The Use of Fuzzy Logic and Geostatistical Techniques for ...

International Journal of Agriculture, Forestry and Fisheries 2015; 3(3): 115-122

Published online May 30, 2015 (http://www.openscienceonline.com/journal/ijaff)

The Use of Fuzzy Logic and Geostatistical Techniques for Spatialization of Soil Texture in Part of Kano Close Settled Zone Kano State, Nigeria

Mohammed Ahmed

Faculty of Earth and Environmental Science, Department of Geography, Bayero University Kano, Kano State, Nigeria

Email address

[email protected]

To cite this article Mohammed Ahmed. The Use of Fuzzy Logic and Geostatistical Techniques for Spatialization of Soil Texture in Part of Kano Close Settled

Zone Kano State, Nigeria. International Journal of Agriculture, Forestry and Fisherie. Vol. 3, No. 3, 2015, pp. 115-122.

Abstract

Lack of spatial information on soil textural classes has attributed to the wrong usage of land for cultivation. The introduction of

Fuzzy logic and Geostatistic offers a convenient tool to help solving the problem of the use of generalization in data analysis

for soil texture using the conventional method. 51 soil samples from a gridded map of the area to the depth of 30 cm were

analysed in the laboratory and imported into Arc GIS 10.1. The purpose of this paper is to present a methodology that provides

the soil texture spatialization by using Fuzzy logic and Geostatistic in the study area. A Geostatistical interpolation algorithm

was employed to determine the prediction performances of the model. Semivariograms were produced for each soil textural

class. The results of the spatial distribution of the textural classes indicated that loamy sand has dominated the area. The study

shows that kriging was the best technique for each soil textural classes. The knowledge of the spatialization of soil properties,

such as the texture, can be an important tool for land use planning.

Keywords

Soil, Kriging, Co-Kriging, GIS, Semivariogram, Geostatistics, Fuzzy Logic

1. Introduction

Soil textural class is one of the important components that

determine the soil nutrient and the available water holding

capacity which has several implications for management in

agricultural. Distribution of soil texture in a cultivated land

can help in identifying areas suitable or unsuitable for a

particular crop. Furthermore understanding the rate of

variation of soil texture in the area has been identified by

Essiet (1995) which shows that sand fraction exceed 70 %

but except in hydromorphic soils that is found within the

river valley with less sandy proportion.

Analyzing the distribution of soil texture has been carried

out using conventional approach which most of the result

were presented in tables, that is to say sandy, loamy, clayey

soil etc, (in percentages). This type of analysis affects the

interpretation of the soil texture in the area because of the use

of mean or average. Value of a particular sample (point) may

go higher than the rest of the values (outlier) which may

probably be as the result of error analysis in the laboratory.

Conventional soil sampling procedures (using soil units as a

criteria) affects the result which causes error in the result

obtained, the values obtained are good, but the interpreter

may realize that many of the values in the field are (outliers)

either less than or greater than the values determined (Mahler

and Tindall, 1990).

Another point is that, soil parameters are continuous data,

therefore care need to be taking when classifying soil

information by delineating boundary as soil units. Yes, soil

may have boundary but that depend on the soil management

practices in the area or other factors of land characteristics

(like geology, lower and upper terrace, farming system or

management etc). The boundary of many geographical

features is an artifact of human perception, not of the real

world (Fisher, 1999). However, the result of this type of

analysis may yield an error because of the uncertainty and

vague this has been stated in Mahler and Tindall (1990). The

knowledge about the uncertainty is crucial for proper

understanding of the content of the map; therefore, if an

agreement on how to express the map quality can be reached,

then the information should be documented and released to

users according to Longley et al., (1999) in Kˇremenová

(2004).

International Journal of Agriculture, Forestry and Fisheries 2015; 3(3): 115-122 116

This study attempted to solve the problem in analyzing and

interpretation of soil texture using mapping technique. The

spatial distribution of geographical phenomenon (like soil

texture) could be mapped and analyzing the distribution and

concentration of a particular textural class. Kˇremenová

(2004) lamented that many spatial phenomena are inherently

fuzzy or vague or possess indeterminate boundaries. Fuzzy

logic has been applied for many areas in GIS such as fuzzy

spatial analysis, fuzzy reasoning, and the representation of

fuzzy boundaries.

