Get GHCNh hourly observations

Downloads Global Historical Climatology Network hourly (GHCNh) data for one or more stations and a date range, returning a parsed data frame ready for use in stream metabolism modelling.

Usage

get_ghcnh(stations, start_date, end_date, quiet = FALSE)

Arguments

stations

Character vector of GHCNh station identifiers (e.g., "USW00023183"). These match the station_id column returned by get_noaa_stations() and closest_noaa_stations().

start_date, end_date

Start and end of the requested period as Date objects or "YYYY-MM-DD" strings.

quiet

Logical. When TRUE progress messages are suppressed. Default FALSE.

Value

A tibble with one row per station-hour. Key columns include:

station_id

Station identifier.

station_name

Station name (when available).

datetime

Observation time as a UTC POSIXct.

temperature

Air temperature in degrees Celsius.

dew_point_temperature

Dew-point temperature (°C).

relative_humidity

Relative humidity (%).

wind_direction

Wind direction (degrees clockwise from north).

wind_speed

Wind speed (m/s).

sea_level_pressure

Sea-level pressure (hPa).

precipitation

Precipitation (mm).

Columns that are entirely NA across all stations and hours are dropped from the result.

Details

GHCNh v1.1.0 data are served as pipe-separated (PSV) files from the NOAA NCEI archive, one file per station per year. Data are current through the present year. The function downloads all station-year combinations covering the requested date range in parallel.

Sentinel values of -9999 are converted to NA. Columns that are all NA are removed.

Multiple stations and years are downloaded in parallel using up to four concurrent connections.

See also

get_noaa_stations(), closest_noaa_stations(), ncei_data()

Examples

if (FALSE) { # \dontrun{
# Fetch WY 2025 hourly data for a station near Konza Prairie
konza <- closest_noaa_stations(
  latitude = 39.1068806,
  longitude = -96.6117151,
  dist_km = 50
)
hourly <- get_ghcnh(
  stations = konza$station_id[1],
  start_date = "2024-10-01",
  end_date = "2025-09-30"
)
} # }