DEWS's DataHub (Drought Early Warning System's DataHub)

Table of contents

• DEWS's DataHub (Drought Early Warning System's DataHub)
  • Description
  • Environment
• Guide
  • Create .env file
    • Production
  • Start Docker container
  • Monitor system activities
  • Create satellite data entries
    • Manual approach: Dataspace Copernicus
      • Import via webpage
    • Alternative approach: Import using Docker
    • Automatic approach: Sentinel Hub API
• DataHub Web App
  • Description
  • Metrics calculation
  • Satellite data aggregation
• PostGIS
  • PostGIS database
• QGIS
  • Import bands/raster from DEWS-DH into QGIS
• Supported satellite data sources
  • Copernicus Browser (Dataspace Copernicus)

Description

DEWS's DataHub was made to import satellite data images from archives downloaded from e.g. Copernicus Browser. It is also possible to use the Sentinel Hub API to request satellite data images. The main goal is to have an overview over your satellite data as well as a PostGIS database to connect with other software.

The satellite mission, product type, thumbnail, etc. will be automatically recognized and displayed in a new entry in the SatData overview tab. When the archive contains satellite images that are raster compatible they will be converted into rasters and imported into the PostGIS database (located at 0.0.0.0:5432).

Overview

Satellite data details

Environment

• Frontend: HTML, CSS, JS (realized with Django Template Language)
• Backend: Python + Django Framework
• Database: PostGIS
• Webserver: Gunicorn + Nginx

Guide

This guide shows how to start the Docker containers locally, how to use the web app, gives information about the PostGIS database, how to use with QGIS and which satellite datasets are supported.

Create .env file

The .env file contains environment variables for the Python Django and the PostGIS environment.
Copy the env file using cp .env.example .env.

The .env.example file is used as a template. For development purposes you do not need to modify .env's content.

Production

For production it is not recommended to leave the .env file unmodified!
Please create a Django secret key using this script

from django.core.management.utils import get_random_secret_key  
print(get_random_secret_key())

or instead use this website.

The .env should look like the following code snippet. Please read the file carefully as the comments speak for themselves!

The Sentinel Hub API key can be created at Sentinel Hub Dashboard - Account Settings. Here you should create a new "OAuth client".

# Settings.py
SECRET_KEY="secret_key_123"  # TODO: Modify on production!
DEBUG=False  # TODO: Modify! (True: development, False: production)

# Create super user
DJANGO_SUPERUSER_USERNAME="dews"  # Forbidden: Do not modify!
DJANGO_SUPERUSER_PASSWORD="dews"  # TODO: Modify on production!
DJANGO_SUPERUSER_EMAIL="dews@dews.de"  # Optional: Modify if needed

# Postgres
PGPASSWORD="dews"  # Optional: Modify if needed

# Sentinel API key
SENTINEL_HUB_API_KEY="key"  # TODO: Modify on production + development!

Start Docker container

1. Install Docker using the official installation guide.
2. Install Docker Compose using the official installation guide.
3. OPTIONAL: Install Docker BuildX.
   • Linux: apt install docker-buildx
   • MacOS: brew install docker-buildx
   • Other OS: Docker BuildX installation guide.
   • INFO: If you experience any issues while building the Docker containers please see this step as necessary!
4. Start a terminal.
5. Change directory to the root of the project (e.g. cd ~/Git/drought-ews).
6. Build the containers:
   • docker-compose up -d --build
   • INFO: In case something went wrong please run sudo rm -rf pgdata/ and try docker-compose up -d --build --force-recreate.
7. If you already build the Docker containers just start them using this command:
   • docker-compose up -d

Monitor system activities

In the root of the repository a log file name django.log mirrors the system's log file in the Docker container. In case of an error or misbehaving you can investiage the problem using the log file.

If you want the contributors to take a look then you should make a copy of the file, create a GitHub issue, attach the django.log file and describe your problem in detail.

