Engine 02 — Soil Analytics & Nutrient Intelligence
Terranova Agriculture Intelligence Platform

Beneath every
harvest lies a story
written in soil

Traditional soil testing captures point-in-time snapshots at sparse sample locations. Engine 02 builds a living, continuous, high-resolution digital model of your soil — mapping every nutrient gradient, every pH shift, every organic carbon pocket — and translates it into variable-rate prescriptions that feed each zone exactly what it needs.

92.9%
AI soil classification accuracy using image-based deep learning
99.9%
Soil texture classification accuracy via neural network models
1.2B
Farm sensors projected globally for IoT-enabled soil monitoring
22%
Average fertilizer cost reduction through precision application
The Foundation

Soil is the most complex biological system on Earth — a single gram contains up to 10 billion microorganisms across thousands of species. Yet most farmers manage this extraordinary ecosystem with a handful of lab tests taken months apart from scattered sample points. The result is uniform fertilizer application across fields where nutrient levels can vary by 300% within a single acre — overfeeding some zones while starving others, wasting billions in chemical inputs, and degrading the very resource that sustains production.

Engine 02 replaces sparse, static soil data with a continuous, high-resolution digital soil model that evolves in real-time — integrating in-field IoT sensors, satellite spectral indices, historical yield maps, and AI-driven nutrient modeling to generate variable-rate prescriptions at sub-field resolution.

Soil Profile Intelligence

Reading every layer

Engine 02 analyzes the complete soil profile — from organic surface matter through mineral horizons to bedrock geology — building a three-dimensional model of nutrient availability, drainage characteristics, and root zone potential.

O
Organic Horizon
Surface litter, decomposing organic matter, humus layer. Critical for carbon sequestration assessment, microbial activity, and moisture retention capacity.
Organic carbon · Microbial biomass · Decomposition rate · C:N ratio
A
Topsoil — Active Root Zone
Primary nutrient exchange zone. Where 80% of crop root activity occurs. Highest concentration of available NPK, microbial populations, and soil organic matter.
NPK availability · pH · CEC · Organic matter % · Texture classification
B
Subsoil — Mineral Accumulation
Illuvial horizon where leached minerals, clays, and iron oxides accumulate. Determines deep-root nutrient access, water holding capacity, and compaction resistance.
Clay content · Fe/Al oxides · Drainage class · Compaction index
C
Parent Material
Weathered bedrock and geological substrate. Determines long-term mineral supply, soil formation rate, and baseline chemistry that governs field-level variability.
Mineralogy · Weathering rate · Geological classification · Base saturation
R
Bedrock — Geological Baseline
Consolidated rock layer. Provides the geological context that explains why adjacent fields with identical management practices can produce dramatically different yields.
Rock type · Permeability · Groundwater interaction · Geochemistry
Analytical Modules

Seven modules. Complete soil intelligence.

Each module fuses IoT sensor networks, satellite spectral imagery, laboratory calibration data, and machine learning models to deliver continuously updated, spatially precise soil analytics.

Module 01
AI Soil Classification & Texture Mapping
92.9% classification accuracy across black, red, clay, and alluvial soils

Soil texture and classification fundamentally govern water holding capacity, drainage behavior, nutrient retention, and workability — yet many farms operate without accurate, spatially resolved soil maps. Engine 02's classification module uses deep learning CNNs trained on thousands of soil images to classify soil types with over 92% accuracy, while neural network models achieve 99.96% accuracy in soil texture classification based on compositional analysis. The system combines satellite-derived bare-soil indices, drone-captured surface imagery, and IoT sensor data to produce field-level texture maps at sub-acre resolution — identifying sand, silt, and clay fractions that determine how every other soil parameter behaves. Digital Soil Mapping leverages these classifications to produce detailed spatial maps of soil features across entire farm operations.

Performance
92.9%
Soil type classification accuracy using CNN image recognition models
99.96%
Texture classification accuracy via ANN compositional analysis
Module 02
NPK Spatial Mapping & Deficiency Detection
Continuous nitrogen, phosphorus, and potassium monitoring at sub-field resolution

Nitrogen, phosphorus, and potassium are the three macronutrients that most directly govern crop yield — and their availability varies dramatically across even a single field due to topography, drainage patterns, historical management, and organic matter distribution. Engine 02 deploys in-situ NPK sensors using ion-selective electrodes that provide real-time nutrient quantification, fused with satellite spectral indices that correlate canopy reflectance patterns to underlying soil nutrient status. AI prediction models achieve mean absolute error as low as 1.975 for NPK estimation — enabling the system to generate high-resolution nutrient maps that reveal deficiency hotspots, over-application zones, and the precise spatial gradients that drive within-field yield variability. These maps update continuously as sensor data streams in, providing a living picture rather than a point-in-time snapshot.

