{"paper_id":"d812b20b-d3e1-47ef-9f94-6ef421d7c63c","body_text":"Windy weather drives social structure in wild zebra finches | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Windy weather drives social structure in wild zebra finches Chris Tyson, Hugo Loning, Noëlle Tschirren, Elke Molenaar, Lysanne Snijders, and 2 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-6502285/v1 This work is licensed under a CC BY 4.0 License Status: Published Journal Publication published 21 Jan, 2026 Read the published version in Movement Ecology → Version 1 posted 9 You are reading this latest preprint version Abstract Background Social connections may provide individuals with multiple benefits. Individuals, however, are often constrained in how they socially organize due to ecological and environmental factors that affect individual space-use and movement patterns. Weather is one such factor that influences individual movements, and thus social structure. While on longer time scales (i.e., seasonally) the impacts of weather are relatively predictable, on shorter time scales (i.e., sub-daily), the impacts of weather on social organization are less predictable yet are largely overlooked. Methods In this study, we examined the influence of short-term weather components, specifically wind and temperature, on the social structure of free-living zebra finches ( Taeniopygia castanotis ) in the Australian arid zone. Our goal was to characterize if social network structure was impacted by hourly changes in these important weather components. To do so, we used an automated radio telemetry system to concurrently track 128 wild zebra finches for 12 consecutive days in the Australian spring in order to examine the relationships between weather components. Using Bayesian network analyses to account for the uncertainty in association strengths among individuals, we examined network structure, as measured by density and modularity, in relation to hourly wind speed, temperature, and time of day. Additionally, to assess if weather impacted the synchronization of group-level movements, we calculated proximity ratios between neighbouring individuals, which we related to wind and temperature. Results We observed that network modularity increased during hours with higher mean wind speed and was highest in the morning and evening hours. In contrast, network density was not related to wind speed. Additionally, neither modularity nor density showed a significant relationship with the moderate temperatures during the tracking period. Group-level movement patterns as measured by proximity ratios between neighbouring individuals showed no relationship with either wind or temperature. Conclusions Our results suggest that short-term changes in wind impact social structure in wild zebra finches. Given the critical role that network modularity plays in social information transfer, increased wind could have significant downstream consequences. Stochastic and more frequent changes in weather due to climate change could thus significantly impact nomadic species that rely on locating ephemeral resources. Social organization wind modularity zebra finch automated radio-tracking Figures Figure 1 Figure 2 Figure 3 Background Among social animals there is large variation in how individuals form groups, which has consequences for multiple key behavioral, ecological and evolutionary processes. At a group level, social network structure may impact predator avoidance, population stability, dispersal, and social evolution ( 1 – 4 ). At an individual level, there is large variation in the quantity and quality of relationships an individual has with conspecifics ( 5 ). Therefore, an individual’s social network position may affect its access to mating opportunities, social support, information about novel resources such as ephemeral food patches, or disease infection ( 6 – 8 ). Social network position can thus have significant fitness implications. Consequently, understanding the drivers of social network variation and the relative influence of the various factors that affect social organization is a major focus of contemporary animal behavior research ( 9 – 12 ). Climate and weather are pervasive influences on social structure in wild populations ( 10 ). At a broad scale, seasonal changes in climate drive shifts in resource distributions, which in turn induces cyclical changes in the spatial organization of numerous species, particularly in the temperate and polar regions ( 13 ). Collective migrations are conspicuous examples of such annual fluctuations in social structure and may be linked with Behavioural changes, such as increased sociality ( 14 , 15 ). In many species, these fluctuations occur annually and as such are regularly experienced. Extreme weather events may also induce changes in social structure more locally and on a more rapid time scale with these changes being either ephemeral or enduring ( 16 , 17 ). For example, hurricanes and cyclones can induce changes in group structure in social primates that persist for months ( 18 , 19 ). While less dramatic, fluctuations in daily weather components may also impact the structure of animal groups over short time scales. In particular, wind and temperature are two primary weather components that may have especially large effects on avian species by increasing the energetic costs of flight ( 20 , 21 ) and thus potentially altering movement and social connectivity. Wind is a critical element for birds that shapes not only long-distance movements in migratory species ( 22 , 23 ), but also affects foraging efficiency by modulating energy expenditure ( 24 , 25 ). For birds with flapping flight, energetic costs are expected to be lowest at intermediate wind speeds ( 26 ) and consequently extreme wind speeds may affect mobility and social interactions. Temperature also shapes energetic landscapes for birds and thus directly impacts locomotion costs as well ( 27 ). On short time scales, temperature may potentially limit the flexibility of social organization and buffering by increasing movement costs and limiting individuals from associating ( 28 ). Alternatively, higher temperatures may lead to increased association rates by constraining individuals to remain near particular resources, such as shade or water ( 29 ). As such, together both weather components may have significant impacts on social network structure in complimentary or opposing directions. In this study, we examine the influence of two primary weather components, wind and temperature, on social network structure and space-use patterns in wild zebra finches ( Taeniopygia castanosis ) over short time scales. Zebra finches live in dynamic, multi-level societies with mated pairs staying almost constantly together over space and time ( 30 ), often form small cohorts of up to 10 individuals ( 31 ) and aggregate in larger groups at resources, social hotspots ( 32 ) or when joining long distance movements during dispersal ( 33 ). The social network of wild zebra finches can extend across years, and relates to reproductive synchrony with others ( 34 ), potentially buffering a pair against the risks of predation during breeding. Zebra finches are native to the Australian arid zone which is characterized by increasingly extreme short- and long-term unpredictability in climate conditions ( 35 , 36 ). Since social connectivity in such dynamic multi-level societies requires active movement of individuals, these social systems provide prime examples to determine if weather and climate change affect social organization. To assess the impact of wind and moderate temperature on social network structure in zebra finches, we used automated radio tracking to concurrently track 128 adult zebra finches and determined how social structure and movement patterns varied in relation to wind, temperature, and time of day. We predicted that birds would fly less during windy periods and would cluster in protected sites, such as dense vegetation, which would result in smaller social groups. During high temperatures (> 35°C), zebra finches reduce activity and remain close to water sources ( 37 ), yet those temperatures are rare during the period in which we conducted our study, and so we expected temperature to have less of an effect than wind on social organization. Zebra finches are typically seen in larger groups during the day ( 33 ) and so we expected to observe decreased social cohesion in the morning and evening. To quantify these changes in social organization, we measured two social network metrics, density and modularity, both of which are linked with processes such as disease and information transfer ( 38 ). We also examined social organization at a broad level by assessing the movement similarity between neighboring individuals, which we predicted would be less cohesive during windier and warmer periods. Methods Study site & automated radio tracking Our study was conducted at Gap Hills within the Fowlers Gap Arid Zone Research Station, western New South Wales, Australia (30°56′58″ S, 141°46′02″ E) in September 2022. The area primarily consists of open Acacia shrubland surrounding a water retention basin (200 x 150 m) and an ephemeral creek system (Fig. 1 A-B). We caught zebra finches using mist nets at three sub-sites (Fig. 1 A). Catching began at sunrise and nets were checked every 15 minutes. We marked all captured birds with an individually numbered metal band from the Australian Bird and Bat Banding Scheme. Birds were then weighed using a 20-g Pesola scale and birds above 10 g were outfitted with a 0.48 g solar-powered radio tag with a nitinol antenna (LifeTag, Cellular Tracking Technologies; New Jersey, USA), which was attached using a stretchable nylon leg-loop harness (Jirinec et al., 2021; Tyson, Loning, et al., 2024; Fig. 1 C). The mean mass of tagged birds was 12.7 g (range 10.4–15.6 g). After tagging, all birds were observed to fly off similarly to untagged individuals and tagged individuals were routinely resighted. Bird handling and tagging was performed in accordance with the Authority (#2018/027) of the Animal Ethics Committee at Macquarie University, and the Australian Bird & Bat Banding Scheme. In total, we tagged 128 zebra finches (53 females, 60 males, and 16 juveniles) from which we obtained tracking data for 12 consecutive days (19 September to 30 September). During this period, birds were tracked for an average of 10.5 days. Variation in tracking duration was due to antennas of the tags breaking after which point the tags were excluded from the analysis. Broken antennas were clearly identifiable by a large change in the signal strength as well as the number of receivers detecting the tag. Tracking began at the start of the period when we annually monitor nest boxes for breeding activity, and we discontinued tracking after 12 days due to the tag malfunctions. During the tracking period, 34% of nest boxes (62 out of 180) had active nests though the breeding status of the birds we tracked was unknown. Tags were programmed to signal every 5 seconds when exposed to sufficient solar energy. During the study period, tags were detected during daylight hours, i.e., between ~ 0600 and ~ 1800 Australian Central Standard Time (ACST), but gaps in detections also occurred when birds entered nest boxes or dense vegetation. On average, localization estimates were obtained every 3 minutes. To track tagged birds, we used an automated radio tracking system which consisted of 93 receivers (Node v2, Cellular Tracking Technologies, New Jersey, USA) and one base station (SensorStation, Cellular Tracking Technologies, New Jersey, USA). Each receiver recorded the tag ID, time of detection, and signal strength, which was relayed to the base station. Receivers were mounted ~ 1.5 m above ground on a steel fence post. Most receivers (n = 73) were spaced 150 m apart on a triangular grid with the remainder placed opportunistically within the three nest box sub-sites (Fig. 1 A). The central base station was placed centrally on top of the water retention basin (Fig. 1 A) and was connected to four 430 MHz Yagi antennas oriented perpendicularly and one 430 MHz omnidirectional antenna oriented vertically, all of which were mounted to the top of a 6 m mast (Fig. 1 D). We used these multiple potentially redundant antennas to minimize the possibility of missed detections, which can greatly impact localization accuracy ( 40 ). Radio tag localization We estimated tag locations using radio signal strength-based (RSS) multilateration ( 41 ). To do so, we first calibrated the signal strength to distance relationship by placing six tags at set distances (1, 2, 5, 10, 15, 25, 50 ,75, 100, 150 m) from four receivers. Tags were mounted horizontally on a cardboard strip that was placed atop a 2 m PVC pole. Tags were held stationary at each distance for two minutes. The exponential decay relationship between the average RSS of detections from each calibration distance was then modelled using the equation: RSS ∼ a × exp(− S × distance) + K; where a is the intercept, S is the decay factor, and K is the horizontal asymptote ( 41 ). Initial parameter values were estimated using the self-starting function ‘ SSasymp,’ which were used to fit an exponential decay model using the function ‘nls’ ( 42 ). The parameter estimates from this model were then used to fit the final exponential decay model ( 41 ). To estimate distance for multilateration, RSS values from each tag at each node were averaged within a 15 second moving-window with a 5 second lag to remove large fluctuations in signal strength which can occur due to various sources of interference ( 30 ). The mean RSS value within each interval at each node was then used to estimate distance between the tag and the corresponding node using the exponential decay relationship given above. In each 15 second interval, nodes within 200 m of the receiver with the strongest average RSS value in that interval were retained and were used to estimate the tag location using RSS-based multilateration. This method uses the estimated distance between a tag and each node (minimally three) to estimate the location of the tag, which can be found using a non-linear least squares model which minimizes the differences between pairwise distances between all nodes and the estimated distance between the tag and each node ( 41 ). The coordinates of the node with the strongest signal were used as initial values for the model. Multiple factors aside from distance will influence the RSS of a tag. Such factors include the relative orientation of the tag and node, physical obstructions between the tag and receiver, and changes to the innate signal strength of the tag over time ( 43 ). As such, a given RSS value will correspond to a range of possible distances between the tag and the receiver, which will increase as the signal strength decreases. To partially account for this variation in the localization estimates, the distance between a tag and a receiver in each interval was estimated 100 times by sampling from ± 1 SE around the mean distance for the mean RSS-value in that interval. These 100 samples from each interval were then averaged to determine a point estimate for that interval. We also calculated an error ellipse corresponding to the \\(\\:\\sqrt{2}\\) -sigma ellipse of a bivariate normal distribution to estimate the uncertainty around each location. The median error (i.e., the difference between the estimated location and the true location) from field calibration tests was 35 m. Weather data Weather data was provided by the Australian Bureau of Meteorology ( http://www.bom.gov.au/nsw/ ) from the Fowler’s Gap Automatic Weather Station (station number 046128), which is 15 km from where birds were tagged at Gap Hills. Data were provided at minute intervals from which hourly average temperature (°C) and wind speed (km h − 1 ) were calculated. Social network analyses We conducted proximity based social network analyses to examine the effects of hourly weather conditions on association rates between zebra finches. For this analysis, we removed localizations where the strongest RSS detected by a receiver was less than − 80 dB since during this interval the tag was not likely detected by the nearest receiver, which increases the localization error ( 40 ). As such, by removing localizations based on weak RSS values, we retained points with less uncertainty. We considered two individuals to be associated when they were within 60 m of one another during the same localization interval. This cutoff is based on previous findings from the same tracking system that social partners, which maintain