🎓 Lesson 20 D5

Integrating CMIP6 Projections into WDN Stress Tests

Using future climate data from global climate models to test how well water pipes and pumps can handle extreme weather like droughts and floods.

🎯 Learning Objectives

  • Explain the role of CMIP6 ensembles in representing climate uncertainty for WDN risk assessment
  • Apply bias correction (e.g., quantile mapping) to CMIP6 precipitation time series for local WDN modeling
  • Analyze simulated pressure head deficits in EPANET under CMIP6-driven demand and supply perturbations
  • Design a stress-test protocol that links SSP-RCP combinations to WDN performance metrics (e.g., % nodes below 20 m head)

📖 Why This Matters

Water utilities face increasing operational strain from climate extremes—2022 European droughts caused over 1,200 WDN pressure failures, while Houston’s 2017 flood damaged 47% of pump stations. Relying on historical weather data alone leaves networks vulnerable. CMIP6 provides the most authoritative, policy-relevant climate projections used by IPCC AR6—and integrating them into WDN stress tests is now required by ISO 56000:2023 for climate-resilient infrastructure certification.

📘 Core Principles

CMIP6 delivers coordinated, multi-institutional climate model outputs across Shared Socioeconomic Pathways (SSPs) and Representative Concentration Pathways (RCPs), enabling probabilistic scenario analysis. For WDNs, three key linkages must be established: (1) downscaling coarse-grid GCM outputs (0.5°–2° resolution) to watershed or district scale using statistical or dynamical methods; (2) bias-correcting variables critical to demand (temperature → ET₀ → peak demand) and supply (precipitation → reservoir inflow → pumping energy); and (3) coupling corrected time series to hydraulic models via demand multipliers and source boundary conditions. Uncertainty propagation—from model spread, emission scenario choice, and internal variability—must be quantified using ensemble statistics (e.g., median ± 90th percentile bounds).

📐 Bias-Corrected Precipitation Time Series

Quantile mapping corrects systematic model biases by matching the cumulative distribution function (CDF) of modeled precipitation to observed (e.g., CHIRPS or NOAA GHCN) CDF. It preserves temporal structure while improving marginal distributions.

Empirical Quantile Mapping

P_{corrected}(t) = F_{obs}^{-1}\left[ F_{model}(P_{model}(t)) \right]

Maps raw CMIP6 precipitation value to observationally constrained distribution using empirical CDFs.

Variables:
SymbolNameUnitDescription
P_{corrected}(t) Bias-corrected precipitation mm/day Precipitation value at time t after quantile mapping
F_{obs}^{-1} Inverse CDF of observed precipitation mm/day Empirical quantile function derived from high-quality gauge/satellite record
F_{model} CDF of raw CMIP6 precipitation dimensionless Cumulative probability distribution of model-simulated precipitation
Typical Ranges:
Semi-arid urban catchment: 0.1 – 45 mm/day
Tropical monsoon zone: 0.2 – 220 mm/day

💡 Worked Example

Problem: Given: 30-year observed daily precipitation at Tucson AZ (CHIRPS, 1991–2020) has 95th percentile = 12.4 mm/day; CMIP6 ensemble mean (MPI-ESM1-2-LR, SSP2-4.5) for same period shows 95th percentile = 18.7 mm/day. A wet-day event on 2055-06-15 is simulated as 22.3 mm/day in raw CMIP6 output. Apply quantile mapping to correct it.
1. Step 1: Compute rank of 22.3 mm in raw CMIP6 CDF → corresponds to ~97.2nd percentile
2. Step 2: Find observed precipitation value at 97.2nd percentile in CHIRPS CDF → 14.9 mm/day
3. Step 3: Assign corrected value = 14.9 mm/day (reducing overestimation by 33%)
Answer: The bias-corrected precipitation is 14.9 mm/day, which falls within the observed 95–99th percentile range (12.4–16.1 mm/day).

🏗️ Real-World Application

The City of Cape Town integrated CMIP6 (UKESM1-0-LL, SSP3-7.0) projections into its 2023 WDN Climate Stress Test Framework. Using 10-member ensemble downscaled to 5 km resolution, they applied quantile-mapped rainfall and temperature to WEAP-MODFLOW to simulate reservoir levels and pumping energy demand through 2100. Results triggered redesign of the Zandvlei booster station—increasing capacity by 35% and adding solar backup—to maintain >99.5% service reliability during ‘Day Zero’-level droughts projected with 28% higher frequency post-2040.

📋 Case Connection

📋 Calibration of Lagos Metropolitan Water Network

Persistent model–field mismatch (>25% pressure error) due to undocumented pipe replacements and unaccounted demand growt...

📋 Climate-Adaptive Reinforcement of Cape Town’s Drought-Resilient Network

System unable to maintain minimum pressure during prolonged low-storage operation and projected 20% rainfall decline

📋 Water Quality Model Validation for Singapore’s Deep Tunnel Sewerage System (DTSS) Supply Branch

Disinfectant residual dropping below 0.2 mg/L at farthest nodes despite design dosing; suspected wall reaction dominance

📚 References