GeoTessera

Python library for accessing and working with Tessera geospatial foundation model embeddings.

🚀 TESSERA v2 is here

The TESSERA v2 code is now live — model weights and inference code are
available in the ucam-eo/tessera
repository. v2 is our next-generation pixel-wise Earth foundation model; see
the preprint, TESSERA v2: Scaling Pixel-wise Earth Foundation Models.

Want to try v2 embeddings early? You can pre-request v2 embeddings for
your region and become an early tester:

• 👉 Submit a v2 Embedding Pre-Request

⚠️ Heads-up: we are still ramping up the compute, storage, and release
infrastructure for v2, so v2 embeddings will be produced slowly at first
and there is no guaranteed turnaround time. If you need embeddings soon,
request v1.1 instead — it is fully available today. Scroll down to
Request missing embeddings to open a standard
v1.1 request.

Overview

GeoTessera provides access to geospatial embeddings from the Tessera
foundation model, which processes
Sentinel-1 and Sentinel-2 satellite imagery to generate 128-channel
representation maps at 10m resolution. These embeddings compress a full year of
temporal-spectral features into dense representations optimized for downstream
geospatial analysis tasks. Read more details about the model.

Request missing embeddings

This repo provides precomputed embeddings for multiple years and regions.
Embeddings are generated by randomly sampling tiles within each region to ensure broad spatial coverage.

If some years (2017–2025) / areas are still missing for your use case, please submit an Embedding Request:

• 👉 Open an Embedding Request
• Please include: your organization, intended use, ROI as a bounding box with four points (lon,lat, 4 decimals), and the year(s).

After you submit the request, we will prioritize your ROI and notify you via a comment in the issue once the embeddings are ready.

Important Notice ⚠️

On 20th August 2025, we updated the data processing pipeline of GeoTessera to resolve the issue of tiling artifacts, as shown below. We have retained the embeddings generated before August 20, as they remain effective for use in small-scale areas. After the 2024 embedding generation is completed, we will reprocess the tiles affected by tiling artifacts. If you observe such artifacts during use and they significantly impact performance, please raise the issue here, and we will prioritize reprocessing your request.

Please note that if the artifacts you observe are slanted, this is not a bug in the pipeline but rather a result of the Sentinel-1/2 satellite trajectories. Currently, Tessera cannot completely eliminate such artifacts, as they reflect the inherent characteristics of the raw data. However, we have observed that they have minimal impact on downstream tasks.

Table of Contents

• Installation
• Architecture
• Quick Start
• Python API
• Cloud-Native Zarr Access
• CLI Reference
• Complete Workflows
• Registry System
• Data Organization
• Contributing

Installation

Requires Python 3.12 or later.

pip install geotessera

For development:

git clone https://github.com/ucam-eo/geotessera
cd geotessera
pip install -e .

Architecture

Core Concepts

GeoTessera is built around a simple two-step workflow:

1. Retrieve embeddings: Fetch raw numpy arrays for a geographic bounding box
2. Export to desired format: Save as raw numpy arrays or convert to georeferenced GeoTIFF files

Coordinate System and Tile Grid

The Tessera embeddings use a 0.1-degree grid system:

• Tile size: Each tile covers 0.1° × 0.1° (approximately 11km × 11km at the equator)
• Tile naming: Tiles are named by their center coordinates (e.g., grid_0.15_52.05)
• Tile bounds: A tile at center (lon, lat) covers:
  • Longitude: [lon - 0.05°, lon + 0.05°]
  • Latitude: [lat - 0.05°, lat + 0.05°]
• Resolution: 10m per pixel (variable number of pixels per tile depending on latitude)

File Structure and Downloads

When you request embeddings, GeoTessera downloads files from the public S3
bucket (using anonymous, unsigned requests) into the output directory you
specify, where they persist for re-use:

Embedding Files (via fetch_embedding)

1. Quantized embeddings (grid_X.XX_Y.YY.npy):

   • Shape: (height, width, 128)
   • Data type: int8 (quantized for storage efficiency)
   • Contains the compressed embedding values

2. Scale files (grid_X.XX_Y.YY_scales.npy):

   • Shape: (height, width) or (height, width, 128)
   • Data type: float32
   • Contains scale factors for dequantization

3. Dequantization: final_embedding = quantized_embedding * scales

4. Persistent Storage: Files are downloaded into your chosen output
   directory and skipped on rerun, so interrupted downloads resume cleanly

Landmask Files (for GeoTIFF export)

