Soil mapping for land-use planning in a karst area of N Thailand with due consideration of local knowledge

2006 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim 1436-8730/06/0306-444

444 DOI: 10.1002/jpln.200521902 J. Plant Nutr. Soil Sci. 2006, 169, 444–452

Soil mapping for land-use planning in a karst area of N Thailand with dueconsideration of local knowledgeUlrich Schuler1*, Chalathon Choocharoen2, Peter Elstner2, Andreas Neef2, Karl Stahr1, Mehdi Zarei1, andLudger Herrmann1

1 University of Hohenheim, Institute of Soil Science and Land Evaluation (310), D-70593 Stuttgart, Germany2 The Uplands Program, Hohenheim Office, Faculty of Agriculture, Chiang Mai University, Chiang Mai 50200, Thailand

Accepted January 1, 2006 PNSS P190/2B

SummaryFor the development of sustainable land-management sys-tems in the highlands of N Thailand, detailed knowledgeabout soil distribution and soil properties is a prerequisite. Yetto date, there are hardly any detailed soil maps available on awatershed scale.In this study, soil maps on watershed level were evaluatedwith regard to their suitability for agricultural land-use plan-ning. In addition to common scientific methods (as underlyingthe WRB classification), participatory methods were used toexploit local knowledge about soils and to document it in a“Local Soil Map”.Where the WRB classification identified eight soil units, thefarmers distinguished only five on the basis of soil color and“hardness”. The “Local Soil Map” shows little resemblance

with the detailed, patchy pattern of the WRB-based soil map.On the contrary, the “Local Soil Map” is fairly similar to thepetrographic map suggesting that soil color is directly relatedto parent material.The farmers’ perception about soil fertility and soil suitabilityfor cropping could be confirmed by analytical data. We con-clude that integrating local soil knowledge, petrographic infor-mation, and knowledge of local cropping practices allows fora rapid compilation of information for land-evaluation pur-poses at watershed level. It is the most efficient way to builda base for regional land-use planning.

Key words: local soil knowledge / petrography / participatory andscientific soil mapping / SE Asia

1 Introduction

The pressure on land and water resources in the highlands ofN Thailand is increasing due to rapid population growth andcommercialization of agriculture. Sound information on envir-onmental qualities including soil characteristics is necessaryfor the development of new management techniques in theseareas. Many numeric models used in scenario development,e.g., FALLOW (Van Noordwijk, 2002), CATCHSCAPE (Becuet al., 2003), CATCHCROP (Perez et al., 2001) need detailedsoil data which are scarce in N Thailand. The sloping lands,in particular, are represented as one undifferentiated “slopecomplex” on the Soil Map of Thailand (Vijarnsorn andEswaran, 2002).

Given this lack of soil information, the question arises how tocollect the necessary soil data in an efficient way. Conven-tional soil mapping approaches rely on grid or transect obser-vations augmented by reference-profile descriptions and lab-oratory analyses (Schlichting et al., 1995). The scale of thesoil map dictates the number of field observations that arerequired. In practice, soil mapping is arduous and time con-suming, and the soil information collected is often precise butnot always relevant (Barrios and Trejo, 2003), particularly inthe developing world. Other mapping efforts in the regionproved also of limited use; e.g., geological maps are insuffi-ciently detailed to provide information on potential soil varia-bility (Schuler et al., 2004).

The challenge is therefore to collect coherent, useful, andsufficiently detailed soil information at watershed level in areasonable time and to present this information in a formatthat the local population can understand and use.

The use of local knowledge in soil appraisal has receivedincreasing attention in the past decades (e.g., Reij et al.,1996) because there is overwhelming evidence that farmerscan provide valuable information on changes in land coverand land-management practices. Several authors suggestthat local knowledge should be blended with conventionallycollected soil information (WinklerPrins, 1999; Ali, 2003;Erkossa et al., 2004).

Many studies use participatory approaches to evaluate localsoil maps (e.g., Oudwater and Martin, 2003) but only fewcombined so-called scientific analysis and local knowledge(e.g., Keiichi and Toshiyuki, 2002). The present study com-pares local soil classification and local soil maps with conven-tionally collected soil information. The results suggest thatincorporating local soil information in soil maps is cost effec-tive and produces maps that are more relevant and easilyunderstood by local farmers (Krasilnikov and Tabor, 2003),which is conducive to (more) sustainable land management.

2 Materials and methods

2.1 Research site

The research site is situated in a Permian limestone area(German Geological Mission, 1979) in Pang Ma Pha District,Mae Hong Son Province, in mountainous N Thailand (Fig. 1).

