Sentinel-1 flows¶

The Sentinel-1 processing available in RS-Python is:

Processing	How to run it	Processor (rs-dpr-service process)
Level-0	`process-l0` or `process-s1-l0` flows, see Level-0 flows. The session is the only input product (`S1CADUS`). In systematic processing, it is started automatically after the staging of the session, see Systematic processing.	`s1_l0`
Analysis Ready Data (ARD)	`process-s1-ard-reference` then `process-s1-ard` flows, see Analysis Ready Data. They can also be run with the `process-generic` flow, with the `s1_ard` processor and one of the pipelines of its task table, see Customise the DPR processing.	`s1_ard`

The default settings of the Sentinel-1 flows are read from the s1l0-default-setting (Level-0) and s1ard-default-setting (ARD) Prefect variables, see Customise the DPR processing.

The task tables of these processors, with their pipelines, inputs and auxiliary files, are in the rs-dpr-service config folder (TaskTable_S1_L0_*.json, TaskTable_S1_ARD_*.json). The Sentinel-1 ARD processor also needs a digital elevation model, read from the EarthDataHub of Destination Earth (edh_api_key parameter).

Analysis Ready Data (ARD)¶

The ARD processing generates products from Sentinel-1 SLC products (IW, EW or SM instrument mode), already converted to Zarr and published in the catalog. The pipelines of the s1_ard task table are:

Pipeline	Description	Output
`ARD_REFERENCE_PIPELINE`	Reference pipeline: computes the reference geometry of a master SLC and writes it in the `REFERENCE_DB`.	`REFERENCE_DB`
`SLC__1_CSL__1`	Co-registered SLC.	`cslcs`
`SLC__1_GSL__1`	Geocoded Single Look Complex.	`gslcs`
`SLC__1_NRB__1`	Normalised Radar Backscatter (mosaicked on the Sentinel-2 tiles, `S2_TILES` auxiliary file).	`nrb`

These pipelines are also available in the DprPipeline enum of rs_client (S1ARD_REF, S1ARD_CSL, S1ARD_GSL, S1ARD_NRB).

The flows are:

Flow	Parameters	Description
`process-s1-ard-reference`	`slc`, `edh_api_key`, `env_args`, `verbose`	Run once on a master SLC to generate the `REFERENCE_DB` used by the nominal flow, with the `ARD_REFERENCE_PIPELINE` pipeline.
`process-s1-ard`	`slcs`, `reference_date`, `edh_api_key`, `env_args`, `pipeline`, `verbose`	Nominal on-demand ARD processing of several SLCs, `reference_date` being the date of the master SLC, with the given `pipeline` (`SLC__1_CSL__1`, `SLC__1_GSL__1` or `SLC__1_NRB__1`), or by default the pipeline of the settings.

slc and slcs are STAC item identifiers of the SLC products, searched in the input_collections of the settings (default: s01sewslc, s01siwslc, s01ssmslc). All the SLCs must share the same sar:instrument_mode, which is given to the processor as the instrument_mode external variable.
reference_date is required when several SLCs are given. For the SLC__1_NRB__1 pipeline with a single SLC, it is deduced from the SLC identifier, to select the S2_TILES auxiliary file.
The ETAD and ORBIT auxiliary files are staged once per input SLC (folder_multiplicity: one_per_input in the task table), see Task table and payload.

The REFERENCE_DB is handled specifically: it is never searched nor staged. For the ARD_REFERENCE_PIPELINE, the flow allocates a new folder s3://<bucket>/<owner>/s1-ard-reference-db/<uuid> for it, the bucket being given by the bucket configuration for the owner, the s1-ard-reference-db collection and the ARD_REFERENCE_DB product type. Its path is written in the logs of the flow.