Fuzzy logic is a mathematical model that is used in

analyzing sets whose elements have degrees of membership.

Fuzzy sets were introduced by Zadeh (1965) as an extension

of the classical notion of set. It is a form of many-valued

logic which deals with reasoning that is approximate rather

than fixed and exact. Compared to traditional binary sets

(where variables may take on true or false values), fuzzy

logic variables may have a truth value that ranges in degree

between 0 and 1. It has been extended to handle the concept

of partial truth, where the truth value may range between

completely true and completely false (Zadeh,1965).

A basic application might characterize sub-ranges of a

continuous variable. For instance, a fertility status of soil

could be determine using one of the factors like soil organic

matter, the measurement may have several separate

membership functions defining the ranges needed to show

the status of the soil. In this case, the analysis could be

presented in either fertile or not fertile which is 1 or 0, but the

problem with this is the possibility of having such kind of

result without defining a boundary or a criteria which will

help in determining the differences in classes of fertility like,

slightly fertile, moderately fertile, highly fertile and so on.

This could be presented using values that ranges from 0 to 1,

pending on the values were categorized in the analysis

(Ahmed et al, 2015).

Fuzzy set theory has been widely used in soil science for

soil classification and mapping, land evaluation and

suitability, soil geostatistics, soil quality indices and decision-

making analyses (McBratney et al., 2003; Lagacherie, 1997;

Busscher et al., 2007; Torbert et al., 2000; Keshavarzi et al.,

2011; Ahmed et al., 2015). While Kˇremenová (2004)

applied a fuzzy model to develop a soil map using the

knowledge of geologists, the study shows that fuzzy expert

system, capable to handle the spatial variability of soil

behavior. In these studies no attempt were made to classify

and analyse the distribution of soil texture.

2. Study Area

Bunkure LGA is located between Latitude 110 34’ 02’’N to

Latitude 110 46’ 05’’N of the Equator and between Longitude

80 26’ 36’’E to Longitude 8

0 46’ 43’’E of the Prime Meridian.

The study area is comprises of fifteen wards (15) with an

aerial extend 9911.22 Km2 and is bordered with Dawakin

kudu and Kura LGAs by the North, Wudil and Garko LGA

by the East while Kibiya at the South Western part of the

study area (Ahmed et al, 2015).

The area experiences four distinct seasons, long wet

(damina) and dry (rani) seasons and short autumn (kaka) and

spring (Bazara) seasons closely associated with the

movement of the ITD. The mean annual Rainfall is about

884mm varying greatly from the northern and southern parts

of the Region (Maryam, Halima and Ummi, (2014) and

Usman, (2014)).

Figure 1. Study Area.

117 Mohammed Ahmed: The Use of Fuzzy Logic and Geostatistical Techniques for Spatialization of Soil Texture in Part of Kano Close

Settled Zone Kano State, Nigeria

Figure 2. Sampling Points.

Figure 3. Converting of the textural classes of soil using fuzzy logic (Adopted from Camarinha et al, 2011).

The geology of Bunkure has been dominated with carse

porhyntiuc biotite and Biotite hornblende Granite which is

under the Pan African Older Granite series within the

Pericambrian to Cambrian Basement Complex (NGSA,

2006). Latosols are the dominant soils in the wind drift of

Kano state. They are well-drained and brownish to reddish in

colour. They are also deep except where iron pans are

exposed or occur near the surface (Olofin, 1985 and Ahmed,

2006). About 75% of the land is cultivated parkland with

average tree densities of less than 25 per hectare (Badamasi,

2013).