* Correspondence: Dr. U. Schuler;e-mail: [email protected]

This area includes the village of Bor Krai and parts of theneighboring village of Jabo.

The Bor Krai research site lies between 540 m asl and1020 m asl. The landscape is characterized by steep, cone-shaped limestone mountains separated by karstic depres-sions. Within the depressions occur some smoother hills thatconsist largely of sand-, silt-, and claystone. The vegetationis mainly mixed deciduous forest with small patches of dryevergreen forest. In 2004, the mean temperature was foundto be 19.9°C, and precipitation amounted to 935 mm. Thelocal monsoon climate has three distinct seasons: a rainyseason from May to October, a cool dry season from Novem-ber to February, and a hot and dry season from March toMay.

The local people belong to the Black Lahu ethnic minoritygroup; they moved from Myanmar to Pang Ma Pha Districtsome 40 years ago. Most villagers are farmers who generateincome from livestock (pigs, cattle, and buffalos) and cropproduction (upland rice, corn, beans, cucumbers).

A first soil survey in the highlands of N Thailand was con-ducted by Hendricks (1981) who reports that soil variability inthe uplands depends largely on soil parent material, climate,and vegetation. Vlassak et al. (1992) classified most soils ongranite and soils on crest positions (with various parent mate-rials) as deeply weathered Acrisols. Shallow and stony Rego-sols are dominant in slope positions on sedimentary andmetamorphic rocks. Soils of valley floors are mostly shallowRegosols or water-saturated Gleysols/Fluvisols. The physicaland chemical properties of soils of the Jabo research station,located in the limestone part of the Bor Krai area, were deter-mined by the “Soil Fertility Conservation project—SFC” (Vlas-sak et al., 1992). The soil of Jabo research station was classi-fied as a Humic Ferralsol. In a transect study close to the sta-tion, Sereke (2002) classified three soil profiles as RhodicNitisols. However, additional chemical and mineralogical ana-lyses suggest that these soils are Acri-Gibbsic Ferralsols andHumi-Umbric Acrisols. In his ethno-pedological soil survey,Tinoco-Ordónez (2003) described eight soil profiles as Luvi-sols and Cambisols.

2.2 Geographical information

The topographic map (Royal Thai Survey Department, 1976)at scale of 1:50,000 and contour lines interval of 20 m resolu-tion and the geological map at scale of 1:250,000 (GermanGeological Mission,1979) were digitized. A GPS was used toobtain coordinates of the research points.

2.3 Soil and petrography mapping

In a first approach, soils were studied along transect linesand at 118 random points. Rugged, steep, and partly inacces-sible terrain in the limestone area was surveyed along trailswith the help of a local guide. Where there were insufficienttrails to obtain adequate information, additional points weredescribed and sampled. Reference soil profiles were de-scribed for the dominant soil units and analyzed in the labora-tory. In total, 341 augerholes and 22 soil-profile descriptionsaccording to WRB and SOTER guidelines (Van Engelen andWen, 1995; FAO, 1998, 2001) were collected in an area ofapproximately 8.5 km2 with a mapping scale of 1:10,000.Most rock types could be determined directly in the field. Therocks were classified according to Dunham (1962), Matthes(1996), and Tucker (1985). Selected rock samples weretaken for X-ray diffraction analysis. After the field survey, thecollected information was used to compile a petrographicmap.

2.4 Local soil classification and mapping

The evaluation of farmers’ soil knowledge was based on a“Participatory Rural Appraisal” (PRA) (Chambers, 1992). Itused semistructured interviews, field- and key-informantinterviews, participatory mapping, and group discussions.The survey was conducted by a multidisciplinary group ofscientists representing soil science, agricultural extension,farming systems, and rural sociology during the dry season(October to May) of 2004/2005.

The first step was to identify farmers with long-term practicalexperience. Next, key informants were asked which soil typesthey distinguish and by which differentiating criteria. Soil clas-

2006 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.plant-soil.com

Figure 1: Location of the Bor Krai research site in N Thailand.

J. Plant Nutr. Soil Sci. 2006, 169, 444–452 Soil mapping with due consideration of local knowledge 445

sification was further refined during field walks with farmers.Soils on the sites chosen by the farmers were describedaccording to local perceptions and the WRB and SOTER sys-tem (Van Engelen and Wen, 1995; FAO, 1998, 2001), andsamples were taken.

