COMMUNICATIONS

New developments in soil classification: the World Reference Base for Soil Resources

Introduction: a brief history of soil classification

The first attempts of soil classification go back to the first farmers that cultivated land: they probably used straightforward criteria related to their ease of recognition and their potential and constraints: colour of the topsoil, consistency and permeability, risk for waterlogging, stoniness, etc.

Since then, soil scientists have done their very best to muddy the waters by elaborating numerous individual nomenclatures reflecting more often than not, national pride, language preference, and personal experience.

One can recognize three linked development stages in the recent history of soil classification:

E N V I R O N M E N T	Pedogenetic
Climate/Topography/... Zonal -Azonal-Intrazonal	P R O C E S S E S	S o i l C H A R A C T E R I S T I C S
Russian School (1890-1930)	Chemical and Physical Podzolization, Leaching, Salinization, Ferralitization	Base saturation, clay mineralogy, depth clay increase...
	W. European classifications (1940 - 1960)	Soil Taxonomy ( 1960 - 1990)

We do know for instance that generally in a temperate, cool climate on sandy materials with a pine cover (=environment) cheluviation takes place (organic matter combines with Fe and Al and migrates in the profile) which leads in turn to the creation of a bleached eluvial horizon over a blackish horizon (=processes and morphology). The eluvial horizon is poorer in C, amorphous Fe or Al (=soil characteristics, together making up a diagnostic horizon).

This emphasis on these different but related aspects of soils in various national classification systems did create an effective blockage to the use of soil names and soil characteristics in other, even closely related, disciplines, such as agriculture . This in turn was detrimental not only to soil science but particularly to its application and use in other sciences as well.

An overview of recent developments in soil classification since the nineteen sixties is given in Table 1.

Table 1

Important dates in the recent history of soil classification

1960		Seventh Approximation published by US Soil Survey Staff
		ISSS Congress in Madison asks FAO to prepare Soil Map of the World.
1967		Russian Classification updated (Ivanova and Rozov)
		The French Soil Classification published (CPCS)
1972		Soil Taxonomy published by Soil Survey Staff
1974		Legend of the Soil Map of the World published by FAO and UNESCO
1980		Last volume and map sheet Soil Map of the World published
1982		ISSS creates the International Reference base for Soil Resources (IRB)
1988		FAO, ISRIC and UNESCO publish the Revised Legend of the Soil Map of the World. (Reprinted in 1990)
1992		IRB Working Goup decides to use the Revised Legend of the Soil Map of the World as a basis for the ISSS system. IRB working group renamed as World Reference Base for Soil Resources (WRB).
1994		WRB produces draft World Reference Base at ISSS Congress in Acapulco.
		Draft Topsoil Characterization presented.
1995		FAO produces digital Soil Map of the World and decides together with ISRIC and UNEP to undertake an update of the map using the principles of the SOTER manual.
1996		Australian Soil Classification system revised (Isbell)
1998		WRB working group produces the World Reference Base for Soil Resources in three volumes endorsed and adopted by IUSS as the official soil correlation system of the International Union of Soil Scientists.
		Second Edition Soil Taxonomy published.
		Référentiel Pédologique published in English (Baise).
1999		European Soil Bureau adopts WRB as correlation system for the Soil Map of Europe.
2000		SOTER adopts WRB as a reference system for classifying soil profiles.

When FAO and UNESCO were asked in 1960 by the International Society of Soil Science (ISSS) to prepare a global soil map it was in the first place required that a legend for this map be prepared. The same year "The Seventh Approximation", the forerunner of Soil Taxonomy was published.

A soil legend is NOT in itself a soil classification system, but it requires a recognition and strict nomenclature of the major soil types recognized on the map. This was therefore a unique occasion to bring together the various soil science schools and come to a major soil correlation system agreeable to all.

This was realized in 1974 by:

• adopting Soil Taxonomy as an objective and language-neutral system as a basis;
• simplifying definitions of horizons and epipedons where possible;
• excluding climate as a factor in the classification;
• maintaining well-known major soil names, dear to specific national classifications (Solonchaks, Chernozems, Rankers);
• maintaining a link with pedogenesis where possible.

