The bifurcation ratio (R_b) is defined as the ratio of streams number of a given order to the number of the streams of the next higher order [57] , and it is computed by:

R b = N u / N u + 1 where (1)

N u = total number of stream segments of order “u”,

N u + 1 = no. of segments of the next higher order.

Bifurcation ratio is elaborated as an index of relief and dissection. Rb values of drainage basins vary from 2 for flat/undulating or rolling terrain, to 6 for catchment where drainage network is distorted heavily by geological structure [57] [58] [59] . High Rb values indicate high overland flow and an early hydrograph peak with a high potential of susceptibility to flash flooding during exceptionally heavy rainstorms [60] , which in turn increase soil erosion rates and sediment discharge in the main channel.

Drainage density (D_d) is computed as the total length of streams in a drainage basin per unit area [57] [61] [62] , or

D d = L u / A (2)

A = the basin area,

L u = is the total stream length.

D_d is a measure of topographic dissection and runoff potential of the drainage basin. High D_d value denotes high runoff, a quick stream response, and in turn, a low infiltration rate. Whereas low drainage density of a catchment implies low runoff and high infiltration [63] .

Stream frequency (F_s) is defined as the ratio of the total number of streams (N_u) of all orders in a catchment to the watershed area (A) [57] . It is expressed by the following equation:

F s = N u / A . (3)

F_s are positively correlated with D_d values of a watershed. Low F_s values indicates a relatively low infiltration rate of surface water, and thus the ground water potential is relatively low [64] . High stream frequency denotes more infiltration, and thus high groundwater potential.

Texture ratio (T_r) T_r is refers to the ratio of the total number of streams of the first order (N1) to the perimeter of the drainage basin. It is considered a highly significant parameter in drainage basin morphometry. T_r variable relies on slope materials and lithology, infiltration capacity and topographic relief [65] . Texture ratio is determined by:

T r = N u / P where, (4)

N u = the total no. of streams of all orders,

P = perimeter (km).

Length of overland flow (L_o) is determined by the equation:

L o = H / L b , where (5)

H = basin relief (B_h) m,

L b = basin length (km).

L_o is the length of water over the ground before it is concentrated into definite stream channels [57] . L_o parameter is one of the most significant independent variables affecting both hydrologic and hydrographic development of drainage basins. This parameter is related inversely to the average slope of the channel and is equivalent to a large extent to the length of sheet flow [57] .

Form factor (R_f) is calculated according to Horton [57] :

R f = A / L b 2 . (6)

R_f refers to the ratio of the area of drainage basin to the square of the basin length [58] . Higher values of form factor denote a more circular shape of a catchment, while smaller R_f values (<0.45) imply that the basin is elongated in shape.

Shape factor (B_s) represent the ratio of the square of the basin length to the area of the basin, or

B s = L b 2 / A . (7)

Shape factor provides a conception regarding the circular character of the catchment. The greater the circular character, the greater the fast response of watersheds following an intense heavy rainstorm [66] .

Elongation ratio (R_e) is defined by Schumm [67] according to the following equation:

R e = 1.128 A / L b . (8)

Low values of R_e denote that the catchment is more elongated. Where the R_e value approaches 1.0, the shape of the watershed becomes a circular shape [67] .

Compactness coefficient (C_c) is developed by Gravelius [68] , and defined as the ratio of perimeter of a watershed to circumference of the circle area, which is equal to the area of the watershed. C_c is computed according to the following equation:

C c = P 2 π A where (9)

P = perimeter of the basin (km),

A = area of the basin (km²).

When C_c value is 1, it indicates that the watershed is a perfect circle. If the C_c value is 1.28, the basin is more square shaped, while the basin is considered highly elongated, when the C_c value > 3.0 [69] .

The circularity ratio (R_c) of a catchment is calculated according to:

R c = 4 π × A / P 2 , where, (10)

(A) is the basin area, and (P) is the perimeter [70] . If R_c is close to 1, the shape of a catchment is circular. Low, medium, and high values of R_c indicate young, mature, and old stages of geomorphic development of the catchment respectively.

