Design like LEGO. Simulate like a twin.
StormNET turns urban water infrastructure design into a live modeling process: place components, simulate response, visualize consequences, estimate cost, and refine decisions in real time. Performance and economics are evaluated together while the design is still flexible.
It sits between natural and built environments — rainfall, runoff, soils, terrain, pipes, channels, storage, controls, flooding, green infrastructure, and infrastructure economics become one interactive digital twin. This is especially powerful for large networked systems where alternatives multiply quickly.
See StormNET bring urban water design into the live modeling loop.
The flipbook is where users see design become simulation: place infrastructure, visualize the digital twin, test hydrology and hydraulics, evaluate LIDs, inspect flood response, view budgets, and connect design choices to cost and decision consequences in real time.
Place infrastructure. See hydraulics and cost respond. Steer.
Five streams. One infrastructure. One design problem.
Urban water systems are tightly coupled — and the coupling crosses water-stream boundaries. Stormwater, sanitary sewer, drinking-water distribution, rainwater harvesting, and treated reuse share infrastructure, share storage, share pumps, and share the same fundamental hydraulics. The SWMM solver doesn't differentiate between stream types; the architectural reality of urban water is that they are one system. A bioretention cell's soil depth changes runoff timing, which changes pipe surcharge risk, which changes pond sizing, which changes excavation cost, which changes whether the project is feasible. A rainwater harvesting cistern reduces both drainage demand (a flood hazard) and drinking-water demand (a resource) — the same water, two systems. Stormwater entering combined sewers becomes sanitary load. Treated reuse cascades from higher-quality to lower-quality applications, then recharges groundwater that becomes drinking water again. The hazard-to-resource transformation — treating stormwater as a resource for capture, treatment, and reuse rather than only as a flood hazard to drain — is one expression of how tightly the streams are coupled. Hydrology, hydraulics, green infrastructure, gray infrastructure, distribution, storage, pumps, controls, treatment, and economics are not separate design problems. They are one design problem — multi-dimensional constraint satisfaction across hydraulic performance, regulatory compliance, resilience, economic feasibility, and constructability.
From placement to consequences in under a minute. One steering loop.
For that, the user draws and places on a data-enabled, georeferenced landscape — polygons for subcatchments, LIDs (Low Impact Development practices — also known as SuDS, WSUD, or sponge-city infrastructure across different regions), and irregular ponds; objects for pipes, channels, sanitary sewers, stormwater drains, water-distribution pipes for harvesting and reuse, manholes, catch basins, and parameterized storage units (surface ponds, underground cube/conduit/vault systems, above-ground storage). From that single act of placement, three peer engines compute in parallel: the EPA SWMM hydrology/hydraulics/water-quality engine, the 3D CAD digital twin engine, and the physics-based cost engine. Multi-mode visualization updates in seconds — 1D time series, 2D map view, Sankey water-balance, dynamic profile with synchronized time slider, 3D CAD immersive (pre- and post-simulation), and hierarchical cost view. A site-scale model simulates in roughly 30 seconds; the full turnaround from a design change to seeing the consequences across all three engines and all visualization modes is typically under a minute. Concept to design to model to economics to visualization, in one cycle. Customizable at every level. Steerable in real time.
This system didn't exist. StormNET is it, anchored by EPA SWMM.
Such a system does not exist as established practice today. The components exist — SWMM-based hydraulic modeling, CAD packages, cost estimation modules, GIS overlays, watershed-scale runoff models, groundwater models, and the data infrastructure built by national agencies and the broader scientific community around the world over decades — often at high quality, each within its own tool. What's not yet established is the architectural integration: one model, one set of inputs, three peer engines computing in parallel, multi-mode visualization in seconds, sub-minute iteration cycle. StormNET is that integrated computational steering system. Computational steering is an established concept in scientific computing — the human directly controlling a running computation, observing and steering it as it evolves rather than waiting for batch results. StormNET applies that principle to urban water infrastructure design, where the loop closes between iterations every ~30 seconds rather than between project phases.
