Choose a source to load its public records. Agency inventories, dated sample reports, forecast fields and community observations are distinct. Missing/retired feeds remain explicit; no silent all-clear or synthetic replacement is supplied.
Bounded longitude/latitude coordinate plot, no basemap; not a navigation, flood-depth or evacuation-routing product. Source points/lines/polygons remain context, not inferred utility dependencies.
Automatic mode uses the available future NOAA forecast and official public terrain, then deduplicates source-defined Census geographic context. It screens planning exposure; it does not establish outages or customers losing service.
Automatic mode populates public context using the forecast or your selected scenario level. Public population is geographic context, not utility customers; treatment volume is published throughput context, not a spill.
Using embedded public context.
Choose Local scenario overrides above to apply these browser-only assumptions. Automatic mode does not use them.
Blank inputs stay not assessable. Values entered here are assumptions supplied in this browser; the console does not verify catchments, costs or grant eligibility. No inputs are stored or sent.
No FEMA offset is assumed. Eligibility, credits, insurance and the applicable declaration must be established independently; entered percentages are arithmetic scenarios, not reimbursement promises. FEMA Public Assistance eligibility guidance.
No example population, cost or funding fraction is assumed.
Datum: the 1.48 ft MLLW-to-NAVD88 offset is specific to NOAA gauge 8726520, tidal epoch 1983–2001, datum table accepted 6 Sep 2022. Applying its plane to all points is a sensitivity test, not a uniform bay-wide flood surface. NHC inundation above ground must not be substituted for MLLW water level.
Interpretation. Terrain at/below a level is not a hydrodynamic flood footprint or a station failure. Wastewater-generation arithmetic is not a forecast spill volume. Population is not measured catchment exposure. Repair allowances are not damage assessments. No illness, casualty, predicted fine or automatic boil-water/evacuation instruction is produced.
Station IDs, ownership, county classifications and basin labels come from the public register. The water-plane comparison uses dated USGS terrain; it is not equipment inundation or operating status. Select a record to review it on the map.
The default filter shows City-owned/maintained assets from the City’s public GIS register. The all-records filter also shows privately owned assets that the agency itself publishes; public provenance and ownership are separate attributes. They are separate from the fictional LS-### network. A current/enabled GIS record does not establish that a pump has power, is operating, or is safe to access.
What is not supplied. The agency register has no ground elevation; independently sourced USGS raster terrain estimates are identified separately where available. Surveyed equipment heights and verified directed pipe/flow connections are not present. Its real locations and evacuation zones do not support the synthetic surge-exposure counts or a station-failure forecast. The ten-metric response checklist offers separate synthetic and public-terrain scenario datasets. Terrain geometry can support screening proxies; real service weights, physical connections and operational consequences remain not assessable without verified inputs.
Pinellas County official Know Your Zone lookup for a real address · Current official county instructions. An evacuation-zone designation differs from a FEMA flood zone and is not an evacuation order.
Real public diagnostics use the exact selected water level, dated station terrain and source-defined Census footprints. Descriptive comparisons are not graded with synthetic calibration thresholds. Unsupported flow topology stays not assessable. ; real per-station failure/flow diagnostics require verified additional inputs.
Each parameter is linked to a decision that belongs at that stage of preparation or recovery. Why now explains that link; the ? button explains the measure and its use at that phase.
T marks an assumed storm-impact time for planning. The hour labels are illustrative review windows, not forecast deadlines. All phases show the same current data and scenario; they do not predict how a parameter changes over time. Follow current official instructions.
Archived calibration findings are retained in the separate historical and synthetic views. Current planning uses the selected public records and their own dates.
The console applies the ten-metric diagnostic suite from a contamination-resilient cluster-randomized-trial (CRT) design to a wastewater lift-station network. At a given scenario level the stations at or below a hypothetical flat water plane are one arm (exposed) and the rest are the other (unexposed). This is an exploratory adaptation of CRT geometric diagnostics to planning scenarios, not a validated equivalence between trial contamination bias and wastewater cascade dynamics. Actual pumped-flow propagation requires verified connections and operating telemetry, which this public inventory does not provide.
