ERSST

ersst

Provides for download, archiving and access to Extended Reconstructed Sea Surface Temperature ‘ersst’ online and local datasets.

Citation

Boyin Huang, Peter W. Thorne, Viva F. Banzon, Tim Boyer, Gennady Chepurin, Jay H. Lawrimore, Matthew J. Menne, Thomas M. Smith, Russell S. Vose, and Huai-Min Zhang (2017): NOAA Extended Reconstructed Sea Surface Temperature (ERSST), Version 5. NOAA National Centers for Environmental Information. doi:10.7289/V5T72FNM [access date, monthly from 2021-04-15].

Requirements

From CRAN…

• R v4+
• rlang
• dplyr
• readr
• terra
• xml2
• httr
• rvest
• stars
• PCICt

Installation

remotes::install_github("BigelowLab/ersst")

Usage

suppressPackageStartupMessages({
  library(dplyr)
  library(sf)
  library(ersst)
  library(stars)
})

Working with points.

We provide a small example of points from locations associated with buoys off of the US southeast coast.

# read in example South Atlantic Bight points
x <- read_sab()

# generate a ersst url for a given date
url <- ersst_url("2018-12-18")

# download
temp_file <- tempfile(fileext = ".nc")
ok <- download.file(url, temp_file)

# open the resource
X <- ncdf4::nc_open(temp_file)

# extract the data 
covars <- ersst::extract(x, X, varname = ersst_vars(X))

# bind to the input
(y <- dplyr::bind_cols(x, covars))

## Simple feature collection with 5 features and 6 fields
## Geometry type: POINT
## Dimension:     XY
## Bounding box:  xmin: 279.815 ymin: 28.508 xmax: 287.752 ymax: 34.714
## Geodetic CRS:  WGS 84
## # A tibble: 5 × 7
##      id name     depth time                            geom   sst  ssta
## * <dbl> <chr>    <dbl> <dttm>                   <POINT [°]> <dbl> <dbl>
## 1 41009 Canveral  -1.5 2022-06-17 14:05:06 (279.815 28.508)  24.9 0.627
## 2 41010 Canaver…  -1.5 2022-06-29 14:05:06 (281.515 28.878)  25.3 0.717
## 3 41002 South H…  -1.5 2022-06-05 14:05:06 (285.064 31.759)  23.2 0.745
## 4 41001 East Ha…  -2   2022-04-30 14:05:06 (287.752 34.714)  22.2 0.704
## 5 41004 EDISTO    -1.5 2022-07-11 14:05:06 (280.901 32.502)  23.4 0.683

Working with bounding boxes (from points or polygons).

Learn more about working with stars objects in the vignettes. Here we drop time and depth as the ERSST data has only one depth and each file holds one time.

# derive the bounding box of the points and convert to sf coordinates
bb = sf::st_bbox(x) |>
  sf::st_as_sfc()

(covars <- ersst::extract(bb, X, varnames = ersst::ersst_vars(X)))

## stars object with 2 dimensions and 1 attribute
## attribute(s):
##          Min.  1st Qu.   Median    Mean  3rd Qu.     Max. NA's
## sst  16.22364 22.10888 23.58391 22.8994 24.44269 25.34757    7
## dimension(s):
##   from to offset  delta refsys x/y
## x    1  6  279.8  1.323 WGS 84 [x]
## y    1  5  34.71 -1.241 WGS 84 [y]

Now let’s see what it looks like on a quick map.

plot(covars, attr = 'sst', axes = TRUE, reset = FALSE)
plot(sf::st_geometry(x), add = TRUE, col = "orange", 
     pch = 19, cex = 2)

# cleanup
ncdf4::nc_close(X)

Local database

If you desire to build a local database of the ERSST dataset, you can use the update_ersst.R script. First you need to set the path for the local database using something like this.

ersst::set_root_path("/mnt/s1/projects/ecocast/coredata/ersst")

Obviously, you would substitute in your own path. Thereafter for any R session you can retrieve that path like this. Note that we need to specify the version we are interested in. At the time of this compostion, we only know about v5 but we are totally ready for v6.

path = ersst_path("v5")

Now you can run the update_ersst.R script to build your database. See the script docs for details. Once you have the database you can read it in.

db = read_database(path) |>
  print()

## # A tibble: 4,118 × 3
##    date       anomaly version
##    <date>     <lgl>   <chr>  
##  1 1854-01-01 FALSE   v5     
##  2 1854-02-01 FALSE   v5     
##  3 1854-03-01 FALSE   v5     
##  4 1854-04-01 FALSE   v5     
##  5 1854-05-01 FALSE   v5     
##  6 1854-06-01 FALSE   v5     
##  7 1854-07-01 FALSE   v5     
##  8 1854-08-01 FALSE   v5     
##  9 1854-09-01 FALSE   v5     
## 10 1854-10-01 FALSE   v5     
## # ℹ 4,108 more rows

This can be filtered for a particular item (or items!)

this_db = filter(db,
                 date %in% as.Date(c("1995-12-01", "1996-01-01", "1996-02-01")))
this_db

## # A tibble: 6 × 3
##   date       anomaly version
##   <date>     <lgl>   <chr>  
## 1 1995-12-01 FALSE   v5     
## 2 1995-12-01 TRUE    v5     
## 3 1996-01-01 FALSE   v5     
## 4 1996-02-01 FALSE   v5     
## 5 1996-01-01 TRUE    v5     
## 6 1996-02-01 TRUE    v5

Now read the data.

x = read_ersst(this_db, path)
x

## stars object with 3 dimensions and 2 attributes
## attribute(s):
##                Min.    1st Qu.      Median        Mean    3rd Qu.      Max.
## sst       -1.800000  1.3255093 14.88280201 13.63851193 25.2284541 30.781897
## sst.anom  -4.763667 -0.3596543 -0.01972961 -0.06662737  0.2058811  2.317597
##            NA's
## sst       15096
## sst.anom  15096
## dimension(s):
##      from  to     offset   delta refsys point x/y
## x       1 180         -1       2 WGS 84 FALSE [x]
## y       1  89         89      -2 WGS 84 FALSE [y]
## time    1   3 1995-12-01 31 days   Date    NA