Multiscale modeling of spring phenology across Deciduous Forests in the Eastern United States

Phenological events, such as bud burst, are strongly linked to ecosystem processes in temperate deciduous forests. However, the exact nature and magnitude of how seasonal and interannual variation in air temperatures influence phenology is poorly understood, and model‐based phenology representations fail to capture local‐ to regional‐scale variability arising from differences in species composition. In this paper, we use a combination of surface meteorological data, species composition maps, remote sensing, and ground‐based observations to estimate models that better represent how community‐level species composition affects the phenological response of deciduous broadleaf forests to climate forcing at spatial scales that are typically used in ecosystem models. Using time series of canopy greenness from repeat digital photography, citizen science data from the USA National Phenology Network, and satellite remote sensing‐based observations of phenology, we estimated and tested models that predict the timing of spring leaf emergence across five different deciduous broadleaf forest types in the eastern United States. Specifically, we evaluated two different approaches: (i) using species‐specific models in combination with species composition information to ‘upscale’ model predictions and (ii) using repeat digital photography of forest canopies that observe and integrate the phenological behavior of multiple representative species at each camera site to calibrate a single model for all deciduous broadleaf forests. Our results demonstrate variability in cumulative forcing requirements and photoperiod cues across species and forest types, and show how community composition influences phenological dynamics over large areas. At the same time, the response of different species to spatial and interannual variation in weather is, under the current climate regime, sufficiently similar that the generic deciduous forest model based on repeat digital photography performed comparably to the upscaled species‐specific models. More generally, results from this analysis demonstrate how in situ observation networks and remote sensing data can be used to synergistically calibrate and assess regional parameterizations of phenology in models.     

A method to estimate phenological variation using data from large‐scale abundance monitoring programmes

Capsule Large‐scale abundance monitoring programmes can be used to estimate annual phenological shifts.

Aims Phenology refers to the timing of any annually repeated biological event. The method developed here aims at measuring phenological variation in an indirect way by modelling seasonal abundance variations. Thus, it provides the opportunity to use a large number of datasets which have rarely been used in phenological studies. Phenological variations computed using this standardized method are comparable between species.

Methods The data used for the development of this method originates from the French Breeding Bird Survey, a large‐scale abundance monitoring programme launched in 2001. For each species, the phenological shift between two seasonal abundance trends is computed using maximum likelihood.

Results Phenological shifts relative to the year 2005 (reference year) were estimated for 46 species over a 5‐year period (2001–6). The standard deviations of the shifts do not differ significantly between species with different migratory status. Moreover, at the species level, the computed phenological shifts relate to the shifts of the mean date weighted by abundance. However, mean date, cannot be used in studies incorporating species with different migratory status (e.g. trans‐Saharan migrant, sedentary) because of ambiguous changes for the same biological shift in timing.

Conclusions The method described here is of particular value in determining how the phenology of common bird species changes in relation to climate. It offers the opportunity to increase the spatial scale of phenological studies and to include multi‐species analyses. This method could be applied to any abundance or constant effort site programme to study the timing of any biological process for which a seasonal distribution is available.     

Who flies first? – habitat‐specific phenological shifts of butterflies and orthopterans in the light of climate change: a case study from the south‐east Mediterranean

1. Insects undergo phenological change at different rates, showing no consistent trend between habitats, time periods, species or groups. Understanding how and why this variability occurs is crucial. 2. Phenological patterns of butterflies and Orthoptera were analysed using a novel approach of standardised major axis (SMA) analysis. It was investigated whether: (i) phenology (the mean date and duration of flight) of butterflies and Orthoptera changed from one survey (1998 and 1999 respectively) to another (2011), (ii) the rate at which phenology changed differed between taxa and (iii) phenological change was significantly different across habitat types (agriculture fields, grasslands, and forests). Using the 2011 dataset, we investigated relationships between habitat‐specific variables and species phenology. 3. For both groups, late‐emerging species had an advanced onset on the second survey while the duration showed no consistent trend for butterflies and did not change for Orthoptera. Although the rate at which phenology changed was consistent between the two groups, at the habitat level, a longer duration of flight period emerged for butterflies in agriculture fields while Orthoptera showed no differentiation in flight duration between habitats. We found an earlier emergence of butterflies in grasslands compared to forests, attributed to habitat‐specific temperature, whereas spatial variation in humidity had a significantly lower effect on butterflies' phenology in grasslands compared to forests. A gradual delay of butterfly appearances as the canopy cover increased was also found. 4. The utility of SMA analysis was demonstrated in phenological studies and evidence was detected that both habitat type and habitat‐specific variables refine species' phenological responses.     

