Search for species occurrence data across many data sources.

Search on a single species name, or many. And search across a single or many data sources.

Usage

occ(
  query = NULL,
  from = "gbif",
  limit = 500,
  start = NULL,
  page = NULL,
  geometry = NULL,
  has_coords = NULL,
  ids = NULL,
  date = NULL,
  callopts = list(),
  gbifopts = list(),
  inatopts = list(),
  ebirdopts = list(),
  vertnetopts = list(),
  idigbioopts = list(),
  obisopts = list(),
  alaopts = list(),
  throw_warnings = TRUE
)

Arguments

query: (character) One to many scientific names. See Details for what parameter in each data source we query. Note: ebird now expects species codes instead of scientific names - we pass you name through rebird::species_code() internally
from: (character) Data source to get data from, any combination of gbif, inat, ebird, vertnet, idigbio, obis, or ala. See vignette(topic = 'spocc introduction') for more details about these sources.
limit: (numeric) Number of records to return. This is passed across all sources. To specify different limits for each source, use the options for each source (gbifopts, inatopts, and ebirdopts). See Details for more. Default: 500 for each source. BEWARE: if you have a lot of species to query for (e.g., n = 10), that's 10 * 500 = 5000, which can take a while to collect. So, when you first query, set the limit to something smallish so that you can get a result quickly, then do more as needed.
start, page: (integer) Record to start at or page to start at. See Paging in Details for how these parameters are used internally. Optional
geometry: (character or nmeric) One of a Well Known Text (WKT) object, a vector of length 4 specifying a bounding box, or an sf object (sfg, sfc, or sf). This parameter searches for occurrences inside a polygon - converted to a polygon from whatever user input is given. A WKT shape written as POLYGON((30.1 10.1, 20 40, 40 40, 30.1 10.1)) would be queried as is, i.e. http://bit.ly/HwUSif. See Details for more examples of WKT objects. The format of a bounding box is min-longitude, min-latitude, max-longitude, max-latitude. Geometry is not possible with vertnet right now, but should be soon. See Details for more info on geometry inputs.
has_coords: (logical) Only return occurrences that have lat/long data. This works for gbif, rinat, idigbio, and vertnet, but is ignored for ebird. You can easily though remove records without lat/long data.
ids: Taxonomic identifiers. This can be a list of length 1 to many. See examples for usage. Currently, identifiers for only 'gbif' for parameter 'from' supported. If this parameter is used, query parameter can not be used - if it is, a warning is thrown.
date: (character/Date) A length 2 vector containing two dates of the form YYY-MM-DD. These can be character of Date class. These are used to do a date range search. Of course there are other types of date searches one may want to do but date range seems like the most common date search use case.
callopts: Options passed on to crul::HttpClient, e.g., for debugging curl calls, setting timeouts, etc.
gbifopts: (list) List of named options to pass on to rgbif::occ_search(). See also occ_options()
inatopts: (list) List of named options to pass on to internal function get_inat_obs
ebirdopts: (list) List of named options to pass on to rebird::ebirdregion() or rebird::ebirdgeo(). See also occ_options()
vertnetopts: (list) List of named options to pass on to rvertnet::searchbyterm(). See also occ_options().
idigbioopts: (list) List of named options to pass on to ridigbio::idig_search_records(). See also occ_options().
obisopts: (list) List of named options to pass on to internal function. See https://api.obis.org/#/Occurrence/get_occurrence and obis_search for what parameters can be used.
alaopts: (list) List of named options to pass on to internal function.
throw_warnings: (logical) occ() collects errors returned from each data provider when they occur, and are accessible in the $meta$errors slot for each data provider. If you set throw_warnings=TRUE, we give these request errors as warnings with warning(). if FALSE, we don't give warnings, but you can still access them in the output.

Value

an object of class occdat, with a print method to give a brief summary. The print method only shows results for those that have some results (those with no results are not shown). The occdat class is just a thin wrapper around a named list, where the top level names are the data sources:

gbif
inat
ebird
vertnet
idigbio
obis
ala

Note that you only get data back for sources that were specified in the from parameter. All others are present, but empty.

