What is a climate battery?
A climate battery — also called an earth battery or ground-to-air heat exchange (GAHE) system — uses the stable thermal mass of the soil beneath a greenhouse as a buffer for heat. While outdoor temperatures swing widely between day and night, the soil a few feet underground stays at a relatively constant temperature year-round.
The system exploits this directly. Instead of letting warm, humid greenhouse air escape through vents during the day, a fan pulls it down through a network of buried perforated pipes. The soil absorbs that heat and moisture. At night, when the greenhouse cools, the stored warmth is drawn back up and released into the growing space.
The result is a greenhouse that stays warmer at night and cooler during peak daytime heat — extending growing seasons without ongoing fuel costs or grid dependency.
Most greenhouse heating solutions require ongoing fuel purchases or reliable grid power. A climate battery, once installed, has no operating cost and no supply chain dependency. For small farms, Indigenous land stewards, and communities building food resilience outside industrial supply chains, this changes the calculus of season extension entirely.
How it works
The system has two operating phases driven by a single low-wattage fan and a buried pipe network. The pipes are typically 4-inch diameter perforated drainage pipe laid 4 to 6 feet below the greenhouse floor in a grid or loop pattern.
The soil as a storage medium
Soil has a volumetric heat capacity of roughly 1.0–1.5 MJ/m³·K — lower than water, but far greater than air. A system with 50–100 linear feet of buried pipe surrounded by moist soil can store and return meaningful amounts of thermal energy across a full diurnal cycle.
Moisture matters. Dry sand holds very little heat per cubic foot. Moist loam or clay holds significantly more. This is one reason sensor data on both temperature and humidity — not temperature alone — is important for understanding system performance.
3D scan of the installation
The model below is a photogrammetric 3D scan of the Greenhouse Twins installation taken April 21, 2025. Sensor nodes are marked and interactive — click any labeled node to view its current readings and open a data panel. Toggle between the textured GLTF photogrammetry scan and the structural STL model using the button below.
Sensor nodes
Four sensor nodes are currently installed in the climate battery. All transmit temperature and humidity data over LoRa mesh to a local base station. Data is fetched live from a Google Sheets log and overlaid directly on the 3D model above.
Node readings are transmitted over a Meshtastic LoRa mesh network to a base station at the site. The base station logs readings to a Google Sheet. This page fetches the most recent rows from that sheet live and renders them on the 3D model as interactive markers.
- Ambient temperature (°C), sampled every 15 minutes
- Ambient relative humidity (%)
- Timestamp and date of each reading
- Sensor error detection: readings outside physical bounds are flagged on the model
- Click any sensor label in the 3D view to open a detailed data panel
The twins methodology
The project uses a paired comparison design: two physically identical greenhouses are built side by side. One receives the full climate battery installation — buried pipes, fan, and sensor network. The other is the control, with no subsurface system and no active heating.
Both greenhouses grow the same crops under the same management. Over a full growing season, sensor data from both structures builds a dataset showing the actual thermal difference the system produces, the energy cost avoided, and how performance varies with soil conditions, season, and crop load.
Climate battery performance claims in the literature vary widely — some installations report 10°F overnight improvements, others show modest gains. Variables like soil type, pipe depth, fan sizing, and local climate all interact. The only way to know what a specific installation actually delivers is to measure it against an identical baseline in the same location at the same time. That is what the twins design provides.
What we are measuring
- Overnight low temperature differential between the two greenhouses
- Daytime peak temperature differential (cooling effect)
- Soil temperature gradient across the TX1 → TX3 → TX7 depth profile
- Energy stored and returned per diurnal cycle (estimated from sensor data)
- Growing season extension in days relative to the control greenhouse
System specifications
| Parameter | Current Installation |
|---|---|
| Active sensor nodes | 4 — TX1, TX3, TX7, SI1A |
| Sensor protocol | Meshtastic / LoRa 915 MHz |
| Data logging | Google Sheets via base station |
| Node power | Solar + battery (no grid connection) |
| Sampling interval | 15 minutes |
| 3D scan format | GLTF (photogrammetry) + STL (structural) |
| 3D scan date | April 21, 2025 |
| 3D rendering | Three.js r163, WebGL, OrbitControls |
| Data fetch | Live from Google Sheets JSON endpoint |