The Legend of the Soil Map of the World contained 26 Major Soil Groups and 106 Soil Units. Not all were represented on the map (e.g. Gelic Planosols), nor was the system fully comprehensive as not all combinations of diagnostic horizons could be classified in it. The soil map itself was finalized in 1980 and a digital version with a number of thematic layers is available from FAO since 1995. It remains a unique inventory more than 20 years after its finalization, and is presently being updated under the Global SOTER programme.

In 1982, the International Soil Science Society, FAO and UNEP initiated a project to construct an internationally accepted soil classification system through which national systems could be correlated and which would hopefully lead to a unique system on soil nomenclature which would make soil information more accessible to other sciences and to the public at large. This initiative was known as the International Reference Base for Soil Resources (IRB).

In 1988, FAO, ISSS and ISRIC issued a Revised Legend of the Soil Map of the World increasing the number of groups from 26 to 28 and the number of soil units from 106 to 153.

It proposed a "third level" of classification and made a first attempt to recognize soils strongly influenced by human activities (Anthrosols) as a separate category.

In 1992 the IRB working group realized that the Revised Legend was a good framework and indicated that by giving it more depth and explaining its background, IRB would reach its aim. The decision was then taken for FAO, ISRIC and ISSS to develop a "World Reference Base for Soil Resources" (WRB) on this basis. By calling on soil specialists of all over the world a first draft was produced for the International Soil Congress in Acapulco in 1994. The structure of the Revised Legend was maintained and a third level added.

Between 1994 and 1998, numerous WRB meetings and field tours resulted in substantive changes in approach as well as in details of the soil classification.

• Thirty Reference Soil Groups were retained as the first level of the system representing the major soil groups with a unique combination of characteristics identifiable to agronomists and other scientists.
• The second level was characterized by the use of UNIQUELY DEFINED MODIFIERS that are used in sequence, resulting in unequivocal definitions of soil units and the possibility to characterize soil profiles fully.

In 1998, at the International Union of Soil Science, the World Reference Base for Soil Resources was presented in 3 volumes and adopted by the IUSS as the reference system of the Organization; its use was recommended as a standard in all peer reviewed scientific journals.

Since then WRB has been translated in eight languages (English, French, Spanish, Italian, German, Chinese, Lithuanian, and Japanese) and has been adopted as the Reference System by the European Soil Bureau. WRB appears to have become in a very short time a true reference system for the soil science community, but it has still a long way to go to be adopted by the other sciences.

Future developments will include the development of a Topsoil Characterization of which a first draft was already presented in the Acapulco Congress and is available from the Web. It will become available as a FAO World Soil Resources Report.

The three WRB books presently being sold commercially should become available next year in a "student" edition at very low cost accompanied with a CD-ROM with numerous soil profile pictures.

A meeting paying special attention to WRB and the classification of Anthrosols will be held in Budapest, 2-7 October 2001.

Regularly Updated News on WRB is available from the FAO Web site at:
www.fao.org/WAICENT/FAOINFO/AGRICULT/AGL/agll/wrb/wrbhome.htm

WRB Objectives

The primary objective of the World Reference Base for Soil Resources is to provide scientific depth and background to the 1988 FAO Revised Legend, incorporating the latest knowledge relating to global soil resources and their interrelationships. To include some of the most recent pedological studies and to expand use of the system from an agricultural base to a broader environmental one, it was recognized that a limited number of important changes to the 1988 Legend was necessary.

More specifically, the other objectives are:

• to develop an internationally acceptable system for delineating soil resources to which national classifications can be attached and related, using FAO's Revised Legend as a framework;
• to provide this framework with a sound scientific basis so that it can also serve as a communication tool for different applications in related fields such as agriculture, geology, hydrology and ecology. Hence WRB aims to promote international use of pedological data, not only by soil scientists but also by other users of soil and land information;
• to recognize within the framework important spatial relationships of soils and soil horizons as characterized by topo- and chronosequences; and
• to emphasize the morphological characterization of soils rather than to follow a purely laboratory-based analytical approach.

WRB is designed to provide an easy means of communication amongst scientists to identify, characterize and name major types of soils. It is not meant to replace national soil classification systems, but be a tool for better correlation among national systems. It aims to act as a common denominator through which national systems can be compared. WRB also serves as a common ground among people with an interest in land and natural resources.