Then within each data source is an object of class occdatind holding another named list that contains:

meta: metadata
- source: the data source name (e.g., "gbif")
- time: time the request was sent
- found: number of records found (number found across all queries)
- returned: number of records returned (number of rows in all data.frame's in the data slot)
- type: query type, only "sci" for scientific
- opts: a named list with the options you sent to the data source
- errors: a character vector of errors returned, if any occurred
data: named list of data.frame's, named by the queries sent

Details

The occ function is an opinionated wrapper around the rgbif, rinat, rebird, rvertnet and ridigbio packages (as well as internal custom wrappers around some data sources) to allow data access from a single access point. We take care of making sure you get useful objects out at the cost of flexibility/options - although you can still set options for each of the packages via the gbifopts, inatopts, etc. parameters.

Inputs

All inputs to occ are one of:

scientific name
taxonomic id
geometry as bounds, WKT, os Spatial classes

To search by common name, first use occ_names() to find scientic names or taxonomic IDs, then feed those to this function. Or use the taxize package to get names and/or IDs to use here.

Using the query parameter

When you use the query parameter, we pass your search terms on to parameters within functions that query data sources you specify. Those parameters are:

rgbif - scientificName in the rgbif::occ_search() function - API parameter: same as the occ parameter
rebird - species in the rebird::ebirdregion() or rebird::ebirdgeo() functions, depending on whether you set method="ebirdregion" or method="ebirdgeo" - API parameters: sci for both rebird::ebirdregion() and rebird::ebirdgeo()
rvertnet - taxon in the rvertnet::vertsearch() function - API parameter: q
ridigbio - scientificname in the ridigbio::idig_search_records() function - API parameter: scientificname
inat - internal function - API parameter: q
obis - internal function - API parameter: scientificName
ala - internal function - API parameter: q

If you have questions about how each of those parameters behaves with respect to the terms you pass to it, lookup documentation for those functions, or get in touch at the development repository https://github.com/ropensci/spocc/issues

iDigBio notes

When searching iDigBio note that by deafult we set fields = "all", so that we return a richer suite of fields than the ridigbio R client gives by default. But you can changes this by passing in a fields parameter to idigbioopts parameter with the specific fields you want.

Maximum of 100,000 results are allowed to be returned. See https://github.com/iDigBio/ridigbio/issues/33

iNaturalist notes

We're using the iNaturalist API, docs at https://api.inaturalist.org/v1/docs/#!/Observations/get_observations

API rate limits: max of 100 requests per minute, though they ask that you try to keep it to 60 requests per minute or lower. If they notice usage that has serious impact on their performance they may institute blocks without notification.

There is a hard limit 0f 10,000 observations with the iNaturalist API. We do paging internally so you may not see this aspect, but for example, if you request 12,000 records, you won't be able to get that many. The API will error at anything more than 10,000. We now error if you request more than 10,000 from iNaturalist. There are some alternatives:

Consider exporting data while logged in to your iNaturalist account, or the iNaturalist research grade observations within GBIF - see https://www.gbif.org/dataset/50c9509d-22c7-4a22-a47d-8c48425ef4a7 - at time of this writing it has 8.5 million observations.
Search for iNaturalist data within GBIF. e.g., the following searches for iNaturalist data within GBIF and allows more than 10,000 records: “

limit parameter

The limit parameter is set to a default of 500. This means that you will get up to 500 results back for each data source you ask for data from. If there are no results for a particular source, you'll get zero back; if there are 8 results for a particular source, you'll get 8 back. If there are 501 results for a particular source, you'll get 500 back. You can always ask for more or less back by setting the limit parameter to any number. If you want to request a different number for each source, pass the appropriate parameter to each data source via the respective options parameter for each data source.

WKT

WKT objects are strings of pairs of lat/long coordinates that define a shape. Many classes of shapes are supported, including POLYGON, POINT, and MULTIPOLYGON. Within each defined shape define all vertices of the shape with a coordinate like 30.1 10.1, the first of which is the latitude, the second the longitude.

