LEED v4.1 Daylight Zone Compliance Guide

Daylight zones aren’t “drawn”—they’re negotiated with physics, blinds, and GBCI’s very specific idea of what “300 lux” means.

I spent three months re-running one daylight simulation—not because the model was wrong, but because I’d modeled the blinds as “fully closed” instead of “partially deployed at 45° tilt.” GBCI flagged it. Not with a comment. Not with a question. Just a hard “No” on EQ Credit 4.1.

That’s how precise this is.

This isn’t about slapping a skylight in and calling it “daylit.” LEED v4.1 EQ Credit 4.1—Daylight—demands documented, defensible, *measured-in-simulation* proof that ≥55% of regularly occupied floor area achieves ≥300 lux for ≥50% of annual occupied hours. And the daylight zone? That’s where the rubber meets the road—and where most submissions get dinged before they even hit the review queue.

Let’s walk through how to actually calculate and document primary and secondary daylight zones—not as theory, but as a working AGi32 workflow that passes GBCI scrutiny. No fluff. No “best practices” without context. Just what works, what fails, and why.

Step 1: Know the zone definitions cold—no interpretation allowed

LEED v4.1 doesn’t let you eyeball it. It defines daylight zones mathematically:

• Primary daylight zone: A band extending 15 ft (4.57 m) horizontally from the glazing plane—but only where direct line-of-sight to the sky exists. This includes view windows, clerestories, and vision panels in doors—but excludes spandrel, fritted, or opaque glazing.
• Secondary daylight zone: The area beyond the primary zone, up to an additional 15 ft (4.57 m), only if it receives ≥300 lux at the workplane (30 in / 762 mm above finished floor) under clear-sky conditions at 9 a.m. and 3 p.m. on the equinox.

Note: That “equinox” requirement isn’t optional window dressing. It’s the basis for the two mandatory time points in your AGi32 simulation run. You don’t get to pick “a sunny Tuesday in April.” You run March 21 and September 21—at 9 a.m. and 3 p.m. local standard time. Period.

I’ve seen teams run simulations at noon only. Or use average overcast sky models. Both get rejected instantly.

Step 2: Glazing assumptions—where 70% of submissions trip

GBCI doesn’t care what your spec sheet says about “VT = 0.42.” They care what’s *in the model*—and whether it matches reality.

The most common failure? Modeling glazing with its *nominal* visible transmittance (VT), but forgetting the interior shading layer. Blinds aren’t decorative—they’re optical filters. And LEED requires you to model them *as deployed*, not stowed.

Here’s what GBCI expects (per LEED v4.1 MPR Supplemental Guidance, updated Jan 2023):

• If automated or manually operated blinds are part of the design intent (and they almost always are in offices), you must model them in their typical operating state during daylight hours.
• “Typical” means: tilted at 45°, not fully open or fully closed. Why? Because that’s the position that balances glare control and light transmission—and it’s the default assumed in ASHRAE 90.1 Appendix G, which LEED references.
• Blind openness factor (OF) must be entered explicitly—not guessed. For a standard aluminum Venetian blind at 45°, OF ≈ 0.35–0.45. Use manufacturer data if available. If not, AGi32’s built-in “Aluminum Blind 45°” material (OF = 0.38) is acceptable—but document that choice.

I once used a generic “fabric shade” material with VT = 0.12—thinking “darker is safer.” Wrong. GBCI rejected it because fabric shades typically have near-zero openness factor (<0.05), so modeling them at VT = 0.12 implied inconsistent optical behavior. They asked for photometric test reports. I didn’t have them. Resubmission delay: six weeks.

Bottom line: Your glazing VT × blind OF = effective daylight transmission. If your window VT is 0.55 and your blind OF is 0.38, your effective VT is 0.21—not 0.55. Model it that way. Document it that way.

Step 3: AGi32 setup—non-negotiable parameters

You can’t wing this in AGi32. Here’s the exact sequence I use—tested across 12 LEED projects, zero daylight-related rejections:

1. Create a dedicated “Daylight” calculation file. Don’t reuse your lighting power density (LPD) model. Daylight demands different surface reflectances, no artificial lights active, and precise sky conditions.
2. Set workplane height to 762 mm (30 in). Not “approx. 30 in.” Not “user-defined.” Enter 762 mm. GBCI checks this.
3. Use CIE Standard Overcast Sky + CIE Clear Sky. AGi32’s “LEED Daylight” sky preset is fine—but verify it loads both skies. You need both for credit compliance (overcast for minimum threshold; clear for peak performance).
4. Run two time points only: March 21 & Sept 21, at 9 a.m. and 3 p.m. local standard time. Time zone must match project location. Daylight saving time? Turn it OFF. GBCI requires standard time.
5. Grid resolution: 2 ft × 2 ft (0.61 m × 0.61 m) max. Finer grids (1 ft) are allowed—but don’t go coarser. GBCI has rejected submissions using 3-ft grids for “insufficient spatial resolution.”
6. Include all adjacent surfaces that affect bounce: ceiling (reflectance ≥0.70), walls (≥0.50), floor (≥0.20). If your actual finish specs fall below those, model them accurately—even if it hurts your results. GBCI compares your model reflectances to spec sheets.

