Impact of Land Use/Cover Changes on the Flow of the Zarqa River in Jordan

This paper investigated the impact of land use/cover changes on the flow of the Zarqa River in Jordan over a period of twenty-eight years. The land use/cover maps were derived using a set of medium spatial images with full scenes for the years 1989, 2002, 2011 and 2017. These images correspond to the river flow data for the same hydrological rainy seasons. The component of the river flow consists of the base-flow, flood and contribution of effluent from treatment plants. Base-flow was separated from hydrographs and effluent contribution was obtained. Runoff coefficient was determined as the ratio of flood volume to rainfall volume. The land use/cover maps were classified as urban fabrics, bare rocks, open rangelands and bare soils, agricultural areas, agro-forestry, and water bodies. During the study period, urban areas increased from 4.87% to 16.14%, and agricultural areas increased from 21.69% to 31.66%. The areas of rangelands and bare soil decreased from 34.91% to 22.57% and bare rocks from 35.98% to 27.57%, respectively. The increase in urban and agricultural areas resulted in runoff coefficient improvement from 1.89% in 1989/1990 to 2.72% for 2016/2017. The results could be useful for planners and decision makers for future flow management in the Zarqa River Basin. The approach and results of this study confirm the findings of similar studies for land and water management.


INTRODUCTION
The land use/cover information is the basic pre-requisite for land and water resource utilization, conservation and management. It is used to protect the water resources necessary for sustaining human life (Demir et al. 2007). The information on land use/cover are available today in the form of thematic maps, where earth features such as water, urban, forest, exhibit typical spectral response pattern that enables for target identification, such as assessing the impact of land use/ cover changes on river flow (Cook 1945; Gautam et al. 2000;Shammout 2003). Land use/cover change and climate variability are two major factors controlling the hydrological responses, but the dominant role in the hydrological responses is played by the land use/cover change (Berihun et al. 2019). It is widely known that changes in landuse/cover, such as conversion of forest land into cropland or grazing land, and urbanization, can increase surface runoff (Chow et al. 1988).
Land use/cover change is a major challenge facing the global environment (Kates and Torrie 1998). This change in land use/cover, when coupled with climate change, could potentially lead to an increased risk of flooding in urban catchments. The influence of land use/cover change on surface runoff has been investigated in many studies. They examined the impacts of land use/ Impact of Land Use/Cover Changes on the Flow of the Zarqa River in Jordan Maisa'a W. Shammout 1* , Khaldoun Shatanawi 2 , Jawad Al-Bakri 3 , Mahmoud M. Abualhaija 4 cover changes on floods, e.g. in Australia (Chen and Yu 2013), and in south-eastern United States (Nagy et al. 2011) as well as studied the water resources and land use and cover, indicating that the forest removal leads to more variable flow. In Italy, the results of land use change impact on flooding areas showed an increase in flood peak and a decrease of the rangelands, forests, and bare lands, indicating a good correlation between flooding areas and land use changes (Apollonio et al. 2016). In the city of Ghent, the results showed that grasslands contribute the most to storm water runoff reduction, and forests controlled the lowest runoff (Li et al. 2020). (Shanableh et al. 2018) confirmed that the expansion of built areas progressively increased the impervious land cover in the coastal city of Sharjah, and significantly increased the runoff coefficient.
