Inputs & assumptions
The green roof is built from standardized modules per Future Green’s Schematic Design for the NYC pilot. A shelter takes the maximum modules that fit its model (no partial coverage). The selected coverage option sets the module dimensions, which in turn fix the soil profile and planting. Cooling and cost values are provisional, to be updated as the Brownsville pilot is monitored.
| Coverage options | Full coverage: 3′9″×3′8.5″ modules (≈13.9 SF each), 1.75″ media + ½″ drain mat, pre-grown sedum mat + native annual seed. Half coverage: 1′10.5″×3′8.5″ modules (≈7.0 SF each), 5.5″ media + ½″ drain mat, native perennial plugs + native annual seed. Both use 3 modules on the 14′ model (≈42 SF full / ≈21 SF half). |
|---|
| Modules per model | From the DOT inventory + NYC JCDecaux dims: 14′ (4-section) → 3, 10′ (2-section short) → 2. Solar models take one fewer (the roof PV panel occupies ~one module’s footprint). Narrow shelters (3′6″ deep) are out of scope: the 3′8.5″ module is too deep to fit. (Solar coverage + the narrow case are best estimates pending Future Green confirmation.) |
|---|
| Soil & storage | Stored-water depth = media×0.30 + drain×0.25 (CNT Green Values storage method; porosities assumed, to confirm against Future Green’s media + drain-mat products). Full → 0.65″ storage; Half → 1.78″. CNT Green Values. |
|---|
| Retention (per storm) | area × min(storm, storage) × 0.62 gal/in/ft². On a 14′ shelter at a 1″ storm: Full ≈ 17 gal (more area, shallow: storage-limited at 0.65″), Half ≈ 13 gal (less area, deep: rain-limited). For storms > ~1.3″ the deeper Half option retains more. |
|---|
| Retention (annual) | area × the NOAA daily-rainfall capture depth at the option’s storage (NOAA GHCN-Daily, Central Park 1991–2020; daily green-roof water balance with FAO-56 Hargreaves ET, calibrated to the ~50% reported for temperate extensive roofs; Mentens et al. 2006). Full’s larger area generally yields more annual gallons; Half’s deeper media captures a higher fraction. |
|---|
| Cooling effect PROVISIONAL | Under-shelter heat-stress reduction, in two cited parts: sheltering ≈4°F (a stainless-steel frame with glass roof, the NYC type, gave the largest heat-stress reduction of four designs, Yang et al. 2024); the green roof adds up to ≈2°F on a 70 ft² reference roof (Wang et al. 2025), scaling with the module area. The two options differ mainly through area. The sedum-vs-perennial transpiration difference is real but stays within this provisional band. Pending Columbia pilot. |
|---|
| Planting & pollinators | Full coverage uses a sedum mat (lower native-pollinator value); Half coverage uses native perennial plugs (higher value). This is shown as a qualitative flag on the Biodiversity lens; it does not change the connectivity score, which is a position-based siting heuristic that doesn’t read the roof’s own plants. |
|---|
| Retrofit cost ROM | $1,400 – $7,000 per shelter (rough order of magnitude). Dominated by fixed structural/labour cost rather than roof area. To be refined with the team. |
|---|
Green-roof specs are from Future Green’s Schematic Design for the NYC Green Roof Bus Shelter Pilot.
Agency lineage
Where possible, lenses align with frameworks already in use at NYC agencies, to avoid duplicate measurement burden and to slot into existing decision-making processes (NPCC standardization work, OMB capital-project intake).
