Roadmap¶

The Open Climate Service has delivered its foundation. An instance can ingest climate and Earth Observation data, keep it current, store it as cloud-native GeoZarr, make it discoverable through STAC, derive new products from it through openEO processes (including climate indices), and chain those steps into reusable workflows.

The focus now shifts from building the core platform to integrating it with the DHIS2 ecosystem, hardening it for production, and broadening the data and analysis it offers.

Delivered¶

Data ingestion and sync

Ingest climate and Earth Observation datasets for a configured spatial extent
Keep datasets current — append new periods and rematerialise releases on sync
Built-in dataset catalogue (CHIRPS, ERA5 / ERA5-Land, WorldPop)

GeoZarr storage and discovery

All datasets stored as cloud-native GeoZarr, with multiscale pyramids for efficient browser-based rendering
Local filesystem storage; S3-compatible and self-hosted (sovereign) object storage for cloud and country deployments
STAC catalogue plus raw Zarr access — direct xarray / stackstac use without API knowledge

Processing and analysis

openEO adopted as the process execution and chaining standard, with both synchronous and batch execution, backed by openeo-pg-parser-networkx and openeo-processes-dask (120+ standard processes)
Custom processes, including climate indices
Derived datasets materialised back into managed GeoZarr stores (e.g. change between two periods, population-weighted exposure)

Workflows

Named, reusable workflows that chain processes end to end — e.g. aggregate to org-unit boundaries, population-weighted aggregation, and change between periods
Workflow outputs published as managed datasets, discoverable like any other

Focus areas¶

These are the priorities for taking the Open Climate Service from a working platform to production-ready shared infrastructure for the DHIS2 climate and health ecosystem. They are largely parallel rather than strictly sequential.

DHIS2 integration¶

The highest-value next step is direct integration with the DHIS2 ecosystem so that climate and EO data flows into the tools country teams already use:

DHIS2 Climate App — consume Open Climate Service datasets and aggregation directly
DHIS2 Maps App — render GeoZarr layers and import aggregated values
CHAP Modelling Platform — supply harmonised climate inputs to models
Push aggregated values into DHIS2 data elements against org-unit hierarchies

Stability¶

Harden ingestion, sync, and job execution against partial failures and stale jobs
Replace the interim JSON-backed artifact store with a proper transactional store
Broaden automated test coverage, including integration tests against running instances

Security and authentication¶

Authentication and authorisation for the API
Per-instance access control for write and admin operations
Secure handling of upstream provider credentials

Performance and scalability¶

Faster ingestion and processing for large extents and long time ranges
Tune chunking, pyramids, and Dask execution for concurrent, long-running jobs
Scale-out execution for batch jobs

Automation and scheduling¶

Scheduled sync and recurring derived-product updates (e.g. monthly normals, weekly anomaly refreshes)
Event-driven cascades — a sync that updates a base dataset triggers downstream derived products automatically

More datasets, processing and workflows¶

Expand the built-in dataset catalogue — more sources, variables, and resolutions
More built-in analysis: climate normals, anomalies, exposure indices, and additional climate indices
More reusable workflows for common climate-and-health use cases

Building competence and partnerships¶

Build climate-data competence within HISP groups and country teams
Grow partnerships with data providers and the wider climate / EO and DHIS2 communities

Documentation, guides, and worked examples
Training materials and presentations for implementers and analysts