The model runs from the first placed object because the platform engineers intelligent defaults that keep the user in decision space, not configuration space — sensible starting conditions for infiltration, routing, surface properties, regional pricing, and management assumptions, so each design increment produces a convergent simulation immediately. Every default is fully editable; the user refines for accuracy where it matters. The user steers across six dimensions: the iteration loop itself, the scenarios being tested, the conceptual model (what's placed where), the numerical model (routing scheme, infiltration method), the cost framework (258 regional presets, materials, labor assumptions), and the visualization itself. The result is something different in kind from running a simulation — the user is inside the iteration loop, designing incrementally while watching hydraulics, visualization, and cost respond together.
Standing on the work of generations. Built for the design space cities actually face.
The urban water modeling foundation already exists — built over decades by national meteorological agencies, soil surveys, hydrologic surveys, and flood mapping programs around the world, alongside open-source numerical engines that have matured into industry standards. EPA SWMM anchors the hydraulic engine: rigorously refined hydrologic, hydraulic, and water-quality methods, validated by the urban water community globally. The data side is similarly mature — national infrastructure (NOAA, NRCS, FEMA, USGS in the US, with parallel infrastructure built by national meteorological, soil, hydrologic, and flood-mapping agencies in many countries) alongside global open-data programs that have made high-quality terrain, soil, and climate products freely available everywhere. The numerical methods matured — kinematic wave, dynamic wave, Manning's roughness, Curve Number, Green-Ampt, full St. Venant equations — all rigorously validated, open-source, in the hands of every urban water modeler, everywhere.
Data-enabled where it matters. Engineered defaults where it speeds you up.
Urban water modeling is fundamentally data-tied — every model parameter has a physical referent in terrain, soils, land cover, infrastructure, climate, or monitoring networks. StormNET assembles all of these at the resolutions urban projects actually use, and pairs them with engineered defaults that converge from the first placed object.
Place a pipe anywhere in an urban project area and StormNET starts from context, not from a blank page.
Do it once, for all
The data preprocessing happens once — at the resolutions urban projects use. No individual project re-ingests SSURGO, re-classifies imagery, or rebuilds the climate stack. The architecture frees every project from work that should never have to be repeated.
DataNET extends, not replaces
The global preprocessed base covers typical urban water workflows. DataNET is the enrichment layer: when local-area data improves on global products — high-resolution LiDAR, project-specific monitoring, municipal GIS — it federates the better data in, dynamically.
Where the bottleneck moves
Urban infrastructure design is combinatorial: pipe sizes × routes × storage locations × LID placements × control rules × phasing. Conventional workflows test a handful of alternatives because each iteration takes weeks. StormNET makes iteration take 30 seconds. The bottleneck moves from setup to the design question that actually matters.
Two foundational layers, then incremental design.
The architecture rests on two foundational layers loaded before the first object is placed, and an incremental design workflow where each placement grows a working model. Add another object, the system grows. Each placement is simultaneously a design move and a model update; the simulation, the 3D twin, and the cost recompute against the new state in ~30 seconds.
The steering system rests on two foundational layers.
Both available before the first object is placed; both editable, both extensible through DataNET:
Spatial fabric
Terrain, land cover, soils, drainage assets — the geometry the design rests on. The SWMM engine reads curve number, infiltration, roughness, and slope from this layer automatically.
Stress framework
The forcing the urban water system responds to — design storms, continuous rainfall, snowmelt, evaporation, plus management inputs and external inflows.
The foundation evolves — refreshed as agencies update, extended through DataNET, deepened with each release.
Plus the engineered defaults that keep the user in decision space.
The platform sets default routing scheme (dynamic wave for hydraulic accuracy), default infiltration method (modified Horton or Green-Ampt depending on context), default surface roughness from land cover, default subcatchment widths and slopes from terrain, default LID layer thicknesses, and default management assumptions. Where engineering trade-offs exist between numerical robustness and parameter accuracy, the platform chooses robustness — robustness for the first placed object; accuracy for the engineer's refinement.
