Skip to main content
ARS Home » Midwest Area » Columbia, Missouri » Cropping Systems and Water Quality Research » Research » Publications at this Location » Publication #325193

Title: A global spectral library to characterize the world's soil

item VISCARRA ROSSEL, RAPHAEL - Commonwealth Scientific And Industrial Research Organisation (CSIRO)
item BEHRENS, THORSTEN - University Of Tubingen
item BEN-DOR, EYAL - Tel Aviv University
item BROWN, D - Washington State University
item DEMATTE, J - Sao Paulo State University (UNESP)
item SHEPHERD, K - International Centre For Research In Agroforestry (ICRAF)
item SHI, Z - Zheijiang University
item STENBERG, B - Swedish University Of Agricultural Sciences
item STEVENS, A - Catholic University Of Leuven
item ADAMCHUK, V - McGill University - Canada
item Sudduth, Kenneth - Ken

Submitted to: Earth-Science Reviews
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 1/25/2016
Publication Date: 2/6/2016
Citation: Viscarra Rossel, R.A., Behrens, T., Ben-Dor, E., Brown, D.J., Dematte, J.A., Shepherd, K.D., Shi, Z., Stenberg, B., Stevens, A., Adamchuk, V., Sudduth, K.A. et al. 2016. A global spectral library to characterize the world's soil. Earth-Science Reviews. 155:198-230. doi:10.1016/j.earscirev.2016.01.012.

Interpretive Summary: Developing a better understanding of soils and how they vary in space and time is important for sustainable agricultural production and ecosystem protection. To do this on a large scale with standard techniques is a daunting task, so there is an increasing interest in using sensing technology. One sensing method that has been useful to characterize important soil attributes across fields and regions is visible-near-infrared (vis-NIR) soil reflectance spectroscopy. In this research, our team assembled a database of vis-NIR reflectance data on soils from around the world and used it to estimate such attributes as soil carbon and clay content. Advanced data mining and machine learning techniques were successfully applied to develop calibration models. Additionally, the compiled database will facilitate future research as new calibration techniques are developed. This research illustrates a sensor-based approach that can help deal with the shortage of data on soil and meet the growing demand for information to assess and monitor soil at scales ranging from regional to global.

Technical Abstract: Soil provides ecosystem services, supports human health and habitation, stores carbon and regulates emissions of greenhouse gases. Unprecedented pressures on soil from degradation and urbanization are threatening agro-ecological balances and food security. It is important that we learn more about soil to sustainably manage and preserve it for future generations. We developed and analyzed a global visible-near infrared (vis-NIR) soil spectroscopic database. It is currently the largest and most diverse database of its kind. We show that the information encoded in the spectra can describe soil composition and be associated to land cover and its global geographic distribution, which acts as a surrogate for global climate variability. We also showed the usefulness of the global database for predicting soil attributes, such as soil organic and inorganic carbon and clay content. Using wavelets as a pretreatment before the spectroscopic modeling helped us to analyze a database with spectra and soil attributes that were measured in different laboratories using different spectrometers and methods. The spectroscopic models were parsimonious and robust. We found that, globally, modeling a diverse set of spectra with a data mining, machine learning algorithm can find the local relationships in the data to produce accurate predictions. Using these spectroscopic models, we derived a harmonized global soil attribute dataset, which might facilitate research on soil at the global scale. We propose that this spectroscopic approach would help to deal with the shortage of data on soil and to meet the growing demand for information to assess and monitor soil at scales ranging from regional to global. We hope this work might reinvigorate our community's discussion towards larger, more coordinated collaborations and encourage other contributions. We also hope that use of the database will deepen our understanding of soil so that we might sustainably manage it and push the research outcomes of the soil, earth and environmental sciences towards applications that we have not yet dreamed of.