Assessing Environmental Sensitivity to Desertification in Heet-Haditha Region-Based on Method Multiple Criteria Decision Analyses

Abstract


Introduction
Desertification in Iraq, especially in the central and southern parts, is worsened due to climate change, the deterioration of the quality of water used for irrigation, and the lack of rain, thus reducing fertility and this has negatively affected the environment, in addition to the human effects. (Jradat, 2003). The rapid urbanization and population increase in Iraq have led to the change of many ecosystems, including land degradation, resulting in a serious environmental crisis (Lee et al., 2019). To face these challenges, several researchers have used modern techniques that help identify the indicators that lead to the emergence of desertification) (Egidi, 2021(. Among the most important of these technologies are geographic information systems and remote sensing (Uzuner and Dengiz, 2020). It is noticed that during the past ten years, researchers in Iraq started using an integrated approach of remote sensing and GIS techniques to assess desertification phenomena in some parts of Iraq. This approach provided integrated information of different parameters that control desertification factors. The techniques have been frequently applied in different global works related to desertification such as Rubio (1998), (Sardinha, 2008) and Egidi et al. (2021).
The multi-criteria decision analysis (MCDA) method was used to evaluate the phenomenon of desertification in the study area, as it provides an integrated approach to appropriate decision-makers and allows the comparison of alternatives with a set of clearly defined criteria that takes into account the most relevant aspects of the decision-making process (Zlaugotne et al., 2020).
Operationally, MCDA supports structuring decision problems, evaluating the performance of alternatives across criteria, exploring tradeoffs, formulating a decision, and testing its strength. Multicriteria decision analysis is especially useful when reducing a multi-objective problem to a singleobjective problem that is either useless or undesirable (Aubert et al., 2022).
There are no published studies on evaluating the phenomenon of desertification by using multivariate statistics including principal component, correlation matrix and cluster analyses in Iraq, and therefore this study is considered the first of its kind. This research aims to assess the environmental sensitivity to desertification in the Heet-Haditha region .

Study Area
The study area is located in the west of Anbar Governorate, at latitudes 42° 48' 47" E to 42° 22' 21" E and longitudes 33° 38' 00" N to 34° 08' 23" N in the west. It is located on the banks of the Euphrates River, and there are many orchards and agricultural areas (Fig.1). The climate in the study area is characterized by high temperatures in summer, with low precipitation (Hammody, 2021).

Geology of the Study Area
Geologically the study area is located in stable shelf within the Salman zone, and the main structural component in the study area is the Abu Al-Jir fault system, which represents the boundary between the stable shelf and the unstable shelf. A stable shelf is a stable unit that was not or slightly affected by the Alpine movements during the Mesozoic and Tertiary periods (Fayyad, 2008). It is characterized by that the Alpine movements did not form surface folds, but rather led to the formation of vertical displacement and some horizontal displacement in the bottom rocks, which led to the formation Grabens and Horsts in different directions through the geologic ages.
The area consists of layers of sedimentary rocks ranging in age from the Eocene to middle Miocene (Table 1), as well as several types of Quaternary deposition, including River terraces (Pleistocene), Slope deposits (Pleistocene -Holocene), Old earth sediments (Pleistocene -Holocene), Flood Plains (Holocene) deposits and Valleys-filled deposits (Holocene). Quaternary deposits are recent deposits dating back to the Glacial and the Pleistocene cover a large area of the study area consisting of friable materials of clay, sand, silt and gravel of various sizes. It has a great economic importance, as it is the main source of gravel, sand and clay, and are a source of many groundwater pools (Al-Alusi, 2001).

Data and Variables
According to the framework of the environmentally sensitive area, nine indicators were selected that determines the level of land sensitivity to desertification in the Heet -Haditha region for three periods 2000-2010-2020. These indices include: Normalized Difference Vegetation Index (NDVI), Normalized Water Difference Index (NDWI), Salinity Index (SI), Land Surface Temperature (LST), Temperature (T), Precipitation (R), Evapotranspiration (Eva), Wind speed (WS), relative humidity (RH). Available climatic data were obtained from the Iraqi Metrological Organization and Seismology (IMOS). Arc-GlS software used to all the base maps and maps required for the study area, based on satellite images Landsat 5 and 8 (Fig.2).

Normalized difference vegetation index (NDVI)
To observing changes in vegetation cover and study desertification, the NDVI was calculated from satellite aerial images. As a result of excessive logging and other human activities brought on by urban growth, the vegetation cover dramatically changes negatively. The NDVI value is calculated by the following equation (Fig.3

Normalized difference water index (NDWI)
The NDWI was used to emphasize the presence of water bodies and waterlogged areas in remotely sensed digital images as well as to delineate them. Several water extraction algorithms have been developed and applied to remote sensing images. The water index is calculated based on the following simple formula (Fig.4): -Bahadeli et al., 2020) Whereas R (G) is spectral reflectance in green band and R (NIR) is spectral reflectance in the infrared ray band.

Salinity index (SI)
Salinization is the process by which water-soluble salts accumulate in the soil, eventually to toxic levels for plants. It may occur naturally or human action or because of inappropriate irrigation methods or overexploitation of aquifers. In compared to collecting field data of salinization-risk areas, remote sensing can predict soil salinity reliably in a short amount of time and with minimal effort.
The Salinity Index (SI), which derives from the green band and red band, was created as a straightforward model to create an image that highlighted regions with low vegetation density and high crusted saline soil (Fig.5).

Land surface temperature (LST)
It is the temperature that can be measured when the Earth's surface is in direct contact with the measuring instrument (Rajendran, and Mani, 2015). The land surface temperature was calculated from Landsat images (Fig.6).