3. Materials and Methods

Soil samples were taken from the depth of 0 – 30 cm using

a grid of 1500 x 1500 meters (Figure 2). In order to facilitate

the field work and to ensure the accuracy of the

examination/sampling points, it were uploaded on Global

Positioning Systems (GPS) Garmin 76csx model and the soil

examined and sampled according to MARDITECH (2011),

Wollenhaupt et al., (1994); Rehm et al., (2001) and Umar

(2011).

The samples were taken to the Standard laboratory (Soil

and Water Lab. of the Department of Geography Bayero

University Kano) air dried and gently crushed with porcelain

pestle and mortar; and then passed through a 2mm sieve to

remove coarse fragments. The fine earth samples (<2mm soil

portion) collected were analyzed. Particle size distribution

was determined using hydrometer method (Gee and Bauder,

1986). Sand, silt and clay were determined by dispersing the

soil samples in 5% calgon (sodium hexametaphosphate)

solution. The dispersed samples were shaken on a

reciprocating shaker after which particle size distribution was

determined with the aid of Bouyoucous hydrometer at

progressive time intervals. The textural classes were

determined with the aid of USDA textural triangle.

The next step is to arrange the textural classes based on the

percentage of clay content (using rating) which could be

converted in numerical values (fuzzy values), using the fuzzy

logic concept (see figure3). Thus, two extremes were

established, representing predominantly sandy soils (values

near from 0) and clayey soils (values near from 1). All

samples were classified numerically according to their degree

of relevance on the texture (Camarinha et al, 2011). The

results were entered into the Microsoft Excel with their

respective coordinates (Latitudes and Longitudes) and

transformed into GIS environment using Arc GIS 10.1

(version) in order to analyse the spatial distribution of the soil

texture in the area according to Mustafa et al, (2011).

Figure 4. Distribution of soil texture.

Table 1. Fuzzy values representing the soil texture for all samples.

Sample Fuzzy value Sample Fuzzy value Sample Fuzzy value Sample Fuzzy value

1 0.04 26 0.16 13 0.12 38 0.36

2 0.04 27 0.2 14 0.12 39 0.4

3 0.08 28 0.24 15 0.12 40 0.4

4 0.08 29 0.24 16 0.12 41 0.44

5 0.12 30 0.28 17 0.12 42 0.44

6 0.12 31 0.28 18 0.12 43 0.45

7 0.12 32 0.32 19 0.12 44 0.5

8 0.12 33 0.36 20 0.16 45 0.5

9 0.12 34 0.36 21 0.16 46 0.5

10 0.12 35 0.36 22 0.16 47 0.54

11 0.12 36 0.36 23 0.16 48 0.6

12 0.12 37 0.36 24 0.16 49 0.7

25 0.16 50 0.8

51 0.9

Sources: Data analysis (2014)



Seminariogram models were tested from Arc GIS 10.1. In

predicting spatial process at nonsampling sites using

geostatistics, it is necessary to decide on a theoretical

variogram based on the experimental variogram. It is vital to

choose an appropriate model to estimate spatial statistics as

each model yields different values for nugget, sill, partial sill,

spatial dependency and range which are essential for

geostatistical analyses (Trangmar and Uehara, 1985).

4. Results and Discussions

By the methodology described previously, all samples

have a numerical value to represent its textural class (fuzzy

value), as presented in the Table 1. Geostatistical extension in

the software was used for spatial interpolation. This is a

value of a variable at unsampled locations. The interpolation

techniques commonly used in earth sciences include Semi-

variogram of Ordinary Kriging which was used to produce

the spatial distribution of the soil texture. The equation of

semivariogram can be defined as stated by Webster and

Oliver (1993):

ý(h) = 1/2N(h)∑ �Z�x��– Z�x� h��

��

Where ( ) i Z x is the value of the regionalized variable Z(x)

at location of i x, Z(x h) i + is the value of the variable Z(x) at

location of (x h) i + , and N(h) is the number of pairs of

sampling sites separated by the lag distance h

The next step was to analyze the spatial characteristics of

the variable (texture) in order to verify the viability of the

spatial interpolation of data by kriging. To apply this

geostatistical method, it is necessary that the spatial

distribution of the variable (texture) meets certain

geostatistical assumptions. Geostatistics uses the technique of

variogram to measure the spatial variability of a regionalized

variable, and provides the input parameters for the spatial

interpolation of kriging (Webster and Oliver, 1993).