In developing countries, the rapid increase in population pressure has pronounced effects on the land use/cover dynamics, mainly through deforestation aimed at increasing agricultural production (Maitima et al. 2009). These changes are causing alterations of the land surface and also have a great impact on hydrological processes such as surface flow, groundwater recharge, infiltration, interception, and evapotranspiration (Costa et al. 2003;Shammout et al. 2013). Most of the studies agree that the expansion of agricultural lands and the growth of urban areas at the expense of vegetation cover and forest significantly increase the surface runoff potential in a given watershed. (Yin et al. 2017) showed that the water resources subjected to land use/cover and climate changes, affected surface runoff differently between decades due to the combined effects of land use/cover and climate changes. The impact of urbanization through assessing the relationship between urbanization, runoff, floods and rainwater harvesting provides a basis for developing sustainable urban storm water management practices for the cities. Hence, proper land management may increase the amount of surface storage, rate of infiltration, and capacity of the soil to store water. Without proper land use management, the alterations of the land surface on hydrological processes will be continued and mainly exhibited in river flow. Moreover, the amount of surface storage, rate of infiltration, and capacity of the soil to store water will be also decreased This study was conducted for the Zarqa River Basin (ZRB). It investigated the impact of land use/cover changes of ZRB on the flow of the Zarqa River over a period of twenty-eight years. This basin is located in Jordan, where surface water resources in Jordan are mainly scattered between 15 basins. One of the major basins is ZRB and it is well recognized as one of the most important basins with respect to its economic, social and agricultural importance. It is the second main tributary to Jordan River. ZRB is the most affected area suffering from several quality and quantity water problems. These problems are due to the scarcity of water which is a result of the wide fluctuations in annual rainfall and climate change, population growth, urbanization, and inappropriate land uses which resulted in watershed degradation. Accordingly, management tools should be adopted by planners to minimize this degradation. For this purpose, the translation of available real historical data of land use/cover images into responses of change for the river flow is essential for understanding, assessing, and managing the ZRB. The specific objectives were (1)

METHODOLOGY Study Site
The Zarqa River Basin (ZRB) drains an area of 4120 km 2 , about 95% of which is within Jordan and only 5% is in Syria. The basin extends from the Syrian city of Salkhad in Jebal al-Arab with an elevation of 1,460 m to south of Amman. Administratively, ZRB is located in five governorates, namely; Amman, Balqa'a, Zarqa, Jerash, and Mafraq. The ZRB hosts more than 50% of Jordan's population (10.5 millions). The basin discharges its water at the confluence of the Zarqa River with the Jordan River at an elevation of about -350 m. The streamflow of the Zarqa River is impounded at King Talal Dam (KTD), the largest Dam in Jordan, South-West of Jerash town. There are several wadis draining in the ZRB, such as Wadi Dhuleil, and Wadi Zaatari. The runoff gauge station at Jerash Bridge on the Zarqa River is the only gauge station in the whole basin (Al-Qaisi 2015) which has the continuous long recorded database and possibly sufficient to be suitable for flood evaluation, as it is located at the entrance of the KTD from the east (Shammout 2003). Figure 1 shows the Zarqa River Basin.
The soils of the ZRB differ widely according to the rainfall and climate. The basin is mainly located in northern highlands dissected limestone and the north Jordan basalt plateau (Al-Bakri 2008). In the west, the prevailing soil textures are clay and clay loams. Toward the east, the soil becomes more immature with silty loam to loamy in texture with very high carbonate content. These soils have weak structure and suffer from low infiltration rate, resulting from the presence of surface crust. Land and water use in the ZRB has undergone considerable changes. Recently, the expansion of Amman and the surrounding towns has been enormous, thus reducing the grazing land and fertile agricultural lands between Amman and other towns. The expansion was developed into a large urban conglomerate. In the north and north western parts, the ZRB is capable of supporting forests and agricultural activities. Natural forests occur in the mountainous part and agriculture within the basin is divided into agroforestry, rain-fed orchards and olives, field crops as well as irrigated agriculture. Historically, irrigation was practiced on the banks of the river, while during the last three decades, it expanded in the northern and middle parts of the basin, particularly in the Mafraq governorate where the irrigated areas nearly tripled during 1990-2014 (Al-Bakri 2008). The total irrigated area of the basin reached 18 thousand ha in year 2014 (Al-Bakri 2016). This resulted in increasing the demand and pressure on the limited groundwater resources of the basin.
The basin includes four municipal wastewater treatment plants (TPs), namely, Khirbet es-Samra Treatment Plant (KSTP), Abu Nusair, Baqa'a, and Jerash. KSTP is the largest wastewater treatment plant located in the ZRB. The treated wastewater discharged from KSTP to Wadi Dhuleil flows downstream to join the Zarqa River near Sukhna. The volume of effluent discharged from the TPs to the Zarqa River is expected to reach 180 MCM by the year 2025, which will lead to the needed of a set of strategies for the expansion, sanitation and storm water management (Heller et al. 2014). The water system in ZRB is characterized by complex characteristics, where its flow mixed with treated effluents from treatment plants and finally stored in KTD, which is located at the outlet of the basin where the dam's water is used for irrigation in the Jordan valley. Only 50 MCM surface fresh water is pumped to the basin from the Jordan Valley for domestic use and the groundwater from different basins is transferred to meet the growing domestic demand (Shatanawi and Shammout 2011). The heavy utilization of the water resources of the ZRB has resulted in reducing the base flow of the Zarqa River from 5 m 3 /s to less than 1 m 3 /s and the discharge of the springs reduces from an of average 317 MCM/year prior to 1985 to less than 130 MCM/year after 2000 (Shammout 2003).

