	Soil Moisture Storage Capacity Derived from the 
	Soil Map of the World


1.	Introduction

	In the context of defining ranges and values for important properties and qualities derived from the Soil Map of the World (FAO-Unesco 1971-81) this text describes a methodology to determine values for the soil moisture storage capacity to a depth of 100cm or to an impermeable layer which ever is shallower (Smax).

	It is recognized that the prediction of Smax is only a part of the solution to the more general problem of soil water available to plants (SWAP).  Indeed Smax is determined in isolation of the prevailing climate and is to a large extent ignoring specific rooting habits of crops.  Furthermore some researchers have rejected the concept of the soil water availability as a static capacity in favour of a continuous variable in relation to atmospheric demand, root extension, water-table depth, rate of water movement to root surfaces, etc.

	Thomasson (1990) reviewed the methods to estimate Smax and SWAP and grouped them in four levels in order of broadly increasing complexity:

Level 1. Standardized soil moisture storage value, e.g. Smax = 40,60,80,100 or 125 mm (undefined profile depth)

	This approach is popular with agrometeorologists as it simplifies the manipulation of spatial atmospheric and crop parameters (Smith, 1976; Thompson et al. 1981; FAO, 1978; Frre and Popov, 1979; Gommes and See, 1992).  The soil information needs are negligible.

Level 2. Standardized soil depth (100cm or impermeable layer) and suction limits (water retained between 0.05 and 15 bar, or 0.1 and 15 bar)

	This method is popular with soil scientists (Hodgson, 1976; Mori, 1982; De Pauw, 1983, 1984) as it emphasizes differences between soils.  It involves rather naive assumptions for plant root behaviour but has the virtue of unambiguous definition as a soil property.  The soil information needed can be summarized as follows:

	-	Horizon thickness
	-	Particle size distribution or class
	-	Pedotransfer functions to calculate available water based on textural class, and/or organic matter, and/or bulk density and horizon class (Madsen and Holst, 1990; Hall et al, 1977; Joshua, 1991; Yerima, 1991)



Level 3.  Soil depth and/or suction limits for patterns of crop abstraction of water

	This method differs from the above in that water retention at 2 bar suction is used to separate easily available water (EAV) from water which is more tightly held at higher suctions and difficult to abstract, especially from deeper subsoils; and in the use of a conceptual model of effective rooting depth.  The approach is operational in the U.K. (e.g. Thomasson and Jones, 1989), Denmark (Madsen and Holst, 1990) and Germany (Renger and Strebel, 1980).  The concept was also applied outside Europe e.g. in Algeria (Nachtergaele, 1985), in Botswana (De Wit et al, 1991) and in Zimbabwe (Hall, 1992).

Level 4.  Dynamic simulation of water movement in the root zone

	This method requires detailed soil data normally not available from routine surveys (e.g. saturated hydraulic conductivity) and is difficult to extrapolate.  For more information see Vereecken et al. (1989), Wosten and Van Genuchten (1988).

	On the basis of the above review the level 1 and level 4 approaches can be rejected out of hand as the former does not take into account soil characteristics, while the latter requires a number of data not available from the 1:5 million scale soil map.  Thomasson (1990) notes that additional soil data requirement to pass from level 2 to level 3 appear slight compared with the gain in precision and relevance to plant behaviour.

2.	Relevant data availability in the Soil Map of the World

	The Soil Map of the World contains the following information when compared to the requirements cited above under level 2:

	a)	Topsoil texture: the texture of the top 30 cm of the dominant soil in the mapping unit is indicated by a symbol (1 = coarse; 2 = medium; 3 = fine) on the SMW.

	b)	Soil depth: an approximation of soil depth can be derived from the soil unit and phase names used on the SMW.  All soil groups are considered to have a root explorable depth of 100 cm.

	Lithosols: 10cm,	
	Rendzinas: 30 cm,	
	Rankers: 30cm

	When the following phases occur, soil depth is considered reduced as follows:


	Lithic phase:		30 cm (for 100% of the soil unit)
	Petrocalcic phase:	30 cm (for 50%), 75 cm (for 50%)
	Petrogypsic phase:	30 cm (for 50%), 75 cm (for 50%)
	Petroferric phase:	30 cm (for 50%), 75 cm (for 50%)
	Duripan phase:	30 cm (for 50%), 75 cm (for 50%)

	c)	Pedotransfer functions: no pedotransfer functions have been established for the textures and units used on the SMW and soil moisture retention. Many of such statistical relationships have been published related to e.g. clay content or to specific textural classes, as defined in the textural triangle. A complication in this respect is the fact that soil physicists disagree on the lower suction limit to use to determine moisture available to crops (0.05, 0.10 bar, 0.33 bar), while the higher suction limit is generally agreed upon and set at 15 bar.



	d)	Other factors which influence the soil moisture storage capacity are: gravel content, textural changes with depth and some particular soils strongly influenced by parent material.