Examples of valid WKT objects:

'POLYGON((30.1 10.1, 10 20, 20 60, 60 60, 30.1 10.1))'
'POINT((30.1 10.1))'
'LINESTRING(3 4,10 50,20 25)'
'MULTIPOINT((3.5 5.6),(4.8 10.5))")'
'MULTILINESTRING((3 4,10 50,20 25),(-5 -8,-10 -8,-15 -4))'
'MULTIPOLYGON(((1 1,5 1,5 5,1 5,1 1),(2 2,2 3,3 3,3 2,2 2)),((6 3,9 2,9 4,6 3)))'
'GEOMETRYCOLLECTION(POINT(4 6),LINESTRING(4 6,7 10))'

Only POLYGON objects are currently supported.

Getting WKT polygons or bounding boxes. We will soon introduce a function to help you select a bounding box but for now, you can use a few sites on the web.

Bounding box - https://boundingbox.klokantech.com/
Well known text - http://arthur-e.github.io/Wicket/sandbox-gmaps3.html

geometry parameter

The behavior of the occ function with respect to the geometry parameter varies depending on the inputs to the query parameter. Here are the options:

geometry (single), no query - If a single bounding box/WKT string passed in, and no query, a single query is made against each data source.
geometry (many), no query - If many bounding boxes/WKT strings are passed in, we do a separate query for each bounding box/WKT string against each data source.
geometry (single), query - If a single bounding box/WKT string passed in, and a single query, we do a single query against each data source.
geometry (many), query - If many bounding boxes/WKT strings are passed in, and a single query, we do a separate query for each bounding box/WKT string with the same queried name against each data source.
geometry (single), many query - If a single bounding box/WKT string passed in, and many names to query, we do a separate query for each name, using the same geometry, for each data source.
geometry (many), many query - If many bounding boxes/WKT strings are passed in, and many names to query, this poses a problem for all data sources, none of which accept many bounding boxes of WKT strings. So, in this scenario, we loop over each name and each geometry query, and then re-combine by queried name, so that you get back a single group of data for each name.

Geometry options by data provider

wkt & bbox allowed, see WKT section above

gbif
obis
ala

bbox only

inat
idigbio

No spatial search allowed

ebird
vertnet

Notes on the date parameter

Date searches with the date parameter are allowed for all sources except ebird.

Notes on some special cases

idigbio: We search on the datecollected field. Other date fields can be searched on, but we chose datecollected as it seemed most appropriate.
vertnet: If you want more flexible date searches, you can pass various types of date searches to vertnetopts. See rvertnet::searchbyterm() for more information
ala: There's some issues with the dates returned from ALA. They are returned as time stamps, and some seem to be malformed. So do beware of using ALA dates for important things.

Get in touch if you have other date search use cases you think are widely useful

Paging

All data sources respond to the limit parameter passed to occ.

Data sources, however, vary as to whether they respond to an offset. Here's the details on which data sources will respond to start and which to the page parameter:

gbif - Responds to start. Default: 0
inat - Responds to page. Default: 1
ebird - No paging, both start and page ignored.
vertnet - No paging implemented here, both start and page ignored. VertNet does have a form of paging, but it uses a cursor, and can't easily be included here via parameters. However, rvertnet does paging internally for you. For example, the max records per request for VertNet is 1000; if you request 2000 records, we'll do the first request, and do the second request for you automatically.
idigbio - Responds to start. Default: 0
obis - Does not respond to start. They only allow a starting occurrence UUID up to which to skip. So order of results matters a great deal of course. To paginate with OBIS, do e.g. obisopts = list(after = "017b7818-5b2c-4c88-9d76-f4471afe5584"); after can be combined with the limit value you pass in to the main occ() function call. See obis_search for what parameters can be used.
ala - Responds to start. Default: 0

Photographs

The iNaturalist data source provides photographs of the records returned, if available. For example, the following will give photos from inat: occ(query = 'Danaus plexippus', from = 'inat')$inat$data$Danaus_plexippus$photos

BEWARE

In cases where you request data from multiple providers, especially when including GBIF, there could be duplicate records since many providers' data eventually ends up with GBIF. See spocc_duplicates() for more.

