Who it's for
Insurers, reinsurers, and brokers quantifying flood exposure across South African property portfolios.
Coverage
Available for portfolios anywhere in South Africa. Locations in areas we have not yet modelled are flagged before anything is run.
Who it’s for
Underwriters and portfolio managers who need an auditable view of flood exposure — not a score of unknown provenance. You send us a property list; we return the losses, the assumptions, and the evidence.
What you get
- Property-level hazard attachment (depth at agreed return periods, exposure tier).
- Expected annual loss (AAL) per property and aggregated at any policy or region grouping you specify.
- Accumulation and concentration metrics at configurable aggregation levels (grid, catchment, municipality, postcode).
- Sensitivity tests — how results change under alternative hydrology assumptions or depth-damage curves.
- Results delivered as a report and accompanying data files (CSV/Parquet, and GIS extracts where relevant).
How it works
- Scope and portfolio intake. We agree the geography, return periods, and depth-damage assumptions. Portfolio data is ingested in the format you use — policy system export, CSV, or geocoded shapefile. We flag rows we could not geocode or for which we do not have hazard coverage, before we run anything.
- Hazard attachment. Each property is intersected with the hazard layers. Where a location falls outside the areas we have modelled, the output records that fact alongside the result; we do not silently fill gaps.
- Loss modelling. Depth at return period drives damage using agreed depth-damage curves (your in-house curve or a transparent blend). AAL is computed by integrating over the return-period set.
- Aggregation. Results roll up to any grouping you specify — policy, region, municipality, reinsurance treaty, custom.
- Review and iteration. You get one round of review and clarification. Re-runs under alternative assumptions are scoped as extensions.
Coverage
The quality of a portfolio result depends on the flood model behind each property, and the deliverable says so:
- Properties in areas we have modelled return their AAL with the model run, the terrain and rainfall data versions and the date attached.
- Properties in areas we have not yet modelled are reported as “not modelled”, never as zero. We can model those areas and re-run them.
A black-box score would hide this. We report the breakdown and let you decide how to treat any unmodelled slice.
How you receive it
Results come as a report and data files (CSV or Parquet, with GIS extracts where relevant). A portfolio can also be uploaded and scored on the Delta Hydro platform, or scored location by location through the API, and scores are refreshed as the model is updated. See how you receive the flood data.
Related consulting services
More flood risk intelligence
Start a conversation
Tell us about your project
Reply within one business day
