Microsoft Planetary Computer Examples

The Global Downscaled Projections for Climate Impacts Research (GDPCIR) dataset, produced by the Climate Impact Lab, provides global, daily climate variables on a 1/4-degree regular global grid.

Key variables included:

• tasmin: Daily minimum air temperature at the surface.
• tasmax: Daily maximum air temperature at the surface.
• pr: Daily cumulative surface precipitation.

Supported scenarios:

• CMIP6 historical
• ssp1-2.6
• ssp2-4.5
• ssp3-7.0
• ssp5-8.5

The dataset covers 25 global climate models.

The planetarycomputerexamples repository provides a collection of notebooks and guides for working with the Planetary Computer Hub. You can use these resources to learn how to access STAC APIs, work with specific datasets (like Sentinel-2, Landsat, or ERA5), and perform advanced geospatial analyses using tools like Dask, Xarray, and Leafmap.

To view these examples without local setup, you can use nbviewer.

Key learning paths include:

• Quickstarts: High-level introductions to topics like reading STAC APIs, Zarr data, or using Leafmap.
• Datasets: Specific implementation examples for various geospatial datasets.
• Tutorials: In-depth guides on complex topics like Cloudless Mosaics, Landcover Classification, and Zonal Statistics.

To begin working with the GDPCIR dataset, you can use the Microsoft Planetary Computer STAC API to discover and access the data. The primary entry point for exploration is the cil-gdpcir-example.ipynb notebook, which demonstrates how to query the dataset via STAC.

See: ./cil-gdpcir-example.ipynb

Beyond basic exploration, you can use the following notebooks to perform advanced climate analysis:

• Building Ensembles: Use ensemble.ipynb to learn how to select specific STAC collections and build climate model ensembles.
• Computing Indicators: Use indicators.ipynb to compute climate and impact indicators using the xclim library.

Notebooks available:
- ./ensemble.ipynb
- ./indicators.ipynb

Sentinel-1 RTC items contain assets representing terrain-corrected gamma nought values for different polarizations. Common asset names include hh, hv, vh, and vv. You can check available polarizations using the sar:polarizations property in the STAC metadata. For quick visualization, use the rendered_preview asset.

# Check available polarizations
print(item.properties["sar:polarizations"])

# Quick visualization using the rendered_preview asset
from IPython.display import Image
Image(url=item.assets["rendered_preview"].href)

Every Sentinel-1 RTC item is derived from a Sentinel-1 GRD item. You can find the original GRD item by following the derived_from link in the RTC item's metadata. RTC products are terrain-corrected, meaning they align more accurately with Earth's surface compared to the original GRD products.

# Get the RTC item
rtc_item = catalog.get_collection("sentinel-1-rtc").get_item("ITEM_ID")

# Access the original GRD item via the 'derived_from' link
import pystac
gr_item = pystac.read_file(rtc_item.get_single_link("derived_from").target)

The localserver service runs a simple Python HTTP server on port 8889. It is used to serve local files from the datasets, quickstarts, and tutorials directories. These directories are mounted to /opt/src inside the container. To start the service, run docker-compose up.

localserver:
        image: python:3.9-alpine
        ports:
            - "8889:8889"
        volumes:
            - ./datasets:/opt/src/datasets
            - ./quickstarts:/opt/src/quickstarts
            - ./tutorials:/opt/src/tutorials
        command: python3 -m "http.server" 8889 -d /opt/src

To access data with more permissive access, you should use a subscription key. If you are running on the Planetary Computer Hub, the environment variable PC_SDK_SUBSCRIPTION_KEY is set automatically. Otherwise, you can manually set your key using pc.settings.set_subscription_key().

import planetary_computer as pc
# Replace <YOUR API Key> with your actual key from the developer portal
pc.settings.set_subscription_key(<YOUR API Key>)

To use the Planetary Computer locally, you must provide an API key. You can do this by setting the PC_SDK_SUBSCRIPTION_KEY environment variable or by using the planetary_computer.settings.set_subscription_key() method in your Python code.

# Option 1: Set environment variable PC_SDK_SUBSCRIPTION_KEY

# Option 2: Use the SDK method
import planetary_computer
planetary_computer.settings.set_subscription_key('<YOUR API Key>')

To run the benchmark notebook, install the required Python packages using pip:

!pip install pandas_path pytorch_lightning cloudpathlib loguru typer

To access data with more permissive access, you should use an API key. You can set the PC_SDK_SUBSCRIPTION_KEY environment variable or use the planetary_computer SDK to set it programmatically. If you are using the Planetary Computer Hub, the key is configured automatically.

Required imports for Sentinel-2 workflows:

from pystac.extensions.eo import EOExtension as eo
import pystac_client
import planetary_computer

import planetary_computer

# Set the subscription key programmatically
planetary_computer.settings.set_subscription_key('<YOUR API Key>')