3.1 Landslide Susceptibility factors analysis
The major Landslide Susceptibility contributing factors such as elevation, slope, aspect, soil type, lithology, distance to stream, land use land cover, rainfall as well as drainage density had been applied for landslide susceptibility identification. The raster values were categorized depending on the susceptibility capability of the Chemoga watershed (Desalegn and Mulu, 2021). All the factors were categorized and weighted depending on their significance. Numerical scales one to five from very low to very high weighted factors were used. The main weighted factors were: lithology, slope, land cover, and aspect. Curvature plain, rainfall, distance to stream, and distance to roads were equally significant with an approach to slope instability(Senouci et al., 2021). For this research, the factors classified from very low to very high weighting were used. The main prominent factors were elevation, slope, aspect, and soil type concerning multi criteria decision making technique a pair-wise comparison of six criteria for AHP process. To produce landslide susceptibility map the sum of all weight multiplied by its influencing weight was carried out in the raster calculator tools by making an allowance for the above equation (2).
The slope classification > 120 has been higher input for landslide incidence. The slope classes >450 are the highest landslide level classes whereas the slope classes < 50 are the smallest ones (Mersha and Meten, 2020). For this research, the less significant slope values were level landscape with similarly, the higher slope values were sharp landscape. Based on their susceptibility to landsliding, slopes were classified into five classes. 0-50, 5-110, 11-190, 19-290 and 29-590. Generally, since the gradient raising, the possibility of landslide incidence as well increases. The other factor was aspect ranks, values of gradient ranks facing near to northeast (22.5-67.5), east (67.5-112.5) with north (0-22.5) were larger than one indicative of advanced possibility of landslide happening. The northeast-facing aspect ranks have gotten the utmost influence otherwise ranking pursued with the east-facing ones (Mersha and Meten, 2020). The chemoga watershed aspect ranks are the values of slope ranks facing near the flat (-1), north (00-22.50), north-east(22.50-67.50), east(67.50-112.50), south-east(112.50-157.50), south(157.50-202.50),south-west(202.50-247.50),west(247.50-292.50),north-west(292.50-337.50),with north(337.50-3600).
In the case of lithology, five components i.e. Ja (sandstone), P2a (Deeply weathered alkaline as well as traditional basalts flows through rare intercalations of tuff often tilted), PNtb (Alkaline to traditional basalts often forming shields volcanoes with minor Trachyte and phonolite flows), ARI (Biotite and hornblende gneisses, granulite and migmatite with minor metasedimentary gneisses), gt4 (Tectonic granite and syenite) have a high possibility of landslide incidence. Ja, P2a and PNtb basalt have been less strong and for this reason susceptible to landslides. The form of land use land cover can be ordered the incident of landslide in the chemoga watershed areas. Land use land cover of chemoga watershed areas was reclassified within nine frequent ranks and was enhanced into the raster layer. The highest weights were examined within the land-use class of grassland, cultivation land, bare land, plantation, wetland, and shrub land indicative of a high possibility of landslide incidence. The correlation among landslide incidence and distance from the river, since distance from river raise, the happening of landslide normally reduces (Mersha and Meten, 2020). In the case of the chemo watershed study, distance from stream land sliding factors, when the distance from stream increase, the incidence of landslide to reduce. Landslide incidence was higher with the following ordered classification of 0-200m, 200-400 m, 400-600m, 600-800m and 800-1000m. As considered to the contributory landslide factors of rainfall, can be classified rainfalls were provided into three classes value, influence with resulting very high landslide value was the least valuable one, viewing amazingly low effect insignificant very low land sliding rate. Consequently, the area with extremely high rainfalls is showed to the classification, 900 – 1200, 1200 – 1500, and1500 – 1800 mm and the study area has been extremely low rainfalls are ranked to class one, as illustrated in the list based on their high value.
The study areas were applied to calculate a drainage density through spatial analyst tool. Depending on a categorization technique, the one that has higher values are extremely affected with landslide and ranked to five (>0.0184 /km) with area coverage of 0.37%, high ranks to four (0.00245–0.0184 /km) with area coverage of 2.91%, moderate ranks to three (0.00185–0.00245 /km) with area coverage of 13.25%, low ranked to two (0.00184–0.00185 /km) with area coverage of 25.09% and extremely low ranked to one (<0.00079 /km) with area coverage of 58.38%.