Design incrementally. The model grows with you.
The design-as-model paradigm is a different methodological commitment from batch simulation. Place a few pipes — see the hydraulic response and the cost. Add a pond — see the response shift. Place LIDs — see runoff reduce and costs adjust. The user is always inside a working model, even when the design is incomplete. The 30-second simulation cycle makes this practical: each increment costs roughly half a minute of machine time, which is short enough that the user keeps designing rather than waiting. This is the methodological commitment that defines the design-as-model paradigm.
Four moves inside the iteration loop.
Explore the site
Virtual site visit: high-resolution terrain, LiDAR, soils, land use, streets, hydrography, surrounding built context — all available before placing a single object.
Place infrastructure
Subcatchments, inlets, manholes, conduits, channels, ponds, culverts, pumps, weirs, LID features — placed directly on the data-enabled landscape, like building with modular objects.
Simulate and visualize
EPA SWMM-based hydrologic + hydraulic + water quality simulation runs in ~30 seconds. Water dynamics, flood extents, and cost estimates all appear inside the same platform, visible across five visualization modes.
Refine and iterate
Add infrastructure, resize, reposition, change LID combinations, swap routing scheme, adjust cost assumptions — each increment reruns and the multi-mode view updates together.
A single integrated coupled modeling system. One model. Many expressions.
The architectural distinctive is the coupling. Three peer engines read the same set of model inputs and compute together on every design change — hydraulics, the 3D digital twin, and physics-based cost. The same model is then visible across six visualization modes, each answering a different design question from the same simulation results. Together, the coupling and the multi-mode legibility are what make implausibility — hydraulic, visual, or economic — immediately visible.
One set of inputs. Multi-dimensional outputs.
Every design object the user places contributes to a single set of StormNET model inputs — dimensions, materials, slopes, layer thicknesses, control rules, land-use assignments. From those same inputs, three peer engines compute in parallel:
EPA SWMM engine
The coupled hydrologic, hydraulic, and water-quality engine. Hydrology: rainfall, infiltration, snowmelt, ET, subcatchment runoff. Hydraulics: kinematic and dynamic wave routing through conduits, channels, storage, and structures. Water quality: pollutant buildup over land uses, washoff during storm events, BMP and LID removal, treatment at storage nodes, routing through the drainage system — similar in concept to how SwaNET generates nonpoint-source loadings at basin scale.
3D CAD digital twin engine
Produces the immersive visualization from the same model inputs. Pre-simulation: buildings, pipes, channels, storage, hydraulic structures, terrain textures, surface context — the design as built. Post-simulation: water dynamics overlay — pond rising, manholes surcharging, LID infiltration visible, flood propagation animated through the storm event.
Physics-based cost engine
Reads the same model inputs as the hydraulic engine: pipe diameter × trench depth × material × regional unit price; pond volume × excavation × liner; LID layer thickness × media cost × planting × maintenance. Across 258 regional presets and 29 cost components, the same engineering physics adapts to different economies. Hierarchical, location-aware cost presentation at four levels — project total, system/zone, component group, and individual items with detailed justification — fully customizable at every level. Cost is not a black box; the user can drill down to see exactly how any number was computed.
The three engines read the same inputs and recompute together on every design change — the no-disconnect property. Any design choice surfaces its consequences in the same iteration cycle: a surcharging pipe, a manhole spilling onto a sidewalk, a cost number that breaks the budget. Real-time steering becomes possible — the user navigates multi-dimensional constraint satisfaction rather than ricocheting between disconnected tools.
Six views of the same model. The user chooses which view answers their current question.
Every visualization mode is fed from the same simulation results — there is no export, no replot, no separate tool. The user moves between modes to read different aspects of the same urban water system:
1D — Time series
Hydrographs at outfalls, water levels at nodes, velocities in conduits, flows through LIDs and storage — read as functions of time at specific points in the network.
