An application of GIS and Remote Sensing for Analysis of Agricultural Development-Induced Changes in Land Use: A case study in Lao PDR pdf

An application of GIS and Remote Sensing for Analysis of
Agricultural Development-Induced Changes in Land Use:
A case study in Lao PDR

By
Boundeth Southavilay1), 2), Teruaki Nanseki 3)

The 30th APAN Meeting
August 2010, in Hanoi, Vietnam
1). Graduate school of Bioresource and Bioenvironmental Science, Kyushu University, Japan.
2).Department of Planning, Ministry of Agriculture and Forestry, Lao PDR.
3). Faculty of Agriculture, Kyushu University, Japan
E-Mails: ;

1


Contents
1.

Introduction

2.

Statement of problems

3.

Objectives

4.

Materials and Methods

5.

Study area

6.

Results and Discussions

7.

Conclusions
2


Introduction
•

In the last decade, in Laos GIS and Remote sensing (RS) has not much used in
countrywide, including agriculture sector did not applied this technique for their
agricultural development and land use planning.

•

A meanwhile, in that times agriculture lands in Laos were transformed from subsistence
farming in uneconomic-sized farms to commercial and market-oriented farms. These
transformed sometimes happens in improperly ways and induced to change in land use
and land covers by despoilment of forest covers and traditional farming system.

•

The problems above due to lack of an appropriate tool in terms of integrated spatial data
on land use/land covers. However, recently GIS and remote sensing has been using in
several types of works in both government and private agencies. As we know, GIS and
remote sensing have an important role in linkage and analysis of such data, in particular
for detection, interpretation, area calculation, monitoring and future estimating. Therefore,
this study applied GIS and remote sensing for analysis the land use pattern changes

3


Statement of Problems
• After 1999, the landscape in the study area has been
changed cause of policy implementation such as rubber
plantation and irrigation system were installed in the area,
and than this place was changed in dynamics way of land
use system
• The forest area was destroyed by increasingly shifting
cultivation and rubber plantation areas
• Lack of an appropriate tool for decision support system in
terms of land use decision
Rubber plantation

Shifting cultivation

4

Objectives
• To illustrate the change detection of land use and land
covers
• To create a tool for decision support system in the
watershed land use planning by created land zoning.

5


Materials and Methods

6


Materials
•

•

•
•

Satellite images:
– Landsat ETM+ (25 January 1999), LIG
format
• Resolution
– 30 m (band 1-5,7)
– 15 m panchromatic
– Landsat ETM+ (12 March 2004), BIL
format

• Resolution
– 25 m (band 1-5,7)
– 15 m panchromatic
GIS data bases with thematic maps (Road
networks, River networks, Village points,
Contour line, DEM and Ground check pointfrom GPS)
Topography map 1:100,000
(Schema F-47-142 and F-47-130)
Software: ArcView3.2a and Idrisi 32
7


Methods
1.

Geometric correction- to georeference maps to a map
coordination system
–

–

2.

Change pixels size- because the pixel sizes of two images
are different (30m and 25m)
–

3.

Image 1999 was registered to local topography maps with 15

ground control points. root-mean-square (RMC) error = 0.45
pixels.
Image 2004 was registered with registered of image 1999
(image to image). RMC= 0.14 pixels.

Change pixel 30m of image 1999 to 25m of image 2004

NDVI compositing utility
– NDVI is useful for identifying of the green leaf from
other objects (water, soil…) It is expressed value -1 to
1 with 0 representing non vegetation
–
NDVI solve the shadow problem

NDVI image

NDVI= (b4-b3)/(b4+b3)
8


Methods (cont.)
•

Images interpretation by Supervised classification

Training area (AOI)
•

Supervised classification
– Maximum likelihood method

The training area from two images 345,
and 2ndvi7 in the1999 and 2004
– classified to 11 classes

•

9

Create zone by overlaid three physical data (Ground data, GIS data and image
classification)


The area is located in the northern Laos,
Lat: 65º07'16" to 67º59'13"
Long: 222º79'96" to 225º56'22".
43 villages

Study area

1250msl

300msl

•

•
Area 696 km2
•
Watershed boundary area = 22 km2

•
Elevation from 300 to 1,235 msl
10
The lowland farms are located between 300 to 450 msl.


Results and Discussions
•

The result of interpretation of two images ETM+1999 and ETM+2004, it provided two
land use maps of 1999 and 2004. In each map was classified into 11 categories of land
use/land cover types

Land covers 1999

Land covers 2004

Intensive of changed areas
11


1999
land use classes

2004

Km2

%

Km2

%

Dark evergreen forest

199.54

28.66

148.7

21.4

Bright evergreen forest

173.94

24.99

134.8

19.4

Disturbed forest/fallow

164.27

23.60

305.1

43.8

Bamboo

22.98

3.30

12.4

1.8

Field crop

32.53

4.67

4.0

0.6

Wet paddy

24.08

3.46

23.6

3.4

0

0

10.5

1.5

22.56

3.24

0

0

Reservoir

0.02

0.00

6.6

0.9

Mekong

6.62

0.95

3.1

0.4

Sandy area

1.70

0.24

4.1

0.6

47.92

6.88

43.2

6.2

696.2

100.00

696.2

Results of Maximum Likelihood
Classification of two images
1999 and 2004

100.0

Irrigation paddy
Bare land/Wet soil

Shrub/other crops
Total

Changes
land use classes

Changed

Dark evergreen forest
Bright evergreen forest
Disturbed forest/fallow
Bamboo
Field crop
Wet paddy
Irrigation paddy
Bare land/Wet soil
Reservoir

Mekong
Sandy area
Shrub/other crops

Km2

Percentage (%)

-50.84
-39.14
140.83
-10.58
-28.53
-0.48
10.5
-22.56
6.6
-3.52
2.4
-4.72

-25.48
-22.50
85.73
-46.04
-87.70
-1.99
-100.00
-53.17
141.18

-9.85

Change rate
(%km2/year)
-5.10
-4.50
17.15
-9.21
-17.54
-0.40
0.00
-20.00
0.00
-10.63
28.24

12

-1.97


Cheng detection
•

•

The change detection is the process of identifying differences in the
state of an object or phenomenon by observing it at different times
(Singh, 1989).
The change detection of land use and land cover of the study area

was analyzed by cross-classification technique– by overlaid of two land use maps

2004

1999

+

=

=

The change detection provides the characteristic changes of each land use type
13


Legend:
DT‐ Disturbed forest/fallow
EF‐ Evergreen forest
FC‐ Field crop
BB‐ Bamboo
BL‐Bare land
WPD‐ Wet paddy
IPD‐ Irrigation paddy
RV‐ Reservoir

14


detection of land use/land cover by crossclassification during 1999 to 2004

No

Land use change 1999-2004

Pixels

Hectares

Km2

%

1

Disturbed > Evergreen Forest

42676

2,667.25

26.67

3.83

2

Evergreen Forest > Disturbed

215062

13,441.38

134.41

19.31

3

Field crop > Disturbed

35353

2,209.56

22.10

3.17

4

Bare land > Disturbed

38975

2,435.94

24.36

3.50

5

Disturbed > Field crop

867

54.19

0.54

0.08

6

Wet paddy > Field crop

932

58.25

0.58

0.08

7

Disturb > Wet paddy

5504

344.00

3.44

0.49

8

Bare land > Wet paddy

7702

481.38

4.81

0.69

9

Disturbed > Irrigation paddy

5883

367.69

3.68

0.53

10

Field crop > Irrigation paddy

476

29.75

0.30

0.04

11

Wet paddy > Irrigation paddy

5398

337.38

3.37

0.48

12

Disturbed > Reservoir

2627

164.19

1.64

0.24

13

Wet paddy > Reservoir

1941

121.31

1.21

0.17

14

Evergreen Forest > Bare land

3297

206.06

2.06

0.30

15

Disturbed > Bare land

20361

1,272.56

12.73

1.83

16

No changes

726786

45,424.13

454.24

During 5 years 3
land use types
were changed to
shifting cultivation:
18,100ha

65.25

Total

1113840

69615

696.15

100.00

15


Land use changed in watershed boundary
May,1999

March, 2004

Dense forest
Open forest
Shifting cultivation
Bamboo
Field crop
Wet paddy
Irrigated paddy field
Reservoir
Shrub land/other