Phenological indices of avian reproduction: cryptic shifts and prediction across large spatial and temporal scales

Climate change‐induced shifts in phenology have important demographic consequences, and are frequently used to assess species' sensitivity to climate change. Therefore, developing accurate phenological predictions is an important step in modeling species' responses to climate change. The ability of such phenological models to predict effects at larger spatial and temporal scales has rarely been assessed. It is also not clear whether the most frequently used phenological index, namely the average date of a phenological event across a population, adequately captures phenological shifts in the distribution of events across the season. We use the long‐tailed tit Aegithalos caudatus (Fig. 1) as a case study to explore these issues. We use an intensive 17‐year local study to model mean breeding date and test the capacity of this local model to predict phenology at larger spatial and temporal scales. We assess whether local models of breeding initiation, termination, and renesting reveal phenological shifts and responses to climate not detected by a standard phenological index, that is, population average lay date. These models take predation timing/intensity into account. The locally‐derived model performs well at predicting phenology at the national scale over several decades, at both high and low temperatures. In the local model, a trend toward warmer Aprils is associated with a significant advance in termination dates, probably in response to phenological shifts in food supply. This results in a 33% reduction in breeding season length over 17 years – a substantial loss of reproductive opportunity that is not detected by the index of population average lay date. We show that standard phenological indices can fail to detect patterns indicative of negative climatic effects, potentially biasing assessments of species' vulnerability to climate change. More positively, we demonstrate the potential of detailed local studies for developing broader‐scale predictive models of future phenological shifts.     

Regional unified model‐based leaf unfolding prediction from 1960 to 2009 across northern China

Using first leaf unfolding data of Salix matsudana, Populus simonii, Ulmus pumila, and Prunus armeniaca, and daily mean temperature data during the 1981–2005 period at 136 stations in northern China, we fitted unified forcing and chilling phenology models and selected optimum models for each species at each station. Then, we examined performances of each optimum local species‐specific model in predicting leaf unfolding dates at all external stations within the corresponding climate region and selected 16 local species‐specific models with maximum effective predictions as the regional unified models in different climate regions. Furthermore, we validated the regional unified models using leaf unfolding and daily mean temperature data beyond the time period of model fitting. Finally, we substituted gridded daily mean temperature data into the regional unified models, and reconstructed spatial patterns of leaf unfolding dates of the four tree species across northern China during 1960–2009. At local scales, the unified forcing model shows higher simulation efficiency at 83% of data sets, whereas the unified chilling model indicates higher simulation efficiency at 17% of data sets. Thus, winter temperature increase so far has not yet significantly influenced dormancy and consequent leaf development of deciduous trees in most parts of northern China. Spatial and temporal validation confirmed capability and reliability of regional unified species‐specific models in predicting leaf unfolding dates in northern China. Reconstructed leaf unfolding dates of the four tree species show significant advancements by 1.4–1.6 days per decade during 1960–2009 across northern China, which are stronger for the earlier than the later leaf unfolding species. Our findings suggest that the principal characteristics of plant phenology and phenological responses to climate change at regional scales can be captured by phenological and climatic data sets at a few representative locations.     

Terrestrial biosphere models need better representation of vegetation phenology: results from the North American Carbon Program Site Synthesis

Phenology, by controlling the seasonal activity of vegetation on the land surface, plays a fundamental role in regulating photosynthesis and other ecosystem processes, as well as competitive interactions and feedbacks to the climate system. We conducted an analysis to evaluate the representation of phenology, and the associated seasonality of ecosystem‐scale CO₂ exchange, in 14 models participating in the North American Carbon Program Site Synthesis. Model predictions were evaluated using long‐term measurements (emphasizing the period 2000–2006) from 10 forested sites within the AmeriFlux and Fluxnet‐Canada networks. In deciduous forests, almost all models consistently predicted that the growing season started earlier, and ended later, than was actually observed; biases of 2 weeks or more were typical. For these sites, most models were also unable to explain more than a small fraction of the observed interannual variability in phenological transition dates. Finally, for deciduous forests, misrepresentation of the seasonal cycle resulted in over‐prediction of gross ecosystem photosynthesis by +160 ± 145 g C m^?2 yr^?1 during the spring transition period and +75 ± 130 g C m^?2 yr^?1 during the autumn transition period (13% and 8% annual productivity, respectively) compensating for the tendency of most models to under‐predict the magnitude of peak summertime photosynthetic rates. Models did a better job of predicting the seasonality of CO₂ exchange for evergreen forests. These results highlight the need for improved understanding of the environmental controls on vegetation phenology and incorporation of this knowledge into better phenological models. Existing models are unlikely to predict future responses of phenology to climate change accurately and therefore will misrepresent the seasonality and interannual variability of key biosphere–atmosphere feedbacks and interactions in coupled global climate models.     