The key informants ranked different soil properties and devel-oped a local soil classification and a local soil map. The topicsfor ranking, e.g., suitability for crop production, fertility, infiltra-tion rate, available water capacity, erodibility, and stickinesswere suggested by us. In a final step, the farmers showed thedistribution of the local soil units on a print-out of the topo-graphic map.

2.5 Soil analysis

Laboratory analyses were done at the University of Hohen-heim, Germany, and Chiang Mai University, Thailand. Thesoils were analyzed according to Klute (1986), VDLUFA(1991), Schlichting et al. (1995), Blume et al. (2000), andHerrmann (2005). The analyses comprised texture, total car-bonate content, cation-exchange capacity (CEC), effectivecation-exchange capacity (CECe), exchangeable cations,Corg and Ntot contents, X-ray diffraction, pH in water and KCl,DCB- and oxalate-extractable Fe contents, water-dispersibleclay content, and levels of “available” P and K.

3 Results and discussion

3.1 Petrography

The petrographic mapping revealed that the area consists for60% of limestone and 39% of clay-, silt-, and sandstone. Theremaining 1% are alluvial deposits, dolomite, freshwater lime-stone, Fe-ore, and andesite (see Fig. 3A). The limestone wascrystalline and mostly massive, only at some places a layer-ing was visible. Occasionally, dolomite intercalation within thelimestone formation was observed. Freshwater limestonewas found around karst springs. The claystone had a yellowweathering color; dark-grey colored claystone was found onlyat a few sites along creeks. In places, layers of silt- and sand-stone are intercalated in the claystone. Andesite was found inthe SE part of the research area.

3.2 Soil mapping according to WRB

Soil mapping according to WRB rules revealed that the areaconsists of approximately 69% Luvisols, 21% Acrisols, 9%Cambisols, and 1% Leptosols. Umbrisols, Ferralsols, Fluvi-sols, and Chernozems each occurred on <0.1% of the area(Fig. 3C). Luvisols are found at stable positions in the land-scape. Their extension on clay-, silt-, and sandstone wasfound to be independent of elevation. Luvisols on limestoneare exclusive to sites below 800 m asl. Acrisols are foundonly on stable limestone above 800 m asl. Ferralsols occur atvery stable positions in the limestone area. Cambisols aretypical of unstable landscape positions, either affected byerosion or by deposition of eroded soil material. An Umbrisolwas identified in a karst depression; it contained charcoaldown to a depth of at least 200 cm which accords with the

colluvial character of the soil parent material. Chernozemsare restricted to the surroundings of karst springs. A Gleysolwas found in a karst depression with a sealed underground.Fluvisols are restricted to valleys and stream beds. Leptosolsoccur in the vicinity of limestone and Fe-ore outcrops. Theuse of CEC as a differentiating feature between Luvisols andAcrisols on the one hand and Cambisols and Ferralsols onthe other hand proved to be a source of insecurity. The WRBclassification system stipulates that Luvisols are character-ized by an Argic horizon with a CEC (1 M NH4OAc bufferedto pH 7.0) ≥ 24 cmol(+) (kg clay)–1 and without any Alic prop-erties. In contrast, Acrisols are characterized by an Argic hori-zon with <24 cmol(+)(kg clay)–1 in some parts. The differen-tiation of the Luvisols and Acrisol is mainly based on the anal-ysis of nine soil profiles, therefore their real distributionremains somewhat unclear. Only one Ferralsol was found,but it can be expected that more Ferralsols exist than the cur-rent soil map suggests. The WRB map was compiled basedon 22 analyzed soil profiles, 341 augerings, information fromthe topographic map, and observations during the field sur-vey. The WRB soil map shows two large homogeneous unitsof Acrisols and Luvisols which are riddled with inclusions ofother soils and limestone outcrops. Especially the Cambisolspots trace back to the land use in this area, responsible fordislocation of soil material. It is expected that the areas ofCambisols, Leptosols, and limestone outcrops will increase inthe future at the expense of Luvisols and Acrisols.

3.3 Local soil map

Farmers in the area differentiate soils according to observa-ble morphological parameters, mainly topsoil color. At first,the farmers distinguished only between two soil types: “BlackSoil” and “Red Soil”. A minority also mentioned texture andwater drainage. However, farmers were well aware of morecomplex soil parameters, like fertility status, suitability for cer-tain crops, and workability.

During the field trips, two additional soil colors were added tothe classification according to color: yellow and orange. Thefarmers identified five to seven soil types, using texture as anadditional criterion to differentiate within the color classes.Local soil classification is not always consistent which wasalso reported by others (e.g., Oudwater and Martin, 2003).This explains the disparity between our results and findingsby Tinoco-Ordónez (2003) who—in the same village—identi-fied five major classes according to color (red, black, yellow,yellow mottled, and grey) and six classes according to tex-ture.

