Soil Organic Carbon Estimation in Croplands by Hyperspectral Remote APEX Data Using the LUCAS Topsoil Database

Fabio Castaldi,Arwyn Jones,Sabine Chabrillat,Bas Van Wesemael,Bart Bomans,Kristin Vreys

doi:10.3390/rs10020153

Abstract

The most commonly used approach to estimate soil variables from remote-sensing data entails time-consuming and expensive data collection including chemical and physical laboratory analysis. Large spectral libraries could be exploited to decrease the effort of soil variable estimation and obtain more widely applicable models. We investigated the feasibility of a new approach, referred to as bottom-up, to provide soil organic carbon (SOC) maps of bare cropland fields over a large area without recourse to chemical analyses, employing both the pan-European topsoil database from the Land Use/Cover Area frame statistical Survey (LUCAS) and Airborne Prism Experiment (APEX) hyperspectral airborne data. This approach was tested in two areas having different soil characteristics: the loam belt in Belgium, and the Gutland–Oesling region in Luxembourg. Partial least square regression (PLSR) models were used in each study area to estimate SOC content, using both bottom-up and traditional approaches. The PLSR model’s accuracy was tested on an independent validation dataset. Both approaches provide SOC maps having a satisfactory level of accuracy (RMSE = 1.5–4.9 g·kg−1; ratio of performance to deviation (RPD) = 1.4–1.7) and the inter-comparison did not show differences in terms of RMSE and RPD either in the loam belt or in Luxembourg. Thus, the bottom-up approach based on APEX data provided high-resolution SOC maps over two large areas showing the within- and between-field SOC variability.

Highlights

Many soil chemical components interact with the electromagnetic radiation within the visible-near infrared (VNIR: 400–1300 nm) and short-wave infrared (SWIR: 1300–2500 nm) spectral regions
The soil organic carbon (SOC) estimation was gained from calibration models obtained by subsetting the Land Use/Cover Area frame statistical Survey (LUCAS) soil database into classes based on 12 soil variables, and the comparison between measured and estimated SOC values showed an overall root mean square error (RMSE) of 4.3 g·kg−1 and a ratio of performance to deviation (RPD) of 2.5 for the calibration
The SOC estimation was gained from calibration models obtained by subsetting the LUCAS soil database into classes based on 12 soil variables, and the comparison between measured and estimated SOC values showed an overall RMSE of 4.3 g·kg−1 and a RPD of 2.5 for the calibration datasets (Table 3)

Summary

Introduction

Many soil chemical components (chromophores) interact with the electromagnetic radiation within the visible-near infrared (VNIR: 400–1300 nm) and short-wave infrared (SWIR: 1300–2500 nm) spectral regions. Overtones of the O–H and H–O–H stretch vibrations in the free water produce two deep absorption features at 1455 nm and 1915 nm, while organic matter has a strong relationship with electromagnetic radiation in the visible region, due to a wide spectral feature centred around 664 nm related to the chlorophyll pigment [3,4]. In order to exploit spectral features, a suitable spectral resolution is necessary, especially for soil variables affecting spectra in narrow spectral regions (e.g., clay and organic matter). Most of the chemometric approaches to estimate soil variables use the whole spectrum; for example, the partial least square regression (PLSR) method exploits both the wavelengths correlated to the target variable (i.e., absorption features) and other bands indirectly related to the target variable [8]. There is an increasing number of papers concerning the estimation of SOC exploiting airborne hyperspectral data [10,11,12,13,14,15,16,17]

Objectives

Methods

Discussion

Conclusion