Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Integrating Meteorology into Research on Migration

Integrating Meteorology into Research on Migration Integrative and Comparative Biology, volume 50, number 3, pp. 280–292 doi:10.1093/icb/icq011 SYMPOSIUM Judy Shamoun-Baranes, Willem Bouten and E. Emiel van Loon Computational Geo-Ecology, Institute for Biodiversity and Ecosystem Dynamics, University of Amsterdam, P.O. Box 94248, 1090 GE Amsterdam, The Netherlands E-mail: shamoun@uva.nl From the symposium ‘‘Integrative Migration Biology’’ presented at the annual meeting of the Society for Integrative and Comparative Biology, January 3–7, 2010 at Seattle, Washington. Synopsis Atmospheric dynamics strongly influence the migration of flying organisms. They affect, among others, the onset, duration and cost of migration, migratory routes, stop-over decisions, and flight speeds en-route. Animals move through a heterogeneous environment and have to react to atmospheric dynamics at different spatial and temporal scales. Integrating meteorology into research on migration is not only challenging but it is also important, especially when trying to understand the variability of the various aspects of migratory behavior observed in nature. In this article, we give an overview of some different modeling approaches and we show how these have been incorporated into migration research. We provide a more detailed description of the development and application of two dynamic, individual-based models, one for waders and one for soaring migrants, as examples of how and why to integrate meteorology into research on migration. We use these models to help understand underlying mechanisms of individual response to atmospheric conditions en-route and to explain emergent patterns. This type of models can be used to study the impact of variability in atmospheric dynamics on migration along a migratory trajectory, between seasons and between years. We conclude by providing some basic guidelines to help researchers towards finding the right modeling approach and the meteorological data needed to integrate meteorology into their own research. resulting in carry-over effects such as an impact on Introduction survival (e.g. Erni et al. 2005; Newton 2006), timing For flying organisms, such as insects, bats and birds, of migration or breeding success. Although global atmospheric dynamics play an important role in patterns of atmospheric circulation may not affect their migratory movements (e.g. Richardson 1990; instantaneous responses of individual migrants Dingle 1996; Liechti 2006; Kunz et al. 2008). directly, they will shape atmospheric conditions at Animals move through a dynamic and heterogeneous smaller scales. At the same time they are likely to environment where conditions from the microscale have more long term and cumulative effects on through the mesoscale and even global circulation migration by affecting migration routes and seasonal patterns are relevant (Drake and Farrow 1988; timing of long distance movements. We provide a Nathan et al. 2005; Kunz et al. 2008). The effects simplified representation of these multi-scale interac- of atmospheric dynamics on migration are complex tions in Fig. 1. and may differ depending on the temporal and spa- Meteorology should be integrated into research on tial scale being considered as well as on the species, migration, especially when trying to understand nat- region, season and year. For example, instantaneous ural variability observed in aspects like the timing responses to changing wind speeds due to microscale of migration; migratory routes; orientation; use of and/or mesoscale dynamics may affect the flight stopover sites; or population trends such as speed, course and energetic cost at that point in effects on survival or breeding success as a result of time, as well as the conditions that will be experi- changes in arrival time or physiological condition. enced later en-route (e.g. Chapman et al. 2010; Meteorology is also of interest at longer time Shamoun-Baranes et al. 2010). These effects may scales when trying to understand the evolution of also be cumulative throughout the season, finally Advanced Access publication April 8, 2010 This is an Open Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (http://creativecommons.org/licenses/ by-nc/2.5/) which permits unrestricted non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited. The Author 2010. Published by Oxford University Press on behalf of the Society for Integrative and Comparative Biology. All rights reserved. For permissions please email: journals.permissions@oxfordjournals.org. Integrating meteorology and migration research 281 Fig. 1 A simplified representation of the different spatio-temporal scales of atmospheric dynamics that may influence instantaneous behavioral responses, resulting in short-term (instantaneous) and longer-term effects. Carry-over effects include not only inter-annual effects (e.g. population size or breeding success), but longer-term effects that may have evolutionary consequences (e.g. shaping migration routes). For example, instantaneous changes in flight behavior would influence instantaneous flight speed, the timing of migration within a migration season and could also lead to carry-over effects such as the timing of breeding or breeding success. particular migratory systems. Thus, integrating mete- Kemp et al. in review), flight altitudes (Bruderer orology and research on animal migration will et al. 1995; Wood et al. 2006, 2010; Reynolds et al. help us better understand both short- and long-term 2009; Schmaljohann et al. 2009), flight strategy (Gibo organismal–environmental linkages, one of the grand and Pallett 1979; Pennycuick et al. 1979; Gibo 1981; Spaar and Bruderer 1997; Spaar et al. 1998; Sapir challenges identified in organismal biology (Schwenk 2009), orientation and trajectories (Thorup et al. et al. 2009). However, linking mechanisms at the 2003; Chapman et al. 2008; Srygley and Dudley individual level to these longer-term, or larger-scale 2008; Chapman et al. 2010), migration intensity or consequences remains challenging. probability (Erni et al. 2002; Reynolds 2006; Cryan During the past few decades numerous empirical and Brown 2007; Stefanescu et al. 2007; van Belle and theoretical studies have addressed the influence et al. 2007; Leskinen et al. 2009), as well as migratory of atmospheric dynamics on animals’ migrations success (Erni et al. 2005; Reilly and Reilly 2009). (for overviews see Richardson 1978, 1990; Drake and Farrow 1988; Dingle 1996; Liechti 2006; Modeling approaches Newton 2008). Atmospheric conditions are known to influence the onset of migration (Shamoun- Different modeling techniques have been used to Baranes et al. 2006; Gill et al. 2009), migration study the influence of atmospheric dynamics on phenology (Hu ¨ ppop and Hu ¨ ppop 2003; Jonzen migration. We would like to distinguish between et al. 2006; Bauer et al. 2008), stopover decisions ‘concept-driven’ and ‘data-driven’ models. The struc- (Akesson and Hedenstro ¨ m 2000; Da ¨nhardt and ture of concept-driven models can only be conceived Lindstro ¨ m 2001; Schaub et al. 2004; Wikelski et al. if extensive prior knowledge of the system is available 2006; Brattstrom et al. 2008), flight speeds (Garland and cannot be discovered in an automated fashion. and Davis 2002; Shamoun-Baranes et al. 2003a; Calibration of parameters of concept-driven models 282 J. Shamoun-Baranes et al. Table 1 An overview of different types of concept- and data-driven models and their characteristics Requirement for creating the model Conceptual Numerical and Observations understanding data processing on state Possibilities for Frequency Name Description of the system skills variables calibration of use Concept-driven SC Static Concept-based model Intermediate Low Few Easy, many methods Intermediate DI Dynamic IBM Intermediate Intermediate Intermediate Difficult, few methods Intermediate DC Dynamic Continuum-based High High Intermediate Intermediate, few methods Low model Data-driven SD Static Data-based model Low Low Intermediate Easy, many methods High DD Dynamic Data-based model Low High Many Intermediate, few methods Low Frequency of use in migration studies is provided in the last column; for references to specific studies see Table 2. In this context, static means that the process being studied is either in steady state or that there is no influence of previous states on the current state. Individual-based: model state variables refer to properties of an individual; continuum based: model state variables refer to population properties. State variables are model-entities which are updated at each model time step with a difference equation in dynamic models and are usually comparable to the dependent variables in static models. is possible if measurements of model output are questions about system feedbacks and scale, compar- available, but this is not always required to use the ing theories and observations, and identifying ave- model. Reasonable values of parameters can often be nues for new research. By modeling the influence found by using expert knowledge or information of atmospheric conditions en-route we can study from experiments or biophysical calculations. emergent patterns of migration at the individual Concept driven models can then be run as thought and population level as well as study the importance experiments, without any observations. of the variability in individual behavior and the vari- The structure of data-driven models, however, can ability in atmospheric conditions, between days, sea- be derived via (highly) automated procedures but sons, years or regions. The models are tools to better does not necessarily represent cause and effect rela- understand underlying mechanisms, but not goals in tions in nature. These models need relatively little themselves. In such models, measurements gathered prior knowledge of the system before they can be from field research or from laboratory experiments constructed, but always require calibration of the can either be implicitly integrated into the models to parameters because the parameters do not necessarily formulate model assumptions and to parameterize match physical entities that can be independently models, or explicitly to compare to model results. observed in nature. Because of the necessity to cali- Furthermore, atmospheric conditions from observa- brate, data-driven models always need observations tions, reanalysis data, numerical models, or artificial on both input and output variables of the model. data (see Table 3 for examples) are needed as input Table 1 specifies three different types of to the models. concept-driven modeling techniques and two differ- In this article we describe the development and ent types of data-driven modeling techniques with application of two studies using dynamic models of some of their main characteristics. These definitions migration (‘concept-driven’, dynamic individual- are of course not specific for models of migration, based models; DI, Table 1) as examples of how but are applicable to ecological models in general. and why to integrate meteorology into research on Table 2 provides several examples of the various migration. We use the models to better understand techniques of measurement and modeling applied underlying mechanisms at the individual level to in studying the influence of atmospheric dynamics help explain the patterns that emerge at the popula- on migration. tion level. We provide some basic guidelines to help We think that especially concept-driven dynamic researchers towards integrating meteorology into simulations of migration provide a suitable frame- their research on migration and we discuss the work for integrating scattered knowledge about opportunities and limitations of different sources of migration, systematically addressing complex data that can be used for such studies. We thus hope Integrating meteorology and migration research 283 Table 2 A selection of studies on the influence of atmospheric conditions on animal migration, including focal species or group, types of data used, geographic region of study, type of model, and relevant reference Effects on migration Species/group Migration data Meteorological variable: data source Geographic region Model type References Flight behavior: altitude Nocturnal migratory Tracking radar Wind : radiosonde NCEP reanalysis data Sahara SC Schmaljohann birds et al. 2009 Flight behavior: altitude Soaring avian Motorized glider Boundary layer height and vertical lift: Israel SD Shamoun-Baranes migrants boundary layer convective model et al. 2003b,c Flight behavior: altitude Nocturnal migratory Radar Various: numerical weather prediction UK SD Wood et al. 2010 insects model, the Unified Model Take off decisions Arctic geese Ringing data Onset of spring proxy: NDVI Palearctic flyway DI Bauer et al. 2008 Take off decisions Bar tailed godwit Satellite telemetry Sea level pressure, wind : NCEP reanalysis Pacific ocean flyway Descriptive no Gill et al. 2009 data GEOS-5 global atmospheric model computer model Take off decisions Not relevant None Wind assistance or no assistance: no data Not relevant SC Weber et al. 1998 Take off decisions Green darners Radio telemetry Wind and temperature: weather station Northeast USA SD Wikelski et al. observations 2006 d a Migration intensity Nocturnal migratory Radar Wind , barometric