The Foundation.
Geospatial, location, Earth observation, climate, and regulatory data — fused for water-network risk.
Risk models are only as defensible as the data behind them. HydroDSS leans on the same multi-jurisdictional data platform that powers QuestFeed's planning intelligence (ZoneDSS) and regulatory intelligence (AuditDSS) products — extended for water utilities with live Earth observation, climate reanalysis, active-radar leak detection, and a regulatory DAG built from named, authoritative sources.
Geospatial · Location & BI · Earth observation · Regulatory DAG.
The foundation is organised in four layers that integrate through a shared spatial reference. Geospatial gives us where physical things are. Location & business intelligence gives us what they are and who uses them. Earth observation gives us what's happening around them. The regulatory DAG gives us what rules apply and what happens when they're broken.
Geospatial & cadastral
Authoritative cadastre, addresses, buildings, transport, zone boundaries. The base layer every other dataset references.
Location & business intelligence
What every building is and who uses it. Places, businesses, hospitals, schools, industrial sites, demographics, vulnerability.
Earth observation & climate
LIDAR topography, gridded climate, ERA5 reanalysis, InSAR ground deformation, SAR leak signatures, surface-water dynamics — live via Google Earth Engine.
Regulatory DAG
Regulations modelled as a directed graph of corpora, rules, obligations, dependencies, and enforcement. Powered by AuditDSS — 320 corpora, 21 jurisdictions.
Where every asset, every customer, every regulation actually is.
Spatial truth is the prerequisite for everything else. The geospatial layer integrates authoritative cadastre, addresses, buildings, transport, and zone boundaries into one PostGIS-aligned base. NSW, QLD, and VIC are live in production via ZoneDSS — 11.1M Australian lots — with New York City and San Francisco added as the first US coverage.
What this gives HydroDSS: the network topology layer for the engine (where every main, pump, valve, junction physically sits), the consequence-of-failure geography for risk and renewals scoring (whose property, whose service, whose neighbourhood is affected), and the development-pipeline context for demand forecasting (what's approved, where, when).
Not just where buildings are — what they are, and who uses them.
Knowing a building's footprint is geography. Knowing it's a hospital, a school, an industrial precinct, or a residential block of flats is intelligence. The location & business intelligence layer maps every place — Overture Places, Google Business Profile, ABS Census demographics, suburb-level safety statistics — and ties them together so every customer in the network is tagged by what it is and who it serves.
What this gives HydroDSS: consequence-weighted criticality. A hospital outage isn't a residential outage; an industrial-precinct disruption isn't a school disruption. Every risk, every alert, every renewals decision can be tied to actual customer impact — quantified in lives served, businesses affected, and revenue exposed — not just hydraulic state. The bedrock of regulator-defensible consequence-of-failure scoring.
What's happening around the network, observed from space and ground.
The agencies that publish these datasets — ECMWF, BoM, NASA, ESA, JRC — operate the world's best Earth-observation systems. We use each dataset as the agency releases it, kept current through live Google Earth Engine pipelines. Every value in HydroDSS traces back to a named, citable agency record — not an in-house model. That's what makes outputs defensible to a regulator on their own terms.
What this gives HydroDSS: elevation-aware hydraulics (LIDAR + DEM), climate-aware demand modelling (temperature, precipitation, drought indices), leak-detection signatures (SAR backscatter anomalies over saturated soil), subsidence-aware asset-failure risk (InSAR mm-scale displacement), and source-water availability (surface-water dynamics, evapotranspiration).
Every regulation, every dependency, every enforcement action — as a graph.
Built and operated by AuditDSS. Each regulatory corpus is parsed into a directed acyclic graph of rules and obligations, with dependency edges between them. A Bayesian layer overlays conditional-probability tables capturing how a failure at one obligation propagates to others. A dynamic layer updates from real enforcement evidence — penalties, fines, regulatory interventions.
What this gives HydroDSS: regulatory-defensibility scoring (which obligations a decision affects, with cross-corpus shadow risk), penalty-aware consequence modelling (real enforcement actions calibrate expected exposure), and AS/NZS 31000 alignment by construction. When HydroDSS surfaces a risk, the regulatory DAG says which rule it intersects and what's happened to others who missed it.
Two destinations. One foundation.
The four layers feed two distinct consumers: the risk layer (which treats data as evidence and outputs calibrated risk — our hydroBAG research line, on the roadmap) and the hydroEngine (which treats data as physical input and outputs hydraulic state). They share the same foundation but use it for different work.
The risk model consumes data as evidence.
- Geospatial → asset and network evidence: pipe material, soil, age, traffic load, building footprint
- Location & BI → consequence evidence: which customers are affected, how vulnerable they are, what their criticality tier is (hospital, school, industrial, residential)
- Earth observation → operational and environmental evidence: SAR leak signatures, InSAR subsidence, climate-driven demand, surface-water availability
- Regulatory DAG → compliance evidence: which obligations a failure breaches, what penalties have been imposed elsewhere
Each evidence stream runs in an independent reasoning path. Independence is enforced — no data point feeds two evidence sources. That's what makes the Bayesian fusion mathematically valid and the reasoning chain decomposable.
The solver consumes data as the network's reality.
- Geospatial → network topology: where mains run, what they connect, what they serve — derived from cadastre, buildings, transport, and DA approvals when the utility's model is incomplete
- Location & BI → demand profiles: customer-type-aware load curves (residential, commercial, industrial, critical) that drive the engine's demand allocation
- Earth observation → boundary conditions: elevation-driven hydraulic gradients (LIDAR), source levels (surface-water dynamics), demand-forcing climate (temperature, precipitation)
- Regulatory DAG → operating constraints: water-sharing plan boundaries, environmental flow obligations, pressure-compliance requirements
The engine doesn't infer — it simulates. Data here is treated as deterministic input. Any uncertainty in those inputs becomes an ensemble dimension propagated through the solver's vectorised batch axis.
The loop closes. The engine produces a hydraulic prediction. The risk layer observes the residual between that prediction and what the data actually shows. New evidence updates the calibrated posterior. The next prediction is sharper. Audit-grade traceability at every step. That real-time risk layer is on the roadmap, held to a research bar; the modelling platform and its data foundation stand on their own today.
Where the data is in production.
NSW · Queensland · Victoria
Full geospatial coverage in production via ZoneDSS — cadastre, addresses, buildings, transport, zone boundaries, planning instruments, DA records. Combined: 11.1M lots across the three states, plus DA and permit records and comprehensive natural-resource and environmental overlays.
New York City · San Francisco
The first US coverage — 1.08M tax lots and parcels with zoning, planning, and assessment data. NYC PLUTO tax lots and San Francisco assessor parcels, modelled into the same lot-intelligence schema as the Australian states.
New Zealand · UK · EU
Geospatial coverage follows the ZoneDSS expansion path. The Earth-observation layer is already global (Overture, ERA5, JRC GSW, Sentinel-1, MODIS, SMAP), and the regulatory DAG is already global — 21 jurisdictions live in AuditDSS, including 116 energy/utility and 66 environmental corpora across UK, US, EU, CA, and international standards bodies.