Migration Mapper is meant to facilitate the technologically challenging analysis of animal movement and migration data. Migration Mapper is meant to be used by researchers, biologists, and managers to analyze fine-scale GPS collar data collected from ungulates. The workflow generally consists of: 1) importing cleaned GPS collar data in ESRI Shapefile format; 2) quantifying whether individual animals are migratory, and if so, identifying the date of the start and end of migration; 3) calculating the utilization distribution of each individual’s migration movements using a Brownian Bridge Movement Analysis; and finally 4) grouping animals into distinct populations and calculating population level migration corridors. Migration Mapper provides the user with a number of outputs including shapefiles of high, medium, and low use migration corridors and stopover sites, and analogous shapefiles for population-level winter ranges. Further, Migration Mapper spits out a table of the start and end of annual spring and fall migration for every individual animal, raster layers of the individual and population-level utilization distributions of migratory corridors and winter ranges, and shapefiles of the GPS locations used in the Brownian Bridge Movement Analysis.
What you need
RUNNING MIGRATION MAPPER
Tab 1 - Import data
Tab 2 - Date column(s)
Tab 3 - Date configuration
Tab 4 - Import results
Tab 5 - Migration dates
Tab 6 - Brownian bridge
Tab 7 - Animal grouping - migrations
Tab 8 - Animal grouping - winter
Tab 9 - Final corridors
Final tab - Analysis parameters
Closing/Exiting Migration Mapper
Updating an existing project
Adding new GPS collar data to an existing project
Understanding and mapping where animals go is of utmost importance to wildlife conservation and management. This is particularly important for migratory wildlife as they rely on connected, seasonal habitats. Given increasing levels of energy development and recreation on public lands, sprawling housing development on private lands, and increasing traffic volumes on roadways, the long-term persistence of migration corridors is uncertain. It is known, for example, that development can alter space use on winter range, affect the speed at which animals migrate, force animals to detour from established routes, and in some cases, impede migration altogether.
In 2009, Sawyer and colleagues devised a method for identifying and prioritizing migration corridors, which provided a way forward to begin mapping and conserving ungulate migrations. The methodology involves a brownian bridge movement analysis of GPS collar data collected from individual animals and a prioritization analysis based on the number of animals migrating through specific areas. The framework has been adopted by the Wyoming Migration Initiative (WMI) to develop their migration assessment program, which includes identifying threats and opportunities for conservation along migration corridors. WMI pioneered this approach by assessing the "Red Desert to Hoback" (RDH) mule deer migration, a 150 mile-long corridor that is the longest yet recorded in the Lower 48. The RDH migration assessment has provided a road map for managers and NGOs seeking to sustain this corridor. It has guided the targeting of conservation easements, fence modifications, highway crossing improvement, and land-use planning on BLM lands.
The issue with these methods, however, is that they are computationally intensive, and require coding and analysis skills that are beyond the reach of most biologists and managers. Many states have GPS collar datasets of ungulates, and many are currently collecting data. Although these datasets exist, technological barriers have stymied the evaluation and adoption of the new methods that would allow any manager or biologist to identify, assess, and prioritize migration corridors in a consistent way.
The goal of Migration Mapper is to simplify the technologically challenging analyses outlined in Sawyer et al. (2009) into a user-friendly software that can be implemented by biologists and managers using their own GPS collar data. Migration Mapper was not developed to be a ‘pull-the-handle-and-see’ program, but to incorporate information from the user in identifying migration corridors. Thus, expect to spend some time working with your data before Migration Mapper spits out a migration corridor map. Link to Table of Contents
Migration Mapper is built using R as the processing environment. R is a command line-based, open-source software for completing complex data analysis, modeling, and visualizations. On top of the R environment, Migration Mapper uses RStudio and Shiny to build an easy to use user-interface. That means you don’t need to look at any R code! RStudio and Shiny provide a framework for interacting with R through a variety of easy to use tools and functions.
