Daily water consumption of a full hydroponic tower by crop and season: ultrasonic level sensor data

Key takeaways

A full hydroponic tower running lettuce in Bangkok dry season consumes between 3.9 and 4.8 litres of water per day at canopy maturity, based on ultrasonic level sensor logs across multiple cycles -- 60 to 90 percent above the 2 to 3 litre daily estimate that appears in most tower setup guides.

Basil consistently outdrinks lettuce on the same tower hardware. At peak growth in dry season, basil consumption reached 5.6 to 6.1 litres per day in our logged cycles -- the highest daily draw of any crop we have run on tower systems.

Wet season reduces daily water consumption by 25 to 40 percent across all crops we have measured, driven by lower vapour pressure deficit in Bangkok's high-humidity monsoon months reducing transpiration demand even when temperatures remain similar to dry season.

The crop growth stage matters more than the crop species in the first two weeks. A full tower of lettuce at day 7 uses less water per day than a half-full tower of mature basil. Consumption curves are not linear -- they accelerate sharply in the week before peak harvest maturity.

Reservoir sizing recommendations based on generic daily consumption estimates will leave a grower with a dry reservoir and a missed top-up in dry season peak. A tower system in Bangkok needs a minimum 20-litre reservoir to run 3 days between top-ups during dry season canopy maturity, not the 12 to 15 litres that entry-level tower kits typically include.

Ultrasonic level sensors resolve consumption tracking to within 1 to 2 millimetres of water level change, which translates to approximately 0.1 to 0.2 litres in a standard 60-litre cylindrical reservoir. Manual dip-stick checks or visual float estimates cannot track the daily consumption curve with enough resolution to catch an accelerating draw before the reservoir empties.

TL;DR

The daily water consumption number in your tower's setup guide is wrong -- not slightly wrong, but wrong by a factor that affects how often you need to top up, how large your reservoir needs to be, and whether you will return from two days away to a dead crop. Ultrasonic level sensor logs across our Bangkok tower cycles show peak dry season consumption running 60 to 90 percent above published estimates, with basil pushing even higher. Season matters as much as crop. And the draw accelerates in the final week before harvest in a way that nothing in the standard documentation warns you about.

The reservoir that should have lasted four days

October, transitioning out of wet season. A tower running 28 heads of lettuce, day 18, reservoir at 14 litres at Monday morning check-in. Based on the setup guide's stated consumption range of 2 to 3 litres per day, I had three to four days of buffer before the next top-up was needed. I was travelling Tuesday and Wednesday. I logged a note in the system to top up Thursday morning. The ultrasonic sensor triggered a low-level alert Wednesday at 11am. The reservoir was at 2.1 litres. The pump was pulling air.

The sensor log showed what happened. Daily consumption on day 16 was 3.2 litres. Day 17 was 3.8 litres. Day 18 was 4.4 litres. The lettuce had entered the final pre-harvest growth surge. In three days it had nearly doubled its daily water draw without any change in temperature, humidity, or light schedule. The setup guide's estimate was not wrong for day 7. It was wrong for day 18. The number that matters for reservoir planning is not the average -- it is the peak, and the peak arrives precisely when the crop looks its best and you are least likely to be watching it closely.

What the level sensor logs actually show

4.8 litres per day

Peak daily water consumption recorded for a full 28-head lettuce tower in Bangkok dry season at canopy maturity (day 18 to 22), based on ultrasonic level sensor data averaged across four dry-season cycles, Plant Mood grow log data, 2023 to 2024. The same tower in wet season peak ran 3.1 litres per day under comparable crop conditions.

6.1 litres per day

Peak daily water consumption for a full basil tower in Bangkok dry season at canopy maturity, the highest single-crop daily draw recorded in our tower logs. Basil's high transpiration rate and larger leaf surface area relative to lettuce produces a meaningfully different water demand profile on identical tower hardware. Plant Mood grow log data, 2024.

Day 14 to 22

The window in which daily water consumption accelerates most sharply in lettuce tower cycles. Before day 14, daily draw in dry season averaged 2.1 litres. After day 14, it climbed at an average rate of 0.35 litres per day. By day 20, consumption was more than double the early-cycle rate. This acceleration window is where most reservoir dry-out events occur in growers who planned top-up schedules based on early-cycle consumption.

25 to 40 percent

Reduction in daily water consumption between dry season peak and wet season peak for the same crop on the same tower hardware, across matched crop stages. The reduction is not driven by temperature difference alone -- Bangkok wet season temperatures are often within 2 to 3 degrees Celsius of dry season. The primary driver is vapour pressure deficit: at 85 to 90 percent relative humidity, the atmosphere's capacity to absorb transpired water vapour is reduced, and the plant's transpiration rate drops accordingly. Plant Mood grow log data, 2023 to 2024.