Additional objectives are to improve:

• the evaluation of soil resources and the potential use of different types of soils;
• the monitoring of soils, particularly soil development which is dependent on the way soils are used by the human community;
• the validation of experimental methods of soil use for sustainable development, which maintain and, if possible, improve the soil's potential; and
• transfer of soil use and management technologies from one region to another.

WRB Principles

The general principles on which the WRB is based were laid down during the early Sofia meetings (Dudal, 1980, Schlichting, 1983), and further elaborated upon by the Working Groups entrusted with its development. These general principles can be summarized as follows:

• The classification of soils is based on soil properties defined in terms of diagnostic horizons and characteristics, which to the greatest extent possible should be measurable and observable in the field.
• The selection of diagnostic horizons and characteristics takes into account their relationship with soil forming processes. It is recognized that an understanding of soil forming processes contributes to a better characterization of soils but that they should not, as such, be used as differentiating criteria.
• To the extent possible at a high level of generalization, it is attempted to select diagnostic features which are of significance for management purposes.
• Climatic parameters are not applied in the classification of soils. It is fully realized that they should be used for interpretation purposes, in dynamic combination with soil properties, but they should not be part of soil definitions.
• WRB is meant to be a comprehensive classification system that may be a starting point for people to accommodate their own national soil classification system. It comprises two tiers of categorical detail:
1. the "Reference base" which is limited to the first level only, having 30 reference soil groups; and
2. the "WRB Classification System" consisting of combinations of a set of prefixes as unique quantifiers (or modifiers) added to the reference soil groups, allowing very precise characterization and classification of individual soil profiles.
• The reference soil units in WRB should be representative of major soil regions so as to provide a comprehensive overview of the world's soil cover.
• The reference base is not meant to substitute for national soil classification systems but rather to serve as a common denominator for communication at an international level. This implies that lower-level categories, possibly a third category of the WRB, could accommodate local diversity at country level. Concurrently, the lower levels could emphasize soil features which are important for land use and soil management.
• The Revised Legend of FAO/UNESCO Soil Map of the World has been used as a basis for the development of the WRB in order to take advantage of the international soil correlation work which has already been conducted through this project.
• Definitions and descriptions of soil units are to reflect variations in soil characteristics both vertically and laterally so as to account for spatial linkages within the landscape.
• The term `Reference Base' is connotative of the common denominator function that the WRB will assume. Its units should have sufficient width to stimulate harmonization and correlation of existing national systems;
• In addition to serving as a link between existing classification systems, the WRB may also serve as a consistent communication tool for compiling global soil databases and for the inventory and monitoring of the world's soil resources.
• The nomenclature used to distinguish soil groups will retain terms which have been traditionally used or which can easily be introduced in current language. These terms are precisely defined in order to avoid the confusion that occurs when names are used with different connotations.

Although the basic framework of the FAO Legend, with its two categoric levels and guidelines for developing classes at a third level, was adopted, it has been decided to merge the lower levels. Each reference soil group of WRB is provided with a listing of possible modifiers in a priority sequence, from which the user can construct the various lower-level units. The broad principles that govern the WRB class differentiation are:

• at the higher categoric level, classes are differentiated mainly according to the primary pedogenetic process that has produced the characteristic soil features; and
• at the lower categoric levels, classes are differentiated according to any predominant secondary soil forming process that has significantly affected the primary soil features. In certain cases, soil characteristics that have a significant effect on use may be taken into account.

An overview of the 30 Reference Soil Groups is given in Table 2.

The World Reference Base for Soil Resources (WRB) classification system

Elements of WRB

For describing and defining the Reference Soil Groups and soil units of the WRB, use is made of soil characteristics, properties and horizons, which in combination will define soils and their interrelationships.

Soil characteristics are single parameters that are observable or measurable in the field, in the laboratory, or can be analyzed by using microscopic techniques. They include characteristics such as colour, texture and structure of the soil, features of biological activity, arrangement of voids and pedogenetic concentrations (mottles, cutans, nodules, etc.) as well as analytical determinations (soil reaction, particle-size distribution, cation exchange capacity, exchangeable cations, amount and nature of soluble salts, etc.).