To begin, R will need to be installed. If you already have R installed you will want to make sure you have at least version 3.5. If you have an earlier version of R you will need to update R, or uninstall your current version and download and install the latest version of R. To Install R, follow these instructions:
After installing R, RStudio needs to be installed. We suggest having at least version 1.1.453 of RStudio installed on your computer. If you have an earlier version of RStudio, it is recommended that you upgrade. To Install RStudio, follow these instructions:
Download the zipped file titled MigrationMapperv1.2.zip from the Wyoming Migration Initiative website under the Projects tab. Once you unzip this file (and, you MUST unzip this file), enter the folder title collarProcessingApp. Within this folder you will see some subfolders.
The app folder has all files that are needed to run Migration Mapper. It is not recommended that you edit these files as you may corrupt your copy of the software. The sampleData folder contains sample datastets that we walk you through in this user guide. Link to Table of Contents
After installing R and Rstudio, you will need to check that RStudio is using the most recent version of R you have on your computer. To do this, open RStudio and go to the tools menu then Global Options. Make sure that the R installation matches the most recent version installed on your computer. Link to Table of Contents
To run Migration Mapper for the first time, you’ll need to open the file called MigrationMapperv1.2.r from the app folder. Double clicking on MigrationMapperv1.2.r will automatically open RStudio, and you will then see some code appear in the RStudio window. If double clicking on MigrationMapperv1.2.r does not automatically open RSudio, then you will manually need to open RStudio, click on File, then Open File, and then navigate to where MigrationMapperv1.2.r is saved and open it. Link to Table of Contents
Next, you will need to select how RStudio will run Migration Mapper. Migration Mapper can be run in your web browser but we suggest running Migration Mapper in your web browser for best results. To select this option choose Run External from the dropdown menu next to Run App. Link to Table of Contents
Now you are ready to start Migration Mapper. Click the text Run App to start. If you’ve not run Migration Mapper before, several packages will be installed when you first click Run App. This could take several minutes as RStudio installs the necessary libraries from the internet. Link to Table of Contents
The image to the right shows the first page you will see after loading Migration Mapper. This is where your GPS collar data will be imported. Migration Mapper was built to import shapefiles of fine-scale GPS collar data. Shapefiles can be imported with several individuals merged together or as unique files for each animal. If importing unique files for each animal, it is necessary that all files have identical columns, data formats and GIS projections. If importing a merged file, it is necessary that the file includes a column delineating unique animal IDs.
You have two choices on this first screen. First, you can start a new project by clicking CLICK TO CHOOSE FOLDER. Alternatively, you can re-load an existing project using the button RESUME EXISTING PROJECT. The following instructions talk you through how to start a new project. Instructions for resuming an existing project are outlined in the section entitled “Updating a current project” below.
To start a new project, you will need to select the folder that contains your data to import. Click the button CLICK TO CHOOSE FOLDER. A dialog box will open that you can use to select the appropriate folder. Note that sometimes this dialogue box does not pop up in front of Migration Mapper, but it will show up on your taskbar. For this demonstration, we will select one of the sample datasets included with Migration Mapper.
Once you select the appropriate folder, you will see a message indicating that you’ve selected an appropriate folder. If you choose a directory that does not contain valid shapefiles, you will receive an error message. If this occurs, check the folder’s content and verify your shapefiles are located there. To choose the file(s) for import click the empty box above the BEGIN FILE IMPORT BUTTON. A menu will appear where you can select the appropriate file(s) for import. After choosing file(s), click BEGIN FILE IMPORT BUTTON to start loading shapefile(s).
Next, you will need to choose an empty working directory or project folder. This project folder is where various files will be written and stored while Migration Mapper runs. Once finished running Migration Mapper, this is also the directory where all the results will be stored and available to you to review in ArcGIS or other GIS software. This directory should be an empty folder on a local drive with sufficient storage space. For instance, an elk dataset from Wyoming with 57 animals required 2 GB of space. To choose this folder click the CLICK TO CHOOSE DIRECTORY button. This will open a dialog you can use to choose the appropriate folder.