pressure, temperature, The Netherlands SD van Belle et al. birds precipitation: weather station 2007 observations d a Migration intensity Nocturnal passerine Radar and visual Wind , temperature, synoptic weather Southeastern USA SD Able 1973 migration observations index: weather station observations Migration intensity Black-cherry aphids Radar and insect Wind : HIRLAM and ECMWF numerical Finland DC Leskinen et al. and diamond-back traps weather prediction models 2009 moths Speed Turkey vulture Satellite telemetry Wind speed, turbulent kinetic energy, cloud Eastern North DD Mandel et al. height: North American regional reana- American flyway 2008 lysis data Speed Red knots Visual Wind : NCEP reanalysis data Afro–Siberian flyway DI Shamoun-Baranes observations et al. 2010 Direction/Orientation Reed warbler Radio telemetry Wind : weather station observations Sweden SD Akesson et al. Direction/Orientation Moths and butterflies Radar Wind : numerical weather prediction UK SD Chapman et al. model, the Unified Model 2010 Timing Soaring avian Visual observa- Barometric pressure, temperature, precipi- Western Palearctic SC Shamoun-Baranes migrants tions, Satellite table water: NCEP reanalysis data (eastern) flyway et al. 2006 telemetry Timing Passerine migrants Ringing data North Atlantic Oscillation Europe and SD Jonzen et al. 2006 Scandinavia (continued) 284 J. Shamoun-Baranes et al. to facilitate further integration of field biology, mete- orology and modeling. Case study of the migration of red knots: The importance of wind From numerous studies over the years it is clear that of all atmospheric conditions wind plays the most important role in avian migration (Akesson and Hedenstro ¨ m 2000; Akesson et al. 2002; Shamoun- Baranes et al. 2003a; van Belle et al. 2007; Schmaljohann et al. 2009; Kemp et al. in review; for review see Liechti 2006). Yet, quantifying the cumulative effect of wind en-route for an entire migration trajectory, as well as the effect of variabil- ity in space and time has rarely been done in avian research (Stoddard et al. 1983; Erni et al. 2005; Vrugt et al. 2007; Reilly and Reilly 2009). The Afro–Siberian red knot (Calidris canutus canutus) is an intensively studied long-distance migrant (e.g. Piersma et al. 1992; Piersma and Lindstrom 1997; van de Kam et al. 2004). These birds migrate north in two non-stop flights of approximately 4400 km each from their wintering grounds in Mauritania via the German Wadden Sea to the Siberian breeding grounds in only four weeks (Piersma et al. 1992; van de Kam et al. 2004). From previous studies, favorable winds were considered essential along this flyway (Piersma and van de Sant 1992). Field observations have shown that red knots erratically use an extra stopover site on the French Atlantic coast (Leyrer et al. 2009). One hypothesis to explain this phenomenon was that birds experiencing unfavorable winds would use this area as an emergency stopover site. Therefore, a dynamic individual-based model (IBM) of north- bound migration, incorporating winds experienced en-route, was developed to study whether use of this intermediate stopover site could be explained by stochastic wind conditions (Shamoun-Baranes et al. 2010). In the model, birds are moved forward in 6-h time steps along the great-circle route between their wintering site in Mauritania and their stopover site at the Wadden Sea. The wind experienced at the beginning of each time step determines the ground speed of the bird which is subsequently used to cal- culate flight times and the birds’ locations at the next time step (Fig. 2). Data on speed and direction of wind at four different levels of pressure were used in this study to represent wind conditions experienced at different altitudes during flight. The data were extracted from the global NCEP-Reanalysis dataset which has a spatial resolution of 2.58 2.58 and a 6-h temporal resolution (Kalnay et al. 1996). The Table 2 Continued Effects on migration Species/group Migration data Meteorological variable: data source Geographic region Model type References Arrival mass Western sandpipers Biometric Wind : weather station observations North American DI Butler et al. 1997 measurements Pacific Coast Route Silver Y (noctuid Radar Wind : numerical weather prediction United Kingdom, DI Chapman et al. moth) model, the Unified Model Northwest Europe 2010 Route Golden Eagles Visual Wind (implicit): digital elevation model Central Pennsylvania DC Brandes and observations Ombalski 2004 Survival Simulated nocturnal Literature Wind : NCEP reanalysis data Western Palearctic DI Erni et al. 2005 passerine migrant migration Mass, population dynamics Houbara bustard Literature Winter severity (implicit): no data Not directly relevant DC Sto ¨ cker and Stonechat Weihs 1998 Type of model is described in more detail in Table 1. The terms used in the column entitled ‘Effects on migration’ are adapted to roughly follow the framework provided in Fig. 1 for comparative purposes and, thus, do not always follow the exact terms used in the original study. Similarly, although many effects may be studied with one model, we generally highlight the effect that was the focus of the study. For suggestions on where to find different sources of meteorological data, some of which are mentioned in this table, see Table 3. Wind speed and direction. White stork, honey buzzard, lesser spotted eagle. NDVI, normalized difference vegetation index. Migration intensity can also be considered a proxy for takeoff decisions. Integrating meteorology and migration research 285 Fig. 2 A forward simulation of the migration of red knots taking off on May 1, 1986. Upward-pointing and downward-pointing triangles indicate wintering site (simulated start location) and Wadden sea stopover site (simulated end location) respectively. The open circle marks the location of the emergency stopover site on the French Atlantic coast. Black circles indicate location at each time step. Arrows indicate the speed and direction of the wind at each location and the dotted line shows the flight trajectory. The shorter the distance between circles, the slower is the ground speed due to disadvantageous winds. advantages of the NCEP reanalysis dataset for this conducted with appropriate modeling techniques type of analysis are the homogenous spatial and tem- and field measurements. One of the main advantages of such a modeling poral coverage, global and long term coverage and free availability on the internet. After running simu- framework is the ability to also explore the effect of lations, results from the migration model can be variability of wind conditions across years, starting compared to observations. dates, and altitudes of flight, all resulting in different A comparison between simulated flight times and wind conditions and flight times. Furthermore, the the number of birds observed stopping over at the potential effects of spatial and temporal auto- French staging site showed how unpredictable winds correlation in wind conditions could be studied affect flight times and that wind is a predominant along the migratory route. For example, wind con- driver of the use of an emergency stopover site along ditions experienced in France were spatially and tem- the French Atlantic coast. Wind clearly plays an porally auto-correlated for418 h and sometimes over important and quantifiable role in this, and probably hundreds of kilometers. Spatio-temporal correlation many other, migratory systems. The study also indi- in atmospheric dynamics could provide migrants cates the importance of this emergency stopover site with information on what to expect further along for conservation. Although the Wadden Sea is an their trajectory and may enable them to fine-tune obligatory staging area in this system, the French their decisions based on their physiological state, Atlantic coast may be essential for ensuring the sur- geographic location, and immediate and expected vival of individuals that have encountered very unfa- environmental conditions. vorable winds en-route. Thus, in the long term, this In the future, the model can be extended by inte- emergency stopover site may influence, and help sta- grating the energetics of flight to model the expen- bilize, migratory population dynamics. These ideas, diture of energy due to the wind conditions however, require further research that can be experienced en-route. These results then can be 286 J. Shamoun-Baranes et al. compared to field measurements. The model is not only designed to enable further extensions but can be applied to other species where sufficient data are available. The model framework presented here is quite similar to meteorological trajectory analysis used to estimate the flight paths of migrating insects (e.g. Scott and Achtemeier 1987; Chapman et al. 2010) as a result of meteorological dynamics experi- enced en-route. Although in contrast to migrating insects, the location of the departure and destination are known for the red knots. Case study of migration by white storks: The importance of thermal convection Atmospheric dynamics strongly influence the migra- Fig. 3 3D trajectories of simulated migration of white storks. Thermals are indicated as grey cylinders; the destination is indi- tion of soaring birds; particularly thermal convection cated by a gray box. Each trajectory represents the movement of which influences daily flight schedules and migration an individual during the simulation. When in a thermal, birds routes (Kerlinger 1989; Leshem and Yom-Tov 1998), climb vertically until they reach the top; they then glide (losing flight speeds (Leshem and Yom-Tov 1996a; altitude) towards the next thermal if it can be sensed and Mandel et al. 2008) and flight altitudes (Leshem reached by the birds. Otherwise the bird glides first to the and Yom-Tov 1996a; Shannon et al. 2002a, 2002b; destination, until another thermal can be utilized. In this Shamoun-Baranes et al. 2003b, 2003c). Many soaring simulation, a bird first searches for the most distant thermal containing other birds. species are known to migrate in large single-species or mixed-species flocks aggregating in both space and time (e.g. Kerlinger 1989; Leshem and Yom-Tov 1996a, 1996b). Large-scale aggregations and select thermals based on several physical con- along well established and narrow migratory corri- straints and on pre-determined behavioral rules dors are often attributed to natural leading lines like which differentiate between thermals with and with- the Appalachian Mountains or to circumvention of out birds (Fig. 3). At each time step birds are either large bodies of water, resulting in geographic bottle- climbing in a thermal or gliding to their current necks such as those seen in Panama, Gibraltar, the target (a thermal or their destination). This design Bosporus and Israel (Kerlinger 1989; Leshem and of the model provides a framework for virtual exper- YomTov 1996b; Bildstein and Zalles 2005; Bildstein iments which can be used to explore the patterns 2006). However, the mechanisms that result in small that emerge due to different decision rules, flight scale convergence and the potential benefits of flock- parameters, convective conditions, or takeoff and ing remain largely unknown. With our model, which destination areas, and to test different scenarios. is described below, we explore the hypothesis that The study shows that under the convective condi- flocking improves the identification and utilization tions simulated, social-decision rules lead to stronger of thermals (e.g. Kerlinger 1989). convergence and slightly more efficient flight then do In order to identify the mechanisms leading to non-social decisions. Furthermore, under equally convergence and the importance of individual deci- dense thermal fields, the spatial distribution of ther- sion rules, a spatially explicit IBM named ‘Simsoar’ mals has a significant impact on the efficiency of was developed to simulate migration of soaring birds migration. (van Loon et al. in review). The model was parame- Although the model was initially run with static terized for the white stork and for atmospheric con- thermal conditions, the model can be extended in ditions in Israel based on extensive information from the future with an additional module to simulate visual observations, motorized glider flights, radar dynamic convective conditions (e.g. Allen 2006). (e.g. Leshem and YomTov 1996a, 1996b, 1998; Currently, we do not expect to have systematic mea- Shamoun-Baranes et al. 2003b, 2003c) and satellite surements of individual thermals; however, the telemetry studies on the migration of this species model can be parameterized with local meteorologi- along the eastern Palearctic flyway (e.g. Shamoun- cal conditions and the properties of landscapes to Baranes et al. 2003a, 2006). In the model, birds provide dynamic information on the density of ther- strive to reach their destination using soaring flight mals, the areas where thermals are most likely to Integrating meteorology and migration research 287 develop, height of the boundary layer, and vertical simulation (e.g. Erni et al. 2003). In order to develop lift (e.g. Shannon et al. 2002a, 2002b; Shamoun- these dynamic IBMs, data are essential, not only to Baranes et al. 2003b, 2003c). Simulated spatial and parameterize models with information such as flight temporal patterns can be compared to field data such speeds, departure dates, but also to develop