Hydroponic tower humidity in Bangkok: what the sensor data actually shows across the full system

Key takeaways

A hydroponic tower in a Bangkok apartment is not one humidity environment -- it is four stacked on top of each other: the reservoir zone, the mid-tower canopy, the upper tower, and the root zone inside each pot site. Each behaves differently and each requires a different intervention.

An uncovered reservoir in a sealed 12 square metre Bangkok room raises ambient humidity by 8 to 14 percentage points above baseline -- but the contribution is largest between 20:00 and 23:00 when the room cools faster than the water. Daytime readings understate the overnight impact by 6 to 9 percentage points. (Plant Mood grow log data, 2024)

The mid-tower zone at 60 to 90cm above the reservoir is the highest mold-risk position in a Bangkok tower -- not the base. AHT20 logs show mid-tower humidity crossing 85 percent for 3 to 4 hours after sunset every night in unventilated setups, coinciding with peak Botrytis spore germination conditions. (Plant Mood grow log data, 2024)

Covering 70 percent or more of the reservoir surface with close-fitting net pot collars or a lid reduced ambient humidity by 5 to 8 percentage points and cut evaporative water loss by 61 percent -- from 2.3 litres per day to 0.9 litres per day -- across a four-week comparison trial. (Plant Mood grow log data, 2024)

Root zone humidity inside tower pot sites runs 18 to 26 percentage points above ambient room air across all logged cycles. At 97 to 99 percent RH inside the pot site during overnight peaks, root zone VPD drops below 0.1 kPa -- suppressing root hair elongation and shifting root morphology toward fewer, longer hairs with 31 percent lower density than roots grown with active pot site airflow.

A single ambient room sensor tells you almost nothing useful about any of the four humidity environments in a Bangkok tower. The reservoir zone, mid-tower canopy, and root zone each require their own sensor placement to produce actionable data.

TL;DR

The humidity problem in a Bangkok hydroponic tower is not one problem. It is four, stacked vertically, each with a different mechanism and a different solution. The reservoir adds moisture to the room continuously -- more at night than during the day, and through surface area not volume. The mid-tower zone accumulates that moisture in a pocket of restricted airflow and crosses the mold threshold for hours every evening. The root zone inside each pot site runs in an entirely separate climate from the room, 18 to 26 points higher, shaping root hair development in ways the ambient sensor never shows. A wall-mounted hygrometer is the wrong tool to manage any of this. The tower needs sensors in the right places and interventions timed to the right windows.

Four trays, four sensors, four different numbers at the same moment

It started as a simple check. I wanted to know whether the ambient room reading -- 74 percent at 14:00 on a March afternoon -- was representative of what was happening inside the tower. So I placed AHT20 sensors at four positions: reservoir level, mid-tower at 70cm, top of tower, and inside a pot site with the sensor element at root depth. I checked all four at the same timestamp. Reservoir level: 93 percent. Mid-tower: 77 percent. Top of tower: 69 percent. Inside the pot site: 95 percent. The ambient room sensor read 74 percent.

Four numbers in the same tower at the same moment, ranging from 69 to 95 percent. The wall reading of 74 percent was accurate for the air at that height. It was not representative of any of the environments the crop was actually living in.

I pulled the same four sensors at 23:00. Reservoir level: 94 percent. Mid-tower: 89 percent. Top of tower: 82 percent. Pot site interior: 97 percent. The ambient room read 80 percent. What had been a 25-point spread at 14:00 had compressed at the base and expanded at the mid and upper positions. The evening cooling had done its work -- the room had lost heat and moisture capacity faster than the reservoir and the pot sites had. Mid-tower had crossed 85 percent and was sitting there for the third consecutive hour. That is a Botrytis window. Every night. And a wall sensor reading 80 percent would never show it.

Why the tower creates four separate humidity environments

A hydroponic tower in a Bangkok apartment is not a uniform humidity environment that the room sensor describes. It is a stack of distinct microclimates, each driven by a different moisture source and each responding to the daily heat cycle in a different way.

