The buried IoT is finally getting real

For decades, farmers have wanted continuous, high-resolution data from the place that matters most: the root zone. Soil moisture, temperature, salinity, nitrate, pH, and redox all change hour by hour a few centimeters below ground, and those changes drive yield, water use, and fertilizer performance. Yet most “smart farming” devices have remained above ground, relying on weather stations, drones, or sparse probes that need frequent servicing. A new wave of underground sensing platforms—some batteryless, some energy-harvesting, and many using unconventional communications—aims to make the buried Internet of Things practical at scale.

Why the underground is hard for electronics

Moving sensors into soil introduces a spectacularly unfriendly environment for radios, power systems, and materials.

• Radio doesn’t like dirt: Water and dissolved ions in wet soil attenuate radiofrequency waves. Even sub‑GHz signals commonly used for long-range farm telemetry lose tens of decibels after a few centimeters, making direct-to-gateway links unreliable for deeply buried nodes.
• Batteries are tough to service: Replacing or recharging hundreds of buried batteries is not realistic during a growing season. Thermal swings, moisture ingress, and pressure cycles accelerate battery degradation.
• Electrochemistry doesn’t sit still: Many soil analytes (nitrate, pH, EC) are measured with ion-selective or ISFET sensors that drift over time and are sensitive to fouling. Maintaining calibration underground is nontrivial.
• Materials must survive—and not contaminate: Encapsulants have to resist water ingress and corrosion for years while avoiding leaching hazardous compounds into soil.

New physics for getting data out of soil

Because conventional over-the-air radio struggles underground, developers are leaning on near-field and acoustic methods that couple energy and data through soil more effectively at short range.

• Magnetic induction (MI): Coils create a quasi-static magnetic field that couples to a nearby receiver coil. At tens to hundreds of kilohertz, magnetic fields pass through soil with far less loss than RF, enabling meter-scale links. A buried sensor can “talk” to a shallow relay or surface hub with modest power, and that hub can forward data via LoRaWAN or cellular.
• Backscatter with underground readers: Batteryless tags modulate the impedance of a coil or antenna to reflect an incident field emitted by a nearby reader. This RFID-style approach removes the need for a power-hungry transmitter, stretching lifetime dramatically.
• Acoustic links: Piezoelectric transducers can send and detect sound through soil. The speed of sound in soil depends on compaction and moisture, but acoustic energy often penetrates better than RF in wet conditions. Short message bursts to a surface stake a meter away can be robust even when the soil is saturated.
• Hybrid meshes: Some designs combine MI or acoustic hops from sensor to surface with standard radios on the surface, forming a “two-layer” network that preserves range without burying high-power transmitters.

A practical pattern is emerging: keep the clever physics underground and the high-power communications above ground. That architecture lowers energy needs in the soil and makes it easier to maintain gateways and firmware.

Where the power comes from

The second breakthrough underpinning buried IoT is energy strategy. Instead of depending on large primary cells, newer systems stretch tiny energy budgets with a mix of harvesting and clever duty cycling.

• Energy harvesting:
  • Microbial fuel cells: Bioelectrochemical cells exploit naturally occurring soil microbes to generate microwatts to low milliwatts. Output depends on soil moisture, temperature, and organic matter, but can sustain low-duty sensing.
  • Thermal gradients: Thermoelectric generators can harvest small temperature differences between deeper soil and the surface, especially in diurnal cycles, to trickle-charge a capacitor.
  • Inductive recharge: An MI reader can provide not just data links but also energy bursts, replenishing a supercapacitor during each interrogation.
• Ultra-low-power electronics: Modern microcontrollers can sleep at sub‑microamp currents and wake for milliseconds to sample and transmit. With event-driven scheduling (for example, sampling more frequently during irrigation), average power draw can fall below 10–20 microwatts.
• Buffered operation: Supercapacitors or thin-film solid-state batteries buffer intermittent harvesting so sensors ride through dry spells or cold snaps.

In practice, a batteryless node might capture a moisture, EC, and temperature snapshot every 15 minutes during irrigation, every few hours otherwise, and store results until a daily MI or acoustic wake-up from a surface reader triggers data upload.

Materials, packaging, and longevity

Keeping electronics alive underground is a materials problem as much as an electrical one. Moisture will eventually work through most seals; ions accelerate corrosion; freeze–thaw cycles stress bonds.

• Encapsulation stacks: Multiple thin barriers—such as parylene conformal coatings under epoxy or polyurethane—can outperform a single thick layer by blocking different permeation paths. Barrier films that incorporate inorganic layers (e.g., ALD-deposited oxides) improve water vapor resistance.
• Selective permeability: Some sensors need direct contact with soil solution. Porous ceramic frits or ion-permeable membranes allow analytes to reach electrodes while keeping bulk water and particles out.
• Biodegradable housings: For short-lived deployments (one season), housings made from PLA or starch-based bioplastics reduce retrieval needs. The electronics core can either be recovered or designed as “transient electronics” that dissolve over time using materials like magnesium and silk fibroin—an active research frontier, with clear environmental tests still needed before mainstream use.
• Mechanical design: Low-profile, flexible forms reduce plow strikes and soil disturbance. Placement below the tillage layer (e.g., >30 cm) or alignment with no-till practices minimizes damage risk.

What these sensors actually measure well

Not all underground measurements are equally mature. The strongest early wins are parameters that are physically straightforward and require minimal calibration.

• Soil moisture (volumetric water content): Capacitance and time-domain methods are proven. Local calibration improves accuracy across soil textures.
• Temperature: Simple and stable, and a valuable input to irrigation and disease models.
• Bulk electrical conductivity (EC): Useful for salinity trends and to flag leaching events, though it conflates several effects (salts, moisture, temperature) without additional modeling.