From the verification and settings of the semivariogram

presented in Figure 4, the Geostatistical Wizard tool is able to

generate the spatial distribution map of the texture, by

interpolation of the spatial data. The result shows the

variability of the texture in whole study area. The

concentration of more soil with clay content while at the

southern part of the area showing moderate concentration and

the extreme Northern site is more of soil with sandy content.

4.1. Spatial Structure of Soil Texture

The best fit model was selected using semivariogram

models to determine the textural variability in the area. From

different models tested it indicated that “Exponential” was

the best fitted with the Root Mean Square Error (RMSE) of

the least value of 0.2027. While the lag size (m), range (m),

Nugget, partial sill and sill of 388.5, 2,798.17m, 0.02550,

0.0173 and 0.0428 respectively (Table 2 and figure 5)

Table 2. Values of different model fittings of semivariogram of soil texture.

Variable Depth

(cm)

Lag size

(m) Model

Nugget

(Co) Sill (Co+C)

Spatial Dependency

Co/(Co+C) (%)

Partial sill

(C)

Range

(m) RMSE

Soil texture 30 388.5 Exponential 0.0255 0.0428 0.596 0.0173 2,798.17 0.2027

Sources: Data analysis (2015)

Data analysis (2015)

Figure 5. Semivariogram model of soil texture.


Figure 6. Fuzzy classification for textural classes.

The semivariogrm obtained from the experimental data

often had a positive value of the intersection with the

variogram axis expressed by the named nugget effect Co.

This can be explained by sampling errors, shortage variability,

and unexplained and inherent variability. It can also indicate

the irresolvable variance that characterizes the micro-

inhomogeneity at the sampling location (Jayeoba et al, 2013).

Some semivariograms are generally well structured with

small nugget effect. It showed that the sampling is adequate

to reveal the spatial structures (McGrath et al, 2004).

The nugget/sill ratio or Spatial dependency Co/(Co+C)

defines the spatial property. The variable is considered as a

strong spatial dependence when the value of Co/(Co+C) is

less than 0.25, a moderate spatial dependence when this

value is between 0.25 and 0.75, and a weak spatial

dependence when the value is more than 0.75 (Cambardella

et al, 1994). The Spatial dependency Co/(Co+C) for the soil

texture at the depth of 30cm, these shows a moderate spatial

dependency with value of 0.596 (within 0.25 to 0.75). This

revealed that the spatial distribution of soil texture at the

30cm depth in the area is dominated by human activities such

as irrigation can cause some variation among the textural

classes. These suggested that the extrinsic factors such as for

plowing and other soil management practices weakened their

spatial correlation after a long history of cultivation.

4.2. Fuzzy Classification for Textural Classes

The fuzzy analysis shows that the soil of the area has been

dominated by mostly sandy particles. Three soil textures

were identified in the area which includes; loamy sand, sandy

loam and sandy clay loam. The soil textural classes has been

affected by it Geological formation and soil type. The soils of

the area according to Esseit (2013) formed over the

Basement Complex rocks which are relatively well structured

and posses sufficient depth to permit the cultivation of most

staple crops. The Basement Complex rocks are quite variable

in size and composition and include schists, shales and

granites among others.

Loamy sand in Figure 5 occupied 72.53% of the land in

the area followed by sandy loam with 17.82% while sandy

clay loam with 9.65%. This shows that places around the

Southern part of the area realized some patchy concentration

of sandy loam soils and also at the central area towards the

Eastern part experienced the presence of sandy loam and

sandy clay loam within an area which indicated a more clay

in the area than other places, this is as the result of closeness

to the river in the area which is more of hydromorphic type

of soils (Essiet 1995). The dominance of loamy sand in the

area has been characterized by Essiet (2013) as ferruginous

type and that is also associated with Basement Complex in

the areas.

