Sentinel-2 monthly NDVI time-series (Earth Search)¶

Shows the raster aggregate= path: pull Sentinel-2 red + NIR over a small AOI for a couple of months, reduce per month into per-window COGs, then compute mean NDVI per month. The download runs live against Earth Search (anonymous).

In [ ]:

import tempfile

from earthlens.aggregate import AggregationConfig
from earthlens.earthlens import EarthLens

out = tempfile.mkdtemp()
el = EarthLens(
    data_source="earth-search",
    start="2024-06-01",
    end="2024-07-31",
    dataset="sentinel-2-l2a",
    variables=["red", "nir"],
    aoi=[-3.72, 40.40, -3.67, 40.45],
    path=out,
    max_items=2,
)
monthly = el.download(
    aggregate=AggregationConfig(freq="1MS", op="mean", out_dir=out + "/monthly")
)
[p.name for p in monthly]

import tempfile from earthlens.aggregate import AggregationConfig from earthlens.earthlens import EarthLens out = tempfile.mkdtemp() el = EarthLens( data_source="earth-search", start="2024-06-01", end="2024-07-31", dataset="sentinel-2-l2a", variables=["red", "nir"], aoi=[-3.72, 40.40, -3.67, 40.45], path=out, max_items=2, ) monthly = el.download( aggregate=AggregationConfig(freq="1MS", op="mean", out_dir=out + "/monthly") ) [p.name for p in monthly]

Each path is one COG per month, named sentinel-2-l2a_mean_1MS_<YYYYMMDD>.tif. Compute the mean NDVI per window:

In [ ]:

import numpy as np
from pyramids.dataset import Dataset

for cog in monthly:
    arr = Dataset.read_file(str(cog)).read_array().astype(float)
    red, nir = arr[0], arr[1]
    ndvi = (nir - red) / (nir + red + 1e-9)
    print(cog.name, "mean NDVI:", round(float(np.nanmean(ndvi)), 3))

import numpy as np from pyramids.dataset import Dataset for cog in monthly: arr = Dataset.read_file(str(cog)).read_array().astype(float) red, nir = arr[0], arr[1] ndvi = (nir - red) / (nir + red + 1e-9) print(cog.name, "mean NDVI:", round(float(np.nanmean(ndvi)), 3))