Examples

if (FALSE) { # \dontrun{
# Single data sources
(res <- occ(query = 'Accipiter striatus', from = 'gbif', limit = 5))
res$gbif
(res <- occ(query = 'Accipiter striatus', from = 'ebird', limit = 50))
res$ebird
(res <- occ(query = 'Danaus plexippus', from = 'inat', limit = 50,
  has_coords = TRUE))
res$inat
res$inat$data
data.table::rbindlist(res$inat$data$Danaus_plexippus$photos)
(res <- occ(query = 'Bison bison', from = 'vertnet', limit = 5))
res$vertnet
res$vertnet$data$Bison_bison
occ2df(res)

# Paging
one <- occ(query = 'Accipiter striatus', from = 'gbif', limit = 5)
two <- occ(query = 'Accipiter striatus', from = 'gbif', limit = 5, start = 5)
one$gbif
two$gbif

# iNaturalist limits: they allow at most 10,000; query through GBIF to get
# more than 10,000
# See https://www.gbif.org/dataset/50c9509d-22c7-4a22-a47d-8c48425ef4a7
# x <- occ(query = 'Danaus plexippus', from = 'gbif', limit = 10100, 
#   gbifopts = list(datasetKey = "50c9509d-22c7-4a22-a47d-8c48425ef4a7"))
# x$gbif

# Date range searches across data sources
## Not possible for ebird
## ala
occ(date = c('2018-01-01T00:00:00Z', '2018-03-28T00:00:00Z'), from = 'ala', limit = 5)
## gbif
occ(query = 'Accipiter striatus', date = c('2010-08-01', '2010-08-31'), from = 'gbif', limit=5)
## vertnet
occ(query = 'Mustela nigripes', date = c('1990-01-01', '2015-12-31'), from = 'vertnet', limit=5)
## idigbio
occ(query = 'Acer', date = c('2010-01-01', '2015-12-31'), from = 'idigbio', limit=5)
## obis
occ(query = 'Mola mola', date = c('2015-01-01', '2015-12-31'), from = 'obis', limit=5)
## inat
occ(query = 'Danaus plexippus', date = c('2015-01-01', '2015-12-31'), from = 'inat', limit=5)


# Restrict to records with coordinates
occ(query = "Acer", from = "idigbio", limit = 5, has_coords = TRUE)

occ(query = 'Setophaga caerulescens', from = 'ebird', ebirdopts = list(loc='US'))
occ(query = 'Spinus tristis', from = 'ebird', ebirdopts =
   list(method = 'ebirdgeo', lat = 42, lng = -76, dist = 50))

# idigbio data
## scientific name search
occ(query = "Acer", from = "idigbio", limit = 5)
occ(query = "Acer", from = "idigbio", idigbioopts = list(offset = 5, limit  = 3))
## geo search
bounds <- c(-120, 40, -100, 45)
occ(from = "idigbio", geometry = bounds, limit = 10)
## just class arachnida, spiders
occ(idigbioopts = list(rq = list(class = 'arachnida')), from = "idigbio", limit = 10)
## search certain recordsets
sets <- c("1ffce054-8e3e-4209-9ff4-c26fa6c24c2f",
    "8dc14464-57b3-423e-8cb0-950ab8f36b6f", 
    "26f7cbde-fbcb-4500-80a9-a99daa0ead9d")
occ(idigbioopts = list(rq = list(recordset = sets)), from = "idigbio", limit = 10)

# Many data sources
(out <- occ(query = 'Pinus contorta', from=c('gbif','vertnet'), limit=10))

## Select individual elements
out$gbif
out$gbif$data
out$vertnet

## Coerce to combined data.frame, selects minimal set of
## columns (name, lat, long, provider, date, occurrence key)
occ2df(out)

# Pass in limit parameter to all sources. This limits the number of occurrences
# returned to 10, in this example, for all sources, in this case gbif and inat.
occ(query='Pinus contorta', from=c('gbif','inat'), limit=10)

# Geometry
## Pass in geometry parameter to all sources. This constraints the search to the
## specified polygon for all sources, gbif in this example.
## Check out http://arthur-e.github.io/Wicket/sandbox-gmaps3.html to get a WKT string
occ(query='Accipiter', from='gbif',
   geometry='POLYGON((30.1 10.1, 10 20, 20 60, 60 60, 30.1 10.1))')

## Or pass in a bounding box, which is automatically converted to WKT (required by GBIF)
## via the bbox2wkt function. The format of a bounding box is
## [min-longitude, min-latitude, max-longitude, max-latitude].
occ(query='Accipiter striatus', from='gbif', geometry=c(-125.0,38.4,-121.8,40.9))