In the second group discussion, the farmers were asked topresent a common soil classification. This was established byasking farmers to sort soil samples according to soil proper-ties. During this ranking process, it became clear that farmersrelate soil properties primarily to soil color. Hence, wefocused on color in the common classification. Farmers dis-tinguish four main soil types, namely “Black”, “Red”,“Orange”, and “Yellow Soils”, which together make up >90%of the preliminary local soil map. Mixed local soil unitsoccurred as well. The use of color as a first criterion is verycommon in local soil classifications (cf., Ettema, 1994; Tala-


446 Schuler, Choocharoen, Elstner, Neef, Stahr, Zarei, Herrmann J. Plant Nutr. Soil Sci. 2006, 169, 444–452

war and Rhoades, 1998). The comprehensive study on eth-nopedology by Barrera-Bassols and Zinck (2003) concludedthat all local soil classifications reviewed used color as aparameter because it is the most obvious and easily distin-guishable property of soils. Further interviews revealed thenecessity to divide the “Red Soil” group into “Hard Red Soils”and “Soft Red Soils”. The local soil map was updated accord-ingly with the help of experienced farmers whereby the preli-minary local soil map and the topographic map were the maincommunication tools. Once this map was made, farmerswere asked for more “Local Soil Unit” properties. In our final“Local Soil Map” (Fig. 3 B) “Black Soil” covers 38% of themap area, followed by “Hard Red Soil” (34%), “Soft Red Soil”(18%), “Orange Soil” (2%), “Yellow Soil” (2%), and “MixedSoils” (7%).

3.4 Comparison of soil properties

The semistructured interviews revealed that the farmers areable to relate a range of soil properties, like infiltration, sticki-ness, erodibility, and crop suitability to the local soil types(Tab. 1). “Black Soil” was considered to have the highestinherent fertility, followed by “Hard Red Soil” and “Soft RedSoil”. In comparison to these soils, the fertility status of“Orange Soil” and “Yellow Soil” was considered as low. “BlackSoil” is characterized by high infiltration rates, erodibility, soilstickiness, and weeds infestation. “Hard Red Soil” is charac-terized by low infiltration rates, soil stickiness, and weedsinfestation; its resistance to erosion is considered to be “med-ium”. “Soft Red Soil” has a low bulk density, high infiltrationrate, negligible erosion hazard, and medium soil stickiness.The difference between “Orange Soil” and “Yellow Soil” is


Figure 2: (A) Cation-exchangecapacity [cmol(+) kg–1], (B) basesaturation [%], (C) plant-availableP [mg kg–1], (D) plant-available K[mg kg–1], (E) organic C [% m], (F)N [% m] of topsoils from differentlocal soil units. BS = “Black Soil”;RH = “Hard Red Soil”; RS = “SoftRed Soil”; OS = “Orange Soil”; YS= “Yellow Soil”.


solely the color. Both these local soil types are characterizedby low infiltration rates, low weed pressure, and high erodibil-ity. Soil stickiness is low. Own analyses of several soil fertilityparameters confirmed the perceived soil fertility status of fourof the five local soil types. Farmers pronounced “Soft RedSoil” clearly more fertile than “Orange” and “Yellow Soils” butall measured fertility parameters indicated that “Soft RedSoil”, “Orange Soil”, and “Yellow Soil” have a similar low ferti-lity status (Fig. 2). Laboratory analysis confirmed the com-paratively high fertility level of “Black Soil”, manifest from thehighest values for CEC of clay, base saturation, and OM andN contents. “Hard Red Soil” also has high base saturationand OM and N contents but the CEC of the clay is extremelylow. The fertility parameters of “Soft Red Soil” are similar tothose of the “Yellow Soils” and “Orange Soils”.

