Skip to main content

Thank you for visiting nature.com. You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.

Influence of individual biological traits on GPS fix-loss errors in wild bird tracking

Abstract

In recent decades, global positioning system (GPS) location data and satellite telemetry systems for data transmission have become fundamental in the study of basic ecological traits in wildlife biology. Evaluating GPS location errors is essential in assessing detailed information about the behaviour of an animal species such as migration, habitat selection, species distribution or foraging strategy. While many studies of the influence of environmental and technical factors on the fix errors of solar-powered GPS transmitters have been published, few studies have focussed on the performance of GPS systems in relation to a species’ biological traits. Here, we evaluate the possible effects of the biological traits of a large raptor on the frequency of lost fixes—the fix-loss rate (FLR). We analysed 95,686 records obtained from 20 Bearded Vultures Gypaetus barbatus tracked with 17 solar-powered satellite transmitters in the Pyrenees (Spain, France and Andorra), between 2006 and 2019 to evaluate the influence of biological, technical, and environmental factors on the fix-loss rate of transmitters. We show that combined effects of technical factors and the biological traits of birds explained 23% of the deviance observed. As expected, the transmitter usage time significantly increased errors in the fix-loss rate, although the flight activity of birds revealed an unexpected trade-off: the greater the proportion of fixes recorded from perched birds, the lower the FLR. This finding seems related with the fact that territorial and breeding birds spend significantly more time flying than non-territorial individuals. The fix success rate is apparently due to the interactions between a complex of factors. Non-territorial adults and subadults, males, and breeding individuals showed a significantly lower FLR than juveniles-immatures females, territorial birds or non-breeding individuals. Animal telemetry tracking studies should include error analyses before reaching any ecological conclusions or hypotheses about spatial distribution.

Introduction

Obtaining a global positioning system (GPS) fix and the reliability of location data are primarily subject to satellite acquisition, a process mainly shaped by technical, environmental, and behavioural factors1. External factors such as GPS satellite geometry (satellite constellation), topography and land surface roughness, vegetation, fix interval (time lapse between successive fixes), or even GPS-tag position and orientation, all limit a transmitters’ ability to make contact with at least three satellites during a period of GPS activation2,3,4, causing GPS misconnections. Some authors have even observed: (1) an association between resource use, habitat selection, and fix-loss rate; and (2) interactions between animal behaviour and local habitat conditions which have to be considered particularly when assessing a species’ habitat use4,5,6,7,8. However, one of the biggest gaps in our understanding of GPS performance is related to species-specific behavioural effects. For example, the position of an individual animal changes the orientation of a receiver, and its performance. Some studies of large mammals have demonstrated that inactive animals have higher fix-loss rates and lower fix accuracy than active ones1,4,9. But very little is known about how, or to what extent, individual biological traits such as sex, age, size, territorial or breeding status, and their associated behaviour and ecology may affect satellite connection, fix-loss and location accuracy10,11,12,13. This kind of information is essential to properly interpret geolocation data and to draw useful conclusions regarding animal movement patterns or species behaviour.

During the last 40 years, Argos Platforms Transmitter Terminals (PTTs) have provided the world’s most commonly used tracking coverage technology for the remote study of free-ranging animal movements, mainly because of their integration of GPS fixes (i.e. satellite locations) with data transmission technologies (i.e. the Argos data transfer system), particularly from the mid-1990s when GPS receivers became able to record high-spatial-resolution tracking data14,15,16,17. However, the raw data registered through GPS-Argos telemetry still suffer from errors and biases (e.g. fix rate bias, fix-loss errors and spatial location errors) that must be considered to avoid drawing incorrect conclusions and making the wrong management recommendations13,15,18. These tracking problems are especially relevant for threatened species where reliable information is particularly important for reintroduction projects and conservation plans.

The endangered Bearded Vulture Gypaetus barbatus represents a good case study for assessing GPS fix-loss errors—measured in this study though the monthly fix-loss rate, FLR (for more details see “Methods” section). In the first instance, this species inhabits rugged mountain landscapes (in the Pyrenees, average home range kernel 90% varying between 63 km2 for territorial individuals to 11,600 km2 for non-territorial ones19) that allows the evaluation of the influence of abrupt topography on GPS fix-loss. Second, the long daylight hours and sunny climatic conditions favour at the same time flying behaviour and the charging of transmitter solar batteries. Third, the territorial behaviour of breeding individuals is very different to the behavioural pattern of non-territorial individuals, which fly over greater distances due to the lack of a nest site acting as a central foraging point19,20. Four, the changing seasonal and weather conditions in the Pyrenees allow the comparison of transmitter performance during different solar radiation conditions. Finally, Bearded Vultures are an endangered species (more specifically, classed by the BirdLife International 201721 as vulnerable in Europe, and globally near threatened) for which accurate GPS data is important to improve management and conservation actions. The species is being reintroduced in several European countries, and GPS transmitter monitoring is one of the main tools used by managers and conservationists to assess its habitat use and reintroduction success22.

