Running irrigation on a schedule - set it Monday, run it Thursday regardless of what the soil actually holds - is one of the more expensive habits in production agriculture. You're either watering ground that doesn't need it yet, or you've already stressed the crop waiting for the calendar to turn. Soil moisture probes have been around long enough that the technology is mature and the costs have dropped to the point where a working network across a mid-size field is within reach for most operations. The hard part was never the hardware. The hard part is placing sensors where the data actually means something, calibrating them to your specific soil, and then building a decision process that turns a number on a screen into a gate opening or staying closed. This guide covers all three of those problems in enough detail to get you started without the common mistakes that turn a probe network into an expensive piece of ignored equipment.
Before you spend money or sink anything in the ground, you need to understand what your sensors are actually measuring, because the answer shapes how you interpret the data.
The two categories you'll encounter most often in field-scale irrigation are tensiometers and capacitance-based sensors, which includes time-domain reflectometry, or TDR.
Tensiometers measure soil water tension - the suction force the soil matrix exerts on water. A ceramic tip is buried in contact with the soil, and as the soil dries it pulls water through the porous tip, creating a partial vacuum inside the tube. You read that vacuum in centibars or kilopascals. At zero, the soil is saturated. As the number climbs, the soil is drying out and the plant is working harder to extract moisture. The appeal of tensiometers is that they're measuring something directly related to plant stress - tension, not just water content. The limitation is they require maintenance. The reservoir needs to be topped up with deionized water. Ceramic tips can dry out and lose contact in cracking soils. In coarse sandy soils they lose sensitivity at low tension because water drains so fast the sensor never registers. They are excellent tools in fine-textured soils with good management attention, and they don't need much calibration because they're not inferring water content from another measurement.
Capacitance sensors and TDR probes take a different approach. They send an electrical signal into the soil and measure how that signal is affected by the dielectric properties of the surrounding material. Water has a dramatically higher dielectric constant than dry soil particles or air, so a wetter soil reads differently than a dry one. The output is typically expressed as volumetric water content - the percentage of the total soil volume that is occupied by water. So when you see a reading of 28%, that means roughly 28 cubic centimeters of water per 100 cubic centimeters of soil. These sensors have no moving parts, no reservoir to refill, and they lend themselves to automated logging and telemetry far more naturally than tensiometers. They also tend to hold up better in very dry soils. The tradeoff is that the raw electrical measurement needs to be converted to a water content value through a calibration equation, and that equation can be significantly wrong if you're using a factory default calibration in a soil that differs from whatever was used to generate it - more on that below.
For most farmers building a field moisture monitoring network with data loggers and remote access, capacitance sensors are the practical choice. They pair cleanly with inexpensive dataloggers, draw little power, and can be daisy-chained at multiple depths on a single probe rod. TDR is more accurate in a laboratory sense but equipment costs remain higher and the signal interpretation is more complex. Unless you have a specific reason to use TDR - close partnership with a university extension office running a calibration study, for example - capacitance probes at multiple depths are where most working farms land.
Placement is where most probe networks fail. A sensor gives you information about the soil immediately surrounding it - roughly a sphere of influence a few inches in diameter - which makes location selection critically important at the field scale.
The goal is to place sensors where they're representative, not where they're convenient. That sounds obvious, but the temptation to put a probe near the field edge, next to the access road, or at the point closest to the power source is strong. Edge effects - different soil compaction, different microclimate, wetter or drier conditions from road runoff or canopy gaps - make edge placements systematically misleading.
Start with a soil map. The NRCS Web Soil Survey is free and gives you the mapped soil series for your field. If your field contains two or three distinct soil map units of meaningful size, you need at least one sensor location in each. A silty clay loam and a sandy loam in the same field will behave completely differently under the same irrigation event. Averaging them into a single number or using one to represent the other produces bad decisions.
Within a soil unit, look for the representative spot - not the wettest area, not the driest corner, but the part of that unit that most closely matches the average conditions. Avoid:
If you have variable rate irrigation capability or manage irrigation by zone, your sensor placement should track the zones, not just the fields. Each independently managed zone should have its own representative location.