continual contact, had a median separation distance of 60 m ( 30 ). Distances between simultaneously localized dyads were calculated using the R package spatsoc ( 44 ). Within each hour interval between 0600 and 1800, we then calculated the number of associations between each pair of tracked birds and the total number of possible associations between the pair (i.e., the total number of intervals when both individuals were simultaneously localized). We removed hour intervals with fewer than 100 dyads (9 out of 145), which left 136 hour-intervals across 12 days. Hourly association rates were analyzed using Bayesian edge-weighted social networks. This method captures the uncertainty in edge weights (i.e., the association strength) between individuals which can arise when all individuals are not uniformly observed ( 45 ). In this framework, the association strength between dyads with a higher sampling effort in each hour will have less uncertainty (i.e., a narrower distribution of edge weights) whereas association strength for dyads with lower sampling effort will have more uncertainty ( 46 ). To quantify edge weight uncertainty of the hourly social networks, we used the R package “bisonR” ( 47 ) and fit one edge weight model per hour interval. In each hour, an edge weight model is constructed to yield a posterior distribution of possible networks with edge weights that vary based on the sampling effort (i.e., the number of observations) for each dyad. We fit edge weight models where association strength was measured by the number of associations between two individuals (the numerator) relative to the total number of times both individuals were simultaneously observed (the denominator). We used the ‘fit_edge’ function in bisonR to fit a count conjugate model with a weakly informative gamma prior for the edge component of the model. Sampling from within this distribution allows for the uncertainty in the ‘real’ network to be quantified and then incorporated into the calculation of network metrics and Bayesian models ( 45 ). From each edge weight model, 400 samples were extracted from the posterior distribution. In bisonR, networks are fully connected since all potential edges have a non-zero probability ( 45 ). To calculate network metrics, however, it is necessary to separate dyads that were unlikely to have interacted from those that did. As such, we set edge weights below the minimum observed edge weight to zero ( 46 ). From these observed networks, we then calculated two metrics: density and modularity. Density is a proportional measure of the number of edges relative to the total number of possible edges (see e.g., Evans et al. 2021), which was calculated using the function ‘extract_metric’ from bisonR. Modularity is a measure of community structure that expresses the relative density of edges within communities with respect to edges outside communities ( 48 ), which we calculated using the functions ‘cluster_louvain’ and ‘modularity’ from the R package igraph ( 49 ). We then constructed Bayesian linear models to assess the relationship between the network density and modularity, hour of the day, and the two-way interaction between mean temperature and mean wind speed. Hour was analyzed as a second-order polynomial as zebra finch social activity is typically highest in the morning and evening ( 33 ). All predictors were z-transformed and for each we set generic weakly informative normal priors with a mean of zero and a standard deviation of one. To incorporate uncertainty in the analysis of each network metric, we created a dataset from each of the 400 random draws extracted from the posterior distribution of each hourly network. As such, each dataset consisted of 136 values: one for each network metric each day-hour interval. In each model, we set max_treedepth = 20 and iter = 10,000. We then used the function ‘brms_multiple’ from the R package ‘brms’ to run one model for each dataset, 400 in total ( 50 ). Model convergence was checked by confirming that Rhat in each model was less than 1.05, and additionally, we conducted posterior predictive checks for a random set of 20 models to confirm that the model generated data and the observed data were comparable ( 51 ). Ultimately, all models were then combined with the function ‘combine_models_c’ in ‘brms’ ( 50 ). If the 95% confidence intervals of the estimated effects encompassed zero, then the data was taken to not strongly support an effect of the covariate on the response. Proximity analysis We additionally assessed the influence of weather on broad scale movement patterns by examining the proximity of neighboring individuals in relation to temperature and wind. To do so, we first fit continuous time movement models to the localization data from each individual using the R package ‘ctmm’ ( 52 ). This method can accommodate irregularly sampled data and can incorporate uncertainty in localization estimates by using the error ellipse information calculated through the repeated multilateration approach described above ( 52 , 53 ). After fitting a suite of continuous time movement models, the top-ranked model based on Akaike’s information criteria adjusted for small samples sizes (AICc) was selected. We then segmented the tracking data into hour intervals each day for each individual. Using these hourly subsets and the corresponding movement models, we then compared the proximity between individuals captured within the same sub-site (i.e., nestbox sites B, D, and E). Specifically, we assessed whether two individuals were closer together or farther apart than expected given independent movement. We did so by calculating the proximity ratio between two individuals using the function ‘proximity’ from the R package ‘ctmm’ which compares the mean-square distance between two individuals to the expected distance if the two individuals were moving independently. Proximity ratios less than one indicate that two individuals are closer together than expected given independent movement and ratios greater than one indicate that individuals are farther apart, while proximity ratios overlapping with one indicate that the distance between two individuals is not significantly different given independent movement. We then assessed the relationship between the proximity ratio point estimates and time of day (hour), hourly mean temperature, hourly mean wind speed, and nestbox section with individual, partner, and day as cross random effects using a Bayesian mixed-effect model with generic weakly informative normal priors with a mean of zero and a standard deviation of one and a skewed t-distribution to account for skewness and the presence of outliers in the proximity ratio estimates. Results We tracked 128 zebra finches for an average of 10.5 days (range 1–12 days) per individual. Each radio tagged individual was localized 1,543 times on average across the 12-day tracking period. Across the tracking period, dyads (i.e., pair-wise associations) were simultaneously localized 98 times on average and were considered to be associated 8 times on average, though there was considerable range in the number of simultaneous localizations (0–4,286) and the number of associations (0–3,456). During the tracking period, hourly mean temperature ranged from 9°C to 28°C and hourly mean wind speed ranged from 1 km/h to 40 km/h. Between 2005 and 2021, the mean hourly temperature in September was 18°C (5th percentile: 11°C, 95th percentile: 24°C) and mean hourly wind speed was 18 km/h (5th percentile: 5 km/h, 95th percentile: 31 km/h). Network density and modularity both showed large variation across the ranges of weather component values that were recorded during the study period (Fig. 2 A and 2 B). Density, however, showed no association with time of day, hourly mean temperature, or hourly mean wind speed (Table 1 ). In contrast, predicted network modularity was significantly related to both time of day and mean wind speed, though not mean temperature (Table 1 ). Modularity was highest at dawn and dusk (estimate hour 2 : 0.38, 0.21–0.55; Fig. 2 B) and increased with mean wind speed (estimate wind speed: 0.03, 0.01–0.04; Fig. 2 B, Fig. 3 ). Table 1 Parameter estimates and 95% CI for Bayesian models of the relationship between time of day and weather components and network density and modularity. Terms in bold indicate parameters (not including the intercept) where the coefficient estimate 95% CI did not overlap zero. Density Modularity Predictors Estimates CI (95%) Estimates CI (95%) Intercept 0.11 0.10–0.11 0.52 0.51–0.54 Hour 0.08 -0.04–0.19 0.03 -0.23–0.30 Hour^2 0.02 -0.05–0.09 0.38 0.21–0.55 Wind speed 0.00 -0.01–0.01 0.03 0.01–0.04 Temperature -0.00 -0.01–0.01 0.01 -0.02–0.03 Wind speed * Temperature -0.00 -0.01–0.00 0.00 -0.01–0.02 Observations 136 136 R 2 Bayes 0.071 0.288 Proximity ratio confidence intervals did not overlap with one for 74%, 70%, and 63% of the comparisons between individuals from the same neighborhood, the sub-sites B, D, and E, respectively, indicating that individuals within the same area of the colony tended to follow similar movement patterns throughout the day. We did not, however, find evidence of a relationship between hour of day, hourly mean temperature, or hourly mean wind speed and proximity ratios (Table 2 ). Proximity ratios were higher in colony section E (estimate: 0.06, 0.04–0.09) compared to the other two sections. Table 2 Parameter estimates and 95% CI for Bayesian models of the relationship between time of day and weather components and proximity ratios of neighbouring individuals. Terms in bold indicate parameters (not including the intercept) where the coefficient estimate 95% CI did not overlap zero. Proximity Predictors Estimates CI (95%) Intercept (Section B) 0.05 0.04–0.07 Hour 0.01 0.00–0.01 Wind speed 0.00 0.00–0.00 Temperature -0.01 -0.01 – -0.01 Wind speed * Temperature 0.00 0.00–0.00 Section D 0.00 -0.02–0.02 Section E 0.06 0.04–0.09 Observations 48,062 Marginal R 2 / Conditional R 2 0.013 / 0.026 Discussion Using automated radio tracking of 128 wild zebra finches across 12 days, we show that social organization was affected by hourly changes in weather. During periods with higher wind, social network modularity increased, indicating stronger separation between sub-groups among the individuals we tracked. In contrast to wind speed, the moderate temperatures during our tracking period did not affect either social network modularity or density. Modularity, however, did change significantly across the day, with higher levels in the mornings and evenings. Together, these findings indicate that the social network structure of wild populations can vary significantly even on short (hourly) time scales, demonstrating substantial social plasticity in response to environmental fluctuations. These fluctuations in social plasticity, especially in network modularity could have relevant consequences for the effective transfer of social information on ephemeral resources and predation threats. Obtaining insights into the extent of social plasticity in wild animal societies is crucial for estimating their capacity to cope with environmental change through mechanisms such as social buffering ( 16 , 54 ). Additionally, knowing baseline social plasticity will help to detect when social structures move beyond their ‘normal’ range and thus when potentially concerning situations arise (Snijders et al., 2017). We observed within-day variation in modularity, which most likely reflects the different activities of zebra finches have during the day. Specifically, the higher modularity early and late in the day most likely results from individuals roosting in pairs or small subgroups at roosting colonies throughout the study area ( 33 ). During the day when not foraging, birds also spend substantial time at various social hotspots and at water sources where they occur in larger groups ( 32 ). These larger gatherings might be relevant for information transfer as individuals mix with others from different subsites. In contrast, we observed that social network density remained relatively constant throughout the day and was unaffected by either wind or temperature. Similarly, we observed that the movements of individuals caught in the same neighborhood were moderately correlated, as indicated by the proximity ratios, suggesting that groups of neighboring individuals were more cohesive than expected given random movement. Multiple factors may explain the observed increased social network modularity in relation to wind. For one, due to their small size, zebra finches experience greater flight costs from strong winds ( 55 ), which may limit movement and increase aggregations at sheltered locations. Additionally, increased wind speeds reduce flight maneuverability ( 56 ), limiting the ability to evade aerial predators, which may reduce the willingness to fly in higher winds. Conceivably, individuals carrying a radio tag might be further deterred from flying given the effects of carrying an increased load on flight performance ( 57 , 58 ). Yet, given that we have observed tagged individuals move around substantially (Tyson et al. 2024) and performing normal behaviors, such as nest building and nestling provisioning, and that we have detected birds more than 24 months after they were tagged, the impacts of the tags appear to be minimal and thus device-effects are unlikely to explain the patterns observed here. Another potential explanation for reduced movement during windier periods is that predator detection may be diminished due to attenuated acoustic cues ( 59 ), which may also limit willingness to fly. Additionally, both reduced flight maneuverability and diminished hearing ( 60 ) may also jeopardies pair cohesion, which is an important component of zebra finch social structure (30) as high wind speeds may make it challenging for zebra finch pairs to remain together and to reunite once separated. Given that zebra finch calls are only audible at distances up to about 14 m ( 61 ), and less in windy conditions, becoming separated may be a severe consequence of higher wind speeds, especially given the uncertainty as to how long windy conditions will persist. The moderate temperatures had no effect on social organization. Animals typically move less at extreme temperatures, seek shade at higher temperatures, and seek warmth and shelter at very low temperatures ( 62 ). While temperatures in the arid zone can well reach beyond 45 degrees in the shade in summer, the moderate temperature of only up to 29 degree C the birds experienced during tracking, most likely did not require any trade off movement and social organization. Taken together for the effects of wind, the higher costs in flying as well the potential threats from increased vulnerability to predators and difficulty in maintaining pair cohesion may explain the observed increase in social network modularity in wild zebra finches in response to higher wind speeds. With respect to the consequences of increased social network modularity, if this change results from individuals reducing connectivity to unfamiliar individuals, it may limit exposure to novel sources of information and thus limit social information transfer ( 38 , 63 ). Moreover, higher modularity could impact collective decisions such as when to initiate nomadic movements in search of favorable conditions or synchronizing breeding activity ( 64 ). Both activities could have potentially significant population impacts. Modularity indeed plays a key role in social transmission, both of information and of infections, although in different (non-linear) ways. Strong modularity is predicted to ‘trap’ simple disease infections into subgroups, but depending on the learning strategy of individuals, modularity does not necessarily affect social transmission of information ( 38 ). Recent modelling work by Evans et al. (2021) demonstrated that highly fragmented structures with small group sizes, such as fission-fusion populations and multi-level societies, e.g., wild zebra finch populations, can be most effective in trading-off reduced infection risk with effective social information transfer. In their modelling, additional variation in modularity did not affect the speed of information transfer when considering relatively higher network densities ( 38 ). However, Evans et al (2021), focused on complex information transfer mechanisms (e.g. copying the majority), while time-sensitive fitness-relevant information (e.g., detection of predators or water and food in resource-scarce environments) is more likely to transfer through knowledgeable individuals rather than the majority. In these cases, an