ReviewMulti-criteria characterization of recent digital soil mapping and modeling approaches
Introduction
Soil properties are continuously modified by internal factors and anthropogenic impacts generating complex spatial soil patterns. Our understanding and ability to describe them have undergone tremendous change as outlined by Hartemink and McBratney (2008) who argued that we are in an age called “soil science renaissance”. Introduction of digital technologies, such as remote and soil sensing, computer processing speed, management of spatial data, quantitative method to describe soil patterns and processes, and scientific visualization methods have provided new opportunities to predict soil properties and processes. The history of digital soil mapping and modeling (DSMM) is marked by adoption of new tools and techniques to analyze, integrate, and visualize soil and environmental datasets (Grunwald, 2006).
To quantify soil patterns mapping or modeling methods can be used. Modeling is defined as “use of mathematical equations to simulate and predict real events and processes”, whereas mapping emphasizes to “make a map” or “to depict something on a map”. Digital soil mapping (DSMa) [or predictive soil mapping] can be defined as “computer-assisted production of digital representations of soil type or soil properties, which involve the creation and population of spatially-explicit information by the use of field and laboratory methods, coupled with spatial and non-spatial soil inference systems”. A comprehensive review of DSMa was provided in McBratney et al. (2003) and an overview of pedometric techniques used in DSMa by McBratney et al. (2000) and Grunwald (2006).
To predict soil properties and classes the SCORPAN approach (Eq. (1)) has been formulated by McBratney et al. (2003) rooted in earlier works by H. Jenny and V.V. Dokuchaev.where
- Sa
soil attribute
- Sc
soil class
- s
soils, other attributes of the soil at a point
- c
climate factor
- o
organisms, vegetation or fauna or human activity
- r
relief (topography, landscape attributes)
- p
parent material, lithology
- a
age, the time factor
- n
space, spatial position
- t
time (where t is defined as an approximate time).
Digital soil modeling (DSMo) is more comprehensive than DSMa because it entails description and/or prediction of soils in space and through time usually implemented using mechanistic simulation models. Pedometrics integrates DSMa and DSMo into a spatially and temporally explicit framework merging soil science, environmental science, GIScience, cartography, statistics, geostatistics, and mathematics for the study of the distribution and genesis of soils.
Hartemink et al. (2001) noted a strong decline in soil mineralogy, soil morphology and soil genesis research in comparison to a strong increase in pedometrics applications from 1967 to 2001. Historically, soil science has been rooted in agriculture, geology, and chemistry. But the emphasis has shifted from classification and inventory to understanding and quantifying spatially and temporally soil patterns in relation to hydrologic cycle and ecosystems health. This environmental-centered approach views soils as integral part of an ecosystem interacting with environmental factors generating complex patterns and processes that co-evolve through time.
In this paper a comprehensive review of DSMM is provided covering a wide range of specialized, mathematical prototype models tested on limited geographic regions and/or datasets and simpler, operational DSMM used for routine mapping over large soil regions.
The aim of this study was to assess the usefulness of recent DSMM approaches to meet specific needs at local, national, and global scales. Such needs are not solely soil-centered, but consider broader issues such as soil and water quality, food security, carbon cycling and sequestration, global climate change, sustainable land management, and more. Strengths and weaknesses of current DSMM, their potential to be operationalized in soil mapping/modeling programs, research gaps, and future trends are discussed.
Section snippets
Materials and methods
Publications of two high-impact and international soil science journals – Geoderma and Soil Science Society of America Journal – were reviewed covering the period 2007–2008 to identify studies focused on DSMM. The Soil Sci. Soc. Am. J. Issues 2007 Vol. 71 No. 6 to 2008 Vol. 72 No. 5 and Geoderma 2008 Vol. 140 Issue 4 to Vol. 146 Issues 1–2 with a total number of 90 papers were considered in the analysis. A selective approach was used to identify published DSMM studies that cover the
Prediction of soil properties and classes at plot/field and coarse landscape scales
Prediction of soil properties and classes at plot/field scale encompassed 35.6% (N: 32) of all reviewed studies, of which 87.5% (N: 28) predicted soil properties and 12.5% (N: 4) soil classes (Table 1). A total of 31.3% (N: 10) of plot/field scale studies used a mechanistic modeling approach (DSMo) to predict soil properties. For example, Bricklemyer et al. (2007) employed the CENTURY model and Causarano et al. (2007) EPIC to model soil organic carbon. At coarser landscape scales a total of
External and internal factors imparting control on soil properties and classes
Most DSMM studies focused on external drivers (controlling factors) such as climate, vegetation, and land use that modulate SOC, TP or other soil properties that were predicted (compare Rivero et al. (2007), Grimm et al. (2008), and Schulp and Veldkamp (2008)). Environmental factors such as land cover and terrain can be rapidly and cost-effectively mapped using remote sensing and digital elevation models covering large areas. But these models did not incorporate intrinsic soil properties such
Conclusion and outlook
A holistic understanding of soil patterns and processes is critical because soils are the matrix which all transport and transformation processes must pass. Scaling of soil properties and processes to an appropriate scale to address societal needs and evaluate ecosystem services will be critical in the future. DSMM can contribute to enhancing the decision making process to resolve local, national, and global environmental problems of significance by providing accurate soil predictions that
Notes
The ranking of DSMM studies into different categories was done to the best knowledge and understanding of the author. High, average or low scores were intended to classify a given study relative to other DSMM studies and there was no intention to reflect on the quality of a specific paper.
Acknowledgements
I would like to thank Dr. Janis Boettinger (organizer and host) and the program committee who invited me to give a keynote talk at the 3rd Global Conference on Digital Soil Mapping organized by the International Union of Soil Sciences, Soil Science Society of America and Utah State University, Sept. 30–Oct. 3, 2008, Logan, UT. This provided the stimulus to conduct the review analysis on DSMM presented in this manuscript. I am very grateful for their encouragement. I would also like to thank Bob
References (112)
- et al.
Soil condition classification using infrared spectroscopy: a proposition for assessment of soil condition along a tropical forest-cropland chronosequence
Geoderma
(2008) Development of class pedotransfer functions of soil water retention — a refinement
Geoderma
(2008)- et al.
Optimization of pedotransfer functions using an artificial neural network ensemble method
Geoderma
(2008) Mapping soil carbon stocks of Central Africa using SOTER
Geoderma
(2008)- et al.
Spectral reflectance based indices for soil organic carbon quantification
Geoderma
(2008) - et al.
Sensitivity of simulated soil heterotrophic respiration to temperature and moisture reduction functions
Geoderma
(2008) - et al.
Soil distribution in the McMurdo dry valleys, Antarctica
Geoderma
(2008) - et al.
Forest soil acidification assessment using principal component analysis and geostatistics
Geoderma
(2007) - et al.
Geostatistical homogenization of soil conductivity across field boundaries
Geoderma
(2008) Using a global VNIR soil-spectral library for local soil characterization and landscape modeling in a 2nd-order Uganda watershed
Geoderma
(2007)