| Heat | NYC CEJ heat metrics. Community framing reports the share of in-scope stops in high-HVI areas (NYC DOHMH Heat Vulnerability Index ≥ 4), alongside the pedestrian demand of the served streets (NYC DOT Pedestrian Mobility Plan — nearest-segment category, most-common across a scope). On the Heat lens the map also flags the six Brownsville stops in the 3M field study: three instrumented test sites (Pitkin Av & Mother Gaston Blvd; Pitkin Av & Bristol St; Dumont Av & Mother Gaston Blvd) and three paired controls (Pitkin Av & Amboy St; Rockaway Av & Livonia Av; Saratoga Av & Livonia Av). |
|---|
| Stormwater | Per-storm retention from the CNT Green Values Calculator green-roof storage method (substrate + drainage water-holding volume); annual retention from NYC’s measured rainfall (NOAA Central Park) via a daily green-roof water balance with evapotranspiration. Community framing reports the share of in-scope stops in high flood-burden areas (NYC MOCEJ FSHRI ≥ 4). Aligns with NYC MOUA conservation metrics (NY State Soil & Water District) + NYC DEP Green Acres standards. |
|---|
| Biodiversity LOW CONFIDENCE | Pollinator habitat connectivity, scored per shelter as a gap-crossing stepping-stone value. The method follows the habitat-availability connectivity work of Saura & Pascual-Hortal 2007 and is anchored to solitary-bee foraging of about 250–300 m (Gathmann & Tscharntke 2002). Read it as a position-based siting heuristic that does not measure a realized benefit: the green-roof-as-habitat evidence is weak, and a shelter roof may sit below the effective stepping-stone threshold. Patches = NYC parks (DPR) + MOUA urban-agriculture sites + DEP green infrastructure (plant-bearing only; concrete-top infiltration basins excluded) + the 2015 street-tree census (alive). GI points and street trees that fall inside a park are dropped, since the park already represents that habitat. Each patch is weighted by habitat capacity, with weight = √(habitat area): park area from geometry, GI from asset_area, gardens a nominal 500 m², each street tree about 12 m². This follows the Probability-of-Connectivity convention (area-weighted patch attributes), so a bridge value = wA·wB·p(dA)·p(dB). A large park therefore outranks a single small GI point, and the √ keeps the weighting sub-linear so a cluster of small GI no longer wins by sheer count. On the map the lens shows green spaces within about 300 m of a stop. In Single-shelter scope, arcs link the selected stop to its nearest reachable patches and street trees. |
|---|
| Air quality | Foregone in this evaluation phase: shelter-scale sensors are physically too large and approval timelines do not fit the current pilot window. |
|---|
What each lens reports
- Heat — the under-shelter heat-stress reduction (sheltering + green roof), the share of in-scope stops in high-HVI areas, and pedestrian demand.
- Stormwater — per-shelter retention per 1″ storm, total retention across the scenario, and the share of in-scope stops in high flood-burden areas.
- Biodiversity — the connectivity score, how many shelters act as stepping stones, and what each bridges (tree, park, garden, GI).
Per-shelter figures (cooling, retention, connectivity) are non-additive across shelters, so multi-shelter scopes report a representative value, such as the median connectivity score. Totals are shown only where they genuinely sum (retention gallons).
Manual overrides. The Heat and Stormwater sidebars each accept a user-entered per-shelter value. On Heat it replaces the under-shelter temperature reduction, and the modeled sheltering-plus-green-roof breakdown is hidden while the override is set. On Stormwater it replaces each viable shelter’s retention, so the scenario total and the average recompute from it (out-of-scope shelters stay at zero). Clearing the field restores the modeled value. An override travels in the shareable URL.
Scope
- All sheltered stops are eligible. The scenario operates on every existing bus-shelter stop (≤25 m shelter–stop match); the viable-typology gate has been removed. Stops with no shelter are shown for context (binary sheltered/unsheltered) but excluded from the scenario.
- Structural capacity — which JCDecaux typologies can carry a green-roof load is still to be confirmed with the team; for now all sheltered stops are modifiable.
- Single-shelter scope defaults to the Brownsville pilot stop (Pitkin Av & Saratoga Av); click any shelter on the map to switch.
- Borough scope aggregates shelters by borough. Finer-grained NTA / Climate Strong Communities geographies are future work.
Open work
- Biodiversity lens — pending Columbia measurement capacity + partner institution coordination.
- NTA & Climate Strong Communities geographies for a finer-than-borough scope.
- Formal DEP "green acres" translation for stormwater community-scale row.
- OMB capital-project key indicators + Technical Appendix B alignment.
- Integrate agency partner metric tables (CEJ heat, MOUA conservation).
- Designed PDF tear sheet.
Data sources
- Bus routes / stops: MTA
- Bus shelters (model, canopy, solar): NYC DOT (JCDecaux inventory)
- Pedestrian demand: NYC DOT Pedestrian Mobility Plan — Pedestrian Demand (nearest-segment category)
- Heat Vulnerability Index (HVI): NYC DOHMH 2024 (by MODZCTA)
- Flood Vulnerability (FSHRI): NYC MOCEJ 2024 (by census tract)
- Biodiversity patches: NYC DPR parks; NYC MOUA urban-agriculture sites; NYC DEP green infrastructure; NYC 2015 Street Tree Census
- Green-roof stormwater retention: CNT Green Values Calculator methodology (per-storm storage volume)
- Rainfall (annual retention model): NOAA NCEI GHCN-Daily, NY City Central Park (USW00094728), daily record 1991–2020
- Cooling: Yang et al. 2024 (bus-shelter thermal study); green-roof pedestrian cooling (Wang et al. 2025) — provisional
- Biodiversity connectivity method: Saura & Pascual-Hortal 2007 (habitat-availability index); Gathmann & Tscharntke 2002 (solitary-bee foraging range)
- 3M heat-study sites (test / control): project field study, Brownsville