Finding NOAA stations

Introduction

Meteorological inputs—temperature, precipitation, and barometric pressure—are essential for computing dissolved oxygen saturation and gas exchange in stream metabolism models. This vignette shows how to find NOAA daily-summary stations near Kings Creek at Konza Prairie Biological Station, Kansas, and download the weather data needed to accompany the built-in kings_discharge dataset (water year 2025: 2024-10-01 through 2025-09-30).

preMetabolizer provides three helpers for station discovery:

get_noaa_stations() searches for GHCND stations via the NCEI Search Service API.
closest_noaa_stations() finds stations within a distance of a target coordinate and ranks them by geodesic distance.
ncei_stations() is the lower-level function underlying both helpers; it can search any NCEI dataset.

Note: All functions in this vignette contact the NCEI API. Set PREMETABOLIZER_RUN_VIGNETTES=true to run the API chunks; they are skipped by default during package installation and checks.

library(preMetabolizer)
library(dplyr)
library(ggplot2)

Study site

Kings Creek drains the Konza Prairie Biological Station near Manhattan, Kansas (39.1069°N, 96.6117°W). The USGS monitoring location USGS-06879650 records daily discharge, gage height, and water temperature throughout water year 2025.

lat_kings  <- 39.1068806
lon_kings  <- -96.6117151
wy_start   <- "2024-10-01"
wy_end     <- "2025-09-30"

Search for stations

Use ncei_bbox() to build a bounding box from a center point and radius, then pass it to get_noaa_stations():

bbox <- ncei_bbox(latitude = lat_kings, longitude = lon_kings, dist_km = 100)
bbox

ks_stations <- get_noaa_stations(bbox = bbox)

glimpse(ks_stations)

Filter to stations that carry the variables you need and span at least part of water year 2025:

wx_stations <- get_noaa_stations(
  bbox       = bbox,
  data_types = c("TMAX", "TMIN", "PRCP"),
  start_date = wy_start,
  end_date   = wy_end
)

wx_stations |>
  select(station_id, station_name, latitude, longitude, start_date, end_date) |>
  arrange(station_name)

Find nearby stations

closest_noaa_stations() builds the bounding box automatically and returns stations sorted by geodesic distance from the target point:

konza_noaa <- closest_noaa_stations(
  latitude   = lat_kings,
  longitude  = lon_kings,
  dist_km    = 50,
  data_types = c("TMAX", "TMIN", "PRCP"),
  start_date = wy_start,
  end_date   = wy_end
)

konza_noaa |>
  select(distance_km, station_id, station_name, latitude, longitude)

Map the candidates before choosing a station:

ggplot(konza_noaa, aes(longitude, latitude)) +
  geom_point(aes(size = distance_km), color = "#2c7fb8") +
  annotate("point", x = lon_kings, y = lat_kings, color = "#d95f0e", size = 3) +
  coord_quickmap() +
  labs(
    x     = "Longitude",
    y     = "Latitude",
    size  = "Distance (km)",
    title = "NOAA stations near Kings Creek"
  ) +
  theme_bw()

Download daily weather data

Pick the nearest station and download daily temperature and precipitation for water year 2025 with ncei_data():

station_id <- konza_noaa |>
  arrange(distance_km) |>
  pull(station_id) |>
  first()

station_id

daily_wx <- ncei_data(
  dataset    = "daily-summaries",
  stations   = station_id,
  start_date = wy_start,
  end_date   = wy_end,
  data_types = c("TMAX", "TMIN", "PRCP")
)

glimpse(daily_wx)

With units = "metric" (the default), ncei_data() returns tmax and tmin in °C and prcp in mm. The date column is already a Date:

ggplot(daily_wx, aes(date)) +
  geom_ribbon(aes(ymin = tmin, ymax = tmax), alpha = 0.3, fill = "#2c7fb8") +
  geom_line(aes(y = (tmax + tmin) / 2), color = "#2c7fb8") +
  labs(
    x     = NULL,
    y     = "Air temperature (°C)",
    title = "Daily temperature range near Kings Creek, WY 2025"
  ) +
  theme_bw()

Barometric pressure and PAR

GHCND covers temperature and precipitation well, but barometric pressure records are sparse in this region. For pressure and photosynthetically active radiation (PAR), use get_nasa_data() to pull modeled values from NASA POWER, or get_ghcnh() to download observed hourly pressure from the GHCNh archive.

stream_data <- tibble(
  dateTime = seq(
    as.POSIXct(paste(wy_start, "00:00:00"), tz = "UTC"),
    as.POSIXct(paste(wy_end, "23:00:00"), tz = "UTC"),
    by = "1 hour"
  )
)

nasa_wx <- get_nasa_data(
  data      = stream_data,
  latitude  = lat_kings,
  longitude = lon_kings,
  elev_m    = 320
)

glimpse(nasa_wx)

The PSC column contains elevation-corrected barometric pressure (kPa), light.obs contains PAR (µmol/m²/s), T2M air temperature (°C), and PRECTOTCORR precipitation (mm/hr).

Use station IDs with GHCNh data

The station_id values returned by closest_noaa_stations() also identify GHCNh files. Pass them directly to get_ghcnh():

station_id

See vignette("ghcnh", package = "preMetabolizer") for a full hourly-data workflow.