4.1. LULC of the Study Area
The Land sat ETM+ of 2016 satellite image was used to classify the LULC map of the study area by using the supervised image classification technique. Hence, the LULC of the study site was classified in to five main classes namely, forest land, cropland land grass, shrub land, and bare land as it is shown in figure (3).
Based on the classification result, forest area covers about 5.8% of the total land mass and mostly forests are found in the highland parts of the study area. Almost greater than half of the total land mass of the study area is covered by crop land. It consists of 68% of the total area. Grass land; on the other hand, share 7% of the whole area, which is the third dominant LULC type and which is observed in the central and eastern parts of the study area. As far as Shrub land concerns, this type of land use/cover is mostly found in the eastern and central low land part of the study area, and the second dominant land use type which consists of 18% of the total areal coverage. Lastly bare land is found in the south east regions of the study area and which accounts a very small proportion shares only 1.21% of the total land masses.
4.1.2 Accuracy Assessment
Based on Ismail and Jusoff (2008) the agreement criteria for Kappa statistics (K) are defined as, poor when K < 0.4, good when 0.4 < K < 0.7 and excellent when K > 0.75. Accordingly, the LULC classification of the study area was 0.86, indicating that the agreement criteria were excellent. Its overall accuracy is 89 % with its kappa coefficient of value of 0.86. Hence a value of 0.86 kappa coefficient can highly support the strong arguments of the land use/cover classification system. In addition to Kappa coefficient producer’s accuracy and user accuracy was also calculated. the user and producer accuracy are minimal in the case of crop land compared to the rest of the other LULC types; this is because of the fact that, the crop land is mostly similar with the grass land while the LULC identifications process was running.
Table 1
Accuracy assessment of the study area LULC classification matrix
LULC
|
Forest
|
Crop
|
Grass
|
Shrub
|
Bare
|
producer accuracy
|
user accuracy
|
Total
|
Forest
|
15
|
0
|
0
|
0
|
0
|
88.2
|
100
|
15
|
crop land
|
1
|
17
|
2
|
2
|
0
|
85
|
77.23
|
184.23
|
Grass
|
1
|
1
|
17
|
0
|
0
|
89.48
|
89.5
|
197.98
|
Shrub
|
0
|
2
|
0
|
14
|
0
|
87.5
|
87.5
|
191
|
bare land
|
0
|
0
|
0
|
0
|
14
|
100
|
100
|
214
|
Total
|
7
|
20
|
19
|
16
|
14
|
|
|
86
|
4.1.3 Flood Risk evaluation of LU/LC Type
According to Morita (2014), the LULC of the study area is the primary concern in flood risk mapping. Land cover which is covered by grassland or the cover of other forests, is better than bare lands for water store to reduce runoff. Hence the availability of vegetation thickness can slow down the runoff. On the other hand, land uses like bare lands, buildings and roads can accelerate soil erosion and intensified flooding since their infiltration capacity is very low. As a result of this, the type of LU/LC plays a key role in determining of flood occurrence for particular topographic features.
Since there is no uniform classification system about LU/LC reclassification for flood cause due to variation in LULC from place to place and even the variation of flood frequency and intensity, the local experts were selected and discussed with the reclassifications of LULC for causing of flooding.
Table 2
Expert determinations of LULC for flood cause
LULC types
|
Risk level
|
Very high risk
|
High risk
|
Moderate
|
Low risk
|
Very low risk
|
Forest
|
|
|
|
|
X
|
Bare land
|
X
|
|
|
|
|
Crop land
|
|
X
|
|
|
|
Grass land
|
|
|
X
|
|
|
Shrub land
|
|
|
|
X
|
|
The experts of Guba Lafto District of Agricultural Office were determining the LULC cause for flooding by considering the characteristics of LULC for runoff resistance. Therefore, by considering their impact for flooding with discussing the selected key informants of Guba Lafto District experts and the nature of each land use to its water holding capacity, it was further reclassified in to five risk classes. Based on this, bare land, cropland, grass, shrub and forest lands are respectively assigned as very high (1), high (2), moderate (3), low (4) and very low (5) for causing of floods. Therefore, the following map shows the land use/cover factor map of the study area.
Figure 4.2 implies that a large proportion of an area is covered by crop lands 68944.2 % (68.3 ha) which is assigned as high flood risk classes, therefore, the probability of an area to be affected by flood risk is higher when flood risk is evaluated by using the LULC parameter as it is supported by the experts of agriculture of Guba Lafto District. Beside this, only 524.2 ha (1 %) is covered by bare land which is already assigned as the most cause for flooding as the capacity of bare land to resist flood is very poor. While, 5855.7 (5.9%) of the total area is covered by forest lands which is very good for flood resistant. Grass land and shrub land is moderately affected floods which consists of the parts of the total land shares (7210.7 ha or 7.4 % and 18338.2 ha or 18.2 %) respectively.
4.2 Soil Types and their Relation to Flooding
There are five soil types identified in the study area, namely cambisols, regosols, pherosols, leptosols and xerosols.
Cambisols have good structure and mainly the degradation activities are not highly affected it since their high resistance for such types of processes. It contains at least some weatherable minerals in the silt and sand fractions. This soil type is mostly found in areas where there is adequate amount of rainfall like in temperate and boreal regions. Cambisols are medium textured and have good structural stability, high porosity, and good water holding capacity and good internal drainage.
FAO (2007) focused up on the concept of Regosols to mean well drained, medium textured, deep mineral soils derived from unconsolidated materials and separated them from shallow soils (Lithosols, Leptosols, etc.) and from those with sandy or coarser textures (Arenosols) and are mainly recent deposit earth materials.
Leptosols are soils, which are limited in depth by continuous hard rock within 30 cm of the soil surface, or contain or overlie within the same depth material with very high calcium carbonate content, or are very grave. Leptosols represent the initial phases of soil formation and are the products of severe erosion. The concept of Leptosols is to comprise all shallow, or very stony, soils overlying rock, partially altered rock or strongly calcareous material, or soils with a limited amount of fine earth material.
Most Leptosols are under natural vegetation, which is generally richer on calcareous ones than on the acid types. The main physical constraint of Leptosols is their low water holding capacity, which makes them very susceptible to drought stress. Leptosols have severe physical limitations for arable cropping, but have a certain potential for trees and for extensive grazing. The better types are the ones developed on limestone under a humid climate. Tree roots find anchorage by entering fissures. In mountain regions, soil erosion is a major problem with Leptosols under arable crops.
Pherosols have a good structure and are generally resistant to erosion. The impact of yiled products is greatly affected if those type of soils are eroded, (FAO, 2007). The favorable physical and chemical properties, especially the stable structure, high porosity and high available water capacity, high levels of organic matter, relative richness in nutrients and medium to high base saturation make these soils excellent farm land.
Xerosols on the other hand is the deserts soils, with low levels of organic matter and mainly Subject to wind erosion and concentration of soluble salts.
4.2.1 Soil flood Risk Assessment for Flooding
Soil type data were again reclassified in to five flood risk classes by using arc tool box in arc GIS 10.3 by considering its water holding capacity for flooding as each type of soils have different impact for flooding.
Based on the nature of soil types for water holding capacity, the existing five soils were further categorized according to their risk class for flooding, (FOA, 2007). Similar studies for instance, Ryutaro (2014) classified soil type in to five risk levels namely very high, high, moderate, low and very low, based on the study area soil characteristics, since the type of soil is unique across regions. As a result of this, by considering the physical characteristics of the available soil to their water holding capacity, the soil types were classified in to five risk levels. Leptosols is assigned as very high risky (1) to flooding while xerosols, regosols, phemosols and Cambisols, are weighted as high (2), moderate (3), low (4) and very low (5) risky for flood risk zone respectively.
Based on the evaluation of soil parameter for flood risk for the study area, almost 31 % of the total land masses are at risk, this is because of the poor water holding characteristics of leptosols and regosols, which consists of 31 % together. As a result of this, leptosols and regosols are assigned as the main cause of flood for the study area since the soil property of water infiltration ability is very low as it is briefly discussed in section three. The following figure shows the flood risk map of soil for the study area.
4.3 Slope and its Risk Assessment for Flooding
The slope of an area plays a substantial role in controlling of water runoff. The nature of slope greatly affects or influences the runoff direction and its amount to reach in a particular drainage. Moreover, slope also influences the infiltration capacity of topographic features. The nature of slope and its relation with soil and other earth materials like drainage and lithological property can greatly affect surface runoff. For example, a smooth/flat surface that permit the water to flow faster and causes flooding, on the other hand, a terrain with roughness and steep slope can slow down its speed. In general, in the case of flood risk impact, Steeper slopes have more probability to damage by surface runoff, while flat terrains are susceptible to water logging.