When exporting to GeoTIFF, additional landmask files are fetched:

• Landmask tiles (grid_X.XX_Y.YY.tiff):
  • Provide UTM projection information
  • Define precise geospatial transforms
  • Contain land/water masks
  • Cached alongside the embedding tiles for re-use

Data Flow

User Request (lat/lon bbox)
    ↓
Parquet Registry Lookup (find available tiles from manifest.parquet)
    ↓
Anonymous S3 Downloads to Output Directory (CRC64NVMe verified)
    ├── embedding.npy (quantized) → output dir
    └── embedding_scales.npy → output dir
    ↓
Dequantization (multiply arrays)
    ↓
Output Format
    ├── NumPy arrays → Direct analysis
    └── GeoTIFF → GIS integration

Storage Note: Only the per-version Parquet manifests (~few MB each) are
cached under ~/.cache/geotessera. Embedding tiles are downloaded on demand
into the output directory you specify and persist there for re-use across runs.

Quick Start

Check Available Data

Before downloading, check what data is available:

# Generate a coverage map showing all available tiles
geotessera coverage --output coverage_map.png

# Generate a coverage map for the UK
geotessera coverage --country uk

# View coverage for a specific year
geotessera coverage --year 2024 --output coverage_2024.png

# Customize the visualization
geotessera coverage --year 2024 --tile-color blue --tile-alpha 0.3

Download Embeddings

Download embeddings as either numpy arrays or GeoTIFF files:

# Download as GeoTIFF (default, with georeferencing)
geotessera download \
  --bbox "-0.2,51.4,0.1,51.6" \
  --year 2024 \
  --output ./london_tiffs

# Download as raw numpy arrays (with metadata JSON)
geotessera download \
  --bbox "-0.2,51.4,0.1,51.6" \
  --format npy \
  --year 2024 \
  --output ./london_arrays

# Download using a GeoJSON/Shapefile region
geotessera download \
  --region-file cambridge.geojson \
  --format tiff \
  --year 2024 \
  --output ./cambridge_tiles

# Download specific bands only
geotessera download \
  --bbox "-0.2,51.4,0.1,51.6" \
  --bands "0,1,2" \
  --year 2024 \
  --output ./london_rgb

Create Visualizations

Generate PCA visualizations and web maps from downloaded GeoTIFFs:

# Create a PCA mosaic from downloaded tiles
geotessera visualize ./london_tiffs pca_mosaic.tif

# Use histogram equalization for maximum contrast
geotessera visualize ./london_tiffs pca_balanced.tif --balance histogram

# Create web tiles and serve interactively
geotessera webmap pca_mosaic.tif --serve

# Serve existing web visualizations locally
geotessera serve ./london_web --open

Python API

Core Methods

The library provides two main methods for retrieving embeddings:

from geotessera import GeoTessera

# Initialize the client
gt = GeoTessera()

# Method 1: Fetch a single tile
embedding, crs, transform = gt.fetch_embedding(lon=0.15, lat=52.05, year=2024)
print(f"Shape: {embedding.shape}")  # e.g., (1200, 1200, 128)
print(f"CRS: {crs}")  # Coordinate reference system from landmask

# Method 2: Fetch all tiles in a bounding box
bbox = (-0.2, 51.4, 0.1, 51.6)  # (min_lon, min_lat, max_lon, max_lat)
tiles_to_fetch = gt.registry.load_blocks_for_region(bounds=bbox, year=2024)
embeddings = gt.fetch_embeddings(tiles_to_fetch)

for year, tile_lon, tile_lat, embedding_array, crs, transform in embeddings:
    print(f"Tile ({tile_lat}, {tile_lon}): {embedding_array.shape}")

Export Formats

Export as GeoTIFF

# Export embeddings for a region as individual GeoTIFF files
# Step 1: Get the tiles for the region
bbox = (-0.2, 51.4, 0.1, 51.6)
tiles_to_fetch = gt.registry.load_blocks_for_region(bounds=bbox, year=2024)

# Step 2: Export those tiles as GeoTIFFs
files = gt.export_embedding_geotiffs(
    tiles_to_fetch=tiles_to_fetch,
    output_dir="./output",
    bands=None,  # Export all 128 bands (default)
    compress="lzw"  # Compression method
)

print(f"Created {len(files)} GeoTIFF files")

# Export specific bands only (e.g., first 3 for RGB visualization)
files = gt.export_embedding_geotiffs(
    tiles_to_fetch=tiles_to_fetch,
    output_dir="./rgb_output",
    bands=[0, 1, 2]  # Only export first 3 bands
)