Temperature (T)
It is one of the most important climatic factors because other elements are closely related to it. Temperatures vary from time to time. In summer, the maximum temperature rises, this rise causes many problems for the soil, such as the drying out of the soil, which lead to the rise of saline groundwater to the surface, and this effect on the phenomenon of desertification (Fig.7). By study the results and comparing the climatic information obtained from the Iraqi Metrological Organization and Seismology (IMOS), it was found that the highest temperature occurred in year 2010, where the mean maximum temperature reached 28.86°in the Heet region and 29.92° in the Haditha region.

Rainfall (R)
It is water droplets that condensed from water vapor in the atmosphere and then fell under gravity. The lack of rainfall or its scarcity causes rapid evaporation and the accumulation of salts in cultivated lands, and thus is a major cause of desertification (Fig.8). The increased rainfall in Heet region during 2020 to 10 mm, while in Haditha region increased during 2000 reached average rainfall 7.80 mm.

Evaporation (Eva)
The study region is characterized by a high temperature in summer and little rainfall in winter, so evaporation increases significantly during the summer due to high temperatures. Evaporation is one of the main causes of desertification (Fig.9). Evaporation increased significantly in the study area during the year, where in 2010 the average evaporation reached 285.5 mm in Heet region and 241.58 mm in Haditha region.

Wind speed (WS)
Wind speed is one of the most important elements of erosion that has an impact on geomorphological forms in arid and semi-arid environments. The studied area is characterized by a dry climate with little rainfall, so the wind has a negative effect that works on drying the land when small amounts of rainfall, and it also works on soil erosion and thus causes the phenomenon of desertification (Fig.10). During the year 2000, the wind speed increased, reaching Haditha 3.27 m/s and 4.12 m/s in Heet region.

Relative humidity (RH)
The RH is defined as the amount of water vapor present in the air at a given temperature, relative to the maximum amount of air it can hold at the same temperature (Fig.10). RH is directly affected by temperature as it represents the percentage of water vapor in the air so, it changes with temperature. Relative humidity is one of the most factors affecting the desertification phenomenon. The relative humidity increased in the Heet region during the year 2020, the mean annual RH of 40.75%, while in the Haditha region, the highest increase in humidity reached 42.41% during the year 2000.

Statistical Analysis
The statistical analysis aims to describe, organize, classify, and present a collection of data in a clear manner in tabular or graphic forms (Witte, 2017). The census provides simple summaries about the sample taken. These summaries may be either quantitative or graphic, and they constitute the initial description of the data (Park and Nemec, 2019). Statistical analysis was used to determine the relationship between the nine indicators for estimating desertification sensitivity. Statistical analysis was performed using the academic software package (STATISTICA -version 13.3) for Windows. The Correlation Matrix Analysis (CMA), Principal Component Analysis (PCA) and Cluster analysis (CA) has been done.

Correlation Matrix Analysis (CMA)
The results of CMA were used to examine correlations among the indicators that affect the phenomenon of desertification in the Heet-Haditha region. Where the positive correlations with strong significance indicate an origin/source with a common effect on desertification, while the negative correlations indicate a weak effect coming from several indicators (Table 2). *Marked correlations are signification at p < 0.5000 A significant negative correlation at p < 0.5000 was found between NDWI-RH (-1.000000), RH-LST (-1.000000) and signification positive correlation between NDWI-LST (1.000000).

Principal Component Analysis (PCA)
PCA is the most used method for data exploration and data analysis across all fields of science and is particularly useful when the data at hand are large (multiple variables) (Kherif and Latypova, 2020). PCA has been done using the academic statistics software package STATISTICA -version 13.3 for Windows. The result of the principal components analysis (PCA) is listed in Table 3. There are two eigenvalues with values higher than one represented 96.68% of the total variance. Factor 1 shows 59.79% of the total variance and has high positive loading on RH and WS, and strong negative loadings on NDWI, R, and LST. The factor 2 explains 36.89% of the total variance with strong positive loading on T and NDVI, and strong negative loading on SI, Eva. Marked loadings are > 0.70

Cluster Analysis (CA)
Statistical techniques are groups that are sequentially created by systematically merging similar clusters together (Yim, and Ramdeen, 2015). On the standardized data set, the hierarchical cluster analysis (HCA) was performed by (Wards method) (McGarigal, 2013). The HCA results for desertification criteria are shown in Fig.3.Two clusters were identified: cluster 1 includes LST, ST, NDWI, R and NDVI, and the cluster 2 includes Eva, WS, RH and T. Cluster 1 includes four subclusters:(ⅰ) LST, (ⅱ) SI, (ⅲ) NDWI, and (ⅳ) R-NDVI and cluster (2) includes three subclusters: (ⅰ) T, (ⅱ) Eva, and (ⅲ) WS-RH. The indicators subcluster (2) greatly affect the desertification phenomenon in the studied region. The result of cluster analysis partially supported of the results of Principal Component Analysis.

Discussion
In this research, the areas sensitive to the desertification phenomenon were evaluated in the Heet-Haditha region based on nine main indicators. The results obtained from the Statistical analysis (correlation Matrix Analysis (CMA), Principal Component Analysis (PCA) and Cluster analysis (CA)) showed that the most influential indicators of the phenomenon of desertification in the study area are Eva, WS, RH and T. These four indicators appear in the cluster (2) and are linked with each other and have a significant impact on the desertification.

Conclusions
The results obtained from the cluster analysis operations and based on the maps that were created using Arc GIS, shows that the most affected area by the desertification phenomenon is in the west and southwest of the studied area (Fig.13). The rate of desertification increased during the year 2010 due to the terrorist operations during that period, and due to the decrease in the rainfall rate and the increase in tree cutting and urban sprawl especially in Haditha region.