increase from intermediate modularity to higher modularity, such as in our study, could hamper information flow, especially in structures with small subgroups ( 65 ). Indeed, transfer of information in theoretical and empirical datasets was revealed most efficient at intermediate modularity levels ( 65 ). Thus, when considering urgent information about the presence of predators or resources, frequent changes in modularity could have population-level effects which would need to be further explored. Social buffering is described as a process through which the negative effects of stress are ameliorated by sociality ( 54 , 66 ), often by individuals aggregating in larger groups or by associating with familiar individuals ( 18 ). Here, we observed that modularity was highest during periods with higher mean wind speeds, suggesting that if social buffering occurs during these periods, it is not through individuals forming larger aggregations. Alternatively, in zebra finches, social buffering may act through small groups, since the preferred social environment of wild zebra finches typically consists of at least two familiar individuals, i.e., the breeding pair, and, most likely other familiar, individuals with overlapping home ranges within the sub-colony ( 30 , 34 ). Conclusions Our findings that weather, specifically wind, drives social structure in wild zebra finches contributes to our understanding of how environmental conditions drive social behavior and the potential for social buffering. While many factors affect animal movements, our findings demonstrate that even short-term and relatively benign influences, such as moderate fluctuations in weather, affect how animal societies are organized, which likely has implications for information flow and effective anti-predator behavior. Our findings are thus a first step in highlighting the direction and degree of social flexibility that currently exists in response to daily environmental fluctuations in a wild population. Furthermore, understanding animal societies, such as zebra finches, that have evolved under unpredictable climatic conditions, may be valuable for assessing how social animals cope with an increasingly extreme and variable climate. Declarations Ethics approval and consent to participate Bird handling and tagging was performed in accordance with the Authority (#2018/027) of the Animal Ethics Committee at Macquarie University, and the Australian Bird & Bat Banding Scheme. Consent for publication Not applicable. Competing interests The authors declare no competing interests. Funding Declaration The project was supported in part by a Data Science / Artificial Intelligence grant to CT and a Dutch Research Council (NWO) grant to MN (ALWOP.334). Availability of data and materials Data and code to reproduce this analysis is available at the anonymized GitHub repository: https://anonymous.4open.science/r/zebby_social_buffering-BF5E/ Author contributions Chris Tyson : Conceptualization, Methodology, Software, Validation, Formal analysis, Investigation, Writing - Original Draft, Visualization, Supervision, Funding acquisition. Hugo Loning : Conceptualization, Methodology, Investigation, Data curation, Reviewing and Editing. Noëlle Tschirren : Conceptualization, Data Curation, Writing - Review & Editing. Elke Molenaar : Investigation, Data curation. Lysanne Snijders : Conceptualization, Writing - Original Draft, Writing - Review & Editing. Supervision. Simon Griffith : Conceptualization, Methodology, Investigation, Data Curation, Reviewing and Editing, Project administration. Marc Naguib : Conceptualization, Methodology, Investigation, Data Curation, Writing - Original Draft, Reviewing and Editing, Supervision, Project administration, Funding acquisition Acknowledgements We are grateful to Lyanne Brouwers and Rita Fragueira for fieldwork assistance. We also thank Garry Dowling and Mark Tilley for assistance installing components of the automated radio tracking system and Vicki Dowling for logistical support. References Kurvers RHJM, Krause J, Croft DP, Wilson ADM, Wolf M. The evolutionary and ecological consequences of animal social networks: Emerging issues. Trends Ecol Evol. 2014;29(6):326–35. Krause J, Ruxton GD. Living in Groups. Oxford University PressOxford; 2002. Papageorgiou D, Farine DR. Multilevel Societies in Birds. Trends Ecol Evol. 2021;36(1):15–7. Snijders L, Blumstein DT, Stanley CR, Franks DW. Animal Social Network Theory Can Help Wildlife Conservation. Trends Ecol Evol. 2017;32(8):567–77. Gartland LA, Firth JA, Laskowski KL, Jeanson R, Ioannou CC. Sociability as a personality trait in animals: methods, causes and consequences. Biol Rev. 2022;97(2):802–16. Beck KB, Farine DR, Kempenaers B. Social network position predicts male mating success in a small passerine. Behav Ecol. 2021;32(5):856–64. Williams AE, Worsley-Tonks KEL, Ezenwa VO. Drivers and consequences of variation in individual social connectivity. Anim Behav. 2017;133:1–9. Aplin LM, Firth JA, Farine DR, Voelkl B, Crates RA, Culina A, et al. Consistent individual differences in the social phenotypes of wild great tits, Parus major. Anim Behav. 2015;108:117–27. Hofmann HA, Beery AK, Blumstein DT, Couzin ID, Earley RL, Hayes LD, et al. An evolutionary framework for studying mechanisms of social behavior. Trends Ecol Evol. 2014;29(10):581–9. Fisher DN, Kilgour RJ, Siracusa ER, Foote JR, Hobson EA, Montiglio PO, et al. Anticipated effects of abiotic environmental change on intraspecific social interactions. Biol Rev. 2021;96(6):2661–93. Brandl HB, Farine DR. Stress in the social environment: behavioural and social consequences of stress transmission in bird flocks. Proc R Soc B Biol Sci. 2024;291(2034). Camerlenghi E, Nolazco S, Farine DR, Magrath RD, Peters A. Social restructuring during harsh environmental conditions promotes cooperative behaviour in a songbird. Proc R Soc B Biol Sci. 2024;291(2021). Dingle H, Migration. The Biology of Life on the Move. Oxford University Press; 2014. Aikens EO, Bontekoe ID, Blumenstiel L, Schlicksupp A, Flack A. Viewing animal migration through a social lens. Trends Ecol Evol. 2022;37(11):985–96. Sayin S, Couzin-Fuchs E, Petelski I, Günzel Y, Salahshour M, Lee C-Y, et al. The behavioral mechanisms governing collective motion in swarming locusts. Sci (80-). 2025;387(6737):995–1000. Blumstein DT, Hayes LD, Pinter-Wollman N. Social consequences of rapid environmental change. Vol. 38, Trends in Ecology and Evolution. Elsevier Ltd; 2023. pp. 337–45. Baigger A, Perony N, Reuter M, Leinert V, Melber M, Grünberger S, et al. Bechstein’s bats maintain individual social links despite a complete reorganisation of their colony structure. Naturwissenschaften. 2013;100(9):895–8. Testard C, Larson SM, Watowich MM, Kaplinsky CH, Bernau A, Faulder M, et al. Rhesus macaques build new social connections after a natural disaster. Curr Biol. 2021;31(11):2299–e23097. Schaffner CM, Rebecchini L, Ramos-Fernandez G, Vick LG, Aureli F. Spider Monkeys (Ateles geoffroyi yucatenensis) Cope with the Negative Consequences of Hurricanes Through Changes in Diet, Activity Budget, and Fission-Fusion Dynamics. Int J Primatol. 2012;33(4):922–36. Nourani E, Safi K, de Grissac S, Anderson DJ, Cole NC, Fell A, et al. Seabird morphology determines operational wind speeds, tolerable maxima, and responses to extremes. Curr Biol. 2023;33(6):1179–e11843. Thorne LH, Clay TA, Phillips RA, Silvers LG, Wakefield ED. Effects of wind on the movement, behavior, energetics, and life history of seabirds. Mar Ecol Prog Ser. 2023;723:73–117. Kranstauber B, Weinzierl R, Wikelski M, Safi K. Global aerial flyways allow efficient travelling. Ecol Lett. 2015;18(12):1338–45. Vansteelant WMG, Bouten W, Klaassen RHG, Koks BJ, Schlaich AE, van Diermen J, et al. Regional and seasonal flight speeds of soaring migrants and the role of weather conditions at hourly and daily scales. J Avian Biol. 2015;46(1):25–39. Cornioley T, Börger L, Ozgul A, Weimerskirch H. Impact of changing wind conditions on foraging and incubation success in male and female wandering albatrosses. J Anim Ecol. 2016;85(5):1318–27. Alerstam T. Conflicting evidence about long-distance animal navigation. Science. 2006;313(5788):791–4. Hedenstrom A, Alerstam T. Optimal flight speed of birds. Philos Trans R Soc Lond Ser B Biol Sci. 1995;348(1326):471–87. Shepard ELC, Wilson RP, Rees WG, Grundy E, Lambertucci SA, Vosper SB. Energy landscapes shape animal movement ecology. Am Nat. 2013;182(3):298–312. Rat M, Mathe-Hubert H, McKechnie AE, Sueur C, Cunningham SJ. Extreme and variable environmental temperatures are linked to reduction of social network cohesiveness in a highly social passerine. Oikos. 2020;129(11):1597–610. Borthwick Z, Quiring K, Griffith SC, Leu ST. Heat stress conditions affect the social network structure of free-ranging sheep. Ecol Evol. 2024;14(2). Tyson C, Loning H, Griffith SC, Naguib M. Constant companions: wild zebra finch pairs display extreme spatial cohesion. Biol Lett. 2024;20(11):2024.03.21.586046. McCowan LSC, Mariette MM, Griffith SC. The size and composition of social groups in the wild zebra finch. Emu. 2015;115(3):191–8. Loning H, Fragueira R, Naguib M, Griffith SC. Hanging out in the outback: the use of social hotspots by wild zebra finches. J Avian Biol. 2023;2023:11–2. Zann RA. The zebra finch: a synthesis of field and laboratory studies. Oxford: Oxford University Press; 1996. Brandl HB, Griffith SC, Farine DR, Schuett W. Wild zebra finches that nest synchronously have long-term stable social ties. J Anim Ecol. 2021;90(1):76–86. Morton SR, Stafford Smith DM, Dickman CR, Dunkerley DL, Friedel MH, McAllister RRJ, et al. A fresh framework for the ecology of arid Australia. J Arid Environ. 2011;75(4):313–29. Loarie SR, Duffy PB, Hamilton H, Asner GP, Field CB, Ackerly DD. The velocity of climate change. Nature. 2009;462(7276):1052–5. Funghi C, McCowan LSC, Schuett W, Griffith SC. High air temperatures induce temporal, spatial and social changes in the foraging behaviour of wild zebra finches. Anim Behav. 2019;149:33–43. Evans JC, Hodgson DJ, Boogert NJ, Silk MJ. Group size and modularity interact to shape the spread of infection and information through animal societies. Behav Ecol Sociobiol. 2021;75(12). Jirinec V, Rodrigues PF, Amaral B. Adjustable leg harness for attaching tags to small and medium-sized birds. J F Ornithol. 2021;92(1):77–87. Tyson C, Fragueira R, Sansano-Sansano E, Yu H, Naguib M. Fingerprint localisation for fine-scale wildlife tracking using automated radio telemetry. Methods Ecol Evol. 2024;2024(April):2118–28. Paxton KL, Baker KM, Crytser ZB, Guinto RMP, Brinck KW, Rogers HS et al. Optimizing trilateration estimates for tracking fine-scale movement of wildlife using automated radio telemetry networks. Ecol Evol. 2022;12(2). R Core Team. R: A Language and Environment for Statistical Computing. Vienna, Austria: R Foundation for Statistical Computing; 2024. Whitehouse K, Karlof C, Culler D. A Practical Evaluation of Radio Signal Strength for Ranging-based Localization. ACM SIGMOBILE Mob Comput Commun Rev. 2007;11(1):41–52. Robitaille AL, Webber QMR, Vander Wal E. Conducting social network analysis with animal telemetry data: Applications and methods using spatsoc. Methods Ecol Evol. 2019;10(8):1203–11. Hart J, Weiss MN, Franks D, Brent L. BISoN: A Bayesian framework for inference of social networks. Methods Ecol Evol. 2023;14(9):2411–20. Pereira AS, Pavez-Fox MA, Hart JDA, Valle JEN-D, Phillips D, Casanova C et al. The impact of kinship composition on social structure. bioRxiv. 2024;(1):2024.01.11.575037. Hart JDA, Franks DW, Brent LJN, Weiss MN. bisonR - Bayesian Inference of Social Networks with R. 2022. Blondel VD, Guillaume J-L, Lambiotte R, Lefebvre E. Fast unfolding of communities in large networks. J Stat Mech Theory Exp. 2008;2008(10):P10008. Csardi G, Nepusz T. The igraph software package for complex network research. InterJournal Complex Syst. 2006;Complex Sy(1695):1695. Bürkner PC. brms: An R package for Bayesian multilevel models using Stan. J Stat Softw. 2017;80(1). Gelman A, Hill J. Data Analysis Using Regression and Multilevel/Hierarchical Models. Cambridge University Press; 2006. Calabrese JM, Fleming CH, Gurarie E. ctmm: an r package for analyzing animal relocation data as a continuous-time stochastic process. Freckleton R, editor. Methods Ecol Evol. 2016;7(9):1124–32. Fleming CH, Drescher-Lehman J, Noonan MJ, Akre TSB, Brown DJ, Cochrane MM et al. A comprehensive framework for handling location error in animal tracking data. 2020. p. 31. Komdeur J, Ma L. Keeping up with environmental change: The importance of sociality. Ethology. 2021;127(10):790–807. Liechti F. Birds: Blowin’ by the wind? J Ornithol. 2006;147(2):202–11. Shepard E, Cole EL, Neate A, Lempidakis E, Ross A. Wind prevents cliff-breeding birds from accessing nests through loss of flight control. Elife. 2019;8:1–15. Tomotani BM, Muijres FT, Johnston B, van der Jeugd HP, Naguib M. Great tits do not compensate over time for a radio-tag-induced reduction in escape-flight performance. Ecol Evol. 2021;11(23):16600–17. Lapsansky AB, Igo JA, Tobalske BW. Zebra finch (Taeniopygia guttata) shift toward aerodynamically efficient flight kinematics in response to an artificial load. Biol Open. 2019;8(6):1–9. Cherry MJ, Barton BT. Effects of wind on predator-prey interactions. Food Webs. 2017;13:92–7. Brumm H, Naguib M. In. Chapter 1 Environmental Acoustics and the Evolution of Bird Song. 2009. pp. 1–33. Loning H, Griffith SC, Naguib M. Zebra finch song is a very short-range signal in the wild: evidence from an integrated approach. Behav Ecol. 2022;33(1):37–46. Sharpe LL, Prober SM, Gardner JL. the Hot Seat: Behavioral Change and Old-Growth Trees Underpin an Australian Songbird’s Response to Extreme Heat. Front Ecol Evol. 2022;10(March):1–19. Duboscq J, Romano V, MacIntosh A, Sueur C. Social information transmission in animals: Lessons from studies of diffusion. Front Psychol. 2016;7(AUG):1–15. Brandl HB, Griffith SC, Schuett W. Wild zebra finches choose neighbours for synchronized breeding. Anim Behav. 2019;151:21–8. Romano V, Shen M, Pansanel J, MacIntosh AJJ, Sueur C. Social transmission in networks: global efficiency peaks with intermediate levels of modularity. Behav Ecol Sociobiol. 