CHANGE DETECTION IN MOSUL DAM LAKE, NORTH OF IRAQ USING REMOTE SENSING AND GIS TECHNIQUES

This study was aimed to demonstrate the spatial changes of Mosul Lake using change detection Techniques. Water bodies, an integral part of the Earth's hydrological cycle, such as rivers, ponds, lakes, wetlands, and reservoirs, have a major effect on climate change and global warming. one of the most imperative vital activities in Iraq for the administration of its water resources is Mosul Dam Lake. The lake has changed in the water surface due to the severe anthropogenic activities, climate change, the monthly inflows received from the Tigris River, and the controlling of water imports by neighboring countries, and the territorial policy. Remote sensing techniques and Geographic Information System ArcGIS 10.3 software were used in the present study for processing the images and managing the database of each image were downloaded from the united states geological survey(USGS). Spectral Water indices, includes Normalized Difference Water Index (NDWI) and unsupervised classification were used to extracting water bodies and computing water area, the images difference technology was used to detect a different image and capture changes area for 20 years, the method has been applied to subsets of the Landsat series images that acquired during April in 2000, 2005, 2010,2015 and 2020. The results showed that there is an increase in the area of Mosul dam lake from its original area in 1986 which was 244 km in a rate of (15.1,31.14,13.11, and 20.9) % for periods (200,2005,2015, and 2020) respectively, and a decrease in water area with a rate of 7.78% for period 2010, that means the water surface area fluctuated over the period given, increasing and decreasing in the water surface. the minimum water surface area was recorded in 2010 of about 225 km, whereas, the maximum area of the lake was found to be 320 km in 2005. the total average increasing and decreasing change detection were 25.4550174 and 19.7851824 km


INTRODUCTION
Typically, Surface water bodies are dynamic in nature as they decrease, increase, and change their appearance or course of the stream with time, due to various natural and human-instigated factors (21). These changes usually cause serious consequences in extreme cases, It can cause floods and droughts; Therefore, it is necessary to distinguish the presence of surface water, extract its extent, evaluate and Calculate the quantity of its volume, and monitor its dynamics (20). Geographically, in the world's driest belt, Iraq is situated (4), This implies that Iraqi agriculture relies on surface and groundwater resources to provide the irrigation water required for agricultural stability (36). These resources are restricted and the greater part of them are imparted to different countries neighboring Iraq, also have faced a lot of threats and damages that caused to decrease the water supply of the country. one of the biggest artificial reservoir in Iraq (26,28), Mosul Dam Lake, which was selected as a case study, is one of the most important Strategic Services Projects in Iraq which supporting the water interest of Mosul, Baghdad, and different urban areas, also provide water for an irrigation project called [North Al-Jazira Irrigation project] (26). The definition of water indices is a recent technique introduced for the recognition of water bodies through changes. Mathematical models improve the water signals in images collected from visible/near-infrared scanning sensors for a given pixel (10). Two bands usually specify these models from the visible (green) and near-infrared portions of the spectrum, When Green and NIR refer to the reflection in the green and near-infrared spectra respectively (14). Among the common water indices are the (NDWI) and the modified normalized difference water index (MNDWI) both of them have been successfully used to delineate surface water area information (12,23). for deeper parts of the water surface, NDWI has a superior outcome, while for shallower parts has bad. In this manner, the utilization of the (NDWI) approach enhance water's reflectance features by reducing the low reflectance of Near Infrared (NIR) and improving the reflectance in green wavelength (25,35). Many studies focused on using remote sensing and GIS techniques to detect the changes in Landsat images, these studies are: Mahir Mahmod H. evaluated water surface area for a period from (1984-2019) The results indicated that the water layer suffered a pattern of changes concerning the water of the surface area (24). Issa  This study was aimed to demonstrate the spatial changes of Mosul dam Lake using change detection in the period from 2000-2020 utilizing multi-temporal Landsat Thematic mapper 5, Enhance Thematic mapper +7, and Operational Land Imager 8 data.

MATERIALS AND METHODS Description of the study area
Mosul Dam is one of the large dams in the world, the second greatest dam in the Middle East, and the largest dam in Iraq, due to the capacity of its reservoir (2). at a maximum operation level with a maximum storage volume and a maximum water depth, The water surface area of its reservoir is 380 km 2 the length is about 45 km and the width ranges between 2 and 14 km (22). Mosul dam lake Figure 1; is an artificial reservoir located between latitude (36˚36'N -36˚50'N) and longitude (42˚27'E -42˚58'E), build on Tigris river in northern Iraq approximately 50 kilometers northwest of Mosul city and 80 kilometers from the Syrian and Turkish borders, also is about 500 km on a straight line distance to the north from Baghdad (27) started operating in 1986. Typically, The climate area may be regarded as being similar to a Mediterranean climate, is hot and dry during the summer, while cold and rainy during winter with occasional snowfall taking place in the mountainous region (16), The average monthly temperatures range between 6°C in January to 34°C in July, but the temperatures decrease toward north regions of Iraq (18), the annual average of wind speed in the northern regions was low about 2.7 m/s compared to the middle and southern regions(3), while (19) splits Iraq into three climatic zones according to the rainfall factor, the study area in Arid and Semi-Arid Zone where the annual rainfall above 400 mm. In April and May, a most seasonal discharge occurs, whereas there's less discharge in October and September. Published research has concluded that the Ilisu Dam project would significantly reduce the inflow to the reservoir of the Mosul Dam. (9,34). The lake has changed in the water surface due to the severe Human activities, climate change, and also the dams and irrigation projects that have been constructed at the headwaters in neighboring countries such as Turkey, Iran, and Syria, where the greater part of the water flowing into the lake comes from Turkey.