More advanced but trickier are chemical measurements:

• Nitrate (NO3‑): Ion-selective electrodes can detect nitrate but drift and fouling are real. Systems often pair them with reference electrodes and periodic in-situ checks, then apply algorithms to correct drift.
• pH and redox: Useful for root-zone health and nutrient availability. Solid-state ISFETs reduce maintenance versus glass electrodes but still need calibration strategy.
• Dissolved oxygen: Relevant in flooded conditions (e.g., rice) and for denitrification monitoring; optical sensors consume more power than electrochemical ones.

In many deployments, arrays of robust physical sensors feed models that estimate harder-to-measure variables, trading direct chemistry for reliability and scale.

From point data to field decisions

Deployments that deliver value tend to share a few patterns:

• Stratified depth sensing: Stacking sensors at multiple depths (for example, 10, 30, and 60 cm) reveals how water and nutrients move through the profile, informing irrigation cutoffs and fertigation timing.
• Coupling with surface and satellite data: Surface micrometeorology and remote sensing (NDVI, thermal, and radar for soil moisture) contextualize sparse underground points, enabling spatial models and zone-level decisions.
• Event logging: Linking sensor time series with irrigation controller logs, fertigation events, and rainfall helps separate management effects from weather—critical for learning and automation.
• Alerts with hysteresis: Rather than chasing every wiggle, systems use thresholds with hysteresis and rate-of-change rules to recommend irrigation starts and stops that avoid oscillation.

The aim is not more data for its own sake, but fewer, better decisions: skipping an unnecessary irrigation set, splitting a nitrogen application to reduce leaching risk, or identifying a clogged drip zone early.

Connectivity and network design on the farm

An underground sensor network typically has three layers:

1. Buried layer: Batteryless or low-power nodes communicate via MI, acoustic, or short-range RF to a nearby surface relay. They spend almost all their time asleep.
2. Surface relays: Stakes or small boxes near or above the soil line aggregate data from several buried nodes. They have bigger antennas or solar panels and forward data via LoRaWAN, private LTE, NB-IoT, or Wi‑Fi.
3. Backhaul and cloud/edge: Gateways push data to farm servers or cloud platforms. Edge software can run irrigation logic locally when links drop.

Standards matter. Open sensor data models (like OGC SensorThings or simple JSON schemas) and interoperability with common farm platforms avoid lock-in. For security, relays authenticate buried nodes, and data are encrypted in the hop to the gateway and onward to the cloud.

Economics: where the ROI comes from

Growers judge these systems on payback, not novelty. The business case usually rests on a few drivers:

• Water savings and yield protection: Avoiding over- or under‑watering reduces stress, fuel, and electricity, while stabilizing yields. In pressurized irrigation, even single-digit percentage savings can cover subscription costs.
• Fertilizer efficiency: Timing nitrogen with soil moisture and temperature reduces leaching and nitrous oxide risk—often a compliance and sustainability win alongside agronomic benefits.
• Labor and monitoring: Early warnings of irrigation failures or salinity build-up save scouting time and reduce crop loss.
• Compliance and carbon reporting: Measured soil data can support water-use reporting and ESG claims, though programs differ by region.

Cost per monitored hectare depends on crop value and management zone size. High-value specialty crops justify denser deployments (dozens of nodes per hectare). Broadacre grains may rely on fewer multi-depth stations per management zone.

Environmental considerations

Embedding electronics in soil raises legitimate end-of-life questions. Responsible designs address:

• Retrievability: Durable cores that last several seasons should be recoverable with GPS and installation maps. Low-profile tethers or detectable markers can help localization.
• Benign materials: Avoiding heavy metals and choosing encapsulants that do not leach harmful additives reduces risk if devices are lost.
• Biodegradation with evidence: Biodegradable housings should be validated for real soil conditions and residues. Claims should match decomposition timelines in the target climate and soil type.

Lessons from the first generation

Early entrants proved both the promise and the pitfalls. Systems that overreached with complex chemistry without a maintenance pathway struggled with drift and cost. Others leaned on conventional radios and batteries, only to face short lifetimes or frequent servicing.

The more durable successes share traits: conservative sensing portfolios at first (moisture, temperature, EC), robust underground links to a serviceable surface relay, strong power budgets, and straightforward agronomic insights rather than dashboards full of raw numbers.

What to watch next

• Deeper MI and acoustic relays: Improved coil geometries and adaptive modulation promise longer underground hops, enabling sensors below tillage that can stay in place for years.
• Batteryless nitrate at scale: Progress in solid-state ion sensors and self-calibrating algorithms could bring reliable multi-season nitrate trends without field servicing.
• Integration with variable-rate systems: Closed-loop irrigation and fertigation that use buried data to drive zone-level prescriptions automatically.
• Standardized underground test beds: Shared evaluation sites and protocols for soil sensor performance across textures, salinities, and climates will make vendor claims comparable.
• Transient electronics, responsibly: If dissolvable devices move from lab to field, expect careful scrutiny of residues, decomposition rates, and agronomic impacts before certification.

Practical starting points for growers

For farms considering underground sensing, a phased approach reduces risk:

• Begin with a few multi-depth stations in representative management zones to validate irrigation thresholds and uncover variability.
• Favor systems that separate buried sensors from serviceable surface relays to minimize underground maintenance.
• Integrate with existing irrigation controllers and recordkeeping from day one; value comes from actions, not dashboards.
• Audit power budgets and estimated lifetimes in your soil and climate; ask vendors for data from sites with similar texture and salinity.

The buried IoT will not eliminate agronomy. It will, however, put the root zone on the record—turning guesswork into measurable cause and effect. As communications, energy harvesting, and materials converge, underground sensors are poised to become as routine as weather stations, quietly shaping decisions beneath our feet.