2018;72(9). Kikusui T, Winslow JT, Mori Y. Social buffering: Relief from stress and anxiety. Philos Trans R Soc B Biol Sci. 2006;361(1476):2215–28. Ethics. declarations. Additional Declarations No competing interests reported. Cite Share Download PDF Status: Published Journal Publication published 21 Jan, 2026 Read the published version in Movement Ecology → Version 1 posted Editorial decision: Revision requested 28 Jun, 2025 Reviews received at journal 15 Jun, 2025 Reviewers agreed at journal 12 Jun, 2025 Reviews received at journal 29 May, 2025 Reviewers agreed at journal 27 May, 2025 Reviewers invited by journal 27 May, 2025 Editor assigned by journal 24 Apr, 2025 Submission checks completed at journal 24 Apr, 2025 First submitted to journal 22 Apr, 2025 You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {\"props\":{\"pageProps\":{\"initialData\":{\"identity\":\"rs-6502285\",\"acceptedTermsAndConditions\":true,\"allowDirectSubmit\":false,\"archivedVersions\":[],\"articleType\":\"Research Article\",\"associatedPublications\":[],\"authors\":[{\"id\":462459937,\"identity\":\"ed9088f7-b180-4b08-ab86-f54292336779\",\"order_by\":0,\"name\":\"Chris Tyson\",\"email\":\"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAAyklEQVRIiWNgGAWjYDACCQaGA1Am4wMwxU6cFgMQkxlMMjAToYUBqoVNgigt/LN7Hx78wfAncX5777Fq3rZ7DOaEtEjcOW5wmIfBIHHDmXNpt3nbihksmwloMZBIYzjMANIikWN2O7ctgcHgMBFagA4zSJw/I8esmGgtB0AOa7iRY8ZMlBaJG0CH8RgYG284c8ZY+s+5BB6CWvhnpDF//FEhJzu/vcfw44yyBDmD4w0E9ECcx+AIU8dDjHowsCda5SgYBaNgFIw8AACAxj1rGmX5UAAAAABJRU5ErkJggg==\",\"orcid\":\"\",\"institution\":\"Wageningen University \\u0026 Research\",\"correspondingAuthor\":true,\"prefix\":\"\",\"firstName\":\"Chris\",\"middleName\":\"\",\"lastName\":\"Tyson\",\"suffix\":\"\"},{\"id\":462459938,\"identity\":\"3a8c80bd-a42c-401b-9db5-9a2c837505a7\",\"order_by\":1,\"name\":\"Hugo Loning\",\"email\":\"\",\"orcid\":\"\",\"institution\":\"Wageningen University \\u0026 Research\",\"correspondingAuthor\":false,\"prefix\":\"\",\"firstName\":\"Hugo\",\"middleName\":\"\",\"lastName\":\"Loning\",\"suffix\":\"\"},{\"id\":462459939,\"identity\":\"50292623-ce1e-4169-b6d9-25badc03979e\",\"order_by\":2,\"name\":\"Noëlle Tschirren\",\"email\":\"\",\"orcid\":\"\",\"institution\":\"Wageningen University \\u0026 Research\",\"correspondingAuthor\":false,\"prefix\":\"\",\"firstName\":\"Noëlle\",\"middleName\":\"\",\"lastName\":\"Tschirren\",\"suffix\":\"\"},{\"id\":462459940,\"identity\":\"18f2217e-fb25-46fb-b09f-f4237bd33fdc\",\"order_by\":3,\"name\":\"Elke Molenaar\",\"email\":\"\",\"orcid\":\"\",\"institution\":\"Wageningen University \\u0026 Research\",\"correspondingAuthor\":false,\"prefix\":\"\",\"firstName\":\"Elke\",\"middleName\":\"\",\"lastName\":\"Molenaar\",\"suffix\":\"\"},{\"id\":462459941,\"identity\":\"5ac0dc1c-f02d-4d05-bd3a-745e67901e60\",\"order_by\":4,\"name\":\"Lysanne Snijders\",\"email\":\"\",\"orcid\":\"\",\"institution\":\"Wageningen University \\u0026 Research\",\"correspondingAuthor\":false,\"prefix\":\"\",\"firstName\":\"Lysanne\",\"middleName\":\"\",\"lastName\":\"Snijders\",\"suffix\":\"\"},{\"id\":462459942,\"identity\":\"ae0ed5b3-22af-4608-b46e-d425dbd566bc\",\"order_by\":5,\"name\":\"Simon Griffith\",\"email\":\"\",\"orcid\":\"\",\"institution\":\"Macquarie University\",\"correspondingAuthor\":false,\"prefix\":\"\",\"firstName\":\"Simon\",\"middleName\":\"\",\"lastName\":\"Griffith\",\"suffix\":\"\"},{\"id\":462459943,\"identity\":\"46829f82-db5d-46c9-8ff6-debe340a6ef7\",\"order_by\":6,\"name\":\"Marc Naguib\",\"email\":\"\",\"orcid\":\"\",\"institution\":\"Wageningen University \\u0026 Research\",\"correspondingAuthor\":false,\"prefix\":\"\",\"firstName\":\"Marc\",\"middleName\":\"\",\"lastName\":\"Naguib\",\"suffix\":\"\"}],\"badges\":[],\"createdAt\":\"2025-04-22 08:53:33\",\"currentVersionCode\":1,\"declarations\":\"\",\"doi\":\"10.21203/rs.3.rs-6502285/v1\",\"doiUrl\":\"https://doi.org/10.21203/rs.3.rs-6502285/v1\",\"draftVersion\":[],\"editorialEvents\":[{\"content\":\"https://doi.org/10.1186/s40462-025-00623-9\",\"type\":\"published\",\"date\":\"2026-01-21T15:57:07+00:00\"}],\"editorialNote\":\"\",\"failedWorkflow\":false,\"files\":[{\"id\":83621215,\"identity\":\"47d37f9c-4527-4d29-a5ce-bad79c92d241\",\"added_by\":\"auto\",\"created_at\":\"2025-05-29 15:16:57\",\"extension\":\"jpg\",\"order_by\":1,\"title\":\"Figure 1\",\"display\":\"\",\"copyAsset\":false,\"role\":\"figure\",\"size\":171118,\"visible\":true,\"origin\":\"\",\"legend\":\"\\u003cp\\u003eA) Map of Gap Hills study area with the 93 automated receivers (white points) and the sensor station (red point). Catching and tagging effort was concentrated at three nest box sites (yellow points). B) Aerial photograph of the study area facing northeast and centered on the water retention basin. C) Male zebra finch wearing LifeTag with nitinol antenna. D) Sensor station and antenna mast mounted with 4 Yagi antennas and 1 omnidirectional antenna.\\u003c/p\\u003e\",\"description\":\"\",\"filename\":\"1.jpg\",\"url\":\"https://assets-eu.researchsquare.com/files/rs-6502285/v1/b41bd75b0ad80aa1a5886f4c.jpg\"},{\"id\":83621216,\"identity\":\"d32605f2-5c8a-4ffd-94db-4c6d49654d1f\",\"added_by\":\"auto\",\"created_at\":\"2025-05-29 15:16:57\",\"extension\":\"jpg\",\"order_by\":2,\"title\":\"Figure 2\",\"display\":\"\",\"copyAsset\":false,\"role\":\"figure\",\"size\":185700,\"visible\":true,\"origin\":\"\",\"legend\":\"\\u003cp\\u003ePredicted (A) density and (B) modularity index values in relation to hour of day, mean hourly temperature, and mean hourly wind speed, after controlling for the other predictors. All predictor values are scaled (z-transformed). Black dots show draws from the posterior distribution (n = 400) and colored bands show the credible interval levels of the predicted index values for each metric.\\u003c/p\\u003e\",\"description\":\"\",\"filename\":\"2.jpg\",\"url\":\"https://assets-eu.researchsquare.com/files/rs-6502285/v1/5910fcc28efd88cb33bf7c4b.jpg\"},{\"id\":83621713,\"identity\":\"8828b7cb-0ab6-41f9-bea4-131f3f84b00e\",\"added_by\":\"auto\",\"created_at\":\"2025-05-29 15:24:57\",\"extension\":\"jpg\",\"order_by\":3,\"title\":\"Figure 3\",\"display\":\"\",\"copyAsset\":false,\"role\":\"figure\",\"size\":404300,\"visible\":true,\"origin\":\"\",\"legend\":\"\\u003cp\\u003eA-B) Localisation estimates for all individuals (white points) and associating individuals (colored points) during an hour with low wind speed (A: 11 km/h) and high wind speed (B: 39 km/h). Each color corresponds to community membership based on Louvain clustering. Localizations are shown for similar times of day (A: 1200 Australian Central Standard Time (ACST), and B: 1300 ACST) and number of interacting individuals (A, 52 individuals; B, 57 individuals). C-D) Representative social network plots depicting low (0.31) and high (0.69) modularity. Networks were constructed using Bayesian edge weight models and edge weights below the minimum observed edge weight were set to zero to avoid constructing fully connected networks. Nodes are colored according to community membership as shown in A and B.\\u003c/p\\u003e\",\"description\":\"\",\"filename\":\"3.jpg\",\"url\":\"https://assets-eu.researchsquare.com/files/rs-6502285/v1/675bee8ca6ca2f3daa46223e.jpg\"},{\"id\":101152840,\"identity\":\"07eb1749-acb1-4402-9409-859c99ae712a\",\"added_by\":\"auto\",\"created_at\":\"2026-01-26 16:13:20\",\"extension\":\"pdf\",\"order_by\":0,\"title\":\"\",\"display\":\"\",\"copyAsset\":false,\"role\":\"manuscript-pdf\",\"size\":1424107,\"visible\":true,\"origin\":\"\",\"legend\":\"\",\"description\":\"\",\"filename\":\"manuscript.pdf\",\"url\":\"https://assets-eu.researchsquare.com/files/rs-6502285/v1/46930cec-ee0e-4289-8148-7e8b3be86c95.pdf\"}],\"financialInterests\":\"No competing interests reported.\",\"formattedTitle\":\"Windy weather drives social structure in wild zebra finches\",\"fulltext\":[{\"header\":\"Background\",\"content\":\"\\u003cp\\u003eAmong social animals there is large variation in how individuals form groups, which has consequences for multiple key behavioral, ecological and evolutionary processes.\\u003c/p\\u003e \\u003cp\\u003eAt a group level, social network structure may impact predator avoidance, population stability, dispersal, and social evolution (\\u003cspan additionalcitationids=\\\"CR2 CR3\\\" citationid=\\\"CR1\\\" class=\\\"CitationRef\\\"\\u003e1\\u003c/span\\u003e\\u0026ndash;\\u003cspan citationid=\\\"CR4\\\" class=\\\"CitationRef\\\"\\u003e4\\u003c/span\\u003e). At an individual level, there is large variation in the quantity and quality of relationships an individual has with conspecifics (\\u003cspan citationid=\\\"CR5\\\" class=\\\"CitationRef\\\"\\u003e5\\u003c/span\\u003e). Therefore, an individual\\u0026rsquo;s social network position may affect its access to mating opportunities, social support, information about novel resources such as ephemeral food patches, or disease infection (\\u003cspan additionalcitationids=\\\"CR7\\\" citationid=\\\"CR6\\\" class=\\\"CitationRef\\\"\\u003e6\\u003c/span\\u003e\\u0026ndash;\\u003cspan citationid=\\\"CR8\\\" class=\\\"CitationRef\\\"\\u003e8\\u003c/span\\u003e). Social network position can thus have significant fitness implications. Consequently, understanding the drivers of social network variation and the relative influence of the various factors that affect social organization is a major focus of contemporary animal behavior research (\\u003cspan additionalcitationids=\\\"CR10 CR11\\\" citationid=\\\"CR9\\\" class=\\\"CitationRef\\\"\\u003e9\\u003c/span\\u003e\\u0026ndash;\\u003cspan citationid=\\\"CR12\\\" class=\\\"CitationRef\\\"\\u003e12\\u003c/span\\u003e).\\u003c/p\\u003e \\u003cp\\u003eClimate and weather are pervasive influences on social structure in wild populations (\\u003cspan citationid=\\\"CR10\\\" class=\\\"CitationRef\\\"\\u003e10\\u003c/span\\u003e). At a broad scale, seasonal changes in climate drive shifts in resource distributions, which in turn induces cyclical changes in the spatial organization of numerous species, particularly in the temperate and polar regions (\\u003cspan citationid=\\\"CR13\\\" class=\\\"CitationRef\\\"\\u003e13\\u003c/span\\u003e). Collective migrations are conspicuous examples of such annual fluctuations in social structure and may be linked with Behavioural changes, such as increased sociality (\\u003cspan citationid=\\\"CR14\\\" class=\\\"CitationRef\\\"\\u003e14\\u003c/span\\u003e, \\u003cspan citationid=\\\"CR15\\\" class=\\\"CitationRef\\\"\\u003e15\\u003c/span\\u003e). In many species, these fluctuations occur annually and as such are regularly experienced. Extreme weather events may also induce changes in social structure more locally and on a more rapid time scale with these changes being either ephemeral or enduring (\\u003cspan citationid=\\\"CR16\\\" class=\\\"CitationRef\\\"\\u003e16\\u003c/span\\u003e, \\u003cspan citationid=\\\"CR17\\\" class=\\\"CitationRef\\\"\\u003e17\\u003c/span\\u003e). For example, hurricanes and cyclones can induce changes in group structure in social primates that persist for months (\\u003cspan citationid=\\\"CR18\\\" class=\\\"CitationRef\\\"\\u003e18\\u003c/span\\u003e, \\u003cspan citationid=\\\"CR19\\\" class=\\\"CitationRef\\\"\\u003e19\\u003c/span\\u003e).\\u003c/p\\u003e \\u003cp\\u003eWhile less dramatic, fluctuations in daily weather components may also impact the structure of animal groups over short time scales. In particular, wind and temperature are two primary weather components that may have especially large effects on avian species by increasing the energetic costs of flight (\\u003cspan citationid=\\\"CR20\\\" class=\\\"CitationRef\\\"\\u003e20\\u003c/span\\u003e, \\u003cspan citationid=\\\"CR21\\\" class=\\\"CitationRef\\\"\\u003e21\\u003c/span\\u003e) and thus potentially altering movement and social connectivity. Wind is a critical element for birds that shapes not only long-distance movements in migratory species (\\u003cspan citationid=\\\"CR22\\\" class=\\\"CitationRef\\\"\\u003e22\\u003c/span\\u003e, \\u003cspan citationid=\\\"CR23\\\" class=\\\"CitationRef\\\"\\u003e23\\u003c/span\\u003e), but also affects foraging efficiency by modulating energy expenditure (\\u003cspan citationid=\\\"CR24\\\" class=\\\"CitationRef\\\"\\u003e24\\u003c/span\\u003e, \\u003cspan citationid=\\\"CR25\\\" class=\\\"CitationRef\\\"\\u003e25\\u003c/span\\u003e). For birds with flapping flight, energetic costs are expected to be lowest at intermediate wind speeds (\\u003cspan citationid=\\\"CR26\\\" class=\\\"CitationRef\\\"\\u003e26\\u003c/span\\u003e) and consequently extreme wind speeds may affect mobility and social interactions. Temperature also shapes energetic landscapes for birds and thus directly impacts locomotion costs as well (\\u003cspan citationid=\\\"CR27\\\" class=\\\"CitationRef\\\"\\u003e27\\u003c/span\\u003e). On short time scales, temperature may potentially limit the flexibility of social organization and buffering by increasing movement costs and limiting individuals from associating (\\u003cspan citationid=\\\"CR28\\\" class=\\\"CitationRef\\\"\\u003e28\\u003c/span\\u003e). Alternatively, higher temperatures may lead to increased association rates by constraining individuals to remain near particular resources, such as shade or water (\\u003cspan citationid=\\\"CR29\\\" class=\\\"CitationRef\\\"\\u003e29\\u003c/span\\u003e). As such, together both weather components may have significant impacts on social network structure in complimentary or opposing directions.\\u003c/p\\u003e \\u003cp\\u003eIn this study, we examine the influence of two primary weather components, wind and temperature, on social network structure and space-use patterns in wild zebra finches (\\u003cem\\u003eTaeniopygia castanosis\\u003c/em\\u003e) over short time scales. Zebra finches live in dynamic, multi-level societies with mated pairs staying almost constantly together over space and time (\\u003cspan citationid=\\\"CR30\\\" class=\\\"CitationRef\\\"\\u003e30\\u003c/span\\u003e), often form small cohorts of up to 10 individuals (\\u003cspan citationid=\\\"CR31\\\" class=\\\"CitationRef\\\"\\u003e31\\u003c/span\\u003e) and aggregate in larger groups at resources, social hotspots (\\u003cspan citationid=\\\"CR32\\\" class=\\\"CitationRef\\\"\\u003e32\\u003c/span\\u003e) or when joining long distance movements during dispersal (\\u003cspan citationid=\\\"CR33\\\" class=\\\"CitationRef\\\"\\u003e33\\u003c/span\\u003e). The social network of wild zebra finches can extend across years, and relates to reproductive synchrony with others (\\u003cspan citationid=\\\"CR34\\\" class=\\\"CitationRef\\\"\\u003e34\\u003c/span\\u003e), potentially buffering a pair against the risks of predation during breeding. Zebra finches are native to the Australian arid zone which is characterized by increasingly extreme short- and long-term unpredictability in climate conditions (\\u003cspan citationid=\\\"CR35\\\" class=\\\"CitationRef\\\"\\u003e35\\u003c/span\\u003e, \\u003cspan citationid=\\\"CR36\\\" class=\\\"CitationRef\\\"\\u003e36\\u003c/span\\u003e). Since social connectivity in such dynamic multi-level societies requires active movement of individuals, these social systems provide prime examples to determine if weather and climate change affect social organization.\\u003c/p\\u003e \\u003cp\\u003eTo assess the impact of wind and moderate temperature on social network structure in zebra finches, we used automated radio tracking to concurrently track 128 adult zebra finches and determined how social structure and movement patterns varied in relation to wind, temperature, and time of day. We predicted that birds would fly less during windy periods and would cluster in protected sites, such as dense vegetation, which would result in smaller social groups. During high temperatures (\\u0026gt;\\u0026thinsp;35\\u0026deg;C), zebra finches reduce activity and remain close to water sources (\\u003cspan citationid=\\\"CR37\\\" class=\\\"CitationRef\\\"\\u003e37\\u003c/span\\u003e), yet those temperatures are rare during the period in which we conducted our study, and so we expected temperature to have less of an effect than wind on social organization. Zebra finches are typically seen in larger groups during the day (\\u003cspan citationid=\\\"CR33\\\" class=\\\"CitationRef\\\"\\u003e33\\u003c/span\\u003e) and so we expected to observe decreased social cohesion in the morning and evening. To quantify these changes in social organization, we measured two social network metrics, density and modularity, both of which are linked with processes such as disease and information transfer (\\u003cspan citationid=\\\"CR38\\\" class=\\\"CitationRef\\\"\\u003e38\\u003c/span\\u003e). We also examined social organization at a broad level by assessing the movement similarity between neighboring individuals, which we predicted would be less cohesive during windier and warmer periods.\\u003c/p\\u003e\"},{\"header\":\"Methods\",\"content\":\"\\u003cdiv id=\\\"Sec3\\\" class=\\\"Section2\\\"\\u003e \\u003ch2\\u003eStudy site \\u0026amp; automated radio tracking\\u003c/h2\\u003e \\u003cp\\u003eOur study was conducted at Gap Hills within the Fowlers Gap Arid Zone Research Station, western New South Wales, Australia (30\\u0026deg;56\\u0026prime;58\\u0026Prime; S, 141\\u0026deg;46\\u0026prime;02\\u0026Prime; E) in September 2022. The area primarily consists of open \\u003cem\\u003eAcacia\\u003c/em\\u003e shrubland surrounding a water retention basin (200 x 150 m) and an ephemeral creek system (Fig.\\u0026nbsp;\\u003cspan refid=\\\"Fig1\\\" class=\\\"InternalRef\\\"\\u003e1\\u003c/span\\u003eA-B). We caught zebra finches using mist nets at three sub-sites (Fig.\\u0026nbsp;\\u003cspan refid=\\\"Fig1\\\" class=\\\"InternalRef\\\"\\u003e1\\u003c/span\\u003eA). Catching began at sunrise and nets were checked every 15 minutes. We marked all captured birds with an individually numbered metal band from the Australian Bird and Bat Banding Scheme. Birds were then weighed using a 20-g Pesola scale and birds above 10 g were outfitted with a 0.48 g solar-powered radio tag with a nitinol antenna (LifeTag, Cellular Tracking Technologies; New Jersey, USA), which was attached using a stretchable nylon leg-loop harness (Jirinec et al., 2021; Tyson, Loning, et al., 2024; Fig.\\u0026nbsp;\\u003cspan refid=\\\"Fig1\\\" class=\\\"InternalRef\\\"\\u003e1\\u003c/span\\u003eC). The mean mass of tagged birds was 12.7 g (range 10.4\\u0026ndash;15.6 g). After tagging, all birds were observed to fly off similarly to untagged individuals and tagged individuals were routinely resighted. Bird handling and tagging was performed in accordance with the Authority (#2018/027) of the Animal Ethics Committee at Macquarie University, and the Australian Bird \\u0026amp; Bat Banding Scheme.