Fertility capability soil classification: a tool to help assess soil quality in the tropics Pedro A. Sancheza,*, Cheryl A. Palma, Stanley W. Buolb a Ecosystem Sciences Division, Department of Environmental Science, Policy and Management, University of California, Berkeley, CA 94720-3110, USA b Soil Science Department, North Carolina State University, Raleigh, NC 27695-7619, USA Abstract The soil quality paradigm was originally developed in the temperate region with the overarching objective of approaching air quality and water quality standards. Although holistic and systems- oriented, soil quality focused principally on issues arising from large nutrient and energy inputs to agricultural lands. Soil quality in the tropics, however, focuses on three overarching concerns: food insecurity, rural poverty and ecosystem degradation. Soil science in the tropics relies heavily on quantitative attributes of soils that can be measured. The emotional, value-laden and ������measure everything������ approach proposed by some proponents of the soil quality paradigm has no place in the tropics. Soil quality in the tropics must be considered a component of an integrated natural resource management framework (INRM). Based on quantitative topsoil attributes and soil taxonomy, the fertility capability soil classification (FCC) system is probably a good starting point to approach soil quality for the tropics and is widely used. FCC does not deal with soil attributes that can change in less than 1 year, but those that are either dynamic at time scales of years or decades with management, as well as inherent ones that do not change in less than a century. FCC attributes can be positive or negative depending on the land use as well as the temporal and spatial scales in question. Version 4 is introduced in this paper. The main changes are to include the former h condition modifier (acid, but not Al-toxic) with ������no major chemical limitations������ because field experience has shown little difference between the two and to introduce a new condition modifier m that denotes organic carbon saturation deficit. Additional modifiers are needed for nutrient depletion, compaction, surface sealing and other soil biological attributes, but there is no sufficient evidence to propose robust, quantitative threshold values at this time. The authors call on those actively involved in linking these attributes with plant growth and ecosystem functions to provide additional suggestions that would enhance FCC. The use of diffuse reflectance spectroscopy (DRS) shows great potential on a wide range of tropical soils. The evolution of soil science from a qualitative art into a 0016-7061/03/$ - see front matter D 2003 Elsevier Science B.V. All rights reserved. doi:10.1016/S0016-7061(03)00040-5 * Corresponding author. The Earth Institute at Colombia University, Lamont Hall, PO Box 1000, Palisades, NY 10964, USA. Tel.: +1-646-244-1720 fax: +1-510-217-9717. E-mail addresses: p.sanchez@cgiar.org (P.A. Sanchez), c.palm@cgiar.org (C.A. Palm), stanley_buol@ncsu.edu (S.W. Buol). www.elsevier.com/locate/geoderma Geoderma 114 (2003) 157���185

quantitative science has progressed well in the tropics. Regressing to qualitative and vaguely defined soil quality attributes would be a step backwards. D 2003 Elsevier Science B.V. All rights reserved. Keywords: Soil attributes Tropical soils Soil taxonomy Integrated natural resource management Organic carbon saturation Diffuse reflectance spectroscopy 1. Introduction The development of the soil quality paradigm stems largely from initiatives in industrialized countries of the temperate zone during the 1990s (Doran et al., 1994 Doran and Jones, 1996 Karlen et al., 1997). Soil quality has been defined as ������the capacity of a soil to function within land use and ecosystem boundaries, to sustain biological productivity, maintain environmental quality and promote plant, animal and human health������ (Doran and Parkin, 1994, 1996). This concept departs from the traditional agricultural approach focusing on the productive functions of soils, shifting to a more holistic one that recognizes the various roles soils play in agroecosystems and natural systems (Karlen et al., 1997 Swift, 1999). Soil quality attributes are biological, chemical and physical parameters that can be quantified at specific temporal scales. Soil science in the tropics developed later than in the temperate regions and operates on somewhat different paradigms (Sanchez, 1994, 1997 Swift, 1999). Many of the myths associated with the geographical distribution and productivity of tropical soils are now dispelled (Lal and Sanchez, 1992). Soils in the tropics are regarded by the international policy community as increasingly important in world development issues such as food security, poverty alleviation, land degradation and the provision of environmental