Dense forest
Open forest

Shifting cultivation
Bamboo
Field crop
Wet paddy
Bare land
Shrub land/other

Hectares

14000

Irrigated paddy (dry season)

12000
Area (ha)

1999

Land use types

Land use/land cover changes from 1999 to 2004

Reservoir

10000

Bare land/wet soil

8000

Field crop

6000

2004
%

Hectares

%

202.44

0.92

0.00

0.00

467.82

2.14

551.38

2.51

0.00

0.00

830.69

3.79

46.44

0.21

831.81

3.79

356.69

1.63

Wet paddy (rainy season)

2000

0.00

Bamboo

4000

0.00

952.56

4.34

740.25

3.37

Shrub/other crops

0
Year

2

Dense forest

Mix-deciduous forest

Bamboo

Shifting cultivation

Field crop

Wet paddy

Irrigation paddy

Bare land/wet soil

Reservoir

1

Shrub/other crops

1596.69

7.28

1524.56

6.95

Mixed-deciduous forest

5286.25

24.10

2827.38

12.8

Dense forest

5813.31

26.50

4112.13

18.7

Shifting cultivation

6061.94

27.64

11657.81

53.1

21935.50

100.0

21935.5

100.0

Total areas

16


Zonation
Ground information 
Combine land use 

type/land holding 
in the villages
Population and 
village location

GIS data
Raster maps: Slope, 
DEM
Vector maps: river, 
boundary

Watershed Zonation 
map 

Remote Sensing
Composite/NDVI 
maps 
Land use/land 
cover

Decision Support 
Land use planning 

• The zone was created by overlaid of three physical information (Ground data,
GIS data and satellite imagery data)
• The zonation can be regarded as a tool for sustainable agricultural
development in the watershed area.
17

Suggestion zones for sustainable of watershed management
Buffer zone

Conservation zone

This zone is designed to link
between development and
conversation zones. Land use
option: field crop, fruit tree and
commercial tree
Area covered 25%

The main purpose of this zone is to
protect forest covers because this
zone included headwater, district
protected area and high forest cover
and biodiversity.
Area covered 43%

Development zone 

The development zone includes
integrated farming systems and more
commonly associated with upland
areas. Located along the river banks
and foothills.
Land use options: paddy, fish ponds,
rice/fish, terraced paddy, grazing,
field crops, fruit trees, commercial
tree.

Area covered: 32%

18


The optional land use for sustainable agricultural development in the
watershed area

x=Restricted potential, xx=Medium potential, xxx=High potential, o=not considered appropriate.
19


The existing of land cover in the study area

Development zone

Conservation zone
Conservation zone

Buffer zone


Conservation zone

Buffer zone

Development zone

21

Conclusions
For 1st objective
1.The total areas of the fallow forest were doubly increased from 16,427
ha in 1999 to 30,510 ha in 2004. This land use type was high potential to
be mixed by different types of land use such as disturbed forest/fallow
forest/shifting cultivation/rubber plantation.
2.Based on our results suggest that after irrigation dam was constructed,
several types of land use areas were changed (decreased/increased) and
fragmented (field crop, evergreen forest, fallow forest) as a result of both
farmers who lost their lands and turned to clear-cut forest areas for upland
rice cultivation, and private investment on commercial tree (rubber
plantation).

22


Conclusions (cont.)
For 2nd objective
1.The tool for decision support system in sustainable of agricultural
development in this study is the land use zoning was created by GIS
and remote sensing technique. By created 3 main land use zones.
1. Conservation zone
2. Buffer zone
3. Development (Agricultural) zone

2.These zones can be the most important tools for agricultural
development planning because it provided integrated information on
social and physical aspects of the study area.

23


Thank you very much for your
kind attention

24