3.1 Analytical Hierarchy Process and Weighting overlay
Analytical Hierarchy Process was applied to estimate the weighting overlay and ratings for class influential weights: elevation, slope, aspect, Soil type, Lithology, Distance to stream, Land use land cover, rainfall and drainage density were obtained from pair-wise comparison matrix. The matrix was applied to the class rating of each constraint and to calculated the Criteria weight value in excel software. As showed in Table 4, elevation was the most influential value factor, with a value of 0.27. The slope also has been the significant value of 0.18, followed by aspect value which was 0.16, soil type which was 0.12, and lithology which was 0.08 the rest ones were listed in Table 4.
Table 9
A matrix of pairwise comparisons of six criteria for the AHP process
|
Elevation
|
Slope
|
Aspect
|
Soil type
|
Lithology
|
Distance to stream
|
Land use/cover
|
Rainfall
|
Drainage density
|
Elevation
|
1
|
5
|
5
|
3
|
3
|
5
|
3
|
5
|
3
|
Slope
|
1/5
|
1
|
3
|
3
|
5
|
3
|
5
|
5
|
3
|
Aspect
|
1/5
|
1/3
|
1
|
3
|
5
|
5
|
3
|
5
|
5
|
Soil type
|
1/3
|
1/3
|
1/3
|
1
|
3
|
5
|
3
|
5
|
5
|
Lithology
|
1/3
|
1/5
|
1/5
|
1/3
|
1
|
3
|
3
|
5
|
3
|
Distance to stream
|
1/5
|
1/3
|
1/5
|
1/5
|
1/3
|
1
|
3
|
3
|
3
|
Land use land cover
|
1/3
|
1/5
|
1/3
|
1/3
|
1/3
|
1/3
|
1
|
3
|
5
|
Rainfall
|
1/5
|
1/5
|
1/5
|
1/5
|
1/5
|
1/3
|
1/3
|
1
|
3
|
Drainage density
|
1/3
|
1/3
|
1/5
|
1/5
|
1/3
|
1/3
|
1/5
|
1/3
|
1
|
Table 10
Pair-wise comparison decimal matrix
|
Elevation
|
Slope
|
Aspect
|
Soil type
|
Lithology
|
Distance to stream
|
Land use/cover
|
Rainfall
|
Drainage density
|
Elevation
|
1
|
5
|
5
|
3
|
3
|
5
|
3
|
5
|
3
|
Slope
|
0.2
|
1
|
3
|
3
|
5
|
3
|
5
|
5
|
3
|
Aspect
|
0.2
|
0.33
|
1
|
3
|
5
|
5
|
3
|
5
|
5
|
Soil type
|
0.33
|
0.33
|
0.33
|
1
|
3
|
5
|
3
|
5
|
5
|
Lithology
|
0.33
|
0.2
|
0.2
|
0.33
|
1
|
3
|
3
|
5
|
3
|
Distance to stream
|
0.2
|
0.33
|
0.2
|
0.2
|
0.33
|
1
|
3
|
3
|
3
|
Land use land cover
|
0.33
|
0.2
|
0.33
|
0.33
|
0.33
|
0.33
|
1
|
3
|
5
|
Rainfall
|
0.2
|
0.2
|
0.2
|
0.2
|
0.2
|
0.33
|
0.33
|
1
|
3
|
Drainage density
|
0.33
|
0.33
|
0.2
|
0.2
|
0.33
|
0.33
|
0.2
|
0.33
|
1
|
Sum
|
3.12
|
7.92
|
10.46
|
11.26
|
19.19
|
22.99
|
21.53
|
32.33
|
31
|
Table 11
Normalized pair-wise matrix calculated
|
Elevation
|
Slope
|
Aspect
|
Soil type
|
Lithology
|
Distance to stream
|
Land use/cover
|
Rainfall
|
Drainage density
|
Elevation
|
0.32
|
0.63
|
0.48
|
0.27
|
0.16
|
0.22
|
0.14
|
0.15
|
0.10
|
Slope
|
0.06
|
0.13
|
0.29
|
0.27
|
0.26
|
0.13
|
0.23
|
0.15
|
0.10
|
Aspect
|
0.06
|
0.04
|
0.10
|
0.27
|
0.26
|
0.22
|
0.14
|
0.15
|
0.16
|
Soil type
|
0.11
|
0.04
|
0.03
|
0.09
|
0.16
|
0.22
|
0.14
|
0.15
|
0.16
|
Lithology
|
0.11
|
0.03
|
0.02
|
0.03
|
0.05
|
0.13
|
0.14
|
0.15
|
0.10
|
Distance to stream
|
0.06
|
0.04
|
0.02
|
0.02
|
0.02
|
0.04
|
0.14
|
0.09
|
0.10
|
Land use land cover
|
0.11
|
0.03
|
0.03
|
0.03
|
0.02
|
0.01
|
0.05
|
0.09
|
0.16
|
Rainfall
|
0.06
|
0.03
|
0.02
|
0.02
|
0.01
|
0.01
|
0.02
|
0.03
|
0.10
|
Drainage density
|
0.11
|
0.04
|
0.02
|
0.02
|
0.02
|
0.01
|
0.01
|
0.01
|
0.03
|
Table 12
Determined relative criterion weights
|
Elevation
|
Slope
|
Aspect
|
Soil type
|
Lithology
|
Distance to stream
|
Land use/ cover
|
Rainfall
|