services (Wood et al., 2000 Sanchez, 2002b). Unlike the temperate zone, the overarching concerns about soils in the tropics relate to food insecurity, rural poverty and their effects on ecosystem degradation (Pinstrup- Andersen et al., 1999). The main soil-environmental concerns in tropical countries are not related so much to nutrient pollution as in the temperate zone but rather to the opposite���nutrient depletion as well as the loss of soil organic matter (SOM) and its related functions (Smaling, 1993 Sanchez et al., 1997). The major environmental concerns the tropics shares with the temperate region are soil erosion, declining biodiversity, watershed hydrology and adaptation to and mitigation of climate change. There are exceptions to all the above statements because the tropics encompass a wide variety of agricultural and social systems. Parts of Brazil, Mexico and South Africa practice the kind of mechanized, large-scale agriculture, typical of North America and Europe, and face similar soil quality challenges, including nutrient pollution. This paper focuses on situations faced primarily by smallholder farmers across the tropics, with emphasis on Africa, because this is where the most acute problems are. The objective of this paper is to indicate how an updated fertility capability soil classification (FCC) system can be used to identify attributes relevant for plant production and overall ecosystem management in the tropics. We first discuss an overall integrated P.A. Sanchez et al. / Geoderma 114 (2003) 157���185 158

natural resource management framework (INRM) that puts soil attributes into a broad context this is followed by quantitative pedology as the entry point for FCC, the fourth version of this system, the search for biological soil attributes and a general discussion. 2. The integrated natural resource management framework (INRM) In our opinion, soil quality must be viewed within a broader context, as a component of an integrated natural resource management framework (INRM), which fits with current soil quality concepts. The soil component is an integral part of this INRM approach (Dumanski et al., 1998). An INRM research framework, shown in Fig. 1, has been developed by the international agricultural research centers, after years of experimentation with ways to do cross-disciplinary research. INRM is defined as ������the sustainable use of the resource base of agriculture, in order to meet the production goals of farmers as well as Fig. 1. Model of the integrated natural resource management research process for use by the international agricultural research community. Source: CIFOR (2000), modified by the authors of this paper. P.A. Sanchez et al. / Geoderma 114 (2003) 157���185 159

the goals of the rest of the community, and the preservation and enhancement of the global environment������ (CIFOR, 2000 Izac and Sanchez, 2001). In this framework, problems are identified in a participatory, quantitative and multi- disciplinary manner, involving farmers and policymakers from day 1 (Step 1). Scientists involved represent various natural and social sciences. Gone are the days when a soil scientist could conceive an idea, test it in replicated small plots within the confines of an experiment station and give the results to extension workers to recommend to farmers in the tropics. Today���s problems are more complex than that. Interdisciplinary research on alternative solutions then follows (Step 2), using the entry points identified in the first step. Soil scientists, plant breeders, plant pathologists, foresters, livestock specialists and others focus on the production dimension as appropriate (Step 3a). Several social science disciplines are involved in the human well-being dimension, particularly economics, anthropology, sociology and geography (Step 3b). Ecologists, including soil and plant ecologists, focus on the ecosystem resilience dimension along with resource economists (Step 3c). Since agroecosystems are driven by the interaction between ecological, economic and social variables, INRM research has to work back and forth across these three dimensions, as well as at various scales in time and space. When results are apparent, a tradeoff analysis is performed between farmer private benefits (both production and human well-being) and global environmental benefits such as biodiversity protection and carbon sequestration (Step 4). This analysis provides a range of flexible, adaptive options (not a silver bullet) for farmers and policymakers to decide. Scaling-up starts, in many cases, with on-farm research (Franzel and Scherr, 2002), followed by Step 5, pilot development projects and large-scale adoption (Franzel et al., 2001). Feedback is needed at all stages, and the loop is closed with Step 6. This framework, with appropriate local modification, fits well with the holistic concept of soil quality (Doran and Parkin, 1996 Doran and Jones, 1996 Karlen et al., 1997) as well with the need for scientific rigor at all stages of the research-development continuum as articulated by Sojka and Upchurch (1999). 