Assessment of land use/cover changes
A set of medium spatial resolution images of Landsat Thematic Mapper (TM), Enhanced Thematic Mapper Plus (ETM+) and Opera-tional Land Imager (OLI) images were used to derive land use/cover maps of the basin. The dataset included full scenes for the years 1989, 2002, 2012 and 2017 for the path/row 173/38 and 174/38. These images corresponded to the same hydrological rainy seasons, except the 1989/1990 season which was only based on the 1989 image instead of the 1990, which was not available. The selected datasets were cloud free images acquired during March -May, when natural vegetation and rain-fed crops are usually at their peak growth. The dataset was mainly downloaded from the official website of Landsat 8 (http://earthexplorer. usgs.gov) at no cost. The data used for mapping land use/cover included the visible and the near infrared (NIR) bands with 30 m spatial resolution. Various image processing techniques were applied to prepare the images for digital and visual interpretation of land use/cover. This included the transformation of images to derive normalized difference vegetation index (NDVI), visual interpretation for urban areas and forests, mosaicking of output layers and clipping of the images to the borders of the study area.
The first step of deriving the land use/cover maps was the derivation of NDVI layers to represent all vegetated areas in the basin. The second step was the use of visual interpretation of images in order to derive layers of agroforestry in the high rainfall areas and urban areas all over the basin. The method of visual interpretation was preferred over the digital classification techniques to avoid classification errors that might result from spectral mixing at the 30 m spatial resolution. Following this step, different functions within the geographic information system (GIS) were applied to append the layers derived from the visual interpretation onto the layers of NDVI. The hydrographs of daily flow in m 3 /sec were constructed in order to separate the baseflow from the flood flow. This was carried out by using the above mentioned flow data. The total base-flow in the Zarqa River consists of one portion of KSTP reaching the gauging station, and the second portion, it is the flow that would exist in the stream without contribution of direct flood (runoff). The separation of the total base-flow was done using the straight line method for each event (Chow et al. 1988), from which the time point of direct surface flood begins to the point where normal baseflow resumes. Between these points, surface flood and base-flow can be separated. This was done by using the stream flow hydrograph in Excel. Once the beginning and ending of each flood were even, the total base-flow were determined and their values were recorded. The actual base-flow was separated from the portion of KSTP effluent. This portion is an equal from effluent of KSTP effluent minus the portion of effluent that has been used upstream of the gauging station for irrigation. The flood flows were determined by integrating the areas of each hydrographs above the base-flow and summing them all to determine the annual flood for each targeted year.
The rainfall data from 10 gauge stations was obtained from Ministry of Water and Irrigation that allow studying the responses of the Zarqa River flow. The ZRB was outlined to define its boundaries, drainage network and outlet using the Digital Elevation Model (DEM). There are 33 rainfall gauges in the ZRB. Only 10 of them have long records over the years from 1989 to 2017. For the selected rain gauges, Thiessen weighted average method was used (Chow et al. 1988) to determine the influence of each rain gauge to the whole basin. Average rainfall was computed according to the following equation  (Chow et al. 1988). In order to study the effect of land use/cover on the Zarqa River flow, each class of the land use/ cover was compared with the runoff coefficient.

Land use/cover changes
Land use/cover maps of the years 1989, 2002, 2011, and 2017 ( Figure 2) showed that the ZRB was dynamic in terms of land use character. open rangelands and non-cultivated areas. These results were confirmed in the previous findings in the basin and in similar areas in Jordan (Al-Bakri et al. 2013).
