rule-based soil erosion modeling in gis

 

Eva Švandová

Department of geography

Masaryk University Brno

Kotlárská 2

611 37 Brno

Czech Republic

tel.:+420-5-42 128 317

evas@ porthos.geogr.muni.cz

 

Summary:

This paper deals with evaluation of susceptibility of the cadaster of the commune Kobylí to soil erosion. The rule-based modeling in geographical information systems (GIS) - MGE - is used. Parameters of relief (slope, aspect), soil (texture), geology (character of bedrock) and human utilization parameter (management and landuse) were taken into account. Relief parameters were calculated by using digital terrain model. Rules for classification of parameters and index overlay were based on general knowledge of soil erosion and reflect the conditions of the cadastre. Index map overlay was performed at three stages: 1. susceptibility to soil erosion based on relief parameters and soil-geology parameters, 2. susceptibility based on parameters of potential erosion (both relief and soil-geology), 3. susceptibility based on both potential and human utilization parameters.

Key words: soil erosion, GIS, digital terrain model, rule-based modeling, index map overlay

 

INTRODUCTION

Soil erosion process effects mainly the agriculture land with intensive utilization. Its extend increased in the last decades due to collectivization being carried out without respect to nature.

 

New technology of data collection and processing - especially geographic information systems and remote sensing - influence the methods of soil erosion research. The aim of this paper is presentation of the landscape evaluation method executed within GIS. Not only potential erosion is the object of research but also human utilization (mostly the main erosion factors) is taken into account.

 

SOFTWARE

Data processing was performed within GIS - MGE (Modular GIS Environment). The main modules - MGE/SX (MGE Base Mapper, MGE Base Administrator, MGE Base Nucleus) and advanced ones (MGE Terrain Analyst, MGE Analyst) were used. MicroStation (the graphics environment of MGE), IRASB, IRASC were applied for production of digital thematic maps.

 

STUDY AREA

Study area - cadastre of the community Kobyli (21 km2) - is located in the Trkmanka catchment in the Middlemoravian Carpathians. Intensive agricultural utilization is distinctive for this catchment - arable land, vineyards and orchards often situated on artificial terraces dominate. High susceptibility to soil erosion is caused also by the character of bedrock - loess, sandstone and claystone prevail.

 

DEFINITION OF SOIL EROSION

Total erosion is delimited as a function of potential erosion parameters (natural conditions) and human utilization parameters (crop management, erosion control practice). The natural conditions include morphological, soil-geological and climatic situation. Soil-geological are expressed by soil and bedrock susceptibility to erosion, morphological by slope, and aspect express through relief climatic condition. Human utilization parameters include susceptibility of vegetation cover, crop management, slope length and technical arrangements (fig.1).

 

 

 

 

A = f (K, G, S, E, C )

 

Fig. 1: Total susceptibility to soil erosion (A) as a function of potential erosion factors and human utilization factors (fig. ).

METHODOLOGY

The rule-based modeling (Meijerink in Solín, Lehotský, 1996) was used for evaluation of susceptibility to soil erosion.

Good knowledge of the relation between landscape components and soil erosion process is the substance of this method. The character of soil erosion process is one of the inputs into the model. Due to lack of the laboratory analysis this character was assessed only on the base of the general assumptions, field survey and the aerial photos interpretation.

Thematic and topographic maps are the other inputs into the model. These were used:

 



















 

 

 

 

 

 

 

 

 

 

 

Fig. 2: Structure diagram of rule based modeling for evaluation of susceptibility to soil erosion in GIS.

 

The map of landuse resulted from the field survey and interpretation of aerial photos.

Following digital maps were constructed:

 

Digital terrain model was created from contours (equidistance of 2 m). Clearing and complexing of contours was followed by tagging and import to the model. TIN model was chosen for better approximation of terrain and direct transfer of the analysis result into MGE project. Breaklines were automatically inferred to avoid generating of flat triangles.

Two relief parameters were assessed:

This background preparation - digital thematic map creation, field survey, aerial photos interpretation, DTM analyses and recapitulation of theoretic knowledge - was followed by the decision stage. On the base of the preparing stage it was necessary to chose parameters entering the model and to formulate decision rules.

These parameters were chosen:

 

Decision rules were formulated to determine erosion indexes of each parameter. Erosion indexes are as follows (tab.1).