reason- as visual observations, radar observations and track- able decision rules. Results from the model, in turn ing of individual birds. The model can also be used can, and should, be compared to measurements to compare different avian species or insects that which can be at the individual or population level use soaring flight during migration (e.g. Gibo and or consider local or more global patterns. The inte- Pallett 1979; Gibo 1981; Garland and Davis 2002). gration of models and measurements has shown that Furthermore, with the appropriate extensions, the atmospheric conditions can play a central role in shaping migratory success and efficiency. Often model can be applied to entire migratory trajectories and help identify the mechanisms that lead to regio- rather simple decision rules can enable animals to adapt to a very dynamic environment. nal and seasonal differences in migration. Guidelines for integrating atmospheric Using IBMs in research on migration conditions into migration models By developing modeling frameworks with flexible structure and explicit spatial and temporal dynamics For many empirical researchers perhaps the biggest we can study the importance of atmospheric condi- hurdle in such an approach is how and where to get tions and individual decision rules in different started. Following, we provide some guidelines, and migratory systems. In the case studies presented although we provide these in a particular order, the above we showed how studying individual responses process is often iterative. to atmospheric dynamics along a trajectory could (1) Data quantity and quality: Consider the quan- help explain emergent patterns such as the use of tity and quality of the available animal and atmo- emergency stopover sites or convergence of flight spheric data. Data are needed as input for models paths as well as understand and quantify the impor- and to compare with the output from models at the tance of the variability in atmospheric conditions relevant space and time. Table 3 provides a brief (within a year, along a trajectory, between years, overview of what types of meteorological data are or at different altitudes). The two case studies we available and for which spatial and temporal scales presented were examples of dynamic IBMs (DI, they would be most suitable. It is important to try to Table 1). This is a relatively large and diverse consider atmospheric conditions at the temporal and group of models with varying ranges of complexity. spatial scale most suitable for the ecological processes Depending on the aim and structure of the model, being studied (Fig. 1, see also Hallett et al. 2004). IBMs may (e.g. the white stork case study), or may (2) Model framework: Consider the aim of your not (e.g. the red knot case study), include interac- model and select the most suitable modeling frame- tions between individuals. IBMs may also include an work. When using models to integrate scientific aspect of heritability, where traits are transferred knowledge, there is a major distinction between the between generations and can evolve during a aim of making adequate (reliable, accurate and Table 3 An overview of the most relevant temporal scales (indicated by an X) for different types of meteorological data that can be incorporated into models of bird migration. Examples of on-line resources for such data are also provided Large eddy Regional numerical Station Global/continental Global circulation Temporal scale simulation mesoscale models observations reanalysis data indices Minutes X X – – – Hourly – X X X – Daily – X X X – Seasonal – – X X X a b c d On-line resource Generally none MM5 ECA&D NCEP reanalysis data NAO index The higher the spatial and temporal resolution of the data, generally the harder it is to find on the internet and such models must be run for the study of interest. PSU/NCAR mesoscale model (MM5); http://www.mmm.ucar.edu/prod/rt/pages/rt.html; Grell et al. 1994. ECA&D European climate and assessment dataset; http://eca.knmi.nl/; Klok and Klein Tank 2009. NCEP-NCAR reanalysis data; http://www.esrl.noaa.gov/psd/data/reanalysis/reanalysis.shtml; Kalnay et al. 1996. NAO (North Atlantic Oscillation) index; http://www.cgd.ucar.edu/cas/jhurrell/indices.html; Hurrell et al. 2003. 288 J. Shamoun-Baranes et al. precise, at the right scale) predictions and the aim of communication. Communication can be facilitated enhancing understanding. When the aim is to make by developing common terminology and using con- adequate predictions, it is generally desirable to ceptual frameworks for the description and design of match the resolution and extent of the model with models (e.g. Grimm et al. 2006; Nathan et al. 2008) the units and domain at which the predictions are as well as by research workshops dedicated to collab- required (‘scale of prediction’ for brevity), as well as oration (Bauer et al. 2009). gather data at the scale of prediction. In this way, errors due to mismatches of scale are avoided. If the Future perspectives scale at which the most important processes operate Several advances in different fields will strongly facil- does correspond with the scale of the prediction, try itate the integration of meteorology into migration to build a concept-driven model. However, if the research. First, meteorological data are becoming scale at which key processes operate is much finer more available and accessible, with numerous sources than the scale of prediction, it will be very hard (if of data freely available on the internet (Table 3). possible at all) to build a concept-driven model from Some of these sources are even archived globally expert knowledge and first principles. In that case a for several decades (e.g. NCEP-NCAR reanalysis data-driven model is the most suitable option. data, Kalnay et al. 1996). More recently, atmospheric In case the aim is to gain understanding about a models that can provide data at the temporal and certain aspect of a migration system, the question spatial scale of interest have been developed and becomes relevant whether you are in an explorative will greatly enhance migration research (Scott or a confirmative phase of your research. In the and Achtemeier 1987; Nathan et al. 2005). At very explorative phase, the aim is to identify patterns, fine scales this will often require that atmospheric attempt to find cause-and-effect relationships and models are run specifically for the research project compare alternative models. Currently it is not fea- (e.g. Shannon et al. 2002a, 2002b; Shamoun-Baranes sible to conduct such explorative activities with et al. 2003b, 2003c; Sapir 2009). Advances in tech- concept-based models because it requires too much nologies to collect data on animal movement effort to generate a single model. In the future, how- (e.g. Robinson et al. in press) will also facilitate the ever, more flexible modeling systems may be built integration of meteorology into migration research. that do, in fact, allow such activities (Taylor et al. Miniaturization of tracking technologies (e.g. 2007). So, data-driven models are the tools of choice Wikelski et al. 2006,2007; Stuchbury et al. 2009) in the explorative research phase. When a specific and collection of precise locations using the Global idea or hypothesis can be formulated, research Positioning system (GPS) improves the tracking of enters a confirmative phase where some sort of individual animals. High-resolution GPS may help formal comparison of that idea against observations revolutionize this field by providing detailed infor- or other ideas has to be made. In this phase, both mation on how animals respond to atmospheric concept- and data-driven models can be used dynamics en-route or even to help reveal animals’ effectively. decision rules. Furthermore, collecting additional Determining which type of model can be used data such as heart rate or 3-axial acceleration can under a given set of research aims and of constraints provide information on behavior and the expendi- with regard to availability of data (the variables that ture of energy (e.g. Ropert-Coudert and Wilson are available, as well as their resolution and extent), 2005; Rutz and Hays 2009) in relation to atmo- cannot be answered in general; it depends a lot on spheric conditions. In addition to individual track- the precise nature of a research question. However, ing, radar is an excellent tool for observing long-term Table 1 specifies the main properties and limitations spatial and temporal patterns of migration at specific of the various modeling techniques and can provide locations and has been used for several decades guidance about which modeling technique to use, in studies of the migration of birds (e.g. Erni et al. after a researcher has specified her/his research ques- 2005; van Belle et al. 2007; Schmaljohann et al. tion. Table 2 provides some examples of studies of migration with a reference to the different types of 2009), bats (e.g. Kunz et al. 2008) and insects (e.g. Chapman et al. 2003; Reynolds et al. 2005;). Weather model applied. (3) Communication and collaboration: During radar networks are particularly promising as they this process and the research itself, consider commu- provide multiple stations and can potentially help nicating and collaborating with the necessary experts, study larger-scale patterns and trajectories more e.g. modelers or meteorologists. Keep in mind that effectively then single locations. Although weather- the models themselves are also vehicles for surveillance Doppler radar has been used in the Integrating meteorology and migration research 289 Bauer S, van Dinther M, Høgda K-A, Klaassen M, Madsen J. United States for such studies (e.g. Diehl et al. 2003; 2008. The consequences of climate-driven stop-over sites Gauthreaux et al. 2008; Westbrook 2008), only changes on migration schedules and fitness of Arctic recently have several radars in the OPERA network geese. J Anim Ecol 77:654–60. in Europe (Operational Programme for the Exchange Bauer S, Barta Z, Ens BJ, Hays GC, McNamara JM, of Weather Radar Information; Kock et al. 2000) Klaassen M. 2009. Animal migration: linking models and been successfully tested for studying bird migration data beyond taxonomic limits. Biol Lett 5:433–5. (Holleman et al. 2008; van Gasteren et al. 2008; Bildstein KL. 2006. Migrating raptors of the world: their Doktor et al. in review) and it will become a valuable ecology and conservation. Ithaca, NY: Cornell University resource when studying Palearctic migration systems. Press. Atmospheric dynamics can affect migration sys- Bildstein KL, Zalles JI. 2005. Old world versus new world tems at many different levels, from instantaneous long-distance migration in accipiters, buteos, and falcons. changes in flight speed and direction to influencing In: Greenberg R, Marra PP, editors. The interplay of migra- breeding success. We hope to see meteorology more tion ability and global biogeography. Johns Hopkins strongly integrated into future research on migration University Press. p. 154–67. Birds of two worlds: the ecol- ogy and evolution of migration. Baltimore, MD. across multiple taxa. Such an interdisciplinary approach will help advance research on migration Bowlin M, et al. In press. Grand challenges in migration biology. Integr Comp Biol. as well as address some of the grand challenges in organismal biology (Schwenk et al. 2009; Bowlin Brandes D, Ombalski DW. 2004. Modeling raptor migration pathways using a fluid-flow analogy. J Raptor Research et al. this issue). 38:195–207. Brattstro ¨ m O, Kjelle ´ n N, Alerstam T, Akesson S. 2008. Effects Acknowledgments of wind and weather on red admiral, Vanessa atalanta, The authors thank M. Bowlin, I. A. Bisson and migration at a coastal site in southern Sweden. Anim Behav 76:335–44. M. Wikelski for organizing, and inviting J.S.B. to give a talk at the Integrative Migration Biology sym- Bruderer B, Underhill LG, Liechti F. 1995. Altitude choice by night migrants in a desert area predicted by meteorological posium at the 2010 Society for Integrative and factors. Ibis 137:44–55. Comparative Biology meeting in Seattle, Butler RW, Williams TD, Warnock N, Bishop M. 1997. Wind Washington. SICB’s Divisions of Animal Behavior, assistance a requirement for migration of shorebirds? Auk Neurobiology, and Comparative Endocrinology all 114:456–66. donated money to the symposium. The authors Chapman JW, Reynolds DR, Smith AD. 2003. Vertical-look- thank J. Leyrer as well as two anonymous reviewers ing radar: a new tool for monitoring high-altitude insect for discussions and constructive feedback on an ear- migration. Bioscience 53:503–11. lier version of the manuscript. They thank M. Chapman JW, Nesbit RL, Burgin LE, Reynolds DR, Duyvendak, for retrieving articles we did not have Smith AD, Middleton DR, Hill JK. 2010. Flight orientation direct access to. Our migration studies are facilitated behaviors promote optimal migration trajectories in high- by the BiG Grid infrastructure for eScience (http:// flying insects. Science 327:682–5. www.biggrid.nl). Chapman JW, Reynolds DR, Mouritsen H, Hill JK, Riley JR, Sivell D, Smith AD, Woiwod IP. 2008. Wind selection and drift compensation optimize migratory pathways in a high- References flying moth. Curr Biol 18:514–8. Able KP. 1973. The role of weather variables and flight direc- Cryan PM, Brown AC. 2007. Migration of bats past a remote tion in determining the magnitude of nocturnal bird island offers clues toward the problem of bat fatalities at migration. Ecology 54:1031–41. wind turbines. Biol Conserv 139:1–11. Akesson S, Hedenstro ¨ m A. 2000. Wind selectivity of migra- ¨ ¨ Danhardt J, LindstromA. 