Skip to contents

Extract ERA5-Land climate variables from Google Earth Engine

Source: R/l4h_era5land.R
l4h_era5land.Rd

Extracts ERA5-Land climate variables for a user-defined region and time range from the Earth Engine datasets ECMWF/ERA5_LAND/DAILY_AGGR (by = "daily") or ECMWF/ERA5_LAND/MONTHLY_AGGR (by = "month"), plus ECMWF/ERA5_LAND/HOURLY for the diurnal bands "rh_day"/"rh_night". Each image is summarized over the region using a chosen statistic (e.g., mean/median), values are converted to conventional units (degrees Celsius, mm, volume fraction), and the function returns an sf or tibble.

ERA5-Land is the land-component replay of the ECMWF ERA5 climate reanalysis (Copernicus Climate Data Store), with global coverage from 1950 to near-present at ~9 km (0.1 degrees) resolution. It complements l4h_chirps() (finer precipitation) and l4h_terra_climate() (finer monthly climatology) by providing daily exposure windows and soil-moisture variables that are key for vector-borne disease modelling (temperature/precipitation lags).

Usage

l4h_era5land(
  from,
  to,
  by = "month",
  band = "t2m",
  region,
  scale = 11132,
  stat = "mean",
  sf = TRUE,
  quiet = FALSE,
  force = FALSE,
  ...
)

Arguments

from

Character or Date. Start date ("YYYY-MM-DD").

to

Character or Date. End date ("YYYY-MM-DD").

by

Character. Temporal resolution. Options:

  • "daily" — daily values from ECMWF/ERA5_LAND/DAILY_AGGR,

  • "month" — monthly values from ECMWF/ERA5_LAND/MONTHLY_AGGR. Default "month". Diurnal bands ("rh_day", "rh_night") always use the hourly collection (see Details), aggregated to the same periods.

band

Character vector. One or more ERA5-Land variables to extract. Supported codes:

  • "t2m" (2m air temperature, K -> degrees C),

  • "d2m" (2m dewpoint temperature, K -> degrees C),

  • "rh" (mean relative humidity, % — derived in Earth Engine from temperature_2m and dewpoint_temperature_2m via the Magnus formula, see Details),

  • "rh_day" (daytime relative humidity, % — mean over 06:00-18:00 Peru time from ECMWF/ERA5_LAND/HOURLY, see Details),

  • "rh_night" (nighttime relative humidity, % — mean over 18:00-06:00 Peru time from ECMWF/ERA5_LAND/HOURLY, see Details),

  • "pr" (total precipitation, m -> mm),

  • "soil" (volumetric soil water layer 1, 0-7 cm, m3/m3),

  • "pet" (potential evaporation, m -> mm). Default "t2m".

region

Spatial object defining the region of interest. Accepts an sf, sfc, or SpatVector object.

scale

Numeric. Reducer scale in meters. Default 11132 (ERA5-Land native pixel ~11,132 m).

stat

Character. Summary statistic per image per region. One of "mean", "median", "min", "max".

sf

Logical. If TRUE, returns an sf; if FALSE, returns a tibble. Default TRUE.

quiet

Logical. If TRUE, suppresses progress bars/messages. Default FALSE.

force

Logical. If TRUE, skips the representativity check. Default FALSE.

...

Additional arguments passed to the extraction backend.

Value

An sf or tibble with columns:

  • date (Date — first day of the period),

  • variable (character — band code, e.g. "t2m", "rh"),

  • value (numeric — degrees C, percent, mm, or volume fraction), plus geometry if sf = TRUE, and any attributes from region.

Details

Relative humidity (band = "rh") is derived server-side in Earth Engine from the bands the function already uses (temperature_2m and dewpoint_temperature_2m), following the Magnus formula for saturation vapour pressure (Wright 1997 coefficients, same as the EU operational product derived from ERA5-Land):

$$e(T) = 611.21 \times \exp\left(\frac{17.502 \times T}{240.97 + T}\right)$$ $$RH = 100 \times \frac{e(Td)}{e(T)}$$

with T (air temperature) and Td (dewpoint) in degrees C, converted from Kelvin inside Earth Engine (- 273.15). The ratio simplifies to 100 * exp(b*c*(Td - T) / ((c + T) * (c + Td))) with b = 17.502 and c = 240.97. Values are clamped to 0–100%.

Note the aggregation order: DAILY_AGGR/MONTHLY_AGGR already store time-averaged T/Td, so rh is the humidity of the averaged temperatures, not the average of instantaneous humidity. This is the standard practice (same as the EU ERA5-Land product). Below 0 °C the water-saturation form is kept (no ice correction), which is acceptable for vector-borne disease modelling but documented here as a limitation.

Day and night are defined in Peru time (PET, UTC-5, no daylight saving): day is 06:00-18:00 PET and night is 18:00-06:00 PET. Since ECMWF/ERA5_LAND/HOURLY timestamps are UTC, "rh_day" averages the UTC hours 11-22 and "rh_night" the UTC hours 23, 00-10. Each hourly image is converted to RH with the same Magnus formula above and then averaged (mean of hourly means with equal weights). Diurnal bands therefore cannot be combined with "rh" or native bands in one call: they come from different collections with different time bases.

Credits

Pioneering geospatial health analytics and open-science tools. Developed by the Innovalab Team. For more information, send an email to imt.innovlab@oficinas-upch.pe.

Follow us on:

References

Muñoz Sabater, J. (2019): ERA5-Land monthly averaged data from 1950 to present. Copernicus Climate Change Service (C3S) Climate Data Store (CDS). doi:10.24381/cds.68d2bb30

Bolton, D. (1980). The computation of equivalent potential temperature. Monthly Weather Review, 108, 1046-1053.

Alduchov, O. A., & Eskridge, R. E. (1996). Improved Magnus form approximation of saturation vapor pressure. Journal of Applied Meteorology, 35, 601-609.

Examples

# \donttest{
if (interactive()) {
library(land4health)
rgee::ee_Initialize()

# ROI simple (EPSG:4326)
region <- st_as_sf(st_sfc(
  st_polygon(list(matrix(c(
    -74.1, -4.4,
    -74.1, -3.7,
    -73.2, -3.7,
    -73.2, -4.4,
    -74.1, -4.4
  ), ncol = 2, byrow = TRUE))), crs = 4326))

# 1. Monthly 2m temperature (degrees C) 2020
out_monthly <- l4h_era5land(
  from   = "2020-01-01",
  to     = "2020-12-31",
  by     = "month",
  band   = "t2m",
  region = region,
  stat   = "mean"
)
head(out_monthly)

# 2. Daily relative humidity (%) + precipitation (mm), one month
out_daily <- l4h_era5land(
  from   = "2020-06-01",
  to     = "2020-06-30",
  by     = "daily",
  band   = c("rh", "pr"),
  region = region,
  stat   = "mean"
)
head(out_daily)

# 3. Daytime vs nighttime relative humidity (%), one week (Peru time)
out_diurnal <- l4h_era5land(
  from   = "2020-06-01",
  to     = "2020-06-07",
  by     = "daily",
  band   = c("rh_day", "rh_night"),
  region = region,
  stat   = "mean"
)
head(out_diurnal)
}
# }