Technological improvements enabling the use of Argos GPS-lightweight PTTs (< 80 g) in marine mammals, birds, or even small animals up to 300 g23,24, have prompted new research into sources of GPS errors associated to wildlife telemetric tracking, especially when fix-loss rate is related to animal behaviour or habitat use. This study focuses on the biological, environmental, and technical factors affecting the fix-loss rate—either caused by GPS misconnections or battery undercharging—in Argos GPS PTTs. We considered specific biological traits of Bearded Vultures including sex, age, territorial and breeding status, and flight activity (derived from the monthly rates of fixes of perched and flying birds) as biological factors. Concurrently, we considered environmental variables including topographic altitude, surface solar radiation, and total precipitation, as well as technical factors considering the transmitter usage time and the duty cycle (i.e. fix recording scheduled regimes), as extrinsic factors. Afterwards, given the flying nature of our case study species and the effect that this kind of movement behaviour has showed over some technological characteristics of the GPS transmitters in previous studies25, we explore the influence of these biological and environmental variables over the flight activity trying to better understand this behavioural parameter and consequently its effect on the FLR (see Fig. 1). Based on previous solar-powered GPS tracking studies13,25,26,27, our hypothesis was that both FLR and flight activity of birds are strongly influenced either by specific biological traits and/or extrinsic factors, especially those related to technical factors. We hypothesized that individuals with greater flight activity (higher rates of fix in flight, RFF—reasonably assumed to be non-territorial individuals, who usually travel larger distances19,20), would be more exposed to direct solar radiation, thus present increased battery charging, and a lower FLR compared to territorial individuals. Considering extrinsic factors, weather conditions will affect fix reception success because periods with more daylight hours (i.e. summer, presenting the highest surface solar radiation and lowest total precipitation) also favour thermal conditions for flight, in contrast to winter, promoting thus birds’ flight activity. Topographic altitude will also probably affect FLR due to the challenging GPS connection in steep terrains. At the same time, it is expected that the transmitter usage time will negatively affect transmitter performance as a consequence of the decreasing battery and electronic system performance of the device.

Figure 1
figure 1

Factors considered to influence Fix Loss Rate (FLR) and bird’s flight activity (composed considering both monthly numbers of fix perched and fix in flight).

Results

A total of 95,686 location results from 20 Bearded Vultures tracked with 17 transmitters in the Pyrenees were recorded from 2006 until January 2019. Of these records: 32.6% were from females and 67.4% from males; 83.8% were from adults and 11.0% from subadults; 4.6% were from immatures and 0.6% from juveniles; 35.6% were from territorial birds and 64.4% from non-territorial ones.

Fix-loss rate (FLR)

We found a substantial FLR variability showing significant differences between individual birds (Kruskal–Wallis, χ2 = 278.13, df = 18, p < 0.001) and also between individual PTTs (χ2 = 251.39, df = 15, p < 0.001). Five PTTs showed an FLR less than a 30%, seven showed FLRs of between 30 and 40%, one had an FLR of 48.2%, and the remaining five registered an FLR equal to or higher than 50% (Table 1). The FLR was highly variable at the individual level: seven birds had an FLR less than 30%; another seven showed FLRs between 30 and 40%; two ranged from 40 to 50%; and four showed an FLR higher than 50% (Table 1). However, FLR barely fluctuated between months, showing no significant differences over the year (Kruskal–Wallis, χ2 = 10.92, df = 11, p = 0.45), ranging between mean values of 0.31 ± 0.24 in May to 0.37 ± 0.23 in October.

Table 1 Basic biological traits and individual measures of fix loss rate (FLR) during a set period of time; rate of fix in flight (RFF) and rate of perched fixes (RPF) (mean ± SD for the monthly FLR, RPF and RFF individual values) for 20 birds tagged with 1770 g solar-powered Argos’ satellite transmitters (PTT/GPS Microwave Telemetry, Inc. Columbia, MD, USA) all bought in 2005–2008.

The total conditional R2 obtained from the GLMM built to evaluate the joint contribution of the biological, technical and environmental factors to the FLR was 0.148 (0.093 of the marginal R2 corresponding to the fixed effects + 0.055 of deviance corresponding to the random effects). The highest was provided by the sum of both technical factors (single effect of 1.5%) and biological traits (with the highest single retained effect of 5.8%) and its interactions with the other groups (− 0.6% shared between both groups, − 0.5% shared between technical factors and environmental factors, 3.3% shared between biological traits and environmental factors, and 13.1% resulting from the interaction of the three groups). Environmental factors retained a single effect of − 7.8% (Fig. 2).

Figure 2
figure 2

Conditional R2 partitions resulting from a partial regression analysis of 17 Microwave PTTs monthly fix-loss rates. Percentages of conditional R2 (deviance explained by the entire model, including both fixed and random effects) explained by each group of variables: Technical (PTT’s lifetime and duty cycle), Biological (rate of perched fixes, sex, age, territorial and breeding status), and Environmental (surface solar radiation, total precipitation, and topographic altitude) and by their interactions. The total conditional R2 of the model is also shown.

Regarding the GLMMs results (Table 2), we found two models from the total selection that met the delta AIC < 2 criterion. The parameters influencing the efficient performance of the transmitters included all of the biological and technical variables tested. Non-territorial birds, males, and breeding individuals showed a significantly lower FLR than females, territorial birds or non-breeding vultures. The rate of perched fixes (RPF) showed a negative relationship with the FLR, while the PTT usage time exerted the opposite effect, so that the higher the RPF and—in parallel—the smaller the PTT usage time, the lower the FLR becomes. Indeed, longer duty cycles provoked also lower FLR, although this variable was only selected for one of the two final models selected. Regarding age, adults and subadults showed significantly lower FLRs. None of the environmental variables were included in the significant GLMMs eventually built. The partial effects of all the explanatory variables included at least in one of the two final models selected are shown in Supplementary Figure S1.

Table 2 Competing GLMMs to evaluate the influence of different biological traits and extrinsic factors (comprising both technical and environmental variables) on the fix loss rate (FLR).

Flight activity

Considering all the data, we observed an average of 64.3 ± 20.0% rate of perched fixes (RPF) and 31.8 ± 16.5% rate of fixes in flight (RFF). Of the 20 Bearded Vultures tagged, 65.0% (n=13) showed a quite homogeneous flight activity pattern, their mean RPF ranging between 87.4 and 68.8%. Three individuals showed rates of 61.3–63.5% and the other four showed perched fix rates lower than 52.8%. Even so, three of the birds exhibited a higher monthly RFF than RPF (Table 1).

The RPFs ranged significantly between 70.4% in summer and 66.4% in winter (Kruskal–Wallis, χ2 = 21.12, df = 3, p < 0.001). The variables selected for the competing GLMMs influencing flight activity (Table 3) were territoriality, breeding status, age, and sex (although, the last two were not always included in the final models). Territorial and breeding individuals showed significantly lower RPFs than non-breeding and non-territorial ones. The mean RPF and RFF were 50.2 ± 25% and 39.3 ± 22.6% for territorial individuals and 72 ± 10.5% and 27.6 ± 10% for non-territorial birds, respectively. The environmental variables were not included in the final models. The partial effects of all the explanatory variables included at least in one of the four final models selected are shown in Supplementary Figure S2.

Table 3 Competing GLMMs for evaluating the influence of different biological traits and environmental variables on birds’ flight activity (computed as a weighted rate of perched fix).

Discussion

Studies of movement ecology often suffer from lost geolocation information due to: (1) technical glitches such as insufficient battery power; (2) environmental factors such as the roughness of the terrain (i.e. the conjunction of vegetation and topography28) or changing climatological parameters; and (3) intrinsic factors (i.e. biological traits) such as the behaviour of individuals25,29. Our findings point out that a combination of technical variables and biological traits gave the best explanation of maximum deviance (22.6%), suggesting that these two groups of factors have a much greater influence on the monthly FLR than environmental factors. This was also one of the main conclusions achieved by Hofman et al.27, in a study where they gathered information of 167 projects deployed on 62 species in 142 study areas worldwide through some questionnaires with the aim of assessing the performance of satellite telemetry units (predominantly collars) tracking terrestrial wildlife. Concretely, they found out that the transmitter performance was strongly influenced by unit and species characteristics while environmental conditions increased the variability, influencing the transmitters’ technique effectiveness. Concurrently, we propose that it could be that technical and biological variables already gather part of the deviance explained by the environmental factors. Such is the case for the breeding status, a parameter directly related to time of year, seasonality, solar radiation, and daylight duration in addition to its biological significance for the species. Another technical variable that affects the FLR, the transmitter battery level, is also related to solar energy availability, and hence to the time of year (through the seasonal variations in solar irradiance received by the device)25,26. Battery power limits the time for the transmitter to search satellites to obtain a location and so influence in the number of satellites acquired for the process1, however it is a parameter only available in the newer GPS models. In this respect, the logger generation, transmitters’ manufacturers, and data receiving system—all of the three uniform parameters for our study case given that all the 17 devices were Microwave solar-powered Argos-GPS bought between 2005 and 2008—are also important technical variables that need to be considered when studying GPS accuracy and location errors25,29,30. On the other hand, our results show that individual flight activity could be one of the most influential factors determining the fix performance of a device. Contrary to previous studies19,36 and our initial hypothesis, the greater the proportion of perched fixes, the lower is the resulting FLR. One possible explanation for this observation could be related to the difficulty of satellite acquisition while a bird is flying, as has been noted for moving animals in various mammal studies29,31,32, perhaps because of changes in the position and orientation of the GPS transmitter. Our findings also confirmed that longer duty cycles (of 2 h compared with those of 1 h) produced lower FLR, probably associated with the fact that more intense duty cycles increase the transmitter energy consumption and consequently reduce the device usage time33. In fact, Silva et al.25 suggested that FLR due to poor GDOP (when Geometric Dilution of Precision limits the transmitter to contact with enough satellites to produce a fix) increased when the birds moved. Nevertheless—considering that the time to obtain a fix increase in dynamic versus static conditions—longer duty cycles (of 2 h compared to 30 s and 15 min interval times) would produce higher fix loss rates while flying, but the opposite situation could happen while the birds are perched, when the length of the fix interval is not so relevant. In addition, as it was predicted, the FLR increases with transmitter usage (as happened in27,34), a relevant information considering that the mean usage time for our PTTs was 5.34 ± 3.03 years (n = 14).

Given the number of studies which point to landscape structure as an important driver of the FLR2,35, we expected the topographic altitude as a variable influencing FLR. However, our monthly-scale analysis could have diluted the effect of this environmental variable and a complementary shorter time-scale FLR study (e.g. daily or hourly) may show a higher influence of this specific variable on the fix loss errors. Notwithstanding these uncertainties, our analyses of the possible effects of Bearded Vulture biological traits on the FLR constitutes a novel approach to the better understanding of the treatment of PTT locations. All the biological variables tested in this study influenced the RPF and also significantly affected the fix reception success. Interestingly, non-territorial Bearded Vultures travel further and later in the daylight than territorial birds20, but exhibit significantly higher RPFs (72 ± 10.5% for non-territorial individuals versus 50.2 ± 25% for territorial birds). At the same time, breeding and territorial adults showed lower FLRs even if they spent less time perched than non-breeding, non-territorial and younger individuals. Probably their daily activity related to parental duties (nest-building, territorial defence, and foraging) results in increased flight activity and a higher proportion of their time spent flying, even if the distances covered are shorter than those of non-territorial birds19,20,36.

