Detailed mapping unit design based on soil–landscape relation and spatial variability of magnetic susceptibility and soil color
Introduction
The demand for detailed information about soil to assist agricultural and politic decision making is increasing (Delden et al., 2011). Detailed information can be used to support sustainable production. Current studies report the requirement for the developing of detailed indexes about global impacts of land use and management (Rockström et al., 2009).
The lack of experienced pedologists (mappers) (Demattê et al., 2007) and governmental resources (Bazaglia Filho et al., 2013) as well as discrepancy among proposed soil maps (Delarmelinda et al., 2011) detracts the spread of this type of information. The soil information provided by groundbreaking projects, such as RADAM Brasil (Brasil, 1981) and the Soil Survey Brazilian Service (Brasil, 1960) meets some requirements of Brazilian regions. However, in São Paulo State, this detailing level is not enough to attend the current demand of the agricultural industry. The forecasts point out a 15.7% raise in the domestic energy supply from sugarcane biomass (Brasil, 2012), which shows the need for more detailed maps to subsidize better production indexes.
For large territories, such as Brazil, Russia and China, the characterization and detailed design of soil mapping units (known as design a draft that defines boundaries for varied characteristic field locations) are hardly obtained. Among the causes is the natural variability of soil conditioning factors (geology, landscape shapes and others) (Legros, 2006). In order to minimize this problem, Soil Science experts manage the cumulative knowledge about soils to better approach the geovariability and pedodiversity. Their aim is to propose an alternative to the deficit of detailed information.
In the USA, soil mapping units were designed based on field variations and, subsequently, taxonomical units were also defined (Soil Survey Staff, 1975). In Brazil, the reverse process occurred. Firstly, taxonomical units were defined, with theoretical boundaries for the diagnostic attribute classification (soil color, base saturation, clay content, etc.). Based on these boundaries, the field unit designs were then established, so that the taxonomical unit does not match with the mapping unit. This problem may be worsened for places with detailed information requirements (Buol, 1990).
Some studies indicate that the solution to the mapping and taxonomical unit definition is in the previous pedodiversity evaluation (Ibáñez et al., 2009, Minasny et al., 2010). Others start from the importance of numerical classification models (Minasny and McBratney, 2007, Trangmar et al., 1985). In this case, the spatial distribution of soil attributes is considered during the boundary identification among soil taxonomical classes and in the detailed information supply about the soil genesis–landscape evolution relation (Hudson, 1992, Marcos, 1982). Our implicit knowledge about landscape models is the major cause of soil survey inefficiencies.
Hence, the understanding of this paradigm may be the base for the detailed design of soil units (Hudson, 1992, Swanson, 1993). In this way, the existing empiric models, based on the implicit knowledge, need to be converted into deterministic models, resulting in a better comprehension of cause–effect relations. In Brazil, the former Soil Committee belonging to the Soil Survey Brazilian Service (Brasil, 1960), currently named EMBRAPA Soils, had already mentioned implicit and explicit relations between soil and relief (Vidal-Torrado et al., 2005).
In this context of pedological study evolution, based on soil–landscape relation, the systematization of the multidisciplinary work proposal for ultra-detailed studies stands out, which is known as structural analysis of the pedological coverage. Such proposal is implemented in two stages: two-dimensional analysis (pedological survey performed perpendicularly to the contour lines) and three-dimensional analysis (using isodifferentiation curves in a defined area) (Ruellan et al., 1989). The hypothesis of this proposal is to integrate two-dimensional, conceptual (Dalrymple et al., 1968, Daniels and Hammer, 1992, Daniels et al., 1971) and mathematical models (Cunha et al., 2005, Pereira et al., 1996) with three-dimensional ones (Hammer et al., 1995, Montgomery, 2003) and geostatistical analyses (Isaaks and Srivastava, 1989, Rossi et al., 1992, Vieira, 2000, Zawadzki et al., 2005) helping in the area design based on field variability.