You now need to select which column in your dataset represents the unique, animal ID. If you imported a single merged shapefile, this column should have been created before importing your dataset. Selecting the appropriate animal ID allows Migration Mapper to create unique migration corridors for each animal. If you imported many unique shapefiles for each animal ID and do not have an animal ID column, choose NA and a unique ID will be created from the filename from each file. After you choose the appropriate option in the menu, click the EXECUTE button to move to the next step. From the sample dataset, we will select the column titled id. Link to Table of Contents
After clicking execute on Tab 1, Migration Mapper will change to Tab 2. This second tab is used to select date and time information from your dataset. Here you can see a sample of attributes from the imported dataset. Note that properly identifying the date and time information is critical for the analyses. Thus, you should be aware of how the date/time stamps for each GPS location are stored.
Based on our experiences, we recommend two formats for date/time information. First, date/time should be stored as text in a single column. For instance, date and times could be stored as “2015-05-26 14:01:52.” It does not matter the order at which the date/time elements are pasted together nor the type of character that separates them. More important is that you, as the user, know how the date time column is formatted, and that the format is consistent throughout the entire column of the shapefile. Second, the elements of dates and times could be stored all as separate columns in numeric format. In other words, the shapefile could contain a column for year, month, day, hour, minute, and second.
Migration Mapper allows for importing date/time data stored in one column or across several columns. Using the checkboxes, you need to select the appropriate column or columns that contain the date and time. After choosing the appropriate columns, click the DONE SELECTING DATE COLUMNS button. Link to Table of Contents
Once you proceed to Tab 3, you will see a reminder about double checking the selected columns. You can see the column(s) you chose on the right side of the screen along with some sample values. Make sure you’ve selected the correct columns before proceeding with this step. If you see you have chosen incorrect columns, go back to tab 2 and add the appropriate columns.
The date configuration tab is very important as it is used to indicate how your date/time information is stored. Migration Mapper has simplified this process and the only thing you need to do is to indicate the order of the date time elements. To do this, you must repeatedly choose date/time components from the drop down meanu. For example, the sample data shows a first date/time value of 2011-02-01 08:00:00. This represents February 1st, 2011 at 08:00:00 AM. We know the order of these values is then year, month, day, hour, minute and finally second. We need to choose this appropriate order from the top menu titled Choose date time elements for: date.
Next, you need to indicate what time zone your date/time stamps are stored in. This sample dataset was collected in UTC which is the same as GMT so no change is needed. After choosing your timezone and making sure you have made correct selections, click the PROCESS DATES button to process the data.
Many calculations happen at this point. These include separating animals by year and id, checking for mortalities, abnormalities, excessive speeds, identifying sequentially connected GPS points (i.e., bursts) and a few other steps. This can take a few minutes and you can monitor the progress in the bar at lower right. Link to Table of Contents
After Migration Mapper finishes processing your dates and importing your data, you will see a page summarizing the import results. This page shows how many individuals were imported along with the time range of imported data. You will also see a summary of points that were found to be a problem. For instance Migration Mapper checks if your animal has not moved for a certain amount of time, or is moving at an unreasonable speed. If you see any errors in the import results on this page that are unexpected, you may need to check your dataset and try importing again. See troubleshooting section for some tips for preparing your data. On the final tab of Migration Mapper (named ANALYSIS PARAMETERS), you can also adjust some of the parameters of how Migration Mapper determines bad GPS locations that are from unreasonable speeds or animal mortality. After you’ve finished reviewing the results you can proceed to step 5 by clicking the tab 5 box in Migration Mapper’s header. Link to Table of Contents
The next tab of Migration Mapper allows you to select start and end dates for the spring and fall migration for each animal, during each year. You can also omit specific seasons and years if the individual is not migratory or there are errors in the dataset. Note that Migration Mapper considers that the biological year begins on February 1st, and ends on January 31st. This way, if there are animals still migrating in January, it can be identified as part of fall migration.