\\u003c/p\\u003e \\u003cp\\u003e \\u003c/p\\u003e \\u003cp\\u003eIn total, we tagged 128 zebra finches (53 females, 60 males, and 16 juveniles) from which we obtained tracking data for 12 consecutive days (19 September to 30 September). During this period, birds were tracked for an average of 10.5 days. Variation in tracking duration was due to antennas of the tags breaking after which point the tags were excluded from the analysis. Broken antennas were clearly identifiable by a large change in the signal strength as well as the number of receivers detecting the tag. Tracking began at the start of the period when we annually monitor nest boxes for breeding activity, and we discontinued tracking after 12 days due to the tag malfunctions. During the tracking period, 34% of nest boxes (62 out of 180) had active nests though the breeding status of the birds we tracked was unknown. Tags were programmed to signal every 5 seconds when exposed to sufficient solar energy. During the study period, tags were detected during daylight hours, i.e., between ~\\u0026thinsp;0600 and ~\\u0026thinsp;1800 Australian Central Standard Time (ACST), but gaps in detections also occurred when birds entered nest boxes or dense vegetation. On average, localization estimates were obtained every 3 minutes.\\u003c/p\\u003e \\u003cp\\u003eTo track tagged birds, we used an automated radio tracking system which consisted of 93 receivers (Node v2, Cellular Tracking Technologies, New Jersey, USA) and one base station (SensorStation, Cellular Tracking Technologies, New Jersey, USA). Each receiver recorded the tag ID, time of detection, and signal strength, which was relayed to the base station. Receivers were mounted\\u0026thinsp;~\\u0026thinsp;1.5 m above ground on a steel fence post. Most receivers (n\\u0026thinsp;=\\u0026thinsp;73) were spaced 150 m apart on a triangular grid with the remainder placed opportunistically within the three nest box sub-sites (Fig.\\u0026nbsp;\\u003cspan refid=\\\"Fig1\\\" class=\\\"InternalRef\\\"\\u003e1\\u003c/span\\u003eA). The central base station was placed centrally on top of the water retention basin (Fig.\\u0026nbsp;\\u003cspan refid=\\\"Fig1\\\" class=\\\"InternalRef\\\"\\u003e1\\u003c/span\\u003eA) and was connected to four 430 MHz Yagi antennas oriented perpendicularly and one 430 MHz omnidirectional antenna oriented vertically, all of which were mounted to the top of a 6 m mast (Fig.\\u0026nbsp;\\u003cspan refid=\\\"Fig1\\\" class=\\\"InternalRef\\\"\\u003e1\\u003c/span\\u003eD). We used these multiple potentially redundant antennas to minimize the possibility of missed detections, which can greatly impact localization accuracy (\\u003cspan citationid=\\\"CR40\\\" class=\\\"CitationRef\\\"\\u003e40\\u003c/span\\u003e).\\u003c/p\\u003e \\u003c/div\\u003e\\n\\u003ch3\\u003eRadio tag localization\\u003c/h3\\u003e\\n\\u003cp\\u003eWe estimated tag locations using radio signal strength-based (RSS) multilateration (\\u003cspan citationid=\\\"CR41\\\" class=\\\"CitationRef\\\"\\u003e41\\u003c/span\\u003e). To do so, we first calibrated the signal strength to distance relationship by placing six tags at set distances (1, 2, 5, 10, 15, 25, 50 ,75, 100, 150 m) from four receivers. Tags were mounted horizontally on a cardboard strip that was placed atop a 2 m PVC pole. Tags were held stationary at each distance for two minutes. The exponential decay relationship between the average RSS of detections from each calibration distance was then modelled using the equation: RSS \\u0026sim; a \\u0026times; exp(\\u0026minus;\\u0026thinsp;S \\u0026times; distance)\\u0026thinsp;+\\u0026thinsp;K; where a is the intercept, S is the decay factor, and K is the horizontal asymptote (\\u003cspan citationid=\\\"CR41\\\" class=\\\"CitationRef\\\"\\u003e41\\u003c/span\\u003e). Initial parameter values were estimated using the self-starting function \\u0026lsquo;\\u003cem\\u003eSSasymp,\\u0026rsquo;\\u003c/em\\u003e which were used to fit an exponential decay model using the function \\u0026lsquo;nls\\u0026rsquo; (\\u003cspan citationid=\\\"CR42\\\" class=\\\"CitationRef\\\"\\u003e42\\u003c/span\\u003e). The parameter estimates from this model were then used to fit the final exponential decay model (\\u003cspan citationid=\\\"CR41\\\" class=\\\"CitationRef\\\"\\u003e41\\u003c/span\\u003e).\\u003c/p\\u003e \\u003cp\\u003eTo estimate distance for multilateration, RSS values from each tag at each node were averaged within a 15 second moving-window with a 5 second lag to remove large fluctuations in signal strength which can occur due to various sources of interference (\\u003cspan citationid=\\\"CR30\\\" class=\\\"CitationRef\\\"\\u003e30\\u003c/span\\u003e). The mean RSS value within each interval at each node was then used to estimate distance between the tag and the corresponding node using the exponential decay relationship given above. In each 15 second interval, nodes within 200 m of the receiver with the strongest average RSS value in that interval were retained and were used to estimate the tag location using RSS-based multilateration. This method uses the estimated distance between a tag and each node (minimally three) to estimate the location of the tag, which can be found using a non-linear least squares model which minimizes the differences between pairwise distances between all nodes and the estimated distance between the tag and each node (\\u003cspan citationid=\\\"CR41\\\" class=\\\"CitationRef\\\"\\u003e41\\u003c/span\\u003e). The coordinates of the node with the strongest signal were used as initial values for the model. Multiple factors aside from distance will influence the RSS of a tag. Such factors include the relative orientation of the tag and node, physical obstructions between the tag and receiver, and changes to the innate signal strength of the tag over time (\\u003cspan citationid=\\\"CR43\\\" class=\\\"CitationRef\\\"\\u003e43\\u003c/span\\u003e). As such, a given RSS value will correspond to a range of possible distances between the tag and the receiver, which will increase as the signal strength decreases. To partially account for this variation in the localization estimates, the distance between a tag and a receiver in each interval was estimated 100 times by sampling from \\u0026plusmn;\\u0026thinsp;1 SE around the mean distance for the mean RSS-value in that interval. These 100 samples from each interval were then averaged to determine a point estimate for that interval. We also calculated an error ellipse corresponding to the \\u003cspan class=\\\"InlineEquation\\\"\\u003e\\u003cspan class=\\\"mathinline\\\"\\u003e\\\\(\\\\:\\\\sqrt{2}\\\\)\\u003c/span\\u003e\\u003c/span\\u003e-sigma ellipse of a bivariate normal distribution to estimate the uncertainty around each location. The median error (i.e., the difference between the estimated location and the true location) from field calibration tests was 35 m.\\u003c/p\\u003e\\n\\u003ch3\\u003eWeather data\\u003c/h3\\u003e\\n\\u003cp\\u003eWeather data was provided by the Australian Bureau of Meteorology (\\u003cspan class=\\\"ExternalRef\\\"\\u003e\\u003cspan class=\\\"RefSource\\\"\\u003ehttp://www.bom.gov.au/nsw/\\u003c/span\\u003e\\u003cspan address=\\\"http://www.bom.gov.au/nsw/\\\" targettype=\\\"URL\\\" class=\\\"RefTarget\\\"\\u003e\\u003c/span\\u003e\\u003c/span\\u003e) from the Fowler\\u0026rsquo;s Gap Automatic Weather Station (station number 046128), which is 15 km from where birds were tagged at Gap Hills. Data were provided at minute intervals from which hourly average temperature (\\u0026deg;C) and wind speed (km h\\u003csup\\u003e\\u0026minus;\\u0026thinsp;1\\u003c/sup\\u003e) were calculated.\\u003c/p\\u003e\\n\\u003ch3\\u003eSocial network analyses\\u003c/h3\\u003e\\n\\u003cp\\u003eWe conducted proximity based social network analyses to examine the effects of hourly weather conditions on association rates between zebra finches. For this analysis, we removed localizations where the strongest RSS detected by a receiver was less than \\u0026minus;\\u0026thinsp;80 dB since during this interval the tag was not likely detected by the nearest receiver, which increases the localization error (\\u003cspan citationid=\\\"CR40\\\" class=\\\"CitationRef\\\"\\u003e40\\u003c/span\\u003e). As such, by removing localizations based on weak RSS values, we retained points with less uncertainty. We considered two individuals to be associated when they were within 60 m of one another during the same localization interval. This cutoff is based on previous findings from the same tracking system that social partners, which maintain continual contact, had a median separation distance of 60 m (\\u003cspan citationid=\\\"CR30\\\" class=\\\"CitationRef\\\"\\u003e30\\u003c/span\\u003e). Distances between simultaneously localized dyads were calculated using the R package spatsoc (\\u003cspan citationid=\\\"CR44\\\" class=\\\"CitationRef\\\"\\u003e44\\u003c/span\\u003e). Within each hour interval between 0600 and 1800, we then calculated the number of associations between each pair of tracked birds and the total number of possible associations between the pair (i.e., the total number of intervals when both individuals were simultaneously localized). We removed hour intervals with fewer than 100 dyads (9 out of 145), which left 136 hour-intervals across 12 days.\\u003c/p\\u003e \\u003cp\\u003eHourly association rates were analyzed using Bayesian edge-weighted social networks. This method captures the uncertainty in edge weights (i.e., the association strength) between individuals which can arise when all individuals are not uniformly observed (\\u003cspan citationid=\\\"CR45\\\" class=\\\"CitationRef\\\"\\u003e45\\u003c/span\\u003e). In this framework, the association strength between dyads with a higher sampling effort in each hour will have less uncertainty (i.e., a narrower distribution of edge weights) whereas association strength for dyads with lower sampling effort will have more uncertainty (\\u003cspan citationid=\\\"CR46\\\" class=\\\"CitationRef\\\"\\u003e46\\u003c/span\\u003e). To quantify edge weight uncertainty of the hourly social networks, we used the R package \\u0026ldquo;bisonR\\u0026rdquo; (\\u003cspan citationid=\\\"CR47\\\" class=\\\"CitationRef\\\"\\u003e47\\u003c/span\\u003e) and fit one edge weight model per hour interval. In each hour, an edge weight model is constructed to yield a posterior distribution of possible networks with edge weights that vary based on the sampling effort (i.e., the number of observations) for each dyad. We fit edge weight models where association strength was measured by the number of associations between two individuals (the numerator) relative to the total number of times both individuals were simultaneously observed (the denominator). We used the \\u0026lsquo;fit_edge\\u0026rsquo; function in bisonR to fit a count conjugate model with a weakly informative gamma prior for the edge component of the model. Sampling from within this distribution allows for the uncertainty in the \\u0026lsquo;real\\u0026rsquo; network to be quantified and then incorporated into the calculation of network metrics and Bayesian models (\\u003cspan citationid=\\\"CR45\\\" class=\\\"CitationRef\\\"\\u003e45\\u003c/span\\u003e). From each edge weight model, 400 samples were extracted from the posterior distribution. In bisonR, networks are fully connected since all potential edges have a non-zero probability (\\u003cspan citationid=\\\"CR45\\\" class=\\\"CitationRef\\\"\\u003e45\\u003c/span\\u003e). To calculate network metrics, however, it is necessary to separate dyads that were unlikely to have interacted from those that did. As such, we set edge weights below the minimum observed edge weight to zero (\\u003cspan citationid=\\\"CR46\\\" class=\\\"CitationRef\\\"\\u003e46\\u003c/span\\u003e).\\u003c/p\\u003e \\u003cp\\u003eFrom these observed networks, we then calculated two metrics: density and modularity. Density is a proportional measure of the number of edges relative to the total number of possible edges (see e.g., Evans \\u003cem\\u003eet al.\\u003c/em\\u003e 2021), which was calculated using the function \\u0026lsquo;extract_metric\\u0026rsquo; from bisonR. Modularity is a measure of community structure that expresses the relative density of edges within communities with respect to edges outside communities (\\u003cspan citationid=\\\"CR48\\\" class=\\\"CitationRef\\\"\\u003e48\\u003c/span\\u003e), which we calculated using the functions \\u0026lsquo;cluster_louvain\\u0026rsquo; and \\u0026lsquo;modularity\\u0026rsquo; from the R package igraph (\\u003cspan citationid=\\\"CR49\\\" class=\\\"CitationRef\\\"\\u003e49\\u003c/span\\u003e). We then constructed Bayesian linear models to assess the relationship between the network density and modularity, hour of the day, and the two-way interaction between mean temperature and mean wind speed. Hour was analyzed as a second-order polynomial as zebra finch social activity is typically highest in the morning and evening (\\u003cspan citationid=\\\"CR33\\\" class=\\\"CitationRef\\\"\\u003e33\\u003c/span\\u003e). All predictors were z-transformed and for each we set generic weakly informative normal priors with a mean of zero and a standard deviation of one. To incorporate uncertainty in the analysis of each network metric, we created a dataset from each of the 400 random draws extracted from the posterior distribution of each hourly network. As such, each dataset consisted of 136 values: one for each network metric each day-hour interval. In each model, we set max_treedepth\\u0026thinsp;=\\u0026thinsp;20 and iter\\u0026thinsp;=\\u0026thinsp;10,000. We then used the function \\u0026lsquo;brms_multiple\\u0026rsquo; from the R package \\u0026lsquo;brms\\u0026rsquo; to run one model for each dataset, 400 in total (\\u003cspan citationid=\\\"CR50\\\" class=\\\"CitationRef\\\"\\u003e50\\u003c/span\\u003e). Model convergence was checked by confirming that Rhat in each model was less than 1.05, and additionally, we conducted posterior predictive checks for a random set of 20 models to confirm that the model generated data and the observed data were comparable (\\u003cspan citationid=\\\"CR51\\\" class=\\\"CitationRef\\\"\\u003e51\\u003c/span\\u003e). Ultimately, all models were then combined with the function \\u0026lsquo;combine_models_c\\u0026rsquo; in \\u0026lsquo;brms\\u0026rsquo; (\\u003cspan citationid=\\\"CR50\\\" class=\\\"CitationRef\\\"\\u003e50\\u003c/span\\u003e). If the 95% confidence intervals of the estimated effects encompassed zero, then the data was taken to not strongly support an effect of the covariate on the response.\\u003c/p\\u003e\\n\\u003ch3\\u003eProximity analysis\\u003c/h3\\u003e\\n\\u003cp\\u003eWe additionally assessed the influence of weather on broad scale movement patterns by examining the proximity of neighboring individuals in relation to temperature and wind. To do so, we first fit continuous time movement models to the localization data from each individual using the R package \\u0026lsquo;ctmm\\u0026rsquo; (\\u003cspan citationid=\\\"CR52\\\" class=\\\"CitationRef\\\"\\u003e52\\u003c/span\\u003e). This method can accommodate irregularly sampled data and can incorporate uncertainty in localization estimates by using the error ellipse information calculated through the repeated multilateration approach described above (\\u003cspan citationid=\\\"CR52\\\" class=\\\"CitationRef\\\"\\u003e52\\u003c/span\\u003e, \\u003cspan citationid=\\\"CR53\\\" class=\\\"CitationRef\\\"\\u003e53\\u003c/span\\u003e). After fitting a suite of continuous time movement models, the top-ranked model based on Akaike\\u0026rsquo;s information criteria adjusted for small samples sizes (AICc) was selected. We then segmented the tracking data into hour intervals each day for each individual. Using these hourly subsets and the corresponding movement models, we then compared the proximity between individuals captured within the same sub-site (i.e., nestbox sites B, D, and E). Specifically, we assessed whether two individuals were closer together or farther apart than expected given independent movement. We did so by calculating the proximity ratio between two individuals using the function \\u0026lsquo;proximity\\u0026rsquo; from the R package \\u0026lsquo;ctmm\\u0026rsquo; which compares the mean-square distance between two individuals to the expected distance if the two individuals were moving independently. Proximity ratios less than one indicate that two individuals are closer together than expected given independent movement and ratios greater than one indicate that individuals are farther apart, while proximity ratios overlapping with one indicate that the distance between two individuals is not significantly different given independent movement. We then assessed the relationship between the proximity ratio point estimates and time of day (hour), hourly mean temperature, hourly mean wind speed, and nestbox section with individual, partner, and day as cross random effects using a Bayesian mixed-effect model with generic weakly informative normal priors with a mean of zero and a standard deviation of one and a skewed t-distribution to account for skewness and the presence of outliers in the proximity ratio estimates.