Climate reanalysis — ERA5-Land monthly 2 m temperature¶
ERA5-Land monthly aggregates resampled to a yearly mean 2 m temperature over the Nile delta — a typical climate-reanalysis pipeline.
Setup¶
First the imports. pyramids provides Dataset (GeoTIFF/NetCDF reading); earthlens provides
the unified EarthLens entry point and the bundled GEE Catalog.
import os
from pathlib import Path
import matplotlib.pyplot as plt
import numpy as np
from pyramids.dataset import Dataset as PyramidsDataset
from earthlens import EarthLens
from earthlens.gee import Catalog
Output directory and credentials¶
Written GeoTIFFs go under a per-notebook out/ directory. The GEE service-account credentials
are read from the GEE_SERVICE_ACCOUNT / GEE_SERVICE_KEY environment variables — both must be
set before running this cell.
OUT_DIR = Path('out') / 'climate-reanalysis'
OUT_DIR.mkdir(parents=True, exist_ok=True)
SERVICE_ACCOUNT = os.environ['GEE_SERVICE_ACCOUNT']
SERVICE_KEY = os.environ['GEE_SERVICE_KEY']
print(f'output directory: {OUT_DIR.resolve()}')
Inspect the catalog entry¶
Before downloading anything, look at what the bundled catalog knows about the asset — bands, cadence, license, provider.
cat = Catalog()
ds = cat.get_dataset('ECMWF/ERA5_LAND/MONTHLY_AGGR')
print(f'title: {ds.title}')
print(f'ee_type: {ds.ee_type}')
print(f'spatial_resolution: {ds.spatial_resolution} m')
print(f'extent.start_date: {ds.extent.start_date}')
print(f'extent.end_date: {ds.extent.end_date}')
print(f'default_reducer: {ds.default_reducer}')
print(f'license: {ds.license}')
print(f'provider: {ds.provider}')
print(f'#bands: {len(ds.bands)}')
print(f'band ids (first 5): {list(ds.bands)[:5]}')
Download¶
Tiny AOI ([29.0, 32.0] lat, [30.0, 33.0] lon) at 11132.0 m, yearly cadence — keeps the synchronous download under EE's 32768-px per-axis cap.
Build the request¶
Construct the EarthLens request first — source, dataset, band (variables), AOI, date window,
cadence, and the synchronous getDownloadURL defaults.
gee = EarthLens(
data_source="gee",
start='2023-01-01',
end='2023-12-31',
dataset='ECMWF/ERA5_LAND/MONTHLY_AGGR',
variables=['temperature_2m'],
aoi=[30.0, 29.0, 33.0, 32.0],
cadence='yearly',
path=str(OUT_DIR),
scale=11132.0,
reducer='mean',
)
Authenticate and download¶
authenticate() resolves the service-account credentials on its own line, then download()
fetches the composite to disk. The try/except keeps the graceful-skip guard the other GEE
example notebooks share.
gee.authenticate(service_account=SERVICE_ACCOUNT, service_key=SERVICE_KEY)
try:
paths = gee.download(progress_bar=False)
print(f'wrote {len(paths)} GeoTIFF(s):')
for p in paths:
print(f' {p} ({p.stat().st_size / 1024:.1f} KB)')
download_ok = True
except Exception as exc:
if False:
print(
f'live EE call failed (tolerated for this category): {type(exc).__name__}: {exc}'
)
paths = []
download_ok = False
else:
raise
Quick preview¶
Load the first written GeoTIFF through pyramids and render the single band. (pyramids.dataset.Dataset is the project's GeoTIFF/NetCDF wrapper.)
Load and mask the band¶
Read the first written GeoTIFF through pyramids, collapse it to a single band, and mask the dataset's nodata value so the colormap isn't pinned to it.
if not download_ok or not paths:
arr = None
print('Skipping preview — no GeoTIFF was written.')
else:
pds = PyramidsDataset.read_file(str(paths[0]))
arr = pds.read_array()
if arr.ndim == 3:
arr = arr[0]
# Mask the dataset's nodata so the colormap doesn't get pinned to it.
nodata = pds.no_data_value
if nodata is not None:
try:
arr = np.ma.masked_equal(
arr, float(nodata[0] if isinstance(nodata, (list, tuple)) else nodata)
)
except (TypeError, ValueError):
pass
Render the band¶
Display the single-band array with a viridis colormap and report its value range.
if arr is None:
print('Nothing to plot.')
else:
fig, ax = plt.subplots(figsize=(6, 5))
im = ax.imshow(arr, cmap='viridis')
ax.set_title(f'{'ECMWF/ERA5_LAND/MONTHLY_AGGR'} / {'temperature_2m'}')
ax.set_xlabel('x (px)')
ax.set_ylabel('y (px)')
fig.colorbar(im, ax=ax, shrink=0.8)
plt.tight_layout()
plt.show()
print(f'value range: [{float(np.nanmin(arr)):.4g}, {float(np.nanmax(arr)):.4g}]')
Tracking submitted jobs (asynchronous export)¶
The download above uses export_via="url" — a synchronous getDownloadURL round-trip. Nothing was queued, so there's no Earth Engine job to track.
To track an export instead, switch to an asynchronous sink (drive / gcs / asset) and pass wait_for_export=False so .download() returns a TaskInfo at submission time rather than blocking until completion. The cells below submit the same (asset_id, band, AOI, scale) request as an export_via="asset" task into the service account's own asset folder, then walk the four jobs-API calls (list_recent_tasks → wait_for_task_id → ee.data.getAsset → ee.data.deleteAsset) to make the job finish and tidy up. See track-batch-exports.ipynb for a deeper worked example.