3.	Parameters used in the estimation of the soil moisture storage capacity

3.1	Texture

	Table 1 applies the results of research in Europe (Thomasson and Jones, 1989) and in Zimbabwe (Hall, 1992) with the textural groupings used in the Soil Map of the World.

-	The textural classes represented on the Soil Map of the World seem to group rather well the available water when compared with the more detailed textural classes.  Two large variations are noted, one the sandy loam class partly grouped under "1", the other is the silt loam class grouped under "2" on SMW which have both a significantly higher value than other members of the group.

-	A more detailed analysis, using textural subclasses based on the composition of the sand grades (Hall, 1992) showed that for all textures coarser than silt loam, the sand composition plays an important role even overruling differences established between traditional textural classes.  This implies that even if the textural class on the SMW would have been based on the 11 classes of the textural triangle, knowledge about the composition of the sand fraction would be required to reach very accurate results.

	Results in tropical soils are generally slightly lower than for comparable classes under temperate conditions.  This confirms other findings (e.g., Yerima, 1991) who proposed a value of 10% less.  This would be related to the higher sesquioxide content and lower CEC of the tropical soils.

3.2	Texture changes with depth

	As the SMW only gives the textural group for the top 30cm a further complication arises as no straightforward information is available for the subsoil horizons.  The Legend gives two indications which may help to interpret changes in texture between top and subsoils.  One is clay illuviation as indicated by a clay increase of:

	a)	>3% if the topsoil has less than 15% clay
	b)	more than 1.2 times the clay content of the topsoil for soils with a topsoil clay content between 15 and 40
	c)	more than 8% if the topsoil contains more than 40% clay

	Soils that are characterized by the above clay increases belong to the Solonetz, Nitosols, Podzoluvisols, Luvisols and Acrisols and to luvic units of Chernozems, Phaeozems, Xerosols, Yermosols and Kastanozems.

	In these soils it is suggested to consider the subsoil textural class as one class heavier than the indication of the topsoil textural class. The same applies to Ferralsols which have by definition a Ferralic B horizon which contains at least 15% clay.


Table 1.	Total and easily available water by textural class

Texture	Texture class		Total available water [1]
group	(Int.texture		(Easily available water [2] SMW		triangle)			A-horizon		B and C horizons
									with medium packing
									density [3]
-----------------------------------------------------------------
1		Sand			12 (8)		6 (5)
		Loamy sand		12 (8)		8 (6)
		Sandy loam [4]		17 (13)		15 (11)
		
		Average			13.2 (9.5)	9.5 (7)
-----------------------------------------------------------------
2		Sandy clay loam	        17 (11)		15 (10)
		Clay loam A[4]		18 (11)		16 (10)
		Silty cl.loam A[4]	19 (10)		17 (10)
		Loam			20 (12)		19 (12)
		Silt loam		23 (15)		22 (14)
		Sandy loam B[4]         17 (13)	        15 (11)

		Average			19 (12)		17.5 (11)
-----------------------------------------------------------------
3		Clay			17 (10)		16 (8)
		Sandy clay		17 (11)		15 (10)
		Silt clay		17 (10)		15 (8)
		Clay loam B[4]		18 (11)		16 (10)

		Average			17.5 (10.4)	16.0 (9.2)



Notes:
[1]	Water held between 0.05 and 15 bar suctions (% vol): 0.1 to 15 bar in loamy sands and sands.
[2]	Water held between 0.05 and 2 bar suction (% vol): 0.1 and 2 bar in loamy sands and sands.
[3]	Packing density between 1.4 and 1.75. Packing density =  Bulk density  + 0.009* Clay % (Renger, 1971)
[4]	Sandy loam a = SL with less than 18% clay, more than 65% sand
	Sandy loam b = SL with more than 18% clay, less than 65% sand
	Clay loam a = clay loam with less than 35% clay
	Clay loam b = clay loam with more than 35% clay
	Silty clay loam a = SiCl with less than 35% clay
	Silty clay loam b = SiCl with more than 35% clay



	The second indicator is the occurrence of an abrupt textural change in Planosols.  This means a significant increase of at least 20% clay or doubling the clay percentage.  Hence it is suggested that all Planosols have a fine textured (3) subsoil.