Tab. 1: Erosion indexes of parameters entering modeling ( *1-the smallest, *4 - the biggest susceptibility to soil erosion)

param.

class, criterion

eros. index

texture

sand

tex1

 

loamy sand

tex2

 

sandy loam

tex2

 

loam

tex3

 

clay loam

tex3

 

clay

tex2

bedrock

fluvial, deluvio-fluvial and deluvial sediments

geo2

 

loess

geo3

 

sandy limestone

geo1

 

sandstone and claystone

geo2

slope

0 - 3o (plateau)

slo1

 

3 - 6o (moderate slope)

slo2

 

6 - 12o (middle slope)

slo3

 

12o + (distinct slope)

slo4

aspect

N

asp1

 

NE

asp1

 

E

asp2

 

SE

asp3

 

S

asp3

 

SW

asp3

 

W

asp2

 

NW

asp1

human utilization

mild susceptibility to soil erosion

- crop on flat position

- nonagricultural utilization

 

util1

 

middle susceptibility

- permanent crop on terraces

- contour cultivation

 

util2

 

high susceptibility to soil erosion

- slope-direction cultivation on middle slope

- slope-direction cultivation on long middle steep slopes

 

util3

 

extremely high susceptibility to soil erosion

- vineyards on distinct slope with slope-direction cultivation

- arable land on distinct slope

 

util4

 

Decision rules are applied on thematic maps and index maps are created:

 

Index maps enter overlay analyses. Decision rules define weight of influence of each parameter evaluating these overlays.

Index map overlays were performed in three stages:

I. Evaluation of susceptibility to soil erosion from the viewpoint of morphological (A) and soil-geology (B) parameters.

A: slope (slo) + aspect (asp) = relief (rel)

B: texture (tex) + bedrock (geo) = soil-geology (sg)

II. Evaluation of susceptibility to soil erosion from view point of natural condition (potential erosion)

soil-geology (sg) +relief (rel) = natural condition (fg)

III. Evaluation of the total susceptibility to soil erosion based on parameters of potential erosion and human utilization parameters.

natural condition (fg) + human utilization (util) = total erosion (erosion)

Index map overlays were evaluated according to following formula:

I = (x1 . v1) + (x2 . v2)

I - index overlay value

x1 - erosion index of the first parameter

x2 - erosion index of the second parameter

v1 - weight of influence of the first parameter

v2 - weight of influence of the second parameter

 

Association of weights with each parameter, evaluation of index map overlays and resultant index association is shown in table 2:

 

Tab. 2: Index map overlay valuation

stage of overlay

weights of param.

erosion index combin.

index overlay value (I)

result. index

characterization

I.A

tex 50%

geo50%

tex1 + geo1

tex1 + geo2

tex2 + geo1

2

3

3

sg1

favorable soil-geological conditions (minimum susceptibility to soil erosion)

   

tex1 + geo3

tex2 + geo2

tex 3 + geo1

4

4

4

sg2

rather favorable soil-geological conditions (moderate susceptibility to soil erosion)

   

tex2 + geo3

tex3 + geo2

5

5

sg3

rather unfavorable soil-geological conditions (middle susceptibility to soil erosion)

   

tex 3 + geo3

6

sg4

high unfavorable soil geological conditions (high susceptibility to soil erosion)

I.B

slo 66%

asp 34%

slo1

 

rel1

very favorable morphological conditions (minimum susceptibility to soil erosion)

   

slo2 + asp1

slo2 + asp2

1,66

2

rel2

favorable morphological conditions (moderate susceptibility to soil erosion)

   

slo2 + asp3

slo3 + asp1

2,34

2,32

rel3

rather unfavorable morphological conditions (middle susceptibility to soil erosion)

   

slo3 + asp2

slo4 + asp1

slo3 + asp3

2,66

2,98

3

rel4

unfavorable morphological conditions (high susceptibility so soil erosion)

   

slo4 + asp2

slo4 + asp3

3,32

3,66

rel5

high unfavorable morphological conditions (very high susceptibility so soil erosion)

II.

rel 70%

sg 30%

rel1+sg1

rel1+sg2

rel1+sg3

rel2+sg1

rel1+sg4

1

1,3

1,6

1,7

1,9

fg1

high favorable natural conditions (no susceptibility to soil erosion)

   

rel2+sg2

rel2+sg3

rel3+sg1

2

2,3

2,4

fg2

favorable natural conditions (moderate susceptibility to soil erosion)

   

rel2+sg4

rel3+sg2

2,6

2,7

fg3

rather unfavorable natural conditions (middle susceptibility to soil erosion)

   

rel3+sg3

rel4+sg1

rel3+sg4

rel4+sg2

3

3,1

3,3

3,4

fg4

unfavorable natural conditions (high susceptibility to soil erosion)

   

rel4+sg3

rel5+sg1

3,7

3,8

fg5

high unfavorable natural conditions (very high susceptibility to soil erosion)

   

rel4+sg4

rel5+sg2

rel5+sg3

rel5+sg4

4

4,1

4,4

4,7

fg6

extremely unfavorable natural conditions (maximum susceptibility to soil erosion)

 

 

 

Objective evaluation of the human utilization parameter was impossible and so for the third stage of overlays the weights were not valuated. Only four resulting indexes were determined (table 3).