Our results showed an overall monthly mean FLR of 34.5 ± 24.72% ranging between a minimum of 3% and maximum of 100% (n = 17). This is lower than the values found for analogous transmitters by Silva et al.25, used on the same Pyrenean and Cantabrian population of Bearded Vultures (FLR = 0.40 ± 0.12), and those recorded by Soutullo et al.24, for lightweight Argos GPS transmitters used on Golden Eagles Aquila chrysaetos in a rocky cliff area in Eastern Spain (FLR = 0.45). In this latter study, breeding season also influenced the FLR (probably through seasonal effects), as was the case in our study. Nevertheless, our findings show the importance of understanding that significant variations in FLR may be due either to variations in individual bird behaviour or to variations in technical glitches affecting each PTT performance. Therefore, it should be expected that both biological and technical factors play a fundamental role in the correct performance of the GPS fix programming.

The significant differences in FLR between male and female birds are not easily explained from a behavioural and ecological perspective (mean values of 0.32 for males cf. 0.39 for females), even if non-territorial males do indeed exploit larger areas and fly over longer distances, as it is the case of territorial females19,20. The specific relationships between the biological traits of this species and RPF or RFF are clear, but even if their influence on the FLR is also obvious, it is more difficult to explain the effect of certain biological variables such as territoriality or sex on FLR. The fix success rate results are most likely due to a synergy between complex interactions; for instance, between flight height and terrain roughness, or between the availability of environmentally optimal flight conditions (which are also favorable for solar battery charging) linked to the likelihood of flight activity and the resulting associated increase in transmitter movement. In any case, it is clear from this study that biological factors such as sex, age, breeding and territorial status have particular effects on FLR and must be considered when studying fix error rates in other flying species (e.g. bird and bat species). Even considering the apparent limitations of working with a single species in a GPS fix loss error study, as it has been shown, our findings can be extrapolated to different medium and large-size animal populations and species. Moreover, technological improvements of materials and both hardware and software enhancements are leading to increasingly better transmitters’ performance with improved location accuracy and reduced FLRs. However, there are still many transmitters in use (in addition to the quite a few already developed) that present scheduled location duty cycles, data receiving Argos-GPS system and device manufacturers similar to the ones evaluated in this study, so these findings obtained remain relevant for long-term conservation studies.

Every animal telemetry tracking study should include an error analysis before reaching any ecological conclusions or hypotheses regarding spatial utilization, since the results can vary substantially depending on extrinsic factors such as GPS transmitter model, retrieval data system, PTT usage time, season, etc., or biological factors such as those analysed in this study. All of these changing elements can influence the data collected and lead to errors in interpreting patterns of movement. Fortunately, these kinds of tracking error, together with accuracy biases in the horizontal plane (x and y coordinates) are being addressed and overcome as transmitter technology improves, thus reducing the potential influence of tracking device shortcomings on the recording and interpretation of basic parameters regarding the spatial ecology of a species26,29,30,37,38.

Methods

Study area

We assessed the GPS fix loss errors resulting from Bearded Vultures studied in the Pyrenees, a steep mountainous region with maximum altitudes of 3400 m, located in the north of the Iberian Peninsula on the border between France and Spain. It includes three different bioclimatic areas (Montane, Sub-Alpine and Alpine) with average annual temperatures between 0 and 20 °C, and a four-season Mediterranean climate with seasonal weather conditions39.

Study species

The Bearded Vulture is a territorial, cliff-nesting vulture specialized in feeding on the bones of medium size ungulates40. In common with other avian scavengers it exploits thermal and orographic updrafts to use the least energy as possible when foraging. It is an endangered species now only found in certain mountainous areas of Europe, Asia, and Africa21,41. In the Pyrenees, the spatial ecology of this species has been studied since the 1980s’, originally using conventional VHF radio tracking42,43,44 and more recently with the solar-powered Argos or GSM data recovery system with GPS-PTTs19,20,45.

Tracking and data origin

Between 2006 and 2019, twenty Bearded Vultures were tagged with 17 different 70 g solar-powered Argos’ satellite transmitters (PTT/GPS Microwave Telemetry, Inc. Columbia, MD, USA, all of the 2005–2008 logger generation)—three of which were reused on new individuals—attached to the bird’s back with a breakaway thoracic junction stitched with cotton thread harness made of 0.64 cm Teflon ribbon (Bally Ribbon Mills, Bally, PA, USA) (for further details see15). The usage time of the transmitters was 5.34 ± 3.03 years (n = 14) on average. To compute this mean value for the three reused PTT, we summed the time usage of each peer of individuals using the same PTT, and for the rest of the PTTs, we excluded the records corresponding to the birds dead on the field (n = 3) since their transmitters could not be recovered and the reason for stopping fix recording was unlikely related with technical causes. We only considered the records of individuals whose PTTs stopped working properly, accounting times from the moment the PTTs were turned on until the moment we stopped receiving location data (see Supplementary Table S1). All of the transmitters were programmed to report hourly GPS fixes between 04:00 and 22:00 UTC hours each day (manufacturer estimated error ± 18 m), except for two individuals whose PTTs transmitted every 2 h. Regarding the biological factors: (1) age of individuals were assigned to four different age classes according to plumage characteristics: juvenile (1 year old); immature (2–3 years old); subadult (4–5 years old); and adult (> 6 years old) (for details see19,20); (2) sex was determined by molecular analysis of blood samples (PCR amplification of the CHD-W gene as described in46); (3) territoriality was described as territorial or non-territorial individuals, depending on their breeding behaviour20; (4) breeding season was defined either as breeding period (1st January to 31st July) or non-breeding period (1st August–31st December)15; and (5) flight activity was defined according to Silva et al.25 by the complementary rates of perched fixes (RPF, calculated from monthly fixes with speeds slower than 1.39 m/s) and fix in flight (RFF, calculated from monthly fixes with speeds equal or faster than 1.39 m/s) (Tables 1 and 2). Regarding the extrinsic factors: for technical variables, (1) we accounted for the device usage time and (2) duty cycle (as mentioned, of 1 or 2 h depending on the individual) and for environmental variables, (3) topographic altitudes were obtained using a Digital Elevation Model (ASTER Global DEM, 1 arc-second spatial resolution); and (4) surface solar radiation and (5) total precipitation were obtained from an interim full-daily at surface forecast (European Centre for Medium-Range Weather Forecasts, 0.75° each 3 h). Monthly means of all three parameters were calculated using the Movebank Env-DATA track data annotation service47,48 (Fig. 1).