Optimizing biocontrol using phenological day degree models: the European earwig in pipfruit orchards

• 1 Phenological day degree models are often used as warning systems for the emergence of arthropod pests in agricultural crops or the occurrence of natural enemies of the pest species. In the present study, we report on a case study of the European earwig Forficula auricularia L., which is an important natural enemy in pipfruit orchards, and describe how such a day degree model can be used to avoid negative effects of crucial orchard management, such as spray applications and soil tillage. A precise timing of these interventions in relation to the phenology of natural enemies will enhance biocontrol.
• 2 Earwig population dynamics are characterized by single‐ and double‐brood populations, each with specific biological characteristics.
• 3 A day degree model capable of predicting the phenology of local earwig populations of both population types was developed. The model was checked for accuracy by comparing the first field observation dates of various life stages with predicted values using temperature data from the nearest weather station. In addition, variation in development time was assessed using field data.
• 4 The model was able to make predictions on a global scale. Although single‐ and double‐brood populations differ in phenology, the predictions of first appearance dates were similar. Variation in development time showed that single‐brood populations were more synchronized.
• 5 Our phenological model provides an accurate tool for predicting and simulating earwig population dynamics, as well as for enhancing the biocontrol of pests in pipfruit orchards.

     

Predicting insect phenology across space and time

Many species appear to be undergoing shifts in phenology, arising from climate change. To predict the direction and magnitude of future changes requires an understanding of how phenology depends on climatic variation. Species show large‐scale spatial variation in phenology (affected by differentiation among populations) as well as variation in phenology from year‐to‐year at the same site (affected predominantly by local plasticity). Teasing apart spatial and temporal variation in phenology should allow improved predictions of phenology under climate change. This study is the first to quantify large‐scale spatial and temporal variation in the entire emergence pattern of species, and to test the relationships found by predicting future data. We use data from up to 33 years of permanent transect records of butterflies in the United Kingdom to fit and test models for 15 butterfly species. We use generalized additive models to model spatial and temporal variation in the distribution of adult butterflies over the season, allowing us to capture changes in the timing of emergence peaks, relative sizes of peaks and/or number of peaks in a single analysis. We develop these models using data for 1973–2000, and then use them to predict phenologies from 2001 to 2006. For six of our study species, a model with only spatial variation in phenology is the best predictor of the future, implying that these species have limited plasticity. For the remaining nine species, the best predictions come from a model with both spatial and temporal variation in phenology; for four of these, growing degree‐days have similar effects over space and time, implying high levels of plasticity. The results show that statistical phenology models can be used to predict phenology shifts in a second time period, suggesting that it should be feasible to project phenologies under climate change scenarios, at least over modest time scales.     

Maintenance of temporal synchrony between syrphid flies and floral resources despite differential phenological responses to climate

Variation in species’ responses to abiotic phenological cues under climate change may cause changes in temporal overlap among interacting taxa, with potential demographic consequences. Here, we examine associations between the abiotic environment and plant–pollinator phenological synchrony using a long‐term syrphid fly–flowering phenology dataset (1992–2011). Degree‐days above freezing, precipitation, and timing of snow melt were investigated as predictors of phenology. Syrphids generally emerge after flowering onset and end their activity before the end of flowering. Neither flowering nor syrphid phenology has changed significantly over our 20‐year record, consistent with a lack of directional change in climate variables over the same time frame. Instead we document interannual variability in the abiotic environment and phenology. Timing of snow melt was the best predictor of flowering onset and syrphid emergence. Snow melt and degree‐days were the best predictors of the end of flowering, whereas degree‐days and precipitation best predicted the end of the syrphid period. Flowering advanced at a faster rate than syrphids in response to both advancing snow melt and increasing temperature. Different rates of phenological advancements resulted in more days of temporal overlap between the flower–syrphid community in years of early snow melt because of extended activity periods. Phenological synchrony at the community level is therefore likely to be maintained for some time, even under advancing snow melt conditions that are evident over longer term records at our site. These results show that interacting taxa may respond to different phenological cues and to the same cues at different rates but still maintain phenological synchrony over a range of abiotic conditions. However, our results also indicate that some individual plant species may overlap with the syrphid community for fewer days under continued climate change. This highlights the role of interannual variation in these flower–syrphid interactions and shows that species‐level responses can differ from community‐level responses in nonintuitive ways.     