Imports and the demo asset folder¶
Bring in ee plus the earthlens.gee task helpers, and derive a Folder asset path under the
current project. The backend writes the image at <asset_id>/<prefix>, so this asset_id is the
parent folder, not the final image path.
import ee
from earthlens.gee import cancel_task, list_recent_tasks, wait_for_task_id
# The asset goes into a `Folder` asset that we own. `GEE._export_via_batch`
# writes the actual image at `<asset_id>/<prefix>`, so `asset_id` here is
# the parent FOLDER (not the final image path). Both must be cleaned up.
_proj = ee.data._get_projects_path().removeprefix('projects/')
DEMO_FOLDER = f'projects/{_proj}/assets/earthlens-demo-climate-reanalysis'
print(f'demo folder: {DEMO_FOLDER}')
Prepare a clean folder¶
Best-effort tear down any leftovers from a previous run, then (re)create the parent folder — Earth Engine requires it to exist before a child write.
# Best-effort cleanup of leftover children from a previous run (so the
# folder is empty before we try to delete it below).
try:
for child in ee.data.listAssets({'parent': DEMO_FOLDER}).get('assets', []):
ee.data.deleteAsset(child['name'])
print(f'cleared leftover child: {child["name"]}')
ee.data.deleteAsset(DEMO_FOLDER)
print(f'cleared leftover folder: {DEMO_FOLDER}')
except Exception:
pass
# Create the parent folder — EE requires it to exist before a child write.
ee.data.createAsset({'type': 'Folder'}, DEMO_FOLDER)
print(f'created folder: {DEMO_FOLDER}')
Submit¶
Same (asset_id, band, AOI, scale) request as the sync download above, just routed through export_via="asset" + wait_for_export=False. download() returns a TaskInfo per submitted bucket at the moment the task is queued — no blocking.
Build the async request and authenticate¶
Same (asset_id, band, AOI, scale) request as the sync download, but routed through
export_via="asset" + wait_for_export=False. Construct it first, then authenticate() on its
own line.
async_gee = EarthLens(
data_source="gee",
start='2023-01-01',
end='2023-12-31',
dataset='ECMWF/ERA5_LAND/MONTHLY_AGGR',
variables=['temperature_2m'],
aoi=[30.0, 29.0, 33.0, 32.0],
cadence='yearly',
path=str(OUT_DIR),
scale=11132.0,
reducer='mean',
export_via='asset',
asset_id=DEMO_FOLDER,
wait_for_export=False,
)
async_gee.authenticate(service_account=SERVICE_ACCOUNT, service_key=SERVICE_KEY)
Submit the export¶
download() returns a TaskInfo per submitted bucket at the moment the task is queued — no
blocking. The try/except keeps the graceful-skip guard.
submitted_ok = False
task_info = None
try:
submitted = async_gee.download(progress_bar=False)
task_info = submitted[0]
submitted_ok = True
print(f'submitted: id={task_info.id} state={task_info.state}')
print(f' description={task_info.description}')
except Exception as exc:
if False:
print(f'async submission failed (tolerated): {type(exc).__name__}: {exc}')
else:
raise
List + wait¶
list_recent_tasks(description_prefix=...) returns every matching task across the current project; wait_for_task_id blocks until the one we care about reaches a terminal state. A real workflow would just poll later from a separate process — the wait here exists so the notebook shows the full success path end-to-end.
if submitted_ok and task_info is not None:
recent = list_recent_tasks(
description_prefix=task_info.description,
max_age_min=10,
)
print(f'list_recent_tasks matched {len(recent)} task(s):')
for t in recent:
print(f' {t.id} {t.state:<12} {t.description}')
try:
final = wait_for_task_id(
task_info.id,
poll_seconds=10,
progress_bar=False,
)
print(f'\nfinal state: {final.state}')
except RuntimeError as exc:
# Raised on FAILED / CANCELLED — cancel-if-still-running
# so we don't leak an in-flight task on notebook restart.
print(f'wait_for_task_id raised: {exc}')
try:
cancel_task(task_info.id)
except Exception:
pass
else:
print('Skipping list/wait — async submission did not succeed.')
Verify + clean up¶
Confirm the produced asset exists on Earth Engine, then delete it (and the surrounding demo folder) so we don't leak storage between notebook runs. The backend wrote the image at <DEMO_FOLDER>/<task description>.
if submitted_ok and task_info is not None:
produced = f'{DEMO_FOLDER}/{task_info.description}'
try:
meta = ee.data.getAsset(produced)
print(f'asset exists: type={meta.get("type")} ' f'name={meta.get("name")}')
ee.data.deleteAsset(produced)
print('asset deleted')
except Exception as exc:
print(f'verify/delete skipped: {exc}')
# Always try to tear down the parent folder.
try:
ee.data.deleteAsset(DEMO_FOLDER)
print(f'folder deleted: {DEMO_FOLDER}')
except Exception as exc:
print(f'folder delete skipped: {exc}')
What's on disk¶
The GeoTIFF (or empty list, if the EE call was tolerated) is left under the per-notebook out/ directory for you to inspect. That directory is .gitignored — re-running the notebook overwrites it.
for p in sorted(OUT_DIR.iterdir()) if OUT_DIR.exists() else []:
print(f'{p} ({p.stat().st_size / 1024:.1f} KB)')