## lots of results, can see how many by indexing to meta
res <- occ(query='Accipiter striatus', from='gbif',
   geometry='POLYGON((-69.9 49.2,-69.9 29.0,-123.3 29.0,-123.3 49.2,-69.9 49.2))')
res$gbif

## You can pass in geometry to each source separately via their opts parameter, at
## least those that support it. Note that if you use rinat, you reverse the order, with
## latitude first, and longitude second, but here it's the reverse for consistency across
## the spocc package
bounds <- c(-125.0,38.4,-121.8,40.9)
occ(query = 'Danaus plexippus', from="inat", geometry=bounds)

## Passing geometry with multiple sources
occ(query = 'Danaus plexippus', from=c("inat","gbif"), geometry=bounds)

## Using geometry only for the query
### A single bounding box
occ(geometry = bounds, from = "gbif", limit=50)
### Many bounding boxes
occ(geometry = list(c(-125.0,38.4,-121.8,40.9), c(-115.0,22.4,-111.8,30.9)), from = "gbif")

## Geometry only with WKT
wkt <- 'POLYGON((-98.9 44.2,-89.1 36.6,-116.7 37.5,-102.5 39.6,-98.9 44.2))'
occ(from = "gbif", geometry = wkt, limit = 10)

# Specify many data sources, another example
ebirdopts = list(loc = 'US'); gbifopts  =  list(country = 'US')
out <- occ(query = 'Setophaga caerulescens', from = c('gbif','inat','ebird'),
    gbifopts = gbifopts, ebirdopts = ebirdopts, limit=20)
occ2df(out)

# Pass in many species names, combine just data to a single data.frame, and
# first six rows
spnames <- c('Accipiter striatus', 'Setophaga caerulescens', 'Spinus tristis')
(out <- occ(query = spnames, from = 'gbif', gbifopts = list(hasCoordinate = TRUE), limit=25))
df <- occ2df(out)
head(df)

# no query, geometry, or ids passed
## many dataset keys to gbif
dsets <- c("14f3151a-e95d-493c-a40d-d9938ef62954", "f934f8e2-32ca-46a7-b2f8-b032a4740454")
occ(limit = 20, from = "gbif", gbifopts = list(datasetKey = dsets))
## class name to idigbio
occ(limit = 20, from = "idigbio", idigbioopts = list(rq = list(class = 'arachnida')))

# taxize integration
## You can pass in taxonomic identifiers
library("taxize")
(ids <- get_ids(c("Chironomus riparius","Pinus contorta"), db = c('itis','gbif')))
occ(ids = ids, from='gbif', limit=20)

(ids <- get_ids("Chironomus riparius", db = 'gbif'))
occ(ids = ids, from='gbif', limit=20)

(ids <- get_gbifid("Chironomus riparius"))
occ(ids = ids, from='gbif', limit=20)

## sf classes
library("sp")
library("sf")
one <- Polygon(cbind(c(91,90,90,91), c(30,30,32,30)))
spone = Polygons(list(one), "s1")
sppoly = SpatialPolygons(list(spone), as.integer(1))

## single polygon in a sf class
x <- st_as_sf(sppoly)
out <- occ(geometry = x, limit=50)
out$gbif$data
mapr::map_leaflet(out)

## single polygon in a sfc class
x <- st_as_sf(sppoly)
out <- occ(geometry = x[[1]], limit=50)
out$gbif$data

## single polygon in a sf POLYGON class
x <- st_as_sf(sppoly)
x <- unclass(x[[1]])[[1]]
class(x)
out <- occ(geometry = x, limit=50)
out$gbif$data

## two polygons in an sf class
one <- Polygon(cbind(c(-121.0,-117.9,-121.0,-121.0), c(39.4, 37.1, 35.1, 39.4)))
two <- Polygon(cbind(c(-123.0,-121.2,-122.3,-124.5,-123.5,-124.1,-123.0),
                     c(44.8,42.9,41.9,42.6,43.3,44.3,44.8)))
spone = Polygons(list(one), "s1")
sptwo = Polygons(list(two), "s2")
sppoly = SpatialPolygons(list(spone, sptwo), 1:2)
sppoly_df <- SpatialPolygonsDataFrame(sppoly, 
   data.frame(a=c(1,2), b=c("a","b"), c=c(TRUE,FALSE),
   row.names=row.names(sppoly)))
x <- st_as_sf(sppoly_df)
out <- occ(geometry = x, limit=50)
out$gbif$data