Mismatch managed? Phenological phase extension as a strategy to manage phenological asynchrony in plant–animal mutualisms

Species‐specific shifts in phenology (timing of periodic life cycle events) are occurring with climate change and are already disrupting interactions within and among trophic levels. Phenological phase duration (e.g. beginning to end of flowering) and complementarity (patterns of nonoverlap), and their responses to changing conditions, will be important determinants of species' adaptive capacity to these shifts. Evidence indicates that extension of phenological duration of mutualistic partners could buffer negative impacts that occur with phenological shifts. Therefore, we suggest that techniques to extend the length of phenological duration will contribute to management of systems experiencing phenological asynchrony. Techniques of phenological phase extension discussed include the role of abiotic heterogeneity, genetic and species diversity, and alteration of population timing. We explore these approaches with the goal of creating a framework to build adaptive capacity and address phenological asynchrony in plant–animal mutualisms under climate change.     

Increasing the robustness of phenological models for Vitis vinifera cv. Chardonnay

Phenological models are important tools for planning viticultural practices in the short term and for projecting the impact of climate change on grapevine (Vitis vinifera) in the long term. However, the difficulties in obtaining phenological models which provide accurate predictions on a regional scale prevent them from being exploited to their full potential. The aim of this work was to obtain a robust phenological model for V. vinifera cv. Chardonnay. During calibration of the sub-models for budburst, flowering and veraison we implemented a series of measures to prevent overfitting and to give greater physiological meaning to the models. Among these were the use of experimental information on the response of Chardonnay to forcing temperatures, restriction of parameter space into physiologically meaningful limits prior to calibration, and simplification of the previously selected sub-models. The resulting process-based model had good internal validity and a good level of accuracy in predicting phenological events from external datasets. Model performance was especially high for the prediction of flowering and veraison, and comparison with other models confirmed it as a better predictor of phenology, even in extremely warm years. The modelling study highlighted a different phenological behaviour at the only mountain station, Cembra. We hypothesised that phenotypical plasticity could lead to growth rates adapting to a lower mean temperature, a mechanism not usually accounted for by phenological models.     

Alpine grassland plants grow earlier and faster but biomass remains unchanged over 35 years of climate change

Satellite data indicate significant advancement in alpine spring phenology over decades of climate warming, but corresponding field evidence is scarce. It is also unknown whether this advancement results from an earlier shift of phenological events, or enhancement of plant growth under unchanged phenological pattern. By analyzing a 35‐year dataset of seasonal biomass dynamics of a Tibetan alpine grassland, we show that climate change promoted both earlier phenology and faster growth, without changing annual biomass production. Biomass production increased in spring due to a warming‐induced earlier onset of plant growth, but decreased in autumn due mainly to increased water stress. Plants grew faster but the fast‐growing period shortened during the mid‐growing season. These findings provide the first in situ evidence of long‐term changes in growth patterns in alpine grassland plant communities, and suggest that earlier phenology and faster growth will jointly contribute to plant growth in a warming climate.     

