In: Wagner W., Székely, B. (eds.): ISPRS TC VII Symposium - 100 Years ISPRS, Vienna, Austria, July 5-7, 2010, IAPRS, Vol. XXXVIII, Part 7B
located in arid areas of Iran. Average annual rainfall in the
study area is 325.38 mm. The significant amount of rainfall
happens during the winter in which agriculture fields and
rangeland vegetation depends. Drought which usually occurs in
this area causes many problems for its ecosystem.
2.2 Image and Metrological Data
The satellite data used in this research consist of ETM+ images
belonging to 19th August 1999 and 10th July 2002. The
meteorological data (monthly rainfall) were collected from 18
rain-gauge stations which have 21 years common statistical
basis. Table 1 present the annual rainfall of the study area for
the considered years. As can be observed from the table, the
rainfall increased considerably in 2002 compared to previous
years.
Year
Annual (mm)
1999
405.5
2000
134
2001
242
2002
458.97
Table 1. The annual rainfall of the study area.
3. METHODOLOGY
3.1 Image Pre-processing
Since the selected images had different acquisition date, sun
angle correction was applied to remove the differences caused
by sun. Atmospheric correction was then performed using
FLASH algorithm in order to obtain the correct reflectance.
After the radiometric pre-processing, the images were geo-
referenced using topographical maps in scale of 1/50000 with
RMSEs equal to 0.47 and 0.2, respectively.
Since the main objective of this study was to assess drought
using different indices, digital numbers recorded by the sensor
were converted to the spectral radiance and reflectance using
gain and offset parameters provided in the header file of the
images.
3.2 Drought Severity by Meteorological Data
Several drought indices based on meteorological data have been
introduced by researchers. This includes Palmer Drought
Severity Index (PDSI), Standardized Precipitation Index (SPI),
Crop Moisture Index (CMI), Reclamation Drought Index (RDI)
and etc. each having advantages and disadvantages. In the
present research, Run-Test method was used because of its
simplicity and also because it only requires annual rainfall. It
can be expressed by the equation (1):
h
p-Jf 0 <0 => Dry
(X-X 0 >0=> Wet
Where X = annual rainfall
P" = average of annual rainfalls
The following parameters can be calculating using this method:
1. Drought duration: the number of consecutive years which
drought occurs.
2. Drought magnitude: the total of X- X 0 amounts in each
period.
3. Drought intensity: the average of X- X 0 amounts in each
period.
4. Drought severity: the maximum amount of X- X 0 in each
period.
3.3 PDI and MPDI Indices
Ghulam et al, in 2006, offered a new index based on spectral
characteristics of surface in red and near infrared spectral space.
As it can be seen from Figure 1, the AD line represents the
changes in surface vegetation from full cover (A) to partial
cover (E) to bare soil (D), while BC refers to an area with a soil
moisture status described as wet (B), drier (D) and extremely
dry (C) (Ghulam, 2006).
Figure 1. NIR-Red space and PDI (Ghulam, 2006)
The soil line is a linear relationship between NIR and Red
reflectance of bare soil (Richardson, 1977). In this paper, in
order to obtain the soil line parameters (slope and intercept),
about 500 pixels of different types of the bare soils were
extracted and were plotted in the feature space of Red-NIR.
Here are the soil line parameters for each image:
R nir = 1.19 R Red + 0.001 (1999 image)
/?nir = 1-19 R Red + 0.003 (2002 image)
PDI can be calculating using the following equation:
PDI = —— (%d+K R m ) (2)
V oc 2 +l
where oc = slope of soil line
R Red and R NIR refer to the atmospherically corrected
reflectance of the Red and NIR bands, respectively (Ghulam,
2006).
Crop growth is directly related to the soil moisture. Where soil
moisture is below a certain level, crops cannot absorb enough
water from the soil and are exposed to drought. Consequently,
the soil moisture is the main factor in remote monitoring of
drought. Soil spectral reflectance decreases with increasing soil
moisture, Therefore, the severity of a drought can be estimated
by the close relationship between soil moisture and soil spectral
reflectance. However, the spectrum received by the sensor is a
mixture reflected or emitted information from different surface
targets. Therefore, both the soil moisture status and the
vegetation status are very important in drought monitoring.