EDITORIAL COMMENTARY Soil spectroscopy: an opportunity to be seized MARCO NOCITA 1, 2 , ANTOINE STEVENS 2 , BAS VAN WESEMAEL 2 , DAVID J. BROWN 3 , KEITH D. SHEPHERD 4 , ERICK TOWETT 4 , RONALD VARGAS 5 andLUCA MONTANARELLA 1 1 Institute for Environment and Sustainability, European Commission, Joint Research Centre, Via E. Fermi 2749, Ispra, Varese, VA, I-21027 Italy, 2 Georges Lema ^ ıtre Centre for Earth and Climate Research, Earth and Life Institute, Universit  e Catholique de Louvain, Louvain La Neuve 1348, Belgium, 3 Whashington State University, 405 Johnson Hall, PO Box 646420, Pullman, WA 99164-6420, USA, 4 World Agroforestry Centre (ICRAF), United Nations Avenue, PO Box 30677, Nairobi 00100, Kenya, 5 Food and Agriculture Organization (FAO), Viale delle Terme di Caracalla, Rome 00153, Italy The trade-off between the growing need for large scale soil information and its high cost could be resolved by a wide- spread use of visible and infrared spectroscopy. A recent workshop by the European Commission – Joint Research Centre (EC-JRC) and the Food and Agriculture Organiza- tion (FAO), focused on the measures to foster the global mon- itoring of soils based on spectroscopy. In 1921 Aldo Leopold wrote: ‘All natural resour- ces...are soil or derivatives of soil. Farms, ranges, crops, livestock, forests, irrigation water and even water power resolve themselves into questions of soil. Soil is therefore the basic natural resource’ (Meine & Knight, 1999). Sci- entists have struggled to provide a quantitative assessment of soil due to the heterogeneity of its con- stituents. In the  epoque of ‘climate change caused by CO 2 emissions’, soil, being the largest terrestrial car- bon pool, suddenly became important for stakehold- ers and policy makers and the demand for high- resolution soil data for large areas grew accordingly. Unfortunately, the availability of such data is scarce. Sanchez et al. (2009) stated that ‘communicating soil information among diverse audiences remains challenging because of inconsistent use of technical jargon, and out- dated, imprecise methods’. Inconsistent, outdated, and imprecise methods also hamper attempts to compare and synthesize soil measurements from different countries, regions, times, and studies. At present, these inconsistencies are the main constraint to merg- ing existent soil data and evaluating the state of soil resources at regional-to-continental scales. The devel- opment of cost-effective soil analysis methods is becoming a priority in contemporary soil science (Grunwald et al., 2011), and this is particularly important when considering the creation of new, har- monized, continental-scale soil datasets. A workshop recently hosted by the European Com- mission – Joint Research Centre (EC-JRC) and the Food and Agriculture Organization (FAO) focused on the potential role of soil spectroscopy in addressing the growing need for low-cost and standardized soil data. Over the past 30 years, diffuse reflectance spectroscopy has proved to be a high-throughput, inexpensive, reproducible and repeatable soil analytical technique (Soriano-Disla et al., 2014). Briefly, spectroscopy mea- sures the amount of light reflected (i.e. reflectance) by a soil sample in the visible (Vis; 0.4 – 0.78 lm) and infra- red spectral regions (IR; 0.78 – 25 lm), reacting to its organic and inorganic composition. Soil moisture, organic matter, carbonates, mineralogy and texture are the main constituents influencing the shape of soil reflectance in the Vis-IR region. While the first studies, in the 70–80’s, focused on the interpretation and classi- fication of soil spectra (e.g., Stoner & Baumgardner, 1981), soil spectroscopy rapidly adopted a quantitative approach to predict many soil properties such as organic carbon (Ben-Dor & Banin, 1995), texture (Søren- sen & Dalsgaard, 2005), cation exchange capacity (Janik et al., 1998) and moisture (Dalal & Henry, 1986). Soil spectroscopy, being an indirect method, generates pre- dictions by multivariate calibration models developed on spectral libraries (Fig. 1). These contain both spectral measurements and corresponding soil properties mea- sured with reference methods of soil analyses. Calibra- tions are not reliable for soils not represented in the spectral library, hence the need for building libraries representative of the soil diversity for a region of interest. While soil spectroscopy estimates of soil properties are not as accurate as reference soil analyses, they can improve regional-to-continental soil resource assess- ments, because more samples can be analyzed for a given budget. Light reflectance, being a physical mea- surement, can provide greater consistency across labo- ratories compared with chemical reference methods. This is the strategy followed, for instance, by the Africa Soil Information Service (AfSIS; Box 1). Once a library is constructed, only a fraction of new samples (10% for the AfSIS project highlighted in Box 1) needs to be sub- mitted for reference laboratory analysis to make reliable predictions. This causes a dramatic drop of costs. 10 © 2014 John Wiley & Sons Ltd Global Change Biology (2015) 21, 10–11, doi: 10.1111/gcb.12632 Global Change Biology