In comparison with the land use distribution from 1989 to 2017, the land use changes have shown that continuous urban fabrics class has increased by 12.23%, and agricultural areas class has increased by 9.97%. On the other hand, there has been a reduction in agroforestry by 0.5%, and a reduction in open rangelands and bare soils by 12.34%. Part of the changes in agricultural areas could be attributed to the differences in rainfall amounts and distribution among the four seasons. This would be expected under the Mediterranean environment, which had variability in rainfall during the season and among the seasons. Urbanization in the basin was seen as an important change that would reflect the water supply and demand. Urbanization was mainly on the expenses of agricultural areas in the high rainfall zones, particularly in the area of Amman, while expan-   Figure 3. This is indicated scientifically for the basins concerning their land uses and flow responses (Bakir and Xingnan 2008). In comparison with the water bodies since 1989 till 2017, there has been no considerable change, whereas the direction of flow change was clearly observed and considerable in the Basin. Table 2 shows the average rainfall, average river flow gauged, average KSTP effluent, reuse of reclaimed water upstream gauging station, contribution of reclaimed wastewater to river flow, average river base-flow, average river flood, and runoff coefficient for the . The water demand is increasing rapidly due to high population growth rates, and prolonged drought periods over the past decades which have reduced the surface and groundwater resources. However, the treated wastewater has become an ever-increasing percentage of the Zarqa River flows since the construction of the KSTP (González 2018). Therefore, due to high population growth rates, the treated wastewater constituted about 39.7% of the total flow of the Zarqa River in 1989/1880, about 45.7% in   Table 2.
These above-mentioned results demonstrate that the impact of land use changes on flow is an issue of considerable importance for land use management (Chow et al. 1988;Archer 2007) that can increase the amount of surface storage, rate of infiltration, and capacity of the soil to store water (Shammout et al. 2013), as water harvesting either for surface storage or for artificial recharge will assist in the sustainable management of water resources (Bakir and Xingnan 2008). Decision makers are challenged to secure more water for an increasing population and agricultural activities. They may implement the land use practices to manage the behavior of the Zarqa river flows (Farajad et al. 2017). Table 2  . This is because of the over-utilization of groundwater upstream and using water for irrigation (Shatanawi and Shammout 2011;Shammout et al. 2013). The base-flow in 1989/1990 was normal, originating from the discharge of the springs along the valley of the river. In the 1990s and afterwards, there was heavy pumping from the groundwater of the basin to the extent that the discharge of springs has dropped and others were completely dried. For the two average years which were similar in rainfall amount and pattern, Table 2 shows that the river flood increased from 11.7 MCM (0.37 m 3 /s) in 1989/1990 to 17.2 MCM (0.55 m 3 /s) in 2016/2017. The increase in flood may be attributed to the land use/cover change, especially the increase in urban and agricultural areas on the expense of range land and forest. This also reflects the runoff coefficient which has increased from 1.89% to 2.72% for the two years, respectively. These results confirmed the previous findings as in (Sriwongsitanon and Taesombat 2011;Li et al. 2020), where, they have shown the influence of land cover on runoff coefficient. It can be useful for land use, flood management of the river basin, and the importance for efficient scenario practices for various aspects of river flow management.

CONCLUSIONS
The Zarqa River Basin (ZRB) is under the pressure of various agricultural, industrial, and commercial activities. The water demand is increasing rapidly due to high population growth rates, and prolonged drought periods over the past decades, which have reduced the ZRB water resources. Hence, the translation of the available real historical data into the river flow responses to land use/cover changes is highly needed. This allows understanding, assessing the river flow and finding the ways towards ZRB management.
It was clearly observed that there was a direction of change of the Zarqa River flow in the basin since 1989/1990 till 2016/2017. The river flow was variable according to land use/ cover changes; the increase of urban and agricultural areas, the reduction in open rangelands and bare soil, and the reduction in agroforestry, resulted in the increase of the river flow. This also influenced the runoff coefficient. On the basis of these conditions, decision makers are challenged to secure additional water for an increasing population and agricultural activities. Nevertheless, the change in land use/cover approaches will allow the validation of future scenarios for flow management, which is an issue of significant importance that can increase the amount of surface storage, rate of infiltration, and capacity of the soil to store water.