2001. Optimal departure decisions tory flight departures in birds. Behav Ecol Sociobiol of songbirds from an experimental stopover site and the 47:140–4. significance of weather. Anim Behav 62:235–43. Akesson S, Walinder G, Karlsson L, Ehnbom S. 2002. Diehl RH, Larkin RP, Black JE. 2003. Radar observations Nocturnal migratory flight initiation in reed warblers of bird migration over the Great Lakes. Auk 120: Acrocephalus scirpaceus: effect of wind on orientation and 278–90. timing of migration. J Avian Biol 33:349–57. Dingle H. 1996. Migration: the biology of life on the move. Allen MJ. 2006. Updraft Model for Development of New York: Oxford University Press. Autonomous Soaring Uninhabited Air Vehicles. Forty Drake VA, Farrow RA. 1988. The influence of atmospheric Fourth AIAA Aerospace Sciences Meeting and Exhibit; structure and motions on insect migration. Ann Rev Reno, Nevada. American Institute of Aeronautics and Entomol 33:183–210. Astronautics. 290 J. Shamoun-Baranes et al. Erni B, Liechti F, Bruderer B. 2003. How does a first Kerlinger P. 1989. Flight strategies of migrating hawks. year passerine migrant find its way? Simulating Chicago: The University of Chicago Press. migration mechanisms and behavioural adaptations. Kock K, Leitner T, Randeu WL, Divjak M, Schreiber KJ. 2000. Oikos 103:333. OPERA: Operational Programme for the Exchange of Erni B, Liechti F, Bruderer B. 2005. The role of wind in Weather Radar Information. First results and outlook for passerine autumn migration between Europe and Africa. the future, Vol. 25. Phys Chem Earth B: Hydrol Oceans Behav Ecol 16:732–40. Atmosphere. p. 1147–51. Erni B, Liechti F, Underhill LG, Bruderer B. 2002. Wind and Klok EJ, Klein Tank AMG. 2009. Updated and extended rain govern the intensity of nocturnal bird migration in European dataset of daily climate observations. Int J central Europe – a log-linear regression analysis. Ardea Climatol 29:1182–91. 90:155–66. Kunz TH, et al. 2008. Aeroecology: probing and modeling the Garland MS, Davis AK. 2002. An examination of Monarch aerosphere. Integr Comp Biol 48:1–11. Butterfly (Danaus plexippus) autumn migration in coastal Leshem Y, YomTov Y. 1996a. The use of thermals by soaring Virginia. American Midland Naturalist 147:170–4. migrants. Ibis 138:667–74. Gauthreaux SA Jr, Livingston JW, Belser CG. 2008. Detection Leshem Y, YomTov Y. 1996b. The magnitude and timing of and discrimination of fauna in the aerosphere using migration by soaring raptors, pelicans and storks over Doppler weather surveillance radar. Integr Comp Biol Israel. Ibis 138:188–203. 48:12–23. Leshem Y, YomTov Y. 1998. Routes of migrating soaring Gibo DL. 1981. Some observations on soaring flight in the birds. Ibis 140:41–52. Mourning Cloak Butterfly (Nymphalis antiopa L.) in south- Leskinen M, Markkula I, Koistinen J, Pylkko ¨ P, Ooperi S, ern Ontario. J New York Entomol S 89:98–101. Siljamo P, Ojanen H, Raiskio S, Tiilikkala K. 2009. Pest Gibo DL, Pallett MJ. 1979. Soaring flight of monarch butter- insect immigration warning by an atmospheric dispersion flies, Danaus plexippus (Lepidoptera: Danaidae), during late model, weather radars and traps. J Appl Entomol published summer migration in southern Ontario. Can J Zool online (doi: 10.1111/j.1439-0418.2009.01480.x). 57:1393–401. Leyrer J, Bocher P, Robin F, Delaporte P, Goulevent C, Gill RE, Tibbitts TL, Douglas DC, Handel CM, Mulcahy DM, Joyeux E, Meunier F, Piersma T. 2009. Northward migra- Gottschalck JC, Warnock N, McCaffery BJ, Battley PF, tion of Afro-Siberian Knots Calidris canutus canutus: High Piersma T. 2009. Extreme endurance flights by landbirds variability in Red Knots numbers visiting stopover sites on crossing the Pacific Ocean: ecological corridor rather than French Atlantic coast (1979–2009). Wader Study Group barrier? Proc R Soc B 276:447–57. Bull 116:145–51. Grell G, Dudhia J, Stauffer D. 1994. A description of the fifth- Liechti F. 2006. Birds: blowin’ by the wind? J Ornithol generation Penn State/NCAR Mesoscale Model (MM5). 147:202–11. NCAR Technical Note. NCAR. p. 117. Mandel JT, Bildstein KL, Bohrer G, Winkler DW. 2008. Grimm V, et al. 2006. A standard protocol for describing Movement ecology of migration in turkey vultures. Proc individual-based and agent-based models. Ecol Model Natl Acad Sci USA 105:19102–7. 198:115–26. Nathan R, Getz WM, Revilla E, Holyoak M, Kadmon R, Hallett TB, Coulson T, Pilkington JG, Clutton-Brock TH, Saltz D, Smouse PE. 2008. A movement ecology paradigm Pemberton JM, Grenfell BT. 2004. Why large-scale climate for unifying organismal movement research. Proc Natl indices seem to predict ecological processes better than Acad Sci USA 105:19052–9. local weather. Nature 430:71–5. Nathan R, et al. 2005. Long-distance biological transport pro- Holleman I, van Gasteren H, Bouten W. 2008. Quality assess- cesses through the air: can nature’s complexity be unfolded ment of weather radar wind profiles during bird migration. in silico? Divers Distrib 11:131–7. J Atmospheric Oceanic Technol 25:2188–98. Newton I. 2006. Can conditions experienced during migration Hu ¨ ppop O, Hu ¨ ppop K. 2003. North Atlantic Oscillation limit the population levels of birds? J Ornithol 147:146–66. and timing of spring migration in birds. Proc R Soc B Newton I. 2008. The migration ecology of birds. Oxford: 270:233–40. Academic Press. Hurrell J, Kushnir Y, Ottersen G, Visbek M. 2003. An over- Pennycuick CJ, Alerstam T, Larsson B. 1979. Soaring migra- view of North Atlantic Oscillation. In: Hurrell J, Kushnir Y, tion of the common crane Grus grus observed by radar and Ottersen G, visbek M, editors. The North Atlantic from an aircraft. Ornis Scand 10:241–51. Oscillation: climate significance and environmental impacts. Washington, DC: American Geophysical Union. ˚ Piersma T, Lindstro ¨mA. 1997. Rapid reversible changes in p. 1–35. organ size as a component of adaptive behaviour. Trends Jonzen N, et al. 2006. Rapid advance of spring arrival Ecol Evol 12:134–8. dates in long-distance migratory birds. Science Piersma T, van de Sant S. 1992. Pattern and predictability of 312:1959–61. potential wind assistance for waders and geese migrating Kalnay E, et al. 1996. The NCEP/NCAR 40-year reanalysis from West Africa and the Wadden Sea to Siberia. Ornis project. Bull Am Meteorol Soc 77:437–71. Svecica 2:55–66. Integrating meteorology and migration research 291 Piersma T, Prokosch P, Bredin D. 1992. The migration system weather at departure sites, onset of migration and timing of Afro-Siberian knots Calidris canutus canutus. Wader of soaring-bird autumn migration in Israel? Global Ecol Study Group Bull 64(Suppl):52–63. Biogeogr 15:541–52. Reilly JR, Reilly RJ. 2009. Bet-hedging and the orientation Shamoun-Baranes J, Baharad A, Alpert P, Berthold P, of juvenile passerines in fall migration. J Anim Ecol YomTov Y, Dvir Y, Leshem Y. 2003a. The effect of wind, 78:990–1001. season and latitude on the migration speed of white storks Ciconia ciconia, along the eastern migration route. J Avian Reynolds AM, Reynolds DR, Riley JR. 2009. Does a ‘turbo- Biol 34:97–104. phoretic’ effect account for layer concentrations of insects migrating in the stable night-time atmosphere? J R Soc Shamoun-Baranes J, Leyrer J, van Loon E, Bocher P, Robin F, Interface 6:87–95. Meunier F, Piersma T. 2010. Stochastic atmospheric assis- tance and the use of emergency staging sites by migrants. Reynolds DR, Chapman JW, Edwards AS, Smith AD, Proc R Soc B published online (doi: 10.1098/ Wood CR, Barlow JF, Woiwod IP. 2005. Radar studies rspb.2009.2112). of the vertical distribution of insects migrating over southern Britain: the influence of temperature inversions Shannon HD, Young GS, Yates MA, Fuller MR, Seegar WS. on nocturnal layer concentrations. B Entomol Res 2002a. American White Pelican soaring flight times and 95:259–74. altitudes relative to changes in thermal depth and intensity. Condor 104:679–83. Reynolds DS. 2006. Monitoring the potential impact of a wind development site on bats in the northeast. Shannon HD, Young GS, Yates MA, Fuller MR, Seegar WS. J Wildlife Manag 70:1219–27. 2002b. Measurements of thermal updraft intensity over complex terrain using American White Pelicans and a Richardson WJ. 1978. Timing and amount of bird migration simple boundary-layer forecast model. Bound-Lay in relation to weather: a review. Oikos 30:224–72. Meteorol 104:167–99. Richardson WJ. 1990. Timing of bird migration in relation to Spaar R, Bruderer B. 1997. Migration by flapping or soaring: weather: updated review. In: Gwinner E, editor. Bird migra- Flight strategies of Marsh, Montagu’s and Pallid Harriers in tion. Berlin: Springer-Verlag. p. 78–101. southern Israel. Condor 99:458–69. Robinson WD, Bowlin MS, Bisson I-A, Shamoun-Baranes J, Spaar R, Stark H, Liechti F. 1998. Migratory flight strategies Thorup K, Diehl RH, Kunz TH, Mabey S, Winkler DW. of Levant sparrowhawks: time or energy minimization? In press. Integrating concepts and technologies to advance Anim Behav 56:1185–1197. the study of bird migration. Front Ecol Environ. Srygley RB, Dudley R. 2008. Optimal strategies for insects Ropert-Coudert Y, Wilson RP. 2005. Trends and perspectives migrating in the flight boundary layer: mechanisms and in animal-attached remote sensing. Front Ecol Environ consequences. Integr Comp Biol 48:119–33. 3:437–44. Stefanescu C, Alarco ´nM, Avila A. 2007. Migration of the Rutz C, Hays GC. 2009. New frontiers in biologging science. painted lady butterfly, Vanessa cardui, to north-eastern Biol Lett 5:289–92. Spain is aided by African wind currents. J Anim Ecol Sapir N. 2009. The effect of weather on Bee-eater (Merops 76:888–98. apiaster) migration. Jerusalem: Hebrew University. Stocker S, Weihs D. 1998. Bird migration – an energy-based Schaub M, Liechti F, Jenni L. 2004. Departure of migrating analysis of costs and benefits. IMA J Math Appl Med European robins, Erithacus rubecula, from a stopover site in 15:65–85. relation to wind and rain. Anim Behav 67:229–37. Stoddard PK, Marsden JE, Williams TC. 1983. Computer Schmaljohann H, Liechti F, Bruderer B. 2009. Trans-Sahara simulation of autumnal bird migration over the western migrants select flight altitudes to minimize energy costs North Atlantic. Anim Behav 31:173–80. rather than water loss. Behav Ecol Sociobiol 63:1609–19. Stutchbury BJM, Tarof SA, Done T, Gow E, Kramer PM, Tautin J, Schwenk K, Padilla DK, Bakken GS, Full RJ. 2009. Grand Fox JW, Afanasyev V. 2009. Tracking long-distance songbird challenges in organismal biology. Integr Comp Biol migration by using geolocators. Science 323:896. 49:7–14. Taylor IJ, Deelman E, Gannon DB, Shields M, editors . 2007. Scott RW, Achtemeier GL. 1987. Estimating pathways of Workflows for e-Science. London: Springer-Verlag. migrating insects carried in atmospheric winds. Environ Thorup K, Alerstam T, Hake M, Kjellen N. 2003. Bird orien- Entomol 16:1244–54. tation: compensation for wind drift in migrating raptors Shamoun-Baranes J, Leshem Y, Yom-Tov Y, Liechti O. 2003b. is age dependent. Biol Lett 270:8–11. Differential use of thermal convection by soaring birds over van Belle J, Shamoun-Baranes J, van Loon E, Bouten W. 2007. central Israel. Condor 105:208–18. An operational model predicting autumn bird migration Shamoun-Baranes J, Liechti O, Yom-Tov Y, Leshem Y. 2003c. intensities for flight safety. J Appl Ecol 44:864–74. Using a convection model to predict altitudes of white van de Kam J, Ens B, Piersma T, Zwarts L. 2004. Shorebirds: an stork migration over central Israel. Bound-Lay Meteorol illustrated behavioural ecology. Utrecht: KNNV Publishing. 107:673–81. van Gasteren H, Holleman I, Bouten W, Van Loon E, Shamoun-Baranes J, van Loon E, Alon D, Alpert P, Shamoun-Baranes J. 2008. Extracting bird migration Yom-Tov Y, Leshem Y. 2006. Is there a connection between 292 J. Shamoun-Baranes et al. information from C-band Doppler weather radars. Ibis Wikelski M, Moskowitz D, Adelman JS, Cochran J, 150:674–86. Wilcove DS, May ML. 2006. Simple rules guide dragonfly migration. Biol Lett 2:325–9. Vrugt JA, van Belle J, Bouten W. 2007. Pareto front analysis of flight time and energy use in long-distance bird migra- Wood CR, Chapman JW, Reynolds DR, Barlow JF, Smith AD, tion. J Avian Biol 38:432–42. Woiwod IP. 2006. The influence of the atmospheric bound- ary layer on nocturnal layers of noctuids and other Weber TP, Alerstam T, Hedenstro ¨ m A. 1998. Stopover deci- moths migrating over southern Britain. Int J Biometeorol sions under wind influence. J Avian Biol 29:552–60. 50:193–204. Westbrook JK. 2008. Noctuid migration in Texas within the Wood CR, Clark SJ, Barlow JF, Chapman JW. 2010. Layers nocturnal aeroecological boundary layer. Integr Comp Biol of nocturnal insect migrants at high-altitude: the influence 48:99–106. of atmospheric conditions on their formation. Agric For Wikelski M, Kays RW, Kasdin NJ, Thorup K, Smith JA, Entomol 12:113–21. Swenson GW Jr. 2007. Going wild: what a global small-animal tracking system could do for experimental biologists. J Exp Biol 210:181–6. http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Integrative and Comparative Biology Pubmed Central

Integrating Meteorology into Research on Migration

Integrative and Comparative Biology , Volume 50 (3) – Apr 8, 2010

Loading next page...
