An Evaluation of Drought Risk Area in Northeast Thailand using Remotely Sensed Data and GIS

3. Methodology
The underlining concept of the paper is that the severity of drought is a function of rainfall, hydrology and physical aspect of landscape. Each of which is a thematic data layer that is a result of a analysis of sub layers or individual layer. The three thematic data layers are then combined to ultimately formulate a composite output layer or drought risk area. The schematic chart of this analysis is shown (Fig 2).

3.1 Rain analysis
Daily rainfall data from 264 stations across northeast Thailand covering at least 15 years record were compiled as point database for analysis. Calculation of mean annual rainfall was made at each station. Kriging interpolation was performed to establish spatial mean annual rainfall across the entire northeast Thailand. The threshold of mean annual rainfall was then identified using decile range to create the decile rainfall.



The decile rainfall of 10 classes was grouped into 4 levels of drought severity that so called meteorological drought.

3.2 Hydrological analysis
Input data layers for hydrological drought include surface water source, irrigated area, stream density in subwatershed and ground water yield and quality. Surface water source and irrigated area were derived from 14 scenes of Landsat TM acquired during December 1998-February 1999 and data collected by the site project of the Royal Irrigation Department respectively. Stream density was calculated from stream length within subwatershed area. Ground water yield and quality at scale of 1:100,000 is available from the Department of Mineral Resources. Each input map layer was digitally encoded in GIS database. To generate the hydrological drought the matrix analysis was performed for every coincidence of classes in two layers (Fig 2). The variables class and severity of drought are given in table 1.

3.3 Physical landscape Analysis
Land form, drainage condition and land use are input variables for analysis of physical drought. Preparation of land form and land use maps were done using the Landsat TM acquired as above. Drainage condition map was compiled from soil map 1:50,000 of Land Development Department. Digital map of each layer was performed using GIS functions. The digital output of the physical drought is a combination of 3 input variables. The analysis and procedure were executed the same as those performed for the hydrological drought. (Fig 2 and table 1)

Table 1 variables,class and drought severity

Variables	Class	Drought severity
Mean annual rainfall (Decile range)	0 - 1130.065 mm.	4
	> 1130.065 - 1276.133 mm	3
	> 1276.133 - 1431.076 mm.	2
	> 1431.076	1
Irrigated area and water source	Water source 0 - 0.5 km2
	Area beyond water source >0.5 km.	4
	Area beyond water source >0.12 - 0.5 km.	2
	Area beyond water source 0 - 0.12 km.	1
	Water source 0.5 - 5 km2
	Area beyond water source >1 km.	4
	Area beyond water source >0.25 – 1 km.	2
	Area beyond water source 0 - 0.25 km.	1
	Water source 5 –10 km2
	Area beyond water source >1.5 km.	4
	Area beyond water source >0.5 – 1.5 km.	2
	Area beyond water source 0 - 0.5 km.	1
	Water source > 10 km2
	Area beyond water source >3 km.	4
	Area beyond water source >1 - 3 km.	2
	Area beyond water source 0 - 1 km.	1
	Area within irrigated area	1
Groundwater yield and TDS.	3 m2 / hr. & TDS. > 1500 mg/l	4
	2 - 10 m2 / hr. & TDS. > 750 - 1500 mg/l	3
	10 –20 m2 / hr. & TDS. <750 mg/l	2
	>20 m2 / hr. & TDS. <750 mg/l	1
Stream density	0 – 120.98 m/ km2	4
	120.99 – 248.17 m/ km2	3
	248.18 – 406.89 m/ km2	2
	>406.89 m/ km2	1
Land form	Mountainous	4
	Dissected erosion surface	3.5
	High terrace	3
	Middle terrace	2.5
	Low Terrace	2
	Flood plain	1
Drainage condition	Excessively drained	4
	Well drained	3.5
	Moderately drained	3
	Somewhat poorly drained	2.5
	Poorly drained	2
	Very poorly drained	1
Land use	Field crop / Deciduous forest / Village	4
	Mixed field crop / Forest and mixed crop	3.5
	Grass land / Shrub / non-use	3
	Mixed paddy / Mixed ever green forest	2.5
	Tree / Fruit tree / Swamp and other	2
	Paddy / Mixed fruit tree / ever green forest	1.5
	Water source / Riparian / Swamp	1

3.4 Drought risk area
The resulting thematic data layers were then combined by matrix analysis to produce the drought risk area. The drought risk area represents the integration of meteorological drought, hydrological drought and physical drought which in turn are assigned according to the drought criteria studied.

4. Results and Discussion
Result of the overall drought risk area are shown in Fig 3d and has 4 classes : very mild, mild, moderate and severe. The severe drought area covers mainly in the southwest of region and has approximately 11.20% of the northeast. The moderate drought area extends around those of severe class and has about 32.06%. The some of very mild areas have occurred in the southwest region where water source and irrigated area was developed. The very mild and mild drought areas are found extensively in the eastern and northeastern part of the study area, particularly along the Mekong river. The areas occupied by the different kinds of the droughts are summerized in table 2.