Morphometric analysis for prioritization of the 23 fourth-order sub-basins related to W.Wala watershed was conducting using topo sheets, ASTER DEM and Arc GIS software. Topographic maps of scale 1:50,000 were acquired from the Royal Jordanian National Geographic Centre (Amman). They were scanned, geo-referenced, and converted to a zone 36N projection system using Arc GIS 10.1, and the associated packages. The W.Wala watershed and the 23 sub-basins were delineated using the topo sheets. Then ASTER DEM (30 m resolution) was utilized to extract the drainage networks using the Arc Hydro tool. Stream order was designated using the stream ordering system developed by Horton [57] and elaborated by Strahler [58] . The W.Wala watershed is classified as a sixth-order drainage basin (Figure 3). Fifteen morphometric parameters were calculated using DEM, GIS, and the mathematical equations developed elsewhere [41] [57] [59] . Five basic parameters were derived to illustrate the morphometric characteristics of the 23 sub-basins, Area (A), Basin length (L_b), Perimeter (P), Stream order (u), and Stream Length (L_u). Furthermore, five linear parameters, and five shape parameters were considered in prioritization of the 23 sub-basins based on morphometric analysis as exemplified earlier. These are: Bifurcation ratio (R_b), Drainage density (D_d), Stream frequency (F_s), Texture ratio (T_r), and Length of overland flow (L_o). Moreover, the five shape parameters used in this research are: Shape factor (B_s), Form factor (R_f), Circularity ratio (R_c), Elongation ratio (R_e), and compactness coefficient (C_c). All morphometric parameters were measured directly from the DEM using GIS software. A land use/land cover map (Figure 4) was compiled using ERDAS IMAGINE 2015 (v.15), LANDSAT 8 (July 2017), and supervised classification. The Maximum likelihood Method of classification techniques was employed to classify land use/cover, based on the classification system proposed by Anderson et al. [71] . A soil map (Figure 5) was digitized from the 1994 National soil survey maps and reports pertaining to the National soil and land use maps [72] . A slope categories map was generated using ASTER DEM (Figure 5). A remarkable variation exists in slope categories. Slope categories of 0˚ - 5˚, 5˚ - 10˚, and 10˚ - 15˚ dominated the eastern part of

the catchment. Whereas, slope categories 15˚ - 20˚, 20˚ - 30˚, and >30˚ stand out in the western part, with the presence of cliffs (>45˚) along the canyon tributaries downstream. The development of an efficient and cost, effective GIS and remote sensing techniques enables researchers to extract, measure, compute and process precisely basic, linear, shape, and relief morphometric parameters of drainage basins. Moreover, the availability of free access Digital Elevation Models (i.e., STRM and ASTER DEMs) with a reasonable resolution (90 m and 30 m respectively) have improved the quantitative analysis approach of drainage networks, and morphometric mapping, thus, expanding the application of morphometric analysis to other fields of research. Discriminant Analysis (DA) was employed to test statistically the validity of priority groups of sub-basins generated based on morphometric analysis, and to determine if they are significantly different from each other, and to help in explaining regional spatial difference among the fourth-order sub-watersheds in term of prioritization. Through this technique, each sub-watershed can be related to a set of discriminant functions by calculating its “scores” on those functions. The mean value of these scores, on each discriminant function, can be calculated for each priority class. The mean values are then utilized to test if the priority classes are significantly different from each other, or in other words, if the resultant priority classes are accepted or not.

Recently, morphometric analysis was employed broadly to assess watersheds of different size (sub-watersheds, mini-watersheds, and micro watersheds) for soil and water conservation [31] [34] [36] [37] [38] [41] [42] [44] [45] . The linear and shape morphometric parameters termed “erosion risk parameters” [31] [41] were utilized for prioritization of the 23 sub-basins connected with W.Wala, and designated 1 - 23 (Figure 6). The linear parameters include(a): bifurcation ratio (R_b), drainage density (D_d), stream frequency (F_s), texture ratio (T_r), and length of overland flow (L_o); and the shape parameters include (6): form factor (R_f), shape factor (B_s), elongation ratio (R_e), compactness coefficient (C_c), and circularity ratio (R_c). Based on the range of computed compound parameter (C_p) values (Table 2), the 23 sub-basins of W.Wala were classified into three priority groups:

1) High priority (3 - 4),

2) Moderate Priority (4 - 5),

3) Low priority (5 - 6).

Using the Arc GIS tool, the spatial distribution of the 23 sub-basins groups based on the compound parameter values (Cp) was generated. Figure 7 illustrates the priority classes of sub-basins for soil and water conservation.

Prioritization was successful in classifying the 23 sub-watersheds into three groups of high, moderate, and low priority for soil conservation measures. The three groups were determined mainly by land use/land cover, and historical exploitation of soil resources. Other factors such as morphology (relief and slope),

and geomorphic agents, i.e., rejuvenation are also influential, although its impact varies spatially from west to east over the W.Wala catchment. The long periods of human intervention, land use abuse, and the destruction of vegetation cover was decisive in maximizing soil erosion over the sub-basins which traditionally have been utilized for rainfed farming, or rangeland transformed to rainfed cultivation since the 1950s in the eastern part of the watershed. Spatially, sub-watersheds are characterized by high priority and are found in the rainfed areas, which occupy a substantial part of the watershed, over-grazed range land, bare land, and the western rejuvenated belt of the watershed.

Accordingly, the highest priority reveals the great degree of erosion risk associated with particular sub-basins, and thus, represents a potential area for adaptation soil and water conservation measures [38] . Out of the 23 sub-basins of W.Wala, 6 sub-basins (26% of the total) (sub-basins nos. 6, 8, 10, 11, 19, 21) are classified as high priority (Figure 7). Eight sub-basins (34.8% of total) (sub-basins nos. 2, 5, 12, 13, 14, 15, 16, 29) come under moderate priority (Figure 7) for soil and water conservation. It is obvious that 62% of the sub-basins are grouped under high and moderate priority for soil and water conservation. Eleven of these sub-basins are located at the eastern-southeastern part of W.Wala watershed, and three sub-basins are located at the northwestern part of the watershed (Figure 7). Their distribution is found in accordance with the land use/land cover pattern, especially overgrazed rangeland, and bare lands where the vegetation cover has been highly degraded (Figure 4).

These sub-basins are characterized by highest relief over the entire watershed (800 > 900 m a.s.l), where slope categories of 0 - 5, 5 - 10, and 10 - 15 are

predominant. The destruction of vegetation cover, caused severe soil erosion since the Neolithic and Chalcolithic periods. Indicators of soil erosion and the existence of agricultural stone terraces, imply that the highlands of rainfed farming (including W.Wala) experienced severe soil erosion at least since the Nabatean period, some 3000 years ago [2] . High relief ratio (R_r) has also been recorded for several sub-basins pertaining to high and moderate priority for soil and water conservation (i.e., sub-basins nos. 6, 10, 11, and 16) which can be considered an essential factor in the assessment of soil erosion and sediment load in W.Wala watershed. Furthermore, high hypsometric integral (HI) values for sub-basins (e.g. nos. 2, 5, 6, 8, 12) classified as high and moderate priority vary from 70% to 89%, which indicate that these drainage basins are influenced by rejuvenation processes, and are considered at the youth-age stage of geomorphic development. Consequently, surface erosion rates are considerably high [75] .

Nine sub-basins (39.2% of the total) (sub-basins nos. 1, 3, 4, 7, 9, 17, 18, 22, and 23) are assigned to be of low priority for soil and water conservation. Three sub-watersheds out of nine consist mainly of wide and flat, undulating table lands overlooking the deep canyons downstream which occupied the rejuvenated belt. Other sub-watersheds (nos. 9, 7, 4, 3, and 1) extend to the northern/northeast part of W.Wala watershed.

The nine sub-watersheds constitute a major part of the rainfed agricultural region of Jordan, where the cultivation of wheat and barley is practiced over these sub-watersheds. Here, the average annual rainfall varies from 250 mm to 350 mm. Scattered irrigated agriculture based on groundwater wells is practiced over the sub-basins categorized as being of high and moderate priority for soil and water conservation. Irrigated farming is also practiced downstream of W.Wala (Al-Haidan area) based on extracting the water from the springs and Al-Haidan groundwater wells (Figure 8(a)). High soil erodibility of Calcaric Cambisol and Eurtic Cambisols soil type (Figure 5), steep slopes (14˚ - 20˚), poor conservation measures and degraded vegetation cover accelerates soil erosion, and exposed the sub-watersheds to high rates of soil erosion loss. However, the transformation of vast areas since the 1990s from rangeland into farming practice, and the expansion of rainfed “mixed farming” towards the marginal areas (Figure 8(b)), accelerate soil erosion and have increase the susceptibility of soil erosion over these sub-watersheds although, they rank under low priority for soil and water conservation. The cultivated land with poor conservation measures exhibits a higher rate of soil erosion and decline in soil fertility. Consequently, all sub-basins ranked under high and moderate priority are vulnerable to soil erosion, thus, they should be prioritized for conservation. Even over the sub-basins with mixed rainfed farming areas, traditional up and down slope tillage without conservation measures (i.e., contouring and terracing) is common in W.Wala. Thus, priority must be given to the protection of sub-watersheds exposed to high soil erosion rates, in order to reduce runoff coefficient, increase infiltration rates and soil moisture in soil profile. Reduction in soil erosion loss has been verified recently by Al-Sheriadeh and Al-Hamdan [26] where they concluded that in the years 1987-1990; the estimated sediments yields were less than the actual yields in the Kink Talal Reservoir of the Zerqa River. They attributed this to the implementation of a soil conservation program launched by the government of Jordan, in cooperation with the German government. Further, Al-Alawi and, Abujamous [78] estimated the average annual sediment loss in Central Jordan (Salt area) at 78 ton∙ha⁻¹∙year⁻¹ before the construction of stone terraces. Twenty years later following construction and tree planting, the estimated average soil loss have been reduced considerably to an average of 33 ton∙ha⁻¹∙year⁻¹ (35.7%). Such encouraging results emphasize the need to execute intensive soil conservation programs over the highland to reduce soil erosion rates, in accordance with enhancing practiced soil conservation techniques, and to modify RUSLE parameters (i.e., C, P, and LS factors which are considered key factors in soil