Create satellite data entries

Manual approach: Dataspace Copernicus

1. Open Copernicus Browser.
2. Log into your account.
3. Zoom to the area of your interest.
4. Selec the pentagon on the right side and then the rectangle or the pencil.
5. Draw an area around the area of your interest.
6. Select the tab Search on the left side panel.
7. Select the satellite mission and the product type of your desire.
   • e.g Sentinel-2 -> MSI -> L2A
8. It is recommended to recude the cloud coverage to maximum 10%.
9. Choose the time range of your desire.
   • e.g. 2023-11-01 - 2023-11-08
10. Click on search.
11. Hover over an entry and a green area is shown on the map which the hovered dataset will represent.
12. Select the entry of your desire and download it using the download button in the bottom right corner of the entry.

Import via webpage

The archive you downloaded from the Datspae Copernicus can be imported manually via the webpage:

1. Open the "DEWS DataHub - Dashboard".
2. Login with the default credentials dews:dews.
3. Visit the "DEWS DataHub - Create SatData Entry" view or click on the "Add Entry" button at "DEWS DataHub - Data Overview" view.
4. Click on "Choose File" and choose the satellite data ZIP archive you want to upload.
5. Click on "Upload Archive" and wait.

After the import is finished an entry will be displayed in the "DEWS DataHub - Data Overview" view. If raster compatible images exist in the archive the processing can take quite a bit of time.
Please be patient! The process cannot be sped up since the official PostGIS script raster2pgsql is used which is already optimized for this kind of imports.

Alternative approach: Import using Docker

An alternative to the manual approach is to upload your archive into the Docker container and then import it using a custom Django admin command:

1. Copy your local archive into the DEWS container's filesystem:
   • docker cp <archive_path> dews:/dews/media/sat_data/archive/
   • Example: docker cp S1A_IW_GRDH_1SDV_20231116T053343_20231116T053408_051238_062E4B_84CA.SAFE.zip dews:/dews/media/sat_data/archive/
2. Import your data into the database.
   1. Log into the DEWS container.
      • docker exec -it dews /bin/bash
   2. Import your archive using the Django admin command importdata.
      • python manage.py importdata <archive_path> -m <sat_mission>
      • Example: python manage.py importdata /dews/media/sat_data/archive/S1A_IW_GRDH_1SDV_20231116T053343_20231116T053408_051238_062E4B_84CA.SAFE.zip -m sentinel-1a

Automatic approach: Sentinel Hub API

This is the easiest and fastest way to create new satellite date entries on the system. The Sentinel Hub API key in the .env file must be correctly set! Follow the instructions to request and import satellite data images via the API:

1. Open the "DEWS DataHub - Create SatData Entry" view.
2. Scroll down to "Request via API".
3. Select the desired mission.
4. Select the desired resolution.
   1. INFO: The shown bands depend on the resolution selected. Not all resolutions support all bands.
5. OPTIONAL: Select the metrics to calculate.
6. Select the desired bands.
   1. Not needed if metrics were already selected because the selection is linked between the metrics and the bands.
   2. Example: If you click on "NDVI" the bands "B04" and "B08" are automatically selected.
      
   3. If a metric is selected and you need more bands, you still can add more bands by clicking the check boxes.
   4. A metric can not be calculated if you uncheck a automatically checked check box.
7. Choose the time range of your desire.
   • For example: 2023-11-01 - 2023-11-08
   • If you need data for one day only the start and the date date should match (e.g. 2024-04-01 - 2024-04-01).
8. Select a bounding box on the map.
   1. Click on the rectangle below the plus and minus buttons.
   2. Select a bounding box which is smaller than 2500px * 2500px. This restriction is made by the Sentinel Hub API.
      1. If your bounding box is too big, the system will tell you.
      2. Then click on the "Edit" icon and rearrange your bounding box or click the "Trash" icon and delete the bounding box.
   3. Example:
      
9. Submit the form.

DataHub Web App

Description

In the "DEWS DataHub - Create SatData Entry" view you can upload official satellite dataset archives that were downloaded from the Copernicus Browser.

After you have downloaded a satellite dataset and upload the ZIP archive or after you have requested satellite data images via the API, the system will process the received data.