For a rectangular center-pivot field, a common and workable approach is three locations along the pivot radius: one near the pivot point, which covers a small area but often receives more passes, one in the middle of the travel arc, and one near the end of the arm where the application rate is highest and the soil may have different characteristics. Three locations won't catch every variation but they're a significant improvement over a single point, and they give you a pattern to watch - if the outer point is consistently drying faster, that's information you can act on.
For drip-irrigated perennial crops, place at least one sensor in the wetted bulb directly under or adjacent to an emitter, and one sensor between emitters to track how far moisture is traveling laterally. The relationship between those two readings tells you whether your emitter spacing is appropriate for the soil's lateral conductivity.
A single sensor at one depth gives you a snapshot of a thin slice of the root zone. For most crops in most soils, you want sensors at a minimum of two depths, and three is better.
The logic is straightforward. A shallow sensor - typically 6 to 12 inches - tells you what's happening in the zone where young roots are densest and where evaporation has the most influence. It responds quickly to rainfall and irrigation events, which makes it useful for tracking the wetting front after an application. A mid-depth sensor at 18 to 24 inches represents the bulk of the active root zone for many annual crops and gives you a picture of what the roots are actually drawing on between irrigation events. A deep sensor at 30 to 36 inches or below serves a different purpose: it tells you whether water is moving past the root zone entirely. If your deep sensor is consistently wet and climbing after every irrigation event, you are over-applying and leaching moisture - and potentially nutrients - beyond where roots can recover them.
Match your probe placement depth to your crop's rooting depth and the profile of your soil. In a shallow soil over fragipan or restrictive clay at 20 inches, there's no reason to put a sensor at 36 inches - it'll be reading conditions roots can't reach. In a deep sandy loam with a vigorous corn root system reaching 36 to 48 inches under irrigation, a three-depth profile makes real sense.
For tree fruits and vines in the Pacific Northwest, where I've seen a lot of probe installs, a useful depth arrangement is 12 inches, 24 inches, and 36 to 40 inches. The 12-inch reading tracks the feeder root zone and the immediate response to irrigation. The 24-inch reading is your primary trigger point - this is what you're managing to. The 36-inch reading is your leaching indicator. If that bottom sensor trends up over the season, you're pushing water through the profile.
When installing a multi-depth probe, take care with backfill. A poorly backfilled installation channel creates a fast-drainage path along the outside of the probe rod, which will give you misleading readings - the soil immediately around the sensor won't represent the surrounding horizon because water is channeling past it. Backfill in stages using the native soil removed from the hole, tamped lightly, and add a bentonite seal at the top few inches to prevent surface water from running straight down the rod.
Factory calibration curves shipped with capacitance sensors are generated in controlled media - often a uniform sandy loam or a mineral-standardized mix that may bear little resemblance to your field soil. The error introduced by applying a factory curve to a high-organic-matter silty clay loam or a gravelly alluvial soil can be substantial. You might see reported volumetric water content that is consistently several percentage points off from the actual condition. In some soils it's worse.
This matters for absolute readings. If you're using the probe to determine whether you're at 50% plant-available water or 35% plant-available water, a systematic error of several points changes your decision. If you're only using relative change - "it dropped 4 points since yesterday, let's irrigate" - factory calibration may be adequate. Know which mode you're operating in.
The most practical field calibration approach is a gravimetric sampling method. You don't need a lab, though a scale accurate to a gram and access to a drying oven helps. The process:
Do this at two or three different moisture conditions - once when the soil is near field capacity after a good rain or irrigation event, once when you're at moderate depletion, and ideally once when conditions are dry. Those paired points - sensor reading versus measured volumetric water content - let you generate a simple linear correction. For most sensors and most soils, a linear offset correction is sufficient.
Soil temperature affects capacitance readings. Most modern sensors include temperature compensation in firmware, but verify this with your manufacturer and check whether the compensation is applied to the reported value or only available as a separate output you have to apply manually. In soils that swing significantly in temperature between spring and late summer, uncompensated sensors can show drift that looks like moisture change but isn't.
Bulk density also matters. Compaction around the probe rod from installation can create a higher-density zone that reads systematically different from undisturbed soil. This is another reason to backfill carefully and, where possible, use hand auguring rather than driving the probe in percussively.
