HomeToolkit

Viewing tracks

The tracks can be accessed from the Project page by clicking ‘Tracks and analysis’ (located on the top panel under the Project menu) , or by clicking the View Tracks icon in the right of the screen. Alternatively, existing tracks can be accessed from Open projects (or projects that you have access rights to) directly from the ZoaTrack Repository. These can be accessed by clicking Browse the Repository in the website header, then searching for your relevant project – e.g. “Moose Translocation” This opens an interactive map showing the relocations contained in the tracking file plotted on a Google maps layer. Also shown are the trajectory the animals took (lines) and the start and end locations (green and red points). The page always faces due north, and a scale is displayed in the bottom left of the map. In the top left of the map are tools to pan (hand) and zoom (+/-) around the map and measure distances (ruler).


Extracting movement metrics

On visualising the animal tracks, movement metrics for each tagged animal are provided in the left hand window. These include:

  1. the date range
  2. the total number of locations for that animal
  3. the mean number of detections per day
  4. the maximum number of detections per day
  5. the distance moved along the track (km) – Estimated using Great circle distances on longitude latitude coordinates (Decimal Degrees)
  6. the mean step length (km)
  7. the mean step speed (km/h)

These field are automatically updated when a new date range is provided, or if tracks are edited using the Edit tracks tool.

How can I upload my data on ZoaTrack ?

If your data is associated with Sweden and if you're willing to release all data with open access, please contact us at zoatrack@bioatlas.se.

ZoaTrack provides a series of tools with which to analyse your animal tracking datasets. These include tools to visualise and extract movement metrics from tagged animals based on a date range (project layers: Trajectory and Detections) and tools to run more complex spatial analyses and extract home range estimates. These tools are available on the Tracks and analysis page by clicking the ‘Analysis’ tab and selecting a one of the tools.


Calculate home range area

The following tools estimate each animal’s home range as a measure of individual space usage. An animal’s home range is the area in which it lives and travels. This area is closely related to (but not identical with) the concept of "territory", which is the area that is actively defended by an individual. There are many variations of home range analysis, each has its own advantages and disadvantages depending on the data. More information and links to the respective publications are obtained by clicking the (?) located next to each home range tool. The choice of home range estimator and the parameter values can have a huge bearing on the final home range estimates. It is also possible to limit the date range (Dates) and/or the Animals from which home ranges and movement metrics are generated. Simply edit the date range and select the animals you are interested in, and then run the analyses. The results are displayed under the Animals tab in the left panel of the screen.


Minimum Convex Polygon

Otherwise known as a convex hull, this approach uses the smallest area convex set that contains the location data (Worton 1998). This calculation is undertaken within R using the adehabitatHR package (Calenge 2008). ZoaTrack will return the MCP calculation in the Analysis Results window. An image of the MCP will be produced for visualisation over the map image, as well as a KML file for viewing in Google Earth.

References

Calenge, C. (2006) The package adehabitat for the R software: a tool for the analysis of space and habitat use by animals. Ecological Modelling, 197, 516-519

Worton, B.J. (1995) A convex hull-based estimator of home-range size. Biometrics, 51, 1206-1215.


Kernel Utilization Distribution

The fixed kernel density estimator is a non-parametric method of home-range analysis, which uses the utilization distribution to estimate the probability that an animal will be found at a specific geographical location. This fixed method of kernel smoothing ignores the temporal sequence whereby locations were obtained, and assumes that all locations from that individual are spatially autocorrelated. This means that the location of an individual at a particular point implies an increased probability that it frequents neighbouring locations as well. The kernel UD accurately estimates areas of high use by the tagged animal, providing that the level of smoothing is appropriate.

These calculations are undertaken within R using the adehabitatHR library of functions (Calenge 2008).

References

Calenge, C. (2006) The package adehabitat for the R software: a tool for the analysis of space and habitat use by animals. Ecological Modelling, 197: 516-519

Seaman, D.E., Powell, R.A. (1996) An evaluation of the accuracy of kernel density estimators for home range analysis. Ecology, 77: 2075-2085.

Silverman, B.W. (1986) Density estimation for statistics and data analysis. Chapman and Hall, London, UK

Worton, B.J. (1989) Kernel methods for estimating the utilization distribution in home-range studies. Ecology 70: 164-168


Kernel Brownian Bridge

The Kernel Brownian Bridge approach calculates the utilization distribution to estimate the probability that an animal will be found at a specific geographical location. Unlike the classical Fixed Kernel approach, the Kernel Brownian Bridge incorporates serial autocorrelation between fixes (i.e. the time the animal took to move between locations) into the calculation (Bullard 1992). Brownian Bridges, therefore, only contribute to utilisation distributions when sequential locations (i.e. XYn at timen and XYn+1 at timen+1) occur close to one another in time (Kie et al. 2010).

This function uses information on the animal’s trajectory, how long it took to move between locations, how fast the animal moves on average (Sig1) and the uncertainty around each location fix (Sig2) to control the degree of kernel smoothing (Bullard 1992).

These calculations are undertaken within R using the adehabitatHR library of functions (Calenge 2008).

