for all Neotropical-wintering migratory species, but especially those wintering in South America, increased information on migratory connectivity will be crucial for the success of these latter efforts.

Please cite this narrative as:  DeSante, D. F., D. R. Kaschube, and J. F. Saracco.  2015.  Vital Rates of North American Landbirds.  www.VitalRatesOfNorthAmericanLandbirds.org: The Institute for Bird Populations.