Estimating abundance and phenology from transect count data with GLMs

Weather anomalies more important than climate means in driving insect phenology

Studies of long-term trends in phenology often rely on climatic averages or accumulated heat, overlooking climate variability. Here we test the hypothesis that unusual weather conditions are critical in driving adult insect phenology. First, we generate phenological estimates for Lepidoptera (moths and butterflies) across the Eastern USA, and over a 70 year period, using natural history collections data. Next, we assemble a set of predictors, including the number of unusually warm and cold days prior to, and during, the adult flight period. We then use phylogenetically informed linear mixed effects models to evaluate effects of unusual weather events, climate context, species traits, and their interactions on flight onset, offset and duration. We find increasing numbers of both warm and cold days were strong effects, dramatically increasing flight duration. This strong effect on duration is likely driven by differential onset and termination dynamics. For flight onset, impact of unusual climate conditions is dependent on climatic context, but for flight cessation, more unusually cold days always lead to later termination particularly for multivoltine species. These results show that understanding phenological responses under global change must account for unusual weather events, especially given they are predicted to increase in frequency and severity.

Quantifying phenological landmarks of migration shows nonuniform use of the Caribbean by shorebirds

Of the boreal- and Arctic-breeding North American shorebirds that migrate south through the Caribbean, most individuals continue farther south. However, for many species, some individuals remain beyond the southbound migration period (i.e., throughout the temperate winter and/or summer). This variation among individuals adds complexity to observation data, obscures migration patterns, and could prevent the examination of the use of different Caribbean regions by various shorebird species during migration and in the nonmigratory seasons. Here, we present a novel method that leverages a well-established statistical approach (generalized additive models) to systematically identify migration phenology even for complex passage migrant species with individuals that remain beyond migration. Our method identifies the active migration period using derivatives of a fitted GAM and then calculates phenology metrics based on quantiles of that migration period. We also developed indices to quantify oversummering and overwintering patterns with respect to migration. We analyzed eBird data for 16 North American shorebird species as they traveled South through the insular Caribbean, identifying separate migratory patterns for Cuba, Puerto Rico, Guadeloupe, Aruba, Bonaire, Curaçao, and Trinidad and Tobago. Our results confirm past reports and provide additional detail on shorebird migration in the Caribbean, and identify several previously unpublished regional patterns. Despite Puerto Rico being farther north and closer to continental North America, most species reached Puerto Rico later than other regions, supporting a long-standing hypothesis that migration strategy (transcontinental vs. transoceanic) leads to geographic differences in migration timing. We also found distinct patterns of migration curves, with some regions and species consistently having either symmetrical or skewed curves; these differences in migration curve shape reflect different migration processes. Our novel method proved reliable and adaptable for most species and serves as a valuable tool for identifying phenological patterns in complex migration data, potentially unlocking previously intractable data.

Multi-season climate projections forecast declines in migratory monarch butterflies

Climate change poses a unique threat to migratory species as it has the potential to alter environmental conditions at multiple points along a species' migratory route. The eastern migratory population of monarch butterflies (Danaus plexippus) has declined markedly over the last few decades, in part due to variation in breeding-season climate. Here, we combined a retrospective, annual-cycle model for the eastern monarch population with climate projections within the spring breeding grounds in eastern Texas and across the summer breeding grounds in the midwestern U.S. and southern Ontario, Canada to evaluate how monarchs are likely to respond to climate change over the next century. Our results reveal that projected changes in breeding-season climate are likely to lead to decreases in monarch abundance, with high potential for overwintering population size to fall below the historical minimum three or more times in the next two decades. Climatic changes across the expansive summer breeding grounds will also cause shifts in the distribution of monarchs, with higher projected abundances in areas that become wetter but not appreciably hotter (e.g., northern Ohio) and declines in abundance where summer temperatures are projected to increase well above those observed in the recent past (e.g., northern Minnesota). Although climate uncertainties dominate long-term population forecasts, our analyses suggest that we can improve precision of near-term forecasts by collecting targeted data to better understand relationships between breeding-season climate variables and local monarch abundance. Overall, our results highlight the importance of accounting for the impacts of climate changes throughout the full-annual cycle of migratory species.

Predicting surface abundance of federally threatened Jollyville Plateau Salamanders (Eurycea tonkawae) to inform management activities at a highly modified urban spring

Urban expansion has contributed to the loss of habitat for range restricted species across the globe. Managing wildlife populations within these urban settings presents the challenge of balancing human and wildlife needs. Jollyville Plateau Salamanders (Eurycea tonkawae) are a range restricted, federally threatened, species of neotenic brook salamander endemic to central Texas. Almost the entire geographic range of E. tonkawae is embedded in the Austin, Cedar Park, and Round Rock metropolitan areas of Travis and Williamson counties, Texas. Among E. tonkawae occupied sites, Brushy Creek Spring has experienced some of the most extensive anthropogenic disturbance. Today the site consists of small groundwater outlets that emerge in the seams within a concrete culvert underlying a highway. Salamanders persist within this system though they are rarely detected. Here, we model the occurrence of salamanders within the surface habitat of Brushy Creek Spring using generalized linear models. In the absence of available data regarding the amount of water that is discharged from the spring, we use accumulated rainfall as a proxy for discharge to estimate salamander abundance. Additionally, we present evidence of reproduction, recruitment, and subterranean movement by E. tonkawae throughout this site. Infrastructure maintenance is inevitable at Brushy Creek Spring. We intend for our results to inform when maintenance should occur, i.e., during environmental conditions when salamanders are less likely to be observed in the surface habitat, to avoid unnecessary impacts to this federally threatened species.