BASELINE / METHODOLOGY BACK TO PORTFOLIO
PUBLIC TECHNICAL TRACE / 2026-07-16

What these demo numbers actually prove.

This page traces the public Baseline scenario from declared inputs to outputs. The pipeline is real. The property and utility data are synthetic. The current case proves internal consistency and product behavior, not real-property predictive accuracy.

SyntheticPROPERTY + UTILITY INPUTS
ReproducibleENERGY + GHG ARITHMETIC
ScreeningECM COSTS + SAVINGS
WithheldNUMERIC LL97 EXPOSURE
01 / CLAIM BOUNDARY

Consistency test, not field validation.

The test generates twelve-month electricity and gas consumption from the same matched archetype EUI that the calibrator receives. The resulting factor is therefore 1.0 by construction. That checks the conversion, calibration, emissions, ECM, and report pipeline without drift; it does not show that the engine predicted a real building.

The repository validation log currently contains synthetic runs. It defines real-property comparison tolerances, but no third-party audit or utility-bill benchmark is represented by this public case.
02 / INPUT LEDGER

Every input shown.

Input Value Provenance
Asset Multifamily midrise, 2005 Synthetic scenario declaration
Size 214,796 SF / 220 units Synthetic scenario declaration
Location ZIP 10001 / climate zone 4A Scenario input + climate-zone matcher
Archetype EUI 14.2 electric + 28.4 gas = 42.6 kBtu/SF/yr Engine snapshot's DOE prototype-derived lookup
Utility data 893,897.4 kWh / 61,002.1 therms Generated from archetype EUI, not observed bills
Rates $0.18/kWh / $1.20/therm Scenario assumptions
PCA Mode flag enabled; no source PCA details Scenario configuration only
03 / FORMULA TRACE

From archetype to 581.6 tCO2e.

  1. 01site EUI = 14.2 + 28.4 = 42.6 kBtu/SF/yr
  2. 02electricity = 14.2 x 214,796 / 3.41214 = 893,897.4 kWh
  3. 03natural gas = 28.4 x 214,796 / 100 = 61,002.1 therms
  4. 04Scope 1 = 61,002.1 x 5.3 kg/therm / 1,000 = 323.3 tCO2e
  5. 05Scope 2 = 893,897.4 x 0.289 kg/kWh / 1,000 = 258.3 tCO2e
  6. 06total = 323.3 + 258.3 = 581.6 tCO2e/yr
The 0.289 NYCW electricity factor and 5.3 kg/therm gas factor are the values in this engine snapshot. A production run must pin the current source version and reporting basis before calculation.
04 / ECM SCREENING

How the measure outputs are built.

Energy + carbon

  • Applicability gates filter measures by asset, vintage, HVAC, and populated delta.
  • Fuel savings equal the lookup delta times calibrated fuel EUI times GFA.
  • GHG savings convert kWh and therm reductions with the same emissions factors.

Cost + finance

  • Installed costs come from versioned unit-cost lookup tables.
  • Incentive assumptions are capped at gross CapEx but still require eligibility and stacking review.
  • Payback, NPV, and IRR use net CapEx, modeled utility savings, escalation, and hold period.
These are preliminary screening outputs. They are not contractor bids, final engineering calculations, realized savings, or eligibility determinations.
05 / POLICY VERSION AUDIT

Why LL97 dollars are withheld.

The jurisdiction matcher correctly identifies NYC Local Law 97 for ZIP 10001. The engine snapshot's exposure calculation still maps the case through an older occupancy-group threshold table. NYC DOB now publishes limits using ENERGY STAR Portfolio Manager property types. Until the property mapping and versioned limits are updated and tested, this portfolio publishes no numeric LL97 exposure or fine avoidance.

06 / CODE PATHS

What a reviewer should inspect.

Concern Code path Review question
Scenario inputs scripts/calibration_stress_test.py Are all assumptions explicit and non-client?
Archetype match engine/archetype/matcher.py Is type, vintage, and climate selection correct?
Calibration engine/calibration/calibrator.py Does confidence mean data quality or proximity to prototype?
Emissions engine/emissions/calculator.py Are factor year, region, and accounting basis pinned?
ECM deltas engine/ecm/runner.py Are fallback cells and source confidence visible?
Financials engine/ecm/financials.py Are incentives, escalation, and hold period reviewable?
Policy routing engine/bps/exposure.py Are current ESPM property types and rules versioned?
Validation docs/validation_log.md Which runs are synthetic versus independently benchmarked?
07 / SOURCE REFERENCES

Primary references.

08 / NEXT VALIDATION GATE

What would make this decision-grade.

Property evidence

  • Real PCA and complete building-system inventory
  • Twelve consecutive months of verified utility bills
  • Current rates, tariffs, and interval data where available
  • Contractor or RSMeans-aligned scope and cost review

Independent checks

  • Compare baseline EUI and GHG to Portfolio Manager or an audit
  • Compare ECM savings to an engineer's calculation
  • Pin current emissions, incentive, and policy source versions
  • Log deltas against declared acceptance tolerances