In this proposal type, there are two limiting factors: the computer-related one and number of required samples. A great number of works are unfeasible because of the sample number (Demattê et al., 2007, McBratney et al., 2002). Thus, despite their quality, this kind of study presents high costs of time and resources as well as a large team for its performance (Lagacherie et al., 1995). As the computational technology advances in Soil Sciences (Hastie et al., 2001, McBratney et al., 2002), together with the use of magnetic susceptibility (MS), some of the limitations of two- and three-dimensional model integration can be solved. The MS and color are covariate attributes of soil formation factors and processes (Maher and Thompson, 1999), showing a representativeness of physical, chemical and mineralogical attributes (Siqueira et al., 2010b, Torrent et al., 2007, Verosub and Roberts, 1995, Zawadzki et al., 2012).
Therefore, the objective of this research was to identify landscape areas with different patterns of variability (detailed mapping unit design) using a statistic protocol with data of magnetic susceptibility (MS) and soil color that are covariate attributes of soil formation factors and processes.
Section snippets
Location, area characterization and sampling scheme
The studied area, of 380 ha, is located in the São Paulo State Northeast, in Guatapará Town, Brazil (Fig. 1a). The geographical coordinates are 21° 28′ S and 48° 01′ W, at 600 m of maximum altitude. According to the Thornthwaite (1948) classification, the local climate can be defined as B1rB′4a′, humid mesothermic with a small water deficiency, with a summer evapotranspiration 70% lower than the annual one.
The natural vegetation was composed of sub-deciduous tropical forest; however, it has been
Two-dimensional analysis: soil transect
The compartment boundaries identified in the field, according to the conceptual model, were top/half slope at 75 m, half slope/shoulder at 825 m, shoulder/slope at 1025 m, slope/transportation foothill at 1440 m and transportation foothill/deposition foothill at 1775 m. The SMW results show that there are five emphasized peaks: the 1st at 150 m; 2nd at 675 m; 3rd at 1075 m; 4th at 1375 m; and 5th at 1825 m. The 2nd, 3rd and 4th peaks indicate the presence of different compartments in areas that were
Conclusion
Magnetic susceptibility (MS) was more efficient in the compartmentalization of the landscape (identification of areas with different patterns of pedogenetic variability) than the hue determined by diffuse reflectance spectroscopy for Oxisols under the transition Basalt and Colluvial–Elluvial–Alluvial Deposits.
The MS and soil–landscape relationship were effective in delineating units of detailed mapping. Due to the lack of protocols for identifying areas based on their pattern of variability,
Acknowledgments
To São Paulo Research Foundation (FAPESP) for granting a graduate scholarship to the first author (Project number 2011/06053-3) <http://www.bv.fapesp.br/pt/bolsas/126039/suscetibilidade-magnetica-e-espectroscopia-de-reflectancia-disfusa-na-identificacao-de-areas-de-mane/> and Brazilian National Council for Scientific and Technological Development – CNPq.
References (120)
- et al.
Procedures using diffuse reflectance spectroscopy for estimating hematite and goethite in Oxisols of São Paulo, Brazil
Geoderma Reg.
(2015) - et al.
Magnetic study of Brazilian peats from São Paulo state
Geoderma
(2004) - et al.
Spatial correlation between the composition of the clay fraction and contents of available phosphorus of an Oxisol at hillslope scale
Catena
(2013) - et al.
Spatial variability of soil elements in two plant communities of NW Spain
Geoderma
(2007) Some factors affecting an increase in magnetic susceptibility of cement dusts
J. Appl. Geophys.
(2001)- et al.
The fractal mind of pedologists (soil taxonomists and soil surveyors)
J. Ecol. Complex.
(2009) - et al.
Effect of different soil conditions on magnetic parameters of power-plant fly ashes
J. Appl. Geophys.
(2001) - et al.
A soil survey procedure using the knowledge of soil patterns established on a previously mapped reference area
Geoderma
(1995) - et al.
- et al.
Improved prediction and reduction of sampling density for soil salinity by different geostatistical methods
Agric. Sci. China
(2007)