\\u003c/p\\u003e\"},{\"header\":\"Results\",\"content\":\"\\u003cp\\u003eWe tracked 128 zebra finches for an average of 10.5 days (range 1\\u0026ndash;12 days) per individual. Each radio tagged individual was localized 1,543 times on average across the 12-day tracking period. Across the tracking period, dyads (i.e., pair-wise associations) were simultaneously localized 98 times on average and were considered to be associated 8 times on average, though there was considerable range in the number of simultaneous localizations (0\\u0026ndash;4,286) and the number of associations (0\\u0026ndash;3,456). During the tracking period, hourly mean temperature ranged from 9\\u0026deg;C to 28\\u0026deg;C and hourly mean wind speed ranged from 1 km/h to 40 km/h. Between 2005 and 2021, the mean hourly temperature in September was 18\\u0026deg;C (5th percentile: 11\\u0026deg;C, 95th percentile: 24\\u0026deg;C) and mean hourly wind speed was 18 km/h (5th percentile: 5 km/h, 95th percentile: 31 km/h).\\u003c/p\\u003e \\u003cp\\u003eNetwork density and modularity both showed large variation across the ranges of weather component values that were recorded during the study period (Fig.\\u0026nbsp;\\u003cspan refid=\\\"Fig2\\\" class=\\\"InternalRef\\\"\\u003e2\\u003c/span\\u003eA and \\u003cspan refid=\\\"Fig2\\\" class=\\\"InternalRef\\\"\\u003e2\\u003c/span\\u003eB). Density, however, showed no association with time of day, hourly mean temperature, or hourly mean wind speed (Table\\u0026nbsp;\\u003cspan refid=\\\"Tab1\\\" class=\\\"InternalRef\\\"\\u003e1\\u003c/span\\u003e). In contrast, predicted network modularity was significantly related to both time of day and mean wind speed, though not mean temperature (Table\\u0026nbsp;\\u003cspan refid=\\\"Tab1\\\" class=\\\"InternalRef\\\"\\u003e1\\u003c/span\\u003e). Modularity was highest at dawn and dusk (estimate hour\\u003csup\\u003e2\\u003c/sup\\u003e: 0.38, 0.21\\u0026ndash;0.55; Fig.\\u0026nbsp;\\u003cspan refid=\\\"Fig2\\\" class=\\\"InternalRef\\\"\\u003e2\\u003c/span\\u003eB) and increased with mean wind speed (estimate wind speed: 0.03, 0.01\\u0026ndash;0.04; Fig.\\u0026nbsp;\\u003cspan refid=\\\"Fig2\\\" class=\\\"InternalRef\\\"\\u003e2\\u003c/span\\u003eB, Fig.\\u0026nbsp;\\u003cspan refid=\\\"Fig3\\\" class=\\\"InternalRef\\\"\\u003e3\\u003c/span\\u003e).\\u003c/p\\u003e \\u003cp\\u003e \\u003c/p\\u003e \\u003cp\\u003e \\u003cdiv class=\\\"gridtable\\\"\\u003e\\u003ctable float=\\\"Yes\\\" id=\\\"Tab1\\\" border=\\\"1\\\"\\u003e \\u003ccaption language=\\\"En\\\"\\u003e \\u003cdiv class=\\\"CaptionNumber\\\"\\u003eTable 1\\u003c/div\\u003e \\u003cdiv class=\\\"CaptionContent\\\"\\u003e \\u003cp\\u003eParameter estimates and 95% CI for Bayesian models of the relationship between time of day and weather components and network density and modularity. Terms in bold indicate parameters (not including the intercept) where the coefficient estimate 95% CI did not overlap zero.\\u003c/p\\u003e \\u003c/div\\u003e \\u003c/caption\\u003e \\u003ccolgroup cols=\\\"5\\\"\\u003e \\u003cdiv align=\\\"left\\\" class=\\\"colspec\\\" colname=\\\"c1\\\" colnum=\\\"1\\\"\\u003e\\u003c/div\\u003e \\u003cdiv align=\\\"left\\\" class=\\\"colspec\\\" colname=\\\"c2\\\" colnum=\\\"2\\\"\\u003e\\u003c/div\\u003e \\u003cdiv align=\\\"left\\\" class=\\\"colspec\\\" colname=\\\"c3\\\" colnum=\\\"3\\\"\\u003e\\u003c/div\\u003e \\u003cdiv align=\\\"left\\\" class=\\\"colspec\\\" colname=\\\"c4\\\" colnum=\\\"4\\\"\\u003e\\u003c/div\\u003e \\u003cdiv align=\\\"left\\\" class=\\\"colspec\\\" colname=\\\"c5\\\" colnum=\\\"5\\\"\\u003e\\u003c/div\\u003e \\u003cthead\\u003e \\u003ctr\\u003e \\u003cth align=\\\"left\\\" colname=\\\"c1\\\"\\u003e\\u0026nbsp;\\u003c/th\\u003e \\u003cth align=\\\"left\\\" colspan=\\\"2\\\" nameend=\\\"c3\\\" namest=\\\"c2\\\"\\u003e \\u003cp\\u003eDensity\\u003c/p\\u003e \\u003c/th\\u003e \\u003cth align=\\\"left\\\" colspan=\\\"2\\\" nameend=\\\"c5\\\" namest=\\\"c4\\\"\\u003e \\u003cp\\u003eModularity\\u003c/p\\u003e \\u003c/th\\u003e \\u003c/tr\\u003e \\u003c/thead\\u003e \\u003ctbody\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c1\\\"\\u003e \\u003cp\\u003e\\u003cem\\u003ePredictors\\u003c/em\\u003e\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003e\\u003cem\\u003eEstimates\\u003c/em\\u003e\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003e\\u003cem\\u003eCI (95%)\\u003c/em\\u003e\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c4\\\"\\u003e \\u003cp\\u003e\\u003cem\\u003eEstimates\\u003c/em\\u003e\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c5\\\"\\u003e \\u003cp\\u003e\\u003cem\\u003eCI (95%)\\u003c/em\\u003e\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c1\\\"\\u003e \\u003cp\\u003eIntercept\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003e0.11\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003e0.10\\u0026ndash;0.11\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c4\\\"\\u003e \\u003cp\\u003e0.52\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c5\\\"\\u003e \\u003cp\\u003e0.51\\u0026ndash;0.54\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c1\\\"\\u003e \\u003cp\\u003eHour\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003e0.08\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003e-0.04\\u0026ndash;0.19\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c4\\\"\\u003e \\u003cp\\u003e0.03\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c5\\\"\\u003e \\u003cp\\u003e-0.23\\u0026ndash;0.30\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c1\\\"\\u003e \\u003cp\\u003eHour^2\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003e0.02\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003e-0.05\\u0026ndash;0.09\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c4\\\"\\u003e \\u003cp\\u003e\\u003cb\\u003e0.38\\u003c/b\\u003e\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c5\\\"\\u003e \\u003cp\\u003e\\u003cb\\u003e0.21\\u0026ndash;0.55\\u003c/b\\u003e\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c1\\\"\\u003e \\u003cp\\u003eWind speed\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003e0.00\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003e-0.01\\u0026ndash;0.01\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c4\\\"\\u003e \\u003cp\\u003e\\u003cb\\u003e0.03\\u003c/b\\u003e\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c5\\\"\\u003e \\u003cp\\u003e\\u003cb\\u003e0.01\\u0026ndash;0.04\\u003c/b\\u003e\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c1\\\"\\u003e \\u003cp\\u003eTemperature\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003e-0.00\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003e-0.01\\u0026ndash;0.01\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c4\\\"\\u003e \\u003cp\\u003e0.01\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c5\\\"\\u003e \\u003cp\\u003e-0.02\\u0026ndash;0.03\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c1\\\"\\u003e \\u003cp\\u003eWind speed * Temperature\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003e-0.00\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003e-0.01\\u0026ndash;0.00\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c4\\\"\\u003e \\u003cp\\u003e0.00\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c5\\\"\\u003e \\u003cp\\u003e-0.01\\u0026ndash;0.02\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c1\\\"\\u003e \\u003cp\\u003eObservations\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colspan=\\\"2\\\" nameend=\\\"c3\\\" namest=\\\"c2\\\"\\u003e \\u003cp\\u003e136\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colspan=\\\"2\\\" nameend=\\\"c5\\\" namest=\\\"c4\\\"\\u003e \\u003cp\\u003e136\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c1\\\"\\u003e \\u003cp\\u003eR\\u003csup\\u003e2\\u003c/sup\\u003e\\u0026nbsp;Bayes\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colspan=\\\"2\\\" nameend=\\\"c3\\\" namest=\\\"c2\\\"\\u003e \\u003cp\\u003e0.071\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colspan=\\\"2\\\" nameend=\\\"c5\\\" namest=\\\"c4\\\"\\u003e \\u003cp\\u003e0.288\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003c/tbody\\u003e \\u003c/colgroup\\u003e \\u003c/table\\u003e\\u003c/div\\u003e \\u003c/p\\u003e \\u003cp\\u003e \\u003c/p\\u003e \\u003cp\\u003eProximity ratio confidence intervals did not overlap with one for 74%, 70%, and 63% of the comparisons between individuals from the same neighborhood, the sub-sites B, D, and E, respectively, indicating that individuals within the same area of the colony tended to follow similar movement patterns throughout the day. We did not, however, find evidence of a relationship between hour of day, hourly mean temperature, or hourly mean wind speed and proximity ratios (Table\\u0026nbsp;\\u003cspan refid=\\\"Tab2\\\" class=\\\"InternalRef\\\"\\u003e2\\u003c/span\\u003e). Proximity ratios were higher in colony section E (estimate: 0.06, 0.04\\u0026ndash;0.09) compared to the other two sections.\\u003c/p\\u003e \\u003cp\\u003e \\u003cdiv class=\\\"gridtable\\\"\\u003e\\u003ctable float=\\\"Yes\\\" id=\\\"Tab2\\\" border=\\\"1\\\"\\u003e \\u003ccaption language=\\\"En\\\"\\u003e \\u003cdiv class=\\\"CaptionNumber\\\"\\u003eTable 2\\u003c/div\\u003e \\u003cdiv class=\\\"CaptionContent\\\"\\u003e \\u003cp\\u003eParameter estimates and 95% CI for Bayesian models of the relationship between time of day and weather components and proximity ratios of neighbouring individuals. Terms in bold indicate parameters (not including the intercept) where the coefficient estimate 95% CI did not overlap zero.\\u003c/p\\u003e \\u003c/div\\u003e \\u003c/caption\\u003e \\u003ccolgroup cols=\\\"3\\\"\\u003e \\u003cdiv align=\\\"left\\\" class=\\\"colspec\\\" colname=\\\"c1\\\" colnum=\\\"1\\\"\\u003e\\u003c/div\\u003e \\u003cdiv align=\\\"left\\\" class=\\\"colspec\\\" colname=\\\"c2\\\" colnum=\\\"2\\\"\\u003e\\u003c/div\\u003e \\u003cdiv align=\\\"left\\\" class=\\\"colspec\\\" colname=\\\"c3\\\" colnum=\\\"3\\\"\\u003e\\u003c/div\\u003e \\u003cthead\\u003e \\u003ctr\\u003e \\u003cth align=\\\"left\\\" colname=\\\"c1\\\"\\u003e\\u0026nbsp;\\u003c/th\\u003e \\u003cth align=\\\"left\\\" colspan=\\\"2\\\" nameend=\\\"c3\\\" namest=\\\"c2\\\"\\u003e \\u003cp\\u003eProximity\\u003c/p\\u003e \\u003c/th\\u003e \\u003c/tr\\u003e \\u003c/thead\\u003e \\u003ctbody\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c1\\\"\\u003e \\u003cp\\u003e\\u003cem\\u003ePredictors\\u003c/em\\u003e\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003e\\u003cem\\u003eEstimates\\u003c/em\\u003e\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003e\\u003cem\\u003eCI (95%)\\u003c/em\\u003e\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c1\\\"\\u003e \\u003cp\\u003eIntercept (Section B)\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003e0.05\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003e0.04\\u0026ndash;0.07\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c1\\\"\\u003e \\u003cp\\u003eHour\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003e0.01\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003e0.00\\u0026ndash;0.01\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c1\\\"\\u003e \\u003cp\\u003eWind speed\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003e0.00\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003e0.00\\u0026ndash;0.00\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c1\\\"\\u003e \\u003cp\\u003eTemperature\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003e-0.01\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003e-0.01\\u0026nbsp;\\u0026ndash;\\u0026nbsp;-0.01\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c1\\\"\\u003e \\u003cp\\u003eWind speed * Temperature\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003e0.00\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003e0.00\\u0026ndash;0.00\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c1\\\"\\u003e \\u003cp\\u003eSection D\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003e0.00\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003e-0.02\\u0026ndash;0.02\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c1\\\"\\u003e \\u003cp\\u003eSection E\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003e\\u003cb\\u003e0.06\\u003c/b\\u003e\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003e\\u003cb\\u003e0.04\\u0026ndash;0.09\\u003c/b\\u003e\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c1\\\"\\u003e \\u003cp\\u003eObservations\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colspan=\\\"2\\\" nameend=\\\"c3\\\" namest=\\\"c2\\\"\\u003e \\u003cp\\u003e48,062\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c1\\\"\\u003e \\u003cp\\u003eMarginal R\\u003csup\\u003e2\\u003c/sup\\u003e\\u0026nbsp;/ Conditional R\\u003csup\\u003e2\\u003c/sup\\u003e\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colspan=\\\"2\\\" nameend=\\\"c3\\\" namest=\\\"c2\\\"\\u003e \\u003cp\\u003e0.013 / 0.026\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003c/tbody\\u003e \\u003c/colgroup\\u003e \\u003c/table\\u003e\\u003c/div\\u003e \\u003c/p\\u003e\"},{\"header\":\"Discussion\",\"content\":\"\\u003cp\\u003eUsing automated radio tracking of 128 wild zebra finches across 12 days, we show that social organization was affected by hourly changes in weather. During periods with higher wind, social network modularity increased, indicating stronger separation between sub-groups among the individuals we tracked. In contrast to wind speed, the moderate temperatures during our tracking period did not affect either social network modularity or density. Modularity, however, did change significantly across the day, with higher levels in the mornings and evenings. Together, these findings indicate that the social network structure of wild populations can vary significantly even on short (hourly) time scales, demonstrating substantial social plasticity in response to environmental fluctuations. These fluctuations in social plasticity, especially in network modularity could have relevant consequences for the effective transfer of social information on ephemeral resources and predation threats.\\u003c/p\\u003e \\u003cp\\u003eObtaining insights into the extent of social plasticity in wild animal societies is crucial for estimating their capacity to cope with environmental change through mechanisms such as social buffering (\\u003cspan citationid=\\\"CR16\\\" class=\\\"CitationRef\\\"\\u003e16\\u003c/span\\u003e, \\u003cspan citationid=\\\"CR54\\\" class=\\\"CitationRef\\\"\\u003e54\\u003c/span\\u003e). Additionally, knowing baseline social plasticity will help to detect when social structures move beyond their \\u0026lsquo;normal\\u0026rsquo; range and thus when potentially concerning situations arise (Snijders et al., 2017). We observed within-day variation in modularity, which most likely reflects the different activities of zebra finches have during the day. Specifically, the higher modularity early and late in the day most likely results from individuals roosting in pairs or small subgroups at roosting colonies throughout the study area (\\u003cspan citationid=\\\"CR33\\\" class=\\\"CitationRef\\\"\\u003e33\\u003c/span\\u003e). During the day when not foraging, birds also spend substantial time at various social hotspots and at water sources where they occur in larger groups (\\u003cspan citationid=\\\"CR32\\\" class=\\\"CitationRef\\\"\\u003e32\\u003c/span\\u003e). These larger gatherings might be relevant for information transfer as individuals mix with others from different subsites. In contrast, we observed that social network density remained relatively constant throughout the day and was unaffected by either wind or temperature. Similarly, we observed that the movements of individuals caught in the same neighborhood were moderately correlated, as indicated by the proximity ratios, suggesting that groups of neighboring individuals were more cohesive than expected given random movement.