 
/lp/pubmed-central/integrating-meteorology-into-research-on-migration-mr9ZueW0Ut

References (104)

Publisher
Pubmed Central
Copyright
© The Author 2010. Published by Oxford University Press on behalf of the Society for Integrative and Comparative Biology. All rights reserved. For permissions please email: journals.permissions@oxfordjournals.org.
ISSN
1540-7063
eISSN
1557-7023
DOI
10.1093/icb/icq011
Publisher site
See Article on Publisher Site

Abstract

Integrative and Comparative Biology, volume 50, number 3, pp. 280–292 doi:10.1093/icb/icq011 SYMPOSIUM Judy Shamoun-Baranes, Willem Bouten and E. Emiel van Loon Computational Geo-Ecology, Institute for Biodiversity and Ecosystem Dynamics, University of Amsterdam, P.O. Box 94248, 1090 GE Amsterdam, The Netherlands E-mail: shamoun@uva.nl From the symposium ‘‘Integrative Migration Biology’’ presented at the annual meeting of the Society for Integrative and Comparative Biology, January 3–7, 2010 at Seattle, Washington. Synopsis Atmospheric dynamics strongly influence the migration of flying organisms. They affect, among others, the onset, duration and cost of migration, migratory routes, stop-over decisions, and flight speeds en-route. Animals move through a heterogeneous environment and have to react to atmospheric dynamics at different spatial and temporal scales. Integrating meteorology into research on migration is not only challenging but it is also important, especially when trying to understand the variability of the various aspects of migratory behavior observed in nature. In this article, we give an overview of some different modeling approaches and we show how these have been incorporated into migration research. We provide a more detailed description of the development and application of two dynamic, individual-based models, one for waders and one for soaring migrants, as examples of how and why to integrate meteorology into research on migration. We use these models to help understand underlying mechanisms of individual response to atmospheric conditions en-route and to explain emergent patterns. This type of models can be used to study the impact of variability in atmospheric dynamics on migration along a migratory trajectory, between seasons and between years. We conclude by providing some basic guidelines to help researchers towards finding the right modeling approach and the meteorological data needed to integrate meteorology into their own research. resulting in carry-over effects such as an impact on Introduction survival (e.g. Erni et al. 2005; Newton 2006), timing For flying organisms, such as insects, bats and birds, of migration or breeding success. Although global atmospheric dynamics play an important role in patterns of atmospheric circulation may not affect their migratory movements (e.g. Richardson 1990; instantaneous responses of individual migrants Dingle 1996; Liechti 2006; Kunz et al. 2008). directly, they will shape atmospheric conditions at Animals move through a dynamic and heterogeneous smaller scales. At the same time they are likely to environment where conditions from the microscale have more long term and cumulative effects on through the mesoscale and even global circulation migration by affecting migration routes and seasonal patterns are relevant (Drake and Farrow 1988; timing of long distance movements. We provide a Nathan et al. 2005; Kunz et al. 2008). The effects simplified representation of these multi-scale interac- of atmospheric dynamics on migration are complex tions in Fig. 1. and may differ depending on the temporal and spa- Meteorology should be integrated into research on tial scale being considered as well as on the species, migration, especially when trying to understand nat- region, season and year. For example, instantaneous ural variability observed in aspects like the timing responses to changing wind speeds due to microscale of migration; migratory routes; orientation; use of and/or mesoscale dynamics may affect the flight stopover sites; or population trends such as speed, course and energetic cost at that point in effects on survival or breeding success as a result of time, as well as the conditions that will be experi- changes in arrival time or physiological condition. enced later en-route (e.g. Chapman et al. 2010; Meteorology is also of interest at longer time Shamoun-Baranes et al. 2010). These effects may scales when trying to understand the evolution of also be cumulative throughout the season, finally Advanced Access publication April 8, 2010 This is an Open Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (http://creativecommons.org/licenses/ by-nc/2.5/) which permits unrestricted non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited. The Author 2010. Published by Oxford University Press on behalf of the Society for Integrative and Comparative Biology. All rights reserved. For permissions please email: journals.permissions@oxfordjournals.org. Integrating meteorology and migration research 281 Fig. 1 A simplified representation of the different spatio-temporal scales of atmospheric dynamics that may influence instantaneous behavioral responses, resulting in short-term (instantaneous) and longer-term effects. Carry-over effects include not only inter-annual effects (e.g. population size or breeding success), but longer-term effects that may have evolutionary consequences (e.g. shaping migration routes). For example, instantaneous changes in flight behavior would influence instantaneous flight speed, the timing of migration within a migration season and could also lead to carry-over effects such as the timing of breeding or breeding success. particular migratory systems. Thus, integrating mete- Kemp et al. in review), flight altitudes (Bruderer orology and research on animal migration will et al. 1995; Wood et al. 2006, 2010; Reynolds et al. help us better understand both short- and long-term 2009; Schmaljohann et al. 2009), flight strategy (Gibo organismal–environmental linkages, one of the grand and Pallett 1979; Pennycuick et al. 1979; Gibo 1981; Spaar and Bruderer 1997; Spaar et al. 1998; Sapir challenges identified in organismal biology (Schwenk 2009), orientation and trajectories (Thorup et al. et al. 2009). However, linking mechanisms at the 2003; Chapman et al. 2008; Srygley and Dudley individual level to these longer-term, or larger-scale 2008; Chapman et al. 2010), migration intensity or consequences remains challenging. probability (Erni et al. 2002; Reynolds 2006; Cryan During the past few decades numerous empirical and Brown 2007; Stefanescu et al. 2007; van Belle and theoretical studies have addressed the influence et al. 2007; Leskinen et al. 2009), as well as migratory of atmospheric dynamics on animals’ migrations success (Erni et al. 2005; Reilly and Reilly 2009). (for overviews see Richardson 1978, 1990; Drake and Farrow 1988; Dingle 1996; Liechti 2006; Modeling approaches Newton 2008). Atmospheric conditions are known to influence the onset of migration (Shamoun- Different modeling techniques have been used to Baranes et al. 2006; Gill et al. 2009), migration study the influence of atmospheric dynamics on phenology (Hu ¨ ppop and Hu ¨ ppop 2003; Jonzen migration. We would like to distinguish between et al. 2006; Bauer et al. 2008), stopover decisions ‘concept-driven’ and ‘data-driven’ models. The struc- (Akesson and Hedenstro ¨ m 2000; Da ¨nhardt and ture of concept-driven models can only be conceived Lindstro ¨ m 2001; Schaub et al. 2004; Wikelski et al. if extensive prior knowledge of the system is available 2006; Brattstrom et al. 2008), flight speeds (Garland and cannot be discovered in an automated fashion. and Davis 2002; Shamoun-Baranes et al. 2003a; Calibration of parameters of concept-driven models 282 J. Shamoun-Baranes et al. Table 1 An overview of different types of concept- and data-driven models and their characteristics Requirement for creating the model Conceptual Numerical and Observations understanding data processing on state Possibilities for Frequency Name Description of the system skills variables calibration of use Concept-driven SC Static Concept-based model Intermediate Low Few Easy, many methods Intermediate DI Dynamic IBM Intermediate Intermediate Intermediate Difficult, few methods Intermediate DC Dynamic Continuum-based High High Intermediate Intermediate, few methods Low model Data-driven SD Static Data-based model Low Low Intermediate Easy, many methods High DD Dynamic Data-based model Low High Many Intermediate, few methods Low Frequency of use in migration studies is provided in the last column; for references to specific studies see Table 2. In this context, static means that the process being studied is either in steady state or that there is no influence of previous states on the current state. Individual-based: model state variables refer to properties of an individual; continuum based: model state variables refer to population properties. State variables are model-entities which are updated at each model time step with a difference equation in dynamic models and are usually comparable to the dependent variables in static models. is possible if measurements of model output are questions about system feedbacks and scale, compar- available, but this is not always required to use the ing theories and observations, and identifying ave- model. Reasonable values of parameters can often be nues for new research. By modeling the influence found by using expert knowledge or information of atmospheric conditions en-route we can study from experiments or biophysical calculations. emergent patterns of migration at the individual Concept driven models can then be run as thought and population level as well as study the importance experiments, without any observations. of the variability in individual behavior and the vari- The structure of data-driven models, however, can ability in atmospheric conditions, between days, sea- be derived via (highly) automated procedures but sons, years or regions. The models are tools to better does not necessarily represent cause and effect rela- understand underlying mechanisms, but not goals in tions in nature. These models need relatively little themselves. In such models, measurements gathered prior knowledge of the system before they can be from field research or from laboratory experiments constructed, but always require calibration of the can either be implicitly integrated into the models to parameters because the parameters do not necessarily formulate model assumptions and to parameterize match physical entities that can be independently models, or explicitly to compare to model results. observed in nature. Because of the necessity to cali- Furthermore, atmospheric conditions from observa- brate, data-driven models always need observations tions, reanalysis data, numerical models, or artificial on both input and output variables of the model. data (see Table 3 for examples) are needed as input Table 1 specifies three different types of to the models. concept-driven modeling techniques and two differ- In this article we describe the development and ent types of data-driven modeling techniques with application of two studies using dynamic models of some of their main characteristics. These definitions migration (‘concept-driven’, dynamic individual- are of course not specific for models of migration, based models; DI, Table 1) as examples of how but are applicable to ecological models in general. and why to integrate meteorology into research on Table 2 provides several examples of the various migration. We use the models to better understand techniques of measurement and modeling applied underlying mechanisms at the individual level to in studying the influence of atmospheric dynamics help explain the patterns that emerge at the popula- on migration. tion level. We provide some basic guidelines to help We think that especially concept-driven dynamic researchers towards integrating meteorology into simulations of migration provide a suitable frame- their research on migration and we discuss the work for integrating scattered knowledge about opportunities and limitations of different sources of migration, systematically addressing complex data that can be used for such studies. We thus hope Integrating meteorology and migration research 283 Table 2 A selection of studies on the influence of atmospheric conditions on animal migration, including focal species or group, types of data used, geographic region of study, type of model, and relevant reference Effects on migration Species/group Migration data Meteorological variable: data source Geographic region Model type References Flight behavior: altitude Nocturnal migratory Tracking radar Wind : radiosonde NCEP reanalysis data Sahara SC Schmaljohann birds et al. 2009 Flight behavior: altitude Soaring avian Motorized glider Boundary layer height and vertical lift: Israel SD Shamoun-Baranes migrants boundary layer convective model et al. 2003b,c Flight behavior: altitude Nocturnal migratory Radar Various: numerical weather prediction UK SD Wood et al. 2010 insects model, the Unified Model Take off decisions Arctic geese Ringing data Onset of spring proxy: NDVI Palearctic flyway DI Bauer et al. 2008 Take off decisions Bar tailed godwit Satellite telemetry Sea level pressure, wind : NCEP reanalysis Pacific ocean flyway Descriptive no Gill et al. 2009 data GEOS-5 global atmospheric model computer model Take off decisions Not relevant None Wind assistance or no assistance: no data Not relevant SC Weber et al. 1998 Take off decisions Green darners Radio telemetry Wind and temperature: weather station Northeast USA SD Wikelski et al. observations 2006 d a Migration intensity Nocturnal migratory Radar Wind , barometric pressure, temperature, The Netherlands SD van Belle et al. birds precipitation: weather station 2007 observations d a Migration intensity Nocturnal passerine Radar and visual Wind , temperature, synoptic weather Southeastern USA SD Able 1973 migration observations index: weather station observations Migration intensity Black-cherry aphids Radar and insect Wind : HIRLAM and ECMWF numerical Finland DC Leskinen et al. and diamond-back traps weather prediction models 2009 moths Speed Turkey vulture Satellite telemetry Wind speed, turbulent kinetic energy, cloud Eastern North DD Mandel et al. height: North American regional reana- American flyway 2008 lysis data Speed Red knots Visual Wind : NCEP reanalysis data Afro–Siberian flyway DI Shamoun-Baranes observations et al. 2010 Direction/Orientation Reed warbler Radio telemetry Wind : weather station observations Sweden SD Akesson et al. Direction/Orientation Moths and butterflies Radar Wind : numerical weather prediction UK SD Chapman et al. model, the Unified Model 2010 Timing Soaring avian Visual observa- Barometric pressure, temperature, precipi- Western Palearctic SC Shamoun-Baranes migrants tions, Satellite table water: NCEP reanalysis data (eastern) flyway et al. 2006 telemetry Timing Passerine migrants Ringing data North Atlantic Oscillation Europe and SD Jonzen et al. 2006 Scandinavia (continued) 284 J. Shamoun-Baranes et al. to facilitate further integration of field biology, mete- orology and modeling. Case study of the migration of red knots: The importance of wind From numerous studies over the years it is clear that of all atmospheric conditions wind plays the most important role in avian migration (Akesson and Hedenstro ¨ m 2000; Akesson et al. 2002; Shamoun- Baranes et al. 2003a; van Belle et al. 2007; Schmaljohann et al. 2009; Kemp et al. in review; for review see Liechti 2006). Yet, quantifying the cumulative effect of wind en-route for an entire migration trajectory, as well as the effect of variabil- ity in space and time has rarely been done in avian research (Stoddard et al. 1983; Erni et al. 2005; Vrugt et al. 2007; Reilly and Reilly 2009). The Afro–Siberian red knot (Calidris canutus canutus) is an intensively studied long-distance migrant (e.g. Piersma et al. 1992; Piersma and Lindstrom 1997; van de Kam et al. 2004). These birds migrate north in two non-stop flights of approximately 4400 km each from their wintering grounds in Mauritania via the German Wadden Sea to the Siberian breeding grounds in only four weeks (Piersma et al. 1992; van de Kam et al. 2004). From previous studies, favorable winds were considered essential along this flyway (Piersma and van de Sant 1992). Field observations have shown that red knots erratically use an extra stopover site on the French Atlantic coast (Leyrer et al. 2009). One hypothesis to explain this phenomenon was that birds experiencing unfavorable winds would use this area as an emergency stopover site. Therefore, a dynamic individual-based model (IBM) of north- bound migration, incorporating winds experienced en-route, was developed to study whether use of this intermediate stopover site could be explained by stochastic wind conditions (Shamoun-Baranes et al. 2010). In the model, birds are moved forward in 6-h time steps along the great-circle route between their wintering site in Mauritania and their stopover site at the Wadden Sea. The wind experienced at the beginning of each time step determines the ground speed of the bird which is subsequently used to cal- culate flight times and the birds’ locations at the next time step (Fig. 2). Data on speed and direction of wind at four different levels of pressure were used in this study to represent wind conditions experienced at different altitudes during flight. The data were extracted from the global NCEP-Reanalysis dataset which has a spatial resolution of 2.58 2.58 and a 6-h temporal resolution (Kalnay et al. 1996). The Table 2 Continued Effects on migration Species/group Migration data Meteorological variable: data source Geographic region Model type References Arrival mass Western sandpipers Biometric Wind : weather station observations North American DI Butler et al. 1997 measurements Pacific Coast Route Silver Y (noctuid Radar Wind : numerical weather prediction United Kingdom, DI Chapman et al. moth) model, the Unified Model Northwest Europe 2010 Route Golden Eagles Visual Wind (implicit): digital elevation model Central Pennsylvania DC Brandes and observations Ombalski 2004 Survival Simulated nocturnal Literature Wind : NCEP reanalysis data Western Palearctic DI Erni et al. 2005 passerine migrant migration Mass, population dynamics Houbara bustard Literature Winter severity (implicit): no data Not directly relevant DC Sto ¨ cker and Stonechat Weihs 1998 Type of model is described in more detail in Table 1. The terms used in the column entitled ‘Effects on migration’ are adapted to roughly follow the framework provided in Fig. 1 for comparative purposes and, thus, do not always follow the exact terms used in the original study. Similarly, although many effects may be studied with one model, we generally highlight the effect that was the focus of the study. For suggestions on where to find different sources of meteorological data, some of which are mentioned in this table, see Table 3. Wind speed and direction. White stork, honey buzzard, lesser spotted eagle. NDVI, normalized difference vegetation index. Migration intensity can also be considered a proxy for takeoff decisions. Integrating meteorology and migration research 285 Fig. 2 A forward simulation of the migration of red knots taking off on May 1, 1986. Upward-pointing and downward-pointing triangles indicate wintering site (simulated start location) and Wadden sea stopover site (simulated end location) respectively. The open circle marks the location of the emergency stopover site on the French Atlantic coast. Black circles indicate location at each time step. Arrows indicate the speed and direction of the wind at each location and the dotted line shows the flight trajectory. The shorter the distance between circles, the slower is the ground speed due to disadvantageous winds. advantages of the NCEP reanalysis dataset for this conducted with appropriate modeling techniques type of analysis are the homogenous spatial and tem- and field measurements. One of the main advantages of such a modeling poral coverage, global and long term coverage and free availability on the internet. After running simu- framework is the ability to also explore the effect of lations, results from the migration model can be variability of wind conditions across years, starting compared to observations. dates, and altitudes of flight, all resulting in different A comparison between simulated flight times and wind conditions and flight times. Furthermore, the the number of birds observed stopping over at the potential effects of spatial and temporal auto- French staging site showed how unpredictable winds correlation in wind conditions could be studied affect flight times and that wind is a predominant along the migratory route. For example, wind con- driver of the use of an emergency stopover site along ditions experienced in France were spatially and tem- the French Atlantic coast. Wind clearly plays an porally auto-correlated for418 h and sometimes over important and quantifiable role in this, and probably hundreds of kilometers. Spatio-temporal correlation many other, migratory systems. The study also indi- in atmospheric dynamics could provide migrants cates the importance of this emergency stopover site with information on what to expect further along for conservation. Although the Wadden Sea is an their trajectory and may enable them to fine-tune obligatory staging area in this system, the French their decisions based on their physiological state, Atlantic coast may be essential for ensuring the sur- geographic location, and immediate and expected vival of individuals that have encountered very unfa- environmental conditions. vorable winds en-route. Thus, in the long term, this In the future, the model can be extended by inte- emergency stopover site may influence, and help sta- grating the energetics of flight to model the expen- bilize, migratory population dynamics. These ideas, diture of energy due to the wind conditions however, require further research that can be experienced en-route. These results then can be 286 J. Shamoun-Baranes et al. compared to field measurements. The model is not only designed to enable further extensions but can be applied to other species where sufficient data are available. The model framework presented here is quite similar to meteorological trajectory analysis used to estimate the flight paths of migrating insects (e.g. Scott and Achtemeier 1987; Chapman et al. 2010) as a result of meteorological dynamics experi- enced en-route. Although in contrast to migrating insects, the location of the departure and destination are known for the red knots. Case study of migration by white storks: The importance of thermal convection Atmospheric dynamics strongly influence the migra- Fig. 3 3D trajectories of simulated migration of white storks. Thermals are indicated as grey cylinders; the destination is indi- tion of soaring birds; particularly thermal convection cated by a gray box. Each trajectory represents the movement of which influences daily flight schedules and migration an individual during the simulation. When in a thermal, birds routes (Kerlinger 1989; Leshem and Yom-Tov 1998), climb vertically until they reach the top; they then glide (losing flight speeds (Leshem and Yom-Tov 1996a; altitude) towards the next thermal if it can be sensed and Mandel et al. 2008) and flight altitudes (Leshem reached by the birds. Otherwise the bird glides first to the and Yom-Tov 1996a; Shannon et al. 2002a, 2002b; destination, until another thermal can be utilized. In this Shamoun-Baranes et al. 2003b, 2003c). Many soaring simulation, a bird first searches for the most distant thermal containing other birds. species are known to migrate in large single-species or mixed-species flocks aggregating in both space and time (e.g. Kerlinger 1989; Leshem and Yom-Tov 1996a, 1996b). Large-scale aggregations and select thermals based on several physical con- along well established and narrow migratory corri- straints and on pre-determined behavioral rules dors are often attributed to natural leading lines like which differentiate between thermals with and with- the Appalachian Mountains or to circumvention of out birds (Fig. 3). At each time step birds are either large bodies of water, resulting in geographic bottle- climbing in a thermal or gliding to their current necks such as those seen in Panama, Gibraltar, the target (a thermal or their destination). This design Bosporus and Israel (Kerlinger 1989; Leshem and of the model provides a framework for virtual exper- YomTov 1996b; Bildstein and Zalles 2005; Bildstein iments which can be used to explore the patterns 2006). However, the mechanisms that result in small that emerge due to different decision rules, flight scale convergence and the potential benefits of flock- parameters, convective conditions, or takeoff and ing remain largely unknown. With our model, which destination areas, and to test different scenarios. is described below, we explore the hypothesis that The study shows that under the convective condi- flocking improves the identification and utilization tions simulated, social-decision rules lead to stronger of thermals (e.g. Kerlinger 1989). convergence and slightly more efficient flight then do In order to identify the mechanisms leading to non-social decisions. Furthermore, under equally convergence and the importance of individual deci- dense thermal fields, the spatial distribution of ther- sion rules, a spatially explicit IBM named ‘Simsoar’ mals has a significant impact on the efficiency of was developed to simulate migration of soaring birds migration. (van Loon et al. in review). The model was parame- Although the model was initially run with static terized for the white stork and for atmospheric con- thermal conditions, the model can be extended in ditions in Israel based on extensive information from the future with an additional module to simulate visual observations, motorized glider flights, radar dynamic convective conditions (e.g. Allen 2006). (e.g. Leshem and YomTov 1996a, 1996b, 1998; Currently, we do not expect to have systematic mea- Shamoun-Baranes et al. 2003b, 2003c) and satellite surements of individual thermals; however, the telemetry studies on the migration of this species model can be parameterized with local meteorologi- along the eastern Palearctic flyway (e.g. Shamoun- cal conditions and the properties of landscapes to Baranes et al. 2003a, 2006). In the model, birds provide dynamic information on the density of ther- strive to reach their destination using soaring flight mals, the areas where thermals are most likely to Integrating meteorology and migration research 287 develop, height of the boundary layer, and vertical simulation (e.g. Erni et al. 2003). In order to develop lift (e.g. Shannon et al. 2002a, 2002b; Shamoun- these dynamic IBMs, data are essential, not