B
BCAL ENERGY
Start a Conversation
Bcal Energy White Paper Series · No. 007

Air Permitting for On-Site Generation:
A California District Reality Check

Thirty-five local air districts, thirty-five rulebooks. The BACT trigger that decides what control hardware a generator must carry is five times stricter across a county line, and non-combustion equipment often needs no district permit at all. Where your site sits shapes what you should build.

The first hard question in on-site generation is not which machine to buy. It is which air district the meter sits in, because California permits generators through thirty-five local rulebooks, and the same engine that passes quietly in one county can require control hardware, purchased offsets, or a redesign across the county line.

Section 01Thirty-five rulebooks, one state

Start with the structure, because it explains everything else. The federal government sets the ambient standards a region must meet. The state's air board writes statewide rules for vehicles, fuels, and toxic hot spots. But the permit that governs a stationary generator comes from neither. California assigns stationary-source and facility permitting to its thirty-five local air districts, each with its own governing board, engineering staff, rulebook, and enforcement posture.1

For an on-site generation project, the sequence is broadly consistent from district to district: an authority to construct before equipment goes in, a permit to operate for its working life, inspection and renewal after that. Everything inside the sequence is local. Each district writes its own new-source-review rule, its own exemption list, its own control-technology triggers, its own offset accounting, its own fees. The two districts most relevant to Northern California's large new loads, the Bay Area district and the San Joaquin Valley district, share a boundary; sites an hour apart by road can answer to different agencies applying thresholds that differ by a factor of five.

The variation is not administrative accident. Districts inherit stringency from geography. The severity of an air basin's federal nonattainment status flows down into how low the permitting triggers sit and how hard the offset provisions bite, and the state's inland basins carry some of the most demanding ozone designations in the country. A permitting story imported from another region, however true it was there, is evidence about the equipment, not about your site.

One more structural fact that owners routinely miss: utility territory and air district boundaries are unrelated maps. A single utility's service area crosses many districts, so the interconnection answer and the permit answer come from different authorities, on different clocks. A project plan that carries only one of them is half a plan.

Section 02Four mechanisms price the permit

District programs move project economics through four mechanisms. Every district tunes them differently, and together they produce the permit line that belongs in any honest technology comparison.

1. Exemption lists decide who plays at all

Every district's permitting rule opens with a list of equipment that does not need a permit, and those lists quietly sort the generation field. The clearest example for this paper's subject: the Bay Area district's general-requirements rule exempts fuel cells that use phosphoric acid, molten carbonate, proton exchange membrane, solid oxide, or equivalent technologies.2 A non-combustion electrochemical plant in that district does not file the application, wait out the evaluation, or carry the permit conditions that an equivalent combustion plant would. The exemption is not immunity; the same rule pulls exempt equipment back into the accounting when a facility is large enough to need a major facility review permit.2 But at the single-project scale, sitting on the exemption list removes an entire regulatory workstream from the schedule.

Exemption lists are district choices, not state policy. What one district exempts, a neighbor may permit. The list must be read for the district the parcel is actually in, in its current form, conditions included.

2. Control-technology triggers set the hardware

Best available control technology, BACT in the trade, is the control standard a permitted unit must meet, and each district sets its own trigger. In the Bay Area, BACT attaches when a new source has the potential to emit ten pounds or more of a covered pollutant on any day.3 In the San Joaquin Valley, the trigger is two pounds in any one day.4 In the South Coast, there is no numeric floor at all: any emission increase of a nonattainment pollutant from new or modified equipment brings the requirement.5 What BACT means in hardware is recorded in each district's published determinations. For combustion generation it generally means add-on controls, selective catalytic reduction and oxidation catalysts, with the reagent supply, monitoring, source testing, and maintenance obligations that ride along.

Two properties of these triggers deserve respect. They are computed on potential to emit, the unit's capability under its enforceable limits, not on the dispatch you expect; a machine you intend to run rarely is priced at what it could run unless the permit itself says otherwise. And they are daily, not annual, so a unit that runs hard on any single day can trip a trigger its annual average would never suggest.