Drainage density
|
Weighted
|
Elevation
|
0.32
|
0.63
|
0.48
|
0.27
|
0.16
|
0.22
|
0.14
|
0.15
|
0.10
|
0.27
|
Slope
|
0.06
|
0.13
|
0.29
|
0.27
|
0.26
|
0.13
|
0.23
|
0.15
|
0.10
|
0.18
|
Aspect
|
0.06
|
0.04
|
0.10
|
0.27
|
0.26
|
0.22
|
0.14
|
0.15
|
0.16
|
0.16
|
Soil type
|
0.11
|
0.04
|
0.03
|
0.09
|
0.16
|
0.22
|
0.14
|
0.15
|
0.16
|
0.12
|
Lithology
|
0.11
|
0.03
|
0.02
|
0.03
|
0.05
|
0.13
|
0.14
|
0.15
|
0.10
|
0.08
|
Distance to stream
|
0.06
|
0.04
|
0.02
|
0.02
|
0.02
|
0.04
|
0.14
|
0.09
|
0.10
|
0.06
|
Land use land cover
|
0.11
|
0.03
|
0.03
|
0.03
|
0.02
|
0.01
|
0.05
|
0.09
|
0.16
|
0.06
|
Rainfall
|
0.06
|
0.03
|
0.02
|
0.02
|
0.01
|
0.01
|
0.02
|
0.03
|
0.10
|
0.04
|
Drainage density
|
0.11
|
0.04
|
0.02
|
0.02
|
0.02
|
0.01
|
0.01
|
0.01
|
0.03
|
0.03
|
Table 13
The weights of all landslide factors value
Landsliding Factor
|
Weighted (%)
|
Elevation
|
27
|
Slope
|
18
|
Aspect
|
16
|
Soil type
|
12
|
Lithology
|
8
|
Distance to stream
|
6
|
Land use land cover
|
6
|
Rainfall
|
4
|
Drainage density
|
3
|
Table 14
Determined Consistency Ratio (CR)
|
Elevation
|
Slope
|
Aspect
|
Soil type
|
Lithology
|
Distance to stream
|
Land use/ cover
|
Rainfall
|
Drainage density
|
Weights sum
|
Weighted
|
Weights sum/weighted
|
Elevation
|
0.27
|
0.90
|
0.80
|
0.36
|
0.24
|
0.30
|
0.18
|
0.20
|
0.09
|
3.34
|
0.27
|
12.37
|
Slope
|
0.05
|
0.18
|
0.48
|
0.36
|
0.40
|
0.18
|
0.30
|
0.20
|
0.09
|
2.24
|
0.18
|
12.44
|
Aspect
|
0.05
|
0.06
|
0.16
|
0.36
|
0.40
|
0.30
|
0.18
|
0.20
|
0.15
|
1.86
|
0.16
|
11.62
|
Soil type
|
0.09
|
0.06
|
0.05
|
0.12
|
0.24
|
0.30
|
0.18
|
0.20
|
0.15
|
1.39
|
0.12
|
11.58
|
Lithology
|
0.09
|
0.04
|
0.03
|
0.04
|
0.08
|
0.18
|
0.18
|
0.20
|
0.09
|
0.39
|
0.08
|
4.87
|
Distance to stream
|
0.05
|
0.06
|
0.03
|
0.02
|
0.03
|
0.06
|
0.18
|
0.12
|
0.09
|
0.64
|
0.06
|
10.66
|
Land use land cover
|
0.09
|
0.04
|
0.05
|
0.04
|
0.03
|
0.02
|
0.06
|
0.12
|
0.15
|
0.6
|
0.06
|
10
|
Rainfall
|
0.05
|
0.04
|
0.02
|
0.02
|
0.02
|
0.02
|
0.02
|
0.04
|
0.09
|
0.32
|
0.04
|
8
|
Drainage density
|
0.09
|
0.06
|
0.03
|
0.02
|
0.03
|
0.02
|
0.01
|
0.01
|
0.03
|
0.30
|
0.03
|
10
|
The RI within grouping by λmax is applied for calculation of Consistency Index and calculates values if < 10 % found judgments are reliable. Within this research area, an effort had been completed to recognize the landslide area of Chemoga watershed using nine spatially distributed landslide hazard parameter factors like elevation, slope, aspect, soil type, lithology, distance to stream, land use land cover, rainfall, as well as drainage density in Saaty’s AHP based MCDA support tools. The influence gets hold of from Saaty’s AHP analysis was applied to obtain ultimate priority of Chemoga watersheds. The reliability of personal decisions could be verified with an approximate reliability ratio that is evaluation among consistency index and random consistency index. The consistency index (CR) has been calculated the following equation: The Consistency Index (CI) can be determined of exit from reliability, was determined with the formula:
CI= \(\frac{\lambda -n}{n-1}\) …………………………………….. …………………………………………… (3)
Where; n is the number of factors (i.e. 9) and λ is the mean value of the reliability vector find out in the above Table (14).