only to Baranes et al. 2003b, 2003c). Simulated spatial and parameterize models with information such as flight temporal patterns can be compared to field data such speeds, departure dates, but also to develop reason- as visual observations, radar observations and track- able decision rules. Results from the model, in turn ing of individual birds. The model can also be used can, and should, be compared to measurements to compare different avian species or insects that which can be at the individual or population level use soaring flight during migration (e.g. Gibo and or consider local or more global patterns. The inte- Pallett 1979; Gibo 1981; Garland and Davis 2002). gration of models and measurements has shown that Furthermore, with the appropriate extensions, the atmospheric conditions can play a central role in shaping migratory success and efficiency. Often model can be applied to entire migratory trajectories and help identify the mechanisms that lead to regio- rather simple decision rules can enable animals to adapt to a very dynamic environment. nal and seasonal differences in migration. Guidelines for integrating atmospheric Using IBMs in research on migration conditions into migration models By developing modeling frameworks with flexible structure and explicit spatial and temporal dynamics For many empirical researchers perhaps the biggest we can study the importance of atmospheric condi- hurdle in such an approach is how and where to get tions and individual decision rules in different started. Following, we provide some guidelines, and migratory systems. In the case studies presented although we provide these in a particular order, the above we showed how studying individual responses process is often iterative. to atmospheric dynamics along a trajectory could (1) Data quantity and quality: Consider the quan- help explain emergent patterns such as the use of tity and quality of the available animal and atmo- emergency stopover sites or convergence of flight spheric data. Data are needed as input for models paths as well as understand and quantify the impor- and to compare with the output from models at the tance of the variability in atmospheric conditions relevant space and time. Table 3 provides a brief (within a year, along a trajectory, between years, overview of what types of meteorological data are or at different altitudes). The two case studies we available and for which spatial and temporal scales presented were examples of dynamic IBMs (DI, they would be most suitable. It is important to try to Table 1). This is a relatively large and diverse consider atmospheric conditions at the temporal and group of models with varying ranges of complexity. spatial scale most suitable for the ecological processes Depending on the aim and structure of the model, being studied (Fig. 1, see also Hallett et al. 2004). IBMs may (e.g. the white stork case study), or may (2) Model framework: Consider the aim of your not (e.g. the red knot case study), include interac- model and select the most suitable modeling frame- tions between individuals. IBMs may also include an work. When using models to integrate scientific aspect of heritability, where traits are transferred knowledge, there is a major distinction between the between generations and can evolve during a aim of making adequate (reliable, accurate and Table 3 An overview of the most relevant temporal scales (indicated by an X) for different types of meteorological data that can be incorporated into models of bird migration. Examples of on-line resources for such data are also provided Large eddy Regional numerical Station Global/continental Global circulation Temporal scale simulation mesoscale models observations reanalysis data indices Minutes X X – – – Hourly – X X X – Daily – X X X – Seasonal – – X X X a b c d On-line resource Generally none MM5 ECA&D NCEP reanalysis data NAO index The higher the spatial and temporal resolution of the data, generally the harder it is to find on the internet and such models must be run for the study of interest. PSU/NCAR mesoscale model (MM5); http://www.mmm.ucar.edu/prod/rt/pages/rt.html; Grell et al. 1994. ECA&D European climate and assessment dataset; http://eca.knmi.nl/; Klok and Klein Tank 2009. NCEP-NCAR reanalysis data; http://www.esrl.noaa.gov/psd/data/reanalysis/reanalysis.shtml; Kalnay et al. 1996. NAO (North Atlantic Oscillation) index; http://www.cgd.ucar.edu/cas/jhurrell/indices.html; Hurrell et al. 2003. 288 J. Shamoun-Baranes et al. precise, at the right scale) predictions and the aim of communication. Communication can be facilitated enhancing understanding. When the aim is to make by developing common terminology and using con- adequate predictions, it is generally desirable to ceptual frameworks for the description and design of match the resolution and extent of the model with models (e.g. Grimm et al. 2006; Nathan et al. 2008) the units and domain at which the predictions are as well as by research workshops dedicated to collab- required (‘scale of prediction’ for brevity), as well as oration (Bauer et al. 2009). gather data at the scale of prediction. In this way, errors due to mismatches of scale are avoided. If the Future perspectives scale at which the most important processes operate Several advances in different fields will strongly facil- does correspond with the scale of the prediction, try itate the integration of meteorology into migration to build a concept-driven model. However, if the research. First, meteorological data are becoming scale at which key processes operate is much finer more available and accessible, with numerous sources than the scale of prediction, it will be very hard (if of data freely available on the internet (Table 3). possible at all) to build a concept-driven model from Some of these sources are even archived globally expert knowledge and first principles. In that case a for several decades (e.g. NCEP-NCAR reanalysis data-driven model is the most suitable option. data, Kalnay et al. 1996). More recently, atmospheric In case the aim is to gain understanding about a models that can provide data at the temporal and certain aspect of a migration system, the question spatial scale of interest have been developed and becomes relevant whether you are in an explorative will greatly enhance migration research (Scott or a confirmative phase of your research. In the and Achtemeier 1987; Nathan et al. 2005). At very explorative phase, the aim is to identify patterns, fine scales this will often require that atmospheric attempt to find cause-and-effect relationships and models are run specifically for the research project compare alternative models. Currently it is not fea- (e.g. Shannon et al. 2002a, 2002b; Shamoun-Baranes sible to conduct such explorative activities with et al. 2003b, 2003c; Sapir 2009). Advances in tech- concept-based models because it requires too much nologies to collect data on animal movement effort to generate a single model. In the future, how- (e.g. Robinson et al. in press) will also facilitate the ever, more flexible modeling systems may be built integration of meteorology into migration research. that do, in fact, allow such activities (Taylor et al. Miniaturization of tracking technologies (e.g. 2007). So, data-driven models are the tools of choice Wikelski et al. 2006,2007; Stuchbury et al. 2009) in the explorative research phase. When a specific and collection of precise locations using the Global idea or hypothesis can be formulated, research Positioning system (GPS) improves the tracking of enters a confirmative phase where some sort of individual animals. High-resolution GPS may help formal comparison of that idea against observations revolutionize this field by providing detailed infor- or other ideas has to be made. In this phase, both mation on how animals respond to atmospheric concept- and data-driven models can be used dynamics en-route or even to help reveal animals’ effectively. decision rules. Furthermore, collecting additional Determining which type of model can be used data such as heart rate or 3-axial acceleration can under a given set of research aims and of constraints provide information on behavior and the expendi- with regard to availability of data (the variables that ture of energy (e.g. Ropert-Coudert and Wilson are available, as well as their resolution and extent), 2005; Rutz and Hays 2009) in relation to atmo- cannot be answered in general; it depends a lot on spheric conditions. In addition to individual track- the precise nature of a research question. However, ing, radar is an excellent tool for observing long-term Table 1 specifies the main properties and limitations spatial and temporal patterns of migration at specific of the various modeling techniques and can provide locations and has been used for several decades guidance about which modeling technique to use, in studies of the migration of birds (e.g. Erni et al. after a researcher has specified her/his research ques- 2005; van Belle et al. 2007; Schmaljohann et al. tion. Table 2 provides some examples of studies of migration with a reference to the different types of 2009), bats (e.g. Kunz et al. 2008) and insects (e.g. Chapman et al. 2003; Reynolds et al. 2005;). Weather model applied. (3) Communication and collaboration: During radar networks are particularly promising as they this process and the research itself, consider commu- provide multiple stations and can potentially help nicating and collaborating with the necessary experts, study larger-scale patterns and trajectories more e.g. modelers or meteorologists. Keep in mind that effectively then single locations. Although weather- the models themselves are also vehicles for surveillance Doppler radar has been used in the Integrating meteorology and migration research 289 Bauer S, van Dinther M, Høgda K-A, Klaassen M, Madsen J. United States for such studies (e.g. Diehl et al. 2003; 2008. The consequences of climate-driven stop-over sites Gauthreaux et al. 2008; Westbrook 2008), only changes on migration schedules and fitness of Arctic recently have several radars in the OPERA network geese. J Anim Ecol 77:654–60. in Europe (Operational Programme for the Exchange Bauer S, Barta Z, Ens BJ, Hays GC, McNamara JM, of Weather Radar Information; Kock et al. 2000) Klaassen M. 2009. Animal migration: linking models and been successfully tested for studying bird migration data beyond taxonomic limits. Biol Lett 5:433–5. (Holleman et al. 2008; van Gasteren et al. 2008; Bildstein KL. 2006. Migrating raptors of the world: their Doktor et al. in review) and it will become a valuable ecology and conservation. Ithaca, NY: Cornell University resource when studying Palearctic migration systems. Press. Atmospheric dynamics can affect migration sys- Bildstein KL, Zalles JI. 2005. Old world versus new world tems at many different levels, from instantaneous long-distance migration in accipiters, buteos, and falcons. changes in flight speed and direction to influencing In: Greenberg R, Marra PP, editors. The interplay of migra- breeding success. We hope to see meteorology more tion ability and global biogeography. Johns Hopkins strongly integrated into future research on migration University Press. p. 154–67. Birds of two worlds: the ecol- ogy and evolution of migration. Baltimore, MD. across multiple taxa. Such an interdisciplinary approach will help advance research on migration Bowlin M, et al. In press. Grand challenges in migration biology. Integr Comp Biol. as well as address some of the grand challenges in organismal biology (Schwenk et al. 2009; Bowlin Brandes D, Ombalski DW. 2004. Modeling raptor migration pathways using a fluid-flow analogy. J Raptor Research et al. this issue). 38:195–207. Brattstro ¨ m O, Kjelle ´ n N, Alerstam T, Akesson S. 2008. Effects Acknowledgments of wind and weather on red admiral, Vanessa atalanta, The authors thank M. Bowlin, I. A. Bisson and migration at a coastal site in southern Sweden. Anim Behav 76:335–44. M. Wikelski for organizing, and inviting J.S.B. to give a talk at the Integrative Migration Biology sym- Bruderer B, Underhill LG, Liechti F. 1995. Altitude choice by night migrants in a desert area predicted by meteorological posium at the 2010 Society for Integrative and factors. Ibis 137:44–55. Comparative Biology meeting in Seattle, Butler RW, Williams TD, Warnock N, Bishop M. 1997. Wind Washington. SICB’s Divisions of Animal Behavior, assistance a requirement for migration of shorebirds? Auk Neurobiology, and Comparative Endocrinology all 114:456–66. donated money to the symposium. The authors Chapman JW, Reynolds DR, Smith AD. 2003. Vertical-look- thank J. Leyrer as well as two anonymous reviewers ing radar: a new tool for monitoring high-altitude insect for discussions and constructive feedback on an ear- migration. Bioscience 53:503–11. lier version of the manuscript. They thank M. Chapman JW, Nesbit RL, Burgin LE, Reynolds DR, Duyvendak, for retrieving articles we did not have Smith AD, Middleton DR, Hill JK. 2010. Flight orientation direct access to. Our migration studies are facilitated behaviors promote optimal migration trajectories in high- by the BiG Grid infrastructure for eScience (http:// flying insects. Science 327:682–5. www.biggrid.nl). Chapman JW, Reynolds DR, Mouritsen H, Hill JK, Riley JR, Sivell D, Smith AD, Woiwod IP. 2008. Wind selection and drift compensation optimize migratory pathways in a high- References flying moth. Curr Biol 18:514–8. Able KP. 1973. The role of weather variables and flight direc- Cryan PM, Brown AC. 2007. Migration of bats past a remote tion in determining the magnitude of nocturnal bird island offers clues toward the problem of bat fatalities at migration. Ecology 54:1031–41. wind turbines. Biol Conserv 139:1–11. Akesson S, Hedenstro ¨ m A. 2000. Wind selectivity of migra- ¨ ¨ Danhardt J, LindstromA. 