So far the system shows the following data:

• Satellite mission
• Product type
• Product start time
• Product stop time
• Archive path in filesystem (if available)
  • e.g. /dews/media/sat_data/archive/<archive>
• Extracted archive path
  • e.g. /dews/media/sat_data/extracted/<mission>/<archive_name>
• Metadata files (if available)
  • manifest.safe, MTD_*.xml, INSPIRE.xml and/or xfdumanifest.xml (if available)
• Thumbnail image (if available)
• Band table names in PostGIS database
  • e.g. 3ed72523-fa4c-447b-b5db-19020d17a7ce_b02_r10m
• Calculated metrics
  • e.g. NDVI, SMI, ...

Metrics calculation

Before the Copernicus Browser had the feature to directly download the True Color, NDVI, SMI, SWIR, etc. layers by selecting a bounding box (see this Copernicus Browser example), the development of the "DEWS DataHub" begun. During the development Copernicus Browser released its feature and put the metrics calculation of the DEWS-DH a little in the shade.

Still the calculation of several metrics is one of the main benefits of the server-side orientation DEWS-DH.

For example to calculate the NDVI (Normalized Difference Vegetation Index), which measures the difference between near-infrared and red light to quantify vegetation health, the band 4 (red) and band 8 (near-infrared) are required.

The .tiff or .jp2 files are loaded as datasets using the Python library rasterio. After that the mathematical equation $\frac{(NIR-Red)}{(NIR-Red)}/$ is applied and the result is turned into a plot image (see Code: dews/sat_data/services/metrics_calc.py). This procedure is applied for all indices that can be calculated.

Example of NDVI image:

Since the Sentinel Hub API does not return a RGB image, it must be calculated by the system. In the context of the DEWS-DH RGB is considered as an index, too.

Example of RGB image:

In the future development the result should be saved as .tiff or .jp2 file.

Satellite data aggregation

Satellite data entries with the same coordinates will be collected and saved to a TimeTravel object which you can find at the "DEWS DataHub - Overview" view's TimeTravel tab.

In a TimeTravel entry you can see the observed area with details about the related SatData entries. The TimeTravel object is linked to the SatData object and vice versa. It is a One-To-Many relationship, so the TimeTravel entry can have multiple SatData entries.

PostGIS

PostGIS database

• Location: 0.0.0.0:5432
• Imported archive's images will be imported as a raster in an own table.
  • The name consists of the sat data id, the band (e.g. B04 which stands for the red band) and the range (meters per pixel), if available.
  • Table name convention: <sat_data_id>_<band> / <sat_data_id>_<band>_<range>
  • Example table name: 3ed72523-fa4c-447b-b5db-19020d17a7ce_b02_r10m

QGIS

Import bands/raster from DEWS-DH into QGIS

1. Visit the "DEWS DataHub - Dashboard" view.
2. Select a satellite data entry whose bands/raster you want to import into QGIS.
3. Scroll down to the section Band Tables.
4. Copy the table name containing the raster you want to import.
   • e.g. 3ed72523-fa4c-447b-b5db-19020d17a7ce_b02_r10m
5. Open QGIS.
6. Select Layer > Add Layer > Add Raster Layer.
7. Select PostgreSQL on the left side panel.
8. Add the PostGIS database.
   • Database: dews
   • User: dews
   • Password: dews
9. You can use the Search options option in the bottom and copy the copied table name or search the table manually.
   • INFO: All tables are located in the public schema!

Supported satellite data sources

Copernicus Browser (Dataspace Copernicus)

• Link: Copernicus Browser

Satellite Type	Instrument	Data Level	Archive Naming Convention	Support Status	Additional Info
Sentinel-1A	C-SAR	Level-0 RAW	S1A_IW_RAW_0SDV_(...).SAFE.zip	✅
		Level-1 SLC	S1A_IW_SLC__1SDV_(...).SAFE.zip	✅
		Level-1 GRD	S1A_S1A_IW_GRDH_1SDV_(...).SAFE.zip	✅
		Level-1 GRD COG	S1A_IW_GRDH_1SDV_(...).SAFE.zip	✅	.dat files
		Level-2 OCN	S1A_IW_OCN_2SDV_(...).SAFE.zip	✅	.nc files
Sentinel-1B			❓	❌	No data found at Copernicus Browser.

