Hydrologic soil group based curve number matrix modeling for Enset-Based land 1 use system in Meki River Watershed, Western Lake Ziway Sub-Basin, Central 2 Rift Valley of Ethiopia

Enset-Based land use system (EBLUS) exhibits good carbon stock and infiltration rate equivalent 23 to forest covered areas, which enhances infiltration and water holding capacity and it can 24 reduce the curve number (CN) of the watersheds but it was not considered in former studies. Therefore, this study is planned to model the hydrologic soil group (HSG) based CN matrix of 26 EBLUS relative to other LUSs with established hydrological characteristics in the Meki river 27 watershed. The soil data is used to determine the HSG of the watershed collected from Ministry 28 of Water, Irrigation and Energy (MOWIE) and verified by Harmonized World Soil Database 29 (HWSD). A Model is developed for CN of EBLUS relative to other LUSs (Alemu’s formula). The 30 model considers both infiltration rate measured using Amozi-meter and carbon stoke of soil 31 weighed as 85% and 15% respectively. HEC-GEO-HMS model is used to consider the CN of EBLUS 32 as a separate LUS to verify the developed CN matrix model to generate CN of the sub- 33 watersheds. Significant reduction in mean CN of the watershed that shows the role of EBLUS in managing 41 the water resources and flood is high. Therefore, escalating EBLUS will reduce the CN of the 42 watershed which reduces runoff volume in the watershed and it ensures the sustainability of 43 Lake Ziway by reducing sedimentation.


Sampling and Measurement techniques 128
In order to get and compare the hydrological characteristics of EBLUS, the following sampling 129 matrix was prepared as shown in Table 1. In the process of sampling, land use class is crucial for 130 the relative comparison of hydrological components (infiltration capacity) with replication. 131 The study area is classified in to eight land cover classes that include forest and natural 132 The vegetation zones are combined with dominant soil types as vegetation zone 1 with soil type 140 1 (Z11) up to vegetation zone 3 with soil type 4 (Z34). Vegetation zones are expressed as Afro-141 alpine as vegetation zone 1, Dry Afro-montane as vegetation zone 2 and Acacia wooded grass 142 land of rift valley as vegetation zone 3 verified in the field as shown in Figure 4 below and the 143 dominant soil types considered are Vertisol, Cambisol, Luvisol and Leptosol and called to be soil 144 type 1, soil type 2, soil type 3 and soil type 4 respectively as shown in Table 1.
There are about 60 sampling possibilities, but EBLUS is not common at the acacia wooded part of the watershed for which zones Z31 to Z34 to all LUSs are not applicable in this research.
The resulting sampling points are Z11 up to Z24 (8 combined zones of two vegetation zones and four dominant soil types) for five LUSs which results in 40 sampling points replicated to four fold to make it representative and to reduce human and instrumental errors.  The water is released to flow down to the ground and time and depth of flow recorded using a stopwatch and gauge fitted to the instrument respectively based on the procedural manual.

Review of Carbon stoke of Land use systems
Published articles are reviewed to get the carbon stock of different land use systems which accounts EBLUS. According to Mesfin et al, 2017, carbon stoke of land uses are measured and reported as shown in Table 2 and considered to develop the CN of EBLUS.

Preprocessing and CN model verification procedure in HEC-GeoHMS
Soil conservation service curve number grid is used by many hydrologic models to extract the curve number for watersheds (Fleming & Brauer, 2018;Merwade, 2012)  preprocessing phase of the project. Basin merge and river merge processes are taken place to increase consistency and convenience of the result output.
According to Fleming and Brauer (2018), in order to perform CN modeling, various types of information are required that includes watershed parameters from DEM, HSG and LUSs and the CN grid is used by many hydrologic models (Merwade, 2012).