# curl debugging
occ(query = 'Accipiter striatus', from = 'gbif', limit=10, 
 callopts=list(verbose = TRUE))
occ(query = 'Accipiter striatus', from = 'inat', 
 callopts=list(verbose = TRUE))
occ(query = 'Mola mola', from = 'obis', limit = 200, 
 callopts = list(verbose = TRUE))

########## More thorough data source specific examples
# idigbio
## scientific name search
res <- occ(query = "Acer", from = "idigbio", limit = 5)
res$idigbio

## geo search
### bounding box
bounds <- c(-120, 40, -100, 45)
occ(from = "idigbio", geometry = bounds, limit = 10)
### wkt
# wkt <- 'POLYGON((-69.9 49.2,-69.9 29.0,-123.3 29.0,-123.3 49.2,-69.9 49.2))'
wkt <- 'POLYGON((-98.9 44.2,-89.1 36.6,-116.7 37.5,-102.5 39.6,-98.9 44.2))'
occ(from = "idigbio", geometry = wkt, limit = 10)

## limit fields returned
occ(query = "Acer", from = "idigbio", limit = 5,
   idigbioopts = list(fields = "scientificname"))

## offset and max_items
occ(query = "Acer", from = "idigbio", limit = 5,
   idigbioopts = list(offset = 10))

## sort
occ(query = "Acer", from = "idigbio", limit = 5,
   idigbioopts = list(sort = TRUE))$idigbio
occ(query = "Acer", from = "idigbio", limit = 5,
   idigbioopts = list(sort = FALSE))$idigbio

## more complex queries
### parameters passed to "rq", get combined with the name queried
occ(query = "Acer", from = "idigbio", limit = 5,
   idigbioopts = list(rq = list(basisofrecord="fossilspecimen")))$idigbio

#### NOTE: no support for multipolygons yet
## WKT's are more flexible than bounding box's. You can pass in a WKT with multiple
## polygons like so (you can use POLYGON or MULTIPOLYGON) when specifying more than one
## polygon. Note how each polygon is in it's own set of parentheses.
# occ(query='Accipiter striatus', from='gbif',
#    geometry='MULTIPOLYGON((30 10, 10 20, 20 60, 60 60, 30 10),
#                           (30 10, 10 20, 20 60, 60 60, 30 10))')

# OBIS examples
## basic query
(res <- occ(query = 'Mola mola', from = 'obis', limit = 200))
## get to obis data
res$obis
## get obis + gbif data
(res <- occ(query = 'Mola mola', from = c('obis', 'gbif'), limit = 200))
res$gbif
res$obis
## no match found
(res <- occ(query = 'Linguimaera thomsonia', from = 'obis'))
## geometry query
geometry <- "POLYGON((8.98 48.05,15.66 48.05,15.66 45.40,8.98 45.40,8.98 48.05))"
(res <- occ(from = 'obis', geometry = geometry, limit = 50))
res$obis

## Pass in spatial classes
## sp classes no longer supported

## Paging
(res1 <- occ(query = 'Mola mola', from = 'obis', limit = 10))
occ_ids <- res1$obis$data$Mola_mola$id
(res2 <- occ(query = 'Mola mola', from = 'obis',
  limit = 10, obisopts = list(after = occ_ids[length(occ_ids)])))
res1$obis
res2$obis
## Pass in any parameters to obisopts as a list
(res <- occ(query = 'Mola mola', from = 'obis', 
   obisopts = list(startdepth = 40, enddepth = 50)))
min(res$obis$data$Mola_mola$minimumDepthInMeters, na.rm=TRUE)
max(res$obis$data$Mola_mola$maximumDepthInMeters, na.rm=TRUE)


# ALA examples
## basic query
(res <- occ(query = 'Alaba vibex', from = 'ala', limit = 200))
## get to ala data
res$ala
occ2df(res)

# geometry search
(x <- occ(query = "Macropus", from = 'ala',
  geometry = "POLYGON((145 -37,150 -37,150 -30,145 -30,145 -37))"))
x$ala
occ2df(x)
} # }