April 20, 2026

Generated lab notebook from daily_summary, plan_journal, and setpoint audit data. It is intentionally chronological and may include in-progress cycles before validation.

🌅 Morning Cycle (6:20 AM) — iris-20260420-0615

Reflection

Daily Summary So Far — this cycle will be validated at the next planning cycle.

Hypothesis

Testing: Compare Monday overcast-to-clear transition (13:00 clearing) vs Tuesday all-day clear to measure cloud cover’s effect on VPD stress onset timing and water consumption. Hypothesis: Monday’s overcast morning delays VPD stress onset by 2-3 hours and reduces water usage by 20-30% vs Tuesday. Expected outcome: Monday: compliance 15-25% (tight bands), heat_stress 3-5h, vpd_high 4-6h, water 250-350 gal, cost USD 7-9. Tuesday: compliance 10-20%, heat_stress 5-7h, vpd_high 6-8h, water 350-450 gal, cost USD 8-11. Wednesday: compliance 25-35%, heat_stress 3-4h, vpd_high 3-5h, water 150-250 gal, cost USD 5-7. 72h total: USD 20-27.

Setpoints

Monday April 20

Primary crop-band changes:

Tactical tunable changes:

Tuesday April 21

Primary crop-band changes:

Tactical tunable changes:

Changed secondary parameters:

🌅 Morning Cycle (8:45 AM) — iris-20260420-0842

Reflection

Daily Summary So Far — this cycle will be validated at the next planning cycle.

Hypothesis

Testing: Modified overcast-to-clear experiment: Monday now clears at 10 AM (3h overcast) vs Tuesday all-day clear. Smaller differential than originally planned but still measurable. Compare VPD stress onset time and water usage 10 AM-2 PM window on both days. Expected outcome: Monday: compliance 12-20% (tight bands + earlier clearing), heat_stress 3-5h, vpd_high 5-7h (was 4-6h before reforecast), water 300-400 gal, cost USD 7-10. Tuesday: compliance 10-18%, heat_stress 5-7h, vpd_high 6-9h, water 350-450 gal, cost USD 8-11. Wednesday: compliance 8-15% (worst day — 86F + wind), heat_stress 5-8h, vpd_high 5-8h, water 250-350 gal, cost USD 6-9. 72h total: USD 21-30.

Setpoints

Monday April 20

Primary crop-band changes:

Tactical tunable changes:

Tuesday April 21

Primary crop-band changes:

Tactical tunable changes:

Wednesday April 22

Primary crop-band changes:

Tactical tunable changes:

Changed secondary parameters:

🌅 Morning Cycle (8:57 AM) — iris-20260420-0850

Result: Score 16.6, compliance 11.6% (temp 17.2%, VPD 28.6%). Heat stress 15.82h dominated — largely structural from tight crop band (temp_high ~67.5°F overnight while slab held greenhouse at 70-72°F). VPD high 16.97h from extreme dry outdoor air (11-14% RH). Cold stress 4.0h (morning band ramp, structural). Peak temp 90.4°F, peak VPD 3.04. Cost USD 9.51 (USD 5.60 gas, USD 2.23 electric, USD 1.68 water). Heat1 had 53 transitions (oscillation), heat2 17 transitions. 4 ESP32 reboots (guru/panic + task wdt). vpd_target_center unconfirmed alerts persisted for hours (validates platform lesson). Multiple firmware relief ceiling and vent latch alerts during afternoon peak. Leak alerts during misting. Water 348 gal (238 mister).

Reflection

Validating previous cycle: iris-20260419-1903

Previous hypothesis: Late-day catch-up plan — previous plan iris-20260419-0022 ran unmodified for 18+ hours with no intraday replanning. Today’s scorecard: 25.5 score, 17.8% compliance, 14.4h cold stress (structural — crop band alignment on spring mornings per lesson #6), 4.2h heat stress, 4.5h VPD high stress. Peak reached 85.6°F/2.02 VPD at 2 PM MDT — better than forecast predicted but still 4+ hours outside band. Evening vpd_high=0.60 was too aggressive — drove south zone to 0.08 VPD (near condensation). This plan corrects the evening posture immediately and covers Mon-Wed.

Monday is the second extreme dry day: overcast morning (100% cloud until 14:00, then clearing to 8%), 82°F peak outdoor, 4% RH. The overcast morning is an advantage — slower heating ramp gives more time before misting is needed. But afternoon clearing will bring a fast spike. Tuesday is the worst day: clear skies, 83°F peak, 9-10% RH, earliest VPD ramp. Wednesday brings relief — overcast with 40-48% RH overnight.