A probe that has to be walked to physically and read manually every few days is better than nothing, but it's close to the bottom of the value ladder. The real payoff from a soil moisture network comes from continuous logging and remote access. You want to see the soil drying curve between irrigation events without being in the field.
The telemetry landscape breaks roughly into three options: cellular, LoRaWAN, and short-range RF to a local hub.
Cellular-connected dataloggers are the most plug-and-play option. A logger with an LTE modem and a SIM card pushes data to a cloud platform on whatever interval you configure - commonly every 15 to 60 minutes. The downside is ongoing data plan cost and the need for cell coverage in the field. In much of the Pacific Northwest, coverage in valley floors is adequate, but hill ground, draws, and anything more than a few miles from a tower can be spotty. Test signal before you commit to a cellular architecture for a location.
LoRaWAN is a low-power, long-range radio protocol well suited to agriculture. Individual nodes - your probe stations - transmit at very low power to a gateway that may be half a mile to several miles away depending on terrain and obstructions. The gateway connects to the internet via cellular or ethernet. The advantage is that a single gateway can serve many sensor nodes across a large acreage, and the per-node cost is lower than individual cellular radios. The limitation is that LoRaWAN networks have to be built or subscribed to - if your region doesn't have LoRaWAN coverage from a service provider, you need to install your own gateway.
Short-range RF systems - 900 MHz spread-spectrum or similar - use the same hub-and-spoke concept as LoRaWAN but typically with proprietary protocols and a shorter range, often a few hundred yards to under a mile. These systems are common in irrigation-specific product lines and work fine on a single farm, but they don't scale across multiple properties as cleanly as LoRaWAN.
Power is straightforward. Most field probe stations run on a small solar panel - something in the 5 to 10 watt range - charging a sealed lead-acid or lithium battery. A station drawing modest power for sensor polling and periodic radio transmission can run through several days of cloud cover without issues if the battery is appropriately sized. Orient panels south in the Northern Hemisphere, clear any obstructions, and check connections annually. Corroded battery terminals are one of the more common causes of station dropout.
Data storage and display platforms vary widely. Some sensor manufacturers bundle their own cloud platforms. Others output data in standard formats that can be ingested by third-party ag data platforms or even into a basic spreadsheet via API. Evaluate platforms on their ability to set threshold alerts - you want an email or text when a reading drops below your refill point - and on data export. Avoid platforms where your data is locked in and you can't get it out in CSV or JSON. Your data is yours.
Once loggers are in the ground and transmitting, you're staring at lines on a chart. Making sense of them requires a few reference points tied to your specific soil and crop.
Field capacity is the moisture level a well-drained soil holds after gravity drainage has stopped - typically 24 to 48 hours after a saturating rain or full irrigation event. In volumetric terms, a silty loam might sit at 32 to 38% at field capacity. A sandy loam might be 18 to 24%. You determine field capacity empirically: run a full irrigation event to saturation, stop irrigating, and watch the readings. After 24 to 48 hours, when the rate of decline has slowed to nearly flat, that reading is your field capacity baseline for that sensor location and depth.
Permanent wilting point is the lower end - the moisture level at which plants can no longer extract water from the soil. You generally don't want to reach this in production, and for most crops you don't need to measure it directly. Use published values for your soil texture and treat them as a backstop, not a target.
Plant-available water is the range between field capacity and permanent wilting point. The refill point is where in that range you want to trigger irrigation. For most crops under normal production pressure, the refill point sits at 40 to 60% depletion of plant-available water. A stress-sensitive crop in active fruit development might trigger at 30% depletion. A drought-tolerant grain crop might tolerate 60 to 65% depletion without yield penalty. Know your crop, know your growth stage, and adjust the trigger point accordingly.
The depletion curve - the slope of the moisture line between irrigation events - tells you daily crop water use under current conditions. If a sensor at 18 inches drops 0.8 percentage points per day across a hot week in July, you can calculate roughly how many days until you hit the refill point, and you have a window to plan your next irrigation set. This is where soil moisture probe data becomes genuinely actionable: not just "are we wet or dry" but "we have approximately four days before we need to irrigate, which means we schedule the system for Wednesday."
