September 18, 2026

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

Planner Execution Ledger

Planner checkpoints can acknowledge that the active plan is still suitable without writing a new public plan ID. Those no-change decisions are part of the audit trail.

🌅 Morning Cycle (12:17 AM) — iris-20260918-0016

Reflection

Validating previous cycle: iris-20260917-0644

Previous hypothesis: {“conditions”:{“outdoor_temp_peak_f”:85.0,“outdoor_rh_min_pct”:31.0,“solar_peak_w_m2”:550.0,“cloud_cover_avg_pct”:100.0,“notes”:“Cloudy forecast, but recent solar underprediction; gate wetting on live VPD.”},“stress_windows”:[{“kind”:“vpd_high”,“start”:“2026-09-17T09:30:00-06:00”,“end”:“2026-09-17T16:00:00-06:00”,“severity”:“high”,“mitigation”:“Band-coupled wet assist near live high edge; fog excess 0.20; dew floor 8F.”},{“kind”:“heat”,“start”:“2026-09-17T13:00:00-06:00”,“end”:“2026-09-17T15:30:00-06:00”,“severity”:“medium”,“mitigation”:“Pre-cool and ventilate; accept capacity-limited residual heat.”}],“rationale”:[{“parameter”:“mister_engage_kpa”,“old_value”:1.6,“new_value”:1.34,“forecast_anchor”:“Corrected VPD rises toward 2.0 kPa by 15:00; current VPD is 0.65.”,“expected_effect”:“Start wet assist at the live high edge and keep controller-attributable VPD stress below 1.5h… Result: Window scorecard: planner score 77.8, temp compliance 85.3%, VPD compliance 71.0%, 0.00 h total stress and 0.00 h dew risk. The forecast dry ramp and heat window were not represented in this partial window, so the structured high-stress prediction is not yet fully testable. Transition audit shows 54 missed and 21 already-at-value applications; tactical causal credit is weak. Current dawn reading is cool/wet at VPD 0.62 kPa with an open house-band-drift wet alert, so today’s plan must stage drying first and only enable wet assist when live VPD recovers. Score: 6/10

New finding: When transition application misses dominate a plan audit, treat apparent success as condition/controller-context evidence rather than parameter causality; carry live-VPD gating into the next plan and verify the dry ramp before escalating moisture. → Added to Lessons Learned

Hypothesis

Testing: Use live VPD to stage moisture: conservative overnight, aggressive only for the forecast dry ramp. Expected outcome: Keep dew-risk hours at 0 and controller-attributable VPD-high stress under 1.5h; avoid unnecessary overnight misting.

Setpoints

Friday September 18

Primary crop-band changes:

Tactical tunable changes:

Changed secondary parameters:

🌅 Morning Cycle (6:48 AM) — iris-20260918-0646

Reflection

Validating previous cycle: iris-20260918-0016

Previous hypothesis: {“conditions”:{“outdoor_temp_peak_f”:85,“outdoor_rh_min_pct”:27,“solar_peak_w_m2”:346,“cloud_cover_avg_pct”:100,“notes”:“Overcast forecast, but corrected VPD peaks near 2.7 kPa; use live-VPD gating and protect dew margin.”},“stress_windows”:[{“kind”:“vpd_high”,“start”:“2026-09-18T10:00:00-06:00”,“end”:“2026-09-18T15:30:00-06:00”,“severity”:“high”,“mitigation”:“Band-coupled wet assist; moisture excess 0.05, all-zone 0.25, fog excess 0.20.”},{“kind”:“heat”,“start”:“2026-09-18T12:00:00-06:00”,“end”:“2026-09-18T15:00:00-06:00”,“severity”:“medium”,“mitigation”:“Pre-cool and ventilate; accept residual heat if outdoor conditions exceed capacity.”}],“rationale”:[{“parameter”:“moisture_engage_vpd_excess_kpa”,“old_value”:0.15,“new_value”:0.05,“forecast_anchor”:“Corrected VPD rises to 1.87-2.09 kPa at noon-15:00… Result: Window score: planner 77.7, temp compliance 72.8%, VPD compliance 77.2%, 1.50 h stress: 0 h heat, 0.41 h VPD-high, 0 h cold/VPD-low, 0 h dew risk. Forecast temp was close (86.0 predicted vs 85.3 actual), but RH/solar were materially mispredicted (27% vs 46% indoor proxy; 392 vs 980 W/m2). Predicted heat window did not occur; VPD-high was limited but present. Transition attribution was weak because later audit records show many missed applications. Score: 6/10

New finding: For cool dawns followed by unexpectedly high solar, forecast RH and solar can be materially wrong even when temperature is close; delay aggressive wet-assist escalation until live VPD confirms the ramp, while keeping band-coupled assist available and accepting that heat risk may be capacity-limited. → Added to Lessons Learned

Hypothesis

Testing: Test live-VPD-gated moisture: dry first after dawn wetness, ramp assist near the high edge, and hold it until measured recovery. Expected outcome: Keep dew-risk hours at 0, controller-attributable VPD-high stress under 1.5 h, and avoid low-VPD overshoot overnight; accept structural heat miss if capacity is exceeded.

Conditions

Cool wet dawn, then a dry hot ramp; use forecast correction and live VPD gating.

Expected stress windows

Parameter rationale

Setpoints

Friday September 18

Primary crop-band changes:

Tactical tunable changes:

Changed secondary parameters:

🌆 Evening Cycle (7:07 PM) — iris-20260918-1905

Reflection

Validating previous cycle: iris-20260918-0646

Previous hypothesis: {“conditions”:{“outdoor_temp_peak_f”:86.0,“outdoor_rh_min_pct”:27.0,“solar_peak_w_m2”:392.0,“cloud_cover_avg_pct”:90.0,“notes”:“Cool wet dawn, then a dry hot ramp; use forecast correction and live VPD gating.”},“stress_windows”:[{“kind”:“vpd_high”,“start”:“2026-09-18T11:00:00-06:00”,“end”:“2026-09-18T18:00:00-06:00”,“severity”:“high”,“mitigation”:“Band-coupled assist near the live high edge; fog excess 0.15-0.20 while safety gates allow.”},{“kind”:“heat”,“start”:“2026-09-18T13:00:00-06:00”,“end”:“2026-09-18T16:00:00-06:00”,“severity”:“medium”,“mitigation”:“Pre-cool and ventilate; accept capacity-limited residual heat if outdoor conditions exceed the box.”}],“rationale”:[{“parameter”:“mister_engage_kpa”,“old_value”:1.6,“new_value”:1.04,“forecast_anchor”:“Current VPD 0.62; corrected afternoon VPD is about 2.47 at raw peak 3.07… Result: Scoped window score 68.5: temp compliance 73.4%, VPD compliance 64.0%, 0.00h heat/cold/VPD-high/VPD-low stress, dew margin stayed safe at 7.1F. The low-wet posture avoided measured VPD-low stress, but compliance remained modest and a house-too-wet alert was active by 00:55; resource cost was unavailable. Score: 6/10

New finding: For cool cloudy overnight plans, suppressing wet assist can avoid VPD-low stress and preserve dew margin, but verify active drying/dehumidification and live VPD recovery before extending the low-wet posture into dawn. → Added to Lessons Learned

Hypothesis

Testing: Overnight cold/humid hold: taper wet assist only after live VPD confirms recovery, not from forecast alone. Expected outcome: Keep dew-risk and VPD-low overshoot at 0h, limit controller-attributable VPD-high stress to under 1h, and reduce heater/vent churn.

Conditions

Cold, cloudy overnight; VPD falls below band before dawn while current dew margin is healthy.

Expected stress windows

Parameter rationale

Setpoints

Friday September 18

Primary crop-band changes:

Tactical tunable changes:

Saturday September 19

Primary crop-band changes:

Tactical tunable changes:

Changed secondary parameters:

End-of-Day Summary

Climate

Interior Light Evidence

• Crop DLI: unavailable
• Reason: source_reading_missing
• Provenance: legacy_invalid_exterior_proxy_plus_fixture_estimate
• Validity: dli-validity-v1; 2024-01-01T00:00:00+00:00 → open
• Qualified-light minutes and photoperiod remain independent of unavailable DLI.

Stress Hours

• Heat stress (>85°F): 3.3h
• VPD stress (>2.0 kPa): 6.6h
• Cold stress (<55°F): 12.8h

Economics

Equipment Runtimes

Water

• Total: 109 gal
• Mister: 23 gal

Hourly Pattern

7-Day Stress Context