Work with NumPy Arrays

# Fetch and process embeddings directly
tiles_to_fetch = gt.registry.load_blocks_for_region(bounds=bbox, year=2024)
embeddings = gt.fetch_embeddings(tiles_to_fetch)

for year, tile_lon, tile_lat, embedding, crs, transform in embeddings:
    # Compute statistics
    mean_values = np.mean(embedding, axis=(0, 1))  # Mean per channel
    std_values = np.std(embedding, axis=(0, 1))    # Std per channel

    # Extract specific pixels
    center_pixel = embedding[embedding.shape[0]//2, embedding.shape[1]//2, :]

    # Apply custom processing
    processed = your_analysis_function(embedding)

Visualization Functions

from geotessera.visualization import (
    create_rgb_mosaic,
    visualize_global_coverage
)
from geotessera.web import (
    create_coverage_summary_map,
    geotiff_to_web_tiles
)

# Create an RGB mosaic from multiple GeoTIFF files
create_rgb_mosaic(
    geotiff_paths=["tile1.tif", "tile2.tif"],
    output_path="mosaic.tif",
    bands=(0, 1, 2)  # RGB bands
)

# Generate web tiles for interactive maps
geotiff_to_web_tiles(
    geotiff_path="mosaic.tif",
    output_dir="./web_tiles",
    zoom_levels=(8, 15)
)

# Create a global coverage visualization
visualize_global_coverage(
    tessera_client=gt,
    output_path="global_coverage.png",
    year=2024,  # Or None for all years
    width_pixels=2000,
    tile_color="red",
    tile_alpha=0.6
)

Cloud-Native Zarr Access

For interactive or large-scale analysis without downloading files, use the Zarr store.
This streams data directly from the cloud:

from geotessera.store import GeoTesseraZarr

gt = GeoTesseraZarr()
print(gt.years)  # [2017, 2018, ..., 2025]

# Sample embeddings at specific points (no download needed)
X = gt.sample_points([(-2.97, 53.44), (0.15, 52.05)], year=2025)
print(f"Shape: {X.shape}")  # (2, 128)

# Read a full region as a mosaic
mosaic, transform, crs = gt.read_region(
    (-3.0, 53.4, -2.9, 53.5), year=2025,
)
print(f"Mosaic shape: {mosaic.shape}")

# Work with individual UTM zones via xarray
ds = gt.open_zone(lon=0.15)
print(ds)

The Zarr store implements the geoemb: convention for geospatial embedding data
and automatically routes queries to the correct UTM zone.

CLI Reference

download

Download embeddings for a region in your preferred format:

geotessera download [OPTIONS]

Options:
  -o, --output PATH         Output directory [required]
  --bbox TEXT              Bounding box: 'lon,lat' (single tile) or 'min_lon,min_lat,max_lon,max_lat'
  --tile TEXT              Single tile by any point within it: 'lon,lat'
  --region-file PATH       GeoJSON/Shapefile to define region
  --country TEXT           Country name (e.g., 'United Kingdom', 'UK', 'GB')
  -f, --format TEXT        Output format: 'tiff' or 'npy' (default: tiff)
  --year INT               Year of embeddings (default: 2024)
  --dataset-version TEXT   Tessera dataset version (e.g. v1, v1.1)
  --dataset-variant TEXT   Tessera dataset variant (default: vultr)
  --bands TEXT             Comma-separated band indices (default: all 128)
  --compress TEXT          Compression for TIFF format (default: lzw)
  --dry-run                Calculate total download size without downloading
  --list-files             List all created files with details
  -v, --verbose            Verbose output

Resume behaviour: Both TIFF and NPY downloads automatically skip files that already exist on disk, so interrupted downloads can be resumed by re-running the same command.

Single tile examples:

# Download a single tile containing a specific point
geotessera download --tile "0.17,52.23" --year 2024 -o ./single_tile

# Same result using --bbox with 2 coordinates
geotessera download --bbox "0.17,52.23" --year 2024 -o ./single_tile

Output formats:

• tiff: Georeferenced GeoTIFF files with UTM projection
• npy: Raw numpy arrays with metadata.json file

visualize

Create PCA visualization from multiband GeoTIFF or NPY format embeddings:

geotessera visualize INPUT_PATH OUTPUT_FILE [OPTIONS]