Satellite Type	Instrument	Data Level	Archive Naming Convention	Support Status	Additional Info
Sentinel-2A	MSI L1C		S2A_MSIL1C_(...).SAFE.zip	✅
	MSI L2A		S2A_MSIL2A_(...).SAFE.zip	✅
Sentinel-2B	MSI L1C		S2B_MSIL1C_(...).SAFE.zip	✅
	MSI L2A		S2B_MSIL2A_(...).SAFE.zip	✅

Satellite Type	Instrument	Data Level	Archive Naming Convention	Support Status	Additional Info
Sentinel-3A	OLCI	Level-1 EFR	S3A_OL_1_EFR____(...).SEN3.zip	✅	.nc files
		Level-1 ERR	S3A_OL_1_ERR____(...).SEN3.zip	✅	.nc files
		Level-2 LFR	S3A_OL_2_LFR____(...).SEN3.zip	✅	.nc files
		Level-2 LRR	S3A_OL_2_LRR____(...).SEN3.zip	✅	.nc files
		Level-2 WFR	S3A_OL_2_WFR____(...).SEN3.zip	✅	.nc files
		Level-2 WRR	S3A_OL_2_WRR____(...).SEN3.zip	✅	.nc files
	SRAL	Level-1 SRA	❓	❌	No data found at Copernicus Browser.
		Level-1 SRA_A	❓	❌	No data found at Copernicus Browser.
		Level-1 SRA_BS	❓	❌	No data found at Copernicus Browser.
		Level-2 LAN	❓	❌	No data found at Copernicus Browser.
		Level-2 WAT	❓	❌	No data found at Copernicus Browser.
		Level-2 LAN_HY	❓	❌	No data found at Copernicus Browser.
		Level-2 LAN_SI	❓	❌	No data found at Copernicus Browser.
		Level-2 LAN_LI	❓	❌	No data found at Copernicus Browser.
	SLSTR	Level-1 RBT	S3A_SL_1_RBT____(...).SEN3.zip	✅	.nc files
		Level-2 AOD	S3A_SL_2_AOD____(...).SEN3.zip	✅	.nc files
		Level-2 FRP	S3A_SL_2_FRP____(...).SEN3.zip	✅	.nc files
		Level-2 LST	S3A_SL_2_LST____(...).SEN3.zip	✅	.nc files
		Level-2 WST	S3A_SL_2_WST____(...).SEN3.zip	✅	.nc files
	SYNERGY	Level-2 SY_SYN	S3A_SY_2_SYN____(...).SEN3.zip	✅	.nc files
		Level-2 SY_V10	S3A_SY_2_V10____(...).SEN3.zip	✅	.nc files
		Level-2 SY_VGP	S3A_SY_2_VGP____(...).SEN3.zip	✅	.nc files
Sentinel-3B	AOD		S3B_SY_2_AOD____(...).SEN3.zip	✅	.nc files
	AUX		S3B_SR___ROE_AX____(...).SEN3.zip	✅	.nc files
	OLCI	Level-1 ERR	S3B_OL_1_ERR____(...).SEN3.zip	✅	.nc files
	SLSTR	Level-2 LST	S3B_SL_2_LST____(...).SEN3.zip	✅	.nc files
	SYNERGY	Level-2-SY-VG1	S3B_SY_2_VG1____(...).SEN3.zip	✅	.nc files
		Level-2-SY-VGP	S3B_SY_2_VGP____(...).SEN3.zip	✅	.nc files

Satellite Type	Instrument	Data Level	Archive Naming Convention	Support Status	Additional Info
Sentinel-5P			❓	❌	Work in progress, but no priority yet because this satellite's main observation goal is air pollution.

Satellite Type	Instrument	Data Level	Archive Naming Convention	Support Status	Additional Info
Sentinel-6			❓	❌	Not available at Copernicus Browser.