Temperature is not the main driver of seasonal water demand

It is partially a temperature story and mostly a vapour pressure deficit story. VPD -- the difference between the amount of water vapour the air can hold at a given temperature and the amount it actually holds -- is what drives transpiration rate in plants. A 33-degree Celsius day at 55 percent relative humidity (typical Bangkok dry season) produces a VPD of approximately 2.0 kPa. A 31-degree Celsius day at 85 percent relative humidity (typical Bangkok wet season) produces a VPD of approximately 0.85 kPa. The plant experiences a fundamentally different atmospheric demand in those two conditions, and its stomata respond accordingly. The wet season plant is not transpiring less because it is cooler. It is transpiring less because the air it is pushing water vapour into is already nearly saturated.

In our sensor logs, the correlation between ambient temperature and daily water consumption was weaker than expected -- an r-squared of approximately 0.41 across matched crop-stage observations. The correlation between calculated VPD and daily water consumption was stronger at approximately 0.68. On wet-season days where temperature spiked unusually close to dry-season levels but humidity remained high, water consumption stayed low -- consistent with VPD as the mechanism, not temperature alone. This matters practically: a Bangkok grower who installs a fan to lower temperature and thinks that will reduce reservoir draw is addressing a secondary variable. Reducing humidity -- or accepting that wet season genuinely reduces consumption -- is the more accurate mental model.

Bangkok dry season means higher temperatures, and higher temperatures mean more water consumption. The seasonal difference in how much a tower drinks is a temperature story.

Why the consumption curve accelerates before harvest

Water consumption in a tower is not a function of the number of plants alone -- it is a function of total transpiring leaf surface area. In a tower system, canopy development follows a roughly exponential curve across the crop cycle. In the first week, the plants are small, the leaf area is limited, and water draw per plant is low. Between days 14 and 22 in a typical lettuce cycle, the outer leaves have filled to full width and the plant is simultaneously maintaining existing tissue, expanding new leaves, and pushing into the pre-harvest growth phase. Total leaf area can double in this window. Daily water consumption follows that doubling.

The level sensor resolves this in the logs in a way that visual inspection never would. Day 7 of a typical dry-season lettuce cycle: the sensor registers a drop of approximately 18 millimetres across the day in a 60-litre cylindrical reservoir -- about 2.0 litres. Day 20 of the same cycle: the drop is 42 millimetres across the same reservoir -- about 4.7 litres. The level drop rate is visible in the slope of the logged curve. Once you have seen that curve across several cycles, you can read where you are in the consumption acceleration just from the sensor data without visiting the tower. A slope change -- the daily drop rate increasing by more than 15 percent versus the previous day -- is a reliable signal that the crop has entered the peak consumption phase. That is when the top-up schedule needs compressing.

Night-time consumption is not zero, which is the other thing the level sensor makes visible that manual checking cannot. In a mature lettuce tower in dry season, the sensor showed a drop of 8 to 12 millimetres overnight between lights-off and the following morning -- roughly 0.9 to 1.3 litres. Root zone respiration continues after dark and the plant continues to lose some water through stomata that are not fully closed. A grower who assumes that overnight top-up buffer is unlimited because the plant is resting is working with an incorrect model. The overnight draw in peak dry season is not negligible relative to the total reservoir volume in a small system.

Lettuce versus basil versus wheatgrass on the same tower hardware

Lettuce (28 heads, full tower)

Moderate canopy density, relatively shallow root zone in tower pockets, peak consumption concentrated in the final third of the crop cycle. Daily draw at peak: 3.9 to 4.8 litres in dry season, 2.6 to 3.1 litres in wet season. Consumption curve is the most predictable of the crops we have logged -- the acceleration window is consistent across cycles and the peak is relatively stable crop to crop.

Lettuce is the most forgiving tower crop for reservoir planning precisely because the consumption curve is predictable once you have run two or three cycles. After the third cycle, you can schedule top-ups proactively based on crop day rather than reservoir level. In dry season, move from every-four-days to every-two-days top-up after day 14 and you will not dry out.

Basil (24 heads, full tower)

Higher transpiration rate per plant than lettuce, driven by larger effective leaf area and higher stomatal density. Peak consumption reached 5.6 to 6.1 litres per day in dry season at canopy maturity. The acceleration curve starts earlier -- around day 10 -- and is steeper. Basil also does not have a clean harvest cutoff the way a lettuce head does, so consumption stays at peak levels for longer if the plants are not cut back.

