Skip to content

Cost-Benefit Analysis

The CBA module (NB08, 08_CBA_0126.ipynb) integrates all three adaptation modules on a common 25-year discounted horizon (3% annual discount rate). Present-value costs and benefits are computed separately for each standalone policy and for explicit combined scenarios.

Hazard track

NB08 reads the active hazard track via HAZARD_TRACK (defaulting to Track B where the city config enables extreme_hazard). All CBA outputs are interpreted within the selected track. For Track-A cities (Rome, Athens, Lisbon) the standard daily-mean workflow drives the policy analysis. For Track-B cities (Copenhagen) the extreme-event workflow is the policy-analysis track, while the standard track is kept only for cross-city comparability with policies disabled.


Benefits

Benefits are measured as heat-attributable deaths avoided annually relative to the current-AC baseline:

  1. Mortality results at the four anchor years (2020, 2030, 2040, 2050) are linearly interpolated across all intervening years to produce annual benefit streams.
  2. For vegetation, avoided deaths are scaled by a tree maturity function so that benefit timing tracks the biological development of planted stock.
  3. Gross avoided deaths → net avoided deaths via mortality displacement factors.
  4. Optionally weighted by residual life expectancy to produce life-years saved.

Costs

Costs are year-specific and incremental (attributable to the policy above the no-intervention trajectory):

Policy Cost components
AC CAPEX + annual maintenance + electricity (dominant component)
Trees CAPEX (distributed linearly over 25 years) + annual maintenance (scaled by maturity profile)
EWS Per-capita per-warning-day operational cost (very low relative to other options)

Standalone vs. combined bookkeeping

NB08 keeps standalone policy rows standalone and reports interactions explicitly, rather than silently folding co-benefits into either policy:

Row Meaning
AC (NET) Standalone AC policy: CAPEX + maintenance + electricity, minus the standalone waste-heat penalty, with no tree discount
AC (NET + trees) Explicit combined scenario: same AC policy but with lower operating cost because trees reduce kwh_per_user, and a reduced incremental waste-heat penalty
Trees (base O&M) Standalone tree policy on direct health benefits only

The AC waste-heat penalty is subtracted from AC benefits as an incremental negative cost (only the incremental penalty from policy-added users is charged). A marginal interaction correction factor λ_y scales that incremental penalty in the combined scenario — it corrects the policy margin only and does not re-solve background waste heat from existing AC users (for Copenhagen λ_y ≈ 0.996, so the omitted background term is negligible).

The vegetation → AC electricity co-benefit (Falchetta, De Cian & Lunghi 2026; config block electricity_feedback) is integrated directly into NB08's AC cost computation: trees lower AC electricity demand per user, reducing the AC operating cost in the combined AC + trees scenario. It is reported as a tree-owned co-benefit (on both current-AC users and all AC users in the combined scenario), not silently subtracted from either policy's standalone cost.


Equity assessment

Mortality benefits are stratified by: - Quintiles of the composite SVI surface (both equal-area and population-weighted). - Public vs. private cost separation — AC costs fall on private households; trees and EWS on public budgets.


Budget optimisation

URBADAPT-HEAT identifies Pareto-efficient adaptation portfolios across the three policy options:

  1. Linear combinations of the three policy scales are evaluated over a dense grid.
  2. The Pareto frontier is identified in (avoided deaths, present-value cost) space.
  3. Incremental cost-effectiveness ratios (ICERs) are reported for transitions along the frontier.
  4. Sensitivity of the optimal mix to budget level and discount rate is reported.

Typical cost-effectiveness ranking

Policy Cost-effectiveness Notes
EWS Most cost-effective Very low operational cost; dominates at low budgets
Trees Intermediate Higher CAPEX/death but co-benefits (equity, amenity, cooling) not fully captured in mortality-only metric
AC Highest absolute benefit Most deaths avoided at scale; dominated by electricity costs

All three policies lie on or near the Pareto-efficient frontier.


Output indicators

Indicator Description
Avoided deaths (gross/net) Annual and cumulative over 25 years
Life-years saved Net avoided deaths × residual life expectancy
PV cost Total discounted cost over 25 years
NPV PV benefits − PV costs (benefits monetised via VSL or cost-per-QALY)
BCR Benefit-cost ratio
Cost per avoided death PV cost / net avoided deaths
Cost per life-year saved PV cost / life-years saved
ICER Incremental cost per additional unit of health benefit
SVI-stratified benefits Avoided deaths by vulnerability quintile

This page is maintained in the URBADAPT-HEAT wiki and synced automatically. Edit it there, not in the website repository.