Strategy: (1) Relax evening posture now — VPD 0.52 at 67°F is fine, stop all humidification. (2) Monday overcast morning = delayed aggression, start misting ramp at 10:00 not 06:00. (3) Monday/Tuesday afternoon = maximum misting with fog_esc 0.3, sealed vent 900s, accept heat stress above 85°F per lesson #5. (4) Wednesday = lightest touch as moisture returns.

{"conditions": [{"day": "Sunday evening", "outdoor_high_f": 69, "outdoor_low_f": 61, "outdoor_rh_pct": 14, "cloud_pct": 0, "severity": "dry_cooling"}, {"day": "Monday", "outdoor_high_f": 82, "outdoor_low_f": 42, "outdoor_rh_pct": 5, "cloud_pct": 60, "severity": "extreme_dry_hot"}, {"day": "Tuesday", "outdoor_high_f": 83, "outdoor_low_f": 48, "outdoor_rh_pct": 10, "cloud_pct": 30, "severity": "extreme_dry_hot"}, {"day": "Wednesday", "outdoor_high_f": 84, "outdoor_low_f": 53, "outdoor_rh_pct": 30, "cloud_pct": 75, "severity": "hot_drying"}], "stress_windows": [{"day": "Monday", "type": "cold", "hours": "04:00-09:00", "mitigation": "bias_heat +4, gas heater stages below 55F"}, {"day": "Monday", "type": "vpd_high+heat", "hours": "13:00-19:00", "mitigation": "fog_esc 0.3, gap 20s, sealed 900s, accept heat above 85F"}, {"day": "Tuesday", "type": "vpd_high+heat", "hours": "10:00-19:00", "mitigation": "fog_esc 0.3, gap 20s, sealed 900s, earliest engage at 1.3 kPa"}, {"day": "Wednesday", "type": "vpd_high", "hours": "12:00-18:00", "mitigation": "moderate misting, humidity recovery expected PM"}], "rationale": "Catch-up from stale plan. Evening posture was over-aggressive (vpd_high 0.60 drove south to 0.08 VPD). Three consecutive extreme dry days with escalating severity. Physics-limited cooling per lesson #5. Monday overcast AM delays the stress window — use it. Tuesday clear sky means earliest ramp. Wednesday moisture returns."}

Result: Score 16.6, compliance 11.6% (temp 17.2%, VPD 28.6%). Heat stress 15.82h dominated — largely structural from tight crop band (temp_high ~67.5°F overnight while slab held greenhouse at 70-72°F). VPD high 16.97h from extreme dry outdoor air (11-14% RH). Cold stress 4.0h (morning band ramp, structural). Peak temp 90.4°F, peak VPD 3.04. Cost USD 9.51 (USD 5.60 gas, USD 2.23 electric, USD 1.68 water). Heat1 had 53 transitions (oscillation), heat2 17 transitions. 4 ESP32 reboots (guru/panic + task wdt). vpd_target_center unconfirmed alerts persisted for hours (validates platform lesson). Multiple firmware relief ceiling and vent latch alerts during afternoon peak. Leak alerts during misting. Water 348 gal (238 mister). Score: 3/10

New finding: Heat stress hours are heavily inflated by tight nighttime crop band (temp_high ~67.5°F) while slab thermal mass holds greenhouse at 70-72°F overnight. Of 15.82h heat stress, estimated 8-10h are structural overnight band misalignment, not actual overheating. Heater oscillation (53 transitions) suggests the controller is fighting to maintain temp_low while simultaneously being above temp_high — an impossible band situation. The 4.5°F nighttime band (63-67.5°F) with slab retention of 70°F+ guarantees continuous out-of-band readings regardless of control strategy. → Added to Lessons Learned

Hypothesis

Testing: Mon/Tue overcast-to-clear paired comparison preserved (Mon clears 10 AM vs Tue all-day clear). Compare VPD stress onset and water usage 10 AM-2 PM both days. Expected outcome: Mon: compliance 12-20%, vpd_high 5-7h, cost USD 7-10. Tue: compliance 10-18%, vpd_high 6-9h, cost USD 8-11. Wed: 8-15% (worst), cost USD 6-9. Thu: 40-60%, cost USD 3-5. Fri: 50-70%, cost USD 4-7. Sat: 30-50%, cold_stress 4-6h, cost USD 6-10 (gas). Sun: 35-55%, cost USD 5-8. 7-day total: USD 39-60.