A standard 12 to 15 litre reservoir on a basil tower in Bangkok dry season will run dry in approximately two to two and a half days at canopy maturity. This is not a worst-case scenario -- it is a reliable prediction based on our peak consumption logs. Basil towers in dry season need a minimum 20-litre reservoir and ideally 25 litres to allow a three-day buffer without monitoring. If you are running basil and planning to be away, do the reservoir capacity calculation before you leave.

Wheatgrass (dense seeding, full tower)

Wheatgrass on a tower is a different water profile than lettuce or basil. The crop is harvested at day 7 to 10 before significant canopy development, which means the peak consumption stage that drives the high daily draws in lettuce and basil never fully arrives. Daily consumption in our wheatgrass tower cycles peaked at 2.2 to 2.8 litres per day in dry season -- below the lettuce and basil peaks -- but the cycle is short enough that reservoir management is simpler. The risk with wheatgrass is not running out of water mid-cycle but running the system correctly across the fast turnover of cycle after cycle.

Wheatgrass is the most reservoir-efficient crop on a tower system on a per-cycle basis. If water availability or top-up frequency is a constraint in your setup, wheatgrass is worth considering on that basis alone, independent of the yield and market considerations.

What the sensor logs showed that guesswork could not

Across our logged tower cycles, a sudden drop in overnight water consumption -- the sensor level change between lights-off and lights-on the following morning falling more than 30 percent below the previous three nights -- appeared as the first quantifiable signal of root zone stress in two separate cycles before any visible above-canopy symptom emerged. In both cases, the root cause was identified as early-stage root rot limiting water uptake. The sensor flagged the change 36 to 48 hours before leaf wilting or discolouration appeared. No visual check would have caught it at that stage.

The overnight consumption rate is a reliable early warning signal for crop health problems

Daily consumption rate and specifically overnight consumption rate are leading indicators of plant health in a tower system, not lagging ones. Logging level sensor data at hourly resolution gives you a health signal the crop itself will not show you for another two days. A grower without level sensor logging is seeing the health story only after the chapter has already been written.

In dry season cycles where the reservoir temperature climbed above 29 degrees Celsius -- which happened in two cycles where the reservoir was positioned near a heat-emitting pump and light fixture -- daily consumption was 8 to 12 percent higher than in cycles where reservoir temperature was held below 27 degrees Celsius by repositioning and insulating the reservoir. The plants were not in better or worse condition. The water was simply evaporating faster from a warmer reservoir surface and through the tower column, adding a non-transpiration component to the sensor-logged daily draw.

Reservoir temperature affected daily consumption rate independently of ambient conditions

Not all of the daily level drop is transpiration. In a warm Bangkok grow space with an exposed reservoir, direct evaporation from the water surface and from wet growing media in the tower pockets contributes to the total daily draw. This component does not benefit the plant. Insulating the reservoir and minimising exposed water surface area reduces daily consumption without any effect on plant growth -- pure efficiency gain.

The months of November and December -- Bangkok's transition from wet to dry season -- showed the highest cycle-to-cycle variability in daily consumption rates of any period we logged. A cycle seeded in late October and harvested in mid-November could see peak consumption 30 to 35 percent lower than a cycle seeded two weeks later and harvested in early December, despite similar crop stages and identical hardware. The atmosphere was drying out underneath the crop cycle, pulling VPD higher across the harvest window.

The wet-to-dry season transition produced the fastest consumption ramp in our logs

Transition-season cycles require more active reservoir monitoring than settled-season cycles because the consumption rate is changing with the weather, not just with crop growth stage. A top-up schedule calibrated to wet season performance will be wrong by late November. The sensor logs make the transition visible. A manual-check routine operating on a fixed schedule will miss it.

Plant Mood grow log data, 2023 to 2024

What growers say when the reservoir keeps running dry

I set up a top-up schedule based on the manual's daily consumption estimate and my reservoir keeps running dry ahead of schedule. I don't know what I am doing wrong.

Nothing is wrong with the operation. The manual's estimate is based on average or early-cycle consumption, not peak consumption at canopy maturity. A schedule calculated on 2.5 litres per day will fail during a dry-season canopy peak running at 4.8 litres per day. The fix is not a better schedule -- it is a schedule that changes across the crop cycle, tightening the top-up interval after day 14 and expanding it again in the first week of a new crop.

My tower uses way more water in February than it did in October. I haven't changed anything. Is something leaking or wrong with the pump?