5. Conclusion and Recommendations

The results of the spatial distribution of the textural classes

indicated that loamy sand has dominated the area. The study

shows that kriging technique is one of the best techniques for

classifying soil textural classes. This approach can improve



on the use of conventional method by saving time, manpower

and energy, also combining kriging and fuzzy logic provide

better data management which helps in modeling. The

knowledge of the spatialization of soil properties, such as the

texture, can be an important tool for land use planning and

agricultural sustainability.

Therefore, the study recommends for proper soil

investigation particularly at the larger scale using GIS and

Remote sensing techniques.

Acknowledgements

This is to acknowledge the contributions of Mallam

Murtala U. Mohammed, Mallam Abubakar of the Soil and

Water laboratory and Dr. Adnan Abdulhamid of the

Department of Geography Bayero University Kano, for their

assistance and guidance especially during data collection,

analysis and presentation.

References

[1] Ahmed M, Jeb D. N, Usman A. K., Adamu G. K and Mohammed M. U. (2015) Spatial Distribution and Assessment of Selected Soil Physiochemical Parameters Using GIS Techniques in Bunkure Kano State, Nigeria International Journal of Plant & Soil Science 5(3): 143-154, IJPSS.2015.068 ISSN: 2320-7035 SCIENCEDOMAIN international www.sciencedomain.org

[2] Ahmed, K. (2006). The Physical Environment of Kano State, www.kanostate.net/physicalenvironment.html.

[3] Badamasi M. M., (2014) Vegetation and Forestry In A.I Tanko and S.B. Mumale (Eds.) Kano Environment, Society and Development. Adonis and Abbey Publishers

[4] Busscher, W., Krueger, E., Novak, J., Kurtener, D. (2007) Comparison of soil amendments to decrease high strength in SE USA Coastal Plain soils using fuzzy decision-making analyses. International Agrophysics 21, 225–231

[5] Camarinha P.I.M. Trannin I.C.B., Simões S.J.C., Bernardes G.P. (2011) Fuzzy Logic and Geostatistical Techniques for Spatialization of Soil Texture in Region with Rough Terrains. Procedia Environmental Sciences 7 (2011) 347–352 Available online at www.sciencedirect.com

[6] Cambardella, C.A., T.B. Moorman, T.B. Parkin, D.L. Karlen, R.F. Turco, and A.E. Konopka.. Field scale variability of soil properties in Central Iowa soils. Soil Sci. Soc. Am. J. 58:1501–1511. 1994

[7] Esseit U. E. (2013) Soils. Geography of Kano region. Tanko and Mumale S, B. (Eds) Kano Environment, Society and Development. London and Abuja, Adonis and Abbey Publishers

[8] Essiet, E.U. (1995), Soil Management and Agricultural Sustainability in the Smallholder Farming System in Northern Nigeria, Journal of Social and Management Studies, Vol. 2, 37-46.

[9] Fisher, P. F. 1999. Models of uncertainty in spatial data. In Longley et. al.: Geographical Information Systems, Volume 1: Principles and Technical Issues. John Wiley & Sons. ISBN: 0471331325. 8, 9, 11, 13, 15, 16, 36

[10] Gee G. W. and J. W. Bauder. (1986): Particle size analysis. In ‘Methods of soil analysis, Part 1’.Vol. 9.(Ed. A. Klute) pp. 91–100.American Society of Agronomy: Madison,WI.

[11] Jayeoba O J, Amana S M, and Ogbe V B. (2013) Spatial Variation of Soil Moisture Content and Total Porosity As Influenced by Land Use Types in Lafia, North Central Nigeria PAT; 10 (1):53-66: ISSN: 0794-5213

[12] Kˇremenová O. (2004) Fuzzy Modeling of Soil Maps. Unpublished Msc. Theses submitted to Helsinki University of technology Department of surveying

[13] Lagacherie P, Cazemier D.R.,. van Gaans, P.F.M Burrough P.A. (1997) Fuzzy k-means clustering of fields in an elementary cathment and extrapolation to a larger area. Geoderma. 77(2-4):197-210