Data processing and statistical analysis

The fix-loss rate (FLR) used in this study was calculated as a monthly value for each individual consisting of the proportion of days per month on which no fixes were recorded. We evaluated the effects of both biological and extrinsic factors (including both technical and environmental variables) on the performance of the 17 transmitters represented by monthly FLRs computed as the number of days per month on which no data were collected, divided by the total number of days on which data were scheduled to be collected. We generated a data set of monthly observations (n = 889), each with its own FLR. Since we reused three of the 17 transmitters to track the movement pattern of 20 birds, we needed to distinguish between two different levels when computing mean FLRs: the PTT/transmitter level and the individual level. For instance, the PTT usage time depends directly on the transmitter but variables related with the biological traits depend uniquely on the individual.

At first, we examined the FLR with some non-parametric explorative analyses to evaluate possible differences among these two levels and to evaluate the influence of the month on the FLR yearly distribution. Secondly, we grouped all the predictor variables: (1) age, sex, breeding and territorial status, and RPF (this latter describing flight activity) as biological factors; (2) PTT usage time per month and duty cycle as the technical factors; and (3) monthly means of topographic altitude, surface solar radiation, and total precipitation as environmental factors (see Fig. 1).

Thirdly, we performed a deviance partitioning analysis49 to evaluate the effect on FLR of the single and joint contributions of each of the three groups of variables comparing by basic algebra the percentage of the explained conditional R2 of each of the best generalized linear mixed models (GLMMs)50 built including the aforementioned biological, technical and environmental factors as fixed factors (where applicable) and the individual as a random factor. Thus, we built seven separate GLMMs to evaluate: (1) the single contribution of the biological factors, (2) the single contribution of the technical factor; (3) the single contribution of the environmental factor; (4) the joint contribution of the biological and technical factors; (5) the joint contribution of the biological and environmental factors; (6) the joint contribution of the technical and environmental factors; and (7) the joint contribution of the biological, technical and environmental factors (see more details about how to perform a deviance partitioning analysis in51). These analyses were computed using R statistical software52 version 3.6.2. For the GLMMs, we applied the “glmer” function of the “lme4” R package53 with a binomial error distribution and logit-link function. All the deviance explained by the different groups of variables was expressed in percentages when we referred to the deviance partitioning results.

Fourthly, to determinate the significant variables influencing the FLR, we constructed the full model with all of the biological, technical and environmental variables as fixed factors and the individual as a random factor considering again a binomial error distribution and logit-link function, made a model selection using Akaike’s Information Criterion (AIC54), and chose the best models with a delta AIC < 2 (Fig. 1).

And fifthly, to better understand the individual flying behaviour and how it could affect FLR, we analysed the influence of all of the same biological and environmental factors on the flight activity of the birds. For this analysis, a weighted RPF (wRFP) was created combining the monthly number of perched fix and monthly number of fix in flight (see Zuur et al.55 for applying binomial generalized models for proportions). Thus, we modelled wRPF using a GLMM (binomial error distribution and logit-link function) with all the biological and environmental variables as fixed factors and the individual as a random factor, and then selected models giving delta AIC < 2 (Fig. 1). Technical factors were not included as predictors in this model because of their obvious absence of influence over the flight activity of the birds.

For all the mixed models built in this study, the relative contributions of the fixed and random factors to R2 were estimated with the “r.squaredGLMM” function from the package “MuMIn”56. We also reviewed for the variance inflation factors (VIF) for all the predictor variables at the first stages of the GLMMs building using the “car” package57 to assess collinearity (accepted VIF values < 3). In fact, we firstly considered season (defined as yearly quarterly periods i.e.: winter, from January to March; spring, from April to June; summer, from July to September; and fall, from October to December) and month for all the GLMMs’ analyses, but they were finally excluded because of their high correlation with breeding season and surface solar radiation. All continuous variables were standardized and centred before modelling using the “scale” R function and all of the non-parametric analyses were performed after checking for the absence of a normal distribution.

Tracking data are inherently auto-correlated, although if fixes are taken infrequently enough so as to be longer than the autocorrelation timescale of the data, data can be considered independent, especially for animals that move long distances in short periods of time5,58. This is the case for our study species in this study, which present minimum duty cycles of 1 h (see also19).