Many factors influence soil fertility; therefore the analysis oftopsoil samples can only provide a general impression. Theevaluation of the analytical results followed Landon (1991).CEC is high in “Black Soil”, “Hard Red Soil”, and “Soft RedSoil” and moderate in “Orange Soil” and “Yellow Soil”. Thesetallies with their median clay-content values, with >35% in“Black Soil”, “Hard Red Soil”, and “Soft Red Soil” and <35%in “Orange Soil” and “Yellow Soil”. “Black Soil” has the highestbase-saturation median values (>50%; Fig. 2A), followed by“Hard Red Soil” (48%), “Yellow Soil” (37%), “Soft Red Soil”(23%), and “Orange Soil” (21%). The pH values of all localsoils are medium; those of “Black Soil” and “Red Soil” arearound 6.5 whereas “Orange Soil” and “Yellow Soil” have val-ues around 6.0. The unifying factor is that nearly all soils aredecalcified. Only one Chernozem and one Gleysol with mea-surable carbonate content were found. The median valuesfor plant-available P are only marginal (6.5–3 mg kg–1) for


Figure 3: (A) Petrographic map ofthe Bor Krai karst area, (B) “LocalSoil Map”; (C) soil map according tothe WRB-classification soil-grouplevel.


“Black Soil” and “Soft Red Soil”, deficient (3–6.5 mg kg–1) for“Hard Red Soil” and “Orange Soil”, and acutely deficient(<3 mg kg–1) for “Yellow Soil” (Fig. 2C). The average pH val-ues suggest low available-P levels in the soil parent materialas the main reason for the low values found. Median valuesfor plant-available K are adequate (117–196 ppm) for “BlackSoil” and “Yellow Soil”; “Hard Red Soil” and “Orange Soil”have marginal (59–117 mg kg–1) available-K median values(Fig. 2D). “Soft Red Soil” is deficient in available K(<59 mg kg–1). The comparably low K levels of “Red Soils”might be explained by the low K content of residual clayminerals from limestone weathering. Organic-C contents of“Black Soil” and “Hard Red Soil” are low with ≈30 g kg–1.Those of “Soft Red Soil”, “Orange Soil”, and “Yellow Soil” arevery low with values <20 g kg–1. The N contents of “BlackSoil” and “Hard Red Soil” are medium to low; “Soft Red Soil”,“Orange Soil”, and “Yellow Soil” have clearly lower N contents(Fig. 2E).

The highest infiltration rates were measured on “Soft RedSoil”, with >200 cm h–1, followed by “Black Soil” (11 cm h–1),“Hard Red Soil” (10 cm h–1), “Orange Soil” (6 cm h–1), and“Yellow Soil” (<1 cm h–1).

3.5 Comparison of “Conventional Soil Map” and“Local Soil Map”

The pattern of the “Local Soil Map” is similar to that of the pet-rographic map (Fig. 3 A, B), whereas the WRB soil mapappears totally different. Poor correlation between WRB-

based soil maps and local soil maps was also reported byothers, e.g., Payton et al. (2003) and Ali (2003). The mainreason cited is the differences in the conceptual bases of thesoil-classification systems used (Niemeijer and Mazzucato,2003). The WRB classification system considers essentialchemical soil parameters (e.g., CEC) and/or properties belowthe soil surface (e.g., mottles) that are invisible to farmerswhereas the local soil types are solely based on visible soilcolor and soil structure. “Black Soil” covers all areas with pre-served dark topsoil; the other local soil types are predomi-nantly marked by subsoil color as surface soils were erodedaway. Strong correlation between local–soil map units andpetrographic map units is found in >55% of the area, wherethe dark topsoil is strongly degraded, and the subsoil color isexposed at the surface. There, “Red Soils” cover >98% of thelimestone area, and “Orange Soils” and “Yellow Soils” cover>99% of the petrographic clay-, silt-, and sandstone unit(Tab. 1). The total area, of “Red Soils” is composed of some40% Acrisols, 46% Luvisols, 12% Cambisols, and 2% otherreference soil groups.

“Hard Red Soil”, “Soft Red Soil”, and “Mixed Black Red Soil”units include a considerable component of Acrisols. The high-est proportion of Cambisols can be found within “Red Soil”units.

“Black Soil”, “Orange Soil”, “Yellow Soil”, and “Mixed Red Yel-low Black Soil” units consist for >80% of Luvisols. The domi-nance of soils with clay illuviation (Acrisols and Luvisols) isexplained by the low surface-soil pH which is conducive to


Table 1: Petrographic units and Reference Soil Groups as components of local soil units.

