plant-light-doctor

Use when choosing plants for a specific window or diagnosing why a houseplant is leggy, pale, or not blooming. Computes real monthly light (DLI) for any window orientation, floor, and latitude, matches plants against light needs, and specs a grow light if nature is not enough.

Overview

"Low light / medium light / bright indirect" is the least useful classification ever applied to plants. Two windows in the same apartment can differ by a factor of 8 in actual light delivered — and the same north window in Oslo vs. Athens is a different planet. Meanwhile people buy fiddle-leaf figs for dark corners, watch them slowly etiolate, and conclude they have a "black thumb".

This skill replaces vibes with physics. It computes DLI (Daily Light Integral, mol·m⁻²·day⁻¹) — the horticulture industry's gold-standard light metric — for a specific window: orientation, floor height, outdoor obstructions, latitude, sky climate. It then matches that DLI against a library of common houseplants' actual requirements, month by month, and delivers one of four verdicts: THRIVES / SURVIVES / STRUGGLES / IMPOSSIBLE. If nature falls short, it sizes a grow light: required PPFD, photoperiod, and realistic wattage.

When to Use

• "What plants will survive in my north-facing window?" (or any orientation)
• "Why is my monstera leggy / why won't my orchid rebloom / why is everything pale?"
• "I live on the 1st floor across from a building — what can I actually grow?"
• "Do I need a grow light? Which one, how strong, how many hours?"
• "Can I grow herbs / lettuce / tomatoes / succulents on my windowsill?"
• Choosing a plant BEFORE buying it, for a specific spot in a specific home.
• Winter light audits ("everything dies in November") — seasonal DLI decline charts.

Don't use for: outdoor garden planning (use seed-calendar-planner style frost-date logic), plant disease diagnosis, or watering/fertilizer questions.

How It Works — the physics in 60 seconds

1. Extraterrestrial irradiance. Solar constant 1361 W/m², adjusted for day angle → clear-sky PPFD at ground level for each month (Bouguer-style model).
2. Day length per month from latitude (solar declination formula).
3. Window capture model. Direct sun hours on vertical glass by orientation (S gets max, E/W half, N ≈ diffuse only), scaled by:
   • floor penalty (ground floor + opposite building = big loss)
   • obstruction factor (user-reported 0-100% blockage during sun hours)
   • glass transmission (~85% typical double-glazed, tinted glass worse)
   • overhang/eave shading (reduces high summer sun more than low winter sun)
4. Sky climate multiplier. Cloudy oceanic (UK, PNW) vs. continental clear (interior US, Mediterranean) — up to 2× swing in delivered DLI.
5. DLI = sum of (PPFD × seconds) over the day → mol·m⁻²·day⁻¹ per month.
6. Plant matching. Each plant in the library has a DLI need range (e.g., herbs 8-14, lettuce 12-17, tomato 20-30+). Verdict rules:
   • window DLI ≥ need → THRIVES
   • within 15% below need → SURVIVES (slow growth, may not bloom)
   • 15-45% below → STRUGGLES (etiolation within months)

   • 45% below or <2 mol floor → IMPOSSIBLE

7. Grow-light sizing. Gap = (need − worst-month DLI). Required PPFD = gap × 10⁶ / (3600 × photoperiod). Then fixture wattage from efficacy (120 lm/W ≈ 2.3 µmol/J typical consumer LED boards).

Files

File	Purpose
scripts/plant_light_analyzer.py	Full CLI analyzer: DLI by month, plant verdicts, grow-light spec
references/plant-dli-table.md	DLI requirements for 40+ plants + symptom-to-light troubleshooting
references/light-physics.md	The full model, formulas, worked examples, measurement methods
README.md	Real-world problem statement

Usage

bash

# Audit a north window in Berlin, 3rd floor, clear view
python3 scripts/plant_light_analyzer.py --orientation N --latitude 52.5 --floor 3

# South window, ground floor, opposite building blocks 70% of sun hours, cloudy city
python3 scripts/plant_light_analyzer.py --orientation S --latitude 41.0 --floor 1 \
    --obstruction 0.7 --climate cloudy

# Can I grow cherry tomatoes? Check a specific plant
python3 scripts/plant_light_analyzer.py --orientation E --latitude 50.1 --floor 2 \
    --check "cherry tomato"

# Include a grow-light recommendation
python3 scripts/plant_light_analyzer.py --orientation NW --latitude 55.7 --floor 5 \
    --grow-light --target-dli 14

Workflow

1. Collect the facts (ask the user, or infer): window orientation (compass direction the window FACES — the direction you look when standing inside), city or latitude, floor number, what's outside the window (open street / trees / opposite wall at X meters), glass type.
2. Run the analyzer with those parameters. Read the monthly DLI table — note the worst month (usually December in the northern hemisphere).
3. Match plants: either --check "plant name" for a purchase decision, or read the ranked table of what thrives in THIS window.
4. If the user wants a plant the window can't support, size the grow light (--grow-light --target-dli <need>) — give PPFD, hours/day, and a realistic wattage + fixture-distance recommendation.
5. For "why is my plant sad" calls: cross-reference the symptom (references/plant-dli-table.md § Symptoms) with the window's verdict.

Common Pitfalls

1. Orientation confusion. Users say "my window gets the morning sun" — that's an EAST window. The analyzer wants where the glass FACES. A window facing south in the northern hemisphere = sun-drenched; in Sydney/Buenos Aires (southern hemisphere) flip everything — the script handles this via latitude sign.
2. "Bright indirect" folk wisdom. Monstera "tolerates low light" means it survives at DLI 4, not that it thrives. Always quote the verdict tier, and set expectations for SURVIVES (slow, smaller leaves, no fenestration).
3. Forgetting winter. A south window in Hamburg is a jungle in July and a cave in December. Always check the monthly table — recommend plants by the WORST month, not the average.
4. Over-trusting the model. ±30% is honest. Nearby buildings, awnings, balconies, screens, and net curtains can dominate. When the user reports a big obstruction, trust --obstruction over theory.
5. Grow-light distance. PPFD falls with the inverse square of distance. The script's wattage figure assumes 30-45 cm above canopy for small fixtures.

Example Session

User: "What can I grow on a north-facing windowsill, 4th floor, Helsinki?"

Run: --orientation N --latitude 60.2 --floor 4 --climate cloudy Output: worst-month DLI ≈ 1.2 (December), summer peak ≈ 4. Verdicts: pothos/ZZ/snake plant SURVIVE, herbs STRUGGLE, lettuce IMPOSSIBLE. Recommendation: accept foliage plants, or add a 20 W full-spectrum bar (≈ 250 µmol/m²/s at 30 cm) for 12 h/day to unlock herbs and lettuce.

Verification Checklist

• Worst month identified and quoted in the answer
• Every plant recommendation tied to a verdict tier, not vibes
• Hemisphere handled (southern latitude flips S/N exposure)
• Grow-light spec includes PPFD + hours + wattage, not just "buy a lamp"