References

Bullard, F. (1991) Estimating the home range of an animal: a Brownian bridge approach. Master of Science, University of North Carolina, Chapel Hill

Calenge, C. (2006) The package adehabitat for the R software: a tool for the analysis of space and habitat use by animals. Ecological Modelling, 197, 516-519

Horne, J.S., Garton, E.O., Krone, S.M. and Lewis, J.S. (2007) Analyzing animal movements using brownian bridge. Ecology, in press

Kie J.G., Matthiopoulos J., Fieberg J., Powell R.A., Cagnacci F., Mitchell M.S., Gaillard J.M., Moorcroft P.R. 2010. The home-range concept: are traditional estimators still relevant with modern telemetry technology? Phil. Trans. R. Soc. B 365, 2221–2231


Alpha Hull

The alpha hull home range estimation is a generalisation of the convex hull but objectively crops low use areas from the polygon surface. Alpha hulls are generated by connecting all locations as a Delauney triangulation, then systematically removing vertices until only those vertices that are shorter in length than the chosen parameter value alpha are retained. The smaller the value of alpha, the finer the resolution of the hull and the greater the exposure of non-use areas. As alpha increases, the polygon surface will increase until it is equivalent to a 100% minimum convex polygon.

This calculation is undertaken within R using the alphahull package (Pateiro-Lopez & Rodriguez-Casal 2011). The analysis is heavy on computing resources and can take up to 20 minutes to calculate depending on the number of location fixes. Be patient.

References

Burgman, M.A. & Fox, J.C. (2003) Bias in species range estimates from minimum convex polygons: implications for conservation and options for improved planning. Animal Conservation, 6, 19-28.


Local Convex Hull

The Local Convex Hull (LoCoH) method estimates individual utilisation distributions based on the local nearest-neighbour convex hulls. These are formed by constructing convex hulls around each location in the animal’s trajectory then jointing these hulls together, iteratively, to form isopleths (Getz 2007). This is a useful home-range estimator when the movements of the animal have been constrained along hard edges such as roads, fences and rivers.

This calculation is undertaken within R using the LoCoH series of functions within the adehabitatHR library of functions (Calenge 2008). Users may either fix the number of nearest neighbours (k-1) to the root point (i.e. the fixed k-LoCoH), or fix the maximum radius from root points when generating local hulls (i.e. the fixed r-LoCoH). This analysis is heavy on computing resources and can take up to 20 minutes to calculate depending on the number of location fixes. Be patient.

References

Calenge, C. (2006) The package adehabitat for the R software: a tool for the analysis of space and habitat use by animals. Ecological Modelling, 197, 516-519

Getz, W.M. & Wilmers, C.C. (2004). A local nearest-neighbor convex-hull construction of home ranges and utilization distributions. Ecography, 27, 489–505.

Getz, W.M., Fortmann-Roe, S.B, Lyons, A., Ryan, S., Cross, P. (2007). LoCoH methods for the construction of home ranges and utilization distributions. PLoS ONE, 2: 1–11.


Heat Map (Point Intensity)

This generates a grid over the study area and uses a coloured gradient to visually identify areas of high usage by the tagged animal. These can be applied to either points or connectivity lines between points. The size of the grid cells (in meters) can be specified. This ZoaTrack tool utilises the spatstat package in R (Baddeley & Turner, 2005)

References

Baddeley, A. & Turner, R. (2005) spatstat: An R package for analyzing spatial point patterns. Journal of Statistical Software, 12,6


Heat Map (Line Intensity)

This generates a grid over the study area and uses a coloured gradient to visually identify areas of high usage by the tagged animal. These can be applied to either points or connectivity lines between points. The size of the grid cells (in meters) can be specified. This ZoaTrack tool utilises the spatstat package in R (Baddeley & Turner, 2005)

References

Baddeley, A. & Turner, R. (2005) spatstat: An R package for analyzing spatial point patterns. Journal of Statistical Software, 12,6


Kalman Filter

This filter applies a state-space model combined with a kalman filter to your location data (logitude and latitude) to help predict the ‘most probable’ track through time. This combines the location and temperature data. More information on this particular method (i.e. the extended kalman filter) can be found in Sibert et al. (2003), Sibert et al. (2006) and the kftrack package documentation on which this code is based at https://code.google.com/p/geolocation/wiki/ArticleKftrack.

References

Sibert, J.R., Musyl, M.K. & Brill, R.W. (2003) Horizontal movements of bigeye tuna (Thunnus obesus) near Hawaii determined by Kalman filter analysis of archival tagging data. Fisheries Oceanography, 12(3):141-151.

Sibert, J. R., Lutcavage, M.E., Nielsen, A., Brill, R.W. & Wilson, S.G. (2006) Interannual variation in large-scale movement of Atlantic bluefin tuna (Thunnus thynnus) determined from pop-up satellite archival tags. Canadian Journal of Fisheries and Aquatic Sciences 63: 2154-2166.


