Pulse Brain · Growing Health Evidence Index
Tier 3 — Observational / field trialPeer-reviewed

PRISMA and Sentinel-2 spectral response to the nutrient composition of grains

Mariana Belgiu, Michael Marshall, Mirco Boschetti, Monica Pepe, Alfred Stein, Andy Nelson

Remote Sensing of Environment · 2023

Read source ↗ All evidence

Summary

This study demonstrates that hyperspectral (PRISMA) and multispectral (Sentinel-2) satellite imagery can effectively predict nutrient concentrations in cereal grain crops before harvest, offering a potentially cost-effective and spatially scalable alternative to laboratory-based chemical analysis. Using field samples from Italy and remote sensing data collected at key crop development stages, the authors developed predictive models achieving moderate to good performance (R² = 0.51–0.73), with Sentinel-2 paired with regression modelling generally outperforming simpler spectral indices. This approach addresses a practical bottleneck in monitoring crop nutrient quality at scale, which is relevant to addressing micronutrient malnutrition globally.

UK applicability

The methodology could potentially be adapted to UK arable systems, though model calibration would be needed for British soil types, climate conditions, and crop varieties. The use of Sentinel-2 imagery is particularly applicable as it provides free, routinely acquired data over the UK, though validation on UK farm conditions would be essential before operational deployment.

Key measures

Coefficient of determination (R²) and root mean square error (RMSE) for predictions of eight macronutrients and micronutrients (Ca, Fe, K, Mg, N, P, S, Zn) across four crop types using two-band vegetation indices (TBVIs) and partial least squares regression (PLSR) modelling approaches

Outcomes reported

The study developed and compared two remote sensing approaches (PRISMA and Sentinel-2 imagery combined with two-band vegetation indices or partial least squares regression) to predict nutrient concentrations (calcium, iron, potassium, magnesium, nitrogen, phosphorus, sulphur, and zinc) in corn, rice, soybean, and wheat grains before harvest. Model performance varied by crop and nutrient, with coefficient of determination values ranging from 0.51 to 0.73 for the best predictions.

Theme
Measurement & metrics
Subject
Crop nutrient density & mineral composition
Study type
Research
Study design
Field trial
Source type
Peer-reviewed study
Status
Published
Geography
Italy
System type
Arable cereals
DOI
10.1016/j.rse.2023.113567
Catalogue ID
SNmov5kr07-s8195i

Topic tags

Pulse AI · ask about this record

Dig deeper with Pulse AI.

Pulse AI has read the whole catalogue. Ask about this record, its theme, or how the findings apply to UK farming and policy — every answer cites the underlying studies.