Pro tip: In AGi32, go to Calculation > Settings > Advanced and enable “Include Direct Sunlight.” Yes, even for overcast sky runs. Why? Because LEED requires “direct sun contribution where present,” and AGi32 handles sun penetration through glazing correctly only when this is toggled. Skip it, and your 3 p.m. clear-sky result will be artificially low.

Step 4: Calculating the zones—how to draw the line (literally)

You don’t “assign” zones. You simulate, then measure where 300 lux hits.

In AGi32:

• Run the four required time/sky combos (clear/overcast × 9 a.m./3 p.m.).
• For each, generate a Contour Plot with contour lines at 300 lux (not 250 or 350). Set “Fill Between Contours” to highlight everything ≥300 lux.
• Export each contour plot as a high-res PNG (300 dpi) with visible grid lines and scale bar.
• Then—this is critical—use AGi32’s Area Calculation Tool to measure square footage within each contour. Do NOT estimate by overlaying CAD layers. Do NOT eyeball from a PDF.

The primary zone is *not* just “15 ft from glass.” It’s the portion of that 15-ft band where ≥300 lux occurs at *both* time points, under *both* sky conditions. If your 9 a.m. clear-sky reading hits 320 lux but your 3 p.m. overcast reading drops to 280 lux? That spot doesn’t count toward compliant area.

Same for secondary zone: It must hit ≥300 lux at *one* of the four time/sky combos—*and* be outside the primary zone—to qualify. But crucially: That secondary zone area only counts if it’s *within the next 15 ft*, and if it’s not obstructed by furniture, partitions, or ceiling-mounted obstructions modeled at true height.

I once modeled a 72-in-high partial-height partition 12 ft from the window—and forgot to include it in the daylight model. The simulation showed 300+ lux beyond 15 ft. Reality? Zero light behind that wall. GBCI caught it in the “model fidelity” review. We had to resimulate with full interior architecture. Two weeks lost.

Step 5: Documentation—what GBCI actually reads

Your submission isn’t the AGi32 file. It’s the narrative report—the one GBCI reviewers open first.

They skim. So make every sentence pull weight.

Required elements (per LEED v4.1 Credit Library, EQ Credit 4.1):

• A table listing all glazing assemblies: orientation, U-value, VT, height/width, head height above floor, adjacent obstruction type (e.g., “2-story plaza, unobstructed” or “1-story building, 45 ft setback”).
• A second table showing blind type, material, openness factor, tilt angle, and control logic (e.g., “automated, photosensor-triggered, tilt adjusts between 30°–60° based on incident light”)
• Four annotated contour plots (one per time/sky combo), with primary/secondary zones clearly labeled and shaded. Add arrows pointing to key obstructions.
• A summary table showing:
  • Total regularly occupied floor area (ROFA)
  • Area meeting ≥300 lux at all four time/sky combos (primary zone)
  • Area meeting ≥300 lux at ≥1 time/sky combo beyond primary zone (secondary zone)
  • Combined compliant area ÷ ROFA = %
• AGi32 export log showing simulation date, software version (v10.2 or newer), and calculation settings (grid size, sky models, workplane height).

What gets flagged every time?

• Mismatched units. One table in metric, another in imperial? Rejected. Convert everything to feet *or* meters—and state which.
• Missing blind documentation. Saying “blinds provided” isn’t enough. GBCI wants OF values, tilt angles, and how they’re controlled.
• ROFA calculated incorrectly. Regularly occupied areas exclude restrooms, mechanical rooms, storage, and corridors unless they’re >6 ft wide and have desks or workstations. I’ve seen submissions include 4-ft-wide corridors as “occupied”—instant fail.

Step 6: The “why” behind the 300 lux threshold

It’s not arbitrary. 300 lux is the Illuminating Engineering Society (IES) recommended minimum for general office task performance—reading, keyboard work, video conferencing—without supplemental lighting.

But here’s what no LEED guide tells you: That 300 lux must be achieved *at the workplane*, not at the ceiling or window head. And it must be uniform *enough* that users don’t constantly adjust blinds or relocate.

AGi32 gives you point-by-point lux values. GBCI doesn’t require uniformity ratios—but if your contour plot shows wild swings (e.g., 600 lux at desk edge, 120 lux 2 ft away), expect a comment asking for glare analysis (which means adding AGi32’s DGPS module and running additional simulations).

This is why modeling blinds correctly matters so much: A blind tilted too far open causes hot spots. Too closed, and you starve the space. 45° is the sweet spot for distribution—not peak transmission.

Final note: This isn’t about “passing.” It’s about proving daylight works.

When you nail the daylight zone calculation, you’re not just checking a box. You’re documenting that people in that office will sit near a window and *not need to flip a switch* for half the year. That the blinds aren’t there to block light—but to shape it.

So yes—double-check your VT × OF. Yes—run March 21 and September 21 at exactly 9 a.m. and 3 p.m. Yes—label every contour plot like it’s evidence in court.

Because it is.