Setpoints

Monday April 20

Primary crop-band changes:

Tactical tunable changes:

Tuesday April 21

Primary crop-band changes:

Tactical tunable changes:

Wednesday April 22

Primary crop-band changes:

Tactical tunable changes:

Thursday April 23

Primary crop-band changes:

Tactical tunable changes:

Friday April 24

Primary crop-band changes:

Tactical tunable changes:

Saturday April 25

Primary crop-band changes:

Tactical tunable changes:

Sunday April 26

Primary crop-band changes:

Tactical tunable changes:

Monday April 27

Primary crop-band changes:

Tactical tunable changes:

Changed secondary parameters:

🌆 Evening Cycle (7:50 PM) — iris-20260420-1945

Result: Evaluated overnight window only (8 PM Apr 20 → 6 AM Apr 21) — plan replaced at sunrise before daytime predictions could be tested. Overnight temps held 62-65°F (hypothesis predicted 65-70°F from slab retention — 3-5°F optimistic). VPD dipped to 0.40 kPa at 2 AM and hit 0.26 kPa during reboot cluster, triggering vpd_extreme warnings — not anticipated by hypothesis. 5 ESP32 reboots between 1:30-2:26 AM (guru/panic + task wdt) disrupted control continuity. bias_cool +5 successfully prevented vent oscillation during heating cycles — heat1 transitions appear reduced vs previous night (53→~30 est). Heat2 ran 2.69h for gas heating. Evening alerts included firmware vent latch and relief ceiling events from daytime carry-over (10-11 PM). Overall: adequate overnight management, slab prediction slightly optimistic, VPD low risk in the 1-3 AM window needs attention in future overnight plans.

Reflection

Validating previous cycle: iris-20260420-0850

Previous hypothesis: JASON-AWAY CONSERVATIVE STAY-IN-BAND: 7-day coverage through Sunday Apr 27. Two distinct regimes:

PHASE 1 (Mon-Wed): Extreme dry-hot. Outdoor 81-86°F, 5-8% RH, VPD 3.0-4.0 kPa. Conservative-aggressive misting: lower engage thresholds (1.3 kPa), generous water budgets (700 gal/day), moderate sealed windows (600s vs 900s — safer without Jason to catch thermal traps). Accept physics-limited heat stress but buy every VPD minute we can. Wednesday worst: 86°F + 27-37 mph wind, shortened sealed windows (480s).

PHASE 2 (Thu): Transition. Overcast, 64°F peak. Minimal intervention needed.

PHASE 3 (Fri-Sun): Cold front. 32-56°F outdoor, 50-97% RH. Full heating regime. Gas heater primary for efficiency. Minimal misting — ambient humidity keeps VPD naturally low. Dehumid may trigger on Sat/Sun (outdoor RH 80-97%). Greenhouse will never freeze (80K BTU combined + slab mass retains 7-8°F above outdoor).

Conservative philosophy: favor band compliance over optimization. Wider misting engagement, generous budgets, shorter sealed windows, longer heat cycles. Accept higher utility cost for safer autonomous operation. Every sunrise waypoint carries full parameter state for reboot resilience.

{"conditions": [{"day": "Monday", "outdoor_high_f": 81, "outdoor_low_f": 48, "outdoor_rh_pct": 5, "cloud_pct": 15, "severity": "extreme_dry_hot"}, {"day": "Tuesday", "outdoor_high_f": 83, "outdoor_low_f": 48, "outdoor_rh_pct": 8, "cloud_pct": 0, "severity": "extreme_dry_hot"}, {"day": "Wednesday", "outdoor_high_f": 86, "outdoor_low_f": 53, "outdoor_rh_pct": 5, "cloud_pct": 60, "severity": "extreme_dry_hot_windy"}, {"day": "Thursday", "outdoor_high_f": 64, "outdoor_low_f": 46, "outdoor_rh_pct": 40, "cloud_pct": 100, "severity": "mild_overcast"}, {"day": "Friday", "outdoor_high_f": 57, "outdoor_low_f": 35, "outdoor_rh_pct": 78, "cloud_pct": 100, "severity": "cold_wet"}, {"day": "Saturday", "outdoor_high_f": 42, "outdoor_low_f": 32, "outdoor_rh_pct": 96, "cloud_pct": 100, "severity": "cold_wet_freezing"}, {"day": "Sunday", "outdoor_high_f": 52, "outdoor_low_f": 32, "outdoor_rh_pct": 86, "cloud_pct": 100, "severity": "cold_wet"}], "stress_windows": [{"kind": "vpd_high", "start": "2026-04-20T10:00-06:00", "end": "2026-04-20T19:00-06:00", "severity": "high", "mitigation": "aggressive misting engage 1.3, gap 20-25s, fog_esc 0.3"}, {"kind": "vpd_high", "start": "2026-04-21T09:30-06:00", "end": "2026-04-21T19:00-06:00", "severity": "extreme", "mitigation": "maximum aggression — gap 15s at peak, 700 gal budget"}, {"kind": "vpd_high", "start": "2026-04-22T10:00-06:00", "end": "2026-04-22T18:00-06:00", "severity": "extreme", "mitigation": "shortened sealed 480s — wind makes sealing less effective"}, {"kind": "heat_stress", "start": "2026-04-22T14:00-06:00", "end": "2026-04-22T18:00-06:00", "severity": "extreme", "mitigation": "86F + wind — accept"}, {"kind": "cold_stress", "start": "2026-04-25T00:00-06:00", "end": "2026-04-26T12:00-06:00", "severity": "moderate", "mitigation": "gas heater primary, 80K BTU, slab buffer"}, {"kind": "vpd_low", "start": "2026-04-25T06:00-06:00", "end": "2026-04-26T10:00-06:00", "severity": "low", "mitigation": "fans dehumid if VPD < 0.35"}], "rationale": [{"parameter": "mister_engage_kpa", "old_value": 1.6, "new_value": 1.3, "forecast_anchor": "away-plan: compliance over optimization", "expected_effect": "Earlier misting reduces peak VPD 0.3-0.5 kPa, +20% water"}, {"parameter": "mist_max_closed_vent_s", "old_value": 900, "new_value": 600, "forecast_anchor": "No human oversight for thermal trap", "expected_effect": "Less thermal risk, slightly less retention"}, {"parameter": "mister_water_budget_gal", "old_value": 500, "new_value": 700, "forecast_anchor": "Apr 13 budget exhaustion lesson", "expected_effect": "40% headroom. FW-9 still backstops VPD>3.0"}]}

