Global Soil Spectral Library, A global reference, spectral library and conversion database for harmonization of legacy soil data for the GlobalSoilMap.net project - World Agroforestry Centre (ICRAF) & ISRIC
1 like675 views
The document outlines the objectives and rationale for creating a global soil spectral library to harmonize legacy soil data for the GlobalSoilMap.net project. This initiative aims to establish a consistent reference method for soil data to improve comparability and facilitate global mapping efforts. Additionally, it discusses the use of infrared spectroscopy for rapid soil characterization and highlights the limitations of current global spectral library initiatives.
Global Soil Spectral Library, A global reference, spectral library and conversion database for harmonization of legacy soil data for the GlobalSoilMap.net project - World Agroforestry Centre (ICRAF) & ISRIC
- 1. Global Soil Spectral Library World Agroforestry Centre (ICRAF) & ISRIC A global reference, spectral library and conversion database for harmonization of legacy soil data for the GlobalSoilMap.net project Towards Global Soil Information: activities within the GEO Task Global Soil Data, FAO, Rome, 19 – 23 March 2012
- 2. Rationale • GlobalSoilMap.net largely based on making effective use of legacy soils data. • Legacy soil data consists of field observations, laboratory analyses and conventional soil maps collected over the last 60-80 years as part of numerous national, regional and institutional soil survey programs. • Much of this legacy soils data is fragmented, non-standard between countries and often even within countries. • Difficult to share or exchange data across national, state or even project borders because of significant differences in standards, methods of analysis, definitions and legends. • Need for soil data to be harmonized to a single consistent reference method in order to be comparable globally and suitable for use in producing global maps and databases. • In addition soil infrared spectroscopy and proximal sensing methods increasingly used for estimating soil properties, but as yet no comprehensive global calibration libraries exists (Viscarra Rossel, 2009; Terhoeven et al.,
- 3. Project Objectives 1. Design and populate a global soil property and spectral reference library Full range of soil types & environmental settings on each continent or node (500 profiles per node) Sample from each horizon analysed using commonly used methods for each soil property 2. Provide inter-method calibrations 1. Provide infrared spectral calibrations 1. Maintain and further build the reference library
- 4. Infrared Spectroscopy for rapid soil characterization • Rapid • Low cost • Reproducible •Predicts many soil functional properties
- 5. What soil properties affect spectral shape? These are the key determinants of soil functional properties: nutrient supply & retention; carbon sequestration; hydraulic properties; erodibility; tillage & engineering properties
- 6. www.africasoils.net Randomization of Sentinel Site locations stratified by climate zone African Soil Information Service
- 7. AfricaSoils Sentinel Site based on the Land Degradation Surveillance Framework a spatially stratified, hierarchical, rand omized sampling framework Sentinel site (100 km2) 16 Clusters (1 km2) 10 Plots (1000 m2) 4 Sub-Plots (100 m2)
- 8. ? Cost surfaces, etc. Covariates Remote Sensing (RS) and Spatial Data Elevation Vegetation Hydrology Topographical properties Climate Landsat Legacy data ASTER Quickbird MODIS 500 m 250 m 28.5 m 15 m 2.4 m 0.6 m
- 9. Local(site-level)Cref High resolution digital mapping 10 km
- 11. Instrumentation Dispersive VNIR FT-NIR FT-MIR Handheld NIR/MIR •Portable •Repeatability? •External service •No validation •Benchtop •Repeatability *** •Self serviceable •Validation in-built •ISO compliant •Industry proven •Multipurpose •Benchtop •Repeatability*** •No gas purging •Some servicing •Robotic •Validation in-built •ISO compliant •Outperforms NIR •Handheld •Sample homogeneity? •Variable moisture? •Repeatability? •Still expensive •Rapidly developing •Need to prepare by developing soil reference libraries
- 12. Current global spectral library initiatives • Soil Spectroscopy Group (Geophysical Research Abstracts, Vol. 11, EGU2009-14021, 2009; NIR News 2009 doi: 10.1255/nirn.1131) • global soil spectral library (VNIR) • community of practice for soil spectroscopy BEN DOR, Eyal - Tel Aviv University; BROWN, David - Washington State University; DEMATTE, Alexandre - Sao Paolo University; SHEPHERD, Keith – ICRAF; STENBERG, Bosse - Swedish Agricultural University; VISCARRA ROSSEL*, Raphael – CSIRO Land & Water • Brown et al 2006: VNIR spectra from NSSC archives: 3768 samples from all 50 U.S. states and two tropical territories and an additional 416 samples from 36 different countries in Africa (125), Asia (104), the Americas (75) and Europe (112) • Terhoeven-Urselmans et al 2010: MIR spectra from ISRIC archives: 4438 soil samples from 754 soil profiles (56 countries) but only 971 samples from 18 countries with complete data
- 13. ISRIC global spectral library Terhoeven-Urselmans T, Vagen T-G, Spaargaren O, Shepherd KD. 2010. Prediction of soil fertility properties from a globally distributed soil mid-infrared spectral library. Soil Sci. Soc. Am. J. 74:1792–1799 Results for 30% independent validation set
- 14. Spectral prediction of TOC, POC, Charcoal-C Australian and Kenya soils Janik LJ, Skjemstad JO, Shepherd KD and Spouncer LR (2007) The prediction of soil carbon fractions using mid-infrared-partial least square analysis. Journal of Australian TOC: alkyl–CH2 stretching modes; carbohydrate overtones of the –COH stretch; carboxylic acid– COOH; amide I and II bands; alkyl–CH2 deformation; aromatic–CH in plane deformation; carbohydrate–COH stretch. CHAR: C=C skeletal vibrations; phenolic, or COO stretching vibrations; ring C–H in plane deformations
- 15. Africa Soil Information Service Spectral library Prediction performance for soil physical properties and total carbon and nitrogen using mid-infrared spectroscopy for sentinel sites analysed to date. Validation results are for a 50% random hold-out sample using boosted regression trees.
- 16. Spectral calibration transfer Robotoic high throughput MIR (Shepherd & Walsh, 2007)
- 17. Current global spectral library limitations • Soil spectral calibrations limited by quality and consistency of the reference data • Large libraries needed to cover the global range in soil conditions • Need to include a number spectral instrument types (VNIR, NIR, MIR) A viable library will be based on conducting new reference measures in one laboratory with excellent quality control
- 18. Proposed plan • Phase 1 (6 months): Scoping, evaluation, feasibility assessment • Survey available profiles; additional sampling? • Export/import procedures; shipping • Suite of methods to be used • Analytical lab for spectral and reference methods • Cost analysis • Full proposal development • Phase 2 (3 years): Sampling and analysis • Full implementation • Global database made available