Ethics statement

All the work was conducted in accordance with relevant national and international guidelines, and conforms to all legal requirements. Captures and blood sample collection were carried out in compliance with the Ethical Principles in Animal Research. Thus, protocols, amendments and other resources were conducted in accordance to the guidelines approved by the Catalan Autonomous Government (Generalitat de Catalunya) following the R.D.1201/2005 (10 October 2005, BOE 21 October 2005) of the Ministry of Presidency of Spain. All experimental protocols were approved by the Catalan Autonomous Government and MAGRAMA (References 15.546 and 25.306).

Data availability

The datasets used and analyzed during the current study are available from AM on reasonable request.

References

  1. Moen, R., Pastor, J., Cohen, Y. & Schwartz, C. C. Effects of moose movement and habitat use on GPS collar performance. J. Wildl. Manag. 60, 659–668 (1996).

    Article  Google Scholar 

  2. Cain, J. W. III., Krausman, P. R., Jansen, B. D. & Morgart, J. R. Influence of topography and GPS fix interval on GPS collar performance. Wildl. Soc. Bull. 33, 926–934 (2005).

    Article  Google Scholar 

  3. Graves, T. A. & Waller, J. S. Understanding the causes of missed global positioning system telemetry fixes. J. Wildl. Manag. 70, 844–851 (2006).

    Article  Google Scholar 

  4. Moen, R., John, P. & Cohen, Y. Effects of animal activity on GPS telemetry location attempts. Alces 37, 207–216 (2001).

    Google Scholar 

  5. D’Eon, R. G. Effects of a stationary GPS fix-rate bias on habitat-selection analyses. J. Wildl. Manag. 67, 858–863 (2003).

    Article  Google Scholar 

  6. Dussault, C., Courtois, R., Ouellet, J. P. & Huot, J. Evaluation of GPS telemetry collar performance for habitat studies in the boreal forest. Wildl. Soc. Bull. 27, 965–972 (1999).

    Google Scholar 

  7. Nielson, R. M., Manly, B. F. J., Mcdonald, L. L., Sawyer, H. & Mcdonald, T. L. Estimating habitat selection when GPS fix success is less than 100 %. Ecology 90, 2956–2962 (2009).

    Article  PubMed  Google Scholar 

  8. Rempel, R. S., Rodgers, A. R. & Abraham, K. F. Performance of a GPS animal location system under boreal forest canopy. J. Wildl. Manag. 59, 543–551 (1995).

    Article  Google Scholar 

  9. Bowman, J. L., Kochanny, C. O., Demarais, S. & Leopold, B. D. Evaluation of a GPS collar for white-tailed deer. Wildl. Soc. Bull. 28, 141–145 (2000).

    Google Scholar 

  10. Jung, T. S. & Kuba, K. Performance of GPS collars on free-ranging bison (Bison bison) in north-western Canada. Wildl. Res. 42, 315–323 (2015).

    Article  Google Scholar 

  11. Recio, M. R., Mathieu, R., Denys, P., Sirguey, P. & Seddon, P. J. Lightweight GPS-tags, one giant leap for wildlife tracking? An assessment approach. PLoS One 6, e28225 (2011).

    ADS  CAS  Article  PubMed  PubMed Central  Google Scholar 

  12. Mattisson, J., Andrén, H., Persson, J. & Segerström, P. Effects of species behavior on global positioning system collar fix rates. J. Wildl. Manag. 74, 557–563 (2010).

    Article  Google Scholar 

  13. Kaczensky, P., Ito, T. Y. & Walzer, C. Satellite telemetry of large mammals in Mongolia: What expectations should we have for collar function?. Wildl. Biol. Pract. 6, 108–126 (2010).

    CAS  PubMed  PubMed Central  Google Scholar 

  14. Harris, R. B. et al. Tracking Wildlife by Satellite: Current systems and Performance. Fish and Wildlife Technical Report https://pubs.er.usgs.gov/publication/70185512 (1990).

  15. Schwartz, C. C. & Arthur, S. M. Radiotracking large wilderness mammals: Integration of GPS and argostechnology. Ursus 11, 261–274 (1999).

    Google Scholar 

  16. Tomkiewicz, S. M., Fuller, M. R., Kie, J. G. & Bates, K. K. Global positioning system and associated technologies in animal behaviour and ecological research. Philos. Trans. R. Soc. B Biol. Sci. 365, 2163–2176 (2010).

    Article  Google Scholar 

  17. Rodgers, A. R. Recent telemetry technology. In Radio Tracking and Animal Populations (eds Marzluff, J. M. & Millspaugh, J. J.) 79–121 (Elsevier, New York, 2001). https://doi.org/10.1016/B978-012497781-5/50005-0.

    Chapter  Google Scholar 

  18. Thomas, B., Holland, J. D. & Minot, E. O. Wildlife tracking technology options and cost considerations. Wildl. Res. 38, 653–663 (2011).

    Article  Google Scholar 

  19. Margalida, A., Pérez-García, J. M., Afonso, I. & Moreno-Opo, R. Spatial and temporal movements in Pyrenean bearded vultures (Gypaetus barbatus): Integrating movement ecology into conservation practice. Sci. Rep. 6, 35746 (2016).

    ADS  CAS  Article  PubMed  PubMed Central  Google Scholar 

  20. García-Jiménez, R., Pérez-García, J. M. & Margalida, A. Drivers of daily movement patterns affecting an endangered vulture flight activity. BMC Ecol. 18, 39 (2018).

    Article  PubMed  PubMed Central  Google Scholar 

  21. BirdLife International. (2017). Gypaetus barbatus (Amended Version of 2017 Assessment). The IUCN Red List of Threatened Species 2017: e.T22695174A118590506. https://doi.org/10.2305/IUCN.UK.2017-3.RLTS.T22695174A118590506.en. Accessed: 12th Mar 2020.

  22. Houston D. C. Reintroduction programmes for vulture species. In Proceedings of the International Conference on Conservation and Management of Vulture populations 1, (eds Houston D. C. & Piper, S. E., 2006). Natural History Museum, University of Crete, Thessaloniki.

  23. Britten, M. W., Kennedy, P. L. & Ambrose, S. Performance and accuracy evaluation of small satellite transmitters. J. Wildl. Manag. 63, 1349–1358 (1999).

    Article  Google Scholar 

  24. Soutullo, A., Cadahía, L., Urios, V., Ferrer, M. & Negro, J. J. Accuracy of lightweight satellite telemetry: A case study in the Iberian Peninsula. J. Wildl. Manag. 71, 1010–1015 (2007).

    Article  Google Scholar 

  25. Silva, R., Afán, I., Gil, J. A. & Bustamante, J. Seasonal and circadian biases in bird tracking with solar GPS-tags. PLoS One 12, e0185344 (2017).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  26. Byrne, M. E., Holland, A. E., Bryan, A. L. & Beasley, J. C. Environmental conditions and animal behavior influence performance of solar-powered GPS-GSM transmitters. Condor 119, 389–404 (2017).

    Article  Google Scholar 

  27. Hofman, M. P. G. et al. Right on track? Performance of satellite telemetry in terrestrial wildlife research. PLoS One 14, 1–26 (2019).

    Google Scholar 

  28. Aubrecht, C. et al. Vertical roughness mapping - ALS based classification of the vertical vegetation structure in forested areas. In Symposium A Quarterly Journal In Modern Foreign Literatures (eds. Wagner, W. & Székely, B.) XXXVIII, 35–40 (2010).

  29. Frair, J. L. et al. Resolving issues of imprecise and habitat-biased locations in ecological analyses using GPS telemetry data. Philos. Trans. R. Soc. B Biol. Sci. 365, 2187–2200 (2010).

    Article  Google Scholar 

  30. Péron, G. et al. The challenges of estimating the distribution of flight heights from telemetry or altimetry data. Anim. Biotelemetry 8, 1–13 (2020).

    Article  Google Scholar 

  31. Cargnelutti, B. et al. Testing global positioning system performance for wildlife monitoring using mobile collars and known reference points. J. Wildl. Manag. 71, 1380–1387 (2007).

    Article  Google Scholar 

  32. Edenius, L. Field test of a GPS location system for moose Alces alces under Scandinavian boreal conditions. Wildl. Biol. 3, 39–43 (1997).

    Article  Google Scholar 

  33. Jurdak, R., Corke, P., Dharman, D. & Salagnac, G. Adaptive GPS duty cycling and radio ranging for energy-efficient localization. In Proceedings of the 8th ACM Conference on Embedded Networked Sensor Systems-SenSys ’10 57–70 (ACM Press, 2010). https://doi.org/10.1145/1869983.1869990.

  34. Gau, R. J. et al. Uncontrolled field performance of Televilt GPS-SimplexTM collars on grizzly bears in western and northern Canada. Wildl. Soc. Bull. 32, 693–701 (2004).

    Article  Google Scholar 

  35. Girard, I. et al. Feasibility of GPS use to locate wild ungulates in high mountain environment. Pirineos 157, 7–14 (2002).

    Article  Google Scholar 

  36. Krüger, S., Reid, T. & Amar, A. Differential range use between age classes of Southern African bearded vultures Gypaetus barbatus. PLoS One 9, e114920 (2014).

    ADS  Article  CAS  PubMed  PubMed Central  Google Scholar 

  37. Augustine, B. C., Crowley, P. H. & Cox, J. J. A mechanistic model of GPS collar location data: Implications for analysis and bias mitigation. Ecol. Modell. 222, 3616–3625 (2011).

    Article  Google Scholar 

  38. Douglas, D. C. et al. Moderating Argos location errors in animal tracking data. Methods Ecol. Evol. 3, 999–1007 (2012).

    Article  Google Scholar 

  39. Cuadrat, J. M. et al. El clima de los Pirineos. Base de datos y primeros resultados. Tiempo Clima 45, 38–41 (2010).

    Google Scholar 

  40. Margalida, A., Bertran, J. & Heredia, R. Diet and food preferences of the endangered Bearded Vulture Gypaetus barbatus: A basis for their conservation. Ibis (Lond. 1859) 151, 235–243 (2009).

    Google Scholar 

  41. del Hoyo, J., Elliott, A., Sargatal, J., Christie, D. A. & Kirwan, G. Handbook of the Birds of the World,2 (Lynx Edicions, Barcelona, 1994).

    Google Scholar 

  42. Antor, R. J. et al. First breeding age in captive and wild bearded vultures Gypaetus barbatus. Acta Ornithol. 42, 114–118 (2007).

    Article  Google Scholar 

  43. Gil, J. A. et al. Home ranges and movements of non-breeding bearded vultures tracked by satellite telemetry in the Pyrenees. Ardeola 61, 379–387 (2014).

    Article  Google Scholar 

  44. Sunyer, C. El periodo de emancipación en el Quebrantahuesos (Gypaetus barbatus): Consideraciones sobre su conservación. In El quebrantahuesos (Gypaetus barbatus) en los Pirineos. Características Ecológicas y Biología (eds Heredia, R. & Heredia, B.) 47–65 (ICONA, Turin, 1991).