Seasonal VPD change. The October reading was in or near wet season -- high humidity, low VPD, low transpiration rate. February is peak dry season -- lower humidity, high VPD, high transpiration rate. A consumption difference of 30 to 40 percent between these two periods on the same hardware with the same crop is expected. Nothing is leaking. The atmosphere changed its demand on the plant.

I added a larger reservoir to my tower but I'm still running out of water faster than expected during peak growing periods.

Reservoir size solves the buffer problem but does not change the consumption rate. If peak daily draw is 4.8 litres and the new reservoir holds 20 litres rather than 12, the buffer has extended from 2.5 days to 4.2 days at peak consumption -- still less than a full week. Growers who expect a larger reservoir to eliminate top-up requirements rather than extend the interval between them will still be surprised by the frequency. The number to manage is the consumption rate at peak, not the reservoir volume alone.

My basil tower seems to be consuming water much faster than my lettuce tower running on the same system. Is this normal?

Yes. Basil's transpiration rate is meaningfully higher than lettuce on equivalent hardware. In our logged cycles, basil at peak dry season canopy drew 5.6 to 6.1 litres per day versus lettuce at 3.9 to 4.8 litres per day. If both towers share the same reservoir, that reservoir is being depleted at a combined rate that neither crop's individual estimate would predict. Separate reservoirs per crop, or separate consumption budgets when shared reservoir capacity is being calculated, is the correct approach.

Managing tower water consumption without being surprised

Any grower setting up a tower system in Bangkok

Size your reservoir to the peak daily consumption of your intended crop in dry season, not the average. For lettuce, design for 5 litres per day. For basil, design for 6.5 litres per day. Multiply by the number of days between available top-up windows and add 20 percent buffer. A three-day buffer for dry-season basil requires a minimum 23-litre effective reservoir capacity.

A reservoir that does not run dry between planned top-ups even at peak consumption. The cost of oversizing a reservoir is carrying slightly more water than you need in wet season. The cost of undersizing is a dead crop at peak growth.

Growers managing top-up schedules manually without level sensors

Operate two top-up intervals: a relaxed interval for days 1 to 13 of the crop cycle (every 3 to 4 days in dry season for lettuce) and a compressed interval for days 14 through harvest (every 1 to 2 days in dry season). The shift between intervals is time-based, not observation-based. By the time consumption acceleration is visible in the reservoir level, you are already behind.

Elimination of the late-cycle reservoir dry-out event that accounts for the majority of lost crops in tower systems running without sensor monitoring.

Growers with a level sensor already installed

Log daily level drop at the same time each morning before the first top-up. When the daily drop rate increases by more than 15 percent over the previous day's rate, flag it as the beginning of the peak consumption window and compress the top-up schedule immediately. Use the same overnight drop rate as a health indicator: a sudden overnight drop of 30 percent or more below the three-night rolling average warrants a root zone inspection regardless of how the canopy looks.

A two-day early warning on crop health problems and a consumption curve that drives top-up decisions rather than a fixed calendar that ignores what the reservoir is actually doing.

Growers running towers through the November to December seasonal transition

Do not carry a wet-season top-up schedule past the first week of November. Check daily consumption against your wet-season baseline at the start of each week from late October onward. When the daily draw has risen 20 percent above your wet-season average, switch to the dry-season interval. The transition is gradual but it is not invisible in the data if you are looking at it.

A top-up schedule that adjusts to seasonal VPD change rather than lagging behind it by two to three weeks -- which is the typical gap between the actual consumption shift and the grower's response when working without logged level data.

Growers running multiple crop types on separate towers sharing a central reservoir

Calculate peak daily consumption for each crop independently, sum them, and treat that sum as the planning number for reservoir sizing and top-up frequency. Do not average across crops or use a single crop estimate. A shared reservoir serving one lettuce tower and one basil tower in dry season at peak canopy is drawing approximately 10 to 11 litres per day -- a number neither crop's individual estimate would suggest on its own.

Accurate reservoir capacity planning for multi-crop setups, eliminating the systematic underestimation that occurs when a single generic consumption estimate is applied to a multi-crop system.

Logging the daily reservoir level at a consistent time before top-up. Most growers check the reservoir when they are nearby, top up when it looks low, and have no data on how much was consumed over what period. Without a timestamped consumption log -- even a manual one written on a whiteboard next to the tower -- the pattern of acceleration before harvest is invisible. It takes four minutes to dip a ruler into a reservoir and write the number down. The growers who do it have a fundamentally different understanding of their system than those who do not.