Climatic effects on the phenology of lake processes

Populations living in seasonal environments are exposed to systematic changes in physical conditions that restrict the growth and reproduction of many species to only a short time window of the annual cycle. Several studies have shown that climate changes over the latter part of the 20th century affected the phenology and population dynamics of single species. However, the key limitation to forecasting the effects of changing climate on ecosystems lies in understanding how it will affect interactions among species. We investigated the effects of climatic and biotic drivers on physical and biological lake processes, using a historical dataset of 40 years from Lake Washington, USA, and dynamic time‐series models to explain changes in the phenological patterns among physical and biological components of pelagic ecosystems. Long‐term climate warming and variability because of large‐scale climatic patterns like Pacific decadal oscillation (PDO) and El Niño–southern oscillation (ENSO) extended the duration of the stratification period by 25 days over the last 40 years. This change was due mainly to earlier spring stratification (16 days) and less to later stratification termination in fall (9 days). The phytoplankton spring bloom advanced roughly in parallel to stratification onset and in 2002 it occurred about 19 days earlier than it did in 1962, indicating the tight connection of spring phytoplankton growth to turbulent conditions. In contrast, the timing of the clear‐water phase showed high variability and was mainly driven by biotic factors. Among the zooplankton species, the timing of spring peaks in the rotifer Keratella advanced strongly, whereas Leptodiaptomus and Daphnia showed slight or no changes. These changes have generated a growing time lag between the spring phytoplankton peak and zooplankton peak, which can be especially critical for the cladoceran Daphnia. Water temperature, PDO, and food availability affected the timing of the spring peak in zooplankton. Overall, the impact of PDO on the phenological processes were stronger compared with ENSO. Our results highlight that climate affects physical and biological processes differently, which can interrupt energy flow among trophic levels, making ecosystem responses to climate change difficult to forecast.     

Chilling and forcing temperatures interact to predict the onset of wood formation in Northern Hemisphere conifers

The phenology of wood formation is a critical process to consider for predicting how trees from the temperate and boreal zones may react to climate change. Compared to leaf phenology, however, the determinism of wood phenology is still poorly known. Here, we compared for the first time three alternative ecophysiological model classes (threshold models, heat‐sum models and chilling‐influenced heat‐sum models) and an empirical model in their ability to predict the starting date of xylem cell enlargement in spring, for four major Northern Hemisphere conifers (Larix decidua, Pinus sylvestris, Picea abies and Picea mariana). We fitted models with Bayesian inference to wood phenological data collected for 220 site‐years over Europe and Canada. The chilling‐influenced heat‐sum model received most support for all the four studied species, predicting validation data with a 7.7‐day error, which is within one day of the observed data resolution. We conclude that both chilling and forcing temperatures determine the onset of wood formation in Northern Hemisphere conifers. Importantly, the chilling‐influenced heat‐sum model showed virtually no spatial bias whichever the species, despite the large environmental gradients considered. This suggests that the spring onset of wood formation is far less affected by local adaptation than by environmentally driven plasticity. In a context of climate change, we therefore expect rising winter–spring temperature to exert ambivalent effects on the spring onset of wood formation, tending to hasten it through the accumulation of forcing temperature, but imposing a higher forcing temperature requirement through the lower accumulation of chilling.     

A new seasonal‐deciduous spring phenology submodel in the Community Land Model 4.5: impacts on carbon and water cycling under future climate scenarios

A spring phenology model that combines photoperiod with accumulated heating and chilling to predict spring leaf‐out dates is optimized using PhenoCam observations and coupled into the Community Land Model (CLM) 4.5. In head‐to‐head comparison (using satellite data from 2003 to 2013 for validation) for model grid cells over the Northern Hemisphere deciduous broadleaf forests (5.5 million km²), we found that the revised model substantially outperformed the standard CLM seasonal‐deciduous spring phenology submodel at both coarse (0.9 × 1.25°) and fine (1 km) scales. The revised model also does a better job of representing recent (decadal) phenological trends observed globally by MODIS, as well as long‐term trends (1950–2014) in the PEP725 European phenology dataset. Moreover, forward model runs suggested a stronger advancement (up to 11 days) of spring leaf‐out by the end of the 21st century for the revised model. Trends toward earlier advancement are predicted for deciduous forests across the whole Northern Hemisphere boreal and temperate deciduous forest region for the revised model, whereas the standard model predicts earlier leaf‐out in colder regions, but later leaf‐out in warmer regions, and no trend globally. The earlier spring leaf‐out predicted by the revised model resulted in enhanced gross primary production (up to 0.6 Pg C yr^?1) and evapotranspiration (up to 24 mm yr^?1) when results were integrated across the study region. These results suggest that the standard seasonal‐deciduous submodel in CLM should be reconsidered, otherwise substantial errors in predictions of key land–atmosphere interactions and feedbacks may result.     