According to (Ozcan 2010), flatter surface slopes are highly vulnerable to flood occurrences compared to steppe slopes. This is because of water from the river always gathers in an area where the slope gradient is usually low as water flow from high elevation to the lower. Area with higher gradient slope cannot able to accumulate water and tends to cause flooding in case of river type of flooding. But in this study, the main concern is mainly with flash flood types and the lower slope is the most vulnerable and vice versa. Hence, the concern is mainly for flash flood, so that the classification and weighted of soil for flood causing constraints is given by considering of it.
The slope with lower value is flatter and the highest value is the steeper feature of a terrain. Hence based on the susceptibility to flooding; slope have been classified into five classes. Therefore, the lowest slope area is very highly affected by flood and then ranked to class 1, which is less than 10 %. Following the very high classes, there is class high (20%) ranked 2, moderate (up to 30%) ranked 3, low (40%) ranked 4 and, and class very low (> 40%) ranked as 5,
As to most literatures, slope can be classified in to different ways according to the criteria and the nature of the study area. Therefore, based on the nature of the study area, this study was classifying the slope factor map of flood in to eight classes, and it again reclassified in to five suitability classes namely, very risky to flooding up to very low risky for flooding for gentle slope to steppe slope respectively in order to identifying the areas of flood risk.
The study result reveals that almost 58 % of the total land mass is at a high risk of flooding, while the 20 % is moderately risky and the remaining 19 % is at low risk of flooding. This means that the study area is the most risk for flooding when slope is the evaluated with the slope criteria. The central and eastern parts of the study area are generally at flood risk levels in the case of slope evaluation parameters, since the area is flatter and characterized by gentle slope.
4.4 Elevation assessment for Flood Risk
Different elevation has different capacity to resist a flood; for instance, When the elevation value of a topography is increase, then its probability to exposed for flood related risk will be low compared to the elevated areas and vice versa. This is because of the fact that runoff is began from the upper part of a surface to the lower part. As a result of this, low laying areas will be victimized from the hit of such hazards and sedimentation of different particles which comes from the upper areas
According to (Morita, 2014), the low laying areas are the most flood hit and the reverse is true for the elevated areas.
The study result indicated that 16090.28 (ha) or 15.96 (%) of the study area is found at a very high-risk zones while 24011.4 (ha) or 23.81 (%) of the study area is at high risk. The remain 19286.52 (ha) or 19.12 (%), 15032.28 (ha) or 14.91 (%) and 26452.32 (ha) or 26.22 (%) of the total land masses are moderately risk, low risk and very low risky respectively using elevation parameters as a criterion. Therefore, flood risk is relatively higher in elevation parameters like slope and LULC as it is described above.
4.5. Assessment of Drainage Density for Flooding
According to Strahler (1964), the higher the density, the higher the area is susceptible to erosion, which is resulting in sedimentation the low laying areas.
Based on this, an area with adequate drainage network has a drainage density value of ≥ 5 whereas the moderate drainage areas have a value of 5 − 1 and the poorest one has < 1. Therefore, in this study, streams of up to 4th order were considered in evaluating the drainage density of the study area. As a result of this, the higher weights were given for higher density areas drainage and lower weights were given to areas with poor drainage areas. The drainage density layer was further reclassified in five sub-groups using the standard classification Schemes (1–5). Areas with very low drainage density are ranked as 5 and those with very high drainage density were ranked with value of 1, and high risk, moderate and low risk areas were assigned as 2, 3 and 4 respectively.
Based on the table 4.6, 4472.91 ha (4.45 %) is characterized as the very high risk where 16632.81 ha (16.52 %) which assigned as very high risk. The reaming 23208.75 ha (23.05 %), 23208.75 ha (23.05 %) and 26788.68 ha (26.60 %) are parts of moderate, low and very low risk areas.
From this, the total land which is under risky zones is almost 20 % of the total areas, which implies that a study area flood risk hazards exposures is relatively lower compared with the rest of the other factor maps as the drainage density factor shows.
4.6 Rainfall
A mean average monthly rainfall for seven years (2001–2007) was considered and interpolated using Inverse Distance Weighting (IDW) to create a continuous raster rainfall. The reclassified rainfall was given a value 5 for least rainfall to 1 for highest rainfall as it is shown in Fig. 4.6 below.
4.6.1 Evaluation of Rainfall for Flooding
According to Woubet (2011), the classification of rainfall for flooding is determined by based on the amount of rainfall receiving of an area; as a result of this, the higher the rainfall an area possess, then the higher the probability of an area to be exposed for flooding. Ended, by considering the rainfall of the study area, Guba Lafto District was classified in to five flood risk classes. In this case the higher rainfall receives areas were assigning as high flood risk zones and the low rainfall receive areas were assigned as low flood risk zones.
As the Fig. 4.6 implies 3620.07 ha (3.58 %), amount of the study area is within a very low risk in terms of rainfall factor is used as a means of causing for flood occurrence. Whereas 13198.41 ha (13.08 %) of the study area ranges to high risk, and 29726.64 ha (29.47 %) is moderately risked for flood. The remaining 30276.27 ha (30.02 %) and 24050.61 ha (23.85 %) are low and very low risk zones respectively for flood hazard. Therefore, in this parameter, the study area is generally having a chance of affected by flooding is up to 19 %, and 30 % of moderately affected by flood.
4.7 Comparisons of each Criterion
Each of the six criteria has different impacts for flood occurrence as it is already notified by many literatures such as Morita (2014), Ryutaro (2014) and so on. Therefore, assessing the potential flood risk areas using either of one or two parameters may not be accurately represents the real situations. Table 4.3 illustrates the flood occurrence potentials of each of the six criteria mentioned above.
Table 3
comparison of the criteria risk level for flooding
Criteria
|
Risk levels and percent of risk classes
|
|
Very high risk
|
High risk
|
Moderate
|
low risk
|
Very low risk
|
LULC
|
1.21
|
68.32
|
7.42
|
18.23
|
5.91
|
Soil
|
28.71
|
3.70
|
17.24
|
49.58
|
0.77
|
Slope
|
30.25
|
28.70
|
20.71
|
13.76
|
6.63
|
Drainage
|
4.45
|
16.52
|
23.05
|
29.38
|
26.60
|
Elevation
|
15.96
|
23.81
|
19.91
|
14.91
|
26.22
|
Rainfall
|
3.58
|
13.08
|
29.47
|
30.02
|
23.85
|
From table 4.3, we understood that Guba Lafto district has very high probability to be affected by flood risk which consist of (1.21% and 68.32%) which is a total of 69.52%. this implies that the district is at high level of flood risk in the case of LULC criteria, whereas about 30.52% of the district has less chance to be affected by flood risks. Moreover, in the case of assessing the flood risk levels of the study area using soil, it is verified that 32.41% of the study area is at risk of flood. Regarding to the slope of a District, 58.95% of the area is laying at risk of flood, and for drainage density 20.97% of the area is at risk. As far as the elevation and rainfall criteria of the study area concerned, 39.77% and 16.66% respectively are found at high risk of flooding.
Generally, LULC shares the great values (35%), followed by slope, rainfall, soil, and elevation (58.95%), (39.77%), (32.41%) and (20.77%) respectively to evaluate the criteria independently. The amount of risk classes of the study area is quite different in different criterions, as a result of this, the study was applied the MCE methods to evaluate the flood risk areas by taking in to account to all criteria’s base on their weights for causing flood.
4.7 Multi-criteria evaluation (MCE) Methods for Ranking and Weighting of Decision Factors for Flood
Analytical Hierarchical Process (AHP) is a decision-making technique which is used to solving of different and complex problems by integrating different methods (parameters) to meet the required objectives, Ryutaro (2014). But each parameter may not be equal in weight, and some parameters may invade the other. As a result of this, deep and expert based literatures are applied for using this approach to assess the flood risk problems in the study area. In addition to this, the nature of the study area has a great role in weighting and ranking of flood causative factors, so that the study area was considered in running weights. For this, selective key informants from Guba Lafto District agricultural and water resource office were organized and interviewed to rank the flood cause factors of the study area. Thus, LULC, soil, slope, rainfall, drainage density and elevation were applied as a parameter and the results are discussed in below.