3. Offsets put a market inside the permit

Past certain thresholds, a project must surrender emission reduction credits to compensate for its new emissions, and here the districts diverge most visibly. The Bay Area requires offsets for organic precursors and nitrogen oxides once a facility's potential to emit exceeds ten tons per year; between ten and thirty-five tons the district itself supplies the credits at one-to-one from a small-facility banking account, and from thirty-five tons up the applicant provides federally enforceable offsets at 1.15 to one.3 The San Joaquin Valley publishes pollutant-by-pollutant annual thresholds in its rule: 20,000 pounds per year for nitrogen oxides and for volatile organics, 29,200 pounds for particulate matter (PM10), 54,750 pounds for sulfur oxides, 200,000 pounds for carbon monoxide.4 The South Coast requires net increases of nonattainment pollutants to be offset at 1.2 to one through emission reduction credits unless a specific exemption in its rules applies.5

Credits are property. Supply is finite, ownership is concentrated in some basins, and price discovery is thin. A project that crosses an offset threshold has added a procurement task to its critical path, and in a tight credit market that task can rival equipment decisions for consequence. The design responses are all legitimate and all have costs: accept enforceable limits that hold the facility under the threshold, buy the credits, or choose equipment whose emissions never engage the requirement.

4. Toxics review decides the neighborhood conversation

Criteria pollutants drive BACT and offsets; toxic air contaminants drive the health-risk review that decides how equipment sits relative to the people around it. Districts screen new units for health risk at nearby receptors, and the analysis tightens with proximity to schools, homes, and care facilities. Diesel particulate is the sharpest case: it is a listed toxic air contaminant in California, which is why the diesel engine, the cheapest standby capacity money buys, faces the most siting friction of any machine in this field. This mechanism stays qualitative in this paper deliberately, because it is the most site-specific of the four. The same engine that permits smoothly in an industrial park can struggle across the street from a schoolyard, at identical emission rates.

Section 03The arithmetic, district by district

Put the two Northern California triggers side by side and the reality check writes itself.

10 lb
Potential to emit per highest day that triggers BACT for a new source under the Bay Area district's New Source Review rule3
2.0 lb
Potential to emit in any one day that triggers BACT under the San Joaquin Valley district's Rule 22014

Same machine, same fuel, same duty cycle, and a five-fold difference in the threshold that decides whether it must carry control hardware. At two pounds per day, essentially any combustion unit intended for continuous service will meet BACT in the Valley; the trigger functions as a design requirement, not a screen. The Bay Area's ten-pound threshold leaves modest room at the small end. The South Coast's zero-floor rule leaves none. And the annual arithmetic connects to the daily: the Valley's 20,000-pound offset threshold is ten tons a year, so run-hours become the variable that walks a facility toward offset territory. Two identical machines with different duty cycles are, to the permit, different projects.

DistrictBACT triggerOffset triggerReading for generation projects
Bay AreaPotential to emit 10 lb or more of a covered pollutant on any day (Reg. 2-2-301).Facility-wide potential to emit above 10 tons per year of NOx or organic precursors; district-supplied credits at 1:1 up to 35 tons per year, applicant-supplied at 1.15:1 beyond (Reg. 2-2-302). Separate provisions govern particulates and sulfur dioxide.Small and non-combustion projects can travel light. Continuous combustion at scale meets BACT, and larger facilities meet the offset ledger.
San Joaquin ValleyPotential to emit exceeding 2.0 lb in any one day (Rule 2201 §4.1).Per pollutant, per year: 20,000 lb NOx; 20,000 lb VOC; 29,200 lb PM10; 54,750 lb SOx; 200,000 lb CO (Rule 2201 Table 4-1).BACT is effectively universal for combustion prime power; annual run-hours drive offset exposure.
South CoastAny emission increase of a nonattainment contaminant from new or modified equipment (Rule 1303).Net increases offset through emission reduction credits at 1.2:1, unless an exemption in Rule 1304 applies (Rule 1303).Controls are table stakes for everything; the real screen is whether credits exist at a price the project survives.

A table like this is a snapshot of moving rules. Districts amend them on public dockets, and the current text controls; the citations below carry access dates for exactly that reason. The table is presented for one purpose: to make visible that the spread between districts is not noise. It is large enough to change which technology wins a fair comparison at a specific site.

The same machine, the same fuel, the same duty cycle. Move it across a county line and the permit answer changes.

Section 04Combustion and non-combustion are different regulatory species

California district rules do not treat generation as one category. They divide it, functionally, into equipment that burns and equipment that does not, and the two families walk different roads through everything described above.

None of this makes combustion unpermittable, and treating the permit as a wall is its own analytical failure. Districts permit engines and turbines routinely; the published control determinations exist precisely because projects clear them. Reciprocating engines remain the low-capital, fast-delivery choice with the deepest service ecosystem, and selective catalytic reduction with oxidation catalysts is mature, effective hardware. Gas turbines and microturbines bring high-grade exhaust heat that rewards sites with genuine thermal demand. The permit's real effect on the combustion case is quantitative: control hardware and reagent add capital and operating cost, source tests and fees recur, offsets add capital where triggered, and the application clock adds schedule. Whether the combustion option survives those additions is a per-site calculation, and at many sites it does.