Land use data preparation for CN grid
Land use map was generated using Land sat image data (30m) supported by Google earth. GPS based field visit is performed to collect data to train ERDAS 2014 to classify the images with maximum likelihood clustering algorithm of supervised classification method (Iliasse and Adil , 2014; Fleming and Brauer 2018) and eight Land use systems are identified as shown in Figure 6 and reclassified into five classes as shown in According to Fleming and Brauer (2018) and Merwade (2012), the Spatial Analyst Tools in Arc Toolbox is used to implement reclassification  According to Fleming and Brauer (2018) and Merwade (2012), the Spatial Analyst Tools in Arc Toolbox is used to implement re-classification In the reclassification window, confirm the Input raster is LULC_2017_March2019Recl field is Class_Name, and then manually assign the new numbers from Table 4 as shown in Figure 7 and the output raster is saved as LULC_2017_March2019Recl.  Add Soil_WZ_2018_1 feature class from spatial dataset collected from Ethiopia MOWIE and in its attribute table create an empty field for storing soil group data as shown in Table 5. Hence the hydrologic soil group data can be populated to Soil_WZ_2018_1 after identifying the type of soil with its corresponding description.
Ethiopian Road Authority (2013) manual is used to generate the relationship of soil type and hydrologic soil group of Meki river watershed and also verified by HWSD viewer and also shape file data acquired from MOWIE as shown in Table 5. Accordingly, the HSG is assigned to each of the six soil types identified in the watershed as shown in Figure 9. During the assignment the HSG of Leptosol is assigned after referring the field characteristics of the soil and it has more similarity to Luvisol with its insignificant areal coverage to influence the value of the curve number and hence HSG of C is assigned to it. The preparation of soil data is over at this point. The next step is to merge/union both soil data and land use data to create polygons that have both soil and land use information. Save the map document.

Merging of Soil and Landuse Data
To merge/union soil and land use data, use the Union tool in Arc Tool box available under Analysis Tools -Overlay. Browse/drag Soil_WZ_2018_1 and LULC_2017_March2019Pol as input features, name the output feature class as "Meki_Soil_LU", leave the default options, and click OK as shown in Figure 11.   Figure 13.

Determination of watershed based curve number
Curve number is extracted from the soil type and land use data considering EBLUS using HEC-GeoHMS, which both affect the infiltration capacity of the soil and the Soil data acquired from the Ethiopian Ministry of Water Irrigation and Energy (MWIE) and Landsat image from USGS (Fleming and Brauer, 2018).
The factors of the CN model developed from the raster format of soil and LUS data with the same coordinate system (UTM WGS 1984 37 0 North) with a spatial resolution of 30m. Finally, the result is extracted and reported for the classified LUS of Meki river watershed and also it is extracted to 34 sub-watersheds and two major growing zones (Enset growing and non-Enset growing zones) of Meki river watershed.  Table 6.

Infiltration capacity of the soil
Sandy loam soil textural class has higher infiltration capacity than other textural classes followed by silt loam in all land use systems. High infiltration capacity is measured in the natural forest covered portion of the watershed followed by EBLUS. EBLUS improves the infiltration and water holding capacity of the soil by increasing the organic matter content of the soil through litter and pseudo stem cover falls and also the root fiber of 2 to 3 meters long away from the edge of the pseudo stem measured in the field that can enhance the void space of the soil to transmit rain water easily to the ground. The presence of wide leaves protects the direct impact of the rain drop (Kebede Wolka, et al, 2015) and permit more through fall which reduces the speed of the rain drop and it will give more time for rain water to infiltrate to the ground. The Upper zone of the watershed has higher mean infiltration rate (11.0125mm/hr) than the Middle zone (10.92625mm/hr) of the watershed. The high mean infiltration rate at the upper zone shows the presence of more forest LUS and EBLUS than the middle zone of the watershed which enhances infiltration and water holding capacity of the soil. In the cultivated LUS, high rate of infiltration is recorded in upper zone of the watershed while middle zone has more infiltration rates in Grass LUS, forest LUS and EBLUS as shown in Figure 15, which is attributed to the high slope of upper zone that influences the carbon stock of the soil. (>α) as shown in Table 7. Therefore, a high mean infiltration rate in Forests and EBLUS show that there is an improvement in hydrological parameters for those land use systems to enhance water absorption to the ground water system. The improvement in infiltration capacity has a direct influence on water resources management. In addition, the increase in infiltration rate has a huge contribution in runoff reduction and alleviation of sedimentation problems in Meki river watershed.  & Fohrer, 2009). According to Mesfin, et al (2017) and Barbora and Jaroslava (2014), compost improves and accelerated both the infiltration and water holding capacity of the soil for a longer period which in turn influences the CN of the watershed.

