guzhongru / above

R package for analysis of animal movement and environment data - primarily for internal collaboration of Animals on the Move / ABoVE project

The above R package

authors: Elie Gurarie and Peter Mahoney and Scott LaPoint and Sarah Davidson and others

To-Do:

• Include simulated data for quick loading and examples
• Improve multimigration (EG)
• Add step-selection functions (PM)
• Annotation code

Background

This R package will collaboratively help participants of the Animals on the Move subproject of the Arctic Boreal Vulnerability Experiment (above) share code for analysis of animal movements and migrations.

For now, it contains:

1. preprocesseing movebank.org functions to get daily means
2. some convenient methods (summary, plot, map.track) to work with movetrack's
3. functions for multi-migration analysis

Example below:

Install package

From GitHub, you need to install first the marcher package, then this package:

require(devtools)
install_github("EliGurarie/marcher")
install_github("ABoVE-AotM/above")

Loading eagle data

ABoVE members with access to some golden eagle data can load it using

login <- movebankLogin(username = "somethingsecret", password = "somethingsecret")

Load a few datasets:

ge1 <- getMovebankData(study="ABoVE: HawkWatch International Golden Eagles", animalName="37307a", login=login) 
ge2 <- getMovebankData(study="Aquila chrysaetos interior west N. America, Craigs, Fuller", animalName="629-26704", login=login) 

These are Move objects, but they're slightly different:

is(ge1)

##  [1] "Move"                   ".MoveTrackSingle"      
##  [3] ".MoveGeneral"           ".OptionalMove"         
##  [5] ".MoveTrack"             ".unUsedRecords"        
##  [7] "SpatialPointsDataFrame" "SpatialPoints"         
##  [9] "Spatial"                "SpatialVector"

is(ge2)

##  [1] "MoveStack"              ".MoveTrackStack"       
##  [3] ".MoveGeneral"           ".OptionalMove"         
##  [5] ".MoveTrack"             ".unUsedRecordsStack"   
##  [7] "SpatialPointsDataFrame" ".unUsedRecords"        
##  [9] "SpatialPoints"          "Spatial"               
## [11] "SpatialVector"

Processing data

For migration analysis we simplify (and get daily averages) using the processMovedata function, which reduces the data to daily average locations (in latitude, longited and x and y):

ge1.simple <- processMovedata(ge1, idcolumn = "deployment_id")
head(ge1.simple)

##   movebank_study        id                time      lon    lat        x
## 1           <NA> 171299043 2003-10-06 04:45:00 -106.411 34.704 -106.411
## 2           <NA> 171299043 2003-10-06 17:44:53 -106.473 34.644 -106.473
## 3           <NA> 171299043 2003-10-06 19:25:04 -106.227 34.215 -106.227
## 4           <NA> 171299043 2003-10-06 20:04:42 -106.111 33.898 -106.111
## 5           <NA> 171299043 2003-10-08 23:44:41 -105.922 31.710 -105.922
## 6           <NA> 171299043 2003-10-09 00:49:38 -105.888 31.705 -105.888
##        y                                            proj4string
## 1 34.704 +proj=longlat +ellps=WGS84 +datum=WGS84 +towgs84=0,0,0
## 2 34.644 +proj=longlat +ellps=WGS84 +datum=WGS84 +towgs84=0,0,0
## 3 34.215 +proj=longlat +ellps=WGS84 +datum=WGS84 +towgs84=0,0,0
## 4 33.898 +proj=longlat +ellps=WGS84 +datum=WGS84 +towgs84=0,0,0
## 5 31.710 +proj=longlat +ellps=WGS84 +datum=WGS84 +towgs84=0,0,0
## 6 31.705 +proj=longlat +ellps=WGS84 +datum=WGS84 +towgs84=0,0,0
##        dateDownloaded
## 1 2017-02-23 00:48:13
## 2 2017-02-23 00:48:13
## 3 2017-02-23 00:48:13
## 4 2017-02-23 00:48:13
## 5 2017-02-23 00:48:13
## 6 2017-02-23 00:48:13

This is a track object (specific for this package), which has some convenient methods:

summary(ge1.simple)

##          id    n               start duration  dt.median
## 1 171299043 1002 2003-10-06 04:45:00 391 days 9.37 hours

Note that x and y are UTM coordinates - you can either provide a proj4 projection string or, by default, it will pick the zone of the midpoint.

The plotting function is similar to the scan.track function in marcher

plot(ge1.simple)

## png 
##   2

For the second data set (ge2) there are three deployments of 1 eagle, each with a unique identifier.

ge2.simple <- processMovedata(ge2, idcolumn = "deployment_id")
summary(ge2.simple)

##          id   n               start duration   dt.median
## 1 196430584 247 1993-01-07 21:58:39 217 days 21.15 hours
## 2 196430597 276 1994-01-27 22:40:46 105 days  9.17 hours
## 3 196430599  98 1995-12-10 02:32:49 338 days 83.52 hours

And we see a simple straight one-time migration in these data.

plot(subset(ge2.simple, id == "196430584"))

## png 
##   2

Migration analysis

Examples to come