Local soil units “Black Soil”[%]

“Hard RedSoil”[%]

“Soft RedSoil”[%]

“OrangeSoil”[%]

“Yellow Soil”[%]

“Mixed RedBlack Soil”[%]

“MixedYellowBlack Soil”[%]

“Mixed RedYellowBlack Soil”[%]

Petrographic units

Alluvium 1 1 0 0 0 0 0 <1

Andesite <1 0 0 0 0 11 0 0

Clay-, silt-, sandstone 75 <1 <1 100 100 54 99 6

Dolomite 0 <1 0 0 0 0 0 0

Freshwater limestone <1 0 <1 0 0 0 0 0

Iron ore 0 1 0 0 0 0 0 0

Limestone 24 98 100 0 <1 34 1 94

Reference Soil Groups

Acrisols <1 39 44 <1 0 25 0 0

Cambisols 5 11 14 1 <1 12 4 17

Chernozems <1 0 0 <1 0 0 0 0

Ferralsols 0 <1 0 <1 0 0 0 0

Fluvisols <1 0 0 <1 0 0 0 0

Gleysols <1 0 0 <1 0 0 <1 0

Leptosols 1 1 2 <1 0 1 0 0

Luvisols 93 49 40 99 100 63 96 83

Umbrisol 0 0 <1 <1 0 0 0 0


clay dispersion and by the seasonal climate with heavy rainshowers that enable mechanical transport of clay down to theargic accumulation horizon. Subsoil color is largely dictatedby the mineral composition of Fe-oxides. Hematite accountsfor the red color of soils on andesite and limestone; goethitecauses the yellow subsoil color of soils in clay-, silt-, andsandstone weathering. Hematite mottles (“ferric properties” inthe WRB system) in the yellow soil matrix of clay-, silt-, andsandstone weathering produce an orange subsurface soil.Small quantities of hematite can already change the soil colorto red. Our results suggest that local climate, high subsoildrainage, low OM content of the subsoil, limestone, and prob-ably accumulation of residual Fe all favor the development ofhematite. X-ray analyses revealed hematite in weatheredandesite and some 10% in Fe. Hematite development seemsto be suppressed on clay-, silt-, and sandstone, probablybecause of impeded drainage. The “Local Soil Map” revealsstrong correlation between parent material and local soilunits. To test the influence of the parent material on essentialsoil properties, cluster analysis with the “nearest-neighbormethod” was applied considering farmers’ rankings of erod-ibility, water-infiltration rate, available water capacity, soilstickiness, nutrient content, nutrient absorption, and soil color(hue) (Fig. 4).

On the dendrogram of the hierarchical cluster analysis, thefirst split at squared Euclidian distance 25 (see Fig. 4 scalebar) leads to three different clusters. The upper cluster com-

prises all mapping units with the exception of alluvial depositsand freshwater limestone. The central cluster representsfreshwater limestone, the lower cluster alluvial deposits. In asecond split at squared Euclidian distance 20, Fe-ores buildtheir own cluster. Andesite forms a single cluster after thethird split squared Euclidian distance 15. After the fourth split,parent material still causes the clearest cluster groups. Weconclude that the cluster analysis accords with farmers’ per-ception that parent rock dictates soil properties.

We could not find any similar case of correspondence be-tween local soil maps and petrographic maps in literature.

4 Conclusions and outlook

The “Local Soil Map” and the WRB-based soil map of the BorKrai study site are strongly dissimilar. While the “Local SoilMap” to a large extent corresponds with the petrographicmap, with exception of the “Black Soil”, the WRB map ismore or less independent from the parent material, becauseLuvisols and Cambisols occur irrespective of parent material.Due to the dominance of limestone above 800 m asl, theindependency of Acrisols and Ferralsols could not be proven.

Farmers, on the other hand, classify their soils according tocolor which first and foremost reflects the parent material thatwas subject to various degrees of weathering and subse-quent erosion. It is expected that population increase and


Figure 4: Hierarchical cluster analysisbased on farmers’ ranking of erodibility,water-infiltration rate, available watercapacity, soil stickiness, nutrient content,nutrient absorption, and color (hue).(Cluster method: nearest neighbor.Measure interval: squared Euclideandistance).


intensification of agriculture in the Bor Krai area will acceler-ate soil erosion. With the loss of the topsoils, the similarity be-tween the Local Soil Map and the petrographic map is likelyto increase. This study confirms that predominant local soiltypes, their properties and features can be efficiently identi-fied on the basis of participatory soil mapping (PSM).Although PSM certainly cannot replace conventional soilmapping, it can reduce fieldwork to some extent. It facilitatesthe first step of getting acquainted with the terrain understudy. An important prerequisite is that surveyors are able tocommunicate with the rural population in personal interviewsand group discussions, which can be facilitated through coop-eration with social scientists. A further advantage of PSM isthat the local soil maps produced are a base for knowledgeexchange between farmers and surveyors. Both have theirown ways of acquiring knowledge; fruitful synergies areanticipated if the expertise of both groups is brought together.Other disciplines (e.g., policy makers and extension workers)can also benefit from the unveiled local knowledge (cf., Bir-mingham, 2003; Ericksen and Ardon, 2003; Ryder, 2003).Finally, local soil maps can help to integrate the views of localpeople into land-use planning and perhaps avert problemswhich confront local communities if their interests are over-looked.