Figure 1. location of Mosul dam lake Data acquisition
The five satellite images of series Landsat which whose path was 170 and Raw was 35 have been downloaded from The United States Geological Survey (USGS) database (11). These images were gathered and employed to give clarifications on lake surface area fluctuations for 20 years, acquired in April from (15/2000) for Landsat 5TM, (21/2005,3/2010 and 14/2020) for landsat7 ETM+, and (25/2015) for landsat8 OLI sensors types with 30 m ground resolution and 185x185 km ground area, the data scenes were all clear of clouds. an efficient tool, the Geographic Information System (ARC GIS10.3) program, is gathering Raw data, processing, analyzing, and presenting spatial data, also evaluate of water resources with the accurate, quick, and implemented spatial technology objectives (6,8)

Methodology
The following tasks were performed to obtain the objectives of the study: data collection, image pre-processing, image processing, the NDWI approach, extraction of the lake surface area in each image, differencing technique, and change detection. Figure 2 illustrates the main techniques proposed in this research to detect changes in the lake surface area.

Pre-processing images
Best and cloud-free satellite images that cover the study area, bad images, or contain defects have been excluded, so registered in the same projection system the Universal Transverse Mercator (UTM) and World Geodetic System (WGS 84). Post processing images Normalized difference water index (NDWI( NDWI an easy and effective technique used to extract water bodies (15), which is calculated from two or more image bands, to identify the differences between water and non-water areas (31). The last one and (1,30) found The performance of NDWI was more accurate for different elevation ranges and water types than other indices and hence it was used to model the lake's spatial Changes. (25) A limit estimation of zero Proposed to extract water bodies from the Raw digital Landsat values, where all positive NDWI values were classified as water and negative values as nonwater. However, this limit estimation does not enable discrimination between built-up surfaces and water pixels. The NDWI for TM, ETM+, and OLI sensors is formulated by Equations below (33): NDWI = (R Green −R NIR )/ (R Green + R NIR ) NDWI 5TM,7ETM+ = (B2-B4)/(B2+B4) ……. (1) NDWI 8OLI = (B3-B5)/(B3+B5) ………. (2) where R NIR and R green are the Reflectance of the near-infrared (band 4) and the green (band2) respectively in Landsat 5TM and 7ETM+ sensors (29) .while, R NIR and R green are the Reflectance of the near-infrared (band 5) and the green (band3) consecutively in land8 OLI sensor

Unsupervised classification
Unsupervised classification is The distinction between water and non-water objects, based on mixtures of two or more spectral bands using different mathematical operations, which was scientifically used to divide the data into two classes. (7). Differencing image technique By using the image differences technique in the ARC GIS10.3 program, can be detected changes in the mentioned area, subtracting pixel values to the same position as two pictures include two different periods. Pixelby-pixel comparisons are made between the two co-registered images, and pixels related to altered areas yield values distinct from those pixels followed by unchanged areas(32). Image differentiation can be mathematically represented as follows: D I = I (T1) − I (T2) ……………. (3) Where D I represents the difference between images and I(T1) and I(T2) represents the captured images during two different periods. the changed images have been Reclassified into three categories, The value (1,0,-1) for an increase, never, and decrease change areas respectively (5).

RESULTS AND DISCUSSION
water body extraction techniques were used to determine decreasing and increasing trends of surface area in Mosul lake to a series of Landsat images between 2000 and 2020 when the values of NDWI ranged between (+1 to -1).in the present study, ArcGIS version 10.3 was applied to classify data into two classes Figure 3and 4 due to the high reflectivity of plant and soil in the range of near-infrared makes the values of NDWI positive for the presence of water areas like (water bodies, rivers, and canals) and are enhanced therefore appear to be bright, While the green and built areas like (Vegetation, bare land) are dark and have negative values or zero and are suppressed. an unsupervised classification technique was applied on the NDWI images, to know the Spatial variation of the water bodies. As well differencing image technique has been employed to capture and detect changes in a mentioned lake. Eventually, the water surface areas of the lake are computing in km 2 .  Figure 5 since 2000 the lake area was 281km 2 increase at a rate of 15.1% from its original area which was 244 km2 in 1986 calculated by (13) until 2005 to achieve its largest area and the most elevated rate during the study period 320 km 2 at a rate of 31.14%. However, after 2005 The lake suffered an apparent decline in its area compared to 2005, and at a rate of 7.786% from its original area to have 225 km 2 in 2010, then the lake area was increase again to reach 276 km 2 and 295 km 2 at a rate of 13.11% and 20.9% in 2015 and 2020 respectively. Table 1 shows the results of satellite images across two decades for the Lake in each year.