American Goldfinch, a common species of semi-open country, brushy fields, and riparian habitats throughout much of North America, was the eighth most abundantly captured of all 158 species included on this website.  A total of 11,496 banded adults and 919 between-year recaptures of adults from 225 stations in 19 BCRs were included in the 1992-2006 MAPS database analyzed here.  As shown in the spatial display of results, the largest numbers of adults, about 15% of the total, were banded in the Eastern Tallgrass Prairie (BCR 22).  Large numbers of adults were also banded in western areas, e.g., the Northern Pacific Rainforest (BCR 5) and Coastal California (BCR 32), and eastern areas, e.g., the Lower Great Lakes/St.  Lawrence Plain (BCR 13) and Appalachian Mountains (BCR 28).  Unlike several other finch species included on this website, the numbers of between-year recaptures of adults and the associated mean recapture probabilities (0.148 and 0.245 from Pradel reverse-time and transient Cormack-Jolly-Seber capture-mark-recapture models, respectively) were not extremely low, thus indicating higher levels of site fidelity for American Goldfinch than for those some other finch species.

Temporal and spatial analyses of 1992-2006 program-wide MAPS data showed weighted mean population density indices for American Goldfinch of 3.2 and 5.4 adults per station, respectively.  These indices were about 14% and 25% higher, respectively, than the analogous mean indices for all species included on the website, and were very slightly higher than those for Pine Siskin and second only to those for Lesser Goldfinch as the highest population density indices of the six finch species included here.  Annual variability in adult population density for American Goldfinch (16.9%) was low, about 28% lower than the analogous mean variability for all species and the lowest of the six finch species.  Spatial variability in population density (43.2%), however, was relatively much higher, about 8% higher than the analogous mean variability for all species, but still was the third lowest of the six finch species.  The linear time model for the index of adult population density, like that for each of the five other finch species, produced a significantly negative Beta of -0.039 adults per station per year, suggesting an overall decreasing population for American Goldfinch.

The weighted geometric mean of the model-averaged annual lambda estimates (0.965, which was not significantly different from 1.0) suggested a non-significantly decreasing population for American Goldfinch, although the time-constant model produced a lambda of 0.950 that was significantly different from 1.0.  The weighted geometric mean of the model-averaged BCR-specific lambda estimates (0.959, which was significantly different from 1.0) also suggested a significantly decreasing population, which agreed with the significantly decreasing index of population density suggested above.  The population trends indicated by these MAPS estimates differed, however, from the program-wide population trend for American Goldfinch from the 1992-2006 North American Breeding Bird Survey (BBS; a lambda of 1.006 which was significantly different than 1.0) which indicated a slightly but significantly increasing population.  The cause of these differences between the MAPS and BBS population trends for American Goldfinch is not clear, especially in light of the linear time model for MAPS lambda which produced a significantly positive Beta of 0.047, suggesting that the MAPS lambda tended to increase during the latter years of the 15 years of study.  Annual variability of lambda for American Goldfinch (31.9%) was about 39% higher than the analogous mean variability for all species, but was the second lowest of the finch species included here.  In contrast, spatial variability of lambda (5.0%) was about 10% lower than the analogous mean variability for all species, and was the lowest of the six finch species.

Temporal and spatial analyses of 1992-2006 program-wide MAPS data produced mean model-averaged estimates of adult apparent survival for American Goldfinch (0.445 and 0.447, respectively) that were very similar to each other and were somewhat lower than expected considering the species’ body mass as compared to the body masses of all species on the website.  This suggests that overall adult apparent survival for American Goldfinch might have been slightly deficient and could have been a possible driver of the population decrease of the species, at least as documented by MAPS data.  Annual variability in adult apparent survival (13.4%) was low, about 43% lower than the analogous mean variability for all species and the second lowest of the six finch species.  Spatial variability in adult apparent survival (16.0%) was relatively higher, about 3% higher than the analogous mean variability for all species, but still was the third lowest of the finch species.