Soil properties are combinations of soil characteristics which are known to occur in soils and which are considered to be indicative of present or past soil-forming processes (e.g. vertic properties are a combination of clayey texture, smectitic mineralogy, gilgai, slickensides, hard consistence, shrinking when dry, sticky consistence and swelling when wet).

Typical newly identified diagnostic horizons in the World Reference Base for Soil Resources are, for example, the vertic horizon, characteristic for the Vertisols, the duric horizon, defining the Durisols, and the plinthic horizon in the Plinthosols.

Table 2

An overview of the WRB Reference Soil Groups, arranged by the environmental
conditions responsible for the soil-forming factors and properties

Soils conditioned by relief	Soils conditioned by parent material	Soils of (semi-) arid regions	Soils of the steppe regions	Soils of the temperate regions	Soils of the tropical regions
FLUVISOLS	ARENOSOLS	CALCISOLS	KASTANOZEMS	LUVISOLS	LIXISOLS
GLEYSOLS	ANDOSOLS	DURISOLS	CHERNOZEMS	PLANOSOLS	ACRISOLS
REGOSOLS	VERTISOLS	GYPSISOLS	PHAEOZEMS	ALBELUVISOLS	ALISOLS
LEPTOSOLS		SOLONETZ		PODZOLS	NITISOLS
			SOLONCHAKS		FERRALSOLS
					PLINTHOSOLS
Organic soils	Soils of limited age		Soils of the arctic regions
HISTOSOLS	CAMBISOLS		CRYOSOLS
	UMBRISOLS
Human-made soils
ANTHROSOLS

Reference Soil Groups are defined by a vertical combination of horizons within a defined depth, and by the lateral organization of these horizons, or by the lack thereof.

Soil horizons and properties are intended to reflect the expression of genetic processes that are widely recognized as occurring in soils. They can therefore be used to describe and define soil classes. They are considered to be "diagnostic" when they reach a minimum degree of expression, which is determined by visibility, prominence, measurability, importance and relevance for soil formation and soil use, and quantitative criteria. To be diagnostic, soil horizons also require a minimum thickness, which must be appraised in relation to bio-climatic factors (e.g. a spodic horizon in boreal regions is expected to be less thick than in the tropics).

In addition, use is made in WRB of diagnostic soil materials, which sometimes define Reference Soil Groups such as Fluvisols (fluvic soil materials), or are used as a modifier in the classification system (e.g. sulfidic soil material to identify Thionic lower level units).

The classification approach in the World Reference Base

The most important innovation in the World Reference Base is the building-block approach. Exclusive use is made of standard definitions for each subdivision name (modifier) of a Soil Reference Group. For instance, while in the Revised Legend the term "dystric" had several meanings (less than 75 percent base saturation in Vertisols, less than 50 percent in different depth-control sections in Cambisols and Planosols) , in WRB "dystric" has a unique meaning, which is : "having a base saturation (by 1M NH4OAc) of less than 50 percent in at least some part between 20 and 100 cm from the soil surface, or in a layer 5cm or more thick directly above a lithic contact in the Leptosols.

The building blocks are the uniquely defined modifiers as described above. There are 121 of these, which compares favourably with the 152 different soil units in the Revised Legend. An overview of all modifiers (qualifiers) is given in Table 3. An example of their definitions is given in Box 1. These building blocks are used to define the lower level subunit as illustrated in the following example:

Example

In Vertisols the following modifiers have been recognized, in order of priority:

1.	Thionic	intergrade with acid sulphate Gleysols and Fluvisols
2.	Salic	intergrade with the Solonchaks
3.	Sodic	intergrade with the Solonetz
4.	Gypsic	intergrade with the Gypsisols
5.	Calcic	intergrade with the Calcisols
6.	Alic	intergrade with the Alisols
7.	Gypsiric	containing gypsum
8.	Pellic	are dark coloured, often poorly drained
9.	Grumic	have a mulched surface horizon
10.	Mazic	have a very hard surface horizon; workability problems
11.	Chromic	are reddish coloured
12.	Mesotrophic	having less than 75 percent base saturation
13.	Hyposodic	having an ESP of 6 to 15
14.	Eutric	having base saturation over 50 pour cent
15.	Haplic	most common one

To classify a reddish coloured Vertisol with a calcic horizon one would follow the priority list and note that modifiers 5 and 11 apply. Therefore, the soil is classified as Chromi-Calcic Vertisol.

When more than two modifiers can be used, they can be added within brackets after the standard name. If, for instance, the Vertisol discussed also has a very hard surface horizon (modifier 10), the soil would be named Mazi-Calcic Vertisol (Chromic).

In addition to the unique modifiers, an opportunity is created to indicate depth (from shallow to deep: Epi, Endo, Bathi) and intensity (from weak to strong: Proto, Para, Hypo, Ortho and Hyper) of features, important for management interpretations. In the example above, one may indicate the occurrence of the calcic horizon within 50 cm from the surface by classifying the soil as Chromi-Epicalcic Vertisol. In cases of polysequential soil profiles, the modifiers Cumuli or Thapto can be used to indicate accumulation or burial.

For each reference soil group there is a defined list of which modifiers may be used and in which a priority order is given. This is illustrated in Table 4.

Table 3

General alphabetical list of modifiers

	Abruptic		Ferralic		Lixic		Rhodic
	Aceric		Ferric		Luvic		Rubic
	Acric		Fibric		Magnesic		Ruptic
	Acroxic		Folic		Mazic		Rustic
	Albic		Fluvic		Melanic		Salic
	Alcalic		Fragic		Mesotrophic		Sapric
	Alic		Fulvic		Mollic	100	Silic
	Alumic		Garbic	70	Natric		Siltic
	Andic	40	Gelic		Nitic		Skeletic
10	Anthraquic		Gelistagnic		Ochric		Sodic
	Anthric		Geric		Ombric		Spodic
	Anthropic		Gibbsic		Oxyaquic		Spolic
	Arenic		Glacic		Pachic		Stagnic
	Aric		Gleyic		Pellic		Sulphatic
	Aridic		Glossic		Petric		Takyric
	Arzic		Greyic		Petrocalcic		Tephric
	Calcaric		Grumic		Petroduric	110	Terric
	Calcic		Gypsic	80	Petrogypsic		Thionic
	Carbic	50	Gypsiric		Petroplinthic		Toxic
20	Carbonatic		Haplic		Petrosalic		Turbic
	Chernic		Histic		Placic		Umbric
	Chloridic		Hortic		Plaggic		Urbic
	Chromic		Humic		Planic		Vetic
	Cryic		Hydragric		Plinthic		Vermic
	Cutanic		Hydric		Posic		Vertic
	Densic		Hyperochric		Profondic		Vitric
	Duric		Hyperskeletic		Protic	120	Xanthic
	Dystric		Irragric	90	Reductic		Yermic
	Entic	60	Lamellic		Regic
30	Eutric		Leptic		Rendzic
	Eutrisilic		Lithic		Rheic
	Where relevant, the names can be defined further using prefixes, for example Epigleyi-, Protothioni-. The following prefixes can be used:
	Bathi		Epi		Orthi		Thapto
	Cumuli		Hyper		Para
	Endo		Hypo		Proto

Box 1. Examples of the unique modifier definitions: Carbi- having a cemented spodic horizon that does not contain enough amorphous iron to turn redder on ignition (in Podzols only). Carbonati- having a soil solution with pH > 8.5 (1:1 in water) and HCO3 > SO4 >> Cl (in Solonchaks only). Cherni- having a mollic horizon more than 30 cm thick, having a strong granular structure and a Munsell colour value and chroma, moist, of less than 2 throughout the upper 15 cm (or immediately below the plough layer). Bulk density is generally between 0.9 and 1.3. Chloridi- having a soil solution (1:1 in water) with Cl >> SO4 > HCO3 (in Solonchaks only). Chromi- having a B horizon that in the major part has a Munsell hue of 7.5 YR and a chroma, moist, of more than 4, or a hue redder than 7.5 YR. Cryi- having a cryic horizon within 100 cm of the soil surface.

Since the conception of the WRB classification system there have been calls for a more strict and reasoned priority listing which would be most helpful, mainly for educational purposes (H. Eswaran, personal communication). On the other hand there is a case to be made not to use the priority listing at all, because if WRB is used as a classification system the order of the building blocks does not matter while if used as a legend, the cartographer should be left the freedom to set his own priority rules.

Furthermore, use is made of standard depths: no other depth limits than 10, 20, 25, 30, 40, 50, 75, 100, 150 and 200 cm have been used. Although still less than satisfactory this is considered an improvement over the Revised Legend.

Table 4

Priority listing of lower level units of reference soil groups

ALBELUVISOLS	ALISOLS	VERTISOLS	ACRISOLS	LUVISOLS
Geli-	Verti-	Thionic-	Plinthi-	Lepti-
Gleyi-	Plinthi-	Salic-	Gleyi-	Verti-
Ali-	Gleyi-	Sodic-	Andi-	Gleyi-
Umbri-	Andi-	Gypsic-	Umbri-	Andi-
Fragi-	Niti-	Calcic-	Areni-	Calci-
Stagni-	Umbri-	Alic-	Stagni-	Areni-
Endoeutri-	Areni-	Gypsiric	Geri-	Stagni-
Abrupti-	Stagni-	Pellic	Albi-	Albi-
Ferri-	Albi-	Grumic-	Humi-	Hyposodi-
Hapli-	Humi-	Mazic-	Veti-	Profondi-
	Abrupti-	Chromic-	Abrupti-	Ferri-
	Lamelli-	Mesotrophic-	Lamelli-	Lamelli-
	Profondi-	Hyposodic-	Profondi-	Rhodi-
	Ferri-	Eutric-	Ferri-	Chromi-
	Hyperdystri-	Haplic-	Alumi-	Hyperochri-
	Rhodi-		Hyperdystri-	Dystri-
	Chromi-		Rhodi-	Hapli-
	Hapli-		Chromi-
			Hyperochri-
			Hapli-

The modifiers can be used to identify specific soil problems as illustrated in the following Table 5.

Implications of WRB for Soil Inventories with special reference to West Africa

Although there are examples where soil classifications have been used as pure legends, it is generally accepted that, although related, the two issues of classification and map legends should not be confused. Soil surveys, and particularly large scale regional soil surveys, perhaps at 1:200 000 or larger, require legend criteria which can not be covered to the last detail by even a very comprehensive soil classification. Some examples of these factors are: slope, drainage criteria, precise colour ranges for certain horizons, classes of stoniness, pH, precise texture ranges including sand classes, etc. These "soil series" criteria have by necessity to be developed by the local surveyors within a nationally agreed system. What should be aimed at is the possibility to incorporate, or at least attach, these "local soil series" within an international system at a higher level. This in turn requires that at the highest level of the national classification, a soil classification system is used that is compatible with WRB. The FAO revised legend, Soil Taxonomy (Soil Survey Staff, 1998 and 1999) and the Référentiel Pédologique Français (AFES, 1995) have relatively little problems with this approach, but more purely pedogenetic classification systems such as the Russian or old French CPCS system may pose more difficulties for soil correlation.

Table 5

WRB modifiers and related soil management problems

In the West African context it is also appropriate to draw attention to the fact that it is a challenge and an opportunity for a number of regional soil services within single countries in West Africa to transcend their differences and adopt WRB as a common national approach to soil classification in line with, and endorsed by, the IUSS.

There has been, for historical reasons, a wide diversity in national West African soil classification systems which results in problems for regional harmonization and transfer of technology.

Table 6

National soil maps of Africa south of the Sahara

Table 7

Operational plan for a World SOTER: 1995-2002

Many francophone countries have since colonial times adopted the French classification system (CPCS) which has now been abandoned by France. The new French classification system does NOT take into account tropical soils and the West African countries concerned are left with a system no longer supported by its originator.

Some countries have adopted Soil Taxonomy, a system that has undergone many changes in the last ten years and continues to rely on complex definitions, expensive laboratory methods while ignoring for all practical purposes human influences in soils and topsoil characteristics most important for agriculture.

Table 6 gives an overview of the most recent national soil maps produced by countries in Africa south of the Sahara.

West Africa is one of the few regions in the world sorely lacking an update of regional soil information, as illustrated in Table 7 It is therefore opportune that all countries in the region accelerate the way they harmonize their soil information by adopting a neutral and simplified correlation system which is able to support regional soil and land degradation inventories, such as the one proposed by WASOTER and WOCAT for the region.

The other issue involved in local soil surveys is the way soil information can be stored and represented. The recent advances in computer and database management and Geographical Information Systems, coupled with the enhanced resolution of satellite images, allows for a more natural ordering of soil and related information than was possible in a classical soil map. There is a general agreement that a land system approach is to be preferred, in which at the highest level large natural landscapes are recognized , at a lower level parent material and other terrain criteria can be brought in , while the soil information sensu stricto can be stored at a third level in which the soil associations are put in their natural context, each identified by a typifying pedon. It is obvious that not all soil units can be represented by profiles, and that each typifying pedon may stand for a number of similar soils in different units in a country or region. This kind of arrangement of soil information has been promoted at national scale for instance by Mitchell and Howard (1978) and CSIRO in Australia (Isbell, 1996), and by the landscape approach of the French school (Duchaufour, 1997). At regional level a similar technique is proposed by Finke et al. (1998) for the European Soils Bureau (ESB), while at the international level the SOTER (UNEP/ISSS/ISRIC/FAO,1995) approach as endorsed by the International Union of Soil Sciences (IUSS), UNEP and FAO is the norm. Although the details among these methods differ, the major difficulty in unifying them appears to be more political and commercial than scientific.

In addition copyright rules on soil data and maps are becoming an issue region. The harmonization of these different approaches and the free accessibility of soil and terrain data remains a priority . Progress and technology transfer is only possible when data are freely available to all, and this is the more true for developing countries.

The concern of many computer scientists to deal with measured observations, rather than with expert opinions, has led to a more systematic inclusion and use of soil profile data in thematic databases. In the best case these databases have been elaborated upon by pedotransfer functions which allow calculation of missing soil parameters with a certain degree of reliability. Although, ideally, each mapping unit is characterized by an actual soil profile (or an association of profiles ) occurring in it, for economic reason the number of soil profiles actually described and measured often remains very limited, and the reliability of extrapolating these results remains debatable. In this context the problem of the format and content of different soil databases can be raised. It is indeed remarkable how many database structures deal nearly exclusively with laboratory data, at the exclusion of morphological profile data. This may limit the possible applications of these inadequately conceived data structures. Therefore the use of broader and more open data storage systems as proposed in SOTER or in the FAO/ISRIC/CSIC multilingual soil profile database (FAO, 1995) is recommended.

Another warning in this context concerns faith often expressed by the general user of soil information in laboratory analyses. Various studies on within- and between- differences of laboratory results have shown a high variability in the results obtained and to base calculations on these may result in error propagation.

CONCLUSION

Soil science has suffered from the lack of a generally accepted system of soil classification that has resulted in a loss of credibility and in a limited interaction with other disciplines. It is hoped that the updating of the Soil Map of the World and the development of the WRB will remedy this situation and will progressively lead to a generally agreed identification of major soil groups and of the criteria to separate them at an international level. The great diversity of the soil cover at country scale justifies national systems at the lower levels. This two-pronged approach will facilitate the establishment of an international consensus.

It is a challenge for the WRB to remedy the widespread ignorance of the soil resources that still prevails. When dealing with the protection of animal species one finds it normal that a distinction be made between endangered rhinos, blue whales or pandas. It is realized that measures must be geared to the specific characteristics and problems of the animal concerned. With soils, even though there is an awareness that they may be different, they are often dealt with as if they were the same. As a result a great deal of misinformation is being generated - about environmental hazards related to agriculture, desertification, degradation - which often leads to decisions which are contrary to sound development. WRB should serve as a first entry into the knowledge of soil diversity and soil distribution, accessible to other disciplines and to a wider public (Dudal, 1996). It is imperative that soil science assert itself in the public debate and in the overall scientific community. A common WRB language could provide the means to do so particularly in Western Africa.

Even though it is generally recognized that soils are an important resource it appears that available soils information is under utilized or even ignored. It is a challenge for soil scientists to influence policy and opinion so that decisions with regard to the effective use of natural resources can be made more rationally. Meeting this challenge will require that soil science broadens its constituency beyond traditional agricultural partners, that it applies itself to develop solutions to problems of soil and land management, that it breaks through a reductionist approach, that it enhances communication with different users. If soil scientists fail to address the problems with the skills at their disposal, then others will, with less knowledge and authority.