\\u003c/p\\u003e \\u003cp\\u003eMultiple factors may explain the observed increased social network modularity in relation to wind. For one, due to their small size, zebra finches experience greater flight costs from strong winds (\\u003cspan citationid=\\\"CR55\\\" class=\\\"CitationRef\\\"\\u003e55\\u003c/span\\u003e), which may limit movement and increase aggregations at sheltered locations. Additionally, increased wind speeds reduce flight maneuverability (\\u003cspan citationid=\\\"CR56\\\" class=\\\"CitationRef\\\"\\u003e56\\u003c/span\\u003e), limiting the ability to evade aerial predators, which may reduce the willingness to fly in higher winds. Conceivably, individuals carrying a radio tag might be further deterred from flying given the effects of carrying an increased load on flight performance (\\u003cspan citationid=\\\"CR57\\\" class=\\\"CitationRef\\\"\\u003e57\\u003c/span\\u003e, \\u003cspan citationid=\\\"CR58\\\" class=\\\"CitationRef\\\"\\u003e58\\u003c/span\\u003e). Yet, given that we have observed tagged individuals move around substantially (Tyson et al. 2024) and performing normal behaviors, such as nest building and nestling provisioning, and that we have detected birds more than 24 months after they were tagged, the impacts of the tags appear to be minimal and thus device-effects are unlikely to explain the patterns observed here. Another potential explanation for reduced movement during windier periods is that predator detection may be diminished due to attenuated acoustic cues (\\u003cspan citationid=\\\"CR59\\\" class=\\\"CitationRef\\\"\\u003e59\\u003c/span\\u003e), which may also limit willingness to fly. Additionally, both reduced flight maneuverability and diminished hearing (\\u003cspan citationid=\\\"CR60\\\" class=\\\"CitationRef\\\"\\u003e60\\u003c/span\\u003e) may also jeopardies pair cohesion, which is an important component of zebra finch social structure (30) as high wind speeds may make it challenging for zebra finch pairs to remain together and to reunite once separated. Given that zebra finch calls are only audible at distances up to about 14 m (\\u003cspan citationid=\\\"CR61\\\" class=\\\"CitationRef\\\"\\u003e61\\u003c/span\\u003e), and less in windy conditions, becoming separated may be a severe consequence of higher wind speeds, especially given the uncertainty as to how long windy conditions will persist. The moderate temperatures had no effect on social organization. Animals typically move less at extreme temperatures, seek shade at higher temperatures, and seek warmth and shelter at very low temperatures (\\u003cspan citationid=\\\"CR62\\\" class=\\\"CitationRef\\\"\\u003e62\\u003c/span\\u003e). While temperatures in the arid zone can well reach beyond 45 degrees in the shade in summer, the moderate temperature of only up to 29 degree C the birds experienced during tracking, most likely did not require any trade off movement and social organization. Taken together for the effects of wind, the higher costs in flying as well the potential threats from increased vulnerability to predators and difficulty in maintaining pair cohesion may explain the observed increase in social network modularity in wild zebra finches in response to higher wind speeds.\\u003c/p\\u003e \\u003cp\\u003eWith respect to the consequences of increased social network modularity, if this change results from individuals reducing connectivity to unfamiliar individuals, it may limit exposure to novel sources of information and thus limit social information transfer (\\u003cspan citationid=\\\"CR38\\\" class=\\\"CitationRef\\\"\\u003e38\\u003c/span\\u003e, \\u003cspan citationid=\\\"CR63\\\" class=\\\"CitationRef\\\"\\u003e63\\u003c/span\\u003e). Moreover, higher modularity could impact collective decisions such as when to initiate nomadic movements in search of favorable conditions or synchronizing breeding activity (\\u003cspan citationid=\\\"CR64\\\" class=\\\"CitationRef\\\"\\u003e64\\u003c/span\\u003e). Both activities could have potentially significant population impacts. Modularity indeed plays a key role in social transmission, both of information and of infections, although in different (non-linear) ways. Strong modularity is predicted to \\u0026lsquo;trap\\u0026rsquo; simple disease infections into subgroups, but depending on the learning strategy of individuals, modularity does not necessarily affect social transmission of information (\\u003cspan citationid=\\\"CR38\\\" class=\\\"CitationRef\\\"\\u003e38\\u003c/span\\u003e). Recent modelling work by Evans \\u003cem\\u003eet al.\\u003c/em\\u003e (2021) demonstrated that highly fragmented structures with small group sizes, such as fission-fusion populations and multi-level societies, e.g., wild zebra finch populations, can be most effective in trading-off reduced infection risk with effective social information transfer. In their modelling, additional variation in modularity did not affect the speed of information transfer when considering relatively higher network densities (\\u003cspan citationid=\\\"CR38\\\" class=\\\"CitationRef\\\"\\u003e38\\u003c/span\\u003e). However, Evans et al (2021), focused on complex information transfer mechanisms (e.g. copying the majority), while time-sensitive fitness-relevant information (e.g., detection of predators or water and food in resource-scarce environments) is more likely to transfer through knowledgeable individuals rather than the majority. In these cases, an increase from intermediate modularity to higher modularity, such as in our study, could hamper information flow, especially in structures with small subgroups (\\u003cspan citationid=\\\"CR65\\\" class=\\\"CitationRef\\\"\\u003e65\\u003c/span\\u003e). Indeed, transfer of information in theoretical and empirical datasets was revealed most efficient at intermediate modularity levels (\\u003cspan citationid=\\\"CR65\\\" class=\\\"CitationRef\\\"\\u003e65\\u003c/span\\u003e). Thus, when considering urgent information about the presence of predators or resources, frequent changes in modularity could have population-level effects which would need to be further explored.\\u003c/p\\u003e \\u003cp\\u003eSocial buffering is described as a process through which the negative effects of stress are ameliorated by sociality (\\u003cspan citationid=\\\"CR54\\\" class=\\\"CitationRef\\\"\\u003e54\\u003c/span\\u003e, \\u003cspan citationid=\\\"CR66\\\" class=\\\"CitationRef\\\"\\u003e66\\u003c/span\\u003e), often by individuals aggregating in larger groups or by associating with familiar individuals (\\u003cspan citationid=\\\"CR18\\\" class=\\\"CitationRef\\\"\\u003e18\\u003c/span\\u003e). Here, we observed that modularity was highest during periods with higher mean wind speeds, suggesting that if social buffering occurs during these periods, it is not through individuals forming larger aggregations. Alternatively, in zebra finches, social buffering may act through small groups, since the preferred social environment of wild zebra finches typically consists of at least two familiar individuals, i.e., the breeding pair, and, most likely other familiar, individuals with overlapping home ranges within the sub-colony (\\u003cspan citationid=\\\"CR30\\\" class=\\\"CitationRef\\\"\\u003e30\\u003c/span\\u003e, \\u003cspan citationid=\\\"CR34\\\" class=\\\"CitationRef\\\"\\u003e34\\u003c/span\\u003e).\\u003c/p\\u003e\"},{\"header\":\"Conclusions\",\"content\":\"\\u003cp\\u003eOur findings that weather, specifically wind, drives social structure in wild zebra finches contributes to our understanding of how environmental conditions drive social behavior and the potential for social buffering. While many factors affect animal movements, our findings demonstrate that even short-term and relatively benign influences, such as moderate fluctuations in weather, affect how animal societies are organized, which likely has implications for information flow and effective anti-predator behavior. Our findings are thus a first step in highlighting the direction and degree of social flexibility that currently exists in response to daily environmental fluctuations in a wild population. Furthermore, understanding animal societies, such as zebra finches, that have evolved under unpredictable climatic conditions, may be valuable for assessing how social animals cope with an increasingly extreme and variable climate.\\u003c/p\\u003e\"},{\"header\":\"Declarations\",\"content\":\"\\u003cp\\u003e\\u003cstrong\\u003eEthics approval and consent to participate\\u003c/strong\\u003e\\u003c/p\\u003e\\n\\u003cp\\u003eBird handling and tagging was performed in accordance with the Authority (#2018/027) of the Animal Ethics Committee at Macquarie University, and the Australian Bird \\u0026amp; Bat Banding Scheme.\\u003c/p\\u003e\\n\\u003cp\\u003e\\u003cstrong\\u003eConsent for publication\\u003c/strong\\u003e\\u003c/p\\u003e\\n\\u003cp\\u003eNot applicable.\\u003c/p\\u003e\\n\\u003cp\\u003e\\u003cstrong\\u003eCompeting interests\\u003c/strong\\u003e\\u003c/p\\u003e\\n\\u003cp\\u003eThe authors declare no competing interests.\\u003c/p\\u003e\\n\\u003cp\\u003e\\u003cstrong\\u003eFunding Declaration\\u003c/strong\\u003e\\u003c/p\\u003e\\n\\u003cp\\u003eThe project was supported in part by a Data Science / Artificial Intelligence grant to CT and a Dutch Research Council (NWO) grant to MN (ALWOP.334).\\u003c/p\\u003e\\n\\u003cp\\u003e\\u003cstrong\\u003eAvailability of data and materials\\u003c/strong\\u003e\\u003c/p\\u003e\\n\\u003cp\\u003eData and code to reproduce this analysis is available at the anonymized GitHub repository: https://anonymous.4open.science/r/zebby_social_buffering-BF5E/\\u003c/p\\u003e\\n\\u003cp\\u003e\\u003cstrong\\u003eAuthor contributions\\u003c/strong\\u003e\\u003c/p\\u003e\\n\\u003cp\\u003e\\u003cstrong\\u003eChris Tyson\\u003c/strong\\u003e: Conceptualization, Methodology, Software, Validation, Formal analysis, Investigation, Writing - Original Draft, Visualization, Supervision, Funding acquisition. \\u003cstrong\\u003eHugo Loning\\u003c/strong\\u003e: Conceptualization, Methodology, Investigation, Data curation, Reviewing and Editing. \\u003cstrong\\u003eNo\\u0026euml;lle Tschirren\\u003c/strong\\u003e: Conceptualization, Data Curation, Writing - Review \\u0026amp; Editing. \\u003cstrong\\u003eElke Molenaar\\u003c/strong\\u003e: Investigation, Data curation. \\u003cstrong\\u003eLysanne Snijders\\u003c/strong\\u003e: Conceptualization, Writing - Original Draft, Writing - Review \\u0026amp; Editing. Supervision. \\u003cstrong\\u003eSimon Griffith\\u003c/strong\\u003e: Conceptualization, Methodology, Investigation, Data Curation, Reviewing and Editing, Project administration. \\u003cstrong\\u003eMarc Naguib\\u003c/strong\\u003e: Conceptualization, Methodology, Investigation, Data Curation, Writing - Original Draft, Reviewing and Editing, Supervision, Project administration, Funding acquisition\\u003c/p\\u003e\\n\\u003cp\\u003e\\u003cstrong\\u003eAcknowledgements\\u003c/strong\\u003e\\u003c/p\\u003e\\n\\u003cp\\u003eWe are grateful to Lyanne Brouwers and Rita Fragueira for fieldwork assistance. We also thank Garry Dowling and Mark Tilley for assistance installing components of the automated radio tracking system and Vicki Dowling for logistical support.\\u0026nbsp;\\u003c/p\\u003e\"},{\"header\":\"References\",\"content\":\"\\u003col\\u003e\\u003cli\\u003e\\u003cspan\\u003eKurvers RHJM, Krause J, Croft DP, Wilson ADM, Wolf M. The evolutionary and ecological consequences of animal social networks: Emerging issues. Trends Ecol Evol. 2014;29(6):326\\u0026ndash;35.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eKrause J, Ruxton GD. Living in Groups. Oxford University PressOxford; 2002.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003ePapageorgiou D, Farine DR. Multilevel Societies in Birds. Trends Ecol Evol. 2021;36(1):15\\u0026ndash;7.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eSnijders L, Blumstein DT, Stanley CR, Franks DW. Animal Social Network Theory Can Help Wildlife Conservation. Trends Ecol Evol. 2017;32(8):567\\u0026ndash;77.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eGartland LA, Firth JA, Laskowski KL, Jeanson R, Ioannou CC. Sociability as a personality trait in animals: methods, causes and consequences. Biol Rev. 2022;97(2):802\\u0026ndash;16.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eBeck KB, Farine DR, Kempenaers B. Social network position predicts male mating success in a small passerine. Behav Ecol. 2021;32(5):856\\u0026ndash;64.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eWilliams AE, Worsley-Tonks KEL, Ezenwa VO. Drivers and consequences of variation in individual social connectivity. Anim Behav. 2017;133:1\\u0026ndash;9.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eAplin LM, Firth JA, Farine DR, Voelkl B, Crates RA, Culina A, et al. Consistent individual differences in the social phenotypes of wild great tits, Parus major. Anim Behav. 2015;108:117\\u0026ndash;27.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eHofmann HA, Beery AK, Blumstein DT, Couzin ID, Earley RL, Hayes LD, et al. An evolutionary framework for studying mechanisms of social behavior. Trends Ecol Evol. 2014;29(10):581\\u0026ndash;9.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eFisher DN, Kilgour RJ, Siracusa ER, Foote JR, Hobson EA, Montiglio PO, et al. Anticipated effects of abiotic environmental change on intraspecific social interactions. Biol Rev. 2021;96(6):2661\\u0026ndash;93.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eBrandl HB, Farine DR. Stress in the social environment: behavioural and social consequences of stress transmission in bird flocks. Proc R Soc B Biol Sci. 2024;291(2034).\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eCamerlenghi E, Nolazco S, Farine DR, Magrath RD, Peters A. Social restructuring during harsh environmental conditions promotes cooperative behaviour in a songbird. Proc R Soc B Biol Sci. 2024;291(2021).\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eDingle H, Migration. The Biology of Life on the Move. Oxford University Press; 2014.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eAikens EO, Bontekoe ID, Blumenstiel L, Schlicksupp A, Flack A. Viewing animal migration through a social lens. Trends Ecol Evol. 2022;37(11):985\\u0026ndash;96.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eSayin S, Couzin-Fuchs E, Petelski I, G\\u0026uuml;nzel Y, Salahshour M, Lee C-Y, et al. The behavioral mechanisms governing collective motion in swarming locusts. Sci (80-). 2025;387(6737):995\\u0026ndash;1000.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eBlumstein DT, Hayes LD, Pinter-Wollman N. Social consequences of rapid environmental change. Vol. 38, Trends in Ecology and Evolution. Elsevier Ltd; 2023. pp. 337\\u0026ndash;45.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eBaigger A, Perony N, Reuter M, Leinert V, Melber M, Gr\\u0026uuml;nberger S, et al. Bechstein\\u0026rsquo;s bats maintain individual social links despite a complete reorganisation of their colony structure. Naturwissenschaften. 2013;100(9):895\\u0026ndash;8.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eTestard C, Larson SM, Watowich MM, Kaplinsky CH, Bernau A, Faulder M, et al. Rhesus macaques build new social connections after a natural disaster. Curr Biol. 2021;31(11):2299\\u0026ndash;e23097.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eSchaffner CM, Rebecchini L, Ramos-Fernandez G, Vick LG, Aureli F. Spider Monkeys (Ateles geoffroyi yucatenensis) Cope with the Negative Consequences of Hurricanes Through Changes in Diet, Activity Budget, and Fission-Fusion Dynamics. Int J Primatol. 2012;33(4):922\\u0026ndash;36.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eNourani E, Safi K, de Grissac S, Anderson DJ, Cole NC, Fell A, et al. Seabird morphology determines operational wind speeds, tolerable maxima, and responses to extremes. Curr Biol. 2023;33(6):1179\\u0026ndash;e11843.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eThorne LH, Clay TA, Phillips RA, Silvers LG, Wakefield ED. Effects of wind on the movement, behavior, energetics, and life history of seabirds. Mar Ecol Prog Ser. 2023;723:73\\u0026ndash;117.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eKranstauber B, Weinzierl R, Wikelski M, Safi K. Global aerial flyways allow efficient travelling. Ecol Lett. 2015;18(12):1338\\u0026ndash;45.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eVansteelant WMG, Bouten W, Klaassen RHG, Koks BJ, Schlaich AE, van Diermen J, et al. Regional and seasonal flight speeds of soaring migrants and the role of weather conditions at hourly and daily scales. J Avian Biol. 2015;46(1):25\\u0026ndash;39.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eCornioley T, B\\u0026ouml;rger L, Ozgul A, Weimerskirch H. Impact of changing wind conditions on foraging and incubation success in male and female wandering albatrosses. J Anim Ecol. 2016;85(5):1318\\u0026ndash;27.