At the small end of the combustion family, some districts publish size or throughput exemptions under which the smallest units escape permitting. Whether a specific microturbine or small engine clears a specific district's floor is a fact to verify in the current rule text, never to assume from a spec sheet.

Emergency engines live in a separate lane, and the lane has walls. Districts condition emergency units on limited non-emergency operation, and the Valley's rule shows the boundary plainly: its offset exemption for emergency standby power equipment applies only where non-emergency operation stays within 200 hours per year, with written logs and permit conditions to hold it there.4 The practical reading is that emergency assets are permitted for emergencies. A plan that quietly re-tasks standby engines as bridge prime power is a category error the permit file will surface, and diesel toxics review shapes where those units can sit regardless of hours.

Non-combustion electrochemical generation walks the other road. Where a district's exemption list covers fuel cells, as the Bay Area's does, the project simply does not enter the permit queue, which removes an entire class of schedule and redesign risk.2 Exemption from the district is not exemption from the state: California law directs the state air board to certify distributed generation technologies that are exempt from district permit requirements, and the certification standard for fossil-fueled units has, since 2007, capped nitrogen oxides at 0.07 pounds per megawatt-hour.6 A machine that avoids the district counter still clears a statewide emissions bar before it can be sold here. For owners, that is comfort rather than loophole: the lighter treatment exists because the equipment class earned it on emissions, and the state audits the claim at certification.

The honest ledger for the exempt classes still has debit lines. Non-combustion electrochemical equipment typically carries a higher capital cost than combustion peers; a gas-fed unit keeps full exposure to fuel price whatever its permit status; stack replacement is a real lifecycle cost that belongs in any comparison; and the air-permit advantage says nothing about carbon, because a gas-consuming device emits carbon dioxide whether or not anything burns. Air districts regulate criteria pollutants and toxics. Carbon accounting runs on separate rails, and a decision that confuses the two lanes will misprice both.

Solar and battery storage sit outside this machinery almost entirely: nothing burns, nothing exhausts, and district generation permitting generally has no purchase on them. That is a genuine advantage worth stating plainly. Their constraints simply live in other files: land and structural capacity, interconnection, and, for storage, fire-safety and land-use review that is its own discipline. For daytime-weighted loads they can be decisive; for continuous loads, capacity factor decides how much of the job they can carry alone.

Section 05What the permit map does to technology choice

The permit line belongs inside the technology matrix, not appended to it. Control hardware is capital cost; reagent, testing, and fees are operating cost; offsets are capital; the application clock is schedule. Schedule deserves particular respect here, because for most owners contemplating on-site generation the entire motivation is a date the grid cannot meet. A path whose permit clock consumes the schedule advantage was never the fast path, whatever its nameplate said.

Facility-wide accounting defeats structural cleverness. The Bay Area's offset test runs on the facility's total potential to emit, and the Valley's runs on the stationary source's post-project total, so splitting one project into several smaller units, or phasing installations across applications, does not reset the arithmetic.34 Districts have seen every version of the attempt. Design the project the load requires, let the thresholds fall where they fall, and price what they imply.

Read the district's own published determinations before pricing anything. Districts maintain public records of what they last required for each equipment class, and that record is the nearest thing to a forward quote the process offers: it tells you the hardware, and therefore much of the cost, that a new application should expect. A comparison matrix that prices technologies without reading those determinations is guessing about the largest open variable.

Duty-cycle limits are legitimate design variables. Potential to emit is computed against enforceable permit conditions, so an accepted limit on annual hours or fuel throughput can hold a unit below a trigger. The price is flexibility, and it should be paid knowingly: a limit that pinches in year three because the load grew is a self-inflicted constraint. Sometimes the all-in combustion case, controls and credits included, still wins. Sometimes the exempt path wins on schedule alone. Sometimes the honest answer is that the district map argues for a different site. The screen exists to surface that answer early, while it is still cheap to act on.

Section 06The eight-question district screen

The screen below runs on public documents and a conversation with the district, and it belongs at the front of any on-site generation evaluation, before equipment conversations begin, because its answers reshape those conversations.