It is also a challenge and an opportunity for a number of provincial autonomous soil services within single countries, particularly in Europe, to transcend their differences in soil classification and adopt WRB as a national reference system.

Fundamental differences between map legends and soil classification systems should not be confused, while trapfalls related to an unwarranted reliance on (pedo) statistics, laboratory results and computer modelling should not be pursued at the cost of less field observations.

Acknowledgement

This article is largely based on a joint effort with my colleagues in the Working Group on the World Reference Base, in particular Jozef Deckers (Leuven University), Otto Spaargaren (ISRIC) and Bob Ahrens (USDA). Parts of the text were published as a paper in Geoderma (2000).

References

AFES. 1995. Référentiel Pédologique. Techniques et pratiques. Baize D. et M.-C. Girard (coord. scient.). INRA Editions. Paris. 332 p.

Duchaufour, P. 1997. Abrégé de pédologie. Sol, végétation, environnement. 5e édition, Masson, Paris, 291 p.

Dudal, R. 1980. Towards an International Reference Base for Soil Classification, ISSS Bulletin Nr. 57, p. 19-20.

Dudal, R. 1990. Progress in IRB preparation: In: Soil Classification. Reports of the International Conference on Soil Classification, 12 - 16 September 1988, Alma-Ata, USSR. Rozanov B.G. (ed.). Centre for International Projects, USSR State Committee for Environmental Protection., Moscow. pp. 69-70.

Dudal, R. 1996. A World Reference Base for Soil Resources: Background, Principles and Challenges. Proceedings International WRB Workshop, Pretoria, 9 p.

FAO. 1974. The legend of the Soil Map of the World. Volume 1, UNESCO, Paris.

FAO/ISRIC/CSIC. 1995. Multilingual Soil Database. World Soil Resources Report No. 81, FAO, Rome.

FAO/ISRIC/ISSS. 1998. World Reference Base for Soil Resources. World Soil Resources Report No. 84. FAO, Rome.

FAO/UNESCO/ISRIC. 1988; 1990. Revised Legend of the Soil Map of the World. World Soil Resources Report No. 60. FAO, Rome.

Finke, P., Hartwich, R., Dudal, R., Ibanez, J., Jamagne, M., King, D., Montanarella, L. and Yassoglou, N. 1998. Georeferenced Soil Database for Europe. Manual of Procedures. Version 1.0. ESB, JRC Ispra, Italy.

Galligani, U. & Magaldi, D. 1997. The translation of the soil name from the EC Soil Map at 1:1 M legend (1985) to the FAO-Unesco Revised Legend . Draft prepared for ESB meeting, Bologna, 3-5 December 1997.

Isbell, R.E. 1996. The Australian Soil Classification. Australian Soil and Land Survey Handbook. CSIRO, Australia, 143 p.

ISSS Working Group RB. 1998a. World Reference Base for Soil Resources. Introduction (J.A Deckers, F.O Nachtergaele, O.C. Spaargaren, eds). First edition. ISSS/ISRIC/FAO. Acco Leuven.

ISSS Working Group RB. 1998b. World Reference Base for Soil Resources. Atlas (E.M. Bridges, N.H.Batjes, F.O Nachtergaele Eds). First edition. ISSS/ISRIC/FAO. Acco Leuven.

Mitchell, C.W. & Howard, J.A. 1978. Land System Classification. a case history : Jordan. AGLT Bulletin 2/78. FAO, Rome.

Nachtergaele, F.O, Spaargaren, O., Deckers J.A. & Ahrens, B. 2000. New developments in soil classification. World Reference Base for Soil Resources. Geoderma 96 (2000): 345-357

Schlichting, E. 1983. Third Meeting of an `International Reference Base for Soil Classification (IRB)', ISSS Bulletin, Nr. 64, p. 20.

Soil Survey Staff. 1998. Keys to Soil Taxonomy. USDA. National Resources Conservation Service. Eighth edition.

Soil Survey Staff. 1999. Soil Taxonomy. A Basic System of Soil Classification for Making and Interpreting Soil Surveys. Second Edition. Agricultural Handbook 436. USDA-NRCS. Washington, DC. 869 p.

UNEP/ISSS/ISRIC/FAO. 1995. Soil and Terrain Database (SOTER). World Soil Resources Report No. 74, rev. FAO, Rome.