Gains:
- Ability to connect the concepts of plant water consumption (ET), soil moisture sensor and water balance to irrigation decisions
- Ability to produce draft irrigation program and calculate water budget from reference ET and crop coefficient with AI
- Ability to prevent water and energy waste by verifying AI irrigation recommendations with sensor data, soil type and system flow rate
Water is the scarcest and most expensive input in agriculture; In many regions, water is the only factor determining productivity. But irrigation is a double-edged sword: too little water stresses the plant and reduces yield, while too much water smothers the roots, washes out nutrients, creates salinization and soil barrenness, and wastes energy and money. Correct watering means knowing how much water the plant really needs and giving it exactly that. This unit teaches the scientific basis for calculating this need—water balance and evapotranspiration—shows you how to draft an irrigation schedule with AI, and tells you how to verify it with sensors and field reality.
Water Balance and Evapotranspiration
Think of a field as a water account. Introduction: rainfall and irrigation. Output: evaporation (from the soil) and transpiration (from the plant). The sum of these two outputs is called evapotranspiration (English Evapotranspiration, ET); that is, the total water loss from the plant-soil system to the atmosphere. The irrigation decision is simply this: how much water must enter to meet outflow (ET).
ET is difficult to measure directly, so it is estimated in two steps. First, the reference ET (ET0) is calculated: the water that a standard grass surface will lose under that day's weather conditions (temperature, humidity, wind, sun). This is purely meteorological. Then this value is multiplied by the crop coefficient (Kc), which reflects the water appetite of that product at that developmental stage:
ETc = ET0 × Kc
ETc is the actual daily water requirement of the product (mm/day). Kc varies from planting to harvest: small (0.3-0.5) in the seedling stage, highest (1.0-1.2) in full development, decreasing again at maturity. Irrigation amount is calculated by subtracting effective rainfall from ETc and taking into account system efficiency.
Tip: Most weather services and agricultural portals publish ET0 data free of charge. You can have the AI say "calculate ETc with this ET0 and Kc", but verify the Kc value from the local source by product and phase; false Kc shifts the entire water budget.
Soil Moisture and Sensors
ET calculation is the theoretical side of the water budget; The soil moisture sensor measures the truth. These sensors continuously read the soil moisture at specific depths and serve to monitor two critical thresholds: field capacity (the most water the soil can hold) and wilting point (the lower limit at which the plant can no longer take up water). Ideal irrigation keeps moisture in the "absorbable water" band between these two. The sensor data and the ET calculation confirm each other: the ET says "how much must be gone", the sensor says "what is actually left". When the two conflict, field control is essential (sensor calibration may be off).
Irrigation System and Yield
The same water need is met with different efficiency by different systems. Drip irrigation delivers water directly to the root zone, its efficiency is high (85-95%); sprinkler medium (70-85%); Flood irrigation is low (40-60%) and prone to salinization. The duration cannot be calculated without knowing the system flow rate (water delivered per hour) and efficiency. When calculating time with AI, be sure to provide system flow and efficiency as input; otherwise the output is meaningless.
Three Mini Cases: By the Numbers
Case 1 - Saving with water budget. A tomato producer was watering on a calendar (fixed time every 2 days). When switching to the ET-based program, ETc was found to decrease on cool and humid days; The number of irrigations decreased by 22% throughout the season, but the yield did not change. The gain was both water and pump energy.
Case 2 - Kc error. For a corn field, AI used Kc = 1.2 throughout the season and recommended 3 times more water during the seedling period. The engineer knew that Kc should be 0.4 at the seedling stage; Too much water would create root suffocation and nitrogen leaching. When I corrected Kc according to the stage, the program became realistic.
Case 3 - Sensor-ET conflict. While the ET calculation said "the soil must have dried out" the sensor was still showing high humidity. We went to the site: the impermeable layer in the lower layer was retaining water, there was a drainage problem. According to ET, if it were watered, the water would pool and root rot would occur; sensor and field reality prevented the error.
Weak Prompt / Strong Prompt
Weak prompt:
How often should I water this field?
