Summary1. Threats to migratory animals can occur at multiple periods of the annual cycle that are separated by thousands of kilometres and span international borders. Populations of the iconic monarch butterfly (Danaus plexippus) of eastern North America have declined over the last 21 years. Three hypotheses have been posed to explain the decline: habitat loss on the overwintering grounds in Mexico, habitat loss on the breeding grounds in the United States and Canada, and extreme weather events. 2. Our objectives were to assess population viability, determine which life stage, season and geographical region are contributing the most to population dynamics and test the three hypotheses that explain the observed population decline. 3. We developed a spatially structured, stochastic and density-dependent periodic projection matrix model that integrates patterns of migratory connectivity and demographic vital rates across the annual cycle. We used perturbation analysis to determine the sensitivity of population abundance to changes in vital rate among life stages, seasons and geographical regions. Next, we compared the singular effects of each threat to the full model where all factors operate concurrently. Finally, we generated predictions to assess the risk of host plant loss as a result of genetically modified crops on current and future monarch butterfly population size and extinction probability. 4. Our year-round population model predicted population declines of 14% and a quasiextinction probability (<1000 individuals) >5% within a century. Monarch abundance was more than four times more sensitive to perturbations of vital rates on the breeding grounds than on the wintering grounds. Simulations that considered only forest loss or climate change in Mexico predicted higher population sizes compared to milkweed declines on the breeding grounds. Our model predictions also suggest that mitigating the negative effects of genetically modified crops results in higher population size and lower extinction risk. 5. Recent population declines stem from reduction in milkweed host plants in the United States that arise from increasing adoption of genetically modified crops and land-use change, not from climate change or degradation of forest habitats in Mexico. Therefore, reducing the negative effects of host plant loss on the breeding grounds is the top conservation priority to slow or halt future population declines of monarch butterflies in North America.

Insect migration may involve movements over multiple breeding generations at continental scales, resulting in formidable challenges to their conservation and management. Using distribution models generated from citizen scientist occurrence data and stable-carbon and -hydrogen isotope measurements, we tracked multi-generational colonization of the breeding grounds of monarch butterflies (Danaus plexippus) in eastern North America. We found that monarch breeding occurrence was best modelled with geographical and climatic variables resulting in an annual breeding distribution of greater than 12 million km 2 that encompassed 99% occurrence probability. Combining occurrence models with stable isotope measurements to estimate natal origin, we show that butterflies which overwintered in Mexico came from a wide breeding distribution, including southern portions of the range. There was a clear northward progression of monarchs over successive generations from May until August when reproductive butterflies began to change direction and moved south. Fifth-generation individuals breeding in Texas in the late summer/autumn tended to originate from northern breeding areas rather than regions further south. Although the Midwest was the most productive area during the breeding season, monarchs that re-colonized the Midwest were produced largely in Texas, suggesting that conserving breeding habitat in the Midwest alone is insufficient to ensure long-term persistence of the monarch butterfly population in eastern North America.

Addressing population declines of migratory insects requires linking populations across different portions of the annual cycle and understanding the effects of variation in weather and climate on productivity, recruitment, and patterns of long-distance movement. We used stable H and C isotopes and geospatial modeling to estimate the natal origin of monarch butterflies (Danaus plexippus) in eastern North America using over 1000 monarchs collected over almost four decades at Mexican overwintering colonies. Multinomial regression was used to ascertain which climate-related factors best-predicted temporal variation in natal origin across six breeding regions. The region producing the largest proportion of overwintering monarchs was the US Midwest (mean annual proportion = 0.38; 95% CI: 0.36-0.41) followed by the north-central (0.17; 0.14-0.18), northeast (0.15; 0.11-0.16), northwest (0.12; 0.12-0.16), southwest (0.11; 0.08-0.12), and southeast (0.08; 0.07-0.11) regions. There was no evidence of directional shifts in the relative contributions of different natal regions over time, which suggests these regions are comprising the same relative proportion of the overwintering population in recent years as in the mid-1970s. Instead, interannual variation in the proportion of monarchs from each region covaried with climate, as measured by the Southern Oscillation Index and regional-specific daily maximum temperature and precipitation, which together likely dictate larval development rates and food plant condition. Our results provide the first robust long-term analysis of predictors of the natal origins of monarchs overwintering in Mexico. Conservation efforts on the breeding grounds focused on the Midwest region will likely have the greatest benefit to eastern North American migratory monarchs, but the population will likely remain sensitive to regional and stochastic weather patterns.

From butterflies to elephants, the rapidly developing science of movement ecology is providing increasingly detailed spatio-temporal data on a wide array of mobile animals. Thus, this discipline also holds great promise for improving the conservation of wildlife. To measure progress toward this promise, we investigated the degree to which movement ecology research is connected to conservation goals as well as the proportion of studies that were incorporated into federal and international status assessments for mobile species at risk. We examined 13,349 "movement ecology" papers published between 1990 and 2014 and found that explicit connections to conservation and management were made in 35% (n = 4, 672) of these papers, with the number of connections increasing over time. We then measured the uptake of movement ecology research into species status assessment and recovery plans (n = 72 documents) produced by three different governance agencies for 12 endangered mobile species. We found that on average 60% of available movement ecology research was used in the status assessment process, demonstrating that when movement ecology research is available, it is generally being utilized in conservation planning. However, for 25% of these species, there was little movement research available to be used, highlighting that knowledge gaps remain for some at-risk species despite the general growth of movement ecology research. We outline opportunities for movement ecology to promote more effective conservation of taxa that move.

Strontium isotopes ((87)Sr/(86)Sr) can be useful biological markers for a wide range of forensic science applications, including wildlife tracking. However, one of the main advantages of using (87)Sr/(86)Sr values, that there is no fractionation from geological bedrock sources through the food web, also happens to be a critical assumption that has never been tested experimentally. We test this assumption by measuring (87)Sr/(86)Sr values across three trophic levels in a controlled greenhouse experiment. Adult monarch butterflies were raised on obligate larval host milkweed plants that were, in turn, grown on seven different soil types collected across Canada. We found no significant differences between (87)Sr/(86)Sr values in leachable Sr from soil minerals, organic soil, milkweed leaves, and monarch butterfly wings. Our results suggest that strontium isoscapes developed from (87)Sr/(86)Sr values in bedrock or soil may serve as a reliable biological marker in forensic science for a range of taxa and across large geographic areas.