3.3	Total depth to be considered

	Thomasson (1990) gives examples of Soil Water Available to Plant (SWAP) as a function of soil depth.  For a general evaluation, as attempted here, 100 cm or the occurrence of an impenetrable layer or rock would seem to be a reasonable maximum rooting depth and is proposed here on this basis.  It is recognized that there are exceptions to this arbitrary rule which should be taken into account when specific crops are evaluated.  

3.4	Stoniness and coarse fragments

	Gravel, stones, boulders and rock fragments when present in the profile reduce considerably the capacity of a soil to store moisture.  Landon (1984) cited in Yerima (1991) indicated a reduction of Smax by about half when the gravel content reached 50%.

	The Legend of the Soil Map of the World uses this criterion when defining the stony phase reflecting the presence of coarse fragments in the surface layers or at the surface to an extent that it makes the use of mechanized equipment impracticable.  Soil groups and units such as Vitric Andosols, Rendzinas and Lithosols generally contain a significant amount of gravel, while in other soils such as Ferralsols (stone line), Regosols, Acrisols, and soils with a petrocalcic or petrogypsic phase gravel occurrence is more common but largely dependent on local conditions.  In the present general approach it is suggested to reduce Smax by 50% when a stony or petric phase is present.  

3.5	Soil groups strongly influenced by parent material

3.5.1	Andosols

	Andosols are soils developed in volcanic ash, tuff, pumice and other volcanic ejecta of various composition.  Due to the presence of an accumulation of amorphous substances such as allophane and imogolite they have particular chemical and physical properties quite distinct from other soils.  These soils often show considerably higher values for soil moisture retention, at field capacity, as well as at wilting point, (Colmet-Daage et al., 1975) but precise limits and ranges are difficult to establish, partly because measurements are strongly influenced by the method applied, in particular the initial moisture content at which the retention is measured and whether the soil has been allowed to dry out or not, and partly because of the inherent variation which depends on the stage of evolution of the soil and its amorphous substance content.

	For the purpose of this study an average soil moisture storage capacity of 200 mm/m has been retained for Andosols except for Vitric Andosols which are considered as similar to Cambisols for Smax evaluation.

3.5.2	Vertisols

	Vertisols are fine textured soils dominated by smectite clay minerals which swell when wet and the soils are very hard and crack when dry.  As with Andosols the soil moisture content at field capacity and at wilting point are both very high which limits partly their storage capacity.  Joshua (1991) estimated the soil moisture storage capacity in these soils at about 135 mm/m.  This figure is provisionally retained here. 

3.6	Soils strongly influenced by groundwater or which are seasonally or permanently flooded

	Water availability to plants grown on Histosols, Gleysols and Fluvisols is mainly a function of groundwater or surface water levels and flooding.  Although an Smax value can be deduced for these soils, this is largely irrelevant for practical purposes.  Hence these soil groups are considered here as "wetlands" and no Smax is determined for them.  
  
4.	Methodology and algorithm development

	The methodology follows the rules explained in section 3.

	As a first step the soil group and soil units are grouped in eight sets which reflect fundamental differences in soil depth, textural changes with depth, influence of parent material or seasonal flooding conditions.  These sets are:

1.	Histosols, Fluvisols and Gleysols: which are considered as wetlands.
2.	Andosols: which due to parent material influence have a very high Smax (except Vitric Andosols).
3.	Vertisols: specific characteristics set this group of soils apart.
4.	Lithosols and miscellaneous land units: Lithosols and "rock" units are characterized by a very limited soil depth.  Water bodies, salt flats and shifting dunes have been grouped with Lithosols on the basis that although they might have different and higher Smaxs they show other characteristics that make them, as Lithosols, unsuitable for all agricultural activity.
5.	Rendzinas and Rankers: both these soil groups are shallow by definition.
6.	Soil groups and soil units with no implied clay increase with depth: this group combines soils in which the topsoil texture is considered representative of the whole profile.  These are: Solonchaks, Regosols, Podzols, Cambisols, Arenosols, Vitric Andosols, Greyzems and the non-luvic soil units of the Xerosols, Yermosols, Kastanozems, Chernozems and Phaeozems.
7.	Soil groups and soil units with an implied clay increase with depth: this set combines soils in which subsoil texture is finer than the topsoil texture.  These are Solonetz, Podzoluvisols, Nitisols, Acrisols, Ferralsols and luvic units of the Xerosols, Yermosols, Kastanozems, Chernozems and Phaeozems.
8.	Planosols:  This soil group is considered to have a fine textured subsoil regardless of the topsoil texture which separates these soils from those discussed under 6 and 7.

	For sets 6 to 8 a distinction is made according to the occurrence of a depth phase which implies the presence of a petrocalcic, petrogypsic, petroferric, or duripan layer within 100cm of the soil surface.  If such a depth phase occurs half of the extent of the soil is considered to be 75 cm deep, the other half to be only 30 cm deep.  If a lithic phase occurs the whole soil extent is considered to be only 30 cm deep.  If no depth phase occurs the soil is considered to be at least 100 cm deep.

	The third step takes into account the topsoil texture indication and the implied or assumed subsoil texture given in step 1 and uses the depth ranges determined in step 2.  For tropical soils (Ferralsols, Acrisols, Nitisols, Ferralic Cambisols and Ferric Luvisols) Smax is decreased by 10% as compared to similar textural classes for other soils.

	In the final step a correction is made for stoniness, and soil units with a stony or petric phase reduced by 50%.

	The procedure is best illustrated with an example:

	For instance, for a chromic Cambisol with topsoil texture indication "2" (medium) and having a petroferric phase, the following calculation would be made:

a)	A chromic Cambisol (Bc) belongs to set 6 and hence the soil is considered to have a medium texture throughout (in topsoil as well as in subsoil).

b)	The petroferric phase is a depth phase and hence half of the area-extent of this soil unit is considered to be 30 cm deep, the other half is considered to be 75 cm deep.

c)	According to Table 1 medium textures have an average soil moisture storage capacity of 19% in the topsoil and 17.5% in the subsoil.
	
	c.1	For the shallow part (50%) of this unit this leads to an Smax = 19 x 300/100 = 57 mm/m.

	c.2	For the deeper part (50%) of this unit Smax = 19 x 300/100 + 17.5 x (750-		300)/100 = 136 mm/m.

d)	As no stony nor petric phase is present no further corrections are needed.

	Given the high variability implied in the texture petrotransfer functions (Table 1) and the numerous assumptions made due to the very small scale of the soil map used, it does not make sense to use these apparently very precise results (57 and 136 mm).  Therefore a number of Smax classes has been defined in order to avoid unwarranted overprecise conclusions.  Please note that as a parallel procedure the easily available water (EAV stored in the upper 100 cm of the soil can also be calculated by using the average EAV values for each textures class (values between brackets in Table 1).  Proposed EAV classes are also given in Table 2.

Table 2.	Smax classes proposed for the interpretation of the Soil Map of the World

Symbol	EAV range		Smax range	Class name
----------------------------------------------------------------
1		Not applicable	Not applicable	W (wetlands)
2		> 120 mm		> 200 mm	A
3		100-120 mm	        150-200 mm	B
4		60-100 mm		100-150 mm	C
5		40-60 mm		60-100 mm	D
6		20-40 mm		20-60 mm	E
7		< 20 mm		        < 20 mm		F
----------------------------------------------------------------

	The algorithms as discussed in this section have been programmed in QBASIC and generated a datafile named SMAX3.BIN (in directory c:\faosoil\data and SMAXOUT.IMG in c:\faosoil\igbp_dis\ of Smax values for the whole world using a specific two-figure code which gives information on the magnitude and distribution of Smax in each cell of the raster as follows:

	The first code number gives the dominant Smax class (60% of the cell), while the second figure indicates the associated (+ 40%) Smax class. When the second number is 0, this means that the whole cell is made up by the Smax class indicated by the first number. For instance, a code of 23 means 60% of the cell has an Smax of less more than 200 mm and 40% of the cell has an Smax between 150 and 200 mm. A code of 50 means that the whole cell is made up by Smax class 5 (60-100 mm/m).

	Users who wish to access the complete distribution may use the datafile called SMAX3.BIN which contains a this information. Information in this format can also be generated by country using the Soil moisture storage option in the country-specific menu under the COUNTRY directory.

	To generate graphics and visual outputs, the simplest solution is to introduce the SMAX3.BIN file in a GIS. However, simplified pictures can also be obtained for the whole world, each region and user-determined areas by running the program c:\faosoil\images.exe and select the soil moisture storage option.


