Zonal Demand
Description
We model demand at the balancing authority level with hourly resolution to capture both intraday shape and seasonal effects. The primary input is the CPUC SERVM 2025 load forecasts, which provide load projections through the forecast horizon and separate demand into:
- Base (economic) load
- Behind-the-meter (BTM) rooftop PV
- Electric vehicles (EVs)
- Large flexible loads:
- Battery manufacturing
- Cryptocurrency mining
- Data center — AI/ML
- Data center — campus infrastructure
- Data center — colocation
- Data center — hyperscale
- EV charging infrastructure
- Hydrogen production
- Semiconductor manufacturing
Forecast
Load profiles
Each load component exhibits a different profile based on function, price responsiveness and any on-site behind-the-meter generation.
Load Sensitivity
The CPUC SERVM load forecast data represents a median (p50) weather year. We extend this by enabling alternative weather-year load shapes to be layered on top — supporting higher/lower demand cases and aligning the demand shapes with our renewable load-factor weather years for consistent scenario analysis.
Assumptions and Caveats
- Demand is modeled at the zonal (balancing authority) level throughout the forecast horizon.
Data Sources
| Source | Description | Link |
|---|---|---|
| CAISO OASIS | Historical load data for splitting CISO to PGAE/SCE/SDGE sub-areas | CAISO OASIS |
| CPUC SERVM 2025 | Forward demand forecasts (raw CSVs + metadata) | CPUC SERVM |
| EIA-930 | BA-level historical load and generation data for backtests | EIA-930 |