erosion). It has been argued that RUSLE parameters can be modified significantly through altering farmers environmental practices [72] with the support of local governments. Slope length and steepness factor (LS), the conservation practice factor (P), and the cover and management factor (C), can be improved noticeably to reduce soil erosion loss and to conserve water in soil on the farm to maintain crop productivity. The LS factor can be modified mainly through the construction of what is locally termed mastaba, or contour stone terraces associated with tree planting on slopes ranging from 5˚ - 25˚. Stone terraces should be aligned in long rows following the contours at various intervals depending on length and steepness of slope. On steep slopes (>25˚) with a shortage of stones, bench terraces are more appropriate [79] [80] . Terrace farming in the W.Wala highlands has been practiced since some 3000 years ago [2] . The construction of check dams across the gullies (1 - 3 m of depth) can decrease runoff coefficient and soil erosion rates under different land utilization [81] . Structural soil conservation measures can be applied on gentle slopes (0˚ - 5˚, 5˚ - 10˚), and moderate to steep slopes (10˚ - 15˚, 15˚ - 20˚) characterize the sub-watersheds of high priority for soil conservation measures (sub-basins nos. 6, 8, 19, and 21), and sub-basins with moderate priority (sub-basins nos. 2, 20, 14, 15, and 16) which are considered of high soil erosion rates, and soil loss-prone areas. It is more appropriate to integrate structural conservation measures with technology improving farming practice (i.e., rotation and contour ploughing) of rainfed cultivation to reduce soil erosion, and to enhance soil moisture in the soil profile. Such recommendations can be applied almost to all sub-basins categorized as of high, moderate, and low priority. The expansion of rainfed farming on rangeland and steep slopes (>20˚) since the 1950s, has stimulated an increase in studies showing that the presence of rock fragments on the surface were highly effective in reducing soil erosion loss. The presence of stone cobbles and fragments of different size at the soil surface at 5% to 15% coverage, caused a momentous reduction in runoff by an average of 17% and 30% respectively. The corresponding reductions in soil loss for both stone streamlets were as large as 35% and 53% respectively [80] . If plantation of trees is accompanied by stone coverage practice, it will decrease soil loss, and increase infiltration rate and soil moisture. Rehabilitation of vegetation cover through planting trees of drought-resistance species is highly recommended to improve the grazing potential of the marginal areas of the entire watershed.

The validity of the three priority classes determined through morphometric analysis, and represent the 23 sub-basins was tested using DA. The intention is to test the hypothesis that there is a significant difference between the priority classes obtained earlier, and if this hypothesis is accepted, to establish a system of coordinate axes which discriminate between the recognized priority groups. With reference to the three classes, it is obvious that there is a significant difference between low priority class, moderate class, and high priority class, with a noticeable concentration of high and moderate priority classes in the eastern part of W.Wala watershed. That is; in accordance with the deteriorated rangeland, bare land, and degraded vegetation cover as clarified in this part of the watershed earlier. Soil resources have been exhausted here over the last 3000 years, with prominent shortage of conservation measures. Statistical testing using Discriminant Analysis was performed on a data matrix representing the three priority groups (i.e. 6 × 11, 8 × 11, and 9 × 11) with the associated ranking values and including the C_p scores. The F test of Wilks Lambda obtained is F ratio is 45.7 with the degrees of freedom V1 = 2 and V2 = 20. Referring to the table of percentage points of the F-distribution, with V1 = 2 and V2 = 20, it is found that at 99.9 percent of confidence, the tabulated value is 9.95, which is significantly exceeded by the computed F ratio value (45.7). Consequently, there is a high significant difference between each of the priority classes (high, moderate, and low), and the three priority groups are highly separated and distinct. Moreover, 100 percent of the difference between the three priority groups is attributed to discriminant function 1 (81.7 percent) and discriminant function 2 (18.3 percent). It was also revealed that discriminant function 1is positively correlated with C_p, T_r, D_d, L_o, F_s, and C_c parameters. By contrast discriminant function 2 is positively correlated with B_s, and R_e parameters. The scores of each sub-basin of the three priority groups on the discriminant functions 1 and 2 were plotted in Figure 9. The plot displays highly separated and distinct priority groups. With reference to the present findings, it can be concluded that prioritization based on morphometric analysis is proven to be statistically valid, consistent and reliable, and of high capacity using the GIS platform. The potential of morphometric analysis method as elaborated by the pioneers, i.e. [31] [36] [41] [42] [81] is proven to be capable for prioritization research.