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eAlerstam T. Conflicting evidence about long-distance animal navigation. Science. 2006;313(5788):791\\u0026ndash;4.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eHedenstrom A, Alerstam T. Optimal flight speed of birds. Philos Trans R Soc Lond Ser B Biol Sci. 1995;348(1326):471\\u0026ndash;87.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eShepard ELC, Wilson RP, Rees WG, Grundy E, Lambertucci SA, Vosper SB. Energy landscapes shape animal movement ecology. Am Nat. 2013;182(3):298\\u0026ndash;312.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eRat M, Mathe-Hubert H, McKechnie AE, Sueur C, Cunningham SJ. Extreme and variable environmental temperatures are linked to reduction of social network cohesiveness in a highly social passerine. Oikos. 2020;129(11):1597\\u0026ndash;610.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eBorthwick Z, Quiring K, Griffith SC, Leu ST. Heat stress conditions affect the social network structure of free-ranging sheep. Ecol Evol. 2024;14(2).\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eTyson C, Loning H, Griffith SC, Naguib M. Constant companions: wild zebra finch pairs display extreme spatial cohesion. Biol Lett. 2024;20(11):2024.03.21.586046.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eMcCowan LSC, Mariette MM, Griffith SC. The size and composition of social groups in the wild zebra finch. Emu. 2015;115(3):191\\u0026ndash;8.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eLoning H, Fragueira R, Naguib M, Griffith SC. Hanging out in the outback: the use of social hotspots by wild zebra finches. J Avian Biol. 2023;2023:11\\u0026ndash;2.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eZann RA. The zebra finch: a synthesis of field and laboratory studies. Oxford: Oxford University Press; 1996.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eBrandl HB, Griffith SC, Farine DR, Schuett W. Wild zebra finches that nest synchronously have long-term stable social ties. J Anim Ecol. 2021;90(1):76\\u0026ndash;86.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eMorton SR, Stafford Smith DM, Dickman CR, Dunkerley DL, Friedel MH, McAllister RRJ, et al. A fresh framework for the ecology of arid Australia. J Arid Environ. 2011;75(4):313\\u0026ndash;29.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eLoarie SR, Duffy PB, Hamilton H, Asner GP, Field CB, Ackerly DD. The velocity of climate change. Nature. 2009;462(7276):1052\\u0026ndash;5.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eFunghi C, McCowan LSC, Schuett W, Griffith SC. High air temperatures induce temporal, spatial and social changes in the foraging behaviour of wild zebra finches. Anim Behav. 2019;149:33\\u0026ndash;43.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eEvans JC, Hodgson DJ, Boogert NJ, Silk MJ. Group size and modularity interact to shape the spread of infection and information through animal societies. Behav Ecol Sociobiol. 2021;75(12).\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eJirinec V, Rodrigues PF, Amaral B. Adjustable leg harness for attaching tags to small and medium-sized birds. J F Ornithol. 2021;92(1):77\\u0026ndash;87.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eTyson C, Fragueira R, Sansano-Sansano E, Yu H, Naguib M. Fingerprint localisation for fine-scale wildlife tracking using automated radio telemetry. Methods Ecol Evol. 2024;2024(April):2118\\u0026ndash;28.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003ePaxton KL, Baker KM, Crytser ZB, Guinto RMP, Brinck KW, Rogers HS et al. Optimizing trilateration estimates for tracking fine-scale movement of wildlife using automated radio telemetry networks. Ecol Evol. 2022;12(2).\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eR Core Team. R: A Language and Environment for Statistical Computing. Vienna, Austria: R Foundation for Statistical Computing; 2024.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eWhitehouse K, Karlof C, Culler D. A Practical Evaluation of Radio Signal Strength for Ranging-based Localization. ACM SIGMOBILE Mob Comput Commun Rev. 2007;11(1):41\\u0026ndash;52.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eRobitaille AL, Webber QMR, Vander Wal E. Conducting social network analysis with animal telemetry data: Applications and methods using spatsoc. Methods Ecol Evol. 2019;10(8):1203\\u0026ndash;11.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eHart J, Weiss MN, Franks D, Brent L. BISoN: A Bayesian framework for inference of social networks. Methods Ecol Evol. 2023;14(9):2411\\u0026ndash;20.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003ePereira AS, Pavez-Fox MA, Hart JDA, Valle JEN-D, Phillips D, Casanova C et al. The impact of kinship composition on social structure. bioRxiv. 2024;(1):2024.01.11.575037.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eHart JDA, Franks DW, Brent LJN, Weiss MN. bisonR - Bayesian Inference of Social Networks with R. 2022.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eBlondel VD, Guillaume J-L, Lambiotte R, Lefebvre E. Fast unfolding of communities in large networks. J Stat Mech Theory Exp. 2008;2008(10):P10008.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eCsardi G, Nepusz T. The igraph software package for complex network research. InterJournal Complex Syst. 2006;Complex Sy(1695):1695.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eB\\u0026uuml;rkner PC. brms: An R package for Bayesian multilevel models using Stan. J Stat Softw. 2017;80(1).\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eGelman A, Hill J. Data Analysis Using Regression and Multilevel/Hierarchical Models. Cambridge University Press; 2006.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eCalabrese JM, Fleming CH, Gurarie E. ctmm: an r package for analyzing animal relocation data as a continuous-time stochastic process. Freckleton R, editor. Methods Ecol Evol. 2016;7(9):1124\\u0026ndash;32.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eFleming CH, Drescher-Lehman J, Noonan MJ, Akre TSB, Brown DJ, Cochrane MM et al. A comprehensive framework for handling location error in animal tracking data. 2020. p. 31.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eKomdeur J, Ma L. Keeping up with environmental change: The importance of sociality. Ethology. 2021;127(10):790\\u0026ndash;807.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eLiechti F. Birds: Blowin\\u0026rsquo; by the wind? J Ornithol. 2006;147(2):202\\u0026ndash;11.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eShepard E, Cole EL, Neate A, Lempidakis E, Ross A. Wind prevents cliff-breeding birds from accessing nests through loss of flight control. Elife. 2019;8:1\\u0026ndash;15.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eTomotani BM, Muijres FT, Johnston B, van der Jeugd HP, Naguib M. Great tits do not compensate over time for a radio-tag-induced reduction in escape-flight performance. Ecol Evol. 2021;11(23):16600\\u0026ndash;17.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eLapsansky AB, Igo JA, Tobalske BW. Zebra finch (Taeniopygia guttata) shift toward aerodynamically efficient flight kinematics in response to an artificial load. Biol Open. 2019;8(6):1\\u0026ndash;9.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eCherry MJ, Barton BT. Effects of wind on predator-prey interactions. Food Webs. 2017;13:92\\u0026ndash;7.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eBrumm H, Naguib M. In. Chapter 1 Environmental Acoustics and the Evolution of Bird Song. 2009. pp. 1\\u0026ndash;33.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eLoning H, Griffith SC, Naguib M. Zebra finch song is a very short-range signal in the wild: evidence from an integrated approach. Behav Ecol. 2022;33(1):37\\u0026ndash;46.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eSharpe LL, Prober SM, Gardner JL. the Hot Seat: Behavioral Change and Old-Growth Trees Underpin an Australian Songbird\\u0026rsquo;s Response to Extreme Heat. Front Ecol Evol. 2022;10(March):1\\u0026ndash;19.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eDuboscq J, Romano V, MacIntosh A, Sueur C. Social information transmission in animals: Lessons from studies of diffusion. Front Psychol. 2016;7(AUG):1\\u0026ndash;15.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eBrandl HB, Griffith SC, Schuett W. Wild zebra finches choose neighbours for synchronized breeding. Anim Behav. 2019;151:21\\u0026ndash;8.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eRomano V, Shen M, Pansanel J, MacIntosh AJJ, Sueur C. Social transmission in networks: global efficiency peaks with intermediate levels of modularity. Behav Ecol Sociobiol. 2018;72(9).\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eKikusui T, Winslow JT, Mori Y. Social buffering: Relief from stress and anxiety. Philos Trans R Soc B Biol Sci. 2006;361(1476):2215\\u0026ndash;28.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eEthics. declarations.\\u003c/span\\u003e\\u003c/li\\u003e\\u003c/ol\\u003e\"}],\"fulltextSource\":\"\",\"fullText\":\"\",\"funders\":[],\"hasAdminPriorityOnWorkflow\":false,\"hasManuscriptDocX\":true,\"hasOptedInToPreprint\":true,\"hasPassedJournalQc\":\"\",\"hasAnyPriority\":false,\"hideJournal\":false,\"highlight\":\"\",\"institution\":\"\",\"isAcceptedByJournal\":true,\"isAuthorSuppliedPdf\":false,\"isDeskRejected\":\"\",\"isHiddenFromSearch\":false,\"isInQc\":false,\"isInWorkflow\":false,\"isPdf\":false,\"isPdfUpToDate\":true,\"isWithdrawnOrRetracted\":false,\"journal\":{\"display\":true,\"email\":\"info@researchsquare.com\",\"identity\":\"movement-ecology\",\"isNatureJournal\":false,\"hasQc\":true,\"allowDirectSubmit\":false,\"externalIdentity\":\"move\",\"sideBox\":\"Learn more about [Movement Ecology](http://movementecologyjournal.biomedcentral.com/)\",\"snPcode\":\"40462\",\"submissionUrl\":\"https://submission.nature.com/new-submission/40462/3\",\"title\":\"Movement Ecology\",\"twitterHandle\":\"@BioMedCentral\",\"acdcEnabled\":true,\"dfaEnabled\":true,\"editorialSystem\":\"em\",\"reportingPortfolio\":\"BMC/SO AJ\",\"inReviewEnabled\":true,\"inReviewRevisionsEnabled\":true},\"keywords\":\"Social organization, wind, modularity, zebra finch, automated radio-tracking\",\"lastPublishedDoi\":\"10.21203/rs.3.rs-6502285/v1\",\"lastPublishedDoiUrl\":\"https://doi.org/10.21203/rs.3.rs-6502285/v1\",\"license\":{\"name\":\"CC BY 4.0\",\"url\":\"https://creativecommons.org/licenses/by/4.0/\"},\"manuscriptAbstract\":\"\\u003ch2\\u003eBackground\\u003c/h2\\u003e \\u003cp\\u003eSocial connections may provide individuals with multiple benefits. Individuals, however, are often constrained in how they socially organize due to ecological and environmental factors that affect individual space-use and movement patterns. Weather is one such factor that influences individual movements, and thus social structure. While on longer time scales (i.e., seasonally) the impacts of weather are relatively predictable, on shorter time scales (i.e., sub-daily), the impacts of weather on social organization are less predictable yet are largely overlooked.\\u003c/p\\u003e\\u003ch2\\u003eMethods\\u003c/h2\\u003e \\u003cp\\u003eIn this study, we examined the influence of short-term weather components, specifically wind and temperature, on the social structure of free-living zebra finches (\\u003cem\\u003eTaeniopygia castanotis\\u003c/em\\u003e) in the Australian arid zone. Our goal was to characterize if social network structure was impacted by hourly changes in these important weather components. To do so, we used an automated radio telemetry system to concurrently track 128 wild zebra finches for 12 consecutive days in the Australian spring in order to examine the relationships between weather components. Using Bayesian network analyses to account for the uncertainty in association strengths among individuals, we examined network structure, as measured by density and modularity, in relation to hourly wind speed, temperature, and time of day. Additionally, to assess if weather impacted the synchronization of group-level movements, we calculated proximity ratios between neighbouring individuals, which we related to wind and temperature.\\u003c/p\\u003e\\u003ch2\\u003eResults\\u003c/h2\\u003e \\u003cp\\u003eWe observed that network modularity increased during hours with higher mean wind speed and was highest in the morning and evening hours. In contrast, network density was not related to wind speed. Additionally, neither modularity nor density showed a significant relationship with the moderate temperatures during the tracking period. Group-level movement patterns as measured by proximity ratios between neighbouring individuals showed no relationship with either wind or temperature.\\u003c/p\\u003e\\u003ch2\\u003eConclusions\\u003c/h2\\u003e \\u003cp\\u003eOur results suggest that short-term changes in wind impact social structure in wild zebra finches. Given the critical role that network modularity plays in social information transfer, increased wind could have significant downstream consequences. Stochastic and more frequent changes in weather due to climate change could thus significantly impact nomadic species that rely on locating ephemeral resources.\\u003c/p\\u003e\",\"manuscriptTitle\":\"Windy weather drives social structure in wild zebra finches\",\"msid\":\"\",\"msnumber\":\"\",\"nonDraftVersions\":[{\"code\":1,\"date\":\"2025-05-29 15:16:53\",\"doi\":\"10.21203/rs.3.rs-6502285/v1\",\"editorialEvents\":[{\"type\":\"communityComments\",\"content\":0},{\"type\":\"decision\",\"content\":\"Revision requested\",\"date\":\"2025-06-28T23:10:00+00:00\",\"index\":\"\",\"fulltext\":\"\"},{\"type\":\"editorInvitedReview\",\"content\":\"\",\"date\":\"2025-06-15T06:07:27+00:00\",\"index\":\"hide\",\"fulltext\":\"\"},{\"type\":\"reviewerAgreed\",\"content\":\"75471600141254585445708260495241620334\",\"date\":\"2025-06-12T06:04:26+00:00\",\"index\":\"hide\",\"fulltext\":\"\"},{\"type\":\"editorInvitedReview\",\"content\":\"\",\"date\":\"2025-05-29T13:13:35+00:00\",\"index\":\"hide\",\"fulltext\":\"\"},{\"type\":\"reviewerAgreed\",\"content\":\"120905355153187553862992888364824053045\",\"date\":\"2025-05-27T11:05:54+00:00\",\"index\":\"hide\",\"fulltext\":\"\"},{\"type\":\"reviewersInvited\",\"content\":\"\",\"date\":\"2025-05-27T08:31:30+00:00\",\"index\":\"\",\"fulltext\":\"\"},{\"type\":\"editorAssigned\",\"content\":\"\",\"date\":\"2025-04-24T12:06:19+00:00\",\"index\":\"\",\"fulltext\":\"\"},{\"type\":\"checksComplete\",\"content\":\"\",\"date\":\"2025-04-24T12:06:11+00:00\",\"index\":\"\",\"fulltext\":\"\"},{\"type\":\"submitted\",\"content\":\"Movement Ecology\",\"date\":\"2025-04-22T08:50:54+00:00\",\"index\":\"\",\"fulltext\":\"\"}],\"status\":\"published\",\"journal\":{\"display\":true,\"email\":\"info@researchsquare.com\",\"identity\":\"movement-ecology\",\"isNatureJournal\":false,\"hasQc\":true,\"allowDirectSubmit\":false,\"externalIdentity\":\"move\",\"sideBox\":\"Learn more about [Movement Ecology](http://movementecologyjournal.biomedcentral.com/)\",\"snPcode\":\"40462\",\"submissionUrl\":\"https://submission.nature.com/new-submission/40462/3\",\"title\":\"Movement Ecology\",\"twitterHandle\":\"@BioMedCentral\",\"acdcEnabled\":true,\"dfaEnabled\":true,\"editorialSystem\":\"em\",\"reportingPortfolio\":\"BMC/SO AJ\",\"inReviewEnabled\":true,\"inReviewRevisionsEnabled\":true}}],\"origin\":\"\",\"ownerIdentity\":\"d706812e-b142-40a7-b85f-551f30a089bb\",\"owner\":[],\"postedDate\":\"May 29th, 2025\",\"published\":true,\"recentEditorialEvents\":[],\"rejectedJournal\":[],\"revision\":\"\",\"amendment\":\"\",\"status\":\"published-in-journal\",\"subjectAreas\":[],\"tags\":[],\"updatedAt\":\"2026-01-26T16:10:37+00:00\",\"versionOfRecord\":{\"articleIdentity\":\"rs-6502285\",\"link\":\"https://doi.org/10.1186/s40462-025-00623-9\",\"journal\":{\"identity\":\"movement-ecology\",\"isVorOnly\":false,\"title\":\"Movement Ecology\"},\"publishedOn\":\"2026-01-21 15:57:07\",\"publishedOnDateReadable\":\"January 21st, 2026\"},\"versionCreatedAt\":\"2025-05-29 15:16:53\",\"video\":\"\",\"vorDoi\":\"10.1186/s40462-025-00623-9\",\"vorDoiUrl\":\"https://doi.org/10.1186/s40462-025-00623-9\",\"workflowStages\":[]},\"version\":\"v1\",\"identity\":\"rs-6502285\",\"journalConfig\":\"researchsquare\"},\"__N_SSP\":true},\"page\":\"/article/[identity]/[[...version]]\",\"query\":{\"redirect\":\"/article/rs-6502285\",\"identity\":\"rs-6502285\",\"version\":[\"v1\"]},\"buildId\":\"8U1c8b4HqxoKbykW_rLl7\",\"isFallback\":false,\"isExperimentalCompile\":false,\"dynamicIds\":[84888],\"gssp\":true,\"scriptLoader\":[]}","source_license":"CC-BY-4.0","license_restricted":false}