Options:
  --n-components INT       Number of PCA components (default: 3)
  --crs TEXT               Target CRS for reprojection (default: EPSG:3857)
  --balance TEXT            RGB balance method: histogram, percentile, or adaptive
  --percentile-low FLOAT   Lower percentile for percentile balance (default: 2.0)
  --percentile-high FLOAT  Upper percentile for percentile balance (default: 98.0)

webmap

Create web tiles and interactive viewer from a PCA mosaic:

geotessera webmap RGB_MOSAIC [OPTIONS]

Options:
  -o, --output PATH        Output directory
  --min-zoom INT           Min zoom for web tiles (default: 8)
  --max-zoom INT           Max zoom for web tiles (default: 15)
  --serve/--no-serve       Start web server immediately
  -p, --port INT           Port for web server (default: 8000)
  --region-file PATH       GeoJSON/Shapefile boundary to overlay
  --force/--no-force       Force regeneration of tiles

coverage

Generate a world map showing data availability:

geotessera coverage [OPTIONS]

Options:
  -o, --output PATH        Output PNG file, or a directory to also receive the
                           coverage.json/globe.html (default: tessera_coverage.png)
  --year INT               Specific year to visualize
  --bbox TEXT              Bounding box: 'lon,lat' (single tile) or 'min_lon,min_lat,max_lon,max_lat'
  --tile TEXT              Single tile by any point within it: 'lon,lat'
  --by-source              Render each (version, variant) source in a distinct colour
  --dataset-version TEXT   Tessera dataset version (e.g. v1, v1.1; or 'all' with --by-source)
  --dataset-variant TEXT   Tessera dataset variant (default: vultr; or 'all' with --by-source)
  --region-file PATH       GeoJSON/Shapefile to focus on specific region
  --country TEXT           Country name to focus on (e.g., 'United Kingdom')
  --tile-color TEXT        Color for tiles (default: red)
  --tile-alpha FLOAT       Transparency 0-1 (default: 0.6)
  --tile-size FLOAT        Size multiplier (default: 1.0)
  --width INT              Output image width in pixels (default: 2000)
  --no-countries           Don't show country boundaries
  --no-multi-year-colors   Disable multi-year color coding

serve

Serve web visualizations locally:

geotessera serve DIRECTORY [OPTIONS]

Options:
  -p, --port INT           Port number (default: 8000)
  --open/--no-open         Auto-open browser (default: open)
  --html TEXT              Specific HTML file to serve

info

Display information about GeoTIFF files or the library:

geotessera info [OPTIONS]

Options:
  --tiles PATH             Analyze tile files/directory (GeoTIFF or NPY format)
  --dataset-version TEXT   Tessera dataset version (e.g. v1, v1.1)
  --dataset-variant TEXT   Tessera dataset variant (default: vultr)
  -v, --verbose            Verbose output

Registry System

Overview

GeoTessera uses a Parquet-based registry system to efficiently manage and access the large Tessera dataset:

• Per-version manifests: Each dataset version has its own manifest.parquet
  listing every (year, lon, lat) tile available for that version's variants
• Fast queries: Uses pandas DataFrames for efficient spatial and temporal filtering
• Block-based organization: Internal 5×5 degree geographic blocks for efficient queries
• Minimal storage: Manifest files are ~few MB each and cached locally
• Integrity checking: End-to-end CRC64NVMe checksums verified against S3's
  x-amz-checksum-crc64nvme response header during each download
  • Always enforced for data integrity — a checksum mismatch (or a missing checksum header) rejects the download

Dataset Versions and Variants

Tessera embeddings are published as dataset versions (e.g. v1, v1.1) and,
within a version, as variants produced by different model runs (e.g. the
default vultr, or cambridge). Select them on the CLI with --dataset-version
and --dataset-variant, or in Python:

gt = GeoTessera(dataset_version="v1.1", dataset_variant="cambridge")

Use geotessera coverage --by-source to render each (version, variant) source
in a distinct colour on the coverage map and globe viewer.

