Skip to main content

California Freight Cleanup → Element 2

Which California power plants are causing the most harm?

We ran a Gaussian plume dispersion model for every California facility above a 1-ton NOx threshold and applied peer-reviewed concentration-response functions to get a ranked list: 68 plants, $3.9–$43.2 M per facility per year in health cost. One biomass facility in Stockton is a 30× outlier worth retiring before any of the natural-gas fleet.

We took CAMPD 2025 self-reported NOx and PM2.5 emissions for every California facility above a 1-ton NOx threshold — 68 pass — and dispersed each one via Gaussian plume onto the same 21,164-cell grid used throughout the study. Three peer-reviewed dose-response functions convert concentrations to health outcomes: PM2.5 to all-cause mortality (Di et al. 2017 / Krewski et al. 2009 blend), O3 to acute mortality (Roman et al. 2019), and NO2 to respiratory hospitalization. Deaths are monetized at the EPA 2024 value-of-statistical-life mode ($11.6M). This investigation runs at point estimates — portfolio-level uncertainty is handled in Elements 3 and 4. Each facility is also tagged with its DAC share: the fraction of affected census tracts carrying CES/AB 617 disadvantaged-community designation. The output is a ranked roster — health cost, deaths per year, health cost per MWh, DAC share — ready to inform permit conditions and shadow pricing.

Fleet total: 11.9 deaths/yr, $138.1M annual health cost across 68 facilities

At the California Freight Cleanup portfolio scale of 11,665 estimated deaths per year from all California PM2.5/O3/NO2 exposure, the 68 gas-adjacent facilities account for roughly 0.1% of statewide burden — but as individually locatable and regulable point sources, they offer a targeted lever that diffuse mobile-source or area-source emissions do not.

DTE Stockton is a biomass outlier: 3.725 deaths/yr, $43.2M, $56.71/MWh — 4.5× the next facility

DTE Stockton (wood-fired, PM2.5 = 1,066.8 t/yr) is included in the CAMPD dataset by the NOx filter, not by fuel type. Its health cost ($43.2M/yr) is 31% of the entire 68-facility total. Any fleet-level siting guidance must separate Stockton from the gas fleet; pooling inflates aggregate statistics with a non-gas outlier. When Stockton is excluded, the remaining 67 facilities produce 8.2 deaths/yr ($94.9M/yr), average $1.46M per facility.

For gas plants, siting location drives rank — not emission rate

Among the 65+ natural gas facilities, emission tonnage does not predict health cost rank. Haynes Generating Station (Los Angeles, 45.6 t NOx) outranks Gateway Generating Station (74.9 t NOx) because its plume reaches a far denser population. CEC siting criteria that reward low-emission locations but ignore population density will systematically misallocate mitigation value. The $1.77–$3.23/MWh externality range for top-10 gas plants provides a defensible benchmark for permit conditions or shadow pricing.

Bay Area Delta/Los Medanos cluster: dual-pathway problem, $16.8M/yr combined

Delta Energy Center (Pittsburg) and Los Medanos Energy Center (Contra Costa) together contribute $16.8M/yr in health cost and a meaningful O3 mortality component (27–33% of their respective totals) in a region where the ozone pathway is positive rather than titrating. Both sit within a terrain corridor where a single-facility Gaussian plume likely underestimates cumulative near-field exposure.

DAC burden tracks California demographics, not a targeted concentration: fleet mean ~20.2% vs. California's 22.9%

Only 1 of 68 facilities (Malburg Generating Station) has more than 50% of its affected tracts DAC-designated. Fleet-average DAC burden (20.2%) tracks California demographics. This does not mean DAC impacts are zero — it means the gas fleet does not exhibit the spatial targeting that would justify a DAC-specific siting restriction above and beyond standard health cost review.

The per-facility health cost outputs feed three downstream elements: