Multi-Temporal Satellite Images and Ground Truth Data Period (1976-2010)
New Challenge about Climate Change or Human Impact
Mohsen Ahadnejad Reveshty
Assistance Professor, Dept. of Geography, Zanjan University, Iran
ahadnejad@gmail.com
Yoshihisa Maruyama
Associated Professor, Chiba University, Japan
ymaruyam@tu.chiba-u.ac.jp
Abstract :
Lake of Urmia is the largest saline lake inside Iran and the second saline lake in the world after
Dead Sea, that average area about 5000 km2 is located in northwestern South Azerbaijan (Iranian Azerbaijan) .
Urmia or (Turkish Language: اورمو, Urmu, Orumiyeh, Urmiye, Urmiya) is a city in Northwestern South Azerbaijan (Iranian Azerbaijan) and the capital of West Azerbaijan Province. The city lies on an altitude of 1,330 m above sea level on the Shahar Chaye river (City River). Urmia is the 10th populated city in iran and 2nd of Azerbaijanian Turks provinces after Tabriz. Urmia is the trade center for a fertile agricultural region where fruit (Specially Apple and Grape) and Tobacco are grown. An important town by the 9th cent. Urmia was seized by the Oghuz Turks (11th cent.), sacked by the Seljuk Turks (1184), and later occupied a number of times by the Ottoman Turks.
The name Urmia or Urmu is thought to have come from Sumerian tongue, the earliest known civilization in the world located in southern Mesopotamia. Ur was a principle Sumerian city. Urmia, situated by a lake and surrounded by rivers, would be the cradle of water. The population of Urmia is predominantly Azerbaijanian Turks (over 90%), but with Kurdish,Assyrian and Armenian minorities.
In recent years this lake
levels affected natural and human factors, including successive droughts in the region and the
construction of dams and the indiscriminate exploitation of water resources in the Basin .The Lake
surface changes that severe fluctuations from 5278 km2 in 1976 has been reached to about 3107.7 km2 in
2009 .
In this paper, using multi-temporal satellite images, including MODIS, ETM+, TM, MSS images,
fluctuations assessment in Urmia Lake in 1976-2009 with using the 18 image series in August and also
using ground truth data methods from water level Lake of Urmia has been studied.
In order to probable predict changes in the lake in the coming years with regard to continuous
fluctuations occurred at this stage, Markov chain and cellular automata methods were used .Based on
results probable survival value for the lake in next ten years has been estimated about 64 percent .Also for
evaluation of role each of the natural factors, including climate change and human impact as major
challenges discussed in this paper was investigated.
Key word: Urmia Lake, Fluctuations, GIS, Satellite imagery, Markov Chain
Introduction
Features and phenomena in the Earth's surface were changed due to over time, the lakes as
one of these phenomena and due to having a closed environment is not exception and due to
climatic changes such as reduced rainfall and increased temperature and uncontrolled use
of surface water resources in watershed areas in agriculture, industrial and drinking ever level
they are exposed to change .Supervision and monitoring changes in these lakes should be
considered as important in the national and regional development and natural resource
management .Currently monitoring the coastal areas and extraction of water level changes at
different intervals is an infrastructure research of interest because the coastal zone management
and dynamic nature of such sensitive ecological environments need to accurate information
about the various intervals(Rasoli,2007). Among the remote sensing data are considered as
useful tools for the continuously monitoring and sequentially compared with traditional methods .
With regard to temporal resolution from half days to one month, and spatial resolution of less
than one meter to several kilometers and multi-spectral resolution of this data and applying
mathematical and statistical methods to detection of changes, the satellite image has become as a
valuable resource for earth sciences specialist for studying earth surface and its changing
(Ahadnejad, 2010)
In the field of application satellite images to monitoring of lakes and lagoon surface
changes much research that is most important, they note :
Ahadnejad et al(2010), in paper entitled “Detecting and Environmental Assessment
of Spatial Changes of Hamun-E-Saberi Lagoon Using Satellite Imagery and GIS studied these
lagoon in the period of 1976-2008 using LANDSAT and MODIS satellite images and analyzes
them with utilizing the Normalized Difference Water Index (NDWI) during the August months,
to assess and evaluate its spatial variations .Al Sheikh et al (2007), in article entitled "coastline
change detection using remote sensing "study changes in coastline Urmia Lake during 1989,
1998 and 2001 and paid to utilizing Landsat satellite images and processing them Coastline
change detection is about Urmia Lake .Ma and Wan (2007),"change in area of Ebinur Lake
during the 1998-2005", they used indicators such as NDWI for detection of water level changes
in this lake .Rasoli et al (2007), in paper “monitoring of Urmia Lake Water level fluctuations
using multi-temporal satellite images processing .Qulin TAN et al (2004), in paper entitled "
measuring Lake water level using multi-source remote sensing combined with hydrological
statistical data for changing Poyang Lake in China and etc.
In this paper using multi-temporal satellite images such as MSS, TM, MODIS data and
using Normalized Difference Water Index (NDWI), firstly occurred changes detected in Urmia
Lake and then using data such as water level measured in ground stations and the amount of
rainfall and water input to the lake to the trend of modeling with integrated remote sensing data
and ground truth data and ultimately Urmia Lake drying reasons will be discussed in recent
years.
Study Area
Urmia Lake as the largest water body in Iranian plateau is located between two major
provinces of East Azerbaijan and west Azerbaijan .The lake is bounded between 37°5´ -38°16´
latitudes and 45°01´ -46° longitudes at 1275 m above sea level .Its surface area ranges from
4750 to 6100 km2 and the average and greatest depths account for 6 and 16 m, respectively
(Azari Takami, 1993) .More than 20 permanent and seasonal rivers as well as a few submarine
streams and springs feed the lake .Average salinity of the lake ranges between 220-300 mg/lit
depending upon temporal and spatial conditions, in recently years it arrived more than 380
mg/lit. Due to the ecological heritage of Urmia Lake it is recorded as a protected habitat in the
world by the United Nations.
Material and methods
-Material
The data used in this paper refer to August month that acquired from Landsat and Terra
satellite sensors data. Table and figure 1 shows characteristic of data used in this paper.
Table1 :The characteristic of data used in this paper
Also in this paper ground truth data such as daily water level data that measured in during
1976-2009 by east and west Azerbaijan water organizations in ground station at Sharaf khaneh
and Golmankhaneh ports .Statistics related to rainfall and water volume input to the lake are
other data that used in this paper .Table 2 shows summarized data used in this article.
Fig1: Satellite image of Lake Urmia in during 1976 - 2009
Methods
-Image processing :
There are many methods for detecting of changes with using satellite images such as
subtraction images, and ratio and difference method, supervised classification, vector change
analysis (VCA), indices and normalized difference ...mentioned.
For detecting of occurred changes in this study satellite images of the area and available
resources, including U.S .Geological Survey were collected .After the initial corrections such as
geometric and radiometric correction changes detection of water level changes has been applied .
Since the separation of water bodies on satellite imagery is done carefully and high
accuracy in compared with other phenomena in the earth surface .In this paper for separation and
detection of water from other phenomena, normalized difference water index were used .In this
index by using near and middle infrared bands in the TM and ETM sensors, green and near
infrared bands in MSS sensor and middle infrared and short wave in MODIS sensor and
applying ratio and difference method water bodies has been separated from other phenomena's in
case study area .The equation number 1 to 3 show normalized difference water index for satellite
data are used in this paper.
Based on normalized difference water index images produced by this index value for water
levels towards desire to +1 value and for other surface without water towards desire -1 value.
- Trend Analaysis
The other object of this paper is to predict the trend of land use changes in the future .Many
methods can be applied to predict the trend .In this paper, two methods are used. Fig3 shows
trend change map of Urmia Lake in during 1976-2009.
(1) Markov chain
The Markov chain method analyzes a pair of water classification images and outputs a
transition probability matrix, a transition area matrix, and a set of conditional probability images .
The transition probability matrix shows the probability that one class will change to the others .
The transition area matrix tells the number of pixels that are expected to change from one class
to the others over the specified period .
The conditional probability images illustrate the probability that each class type would be
found after a specific time passes. These images are calculated as projections from the two input
land cover images .The output conditional probability images can be used as direct input for
specification of the prior probabilities in Maximum Likelihood Classification of remotely sensed
imagery (such as with the MAXLIKE and BAYCLASS modules) .A raster group file is also
created listing all the conditional probability images .
In this study, a series of image processing was performed to predict the trend of Urmia
Lake change in 2019 .
Fig2 :Images resulted from NDWI reclassify for separated land from water (1976-2009)
(2) Combination of Cellular Automata and Markov Chain
To know the changes that have occurred in the past may help to predict future changes .
Combination of Cellular Automata and Markov Chain is often employed to predict Urmia Lake
change estimation .
In order to predict the trends of Lake Changes, first 1976 and 2009 Lake Map was
analyzed with Markov Chain .Then, combined method of Cellular Automata and Markov Chain
was used for forecasting land use change in 2019 .According to the results Urmia Lake areas
decrease from 3107.78 Km2 in 2009 to 2095.44 km2 in 2019. Fig4 shows predicted map
The results of satellite images processing show that most changes occurred in the southern
and eastern part of lake that indicates the water depth is low in these areas compared with other
areas of the lake .The lowest lake retreat occurred in the north and northwest of lake. However
high the river water from flowing into the area to the lake but not much depth of water in these
areas has caused a retreat in this section are vertically regions and less in these regions compared
with southern parts of East coverts to salty land .
Notable in recent years especially in 2008 and 2009 connecting the Aspire and Ashk
islands in the middle part of Urmia Lake has caused this intensification and increasing areas of
salt in this area .The resulting map method based on Markov chain and the Cellular Automata
with the likely trend of the islands of the East Lake are connected to the land where the eastern
and southern areas of the lake completely dry and this can be associated irreparable
environmental effects.
Change trends analysis using ground truth data and image processing
Based on existing data in Table 2 can be realized that Urmia Lake long-term average
water level in the periods 1976-1998 about 1276.042 m above sea level, except in 1998 than the
long-term average of about 0.365 m high in the rest of the years. From 1999 to 2009 the lake
water level has fallen and garlic to the long-term average of about 4.898 meters has decreased.
Based on the predictions done based on time series method if this trend continues to be in the
lake water level in 2019 decreased to 1267 meters and this will mean that the level of the lake
level trend.
Reducing the lake water level will be reduced lake area. Especially in the southern half and
eastern parts of the lake that available evidence shows to be shallow in these areas than the
northern half of the lake. According to the results obtained from satellite image processing in the
long term average lake area of about 5277 sq. km area is that from 1999 to 2009 had reduced the
garlic so the lake area in 2009 reached approximately 3107 sq. km with average long-term
reduction of about 2119 sq. km. Based on the analysis carried out using the Markov chains and
Cellular Automata analysis and time series until 2019 this trend with regard to the lake area
decreased by approximately 2000 square kilometers. Figure 4 shows predicted changes in 2019
between water level and area in Urmia Lake.
Fig7: The comparison plot between Water Level and Area (KM2)
Data of rainfall in Table 2 shows that the long-term average rainfall in the Basin of
Urmia Lake is about 281 mm. During 1998 to 2001 for three consecutive years the amount of
rainfall markedly decreased in the years 1998-1999 and reaches about 165 mm. this decreasing
in rainfall is starting point in Lake water level reductions. Because of concern that has caused
droughts in dams after this year will be built or existing dams will be save water. During after
2001 significantly on the amount of rainfall in Urmia Lake basin been increased and many
long-term average of these years has been even higher. Then with consider to statistics such as
rainfall, climate change has been not considered only factor in Urmia lake water level
reductions. But also uncontrolled use of water resources in the basin has led in recent years; the
lake water level was decline. Finally, we can say that the role of human factors and impacts
is more than natural factors in the destruction of lake.
Conclusion
The results of this paper shows that human effects and uncontrolled exploitation of water
decade so that this period approximately 5 meters reduced lake water and lake area of 5200
square kilometers in 1998 reduced to about 3107 square kilometers in 2009. According to
analysis conducted in this paper include the use of Markov Chains, Cellular Automata and time
series if this trend continues, lake area in 2019 will be reduced to about 2000 sq. km. The issue
that caused irreparable environmental effects of increased salt in the region, the loss of
agricultural lands adjacent to the lake of salt transport by the winds and thus cause large
economic losses will be happen in this region. on other hand reduce the water level increases the
amount of saturated salt water will face that the amount currently reached 380 mg/lit, causing
destruction of the only existing live Artemia in the lake that as food for migratory birds.
Also in this article the role and importance of remote sensing data and processing them
for purposes such as monitoring and continuous monitoring, even during the days, weeks or
months can be considered the traditional methods no such ability and speed to act and sometimes
due to natural and human problem is not possible quickly data collecting. References
[1]Ahanejad,M and et al,2010. Detecting and Environmental Assessment of Spatial Changes of Hamun-
E-Saberi Lagoon Using Satellite Imagery and GIS, ICEST2010,Bagkok, Thailand, April2010.
[2]Ahanejad,M and et al, 2010. Evaluation and forecast of human impacts based on land use changes
using multi-temporal satellite imagery and GIS: A case study on Zanjan, Iran Journal of the Indian
Society of Remote Sensing , Pages -659-669
[3]Alesheikh, A and et al, 2007. Coastline change detection using remote sensing, Int. J. Environ. Sci.
Tech., 4 (1), pp.61-66.
[4]Azari Takami,G.,1993. Uraemia Lake as a valuable source of Artemia for feeding sturgeon fry.J.Vet,
Fac, Univ, Tehran,47.2-14.
[5]Cosh, M., E. R. Hunt, Jr., T. J. Jackson, and T. M. Yilmaz, 2009. SMEX04 Landsat TM/ETM+ NDVI
and NDWI. Boulder, Colorado USA: National Snow and Ice Data Center. Digital media.
[6]Gu, Y., Hunt, E., Wardlow, B., Basara, J.B., Brown, J.F., Verdin, J.P,2008. Evaluation of MODIS
NDVI and NDWI for vegetation drought monitoring using Oklahoma Mesonet soil moisture data.
Geophysical Research Letters 35.
[7]http://www.azarwater.ir/MSathi_Rpt.asp
[8]http://www.agrw.ir/Farsi/Orumieh.asp?Id=11#P-1-1
[9]Liu Cheng-lin, Wu Jian-jun, 2008. Crop Drought Monitoring using MODIS NDDI over Mid-
Territory of China, International Geoscience & Remote Sensing Symposium.
[10] Mcfeeters,S.k, 1996. The use of the Normalized Difference Water Index (NDWI) in the delineation of
open water features , International Journal of Remote Sensing, Volume 17, Issue 7 May 1996 ,
pp.1425-1432.
[11] Ma,M and et al, 2007. Change in area of Ebinur Lake during the 1998-2005 period , International
Journal of Remote Sensing, Vol.28,No.24, pp.5523–5533.
[12] Maktav, D and et al. Landsat Thematic Mapper Monitoring Lake Salda in Turkey ASPRS ACSM,
1994.
[13] Qulin TAN, Siwen Bi, Jiping Hu, Zhengjun Liu, 2004. Measuring Lake water Level Using Multi-
source Remote Sensing Combined with Hydrological Statistical Data, Geosciences and Remote
Sensing Symposium.
[14] Rasoli, A and et al, 2007. Monitoring of Urmia Lake water level fluctuations using multi-temporal
satellite images processing.J, Modaress, No2, summer 2007.
[15] Robert Gilmore Pontius Jr. and Hao Chen, 2008. "Land Change Modeling with GEOMOD", Clark
University.
[16] Ronald Eastman J, 2008. "Idrisi Andes Tutorial ", Clark University.
[17] Wang, S and et al, 2005. Suspended Substance Content Inversion in Lake Taihu Using Remote
Sensing Data, Geosciences and Remote Sensing Symposium.