Phenological modularity in amphibian calling behaviour: Geographic trends and local determinants

Phenology of species, the coupling of vital activities to specific times of the year, plays a main role in ecosystem functioning and is expected to be affected by global change. We analysed the temporal structure of 52 amphibian communities in South America encompassing a latitudinal range from 7º to 34º south. Phenological modularity – species tendencies to aggregate along the months – is here introduced as a ubiquitous property of biodiversity architecture. Further, we identified an increase in phenological modularity with species richness, available energy and in communities with lower thermal dependence (i.e. the rate of change in the number of species active along the year associated with the environmental temperature). These patterns are in agreement with predictions derived from several ecological hypotheses: complexity‐stability, species‐energy and metabolic ecology. However, no direct association between modularity and the phylogenetic structure of communities was observed. A structural equation model that outperformed all the plausible alternative models considered supports these results. Modularity is reported here as a main feature of the phenology of communities that depends on environmental conditions. Here, we report for the first time a putative connection between community species richness and the degree of temporal structure – phenological modularity; the thermal dependence shows that communities at low latitudes are more vulnerable to climate change; energetic environments also promote communities with phenological modularity; and latitudinal patterns of phylogenetic community structure can give us clues of which species would be important to the conservation of community processes. These results call for further theoretical analyses to support the connection between phenological modularity, community stability and vulnerability to global change.     

Mismatches between breeding phenology and resource abundance of resident alpine ptarmigan negatively affect chick survival

相似文献   

Predicting the sensitivity of butterfly phenology to temperature over the past century

Studies to date have documented substantial variation among species in the degree to which phenology responds to temperature and shifts over time, but we have a limited understanding of the causes of such variation. Here, we use a spatially and temporally extensive data set (ca. 48 000 observations from across Canada) to evaluate the utility of museum collection records in detecting broad‐scale phenology‐temperature relationships and to test for systematic differences in the sensitivity of phenology to temperature (days °C^?1) of Canadian butterfly species according to relevant ecological traits. We showed that the timing of flight season predictably responded to temperature both across space (variation in average temperature from site to site in Canada) and across time (variation from year to year within each individual site). This reveals that collection records, a vastly underexploited resource, can be applied to the quantification of broad‐scale relationships between species' phenology and temperature. The timing of the flight season of earlier fliers and less mobile species was more sensitive to temperature than later fliers and more mobile species, demonstrating that ecological traits can account for some of the interspecific variation in species' phenological sensitivity to temperature. Finally, we found that phenological sensitivity to temperature differed across time and space implying that both dimensions of temperature will be needed to translate species' phenological sensitivity to temperature into accurate predictions of species' future phenological shifts. Given the widespread temperature sensitivity of flight season timing, we can expect long‐term temporal shifts with increased warming [ca. 2.4 days °C^?1 (0.18 SE)] for many if not most butterfly species.     

An empirical comparison of models for the phenology of bird migration

Bird migration phenology shows strong responses to climate change. Studies of trends and patterns in phenology are typically based on annual summarizing metrics, such as means and quantiles calculated from raw daily count data. However, with irregularly sampled data and large day‐to‐day variation, such metrics can be biased and noisy, and may be analysed using phenological functions fitted to the data. Here we use count data of migration passage from a Finnish bird observatory to compare different models for the phenological distributions of spring migration (27 species) and autumn migration (57 species). We assess parsimony and goodness‐of‐fit in a set of models, with phenological functions of different complexity, optionally with covariates accounting for day‐to‐day variability. The covariates describe migration intensities of related species or relative migration intensities the previous day (autocovariates). We found that parametric models are often preferred over the more flexible generalized additive models with constrained degrees of freedom. Models corresponding to a mixture of two distinct passing populations were frequently preferred over simpler ones, but usually no more complex models are needed. Slightly more complex models were favoured in spring compared to autumn. Related species’ migration activity effectively improves the model by accounting for the large day‐to‐day variation. Autocovariates were usually not that relevant, implying that autocorrelation is generally not a major concern if phenology is modelled properly. We suggest that parametric models are relatively good for studying single‐population migration phenology, or a mix of two groups with distinct phenologies, especially if daily variation in migration intensity can be controlled for. Generalized additive models may be useful when the migrating population composition is unknown. Despite these guidelines, choosing an appropriate model involves case‐by‐case assessment or the biological relevance and rationale for modelling phenology.