It is to be expected that local soil-classification systems in NThailand differ between environments and ethnic groups. Ifour results are confirmed elsewhere, integration of local andforeign soil knowledge can be attempted at the landscapeand regional levels.

Acknowledgments

We are grateful to the Black Lahu farmers of Bor Krai whospent their time participating in the survey and welcomed uswith great hospitality. We are particularly indebted to our localresearch assistant Somchay Sookreuntong for his help dur-ing the soil survey and for facilitating the communication be-tween the local people and ourselves. Finally, we express ourgratitude for the helpful comments of Paul M. Driessen andan anonymous reviewer on an earlier draft of this paper.

The financial support of the Deutsche Forschungsge-meinschaft in carrying out this study is gratefully acknowl-edged.

ReferencesAli, A. M. S. (2003): Farmers’ knowledge of soils and the sustain-

ability of agriculture in a saline water ecosystem in SouthwesternBangladesh. Geoderma 111, 333–353.

Barrera-Bassols, N., Zink, J. A. (2003): Ethnopedology: a worldwideview on the soil knowledge of local people. Geoderma 111,171–195.

Barrios, E., Trejo, M. T. (2003): Implications of local soil knowledgefor integrated soil management in Latin America. Geoderma 111,217–231.

Becu, N., Perez, P., Walker, A., Barreteau, O., Le Page, C. (2003):Agent based simulation of a small catchment water managementin northern Thailand. Description of the CATCHSCAPE model.Ecol. Model. 170, 319–331.

Birmingham, D. M. (2003): Local knowledge of soils: the case ofcontrast in Côte d'Ivoire. Geoderma 111, 481–502.

Blume, H.-P., Deller, B., Leschber, R., Paetz, A., Schmidt, S., Wilke,B.-M. (2000): Handbuch der Bodenuntersuchung. Beuth, Berlin,p. 4292.

Chambers, R. (1992): Rural appraisal: rapid, relaxed and partici-patory. Institute of Development Studies (University of Brighton).IDS Discussion paper No. 331, Sussex, UK.

Dunham, R. J. (1962): Classification of carbonate rocks according todepositional texture, in Ham, W. E.: Classification of carbonaterocks – a symposium. AAPG Memoirs 1, Tulsa, pp. 108–121.

Ericksen, P. J., Ardon, M. (2003): Similarities and differences be-tween farmer and scientist views on soil quality issues in centralHonduras. Geoderma 111, 233–248.

Erkossa, T., Stahr, K., Gaiser, T. (2004): Participatory soil qualityassessment: The case of smallholder farmers in Ethopian high-lands. Aust. J. Soil Res. 42, 793–797.

Ettema, C. H. (1994): Indigenous soil classifications. What are theirstructure and function and how do they compare with scientific soilclassifications. Institute of Ecology, University of Georgia, Athens,Georgia, USA.

FAO (1998): World Reference Base for Soil Resources. World SoilResources Reports 84, Rome, p. 81.

FAO (2001): Lecture Notes on the Major Soils of the World. WorldSoil Resources Reports 94, Rome, p. 334.

German Geological Mission (GGM) (1979): Geological Map ofNorthern Thailand 1:250000. Federal Institute for Geosciencesand Natural Resources, Germany.

Hendricks, C. A. (1981): Soil-vegetation relations in the north conti-nental highland region of Thailand; a preliminary investigation ofsoil-vegetation correlation. Technical Bulletin No. 32. Soil SurveyDivision, Department of Land Development; Ministry of Agricultureand Cooperatives, Thailand.

Herrmann, L. (ed.) (2005): Das kleine Bodenkochbuch (Version2005). Institut für Bodenkunde und Standortslehre der UniversitätHohenheim, Stuttgart, p. 116.

Keiichi, H., Toshiyuki, W. (2002): Sustainable soil fertilitymanagement by indigenous and scientific knowledge in the Sahelzone of Niger. Symposium 15, Paper 1251, 17th World Congressof Soil Science, Bangkok, Thailand.

Klute, A. (1986): Methods of soil analysis, in Black, C. A., Klute, A.:Physical and mineralogical methods. American Society ofAgronomy 9.1, Madison, WI, p. 1188.