As far as the significance of the parameters weight concerned, Eigenvector techniques were applied to give the weight of the standardized raster layers.
Table 4
Square pair wise comparison matrix of the criteria
Criteria
|
Rainfall
|
Soil
|
LULC
|
D. density
|
Elevation
|
Slope
|
Rainfall
|
1
|
1/3
|
1/7
|
4
|
1/5
|
¼
|
Soil
|
3
|
1
|
1/3
|
4
|
1/6
|
½
|
LULC
|
7
|
3
|
1
|
9
|
2
|
2
|
Drainage density
|
¼
|
¼
|
1/9
|
1
|
1/8
|
1/7
|
Elevation
|
5
|
6
|
½
|
8
|
1
|
2
|
Slope
|
4
|
2
|
½
|
7
|
½
|
1
|
According to the Nine-point pair wise comparison scale; 1 means, one factor is equally important to the other, 2 mean it is slightly moderately important than the other, 3 mean it is almost moderately important than the other and 9 means one factor (parameter) is extremely importance than the other factor in their relative weights.
Based on this, as the above table 4.3 indicates about the weight of parameters, soil is moderately important with rainfall factors, while LULC is very strongly importance than rainfall, but rainfall is strongly importance than drainage density and elevation. On the other hand, slope and elevation is strongly important than rainfall.
Table 5
Normalized Matrixes of the criteria
Criteria
|
Rainfall
|
Soil
|
LULC
|
Drainage density
|
Elevation
|
Slope
|
Weight
|
%
|
Rainfall
|
0.04933
|
0.02649
|
0.055215
|
0.12
|
0.05
|
0.04
|
0.057
|
5.7
|
Soil
|
0.1481
|
0.0794
|
0.129
|
0.12
|
0.04
|
0.08
|
0.1007
|
10.07
|
LULC
|
0.34
|
0.23
|
0.38
|
0.27
|
0.50
|
0.33
|
0.347
|
34.7
|
Drainage density
|
0.012
|
0.019
|
0.04
|
0.03
|
0.031315
|
0.024
|
0.0268
|
2.68
|
Elevation
|
0.24
|
0.476
|
0.193
|
0.24
|
0.250
|
0.33
|
0.2915
|
29.15
|
Slope
|
0.197
|
0.158
|
0.193
|
0.212121
|
0.125261
|
0.169
|
0.176
|
17.61
|
Grand
Total
|
1
|
1
|
1
|
1
|
1
|
1
|
1
|
100
|
4.8 Calculating CR
The consistency ratio is computed with RI obtained from Pairwise comparison table. Hence, the result of CI is by far lower than the threshold value of 1.24 which is equal to 0.0. This implies there is a high level of consistency in the pair wise opinion and no need of reviewing the criteria since it is accurate.
4.9 Ranking Method
Inverse ranking method was applied for this study in order to rank the criteria consideration in addition to the factor maps. In the case of flood causing risk level, the list important is ranked as 1 while the most important criteria are ranked as 5. The eigenvector of the pair wise comparison matrix is used produce a best fit to the weight set. Weight values in this case refer to the priorities which are absolute numbers between zero and one. Their sum is equal to 1, and their weight is given with their perspective influence to flood occurrence.
A higher weight value of the factors represents more priority than the rest of the other factor. From the factor weights found for this study area, it is clear that the LULC characterized by different water holding capacity, have the highest weights, showed that LULC have more contribution to the occurrence flooding in the study area as compared to the other factors or elements followed by elevation and slope.