Watch the shape of the depletion curve, not just the absolute value. A curve that suddenly steepens indicates higher-than-expected evapotranspiration, often corresponding to a heat spike or wind event. A curve that flattens unexpectedly during a period when ET should be high might indicate roots are experiencing stress and water uptake has actually slowed, which can be an early warning sign.
The data is only as valuable as the decisions it drives. Here is how to build a practical decision workflow that isn't dependent on remembering to check a dashboard at the right moment.
Set threshold alerts in your platform. Define your refill point for the primary management depth at each sensor location. When that reading crosses the threshold, you get a notification. Don't rely on checking the dashboard proactively - alerts catch the moments you'd otherwise miss during a busy week.
When an alert fires, the question isn't just whether to irrigate, but how much. Your target is to refill the root zone to field capacity without pushing water below the root zone. In plain terms, you take the gap between field capacity and the current reading, multiply it across the depth of the root zone you're refilling, and correct for the soil's bulk density to get inches of net water to apply.
In practical terms, if your 18-inch sensor is at 22% and field capacity is 32%, that's a 10-point deficit across 18 inches of root zone. For a typical mineral soil that works out to roughly 1.5 to 2 inches of net water application. Account for your system's distribution uniformity - if your drip or sprinkler system applies unevenly, the average application needs to be higher to bring the dry spots up without drowning the wet ones.
Watch the post-irrigation response at each depth. The shallow sensor should respond quickly - within an hour or two for a drip system, faster for overhead. The mid-depth sensor should show a rise over the course of the event and the hours following. The deep sensor should remain relatively stable. If the deep sensor rises sharply during or immediately after irrigation, you applied too much too fast for the soil's infiltration rate, or your total volume was excessive.
Adjust your refill point seasonally. Early in the season when ET is low and roots are shallow, a 50% depletion trigger is conservative enough. At peak ET in midsummer with full canopy and high temperatures, you may want to tighten that trigger to 35% to stay ahead of stress. Late in the season as the crop matures and you're managing for harvest timing, you might intentionally allow higher depletion to firm up soils or stress-harden the crop - in which case, move the threshold deliberately rather than ignoring the data.
When irrigation scheduling data from multiple probe locations diverge significantly - one location reading at 55% depletion while another reads at 25% - don't average them. Investigate. The divergence might reflect actual soil variability that justifies split irrigation management. Or it might mean one sensor has a calibration drift, has lost soil contact, or has a connectivity issue. You can't tell without looking at historical patterns and checking the sensor.
A probe network you can't trust is worse than no network, because it breeds false confidence. Here are the failure modes I see most often and how to catch them before they corrupt a season's decisions.
Soil gap formation is the most common problem with capacitance sensors. When a soil cracks in dry conditions, a gap opens around the probe rod and the sensor reads artificially dry - it's measuring air instead of soil. This is most severe in high-clay soils with significant shrink-swell. You can identify it by comparing the probe reading to your field observations and to ET-based estimates. If the sensor shows extreme dryness but the crop shows no stress and your ET model suggests adequate water, suspect a gap. The fix is to thoroughly wet the soil around the probe to allow shrink-swell soils to close back around the rod.
Electrical interference can affect readings, particularly in fields with variable-rate drive systems, center-pivot electrical infrastructure, or ground current from aging irrigation system wiring. If you see erratic spikes or drops in readings that don't correspond to weather or irrigation events, check for ground fault issues in nearby electrical systems before concluding the sensor is bad.
Fouling of the sensor element can occur in soils with high iron content, high salinity, or significant biological activity. Iron oxide deposits on the sensor surface alter the dielectric measurement. Some sensors are more susceptible than others. Check manufacturer guidance on cleaning procedures, and do a visual inspection annually when you'd otherwise pull sensors for the off-season.
Battery and solar panel issues are routine. Build a simple annual maintenance schedule: clean the panel surface, check connection integrity, test battery voltage under load, inspect for rodent damage to cables. Rodents chewing through sensor cables is a real and recurring problem on many farm installations - conduit on the above-ground portion of the cable run is cheap insurance.