2D — Map view
Animated through time: flooding nodes, surcharging manholes, surface ponding, velocities, water levels color-coded across the entire network — the spatial story of how the storm propagates.
Sankey — Water balance
Whole-simulation totals: volumes in, out, and stored at each node, plus LID-specific threads (rain gardens, green roofs, pervious pavers, infiltration trenches). The scenario-comparison instrument.
Dynamic profile
One path through the system: DEM + infrastructure + water levels animated with a time slider, synchronized with the rain intensity chart. Hydraulics legibility at a glance.
3D CAD immersive
Pre-simulation: inspect the design as built — buildings, pipes, channels, storage, textures, terrain. Post-simulation: water dynamics overlay — pond rising, manholes surcharging, LID infiltration. The digital twin in two modes.
Cost view
Physics-based estimate from model inputs — pipe diameter, trench depth, pond volume, LID layer thickness, pump curve — multiplied by regional unit costs across 258 regions. Hierarchical at four levels: project total → system/zone → component group → individual items with detailed justification. Drill down to see exactly how any number was computed; customize at any level. The economics surface of the same design.
Six scales of steering. Implausibility immediately visible.
This section shows what the architecture distinctively enables — steering happens at six different scales, and because hydraulics and cost recompute together from the same model inputs, implausibility (hydraulic or economic) becomes immediately visible. Each property, numerical, conceptual, or cost change triggers both hydraulic and cost recompute together because both engines read the same model inputs.
What's shown, how it's read, what drives it, how it's computed, what's designed, and what it costs.
The six scales correspond to six dimensions the engineer adjusts: display, 3D twin perspective, parameters, numerical methods, infrastructure layout, and cost framework. Each adjustment cascades through hydraulics and cost together.
Display steering
Which visualization mode to read. Toggle between 1D time series, 2D map view, Sankey water-balance, dynamic profile with time slider, 3D CAD immersive, and cost view — without rerunning the simulation. Fast — no big-data operations triggered.
3D digital twin steering
Spatially explicit understanding in two modes. Pre-simulation: rotate, tilt, and inspect the design as built — buildings, pipes, channels, storage, hydraulic structures, terrain textures, surface context. Post-simulation: the same environment with water dynamics overlaid — ponds rising, manholes surcharging, LID infiltration, flood propagation. The user navigates the digital twin to read the design and the response — fast, no big-data operations triggered, just diagnostic attention.
Property steering
Parameter values. Pipe diameter, conduit roughness, pond depth, LID layer thickness, infiltration parameters, control rules — change them, the platform reruns the SWMM simulation against the existing design and the cost engine reruns against the same inputs. Hydraulic and cost consequences appear together. Fast — no fabric operations, just the coupled rerun in ~30 seconds.
Numerical steering
Method choices. Switch routing scheme (kinematic wave ↔ full dynamic wave), infiltration method (Modified Horton ↔ Green-Ampt ↔ Curve Number), surface routing approach. The platform reruns the simulation with the new methods against the existing design. Fast — coupled rerun, no fabric operations.
Conceptual steering
The infrastructure itself. Add a pipe, move a manhole, reroute a conduit, place LIDs, resize a pond, add a control structure, change phasing. The platform updates the simulation model and the cost model directly from the new design. You build incrementally — the model grows with each placed object, not in a batch at the end.
Cost steering
The economic framework. Change regional preset (258 regions available), adjust material costs, update labor assumptions, override unit prices with project-specific bids, tune lifecycle parameters. The same engineering physics adapts to different economies — the design doesn't change, but the cost reads differently. Navigate the cost at four hierarchical levels — project total, system/zone, component group, individual items with detailed justification — to identify which assumptions matter, collect better local data, and refine for the next planning phase.
The complete coupling, all in view together.