    Google Scholar 

  45. Margalida, A. et al. Uneven large-scale movement patterns in wild and reintroduced pre-adult Bearded Vultures: Conservation implications. PLoS One 8, e65857 (2013).

    ADS  CAS  Article  PubMed  PubMed Central  Google Scholar 

  46. Ellergren, H. First gene on the avian W chromosome (CHD) provides a tag for universal sexing of non-ratite birds. Proc. R Soc. Lond. Ser. B Biol. Sci. 263, 1635–1641 (1996).

    ADS  Article  Google Scholar 

  47. Cruz, S., Proaño, C. B., Anderson, D., Huyvaert, K. & Wikelski, M. Data from: The Environmental-Data Automated Track Annotation (Env-DATA) System: Linking animal tracks with environmental data. (2013). https://doi.org/10.5441/001/1.3hp3s250.

  48. Dodge, S. et al. The environmental-data automated track annotation (Env-DATA) system: Linking animal tracks with environmental data. Mov. Ecol. 1, 3 (2013).

    Article  PubMed  PubMed Central  Google Scholar 

  49. Cuscó, F., Cardador, L., Bota, G., Morales, M. B. & Mañosa, S. Inter-individual consistency in habitat selection patterns and spatial range constraints of female little bustards during the non-breeding season. BMC Ecol. 18, 1–12 (2018).

    Article  Google Scholar 

  50. Nakagawa, S. & Schielzeth, H. A general and simple method for obtaining R2 from generalized linear mixed-effects models. Methods Ecol. Evol. 4, 133–142 (2013).

    Article  Google Scholar 

  51. Anadón, J. D. et al. Factors determining the distribution of the spur-thighed tortoise Testudo graeca in south-east Spain: A hierarchical approach. Ecography (Cop.) 29, 339–346 (2006).

    Article  Google Scholar 

  52. R Foundation for Statistical Computing. R Core Team. R: A Language and Environment for Statistical Computing. https://www.R-project.org/ (2019).

  53. Bates, D., Maechler, M. & Dai, B. lme4: Linear mixed-effects models using S4 classes. 2009. R package version 0.999375-31. https://CRAN.R-project.org/package=lme4 (2009).

  54. Burnham, K. P. & Anderson, D. R. Model Selection and Multimodel Inference: A Practical Information-Theoretic Approach Vol 2 (Springer, Berlin, 2002).

    MATH  Google Scholar 

  55. Zuur, A., Ieno, E. N., Walker, N., Saveliev, A. A. & Smith, G. M. Mixed Effects Models and Extensions in Ecology with R (Springer, Berlin, 2009).

    MATH  Book  Google Scholar 

  56. Barton, K. Package ‘MuMIn’. R package version 1.43. 15. https://CRAN.R-project.org/package=MuMIn (2019).

  57. Fox, J. & Weisberg, S. An R Companion to Applied Regression (Sage publications, Thousand Oaks, 2018).

    Google Scholar 

  58. Mitchell, L. J., White, P. C. & Arnold, K. E. The trade-off between fix rate and tracking duration on estimates of home range size and habitat selection for small vertebrates. PLoS One 14, e0219357 (2019).

    CAS  Article  PubMed  PubMed Central  Google Scholar 

Download references

Acknowledgements

V. García-Matarranz, J. Canut, D. García, R. Heredia, I. Afonso, E. Vega, S. Pinedo and the forestry rangers of Pallars Jussà, Alta Ribagorça and Pallars Sobirà helped with the captures. We thank D. García (Generalitat of Catalunya), L.M. González and R. Moreno-Opo (Ministry for the Ecological Transition), M. Alcántara and J.M. Martínez (Government of Aragón), M. Razin (LPO), P. Muñoz, M. López Liberal, J.M. Vadillo, D. Campión and A. Llamas (Government of Navarra), and R. Heredia for their support.

Funding

This work was funded by the Spanish Ministry of Economy, Industry and Competitiveness (project CGL2015-66966-C2-2-R), the Ministry of Science, Innovation and Universities (RTI2018-099609-B-C22). RGJ was supported by pre-doctoral Grant (FPI/BES-2016-077510) granted by the Spanish Ministry of Economy and Competitiveness.

Author information

Authors and Affiliations

Authors

Contributions

All authors designed the study. A.M. conducted the fieldwork. R.G.J. wrote the manuscript with contributions from J.M.P.G. and A.M.; R.G.J. performed exploratory and statistical analysis under the supervision of J.M.P.G. All authors read and approved the final manuscript.

Corresponding authors

Correspondence to Ruth García-Jiménez, Antoni Margalida or Juan M. Pérez-García.

Ethics declarations

Competing interests

The authors declare that no competing interests.

Additional information

Publisher's note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary information

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

García-Jiménez, R., Margalida, A. & Pérez-García, J.M. Influence of individual biological traits on GPS fix-loss errors in wild bird tracking. Sci Rep 10, 19621 (2020). https://doi.org/10.1038/s41598-020-76455-x

Download citation

  • Received:

  • Accepted:

  • Published:

  • DOI: https://doi.org/10.1038/s41598-020-76455-x

Comments

By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.

Search

Quick links

Nature Briefing

Sign up for the Nature Briefing newsletter — what matters in science, free to your inbox daily.

Get the most important science stories of the day, free in your inbox. Sign up for Nature Briefing