Performance
MAE 1.97
NPK prediction accuracy via ANN models with IoT sensor integration
Sub-acre
Spatial resolution for nutrient gradient mapping across field operations
Module 03
Organic Carbon & Microbial Intelligence
Soil organic carbon tracking for health assessment and carbon credit verification

Soil organic carbon is the single best indicator of overall soil health — governing water retention, nutrient cycling, microbial activity, and structural stability. It is also the foundation of agricultural carbon credit systems, where verified increases in soil organic carbon can generate meaningful revenue from regenerative practices. Engine 02 tracks soil organic carbon through a combination of near-infrared spectroscopy sensors, satellite-derived soil indices on bare fields, and calibrated laboratory analysis at reference points. The system monitors changes in organic carbon over seasons and years, correlating management practices — cover cropping, reduced tillage, compost application — with measurable carbon accumulation rates. AI models predict future carbon trajectories based on current practices, enabling farms to optimize both soil health and carbon credit potential simultaneously.

Performance
MRV
Satellite-verified measurement, reporting, and verification for carbon credits
±0.1%
Organic carbon change detection sensitivity over seasonal time horizons
Module 04
pH & Micronutrient Analysis
pH prediction accuracy MAE 0.103 with integrated micronutrient profiling

Soil pH controls the bioavailability of virtually every nutrient — a half-point pH shift can lock up or release critical micronutrients like iron, manganese, zinc, and boron that determine crop quality and disease resistance. Engine 02 maps pH variability at sub-field resolution using in-situ ISFET sensors that stabilize within 100 seconds, delivering real-time measurements without laboratory delays. AI models trained on regional soil datasets achieve pH prediction MAE of just 0.103, enabling the system to identify liming requirements, detect acidification trends, and map micronutrient availability zones that traditional sampling programs miss entirely. The module integrates soil electrical conductivity data to detect salinity gradients that threaten crop establishment in irrigated systems, and monitors the secondary and trace elements — calcium, magnesium, sulfur, zinc, boron, copper, manganese — that are increasingly recognized as yield-limiting in intensive production.

Performance
MAE 0.10
pH prediction accuracy using neural network regression models
100sec
ISFET sensor stabilization time for in-situ pH and nutrient measurement
Module 05
Variable-Rate Fertilizer Prescription
Zone-specific NPK prescriptions reducing fertilizer cost by 22% average

The culmination of Engine 02's analytical capabilities is the variable-rate prescription — a georeferenced map that tells precision application equipment exactly how much of each nutrient to apply at every point in the field. The system combines current soil nutrient status, crop-specific demand curves, target yield goals, and removal rates from previous harvests to calculate the optimal fertilizer blend and rate for each management zone. AI advisory systems analyze current weather, growth stage, and soil conditions to generate real-time fertilizer recommendations that adapt to changing conditions. The prescriptions are exported directly to precision spreader and sprayer equipment as ISOBUS-compatible application maps, eliminating manual rate calculations and enabling fully automated variable-rate application that reduces fertilizer expenditure by an average of 22% while maintaining or improving yield targets.

Performance
22%
Average fertilizer cost reduction through AI-optimized variable-rate prescriptions
ISOBUS
Direct prescription export to precision application equipment systems
Module 06
Soil Health Trending & Degradation Alerts
Multi-season trend analysis detecting compaction, erosion, and acidification

Soil degradation is agriculture's slow-motion crisis — compaction, erosion, acidification, salinization, and organic matter depletion advance incrementally, often invisible until yield declines become irreversible. Engine 02 tracks soil health indicators across multiple seasons and years, building trend lines that reveal whether management practices are building or depleting the soil resource. The system generates automated degradation alerts when key indicators cross critical thresholds: pH dropping below 5.5, organic matter declining more than 0.2% year-over-year, compaction indices rising above root-limiting values, or electrical conductivity exceeding salt-tolerant crop thresholds. AI-driven analysis correlates soil health trends with specific management practices — tillage intensity, crop rotation, cover crop usage, irrigation patterns — providing actionable recommendations for reversing degradation trajectories before they impact yield.

Performance
Multi-yr
Longitudinal soil health trending with season-over-season comparison
Auto
Degradation alerts triggered when indicators breach critical thresholds
Module 07
3D Digital Soil Modeling
Volumetric soil models integrating depth, spatial, and temporal dimensions

The ultimate expression of Engine 02 is the three-dimensional digital soil model — a volumetric representation of every soil property at every point in the field, across depth and time. Digital Soil Mapping leverages machine learning algorithms to produce detailed spatial maps of soil features across extensive land areas, offering accurate information on nutrient concentrations, texture, moisture, organic carbon, and pH at resolutions that traditional sampling could never achieve. Engine 02's 3D model integrates surface sensor data, multi-depth probe measurements, satellite spectral indices, elevation models, and geological survey data to build a complete digital twin of the soil resource. This model becomes the decision foundation for every agronomic choice — crop selection, variety placement, tillage strategy, drainage investment, and long-term land management planning. The model evolves with every new data point, becoming more accurate with each season of operation.