The differentiator is not any single instrument — it is the complete coupling. One can easily come up with a design that is hydraulically reasonable but impossible to implement in construction or energy cost — and that will show immediately in the multi-mode environment. Hydraulic performance, water balance, visual realism, and economic feasibility are read together because they all derive from the same model inputs. The user catches implausible designs in seconds rather than discovering them in months.
infiltration · LID · storage · sewer routing · flooding · treatment
The Sankey shows the whole-simulation water balance: volumes in, out, and stored at each node, plus separate threads for each LID type (rain gardens, green roofs, pervious pavers, infiltration trenches). It changes between scenarios — different storms, designs, or LID combinations reshape the flow distribution. That is how a designer compares alternatives.
Tight coupling, immediately legible.
The three peer engines reading one set of inputs make the consequence direct: every Property, Numerical, or Conceptual change triggers all three reruns together. A design that is hydraulically reasonable but economically absurd, or visually plausible but hydraulically broken, shows it immediately — not after months of separate engineering. The coupling is architectural, not a project-specific custom build.
Scenario response visible
Change the storm event, the design, or the LID combination — the Sankey rebalances, the flood map shifts, the cost adjusts, all at once. The combinatorial design space becomes navigable rather than overwhelming. Conventional workflows test a handful of alternatives; StormNET enables searching the space.
Implausibility — hydraulic and economic — shows immediately.
Real-time multi-mode visualization plus tight model+cost coupling means the user cannot escape the consequences of their design choices. Place a pipe that surcharges under a 10-year storm — the 2D map shows the flooding, the profile shows the hydraulic grade line above the rim, the 3D twin shows the manhole spilling. Choose a regional preset that triples the construction cost — the cost view shows it before the user finishes placing the next LID. When cost is evaluated late, design options are already constrained — oversizing accumulates, green-gray tradeoffs go unexplored, alternatives that would have been better never get tested. StormNET puts cost in the design loop as a peer model: every increment shows its economic consequences immediately, while the design is still flexible. The user shifts from sequential engineering to continuous decision-making — and for large urban infrastructure, that typically translates into lower CAPEX, reduced OPEX, and improved lifecycle performance.
Machines compute. Humans decide.
Urban water design is fundamentally multi-dimensional constraint satisfaction — hydraulics across streams, regulations, resilience, economics, constructability. Conventional workflows force the designer to navigate these dimensions sequentially across separate tools, ricocheting between disconnected environments. When the model is tightly coupled — one set of inputs, three peer engines, six visualization modes — the designer can do something different: navigate the design space looking for solutions that satisfy multiple constraints simultaneously, because the tradeoffs are visible at every step rather than discovered late. Machines stay on machine work (SWMM, cost, 3D rendering, Sankey rebalancing); humans stay on human work (judgment, system intuition, weighing tradeoffs, stakeholder explanation). They connect through the 30-second iteration loop. The designer stays in the sense-making space.
Cities don't stand alone. Neither does your work.
StormNET is one of five expressions of the MAGNET4WATER architecture. SwaNET and StormNET are sister watershed platforms at complementary scales (rural-basin and urban). StormNET ⊃ ConduitNET is the sister pipe-network relationship — pipe-network mathematical superset. IGW-NET will couple to StormNET next release, providing water-table state to subcatchments and to the cost model. DataNET federates additional context worldwide, alongside the always-on preprocessed multi-resolution database that direct-links StormNET to essential layers. The subsections below cover each relationship; the homepage Cross-Platform Guide documents the full family. And — when you choose — your work joins the Global Model Network and the Model Observatory, where your published work becomes the foundation for the next person's work.
Two watershed modeling platforms, complementary scope and physics.