3. Quantitative pedology: the entry point Quantitative pedology is based on soil taxonomy (Soil Survey Staff, 1999), the World Reference Base for Soil Resources (Deckers et al., 1998) and the digitized world soils map (FAO, 1995a). The parameters measured in soil taxonomy have been carefully selected as those necessary for classifying soils as natural bodies. Through soil surveys and geo- graphic information systems, pedon data is scaled-up to larger spatial scales, including higher resolution national, provincial and municipality maps, agroecological zone map- ping (FAO, 1981) and digitized soil and terrain databases (FAO, 1995b). Differences in quantitative pedology are important to establish the broad picture of soils as natural bodies, but what does it mean in agronomic or ecological terms? The limitation of soil taxonomy, the FAO legend and the World Reference Base for Soil Resources is that they quantify only permanent soil attributes, most of which are located in the subsoil. In a quest to identify undisturbed soil and cultivated soil in the same taxa, these soil classification P.A. Sanchez et al. / Geoderma 114 (2003) 157���185 160

systems ignore many inherent or dynamic attributes crucial to plant productivity. These are located mostly in the topsoil, where the majority of plant roots are, both in natural ecosystems and agroecosystems. 4. Fertility capability classification and its relevance to soil quality To overcome this limitation, the fertility capability soil classification (FCC) system was developed over 25 years ago to interpret soil taxonomy and additional soil attributes in a way that is directly relevant to plant growth (Buol et al., 1975 Buol and Couto, 1981 Sanchez et al., 1982). The initial version of FCC (Buol et al., 1975 Buol and Couto, 1981 Sanchez et al., 1982) was succeeded by a second version (Sanchez and Buol, 1985 Buol, 1986), which includes specific interpretations for wetland rice soils. A third version (Smith, 1989 Smith et al., 1990) added a new condition modifier for permafrost and subdivisions of some existing ones. Smith (1989) also developed a thorough rationale for each FCC class and provided detailed interpretations for tropical food crops, pastures and tree crops. An algorithm of this third version was later developed by Yost et al. (1997) with software that converts soil profile data into FCC units plus a series of automatic interpretations and recommendations (http://www.fao.org/ag/AGL/AGLL/fcc3/faorep.htm). The FCC system is widely used. It is included in the worldwide FAO digitized soils map at a resolution of 10 km2 (FAO, 1995a). It is used at national and provincial scales in several countries such as Brazil (Oliveira, 1978), Venezuela (Avilan �� et al., 1979 Brito et al., 1979), Taiwan (Lin, 1984, 1985), United States (Denton et al., 1986), Thailand (Euimnoh, 1984), Indonesia (Sornsumran, 1985), Peru (Paredes, 1986) and Cambodia (White et al., 1997). At the regional scale, FCC has been used in South America (Cochrane et al., 1985) and in the Caribbean for pine plantations (Liegel, 1986). FCC is used at the global scale by the World Resources Institute (1990, 1992), by FAO (1995a) and in the pilot analysis of global agroecosystems (Wood et al., 2000). But is FCC relevant to soil quality? One important aspect of soil quality is that it deals with soil attributes across temporal and spatial scales. Although dynamic, quantitative attributes most sensitive to changes in land use are the most desirable as soil quality indicators (Doran and Parkin, 1996), the importance of inherent, albeit static, soil attributes is recognized as an important component of soil quality (Karlen et al., 1997). FCC was originally conceived as dealing only with inherent soil properties that are the product of soil genesis and cannot be easily changed with time (Sanchez et al., 1982). FCC considers topsoil parameters as well as specific subsoil properties. This is why the FCC system does not include routine soil tests used for N and P fertilizer recommendations. A further reason is that such soil tests are not very useful in farming systems where fertilizer use is not the main nutrient input (Smithson and Sanchez, 2001). The question is at what time scale are we referring to���days, months, years, decades, centuries? The FCC system may represent a useful approach to tackling soil quality in the tropics on a quantitative basis. FCC is not equal to soil quality, which is in many instances a difficult concept to put in practice with sufficient scientific rigor. The next section presents the updated version of FCC. P.A. Sanchez et al. / Geoderma 114 (2003) 157���185 161