[14] Longley, P. A., Goodchild, M. F., Maguire, D. J., Rhind D. W. (1999) Geographical information Systems, Volume 1: Principles and Technical Issues. John Wiley & Sons. ISBN: 0471331325. 16, 29

[15] Mahler, R.L. and T.A. Tindall (1990) Soil sampling. Univ. Idaho Coop. Ext. Bull. 704 (revised)

[16] MARDITECH (2011). Development of a GIS-Based Soil Suitability Classification for Rice Production in Kano State, Nigeria, Unpublished Interim Report submitted to the Kano State Government, MARDITECH, Kualar Lumpur, Malaysia

[17] Maryam L, Halima A. Idris and Ummi K. Mohammed (2014). Weather and Climate. In A.I Tanko and S.B. Mumale (Eds.) Kano Environment, Society and Development. London and Abuja, Adonis and Abbey Publishers

[18] McBratney A, Inakwu, O.A. Odeh. The application of fuzzy sets in soil science: fuzzy logic, fuzzy measurements and fuzzy decisions. Geoderma. 77(2-4):85-113. 1997.

[19] McGrath, D.Zhang, C. S and Carton O. T, Geostatistical analyses and hazard Assessment on soil lead in Silvermines area, Ireland. Environmental Pollution. vol,127, 2004. pp.239-248

[20] Mustafa A.A, Man Singh, R. N Sahoo, Nayan Ahmed, Manoj Khanna, A. Sarangi and A. K. Mishra (2011) Land Suitability Analysis for Different Crops: A Multi Criteria Decision Making Approach using Remote Sensing and GIS. Researcher, 2011; 3 (12) http://www.sciencepub.net/researcher

[21] NGSA, Nigerian Geological Survey Agency (2006) Geological map of Kano State

[22] Olofin E. A. (1985). Human Responses to the Natural environment in the Kano Region. In: Barkindo, A. A. (ed). Kano and its Neighbours. ABU Press, Zaria.

[23] Rehm, G.W., A.P. Mallarino, K. Reid, D. Franzen, and J. Lamb. (2001). Soil sampling for variable-rate fertilizer and lime application. North Central Multistate Report 348 - NCR-13 Committee. Minnesota Agricultural Experiment Station Bull. 608-2001. Univ. of Minnesota, St. Paul., MN.

[24] Torbert, H.A., Searcy, S.W., Kenimer, A.L., Roades, J. (2000) Precision farming effects on corn productivity and water quality. In: Second international conference on geospatial information in agriculture and forestry, Lake Buena Vista, Florida


[25] Trangmar, B.B. Yost, R.S. Uehara G (1985) Application of geostatistics to spatial studies of soil properties. Adv Agron 38: 1985. 45–94pp

[26] Umar, G. (2011) Assessment of the Fertility Status of some Irrigated Fluvisols in Northern Guinea Savannah of Nigeria. Volume 6 (1); Savannah Journal of Agriculture ISSN 1597 9377 Faculty of Agriculture, Bayero University

[27] Usman A (2014) Rainfall variability in Kano region. Unpublished Msc. Theses submitted to the Department of Geography Bayero University Kano, Nigeria

[28] Webster, R. and Oliver, M.A. (1993) “How large a sample is needed to estimate the regional variogram adequately?” In: Soares, A. (ed) Geostatistics Troia '92. Kluwer, Dordrecht, p. 155-166.

[29] Wollenhaupt, N.C., R.P. Wolkowski, and M.K. Clayton. (1994) Mapping soil test phosphorusand potassium for variable-rate fertilizer application. J. Prod. Agric. 7:441-448.

[30] Zadeh L. A. (1965) "Fuzzy sets". Information and Control 8 (3) 338–353. http://www.sjsu.edu/faculty/watkins/fuzzysets.htm "Fuzzy Logic: The Logic of Fuzzy Sets"

The Use of Fuzzy Logic and Geostatistical Techniques for ...

Documents