The reservoir at the base is a continuous evaporative source. Water at 26 to 30 degrees Celsius in a Bangkok room evaporates into the air immediately above it, holding that zone at 88 to 95 percent relative humidity across all logged cycles regardless of what the ambient room reads. This is the source that drives the whole system. Every percentage point of humidity that reaches mid-tower and beyond originated at the reservoir surface.

The mid-tower zone is where the humidity accumulates. Warm, moist air rising from the reservoir encounters the densest canopy on the tower at this height. Canopy restricts airflow. In restricted airflow, evaporated moisture pools rather than dispersing. The mid-tower position at 60 to 90cm above the reservoir is where elevated humidity meets the lowest air movement -- the combination that the controlled environment research consistently identifies as highest infection pressure, not just highest absolute humidity.

The Bangkok daily heat cycle adds a second mechanism on top of this baseline. During peak afternoon heat, ambient temperature is highest, the air has the most capacity to hold moisture, and the humidity gradient from base to top compresses. After sunset, the room cools faster than the reservoir -- water holds heat longer than air. The reservoir continues evaporating at nearly the same rate while the room air loses the capacity to hold that moisture quietly. Mid-tower and upper-tower readings spike. The post-sunset rebound is not a secondary effect. It is the window when mold pressure is highest, and it is invisible to a grower who checks conditions once in the morning.

Inside each pot site, a fourth environment operates by entirely different rules. The pot site cavity is partially enclosed -- 8 to 12 percent open area is typical in commercial tower designs. Moisture enters from root zone transpiration, from the growing medium, and from the root surface itself. It exits slowly through the small opening fraction. The ambient room conditions influence the pot site only at the margin. At 14:00 when the room reads 74 percent, the pot site reads 94 to 96 percent. At 23:00 when the room reads 80 percent, the pot site reads 97 to 99 percent. And critically: the pot site peak occurs 60 to 90 minutes after the ambient peak, because the enclosed geometry delays the response to ambient changes. Ventilation timed to the ambient peak is ending before the pot site has cleared.

In a four-row tower cycle logged in March 2024, mid-tower humidity at 70cm ranged from 73 percent at 14:00 to 89 percent at 23:00 across a five-day window. Base humidity was stable between 91 and 94 percent across the same period. Top-tower humidity ranged from 67 percent at 14:00 to 81 percent at 23:00. A grower reading the 14:00 mid-tower number would see 73 percent and draw no concern. The same grower logging at 23:00 would see that mid-tower had been above 85 percent for three consecutive hours. The problem was not in the number. It was in the timestamp. (Plant Mood grow log data, 2024)

The mold threshold cited throughout this article -- 85 percent relative humidity -- is the commonly referenced value for Botrytis cinerea spore germination in controlled environment agriculture pathology literature (Elad et al., 2004, Phytoparasitica). Pythium species become active at lower thresholds, above approximately 85 percent at the substrate surface for sustained periods. Neither pathogen is introduced by the grower -- both are present in the environment. Managing the conditions is the only available intervention at the tray or root zone level without chemical treatment.

What the logged data shows across each humidity zone

A 48-hour baseline trial comparing an empty reservoir room against a full reservoir room in the same sealed 12 square metre space showed a delta of 4 percentage points at 14:00 and 13 percentage points at 23:00 on the same day. The reservoir was not contributing more moisture after sunset -- the room was losing the capacity to hold the moisture that had been accumulating all day. Over five consecutive sealed days, the full-reservoir room showed morning baseline humidity drifting upward by 2 to 3 percentage points per day -- a ratchet effect that does not appear in the empty-reservoir control. By day four, the full-reservoir room was starting each morning 8 to 10 percentage points above where it started on day one. (Plant Mood grow log data, 2024)

The reservoir raises ambient humidity most after sunset, not during the day

Growers who measure ambient humidity during the day and find it acceptable may be measuring before the ratchet has loaded. The relevant measurement is at 23:00 in the first week of a cycle, not at midday. A room that reads 72 percent at 14:00 can be running 83 percent at 23:00 and drifting upward across days -- creating mold conditions that no single daytime reading shows.