Powerful prompt:
Your role: Irrigation specialist agricultural engineer. Calculate an irrigation schedule DRAFT.Inputs:- Product: [crop], developmental stage: [stage], Kc (from local source): [value]- Last 7 days ET0 (mm/day): [array], effective rainfall (mm): [array]- Soil type: [sand/loam/clay], system: [drip, efficiency 85%], flow rate: [L/hour]Task:- Daily ETc = ET0 × Calculate Kc, with table show.- Find 7-day net water requirement (ETc - precipitation) and gross water with system efficiency.- Subtract irrigation duration based on flow rate.- Note that it must be VERIFIED with soil moisture sensor data.- Show each step of the calculation; Mark where you make an assumption.
Powerful prompt clarifies all inputs, formula, yield correction and sensor verification requirement; It prevents AI from making hidden assumptions.
Four Copiable Templates
1) ETc account:
Calculate and add daily ETc with the following ET0 sequence and Kc value. Show each step. Write that I need to confirm whether Kc is correct for the product/phase. ET0: [...], Kc: [...]
2) Net and gross water need:
Subtract the effective precipitation (net requirement) from the 7-day ETc total. Find the gross water by dividing the system efficiency% by [..]. Give the result in mm and m3/da; Show unit conversion.
3) Sensor-ET comparison:
According to the ET calculation, how much is soil moisture expected to decrease? Compare this to the sensor reading I gave you. If the difference is large, list possible causes (drainage, calibration, leakage, precipitation). Decision making; checklist appears.
4) Drought/limited water scenario:
If water is limited (available: [m3]), indicate which developmental stage of this crop is most sensitive to water and propose an outline that prioritizes water at that critical stage. Include the yield risk.
System
Typical yield
suitable situation
Risk
drop
85-95%
Row crop, limited water
blockage, investment
sprinkler
70-85%
grain, fodder plant
Wind loss, disease moisture
keel
40-60%
Plenty of water, flat land
salting, washing
Fertigation
85-95%
Fertilizer + water together
Dose/calibration precision
Common mistakes
- Watering by calendar. Fixed day range ignores weather; It provides more water on cool days and less water on hot days.
- Using constant Kc. Kc varies with developmental stage; The only value is making a big mistake in seedling or maturation.
- Bypassing system efficiency. If you do not convert the net requirement to gross, less water will actually reach the root zone.
- Blind trust in ET. Drainage, leakage and calibration distort the ET calculation; Cross-checking with sensor and field is a must.
- Thinking that excess water is safe. Excessive water causes root suffocation, salinization, nutrient leaching and energy waste.
Attention: The irrigation decision is not just about yield, it is about the long-term health of the soil and the sustainability of the water supply. Overdraft reduces groundwater; Excessive watering makes the soil barren. AI calculates the water budget, but sustainability judgment is up to the engineer.
In summary
The science of irrigation is water balance: how much water must go in to meet outflow (evapotranspiration). The formula ETc = ET0 × Kc gives the actual need of the product; System efficiency determines gross water, sensor data determines reality. AI quickly drafts this calculation, but it is necessary to verify Kc locally, enter the system efficiency and cross-check the result with the sensor/field. Both too little and too much water is harmful; Correct irrigation is exactly the amount of water needed and the final decision is made by the engineer's sustainability judgment.
Application task
Choose a product and development stage; Find a Kc value with 7 days of ET0 and precipitation data (actual or representative). Create a weekly water budget for AI with the "ETc calculation" and "net and gross water requirement" templates in this unit. Then repeat the same calculation, deliberately giving Kc wrong (out of phase), and compare the two results: by how much did the wrong Kc shift the water requirement? Write how you would set up the verification step with the sensor.
checklist
- [ ] I calculated ETc by ET0 × Kc and verified Kc by product/phase.
- [ ] I moved from net need to gross water with system efficiency; I checked the units.
- [ ] I calculated the irrigation time according to the system flow rate.
- [ ] I cross-checked the ET calculation with the soil moisture sensor and the field.
- [ ] I evaluated the risk of excess water and sustainability and left the final decision to the engineer.