Curve number modeling result
EBLUS exhibited a good carbon sequestration, which is equivalent to high-vegetation areas (Mesfin et al., 2017). Hence, Carbon stock considered as criteria to model the curve number of EBLUS next to the infiltration capacity of the soil as reported in Table 8.  Ci and CSq are coefficients for the relative influence of infiltration capacity and Carbon stock on curve number of land use systems respectively at 85% to 15% proportion.
Accordingly, the curve numbers are generated for EBLUS for each hydrologic soil group with their respective infiltration capacity and Carbon stock relative to other LUSs for which curve number is encoded and presented in Table 9. Carbon stock is modified by growth period and energy production of those LUSs. Hence, infiltration capacity of the soil powers 85% of the curve number of LUSs while 15% of the curve numbers of LUSs are influenced by carbon stock which is computed.  Table 10 and fed to HEC-GEO-HMS together with that of the union of LUS and soil information in order to generate the curve number grid.  Figure 16. This CN grid can be used in different models of rainfall-runoff modeling purposes and also researchers and runoff flow simulators. Figure 16: Average soil loss from sub-watersheds with respect to Rift valley limit

Zone and sub-watershed based CN determination
CN of EBLUS is evaluated for two enset growing zones of Meki river watershed and the mean values of the upper zone is less than the middle one due to the high proportion of forest and EBLUS in the upper zone of the watershed as shown in Table 11.  Figure 17.

Scenarios on mean CN of Meki river watershed with respect to EBLUS
Meki river watershed has a mean CN of 78.21 considering EBLUS as it has its own CN matrix while a mean CN of 81.72 by considering EBLUS as cultivated LUS as shown in the Table 12 below.  The CN of EBLUS is computed with respect to other LUSs. According to Chow et al, (1988) and Merwade (2012), the precipitation excess is a function of cumulative precipitation, soil type, land use systems and antecedent moisture. Considering the initial loss and the potential maximum retention, the precipitation excess can be calculated; the maximum retention and the basin characteristics are related through the curve number. The standard SCS curve number method is based on the following relationship between rainfall depth, P, and runoff depth, Q an input.
For the annual daily average rainfall of 50mm for Meki river watershed (ENMA, 2017), the runoff considering and without considering EBLUS is computed respectively as: Where: Q is surface runoff (mm), P is precipitation (mm), S is the soil retention (Sorpitivity) (mm) and Ia is initial abstraction or initial loss (mm) and CN is curve number Hence a difference of runoff occurs due to consideration of EBLUS by 3.59 mm which can be cumulate to a huge volume of runoff (8340916m 3 ) from the whole area of Meki river watershed which can be a sign post for the improvement of the hydrological characteristics of the watershed by increasing the infiltration capacity of the soil.

Conclusion& Recommendation Conclusion
Meki river watershed is dominantly covered by cultivated LUS (41.5%) followed by Bush and Therefore, a high mean infiltration rate in Forests and EBLUS shows that there is an improvement in hydrological parameters for those LUSs to enhance water absorption into the ground water system. The improvement in infiltration capacity has a direct influence on water resources management. In addition, the increase in infiltration rate has a huge contribution in runoff reduction and alleviation of sedimentation problems in the Meki river watershed.

Recommendation
A significant CN reduction due to EBLUS shows there will be an improvement in the hydrological characteristics of the watershed by increasing the infiltration capacity of the soil and also by increasing the canopy cover of the area. Hence, increasing the coverage of EBLUS can reduce the CN and runoff volume which could be the cause of flooding in different parts of the watershed and also protect Lake Ziway against sedimentation.
Therefore, creating separate land use policy to EBLUS and incorporating other fruit as an agroforestry to it will create harmony to the existing ecology. The extension program has to be initiated to the expansive production and processing of EBLUS which can help the production of sufficient inputs to the industries to be established in the area.