SwaNET and StormNET are both watershed modeling platforms, related architecturally but with no model-to-model data flow between them. They share the same dimensional architecture: hydraulic state is lumped (0D + time) within natural or engineered cells; routing connects cells; flood depth and inundation are 2D derived from cell state minus DEM. They diverge in three dimensions: routing physics (SwaNET hydrology-based — Muskingum, variable-storage; StormNET full 1D unsteady Saint-Venant on a grid), storage element treatment (SwaNET lumped effective within subwatersheds; StormNET explicit per element for individual ponds, lakes, detention, retention), and subsurface depth (SwaNET multi-layer soil profile sophisticated enough for crop root uptake; StormNET simpler single vadose layer plus one aquifer term). LIDs in StormNET parallel SwaNET's BMP architecture — lumped within subcatchments as a percentage — with an added explicit site-scale treatment where the LID is the design element and is visualized in the 3D CAD twin. SwaNET applies at larger scales (river basins, rural basins, agricultural watersheds); StormNET applies primarily to urban systems. Pick the platform whose physics and scale match the dominant question.
Regional water-table state will arrive as aquifer field at subcatchments.
Urban drainage interacts with groundwater in two directions: shallow groundwater feeds infiltration/inflow into storm sewers and sanitary systems, and the depth-to-water-table controls how much infiltration LIDs can actually achieve. Coming next release, IGW-NET will provide water-table information to StormNET for two purposes: as initial water table at subcatchments — needed to compute groundwater flux to drainage nodes (urban groundwater interactions, baseflow contributions, infiltration); and as water-table depth for the cost model — driving excavation depth, trench dewatering, and infrastructure-cost adjustments. Critical in coastal cities and alluvial systems where the water table strongly affects both the storm response and the cost of building drainage. Same architectural pattern as the SwaNET-IGW-NET coupling: requires identical computational domain and projection system on both sides, with the projection chosen from the intersection of IGW-NET's comprehensive support and StormNET's narrower selection — typically a regional UTM zone.
A mathematical superset for pipe networks. Choose the right platform from the start.
StormNET and ConduitNET overlap in pressurized pipe-network modeling, with no model-to-model data flow between them. The relationship is a mathematical superset: ConduitNET is a pure pipe-network hydraulics platform (circular pipes, always pressurized, pumps, tanks, valves, demand patterns, control rules) built on the EPA EPANET engine — an efficient way to solve Saint-Venant + tank water balance for pressurized pipe networks. Storage change in pipes is ignored because it's negligibly small on distribution timescales; storage change in tanks is integrated transiently. The system itself is fully transient: demand patterns evolve, tank levels respond, pumps cycle, water quality advances. StormNET handles a broader regime — pressurized AND free-surface AND transitions between them — with 24 standard cross-section shape types (EPA SWMM 5 User's Manual Table 3-1) plus user-defined transects and shape curves, full 1D unsteady Saint-Venant. The relationship is the EPANET engine = the SWMM engine with storage change in pipes ignored — negligible error on distribution timescales.
There is no file handoff between the two platforms. Large-scale water distribution and long-distance transfers often involve mixed networks — pressurized pipes in complex terrain combined with free-surface gravity sections where topography permits (cost-effective and energy-efficient), partly-full conduits, regime transitions where pipes surcharge or unsurcharge, open channels into reservoir intakes, harvested-water inflows. For these mixed systems, build directly in StormNET from the start — its 1D unsteady Saint-Venant solver handles pressurized and free-surface flow in one framework, including the transitions between them. Use ConduitNET when the network is purely pressurized. Architectural principle: don't reach for the more general solver when the simpler one is correct, and don't try to migrate when you can pick the right platform at the start.
Scope boundary — water hammer. Both ConduitNET and StormNET assume incompressible fluid; neither models water hammer. Water hammer is a different physics regime — fluid compressibility, pipe elasticity, millisecond timescales (the method of characteristics on the elastic-water-column equations) — and is typically handled in real systems through surge-protection devices (surge tanks, air chambers, pressure-relief valves, slow-closing valve operation, soft-start pumps) rather than through network simulation. When explicit water-hammer simulation is required (e.g., for surge-protection sizing), specialized water-hammer tools are appropriate.
One architecture. Five expressions.