Firmware and calibration drift are less common but worth tracking. If a sensor that previously tracked well against your gravimetric samples starts showing systematic bias, either the sensor element is degrading or firmware was updated and the factory calibration changed. Maintain a log of sensor readings versus periodic gravimetric samples - even once or twice a season - as a check on data integrity.
If you're starting from zero, here's a sequence that avoids the most common first-year mistakes.
Start with one or two locations, not ten. Pick your most problematic irrigation block - the one where you're least confident in your scheduling decisions, or where soil variability is highest - and instrument it first. A three-depth capacitance probe station with a cellular logger and a small solar panel is a reasonable first installation. Get that first station running, validate its data against your field observations for a full irrigation season, and calibrate it with at least two gravimetric samples. Then expand.
Before you buy hardware, confirm connectivity. If you're planning cellular telemetry, go to the candidate locations with your phone and check signal. Check both data signal and voice - LTE data requires stronger signal than voice calls. If signal is marginal, you'll want either a higher-gain antenna option or a different telemetry approach.
Choose a platform that supports threshold alerting and data export. Lock in on this before you commit to hardware, because many sensors are tied to specific platforms and switching later is expensive.
Document your reference points for each sensor location during the first season:
These reference points are what make the raw numbers actionable. Without them, you're watching lines move without context.
After the first season, do a retrospective. Did the sensor locations you chose actually capture the variability you needed to see? Were there irrigation events where the probe data and your field observations disagreed, and if so, why? Did any stations drop out, and for how long? Use those answers to refine placement and maintenance for the following year.
Expand coverage as budget and confidence allow. A fully instrumented field with sensors in each major soil management zone and reliable telemetry represents a meaningful improvement in irrigation efficiency over time. The payoff isn't always visible in a single season, but across several years of better-timed applications you'll see it in reduced pumping hours, reduced leaching of fertilizer below the root zone, and in the quiet confidence of knowing what your soil actually holds rather than guessing.
Soil moisture probes don't make irrigation decisions. You do. What they give you is the data needed to make those decisions with something closer to certainty rather than educated guesswork. The gap between a well-placed, well-calibrated capacitance sensor network and a schedule-driven irrigation approach is the gap between managing to what the soil actually holds and managing to what you assume it holds.
The technical barriers to building a functional field moisture monitoring network are lower than they've ever been. The sensors are reliable, the telemetry options are mature, and the cost per location has dropped to the point where it's comparable to other farm instrumentation you'd consider routine. The remaining challenge is the agronomic and interpretive work: placing sensors where they're representative, calibrating them to your soil, establishing the reference points that give readings meaning, and building a decision workflow that actually changes when and how much you irrigate.
Do that work carefully in the first season, and you'll end up with a network you trust. One you trust is one you'll use. And one you use will pay back the investment in fuel, water, and yield quality more reliably than almost any other field instrumentation you can put money into.
Trigger irrigation at your refill point, which for most crops under normal pressure sits at 40 to 60 percent depletion of plant-available water, the range between field capacity and permanent wilting point. A stress-sensitive crop in fruit development might trigger at 30 percent depletion, while a drought-tolerant grain crop can tolerate 60 to 65 percent without yield penalty. Adjust the point by crop and growth stage.
Use at least two depths, and three is better. A shallow sensor at 6 to 12 inches tracks young roots and the wetting front, a mid-depth sensor at 18 to 24 inches represents the active root zone you manage to, and a deep sensor at 30 to 36 inches acts as a leaching indicator. If it climbs after every event, you are over-applying.
Yes, if you rely on absolute readings. Factory curves are generated in controlled media and can be several percentage points off in a high-organic or gravelly soil. A gravimetric field calibration fixes it: record the sensor reading, pull cores, weigh them wet, dry at 105 degrees Celsius for 24 hours, and reweigh. Pair those points at two or three moisture conditions for a linear correction.
Tensiometers measure soil water tension, the suction the soil exerts on water, read in centibars, which relates directly to plant stress but requires refilling the reservoir and can lose contact in cracking soils. Capacitance sensors send an electrical signal and report volumetric water content, have no reservoir, and log automatically, but their raw reading needs a calibration equation. Most field networks use capacitance probes.
Join our list for practical guides on farm tech, precision agriculture, and tools that work.