Please note: Starting with release v0.2.0, the New CDS platform is now supported.
bmi_era5 package is an implementation of the Basic Model Interface (BMI) for the ERA5 dataset. This package uses the CDS API to download the ERA5 dataset and wraps the dataset with BMI for data control and query. It currently supports 3-dimensional ERA5 datasets defined with dimensions such as valid_time (or date), latitude, and longitude.
This package is not implemented for people to use but is the key element to convert the ERA5 dataset into a data component (pymt_era5) for the PyMT modeling framework developed by Community Surface Dynamics Modeling System (CSDMS).
If you have any suggestion to improve the current function, please create a GitHub issue here.
The bmi_era5 package and its dependencies can be installed with pip
$ pip install bmi_era5
or conda
$ conda install -c conda-forge bmi_era5
After downloading the source code, run the following command from top-level folder to install bmi_era5.
$ pip install -e .
Below shows how to use two methods to download the ERA5 datasets.
You can learn more details from the tutorial notebook. To run this notebook, please go to the CSDMS EKT Lab and follow the instruction in the "Lab notes" section.
import cdsapi
import xarray
import matplotlib.pyplot as plt
c = cdsapi.Client()
c.retrieve(
"reanalysis-era5-single-levels",
{
"product_type": "reanalysis",
"format": "netcdf",
"variable": ["2m_temperature", "total_precipitation"],
"year": "2021",
"month": "01",
"day": "01",
"time": ["00:00", "01:00", "02:00"],
"area": [41, -109, 36, -102],
"grid": [0.25, 0.25],
},
"download.nc",
)
# load netCDF data
dataset = xarray.open_dataset("download.nc")
# select 2 meter temperature on 2021-01-01 at 00:00
air_temp = dataset.t2m.isel(valid_time=0)
# plot data
air_temp.plot(figsize=(9, 5))
plt.title("2 metre temperature in Colorado on Jan 1st, 2021 at 00:00")
from bmi_era5 import BmiEra5
import numpy as np
import matplotlib.pyplot as plt
data_comp = BmiEra5()
data_comp.initialize("config_file.yaml")
# get variable info
for var_name in data_comp.get_output_var_names():
var_unit = data_comp.get_var_units(var_name)
var_location = data_comp.get_var_location(var_name)
var_type = data_comp.get_var_type(var_name)
var_grid = data_comp.get_var_grid(var_name)
var_itemsize = data_comp.get_var_itemsize(var_name)
var_nbytes = data_comp.get_var_nbytes(var_name)
print(f"{var_name=}")
print(f"{var_unit=}")
print(f"{var_location=}")
print(f"{var_type=}")
print(f"{var_grid=}")
print(f"{var_itemsize=}")
print(f"{var_nbytes=}")
# get time info
start_time = data_comp.get_start_time()
end_time = data_comp.get_end_time()
time_step = data_comp.get_time_step()
time_unit = data_comp.get_time_units()
time_steps = int((end_time - start_time) / time_step) + 1
print(f"{start_time=}")
print(f"{end_time=}")
print(f"{time_step=}")
print(f"{time_unit=}")
print(f"{time_steps=}")
# get variable grid info
grid_rank = data_comp.get_grid_rank(var_grid)
grid_size = data_comp.get_grid_size(var_grid)
grid_shape = np.empty(grid_rank, int)
data_comp.get_grid_shape(var_grid, grid_shape)
grid_spacing = np.empty(grid_rank)
data_comp.get_grid_spacing(var_grid, grid_spacing)
grid_origin = np.empty(grid_rank)
data_comp.get_grid_origin(var_grid, grid_origin)
print(f"{grid_rank=}")
print(f"{grid_size=}")
print(f"{grid_shape=}")
print(f"{grid_spacing=}")
print(f"{grid_origin=}")
# get variable data
data = np.empty(grid_size, var_type)
data_comp.get_value("2 metre temperature", data)
data_2D = data.reshape(grid_shape)
# get X, Y extent for plot
min_y, min_x = grid_origin
max_y = min_y + grid_spacing[0] * (grid_shape[0] - 1)
max_x = min_x + grid_spacing[1] * (grid_shape[1] - 1)
dy = grid_spacing[0] / 2
dx = grid_spacing[1] / 2
extent = [min_x - dx, max_x + dx, min_y - dy, max_y + dy]
# plot data
fig, ax = plt.subplots(1, 1, figsize=(9, 5))
im = ax.imshow(data_2D, extent=extent)
cbar = fig.colorbar(im)
cbar.set_label("2 metre temperature [K]")
plt.xlabel("longitude [degree_east]")
plt.ylabel("latitude [degree_north]")
plt.title("2 metre temperature in Colorado on Jan 1st, 2021 at 00:00")
# finalize the data component
data_comp.finalize()