The suite has three pillars: Statistical Interchangeability (metrics 1 to 3), Topological and Spatial Cohesion (4 to 7) and Contamination Resilience (8 to 10). Each metric has a threshold; when it is crossed, the console shows the sub-optimal condition, an analytical mitigation (how to read the numbers) and an operational mitigation (what to do in the field). The response plan collects the operational mitigations of the sub-optimal metrics into timeline phases.
Live feeds set context only: the gauge reading and the OFS forecast map onto the ladder, and the NHC track suggests a phase. Nothing here predicts which station fails.
The network is generated in your browser from a fixed seed (mulberry32, seed 20260904), so every visitor sees the same 130 stations. Ground elevations are drawn inside bands whose counts reproduce the real network's surge ladder and storm planes exactly, with a few fixed values pinning the published quantiles. Stations sit on a fictional peninsula in a local kilometer frame (barrier islands and keys, causeways, tidal creeks, an inland ridge): lower stations on islands, shorelines and creek lowlands, higher ones inland. The fictional 15-edge graph, evacuation-zone distribution and criticality flags preserve archived calibration summaries. The predecessor inferred its 15 directional pairs from terrain, rather than verified pumped flow; that geometry is kept only as a synthetic exercise. The three treatment plants are the terminals: each force-main sink links, schematically, to its basin's plant (these links are not counted as edges). The archived calibration contains 130 elevation-linked records. The separate current public register has 131 records, including one with no FACILITYID; the newly retrieved terrain snapshot covers all 131. The fictional 130-station calibration is retained for reproducibility and is not presented as the current public register. The table below is recomputed from the generated network each time the page loads.
Synthetic zone assignments, not an address lookup. The public model intentionally leaves 37 of 130 fictional stations unassigned so its counts match the archived network-level summary. Assigned letters are modeled scenario labels; a missing letter is not a failed live lookup, and a fictional station cannot be resolved against a real evacuation map. For a real property, use Pinellas County’s official Know Your Zone address lookup. An evacuation zone is different from a FEMA flood zone; follow current county instructions.
No nonpublic facility coordinates or nonpublic per-station data are published. The separate Public stations view contains agency-published asset records from official public GIS, with an ownership filter plus official county zone overlays, with their provenance and limits. Its elevations are independently retrieved USGS raster-terrain estimates with NAVD88 datum and acquisition metadata, not surveyed equipment heights. It contains no operational telemetry or verified flow connections. The synthetic scenario remains unlinked to those official asset IDs; its per-station elevations, zones, topology and criticality are fictional. Archived private-network evidence remains at network-level aggregate only: NOAA gauge readings, storm-level backtest counts, publicly reported sewage volumes, and network-level summaries (counts, shares, the ten metric values). The synthetic IDs (LS-001 to LS-130) and names are neutral and do not correspond to real facilities.
All feeds are keyless and fetched by your browser. Nothing is stored or sent anywhere else.
The script defines STATUS_OVERRIDE = {}. A utility could wire it to SCADA: keys are station IDs and values look like {status: 'down' | 'surcharged' | 'ok', note: '...'}. An overridden station is drawn with a cross on the map and its status appears in the station panel. In this public copy it is empty, so nothing is overridden.
Built by Patrick Iben. This case study in research engineering and systems architecture adapts the ten-metric diagnostic suite from his contamination-resilient CRT design (preprint, doi:10.5281/zenodo.23486177) and uses directed source/sink topology accounting informed by his directed-contagion research. The directed component is source/sink accounting, not a port of that paper's VAR/FEVD estimation. Live NOAA/NHC/NWS information, archived network-level aggregates, computations on the synthetic public network, and interpretive guidance are labeled separately. The private console and response plan this public version generalizes were built for a Gulf-coast city's wastewater network.