The most effective way of indicating start and end dates is to look at a graph of displacement over time. Displacement is calculated by identifying the mean x and y location from the first few days of monitoring in February or whenever you first captured your animal, and then calculating the distance (in km) each GPS location is from that location. In other words, these initial february points should theoretically represent the location of the animal’s winter range, and displacement allows you to assess whether the animal has distinct seasonal ranges. An animal that migrates should display a displacement curve that has a hump’ shape, where the animal is close to its winter range in late winter, then moves to a distinct summer range in spring that is a significant distance from the winter range, and then at the end of summer moves back to its winter range. There is a wide variety of displacement curves, and we recommend that you review the references noted at the end of this document. Note that you can plot displacement or Net Squared Displacement, whatever your preference, by adjusting the toggle on the analysis parameters tab.
This page will load with the first animal and first year present in your database. From here you must use the sliders to choose the start and end dates for both the spring and fall migration.
To get started, view the NSD_ANNUAL graph and use the sliders to delineate the period showing the greatest change in displacement for each season. Link to Table of Contents
After choosing the general period of the spring and fall migration, you can then view the NSD SPRING and NSD FALL graphs to better refine the period you’ve chosen. You will notice as you select the migration periods, the map will also update. Spring migration points will be shown in green while fall migrations will be shown in orange. The map can be very useful for identifying animals with complicated movement patterns. You can zoom in to the map and more closely review the movement trajectories of the animal. Each point in the map represents the first location for the animal on a given day. The background of the map can be changed by clicking the upper right hand corner of the map.
After you finish delineating spring and fall migration for each animal/year, click the NEXT ANIMAL/YEAR button to move to the next individual or next year of the same individual.
It is inevitable that you will come across animals or years with erroneous data or when no migration occurred. If this happens, be sure to select the No migration present options under the headings of Does spring migration occur? and Does summer migration occur?.
Once you have finished all animals and years in your dataset, you will be prompted to proceed to Tab 6 - Brownian Bridge. If you have missed any animals, you will get a message telling you which individuals still need their migration dates identified. Link to Table of Contents
After proceeding to tab 6, you will see a blank map, which will be populated after the Brownian Bridge Movement Analysis has been completed. Click the Start Brownian Bridge button to start running the brownian bridge analysis. Using the Analysis Parameters tab, you can choose between regular Brownian Bridge or Dynamic Brownian Bridge, and the raster resolution (or cell size) that the Brownian Bridge analysis uses. The regular Brownian Bridge analysis takes much longer than the Dynamic Brownian Bridge. Further, the smaller the raster resolution you choose, the slower the Brownian Bridge analysis will take to complete. The analysis can take a few minutes to several hours. Running the sample dataset on a GIS workstation with a computer processor with 12 cores and 20gb RAM completed in 2 minutes using the Dynamic Brownian Bridge option and a raster resolution of 150 meters. Another trick to making the Analysis run faster is by properly specifying the number of CPU cores that your computer has on the Analysis Parameters page.
After the brownian bridge analysis is complete, the output on this tab provides you a map of each individual’s migration corridors (and winter ranges, if you specified that you also wanted winter ranges to be calculated on the Analysis Parameters tab). See below for an example. Your results may look different depending on what dates were selected for each migration period. You can click on individual polygons to retrieve the animal ID associated with each polygon.