Result: Evaluated overnight window only (8 PM Apr 20 → 6 AM Apr 21) — plan replaced at sunrise before daytime predictions could be tested. Overnight temps held 62-65°F (hypothesis predicted 65-70°F from slab retention — 3-5°F optimistic). VPD dipped to 0.40 kPa at 2 AM and hit 0.26 kPa during reboot cluster, triggering vpd_extreme warnings — not anticipated by hypothesis. 5 ESP32 reboots between 1:30-2:26 AM (guru/panic + task wdt) disrupted control continuity. bias_cool +5 successfully prevented vent oscillation during heating cycles — heat1 transitions appear reduced vs previous night (53→~30 est). Heat2 ran 2.69h for gas heating. Evening alerts included firmware vent latch and relief ceiling events from daytime carry-over (10-11 PM). Overall: adequate overnight management, slab prediction slightly optimistic, VPD low risk in the 1-3 AM window needs attention in future overnight plans. Score: 5/10

New finding: Overnight slab retention in spring with outdoor 50-60°F delivers 62-65°F, not 65-70°F as commonly predicted. Use 62-66°F as the overnight slab equilibrium range when outdoor is 48-55°F. VPD can dip below 0.35 kPa between 1-3 AM when RH climbs above 75% — consider whether vpd_low should be relaxed overnight to avoid unnecessary dehumid cycling. → Added to Lessons Learned

Hypothesis

Testing: Tue vs Wed paired comparison under near-identical heat/dryness (84°F/6% vs 85°F/5%) but different cloud profiles (Tue clear AM/variable PM vs Wed overcast AM/clearing PM) and wind (Tue light vs Wed 20+ mph). Compare: (1) VPD stress onset time, (2) fog activation count, (3) water usage, (4) whether Wed wind improves or degrades compliance vs Tue calm. Expected outcome: Tuesday: compliance 10-15%, heat_stress 6-9h (structural band + physics), vpd_high 8-12h, water 300-450 gal, cost USD 7-11. Wednesday: compliance 12-18% (cloudier morning delays onset), heat_stress 5-8h, vpd_high 7-10h, water 250-400 gal, cost USD 6-9. Thursday: compliance 25-35% (much milder), heat_stress 0-2h, vpd_high 2-4h, cost USD 4-7 (higher gas from cold overnight). 72h total: USD 17-27.

Setpoints

Monday April 20

Primary crop-band changes:

Tactical tunable changes:

Tuesday April 21

Primary crop-band changes:

Tactical tunable changes:

Wednesday April 22

Primary crop-band changes:

Tactical tunable changes:

Thursday April 23

Primary crop-band changes:

Tactical tunable changes:

Changed secondary parameters:

Daily Summary So Far

Climate

Stress Hours

• Heat stress (>85°F): 17.7h
• VPD stress (>2.0 kPa): 17.4h
• Cold stress (<55°F): 1.0h

Economics

Equipment Runtimes

Water

• Total: 565 gal
• Mister: 133 gal

Crop Health (Gemini Vision)

Hourly Pattern

7-Day Stress Context