Table 6
summary of weighting of flood risk criteria
Decision factor
|
relative weight of each decision criteria
|
Decisions of sub factor ranking
|
Ranking
|
Rainfall (mm)
|
0.058
|
1.37862968-35.59924507-
|
5
|
35.59924507–73.76876831
|
4
|
73.76876831–113.895703
|
3
|
113.8957031–162.8309892
|
2
|
162.8309893–238.19133
|
1
|
Soil (type)
|
0.100
|
Cambisols
|
5
|
Pheamosols
|
4
|
Regosols
|
3
|
Xerosols
|
2
|
Leptosols
|
1
|
LULC (type)
|
0.347
|
Foerst land
|
5
|
Shrub land
|
4
|
Grass land
|
3
|
Crop land
|
2
|
Bare Land
|
5
|
Drainage density (Km)
|
0.026
|
0-0.69352
|
1
|
0.69352–1.386504
|
4
|
1.386504–2.079757
|
3
|
2.079757–2.773009
|
2
|
2.773009–3.466261
|
1
|
Elevation (M)
|
0.291
|
1,324–1,786
|
1
|
1,786.000001–2,207
|
2
|
2,207.000001–2,665
|
3
|
2,665.000001–3,158
|
4
|
3,158.000001–3,802
|
5
|
Slope (%)
|
0.176
|
0–10.25183632
|
1
|
10.25183633–19.86293287
|
2
|
19.86293288–30.75550896
|
3
|
30.75550897–43.89067425
|
4
|
43.89067426–81.69432068
|
5
|
The flood risk map in this study was generated by using the weighting and ranking technique. The nature of the method is solely depending up on the nature of the biophysical factors which are correlated with potential flood causing capacity. The factors were weighted according to their relative importance to each other and to their expected importance in causing floods. In addition to this, each factor was classified into five sub-factors, each of which was weighted and ranked. The higher-ranking values the higher susceptible to the occurrence of floods. LULC, Rainfall, elevation, slope, drainage density, and soil type decision parameters were identified as the most important factors for controlling the potential flood susceptible and its sub-factors as well as reclassification system was run depending on the previous flood related studies, (Tanavud et al., 2001).
Among all the six decision factors, a high weight was given for LULC followed by elevation, slope, soil type, and rainfall and drainage density.
Each of the six factors was further classified in to five classes based on their risk classes and given a ranking value accordingly. Finally, by using weighted overlay methods, it was reach to five categorical classes’ very low, low, moderate, high, and very high.
Table 7
Final weighted overlay flood risk classes
No
|
Risk classes
|
Area coverage
|
1
|
Very high risky
|
5.76 (0.35)
|
2
|
High risky
|
25398.45 (25.58)
|
3
|
Moderately risky
|
61065.91 (61.41)
|
4
|
Low risky
|
12752.28 (12.81)
|
5
|
Very low risky
|
83.72 (0.08)
|
Grand Total
|
|
99306.09 (100)
|
Area coverage: number in ha; ( ) = % ( percent coverage) |
As table 4.8 shows from the total land (99306.09 ha or 100 %), although there is a very small proportion of very high-risk areas (0.085) % of the total area, but there is a significant value of high risky 25398.45 ha (25.58) %. Those most risk areas are located in the central and eastern as well as south east parts of the study area. In addition to this, almost greater than a half 61065.91 ha (61.41) % is under moderately risky zones which are found in the central south and west parts of the study area including the northern regions. While the remains substantial area (12.81) % and very small area (0.08) % are low risky and very low risky respectively for flooding.
Based on the final weighted analysis of flood risk map of the study area, most highly flood risk kebeles were identified. 26064.29
Table 8
Potential identified high flood risk kebeles in Guba Lafto District
Kebeles
|
Area coverage
|
Gugsa
|
2184.57 (2.14)
|
Lasgerado
|
453.3 (3.41)
|
Hara
|
449.82 (0.45)
|
Lay Alhua
|
2939.13 (2.95)
|
Gubarija
|
3186.27 (3.2)
|
Shewit
|
4152.6 (4.18)
|
Sagat
|
5281.2 (5.31)
|
Gebre Amba
|
1569.6 (1.57)
|
Anova
|
2852.37 (2.87)
|
Area coverage: number in ha; ( ) = % ( percent coverage)
32 kebeles in Gub Lafto District, nine kebeles were identified as the highest flood risk areas which consists of 26064.29 ha (26.25%) from the total 99306.09 ha (100%). From those high flood risk kebeles of the District, Sagat shewit, and Anova are found in the southern parts of the study area. Whereas Lay Alwha, Gebre Amba and Hara kebeles are located in the central part of the district. Moreover, the remain Gubajira, Gugsa and Las Gerado are found around in the eastern and north east part of the district.
Generally, it was possible to identify the highest flood risk areas particularly the flood prone kebeles. The result shows that there are adequate flood prone areas in Guba Lafto District and the most sensitive areas to flooding were justified. The result also shows that one fourth of the study area (25%) is prone to flooding, which is very significant areas are under the risk of flooding. The identified prone areas are those which are bare lands, areas with low elevation and flatter slopes, closed to drainages and agricultural practiced area.