Krasilnikov, P. V., Tabor, J. A. (2003): Perspectives on utilitarianethnopedology. Geoderma 111, 197–215.

Landon, J. R. (1991): Booker tropical soil manual. Longman &Booker, England, p. 474.

Matthes, S. (1996): Mineralogie – Eine Einführung in die spezielleMineralogie, Petrologie und Lagerstättenkunde. Springer, Berlin,p. 499.

Niemeijer, D., Mazzucato, V. (2003): Moving beyond indigenous soiltaxonomies: local theories of soils for sustainable development.Geoderma 111, 403–424.

Oudwater, N., Martin, A. (2003): Methods and issues in exploringlocal knowledge of soils. Geoderma 111, 387–401.

Payton, R. W., Barr, J. J. F., Martin, A. Sillitoe, P., Deckers, J. F.,Gowing, J. W., Hatibu, N., Naseem, S. B., Tenywa, M., Zuberi, M. I.(2003): Contrasting approached to integrating indigenousknowledge about soils and scientific soil survey in East Africa andBangladesh. Geoderma 111, 335–386.



Perez, P., Ardlie, N., Kuneepong, P., Dietrich, C., Meritt, W. S.(2001): CATCHCROP: modeling crop yield and water demand forintegrated catchment assessment in northern Thailand. Environ.Model. Software 17, 251–259.

Reij, C., Scoones, I., Toulmin, C. (1996): Sustaining the Soil. Indi-genous Soil and Water Conservation in Africa. Earthscan Publica-tions, London, p. 224.

Royal Thai Survey Department (1976): Topographic Map 4648 III –Ban Mae Lana, Scale 1:50,000. Bangkok.

Rudebjer, P., Taylor, P., Del Castillo, R. A. (eds.) (2001): A Guide toLearning Agroforestry. A framework for developing Agroforestrycurricula in Southeast Asia. Training and Education Report No. 51,World Agroforestry Center (ICRAF), Bogor.

Ryder, R. (2003): Local soil knowledge and site suitability evaluationin the Dominican Republic. Geoderma 111, 289–305.

Schlichting, E., Blume, H. P., Stahr, K. (1995): BodenkundlichesPraktikum. Pareys Studientexte 81, Berlin, p. 295.

Schuler, U., Spohrer, K., Herrmann, L., Stahr, K. (2004): Variability ofsoils in a karst catchment of northern Thailand. Proceedings ofTrans-Karst 2004 – International Transdisciplinary Conference onDevelopment and Conservation of Karst Regions, Vietnam.

Sereke, F. (2002): Land evaluation for sustainable highland agri-culture in NW-Thailand (Pang Ma Pha) – with respect to soil andwater resources. M.Sc. thesis, University of Hohenheim, Stuttgart.

Talawar, S., Rhoades, R. E. (1998): Scientific and local classificationmanagement of soils. Agriculture and Human Values 15, 3–14.

Tinoco-Ordónez, R. (2003): Steps towards sustainable agriculture:an ethnopedological soil survey in a limestone area of NorthernThailand. Master thesis, University of Hohenheim, Stuttgart.

Tucker, M. (1985): Einführung in die Sedimentpetrologie. FerdinandEncke-Verlag, Stuttgart, p. 275.

Van Engelen, V., Wen, T. T. (1995): Global and national Soils andterrain digital databases (SOTER). Procedures Manual (revisededition). ISRIC, Wageningen, p.125.

Van Noordwijk, M. (2002): Scaling trade-offs between crop produc-tivity, carbon stocks and biodiversity in shifting cultivation land-scape mosaics: the FALLOW model. Ecol. Mod. 149, 113–126.

VDLUFA (1991): Methodenbuch – Die Untersuchung von Böden.VDLUFA-Verlag, Darmstadt.

Vijarnsorn, P., Eswaran, H. (2002): The soil resources of Thailand.Land Development Department, Bangkok, Thailand, p. 264.

Vlassak, K., Ongprasert, S., Tancho, A., Van Look, K., Turkelboom,F., Ooms, L. (1992): Soil Fertility Conservation Research Report1989-1992. SFC project, Chiang Mai, Thailand, p. 255.

Vlassak, K. (1994): Soil Fertility Conservation Research Report1993. SFC project, Chiang Mai, Thailand, p. 138.

WinklerPrins, A. M. G. A. (1999): Local soil knowledge: a tool forsustainable land management. Soc. Natural Resourc. 12,151–161.



Documents

Soil mapping for land-use planning in a karst area of N Thailand with due consideration of local knowledge