λ = (12.37+12.44+11.62+11.58+4.87+10.66+10+8+10)/9 =10
Depend on the above equation, CI =10-9/9-1=0.13
To evaluate the strength of the author's outlook the consistency ratios (CR) were calculated with equation (1). Where, RI is random inconsistency index values based on number (n) of factors being balanced; for n = 9, RI = 1.45 as showed in Table 8 above.
$$CR=\frac{0.13}{1.45}=0.09$$
As a result, 0.09 < 0.1, this value point out there was a pragmatic degree of reliability within the pair wise comparison and therefore, the criteria weight values were 0.27, 0.18, 0.16, 0.12,0.08, 0.06, 0.06, 0.04, and 0.03 had been allocated to Elevation, Slope, Aspect, Soil type, Lithology, Distance to stream, Land use land cover, Rainfall and Drainage density respectively.
3.2 Landslide Susceptibility Maps
In the chemoga watershed research areas were used, GIS-based Analytical Hierarchy Process as multi-criteria analysis approach to mapping the prospective landslide incidences. The analytical hierarchy process model was predictably dependent on the ranking classification afforded with expert judgment. The expert outlook was incredibly helpful within determine difficult harms seems like landslide incidence. Depend on landslide susceptibility maps results were categorized into five ranks: very low, low, moderate, high and very high as showed in Figure 12 below. A very low area susceptible to landsliding coverage was 550.61km2(46.52%), low areas susceptible to landsliding coverage was 163.72km2(13.83%), moderate areas susceptible to landsliding coverage was 221.51km2(18.71%), high areas susceptible to landsliding coverage was 182.15km2(15.39%), very high areas susceptible to landsliding coverage was 65.64km2(5.55%) of the total chemoga watershed research area in Figure12 & Table 15.
Table 15
Chemoga watershed area of landslide susceptibility map ranks
Landsliding susceptibility ranks
|
Area in (km2)
|
Area in (%)
|
Very Low
|
550.61
|
46.52
|
Low
|
163.72
|
13.83
|
Moderate
|
221.51
|
18.71
|
High
|
182.15
|
15.39
|
Very High
|
65.64
|
5.55
|
3.3 Validation of the model
A final landslide susceptibility map of the chemoga watershed should be validated. Without model validation, the susceptibility map does not give significance (Mersha and Meten, 2020). A predictive model map was constructed by using ArcSDM with the extension of ArcGIS tools to generate AUC and overlying of the landslides with GPS field data and Google earth data over the causative factor. Depend on Area under curve calculation (AUC) a value of AUC were classified regarding 0.9-1.0 ranks as test quality was excellent, 0.8- 0.9 ranks as test quality was very good, 0.7- 0.8 ranks as test quality was good, 0.6- 0.7 ranks as test quality was satisfactory, 0.5- 0.6 unsatisfactory, under this criteria, since the spatial efficiency of the produced landslide susceptibility maps verified with AUC (76.7% of accuracy), which was observed that applied model capitulated a good outcome for landsliding susceptibility maps within research areas (El Jazouli et al., 2019). As a result, chemoga watershed landslide susceptibility map result showed very good accuracy of AUC value 81.45%.