 

Tab. 3: Evaluation of index map overlay for total erosion

stage of overlay

erosion index combination

resultant index

characterization

III.

h1+FG1, h1+FG2, h1+FG3, h1+FG4, h1+FG5, h1+FG6, h2+FG1, h3+FG1, h4+FG1, h2+FG2

erosion1

minimum susceptibility to soil erosion

 

h2+FG3, h2+FG4, h2+FG5, h3+FG2

erosion2

moderate to middle susceptibility to soil erosion

 

h2+FG6, h3+FG3, h3+FG4, h4+FG2

erosion3

high susceptibility to soil erosion

 

h3+FG5, h3+FG6, h4+FG3, h4+FG4, h4+FG5, h4+FG6

erosion4

extremely high susceptibility to soil erosion

 

 

Tools of MGE Analyst were used to perform index map overlays. Topo files were created over index maps entering each overlay and spatial querying followed (table 4). Query results were transferred to graphic and following synthetics maps were created:

 

 

 

Tab. 4: Spatial querying structure

query name

structure of query

slo1

slope where mini=0

slo2

slope where mini=3

slo3

slope where mini=6 or 9

slo4

slope where mini 9

asp1

aspect where mini=22 or 292 or 337

asp2

aspect where mini=67 or 247

asp3

aspect where mini=112 or 157 or 202

tex1

texture where erozní_index=1

tex2

texture where erosion_index=2

tex3

texture where erosion_index=3

geo1

bedrock where erosion_index=1

geo2

bedrock where erosion_index=2

geo3

bedrock where erosion_index=3

rel1

slo1

rel2

(slo2 intersect asp1) union (slo2 intersect asp2)

rel3

(slo2 intersect asp3) union (slo3 intersect asp1)

rel4

(slo3 intersect asp2) union (slo4 intersect asp1)

rel5

(slo3 intersect asp3) union (slo4 intersect asp2) union (slo4 intersect asp3)

sg1

(tex1 intersect geo2) union (tex2 intersect geo1)

sg2

(tex1 intersect geo3) union (tex2 intersect geo2) union (tex3 intersect geo1)

sg3

(tex2 intersect geo3) union (tex3 intersect geo2)

sg4

tex3 intersect geo3

FG1

(rel1 intersect sg1) union (rel1 intersect sg2) union (rel1 intersect sg3) union (rel1 intersect sg4) union (rel2 intersect sg1)

FG2

(rel2 intersect sg2) union (rel2 intersect sg3) union (rel3 intersect sg1)

FG3

(rel2 intersect sg4) union (rel3 intersect sg2)

FG4

(rel3 intersect sg3) union (rel4 intersect sg1) union (rel3 intersect sg4) union (rel4 intersect sg2)

FG5

(rel4 intersect sg3) union (rel5 intersect sg1)

FG6

(rel4 intersect sg4) union (rel5 intersect sg2) union (rel5 intersect sg3) union (rel5 intersect sg4)

util1

utilization where erosion_index = 1

util2

utilization where erosion_index = 2

util3

utilization where erosion_index = 3

util4

utilization where erosion_index = 4

erosion1

(util1 intersect FG1) union (util1 intersect FG2) union (util1 intersect FG3) union (util1 intersect FG4) union (util1 intersect FG5) union (util1 intersect FG6) union (util2 intersect FG1) union (util2 intersect FG2) union (util3 intersect FG1) union (util4 intersect FG1)

erosion2

(h2 intersect FG3) union (h2 intersect FG4) union (h2 intersect FG5) union (h3 intersect FG2)

erosion3

(h2 intersect FG6) union (h3 intersect FG3) union (h3 intersect FG4) union (h4 intersect FG2)

erosion4

(h3 intersect FG6) union (h3 intersect FG6) union (h4 intersect FG3) union (h4 intersect FG4) union (h4 intersect FG5) union (h4 intersect FG6)

 

 

 

 

 

SUMMARY AND DISCUSION

Universal soil loss equitation (USLE) (Kirkby, Morgan, 1980) is still the most used method in soil erosion research. It brings the total amount of eroded material (absolute value) on the spot. The method discussed (rule-based modeling of soil assessment using GIS) enables to determine relative values of erosion in the study area.

Map of susceptibility of natural conditions to soil erosion (Map7) shows the spatial variation of factors positively affecting surface runoff and contributing soil erosion process. In this way it determines natural limits which should be respected in the landuse planning.

Map of total susceptibility to soil erosion (Map 8) evaluate human activity in the relationship to natural condition. It points out areas where landuse is in the contrast with the natural landscape character and where the antierosion arrangements should be of the main importance.

The results of the model are applicable in the regional scale. Precision of the model reflects the quality of input data. The better quality and more input data, the better precision of the model.

This study is about to show the method of soil erosion research. Commonly accessible data are processed. To get the better results by using this method I suggest following adjustment of input data:

 

REFERENCES: