04_use_case.md

Use case example

Prerequisites

On top of the requirements in the first workshop, the following is needed:

• having an account on https://creodias.eu/
• installing OceanDatalab's SEAScope (installation instructions for Linux, MacOS and Windows are provided on the website).

Setup

Just like in the first workshop, let's run the database and main containers by running the following commands from the geospaas_workshops directory.

docker-compose up -d
docker-compose logs -f main

Once you see the line Quit the server with CONTROL-C., the containers are ready for use.

Data harvesting

We are going to work with Sentinel 3 OLCI and SLSTR data from Creodias. For convenience, the harvested data can be imported in the database using the following command:

docker-compose exec main python geospaas_project/manage.py loaddata use_case_sample.json

To truly harvest the data, you can use the following commands.

docker pull nansencenter/geospaas_harvesting:latest

docker run -d \
--name geospaas_workshops_harvesting \
-v "$(pwd)/resources/use_case/harvest.yml:/etc/harvest.yml" \
--network 'geospaas-workshops_default' \
-e 'GEOSPAAS_DB_HOST=geospaas-workshops_postgis_db_1' \
-e 'GEOSPAAS_DB_PORT=5432' \
-e 'GEOSPAAS_DB_NAME=geodjango' \
-e 'GEOSPAAS_DB_USER=geodjango' \
-e 'GEOSPAAS_DB_PASSWORD=geospaas123' \
-e "COPERNICUS_OPEN_HUB_PASSWORD=$COPERNICUS_OPEN_HUB_PASSWORD" \
nansencenter/geospaas_harvesting:latest \
-m geospaas_harvesting.harvest -c /etc/harvest.yml

docker logs -f geospaas_workshops_harvesting

Finding relevant data

Once the data is harvested, we can start looking for interesting datasets.

The following command opens a Django shell in the main container.

docker-compose exec main python geospaas_project/manage.py shell

The following code is an example of how to find relevant datasets.

from geospaas.catalog.models import Dataset

Dataset.objects.count()

olci_datasets = Dataset.objects.filter(source__instrument__short_name='OLCI')
slstr_datasets = Dataset.objects.filter(source__instrument__short_name='SLSTR')

olci_datasets.count()
slstr_datasets.count()

lofoten_polygon = 'POLYGON((-5 75, 20 75, 20 65, -5 65, -5 75))'

lofoten_olci_datasets = olci_datasets.filter(geographic_location__geometry__intersects=lofoten_polygon)
lofoten_slstr_datasets = slstr_datasets.filter(geographic_location__geometry__intersects=lofoten_polygon)

for od in lofoten_olci_datasets:
    for sd in lofoten_slstr_datasets:
        # Find datasets whose spatial and temporal coverage intersects
        if (od.time_coverage_start <= sd.time_coverage_end
                and od.time_coverage_end >= sd.time_coverage_start
                and od.geographic_location.geometry.intersects(sd.geographic_location.geometry)):
            intersection = od.geographic_location.geometry.intersection(sd.geographic_location.geometry)
            if (intersection.area / od.geographic_location.geometry.area >= 0.99
                    or intersection.area / sd.geographic_location.geometry.area >= 0.99):
                print(od.id, od.entry_id)
                print(sd.id, sd.entry_id)
                print()

Downloading the datasets

Let us download some of these datasets:

• go to https://finder.creodias.eu/
• log in if necessary
• enter the entry_id of the dataset you want in the "Product identifier or path" box
• click search
• click on the result in the list (there should be only one)
• click on the download button

Save the dataset files in the geospaas_workshops/resources/use_case/ folder.

Converting to IDF

Next, we need to convert them to the IDF format supported by SEAScope.

The following command opens a Python shell in a container where geospaas_processing is installed.

docker pull nansencenter/geospaas_processing_worker:latest

docker run -it --rm \
--name geospaas_workshops_conversion \
-v "$(pwd)/resources/use_case/:/var/use_case/" \
--network 'geospaas-workshops_default' \
-e 'GEOSPAAS_DB_HOST=geospaas-workshops_postgis_db_1' \
-e 'GEOSPAAS_DB_PORT=5432' \
-e 'GEOSPAAS_DB_NAME=geodjango' \
-e 'GEOSPAAS_DB_USER=geodjango' \
-e 'GEOSPAAS_DB_PASSWORD=geospaas123' \
--entrypoint python \
nansencenter/geospaas_processing_worker:latest

Here is the code to convert our two downloaded datasets to IDF.

from geospaas_processing.converters import IDFConversionManager
conversion_manager = IDFConversionManager('/var/use_case')
conversion_manager.convert(1795, 'S3A_OL_2_WFR____20200822T121500_20200822T121800_20200822T142255_0180_062_052_1620_MAR_O_NR_002.zip')
conversion_manager.convert(5472, 'S3A_SL_2_WST____20200822T121500_20200822T121800_20200822T143516_0180_062_052_1620_MAR_O_NR_003.zip')

After this, there should be new dataset folders in geospaas_workshops/resources/use_case/sentinel3_olci_l2_wfr and geospaas_workshops/resources/use_case/sentinel3_slstr_l2_wst.

Visualization in SEAScope

Let us copy these folders to the data/ directory of SEAScope and start SEAScope.

If you want to adjust the visualization parameters, you can modify the config.ini file in each collection directory. The available parameters are described in the SEAScope user manual.