Bibliography

Colmet-Daage,   F.,   Cucalon,   F.,  Delaune,  M.,  Gautheyrou  J.  et  M.,  Moreau,  B. (1975)
Caractristiques de quelques sols d'quateur drivs de cendres volcaniques. 1re Partie: Essai de caractrisation des sols des rgions tropicales humides.

De Pauw, E.  (1983) 
Agricultural  potential  and constraints in relation to growing periods and their availability in Tanzania. Pedologie, 1983, pp. 147-170. RUG, Ghent, Belgium.

De Pauw, 1984
Soils, physiography and agroecological zones of Tanzania.
Crop monitoring and Early Warning Systems project, Dar es Salaam, GCP/URT/047/NET, FAO, 158 pp.

De Wit, P.V., Radcliffe, D. and Tersteeg, J.  (1991) 
CYSLAMB (in preparation) 
AG/DP/BOT/92/001. Field Document, Gaborone, Botswana.

FAO  (1971-1981) 
FAO-Unesco Soil Map of the World. Vols. I-X. Paris, Unesco

FAO (1978) 
Report on Agro-Ecological Zones Project. Vol. 1. Methodology for Africa. World Soil Resources Report No. 48. FAO, Rome.

Frre and Popov (1979)

Gommes and See (1992)

Hall, D.G.M., Reeve, M.J., Thomasson, A.J. and Wright, V.F. (1977) 
Water retention, porosity and density of field soils.  
Soil Surv. Tech. Monogr. 9, Harpenden, UK.

Hall, D.G.M. (1992) 
Soil Water reserves available for plants in Zimbabwe (in preparation).

Hodgson, J.M. Editor. (1976) 
Soil Survey Field Handbook. Soil Survey Tech. Mon. No.5, Harpenden, UK.

Joshua, W. (1991) 
Physical properties of the soils of Botswana.  Field Document 33. AG/BOT/85/011, Gaborone, Botswana.

Landon, J.R. (1984) 
Booker Tropical Soil Manual.  Longman Inc. New York, USA.

Madsen, H.B.,  and  Holst, A.H. (1990)
Mapping of irrigation need based on computerized soil and
climatic data. Agricultural Water Management 17. Elsevier, Amsterdam.

Mori, A. (1982) 
Notices explicatives de la carte des sols de France affects par l'hydromorphie et de la carte des rserves en eau utile des sols de France (chelle: 1:2.000.000), SESCPF, Orlans, INRA, Paris, 54p.

Nachtergaele, F.O.F. (1985) 
Land evaluation in northern Algeria based on small scale maps. Ph.D. Thesis. RUG. Ghent, Belgium.

Renger,   M.  and    Strebel,   O. (1980)  
Besegnungsbedarf  Landuvirtschaftlicher  Kulturen.   In: Abhngigkeit vom Boden. Wasser und Boden 32. p. 572-576.

Smith, L.P.  The agricultural climate of England and Wales.  MAFF Tech. Bull. 35. HMSO, London, UK. 

Thomasson, A.J. and Jones, R.J.A. (1989) 
Land evaluation at regional scale.  Land qualities in space and time.  Proceedings of a symposium by the ISSS. Editors, J. Bouma and A.K. Bregt. Pudoc, Wageningen, Netherlands.

Thomasson,  A.J. (1989) 
The principles for Assessment of Soil and Water Reserves available for Plants (SWAP).  Discussion paper prepared for EEC. Soil and GIS Group. ISPRA.

Thomasson, A.J. (1990) 
Soil Climate Classification. Draft Working Paper. FAO, Rome.

Thompson,  N.,  Barrie,  I.A.  and  Ayles,  M. (1981)   
The  Meteorological  Office  rainfall  and evaporation calculation system.  MORECS.  Hydrological Memorandum No. 45. Meteorological Office, Bracknell, UK.

Vereecken, H., Maes, J., van Orshoven, J. and Feyen, J. (1989)
Deriving pedotransfer functions of soil hydraulic properties  Land Qualities in Space and Time.  Proceedings of a symposium by the ISSS.  Editors, J. Bouma and A.K. Bregt, Pudoc, Wageningen, Netherlands.

Worsten, J.H.M. and van Genuchten, M.T.L. (1988) 
Using texture and other soil properties to predict unsaturated soil hydraulic functions.  Soil Sci. Soc. Am. J. 52.

Yerima, B. (1991) 
Manuel sur les interprtations pdologiques des analyses des sols. AG/DP/GAF/82/007, Rapport technique no. 13. FAO, Bangui, Republique Centrafricaine.




 

 