Temporal and spatial analyses yielded mean productivity indices for American Goldfinch (0.004 and 0.031, respectively, from the selected models) that were essentially the lowest of all 158 species on the website.  Only Cedar Waxwing had productivity indices that were nearly as low as those for American Goldfinch.  The very low productivity indices for both of these species were caused by the fact that their young do not fledge and become independent from their parents, and thus become available for capture in mist nets, until very late in summer, just when the operation of MAPS stations ends each year.  Thus, despite these very low productivity indices, it is not clear whether or not overall productivity for American Goldfinch was deficient and could have been a possible driver of the population declines documented by MAPS.  Annual variability in productivity for American Goldfinch (28.1%) was low, about 34% lower than the analogous mean variability for all species and the second lowest of the finch species, while spatial variability in productivity (56.3%) was relatively higher, nearly 20% higher than the analogous mean variability for all species, but still the third lowest of the six finch species.

Temporal analyses for American Goldfinch showed that lambda was strongly and highly significantly positively correlated with post-breeding effects, rather weakly and non-significantly positively correlated with adult apparent survival, and not correlated at all with productivity.  These results suggest that annual variation in lambda for American Goldfinch was driven primarily by annual variation in post-breeding effects and secondarily by annual variation in adult apparent survival, but was not at all driven by annual variation in productivity.  Because of this species’ low adult apparent survival, it seems likely that the annual variation in post-breeding effects more strongly reflected annual variation in first-year survival of young birds during the non-breeding season than annual variation in the recruitment of surviving young during the subsequent breeding season.  Post-breeding effects were strongly and significantly negatively correlated with productivity, but were only weakly and non-significantly negatively correlated with adult apparent survival, which may provide some support for the above suggestion regarding recruitment of young.  As perhaps expected from the latter two correlations, adult apparent survival was moderately but not significantly positively correlated with productivity.

Temporal analyses also showed that both lambda and adult apparent survival were essentially not correlated at all with the index of adult population density, suggesting that very little or no density dependence was involved in American Goldfinch population regulation.  In addition, productivity was moderately but non-significantly negatively correlated with adult population density, while post-breeding effects were moderately and non-significantly positively correlated with population density.  These results suggest that if any density dependence did occur in the population dynamics of American Goldfinch, it likely was effected through productivity.  These results also indicate that the vital rate whose annual variation was the primary driver of annual variation in lambda (post-breeding effects) was not the vital rate through which density dependence was primarily effected.

Spatial correlations between lambda and the three other vital rates for American Goldfinch were similar, at least in sign and relative strength of correlation, to the analogous temporal correlations.  Thus, lambda was rather strongly positively correlated with post-breeding effects, moderately positively correlated with adult apparent survival, and strongly negatively correlated with productivity, although all three of these spatial correlations were non-significant.  These results therefore suggest that the drivers of spatial variation in lambda were the same as the drivers of annual variation in lambda, a rather uncommon situation among the landbird species included on this website.  Interestingly, however, the three spatial correlations among adult apparent survival, productivity, and post-breeding effects differed in some respects from the analogous temporal correlations.  Thus post-breeding effects were extremely strongly negatively correlated with productivity and very strongly positively correlated with adult apparent survival, while adult apparent survival was strongly negatively correlated with productivity; moreover, all three of these correlations were significant.

Summary of research and management hypotheses – Although BBS results suggest a small but significant overall population increase for American Goldfinch, MAPS results suggests a rather strong and possibly significant overall population decline.  Thus, we suggest that demographic monitoring of American Goldfinch populations continue, and if further population declines are confirmed, research and management effort to reverse the declines and maintain stable or increasing populations should be implemented.  We suggest that these efforts should first aim to determine and manage for habitat and environmental characteristics that promote higher survival of adults and, especially, young birds during the non-breeding season, especially during years when, and within regions where, survival of young and adult birds is low and leads to population declines.  Because the post-breeding effects and adult apparent survival that drive both temporal and spatial variation in lambda could involve recruitment of young and emigration of adults during the breeding season, as well as survival during the non-breeding season, we suggest that habitat and environmental conditions that promote high recruitment rates and low emigration rates also be determined and managed.  We also suggest that additional demographic monitoring be undertaken to determine whether or not American Goldfinch productivity is deficient and, if so, management to enhance it should be implemented.  Finally, we suggest that to achieve effective conservation results for American Goldfinch, considerations of weather and climate change will need to be included in the research and management efforts outlined above, and considerable information on the migratory connectivity of their populations will need to be obtained.

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.