Performance
3D
Volumetric soil modeling across spatial, depth, and temporal dimensions
Evolving
Model accuracy improves continuously with each season of sensor data
Proven Impact

Every dollar of fertilizer, precisely placed

Three deployments demonstrating how sub-field soil intelligence transforms the economics and sustainability of nutrient management.

Corn Belt — 320,000 Acres — Iowa & Illinois
Multi-state corn operation saves $11.2M annually by replacing uniform application with AI-driven variable-rate prescriptions
A major corn producer managing 320,000 acres across Iowa and Illinois had applied nitrogen at a uniform 180 lbs/acre for over a decade — the rate their average yield required. Engine 02 revealed that actual nitrogen needs ranged from 120 to 220 lbs/acre within individual fields, with hilltops requiring 30% less and bottom-ground requiring 20% more than the flat-rate average. Variable-rate nitrogen prescriptions, updated weekly based on sensor data and weather-adjusted crop demand models, reduced total nitrogen use by 18% while increasing average yield by 4.2 bu/acre — because the zones previously over-applied were actually experiencing nitrogen toxicity that suppressed yield, while under-applied zones were leaving potential on the table.
$11.2M
Annual fertilizer savings
18%
Nitrogen reduction
+4.2bu
Per-acre yield gain
320K
Acres optimized
Irrigated Agriculture — 18,000 Acres — Central Valley, California
Specialty crop grower reduces nutrient runoff 42% while maintaining premium grade on 18,000 acres of almonds and tomatoes
California's Central Valley faces escalating regulatory pressure on nitrate contamination of groundwater — a crisis driven by decades of excessive nitrogen application to irrigated crops. A major grower deployed Engine 02 across 18,000 acres of almonds and processing tomatoes, using the pH and micronutrient module to map zinc deficiency zones that had been masked by blanket micronutrient applications, and the NPK spatial mapping module to identify areas where nitrogen was accumulating below the root zone and leaching into groundwater. Variable-rate fertigation prescriptions — delivered through the drip irrigation system — reduced nitrogen application 28% and nutrient runoff 42% while maintaining the premium grade standards that drive commodity pricing in specialty crops.
42%
Nutrient runoff reduction
28%
Nitrogen reduction
100%
Premium grade maintained
18K
Acres monitored
Regenerative Transition — 4,800 Acres — UK Arable
Carbon credit generation produces £420K new revenue as regenerative practices verified by satellite-validated soil carbon tracking
A UK arable farm transitioning to regenerative practices needed to demonstrate measurable soil health improvement to qualify for carbon credit programs and to justify the short-term yield risk of practice changes to their lending bank. Engine 02's organic carbon and soil health trending modules provided continuous monitoring across 4,800 acres, documenting a 0.3% increase in soil organic carbon over three seasons of cover cropping and reduced tillage — verified through satellite MRV protocols that carbon registries accept. The soil health trending module simultaneously demonstrated improvements in aggregate stability, water infiltration rate, and earthworm counts that correlated with the carbon gains. The verified carbon credits generated £420K in new revenue over three years, and the farm's lending bank accepted the Engine 02 data as evidence of improving land asset quality.
£420K
Carbon credit revenue
+0.3%
SOC increase in 3 years
MRV
Satellite verification
4,800
Acres documented
From the Field

The agronomists who trust Engine 02

We thought we knew our soils. We'd been farming these fields for three generations. Engine 02 showed us nitrogen variability within a single field that ranged from 120 to 220 pounds per acre. We'd been putting on 180 everywhere. That uniform rate was simultaneously burning some zones and starving others — and we had no idea.

TW
Tom Westergaard
Owner-Operator, Westergaard Farms, Iowa

The regulators were going to mandate a 30% nitrogen reduction across the entire Central Valley. Engine 02 allowed us to demonstrate that we could achieve a 42% nutrient runoff reduction through precision application without any regulatory mandate — and without losing a dime of yield. That data changed the conversation entirely.

MR
Dr. Maria Reyes-Gutiérrez
Sustainability Director, Central Valley Agricultural Consortium

Our bank wanted to see proof that regenerative practices were actually improving soil value, not just idealism. Engine 02 gave us satellite-verified carbon data, soil health trend lines, and a revenue stream from carbon credits. The bank didn't just renew our facility — they increased it. They could see the asset improving.

JP
James Pemberton
Farm Director, Pemberton Estate, Oxfordshire

Your soil already
holds the answers.
Read them.

Deploy Engine 02 to build a living digital model of your soil — and transform every acre from guesswork to precision.

Enterprise deployment · Variable-rate integration · Carbon verification · API access