After exploring the results, you can choose tab 7 to proceed to the animal grouping page.Link to Table of Contents
Calculating a population-level migration corridor is dependent on your definition of which animals constitute a population. In most cases, your dataset will include only animals captured in the same area, thus they can be classified as a single population. However, in other cases, you may want to group your animals into different populations before producing the final corridors. If your animals constitute a single population, you can simply skip this step and click on tab 8 (if you are calculating winter ranges as well) or tab 9. However, if you would like to assign individuals into 2 or more populations, all you have to do is identify the names of these groups (i.e., populations) and then assign each individual animal into one of those groups. Once each individual has been assigned to a group, you can proceed. If you don’t assign an individual to a group, it will automatically be given its own ‘blank’ group name. Below is an example of a dataset that we deemed individuals to be a part of two different groups:
By specifying so on the Analysis Parameters tab, you can calculate winter range polygons as well. This analysis mirrors that of the migration corridor analysis but is geared towards calculating core and high, medium, and low use winter ranges. On tab 8, you can create groups (i.e., populations) as in tab 7 for migration corridors. Link to Table of Contents
For each group you specify for migration corridors and winter ranges, Migration Mapper spits out the following final files in the PopulationShapefiles folder within the Project folder:
The ANALYSIS PARAMETERS tab has been created to allow you to customize some of the analysis parameters. Default values are reasonable, but you can adjust as you see fit. Detailed information about each parameter is listed below:
If at any time if you would like to start over, or you want to close Migration Mapper, here is how to do it. First, you must close Migration Mapper by clicking the ‘X’ in the upper right hand corner of the window. Second, you must click the ‘X’ in the upper right hand corner of RStudio. It will ask you whether you want to save the workspace image. Click no. Link to Table of Contents
If you have completed a project, and would like to update it in any way here is how you do it. Open Migration Mapper as instructed at the beginning of this document. However, instead of clicking CLICK TO CHOOSE FOLDER, you can re-load an existing project using the button RESUME EXISTING PROJECT. After clicking this button, you must navigate to the folder where the project exists. After the project has been loaded, you can adjust analysis parameters and rerun the Brownian Bridge analysis or you can adjust the dates of start and end of migration and then rerun the Brownian Bridge analysis. Link to Table of Contents
If you have been working on a project, and new data come in, you can integrate these new data into your project. To do this, you must first open Migration Mapper and click on RESUME EXISTING PROJECT. Navigate to your project’s folder and let Migration Mapper load up your existing project. Once your project is loaded up, then click on CLICK TO CHOOSE FOLDER and navigate to the folder where your new data exists. You then follow the same process of identifying the animal id and date/time information for these new data. Once that is complete Migration Mapper will merge the two datasets, and provide an updated tab 4 (IMPORT RESULTS). You can then continue using Migration Mapper as normal. Link to Table of Contents
Migration Mapper provides you two different datasets for testing. Both of the datasets are in ArcGIS shapefile format and located in the 'sampleData' folder within the MigrationMapperv1.2 folder. These datasets include GPS locations of female mule deer captured in Wyoming. The first dataset 'MuleDeer_PlatteValley_testdata' represents 10 individuals captured in Platte Valley. The second dataset 'gps_XXX' represents 15 individuals captured near Pinedale. The second dataset is broken into multiple files, each representing a single individual. These datasets are only to be used for testing and practicing purposes. They are not to be used for any other purpose or sent to any third party user. Link to Table of Contents
Please direct questions related to Migration Mapper to Jerod Merkle at . Link to Table of Contents
Börger, L., & Fryxell, J. M. (2012). Quantifying individual differences in dispersal using net squared displacement. Dispersal ecology and evolution, 222-230.
Bunnefeld, N., Börger, L., van Moorter, B., Rolandsen, C. M., Dettki, H., Solberg, E. J., & Ericsson, G. (2011). A model‐driven approach to quantify migration patterns: individual, regional and yearly differences. Journal of Animal Ecology, 80(2), 466-476.
Horne, J. S., Garton, E. O., Krone, S. M., & Lewis, J. S. (2007). Analyzing animal movements using Brownian bridges. Ecology, 88(9), 2354-2363.
Kranstauber, B., Kays, R., LaPoint, S. D., Wikelski, M., & Safi, K. (2012). A dynamic Brownian bridge movement model to estimate utilization distributions for heterogeneous animal movement. Journal of Animal Ecology, 81(4), 738-746.
Sawyer, H., Kauffman, M. J., Nielson, R. M., & Horne, J. S. (2009). Identifying and prioritizing ungulate migration routes for landscape-level conservation. Ecological Applications, 19(8), 2016-2025.
Singh, N. J., Allen, A. M., & Ericsson, G. (2016). Quantifying migration behaviour using net squared displacement approach: Clarifications and caveats. PloS one, 11(3), e0149594.
Support for the development of Migration Mapper has come from The Wyoming Migration Initiative, Western Association of Fish and Wildlife Agencies, The Pew Charitable Trust, and The Knobloch Family Foundation. We especially thank Josh Gage (with Gage Cartographics LLC) and Jerod Merkle for writing the majority of code and leading the development of the App. Link to Table of Contents
Copyright 2017 University of Wyoming. All Rights Reserved. Please click here to read the license agreement.