2001. Optimal departure decisions tory flight departures in birds. Behav Ecol Sociobiol of songbirds from an experimental stopover site and the 47:140–4. significance of weather. Anim Behav 62:235–43. Akesson S, Walinder G, Karlsson L, Ehnbom S. 2002. Diehl RH, Larkin RP, Black JE. 2003. Radar observations Nocturnal migratory flight initiation in reed warblers of bird migration over the Great Lakes. Auk 120: Acrocephalus scirpaceus: effect of wind on orientation and 278–90. timing of migration. J Avian Biol 33:349–57. Dingle H. 1996. Migration: the biology of life on the move. Allen MJ. 2006. Updraft Model for Development of New York: Oxford University Press. Autonomous Soaring Uninhabited Air Vehicles. Forty Drake VA, Farrow RA. 1988. The influence of atmospheric Fourth AIAA Aerospace Sciences Meeting and Exhibit; structure and motions on insect migration. Ann Rev Reno, Nevada. American Institute of Aeronautics and Entomol 33:183–210. Astronautics. 290 J. Shamoun-Baranes et al. Erni B, Liechti F, Bruderer B. 2003. How does a first Kerlinger P. 1989. Flight strategies of migrating hawks. year passerine migrant find its way? Simulating Chicago: The University of Chicago Press. migration mechanisms and behavioural adaptations. Kock K, Leitner T, Randeu WL, Divjak M, Schreiber KJ. 2000. Oikos 103:333. OPERA: Operational Programme for the Exchange of Erni B, Liechti F, Bruderer B. 2005. The role of wind in Weather Radar Information. First results and outlook for passerine autumn migration between Europe and Africa. the future, Vol. 25. Phys Chem Earth B: Hydrol Oceans Behav Ecol 16:732–40. Atmosphere. p. 1147–51. Erni B, Liechti F, Underhill LG, Bruderer B. 2002. Wind and Klok EJ, Klein Tank AMG. 2009. Updated and extended rain govern the intensity of nocturnal bird migration in European dataset of daily climate observations. Int J central Europe – a log-linear regression analysis. Ardea Climatol 29:1182–91. 90:155–66. Kunz TH, et al. 2008. Aeroecology: probing and modeling the Garland MS, Davis AK. 2002. An examination of Monarch aerosphere. Integr Comp Biol 48:1–11. Butterfly (Danaus plexippus) autumn migration in coastal Leshem Y, YomTov Y. 1996a. The use of thermals by soaring Virginia. American Midland Naturalist 147:170–4. migrants. Ibis 138:667–74. Gauthreaux SA Jr, Livingston JW, Belser CG. 2008. Detection Leshem Y, YomTov Y. 1996b. The magnitude and timing of and discrimination of fauna in the aerosphere using migration by soaring raptors, pelicans and storks over Doppler weather surveillance radar. Integr Comp Biol Israel. Ibis 138:188–203. 48:12–23. Leshem Y, YomTov Y. 1998. Routes of migrating soaring Gibo DL. 1981. Some observations on soaring flight in the birds. Ibis 140:41–52. Mourning Cloak Butterfly (Nymphalis antiopa L.) in south- Leskinen M, Markkula I, Koistinen J, Pylkko ¨ P, Ooperi S, ern Ontario. J New York Entomol S 89:98–101. Siljamo P, Ojanen H, Raiskio S, Tiilikkala K. 2009. Pest Gibo DL, Pallett MJ. 1979. Soaring flight of monarch butter- insect immigration warning by an atmospheric dispersion flies, Danaus plexippus (Lepidoptera: Danaidae), during late model, weather radars and traps. J Appl Entomol published summer migration in southern Ontario. Can J Zool online (doi: 10.1111/j.1439-0418.2009.01480.x). 57:1393–401. Leyrer J, Bocher P, Robin F, Delaporte P, Goulevent C, Gill RE, Tibbitts TL, Douglas DC, Handel CM, Mulcahy DM, Joyeux E, Meunier F, Piersma T. 2009. Northward migra- Gottschalck JC, Warnock N, McCaffery BJ, Battley PF, tion of Afro-Siberian Knots Calidris canutus canutus: High Piersma T. 2009. Extreme endurance flights by landbirds variability in Red Knots numbers visiting stopover sites on crossing the Pacific Ocean: ecological corridor rather than French Atlantic coast (1979–2009). Wader Study Group barrier? Proc R Soc B 276:447–57. Bull 116:145–51. Grell G, Dudhia J, Stauffer D. 1994. A description of the fifth- Liechti F. 2006. Birds: blowin’ by the wind? J Ornithol generation Penn State/NCAR Mesoscale Model (MM5). 147:202–11. NCAR Technical Note. NCAR. p. 117. Mandel JT, Bildstein KL, Bohrer G, Winkler DW. 2008. Grimm V, et al. 2006. A standard protocol for describing Movement ecology of migration in turkey vultures. Proc individual-based and agent-based models. Ecol Model Natl Acad Sci USA 105:19102–7. 198:115–26. Nathan R, Getz WM, Revilla E, Holyoak M, Kadmon R, Hallett TB, Coulson T, Pilkington JG, Clutton-Brock TH, Saltz D, Smouse PE. 2008. A movement ecology paradigm Pemberton JM, Grenfell BT. 2004. Why large-scale climate for unifying organismal movement research. Proc Natl indices seem to predict ecological processes better than Acad Sci USA 105:19052–9. local weather. Nature 430:71–5. Nathan R, et al. 2005. Long-distance biological transport pro- Holleman I, van Gasteren H, Bouten W. 2008. Quality assess- cesses through the air: can nature’s complexity be unfolded ment of weather radar wind profiles during bird migration. in silico? Divers Distrib 11:131–7. J Atmospheric Oceanic Technol 25:2188–98. Newton I. 2006. Can conditions experienced during migration Hu ¨ ppop O, Hu ¨ ppop K. 2003. North Atlantic Oscillation limit the population levels of birds? J Ornithol 147:146–66. and timing of spring migration in birds. Proc R Soc B Newton I. 2008. The migration ecology of birds. Oxford: 270:233–40. Academic Press. Hurrell J, Kushnir Y, Ottersen G, Visbek M. 2003. An over- Pennycuick CJ, Alerstam T, Larsson B. 1979. Soaring migra- view of North Atlantic Oscillation. In: Hurrell J, Kushnir Y, tion of the common crane Grus grus observed by radar and Ottersen G, visbek M, editors. The North Atlantic from an aircraft. Ornis Scand 10:241–51. Oscillation: climate significance and environmental impacts. Washington, DC: American Geophysical Union. ˚ Piersma T, Lindstro ¨mA. 1997. Rapid reversible changes in p. 1–35. organ size as a component of adaptive behaviour. Trends Jonzen N, et al. 2006. Rapid advance of spring arrival Ecol Evol 12:134–8. dates in long-distance migratory birds. Science Piersma T, van de Sant S. 1992. Pattern and predictability of 312:1959–61. potential wind assistance for waders and geese migrating Kalnay E, et al. 1996. The NCEP/NCAR 40-year reanalysis from West Africa and the Wadden Sea to Siberia. Ornis project. Bull Am Meteorol Soc 77:437–71. Svecica 2:55–66. Integrating meteorology and migration research 291 Piersma T, Prokosch P, Bredin D. 1992. The migration system weather at departure sites, onset of migration and timing of Afro-Siberian knots Calidris canutus canutus. Wader of soaring-bird autumn migration in Israel? Global Ecol Study Group Bull 64(Suppl):52–63. Biogeogr 15:541–52. Reilly JR, Reilly RJ. 2009. Bet-hedging and the orientation Shamoun-Baranes J, Baharad A, Alpert P, Berthold P, of juvenile passerines in fall migration. J Anim Ecol YomTov Y, Dvir Y, Leshem Y. 2003a. The effect of wind, 78:990–1001. season and latitude on the migration speed of white storks Ciconia ciconia, along the eastern migration route. J Avian Reynolds AM, Reynolds DR, Riley JR. 2009. Does a ‘turbo- Biol 34:97–104. phoretic’ effect account for layer concentrations of insects migrating in the stable night-time atmosphere? J R Soc Shamoun-Baranes J, Leyrer J, van Loon E, Bocher P, Robin F, Interface 6:87–95. Meunier F, Piersma T. 2010. Stochastic atmospheric assis- tance and the use of emergency staging sites by migrants. Reynolds DR, Chapman JW, Edwards AS, Smith AD, Proc R Soc B published online (doi: 10.1098/ Wood CR, Barlow JF, Woiwod IP. 2005. Radar studies rspb.2009.2112). of the vertical distribution of insects migrating over southern Britain: the influence of temperature inversions Shannon HD, Young GS, Yates MA, Fuller MR, Seegar WS. on nocturnal layer concentrations. B Entomol Res 2002a. American White Pelican soaring flight times and 95:259–74. altitudes relative to changes in thermal depth and intensity. Condor 104:679–83. Reynolds DS. 2006. Monitoring the potential impact of a wind development site on bats in the northeast. Shannon HD, Young GS, Yates MA, Fuller MR, Seegar WS. J Wildlife Manag 70:1219–27. 2002b. Measurements of thermal updraft intensity over complex terrain using American White Pelicans and a Richardson WJ. 1978. Timing and amount of bird migration simple boundary-layer forecast model. Bound-Lay in relation to weather: a review. Oikos 30:224–72. Meteorol 104:167–99. Richardson WJ. 1990. Timing of bird migration in relation to Spaar R, Bruderer B. 1997. Migration by flapping or soaring: weather: updated review. In: Gwinner E, editor. Bird migra- Flight strategies of Marsh, Montagu’s and Pallid Harriers in tion. Berlin: Springer-Verlag. p. 78–101. southern Israel. Condor 99:458–69. Robinson WD, Bowlin MS, Bisson I-A, Shamoun-Baranes J, Spaar R, Stark H, Liechti F. 1998. Migratory flight strategies Thorup K, Diehl RH, Kunz TH, Mabey S, Winkler DW. of Levant sparrowhawks: time or energy minimization? In press. Integrating concepts and technologies to advance Anim Behav 56:1185–1197. the study of bird migration. Front Ecol Environ. Srygley RB, Dudley R. 2008. Optimal strategies for insects Ropert-Coudert Y, Wilson RP. 2005. Trends and perspectives migrating in the flight boundary layer: mechanisms and in animal-attached remote sensing. Front Ecol Environ consequences. Integr Comp Biol 48:119–33. 3:437–44. Stefanescu C, Alarco ´nM, Avila A. 2007. Migration of the Rutz C, Hays GC. 2009. New frontiers in biologging science. painted lady butterfly, Vanessa cardui, to north-eastern Biol Lett 5:289–92. Spain is aided by African wind currents. J Anim Ecol Sapir N. 2009. The effect of weather on Bee-eater (Merops 76:888–98. apiaster) migration. Jerusalem: Hebrew University. Stocker S, Weihs D. 1998. Bird migration – an energy-based Schaub M, Liechti F, Jenni L. 2004. Departure of migrating analysis of costs and benefits. IMA J Math Appl Med European robins, Erithacus rubecula, from a stopover site in 15:65–85. relation to wind and rain. Anim Behav 67:229–37. Stoddard PK, Marsden JE, Williams TC. 1983. Computer Schmaljohann H, Liechti F, Bruderer B. 2009. Trans-Sahara simulation of autumnal bird migration over the western migrants select flight altitudes to minimize energy costs North Atlantic. Anim Behav 31:173–80. rather than water loss. Behav Ecol Sociobiol 63:1609–19. Stutchbury BJM, Tarof SA, Done T, Gow E, Kramer PM, Tautin J, Schwenk K, Padilla DK, Bakken GS, Full RJ. 2009. Grand Fox JW, Afanasyev V. 2009. Tracking long-distance songbird challenges in organismal biology. Integr Comp Biol migration by using geolocators. Science 323:896. 49:7–14. Taylor IJ, Deelman E, Gannon DB, Shields M, editors . 2007. Scott RW, Achtemeier GL. 1987. Estimating pathways of Workflows for e-Science. London: Springer-Verlag. migrating insects carried in atmospheric winds. Environ Thorup K, Alerstam T, Hake M, Kjellen N. 2003. Bird orien- Entomol 16:1244–54. tation: compensation for wind drift in migrating raptors Shamoun-Baranes J, Leshem Y, Yom-Tov Y, Liechti O. 2003b. is age dependent. Biol Lett 270:8–11. Differential use of thermal convection by soaring birds over van Belle J, Shamoun-Baranes J, van Loon E, Bouten W. 2007. central Israel. Condor 105:208–18. An operational model predicting autumn bird migration Shamoun-Baranes J, Liechti O, Yom-Tov Y, Leshem Y. 2003c. intensities for flight safety. J Appl Ecol 44:864–74. Using a convection model to predict altitudes of white van de Kam J, Ens B, Piersma T, Zwarts L. 2004. Shorebirds: an stork migration over central Israel. Bound-Lay Meteorol illustrated behavioural ecology. Utrecht: KNNV Publishing. 107:673–81. van Gasteren H, Holleman I, Bouten W, Van Loon E, Shamoun-Baranes J, van Loon E, Alon D, Alpert P, Shamoun-Baranes J. 2008. Extracting bird migration Yom-Tov Y, Leshem Y. 2006. Is there a connection between 292 J. Shamoun-Baranes et al. information from C-band Doppler weather radars. Ibis Wikelski M, Moskowitz D, Adelman JS, Cochran J, 150:674–86. Wilcove DS, May ML. 2006. Simple rules guide dragonfly migration. Biol Lett 2:325–9. Vrugt JA, van Belle J, Bouten W. 2007. Pareto front analysis of flight time and energy use in long-distance bird migra- Wood CR, Chapman JW, Reynolds DR, Barlow JF, Smith AD, tion. J Avian Biol 38:432–42. Woiwod IP. 2006. The influence of the atmospheric bound- ary layer on nocturnal layers of noctuids and other Weber TP, Alerstam T, Hedenstro ¨ m A. 1998. Stopover deci- moths migrating over southern Britain. Int J Biometeorol sions under wind influence. J Avian Biol 29:552–60. 50:193–204. Westbrook JK. 2008. Noctuid migration in Texas within the Wood CR, Clark SJ, Barlow JF, Chapman JW. 2010. Layers nocturnal aeroecological boundary layer. Integr Comp Biol of nocturnal insect migrants at high-altitude: the influence 48:99–106. of atmospheric conditions on their formation. Agric For Wikelski M, Kays RW, Kasdin NJ, Thorup K, Smith JA, Entomol 12:113–21. Swenson GW Jr. 2007. Going wild: what a global small-animal tracking system could do for experimental biologists. J Exp Biol 210:181–6.

Journal

Integrative and Comparative BiologyPubmed Central

Published: Apr 8, 2010

There are no references for this article.