  1. Which district holds the parcel?District boundaries follow air basins and county lines, not utility territory or city limits, and some districts split counties. Map the actual site address before anything else.
  2. Is the candidate equipment on the exemption list?Read the named district's current general-requirements rule, including every condition attached to the exemption and how exempt equipment is counted at major-facility scale.
  3. What is the potential to emit, by pollutant, by day and by year?Permits price capability under enforceable limits, not the dispatch you expect. Compute potential to emit first; it decides which rules engage.
  4. Does potential to emit cross the BACT trigger, and what did this district last call BACT for this equipment class?Published determinations are the closest thing to a forward quote. Price the listed controls into capital and operating cost before comparing technologies.
  5. Where does the facility land against offset thresholds, and who supplies credits at what ratio?The arithmetic is facility-wide, not per-unit. Identify the credit source, the ratio, and whether a district banking mechanism applies at your size.
  6. What does toxics review ask at this site?Inventory the receptors: schools, residences, care facilities. Health-risk screening is the mechanism most sensitive to the map around the fence line.
  7. Does the permit clock fit the power deadline?Ask the district about current processing realities, including public-notice paths. If the clock and the deadline disagree, that finding outranks any efficiency number.
  8. Which way is the rulebook moving?Districts amend rules and determinations on a public calendar, and a unit permitted at today's threshold operates for decades. Check the active rulemaking docket before committing a design.

An answer sheet with all eight filled in does not choose the machine. It prices the paths honestly, which is the only thing a permit screen owes a decision.

Section 07District facts are site facts

Soil reports, seismic class, utility capacity: no one prices a building without the site facts. The air district is a site fact of exactly that rank for on-site power, and it is the one most often left to the end of the analysis, where it does the most damage. Named early, it is arithmetic: a trigger, a threshold, a credit market, a clock. Named late, it is a redesign.

None of this crowns a technology. Combustion equipment permits across California every year, with controls the districts themselves judged achievable. Non-combustion equipment earns its lighter treatment by emitting almost nothing the criteria-pollutant system needs to forgive. The right answer for a specific site follows from the load, the fuel path, the deadline, and the district together, and what the district map removes is the fiction that the decision can be made statewide. The first paper in this series put a test question to any energy recommendation: a study that never names your air district never studied your site. This paper is the reason that question earns its place.

Sources

  1. California Air Resources Board, "California Air Districts" (stationary source and facility permitting authority of the 35 local air districts). arb.ca.gov. Accessed August 9, 2026.
  2. Bay Area Air Quality Management District, Regulation 2, Rule 1 (General Requirements), Section 2-1-128.22 (fuel cell exemption) and Section 2-1-129. baaqmd.gov. Accessed August 9, 2026.
  3. Bay Area Air Quality Management District, Regulation 2, Rule 2 (New Source Review), Sections 2-2-301 and 2-2-302. baaqmd.gov. Accessed August 9, 2026.
  4. San Joaquin Valley Unified Air Pollution Control District, Rule 2201 (New and Modified Stationary Source Review), Sections 4.1, 4.5.3 (Table 4-1), and 4.6.2. valleyair.org. Accessed August 9, 2026.
  5. South Coast Air Quality Management District, Rule 1303 (Requirements), Regulation XIII. aqmd.gov. Accessed August 9, 2026.
  6. California Air Resources Board, Distributed Generation Certification Regulation, Final Regulation Order, title 17 CCR sections 94200-94214 (authority: Health and Safety Code section 41514.9; section 94203(b), January 1, 2007 fossil fuel emission standards). arb.ca.gov. Accessed August 9, 2026.
The series

One paper. Every day.

The Bcal Energy White Paper Series covers the decisions, technologies, and market evidence behind time-to-power. New research publishes continuously in the library.

Browse all papers
The study

Run this test on your own site.

The Power Readiness Study is our fixed-fee written analysis of every credible path to power for one specific site: $25,000, technology-neutral by design, sold with no equipment margin behind it. A free 20-minute conversation comes first.

info@bcalenergy.com

About Bcal Energy. Bcal Energy is an independent, founder-led California firm. We prepare technology-neutral power readiness studies for organizations facing time-to-power decisions, on the owner's side of the table. We sell the decision, not equipment. Author: Bharath Ramanidharan, Founder. Contact: info@bcalenergy.com.

Disclaimer. This paper is general information, not engineering, legal, tax, or investment advice, and not an offer of services on any specific terms. Figures described as illustrative are estimates. Statutory, tariff, and program references are current as of the publication date only; confirm status with qualified counsel and advisors before acting. Bcal Energy provides no guarantee of savings, output, performance, or timelines. © 2026 Bcal Energy.