A four-week side-by-side comparison of covered versus uncovered reservoirs in matched tower setups showed the covered reservoir room averaged 6 percentage points lower ambient humidity during weeks one and two. By week three, with canopy fully established, the differential had compressed to 2 to 3 percentage points as crop transpiration became the dominant moisture source. Water loss from the uncovered reservoir averaged 2.3 litres per day versus 0.9 litres per day from the covered reservoir -- a 61 percent reduction that also meant 0.3 mS/cm per day EC drift in the uncovered system requiring daily correction versus every 3 to 4 days in the covered system. Algae coverage on uncovered reservoir interior surfaces reached 35 to 45 percent by day 14; the covered reservoir showed less than 2 percent at day 28. (Plant Mood grow log data, 2024)

Covering the reservoir reduced water loss by 61 percent and algae coverage from 35-45 percent to under 2 percent -- but the humidity benefit compressed to near zero by week three

Cover the reservoir for algae control and water retention first. The humidity benefit is real but time-limited within a cycle. By week three, the canopy is the dominant moisture source and the reservoir intervention is no longer the relevant lever. Do not use reservoir coverage as the primary humidity strategy and expect it to hold through a full crop cycle.

AHT20 logs from a four-row tower cycle in March 2024 showed the mid-tower zone at 70cm above the reservoir reading above 85 percent relative humidity from approximately 20:00 to 23:00 on five consecutive nights -- 3 to 4 hours per night. The base zone was stable at 91 to 94 percent across the same period. The top zone ranged from 67 percent at 14:00 to 81 percent at 23:00. The mid-tower zone was the only position that consistently crossed the Botrytis germination threshold and held there for a duration sufficient for spore germination. A directional circulation fan creating cross-tower airflow at mid-canopy height reduced the base-to-top gradient from 14 to 19 percentage points to 7 to 10 percentage points. Passive ventilation with the door open produced no meaningful gradient reduction. (Plant Mood grow log data, 2024)

The mid-tower zone crosses 85 percent for 3 to 4 hours every night in unventilated setups

The fan is not optional in a Bangkok tower setup. It is the only intervention that consistently reduces the mid-tower humidity peak to below the mold threshold. Position matters more than size -- a single fan creating cross-flow at mid-canopy height outperforms two fans moving air parallel to the tower face.

AHT20 sensors mounted inside tower pot sites with the element at mid-root-zone depth logged 18 to 26 percentage points higher relative humidity than the ambient room sensor across all Plant Mood cycles. Pot site interior humidity reached 97 to 99 percent for 4 to 6 hours per night during the post-sunset window. Root hair density in crops from high pot site humidity cycles averaged 31 percent lower than in cycles with active pot site airflow from a repositioned fan. Pot site humidity peaks occurred 60 to 90 minutes after the ambient humidity peak, meaning ventilation timed to the ambient peak was ending before the pot site had cleared. (Plant Mood grow log data, 2024)

Root zone humidity inside pot sites runs 18 to 26 percentage points above ambient and peaks 60 to 90 minutes after the ambient peak

Root zone management requires root zone data. An ambient sensor cannot tell you what is happening inside a pot site. The 18 to 26 point gap is consistent and large enough that ambient VPD calculations systematically overestimate the driving gradient at the root surface by a factor of 10 in peak conditions. Root hair morphology at harvest -- dense and short versus sparse and elongated -- is a visible retrospective indicator of whether root zone humidity was within a functional range across the cycle.

Plant Mood grow log data, 2023-2024

The highest-humidity position in the tower is not where mold events occur

The base of the tower is saturated, but saturation there is accompanied by natural convective airflow from evaporation. Water evaporating from the reservoir surface creates upward air movement. The base is wet, but it is not stagnant. Root rot events trace to the base through an irrigation and drainage management mechanism -- not primarily through a humidity-stagnation mechanism.

The mid-tower zone at 60 to 90cm is where elevated humidity meets the most restricted airflow. Dense mid-canopy growth blocks the convective upward movement from the base. Humidity accumulates in a pocket with no exit. The combination that pathology research identifies as highest infection pressure -- elevated humidity plus restricted airflow -- is concentrated at mid-tower, not at the base where the humidity source sits.

The post-sunset rebound amplifies this. When the room cools and mid-tower humidity spikes, the base barely moves -- it was already at near-saturation. The mid-tower zone is the position experiencing the most volatile humidity swing across the daily cycle, and the position where that swing crosses and holds the mold threshold longest. Growers who focus remediation effort on the base, cover the reservoir more thoroughly, or add aeration to the root zone at base level are addressing the wrong position for mid-tower mold events.

Crown rot events in tower cycles at Plant Mood have traced to row two -- mid-tower -- in the majority of instances, not to the base row despite the base running 15 to 20 percentage points higher absolute humidity. Pulling the AHT20 logs at 23:00 for the five days before a mid-tower event consistently shows 3 to 4 hours per night above 85 percent at that specific height. The same logs show the base stable at 92 to 94 percent -- above the threshold, but not spiking and not the position where the event occurred. (Plant Mood grow log data, 2024)

The reservoir base of the tower, sitting at 91 to 95 percent relative humidity, should be the primary mold concern. It is the wettest zone, it has the most direct water contact, and it is where the moisture load originates. Managing the base should be the priority.

One sensor versus four: what each setup tells you and what it misses

Single thermohydrometer at room level

This is still the most common monitoring setup among small Bangkok tower growers. It integrates ambient air that has mixed across the full space. At 14:00 it may read 72 percent while the reservoir zone reads 93 percent, the mid-tower reads 77 percent, and the pot site reads 95 percent. At 23:00, it may read 80 percent while the mid-tower is at 89 percent and the pot site is at 97 percent. The room reading has essentially no predictive value for mold events at specific tower positions.

Growers using this setup will see mold events after they are established and will not be able to connect them to a specific time window or height zone. Every intervention will be reactive. The most common complaint pattern -- 'my room humidity is fine but I keep getting mold' -- is almost always this setup.

Two AHT20 sensors -- one at base, one at top

Captures the gradient endpoints and shows whether the differential is widening or compressing across the day. Misses the mid-zone accumulation that is the highest mold-risk position, and cannot resolve the pot site interior environment. Better than one sensor but still blind to the most operationally important zone.

The grower can see that a gradient exists and can track whether ventilation changes affect the overall differential. Cannot identify which specific row position needs intervention or what the root zone is experiencing.

Four AHT20 sensors -- reservoir level, mid-tower 60cm, upper tower 110cm, pot site interior

This setup produced the Plant Mood gradient data. Four sensors at consistent positions across cycles generate a profile showing the gradient shape, the daily cycle behaviour at each zone, and the specific position where humidity peaks after sunset. AHT20 units cost approximately 80 to 120 Thai Baht each. The information gap between two sensors and four sensors is disproportionate to the cost difference.

A grower with this setup can identify the specific height zone generating mold risk, time ventilation to the evening rebound window, detect when a cycle is trending toward a high-risk gradient profile before mold establishes, and log root zone conditions that ambient monitoring never captures. This is the minimum setup for understanding what the tower is actually doing.

Covered versus uncovered reservoir

Covered reservoir: 6 percentage points lower ambient humidity in weeks one and two (compressing to 2 to 3 points by week three), 61 percent less water loss, 0.1 mS/cm per day EC drift versus 0.3 per day uncovered, less than 2 percent algae coverage at day 28 versus 35 to 45 percent. Tradeoffs: 0.8 degrees Celsius warmer water temperature, lower dissolved oxygen without active aeration, slower root zone problem detection without scheduled inspection.

Cover the reservoir and add an airstone simultaneously. Schedule twice-weekly lid-off root zone inspections. Do not expect the humidity benefit to persist through week three -- the canopy overtakes the reservoir as the dominant moisture source at that point and a different intervention is needed.

What growers say when the tower humidity problem cannot be explained

My room humidity reads fine all day. I still get mold at the same row position every cycle.

The room reading and the mid-tower reading are not the same environment. A room averaging 70 percent at midday can sustain a mid-tower zone at 87 to 91 percent for 3 to 4 hours after sunset. The mold is recurring at the same row because that position sits inside the highest-risk humidity zone in the tower -- the overlap of elevated humidity and restricted canopy airflow -- and the ambient sensor has never captured that zone accurately. Pull the AHT20 log at 23:00, not at 14:00.

I covered the reservoir and the humidity is still high by week three.

Reservoir coverage worked for weeks one and two. By week three, the canopy is the dominant ambient moisture source. The reservoir lid reduced one source; the established crop canopy has replaced it with a larger one. The intervention is correctly timed for the wrong growth stage. Shift from reservoir management to canopy-level airflow intervention when the tower reaches full canopy -- typically day 10 to 14.

My roots look white and healthy but my plants are showing calcium deficiency symptoms despite correct EC.

Calcium is transported primarily via the transpiration stream. When root zone VPD is near zero -- a condition measured inside tower pot sites at 97 to 99 percent humidity -- transpiration at the root surface is suppressed, weakening the stream that pulls calcium from root to developing leaf tissue. The nutrient solution calcium is adequate. The delivery mechanism is compromised by near-zero root zone VPD. The fix is pot site airflow, not nutrient programme adjustment. (Plant Mood grow log data, 2024)

I increased ventilation and the mold improved for a few weeks but came back.

Ventilation timed to active growing hours addresses the daytime humidity gradient, which is not when mold pressure is highest. If the fan timer reduces overnight or the manual habit is to ventilate during the day, the post-sunset rebound window is unaddressed. The mid-tower zone is above 85 percent for 3 to 4 hours every night that this window is not covered. The pot site interior peaks 60 to 90 minutes after the ambient peak -- ventilation that ends at the ambient peak ends before the pot site has cleared.

I added a second fan and it made almost no difference to mold rate.

A second fan positioned similarly to the first doubles airflow in the same direction without necessarily creating cross-tower flow through the mid-canopy zone where the gradient is worst. Fan positioning matters more than fan quantity. A single fan creating horizontal cross-flow at mid-canopy height flattens the gradient more effectively than two fans moving air parallel to the tower face from the same side.

What the logs reveal that standard tower guides never mention

In sealed growing periods, the morning baseline ambient humidity drifts upward by 2 to 3 percentage points per day with a full uncovered reservoir. By day four, the room is starting each morning 8 to 10 percentage points above where it started on day one. A grower who measures ambient humidity at the start of a cycle and finds it acceptable may be measuring before the ratchet has loaded. Humidity management needs to account for cumulative loading across the cycle, not just point-in-time readings. Continuous logging with comparison across the first week is the only way to detect this before it becomes a mold problem.

The overnight ratchet effect is cumulative across days

A 40-litre reservoir with 800 square centimetres of exposed water surface contributes less ambient humidity than a 20-litre reservoir with 1,400 square centimetres of exposed surface. Evaporation is driven by vapour pressure gradient at the water surface, not by the volume of water below it. Growers who reduce reservoir volume to lower humidity without changing lid geometry and net pot collar coverage are changing the wrong variable. The same applies to growers who delay top-ups to keep levels lower -- the evaporative surface area is unchanged. (Plant Mood grow log data, 2024)

Reservoir surface area matters more than volume

Evaporative cooling from the open water surface dissipates submersible pump heat. Covering the reservoir removes that pathway. The covered reservoir averaged 0.8 degrees Celsius warmer than the uncovered reservoir during daytime in four-week comparison trials. Dissolved oxygen in the covered system averaged 6.1 mg/L versus 6.8 mg/L at midday during week two, trending downward through weeks three and four. In Bangkok where water temperature already limits dissolved oxygen, removing passive surface exchange without replacing it with an airstone is trading algae and water loss problems for a root zone oxygen problem. Every guide that recommends covering the reservoir fails to mention this. (Plant Mood grow log data, 2024)

Covering the reservoir without adding aeration creates a dissolved oxygen problem

When ambient humidity peaks after sunset, the pot site interior is still loading. The enclosed geometry delays the response. A pot site that reaches 96 percent at 23:00 does not return to 85 percent by midnight just because ambient conditions have improved. It holds the humid microclimate until sufficient air exchange has occurred through the limited opening area. Ventilation timed to the ambient peak is ending before the pot site has cleared. Extending overnight ventilation 90 minutes beyond the ambient peak is the starting adjustment -- then verify with pot site logging that the extension is actually clearing the interior before terminating.

The pot site peak lags the ambient peak by 60 to 90 minutes

Where Bangkok tower humidity management is heading

Multi-height humidity logging will replace single-point ambient monitoring as the standard practice for Bangkok tower growers within two years, as AHT20 costs fall and grow log sharing spreads data on the gradient pattern

The cost barrier to four-sensor tower monitoring is approximately 400 to 500 Thai Baht in hardware -- less than the value of one crop cycle. As ESP32-based sensor mesh setups become better documented and the gradient data is shared across growing communities, growers who build a multi-height baseline will have comparative data that single-point growers cannot replicate retroactively. The pattern is consistent and reproducible enough that it will be recognised as standard once enough growers measure it.

A published Bangkok tower grow guide or Thai-language growing community resource recommending multi-height sensor placement as standard practice, appearing before mid-2026.

Root zone VPD will become a distinct monitoring parameter in commercial Bangkok tower operations within three years, displacing ambient VPD as the primary root health environmental indicator

The gap between ambient VPD and root zone VPD inside a pot site is large enough -- ambient VPD of 1.06 kPa versus root zone VPD of 0.10 kPa at the same moment -- that any operation running root zone sensors will quickly find ambient calculations useless for predicting root zone conditions. Once the finding spreads through operations selling to buyers who audit environmental data, the measurement shift becomes a market requirement.

Published CEA certification or buyer audit frameworks for Southeast Asian hydroponic produce specifying root zone VPD or humidity targets distinct from ambient room targets, appearing between 2025 and 2027.

A named failure pattern for tropical tower growing -- characterised by elongated sparse root hair morphology, intermittent wilting without visible root rot, and calcium deficiency in adequate nutrient solution -- will be documented in growing communities within 18 months

The pattern is already present across multiple Plant Mood cycle logs and is consistent enough to be reproduced. It does not yet have a name because most growers are not measuring root zone humidity separately from ambient conditions. Once multi-height and pot site logging becomes more common, the pattern will be recognised across enough operations simultaneously to acquire a named mechanism description.

Documented discussion of this specific root morphology pattern in connection with elevated root zone humidity in tropical tower growing communities, forums, or published grow guides by mid-2026.

The intervention most tower growers reach for first is the one that matters least

Growers who encounter a humidity problem in a Bangkok tower operation almost universally reach first for a dehumidifier or a bigger fan. Both are understandable responses. Neither addresses the specific mechanism that is producing mold events in most cases. The mid-tower humidity peak occurs during a 3 to 4 hour window after sunset. A dehumidifier running all day reduces the ambient baseline but does not eliminate the evening spike at mid-canopy height. A bigger fan pointed in the wrong direction does less than a small fan pointed correctly. The intervention that consistently shows up in the data as moving the mid-tower number is a directional fan creating cross-flow at mid-canopy height, timed to the evening window, in a setup where the reservoir surface area has been reduced to lower the base evaporation load. That combination is achievable for under 500 Thai Baht before counting the cost of sensor placement.

That barrier is real for some growers and should not be dismissed. But a standalone digital hygrometer placed inside a pot site -- no wiring, no ESP32, no logging -- costs less than 200 Thai Baht and immediately shows the gap between pot site conditions and the ambient reading. That single observation changes the framing of every subsequent intervention decision. The logging setup comes later. The observation that the pot site is 20 points higher than the room can happen in the next hour with a cheap thermometer.

The strongest counterargument is that sensor placement and multi-height logging require technical confidence that many small-scale growers do not have. For a grower managing one tower in a spare room without electronics experience, recommending an AHT20 wired into a pot site is recommending a barrier they will not cross. A dehumidifier is easier to implement and produces a visible number change.

One thing to do before your next cycle starts

Before the next cycle: position a sensor -- any sensor -- at mid-tower height and set it to log or record continuously for the first five days. Check it at 23:00 on day three. If it is above 85 percent for more than two hours, you have a mold window that your daytime checks have never shown you. That one reading does not fix the problem, but it tells you whether the mid-tower gradient is worth managing before you commit another crop to the same conditions. If it is above threshold, add a fan creating cross-flow at that height and check again on the same schedule. The improvement will be visible in the numbers within 24 hours.