Registry Sources

The registry can be loaded from multiple sources (in priority order):

1. Local file (via registry_path parameter)
2. Local directory (via --registry-dir or registry_dir parameter, looks for manifest.parquet, falling back to the legacy registry.parquet)
3. Remote URL (via registry_url parameter)
4. Default remote (from https://s3.us-west-2.amazonaws.com/tessera-embeddings/{version}/manifest.parquet)

# Use local manifest file
gt = GeoTessera(registry_path="/path/to/manifest.parquet")

# Use local registry directory
gt = GeoTessera(registry_dir="/path/to/registry-dir")

# Use default remote manifest (downloads and caches automatically)
gt = GeoTessera()  # Default behavior

Registry Structure

The Parquet manifest contains columns for:

• Coordinates: lon, lat (tile center coordinates)
• Year: year (data year, 2017-2025)
• Size: file_size (file size in bytes for download planning)

# Example manifest query
import pandas as pd
manifest = pd.read_parquet("manifest.parquet")
print(manifest.head())

How Registry Loading Works

1. Load Parquet manifest → Download and cache the version's manifest (if not local)
2. Request tiles for bbox → Query DataFrame for tiles in region
3. Filter by year and variant → Select tiles matching the requested year/variant
4. Find available tiles → Return list of matching tiles
5. Anonymous S3 download → Fetch tiles on demand into the output directory, verified with CRC64NVMe
6. Persist → Downloaded tiles stay in the output directory and are skipped on rerun

Data Organization

Tessera Data Structure

Remote Server (https://s3.us-west-2.amazonaws.com/tessera-embeddings)
├── v1/                                        # Dataset version 1.0
│   ├── manifest.parquet                       # Per-version tile manifest
│   ├── landmasks.parquet                      # Landmask manifest
│   ├── global_0.1_degree_representation/      # vultr variant (default)
│   │   └── 2024/grid_0.15_52.05/grid_0.15_52.05{,_scales}.npy
│   └── global_0.1_degree_tiff_all/
│       └── grid_0.15_52.05.tiff               # Landmask with projection info
└── v1.1/                                      # Dataset version 1.1
    ├── manifest.parquet
    ├── landmasks.parquet
    └── global_0.1_degree_representation.cambridge/
        └── 2024/grid_0.15_52.05/grid_0.15_52.05{,_scales}.npy

Local Cache Structure

~/.cache/geotessera/                 # Default cache location (manifests only)
├── v1/
│   ├── manifest.parquet             # Cached per-version manifest (~few MB)
│   └── landmasks.parquet
└── v1.1/
    ├── manifest.parquet
    └── landmasks.parquet

# Note: Embedding and landmask tiles are NOT stored here. They are downloaded
# into the output directory you specify and persist there for re-use.

Coordinate Reference Systems

• Embeddings: Stored in simple arrays, referenced by center coordinates
• GeoTIFF exports: Use UTM projection from corresponding landmask tiles
• Web visualizations: Reprojected to Web Mercator (EPSG:3857)

Cache Configuration

GeoTessera caches only the per-version Parquet manifests (~few MB each). Embedding and landmask tiles are downloaded into the output directory you specify and persist there for re-use across runs.

Python API

from geotessera import GeoTessera

# Use custom cache directory for registry
gt = GeoTessera(cache_dir="/path/to/cache")

# Use default cache location (recommended)
gt = GeoTessera()

CLI

# Specify custom cache directory
geotessera download --cache-dir /path/to/cache ...

# Use default cache location
geotessera download ...

Default Cache Locations

When cache_dir is not specified, the registry is cached in platform-appropriate locations:

• Linux/macOS: $XDG_CACHE_HOME/geotessera or ~/.cache/geotessera
• Windows: %LOCALAPPDATA%/geotessera

Hash Verification

GeoTessera verifies end-to-end CRC64NVMe checksums for all downloaded files (embeddings, scales, and landmasks) against S3's x-amz-checksum-crc64nvme response header to ensure data integrity. This check is always enforced: a download whose checksum does not match — or whose S3 object is missing the checksum header — is rejected rather than used, so corrupt or truncated files never reach the cache.

Contributing

Contributions are welcome! Please see our Contributing Guide for details.
This project is licensed under the MIT License - see the LICENSE file for details.

Citation

If you use Tessera in your research, please cite the arXiv paper:

@misc{feng2025tesseratemporalembeddingssurface,
      title={TESSERA: Temporal Embeddings of Surface Spectra for Earth Representation and Analysis}, 
      author={Zhengpeng Feng and Clement Atzberger and Sadiq Jaffer and Jovana Knezevic and Silja Sormunen and Robin Young and Madeline C Lisaius and Markus Immitzer and David A. Coomes and Anil Madhavapeddy and Andrew Blake and Srinivasan Keshav},
      year={2025},
      eprint={2506.20380},
      archivePrefix={arXiv},
      primaryClass={cs.LG},
      url={https://arxiv.org/abs/2506.20380}, 
}

Links

• Tessera Foundation Model
• Tessera Interactive Notebook
• Tessera Examples
• Documentation
• PyPI Package
• Issue Tracker

Star History