Kalman Filter (SST)

If tag-recorded sea-surface temperature (SST) data are available for each location, this filter applies a state-space model combined with a kalman filter to your location data (logitude, latitude) and SST to help predict the ‘most probable’ track through time. In an attemt to improve the model’s predictions, the tag-recorded SST is matched with external SST data collected by the National Oceanic and Atmospheric Administration (NOAA). This model is adapted from the kfsst function, in the ukfsst R package https://code.google.com/p/geolocation/wiki/ArticleUkfsst. More information on this particular method can be found in Lam, Nielsen & Sibert (2008).

References

Lam, C.H., Nielsen, A. & Sibert, J.R. (2008) Improving light and temperature based geolocation by unscented Kalman filtering. Fisheries Research, 91: 15-25.


Exploring the data in Google Earth

To gain a fuller understanding of the movements of the animals, it is useful to visualise the animal’s trajectory through time.

  1. If not already installed, download Google earth from http://www.google.com/earth/index.html
  2. In ZoaTrack, make sure you are still on the ‘Animals’ tab in the Tracks and analysis page.
  3. Below an animal ID, click ‘KML’ in the Trajectory box. This will convert the ZoaTrack -generated trajectory into a Google earth file.
  4. Click the downloaded .kml file to open and view in Google earth.
  5. The tracks in your ZoaTrack project are now visible with the last location represented as an arrow. You can visualise the animal’s trajectory through time by moving the time slider in the top left corner of Google Earth.
  6. To visualise the home ranges, click ‘KML’ in the MCP results box or that of any other home ranges you may have generated. This will convert the ZoaTrack -generated home range polygon into a Google Earth file.
  7. Click the downloaded .kml file to open and view in Google earth.

Speed filter

You can select a speed filter to remove unlikely locations (i.e. those where the animal would have to attain a certain sustained velocity to achieve: e.g.> 50 km/h). To use this tool, click Speed filter and type in a Maximum speed that you animal could hypothetically obtain. Click Apply filter and those relocations exceeding this maximum will have been removed. When you are happy with your tracks, click the arrow next to ‘Back to Project’ at the top of the left panel and select ‘Tracks and analysis’.


Argos Location Class and Dilution of Precision Filter

If your location dataset had a column containing the Argos Location Class (Argos tracking data only) or a DOP Class (GPS data only), this filter can be applied to remove locations with a low estimated accuracy. To use this tool, click the relevant filter (Argos or DOP) and apply the minimum accuracy with which you wish to visualise and run subsequent analyses on. Click Apply filter and those relocations exceeding this minimum will have been removed. When you are happy with your tracks, click the arrow next to ‘Back to Project’ at the top of the left panel and select ‘Tracks and analysis’


Kalman filter and Kalman filter SST

Often telemetry technologies with low precision and accuracy can result in highly improbable animal trajectories. For example, studies on the accuracy of light-based geolocation have recognized that raw geolocations are often imprecise and biased, particularly for estimates of latitude during equinox periods. The state-space Kalman filter approach can estimate the “most probable” track from imprecise and biased location estimates (and sea surface information if it is available from the tag sensor). These models are adapted from the kftrack and ukfsst functions developed by Nielsen and Sibert, 2004 and Lam, Nielsen & Sibert, 2008. Click Kalman filter and use your known information on the start and end date and location to match the actual days when the tag was deployed and retrieved. By providing a known start and end location, this can often provide a more realistic track for a given set of noisy data. Note, please enter the numbers only, without units (e.g. without ° E) Click Run filter to run the kalman filter (kftrack) on the raw geolocation data. The spinning wheel will show that ZoaTrack is processing your request. Once the Kalman filter is complete, the ‘most probable’ track (white triangles) will be overlaid on the original track. On completion the model parameters and model results will be displayed on the left of the map Models can also be re-run with a new set of parameters by editing the Advanced parameters fields. The systematic error (or bias) in the estimation of position of Longitude and Latitude can be changed to 0 degrees by deselecting bx.active and by.active buttons. The systematic error can be adjusted on the Longitude and Latitude by entering the predicted error. Click Run filter to re-run the kalman filter To replace the original track and calculate the movement metrics for this filtered track (i.e. track length, step length, speed), click Replace track. The original data points now appear as red crosses. When you are happy with your tracks, click the arrow next to ‘Back to Project’ at the top of the left panel and select ‘Tracks and analysis’. The movement metrics for the Kalman filtered track are now displayed in the left hand window


Using and Citing data from the ZoaTrack data repository

If you use data from ZoaTrack in any type of publication then you must cite the project DOI (if available) or any published peer-reviewed papers associated with the study. Please contact the data custodians to discuss data usage and appropriate accreditation.

Cite the ZoaTrack platform

If you publish data from the ZoaTrack data repository or use any of the analysis tools to process and sythesise your animal tracking data then please citethe following paper:

R. G. Dwyer, C. Brooking, W. Brimblecombe, H. A. Campbell, J. Hunter, M. E. Watts, C. E. Franklin, "An open Web-based system for the analysis and sharing of animal tracking data", Animal Biotelemetry 3:1, 29 Jan 2015, DOI 10.1186/s40317-014-0021-8.

References for specific analysis tools can be found here.