DataNET FEDERATED ROUTE (next release)
The federated route to worldwide water-data services, alongside the always-on preprocessed multi-resolution database that already direct-links StormNET to essential layers (terrain, base drainage, base imperviousness). Direct in-platform DataNET integration for StormNET coming next release — municipal GIS, regional climate downscalings, sub-1m LiDAR DEM, regional IDFs (GSDR-IDF, ECCC IDF), and project-specific overrides federated dynamically. The two data routes are complementary by design: the preprocessed database for the essentials (working model in minutes); DataNET for everything else.
Explore DataNET →SwaNET SISTER PLATFORM
Watershed modeling at larger scales — river basins, rural watersheds, agricultural watersheds. Sister watershed platform of StormNET (shared dimensional architecture, no model-to-model data flow). SwaNET applies multi-layer soil profile sophisticated enough for crop root uptake; StormNET applies full 1D unsteady Saint-Venant at urban scale.
Explore SwaNET →IGW-NET COUPLING (next release)
3D groundwater modeling. Coming next release: IGW-NET will provide water-table information to StormNET for two purposes — as initial water table at subcatchments (for groundwater flux to drainage nodes, baseflow contributions, urban groundwater interactions) and as water-table depth for the cost model (excavation, trench dewatering, infrastructure-cost adjustments). Critical in coastal cities and alluvial systems.
Explore IGW-NET →ConduitNET SISTER PLATFORM
Pure pipe-network hydraulics platform for pressurized supply networks across different scales. Sister pipe-network platform of StormNET — mathematical superset/subset relationship; right equations for the pressurized regime. StormNET handles the broader regime (free-surface, partly-full, regime transitions) when needed, but build directly in StormNET from the start for mixed systems. No file handoff between the two. Both share a physics-based, bottom-up, location-aware, water-table-aware cost-model architecture; StormNET extends cost coverage to LIDs, subcatchment-scale elements, storage units, and hydraulic structures.
Explore ConduitNET →Cost model deep dive
Detailed cost model documentation: 29 components, 258 regional presets, layer-by-layer LID costs, interface-tunable assumptions, lifecycle cost approach.
Open cost model →Your work joins the network.
Your StormNET designs can stay private to your project, shared with your team, or published to the Global Model Network. Published designs become accessible to other urban water practitioners and to students learning the craft. Local-area data served from your own GeoServer node — visualized in DataNET by URL, transferred directly to any modeling platform — stays on your own infrastructure with access under your control. This is how infrastructure grows: every new participant who shares, makes the network slightly richer for everyone else.
Model Observatory
Publish your StormNET design to the Observatory in under a minute — a feature shared across all five platforms. Your work appears as a pin on a global map, styled by model type.
Click any pin → a local Observatory page opens with:
- Your design graphics and downloadable model file
- An AI-generated report grounded in the model file (rainfall scenarios, network configuration, default parameters, performance)
- A live USGS National Water Network overlay — urban gauges, streamflow, water-quality time series — as statistics, historical series, and realtime
Private or public — your choice. The USGS overlay appears on every Observatory page regardless of platform.
On the roadmap: auto-generated DataNET fusion, Cesium 3D; direct in-platform LiDAR linking coming next release.
Browse the Observatory →Education and design competition
Real-time cost makes urban water design teachable as engineering. Students design complete systems, verify hydraulic performance, test green infrastructure, and compete on lifecycle cost — meeting explicit regulatory and resilience constraints.
View case studies →Urban water design becomes continuous decision-making.
The components exist — extraordinary infrastructure built over decades by national agencies and the broader scientific community across the world. What's not yet established as practice is the architectural integration across the full urban water system — stormwater, sanitary, drinking-water distribution, rainwater harvesting, and treated reuse. StormNET is that integration. Urban water design becomes a continuous, multi-mode, cost-aware discovery process — one that honors the work the field has already done, makes implausibility visible immediately, and lets your own work join the network.