3.1. Factors Suitability Class
Topography is one of the main factors affecting the efficiency of irrigation practices. It is described in terms of landscape slope calculated from a DEM. Worqlul et al. [14
] considered four slope suitability classes: highly suitable (0–2%), moderately suitable (2–8%), marginally suitable (8–12%), less suitable class (12–30%) and unsuitable class (above 30%). On the other hand, land cover is also one of the potential factors for irrigation land suitability assessment. Cultivated land is the dominant land use in Lake Tana Basin (56%) followed by water body (about 20%) and grassland (10.6%). Land use types were classified into suitability class as follows: cultivated land (S1); grassland (S2); shrubs and bare land (S3); and forest and plantation (S4) [16
]. Topography and land use types of Lake Tana Basin were reclassified into suitability classes (Figure 4
Proximity to river networks is the most important factor for irrigable land assessment from the surface water source. Agricultural land close to river networks can get access to water easily for crop production. Euclidean distances were calculated from each river and then an equal interval method was used for classification into suitability classes (Figure 5
a). The soil is also one of the potential factors that influence irrigation practices. Awulachew et al. [72
] explained soil texture suitability classes: highly suitable (fine to medium), moderately suitable (heavy clay), marginally suitable (course or poorly drained) and unsuitable class (very coarse) or shallow depth. Baniya [45
] classified soil depth into various suitability classes: highly suitable (above 80 cm), moderately suitable (50–80 cm), marginally suitable (30–50 cm) and unsuitable (below 30 cm). Soil salinity was considered based on Awulachew et al. [72
] classification: highly suitable (0–0.5 dS/m), moderately suitable (0.5–2 dS/m), marginally suitable (2–5 dS/m) and unsuitable (above 5 dS/m).
Lake Tana Basin is dominated by Haplic Luvisols (21%) followed by Chromic Luvisols (16%), which are characterized as moderately well drained soils. The effect of soil characteristics was expressed using soil capability index (SCI), which is described as the ability of soil for irrigation practices. First, soil texture, depth, salinity, and drainage classes were used to compute the soil capability index (SCI). Physical and chemical properties of the soil types over the basin were obtained from the FAO soil database (Table 3
) and rated out of 100% based on O’Geen [59
] and Teka et al. [29
]. Rating soil properties were made considering the capability of individual soil properties from the perspective of irrigation practice. SCI was then reclassified into different suitability classes based on Teka et al. [29
] as follows: greater than 80 (S1); 60–80 (S2); 45–60 (S3); 30–45 (S4%) and less than 30 (N1) (Figure 5
The market outlet is one of the main factors that affect irrigation suitability from all water sources because agricultural products need to get access to markets. The market outlet was considered based on proximity to road networks and urban centers together. All urban centers that were included in the nation town map within Lake Tana Basin are considered in this study. Agricultural lands close to road networks and urban centers are most likely to get access to markets. Euclidean distances were calculated from road networks and urban centers and then classified into suitability classes based on the equal interval approach (Figure 6
Groundwater depth and salinity are important factors when considering irrigation from groundwater sources. Spline interpolation was used to find the spatial distribution of groundwater level and salinity over the basin. Groundwater table elevation was deducted from DEM to obtain groundwater depth from the surface. According to Gebregziabher [60
], depth to groundwater less than 6 m (S1), 6–10 m (S2), 10–30 m (S3) and more than 30 m (S4). Similarly, Kumar et al. [61
] classified groundwater salinity as follows: less than 250 µS/cm (S1), 250–750 µS/cm (S2), 750–2000 µS/cm (S3) and 2000–3000 µS/cm (S4). Groundwater depth and salinity were then reclassified based on suitability classes (Figure 7
Mean annual rainfall depth is one of the main factors considered for irrigation suitability analysis from harvested rainwater sources. Considering mean monthly rainfall, Addis Zemen station receives the highest rainfall in July followed by Debre Tabor and Bahir Dar stations (Figure 8
a). Annual rainfall analysis showed that Dangila station receives the highest mean annual rainfall followed by Debretabor and Bahir Dar stations (Figure 8
b). Delgi station receives the lowest rainfall considering both monthly and annual rainfall amounts. The highest rainfall record was observed in the southern and eastern part, whereas the lowest rainfall was observed in the northern and northwest part of the basin (Figure 8
b). The rainy season was from June to September (Figure 8
The mean annual rainfall ranges from 862 mm to 1637 mm over the basin. According to Nketiaa et al. [62
], mean annual rainfall can be classified as follows: above 900 mm (S1), 790–900 mm (S2), 640–790 mm (S3) and less than 640 mm (S4). Piche evaporation [73
] measurement was obtained from five climate stations: Adet, Bahir Dar, Debretabor, Gondar, and Shahura. Penman–Monteith average monthly evapotranspiration was computed for Bahir Dar station and compared with mean monthly Piche evaporation measurements (Figure 9
a). Piche evaporation was found overestimated during the dry season (November to May) and underestimated during the wet season (June to October). According to Stanhill [73
], a linear relationship exists between Piche evaporation and Penman potential evapotranspiration indicating the possibility to derive monthly adjustment coefficients (Figure 8
a). Monthly adjustment coefficients were derived and used to adjust the mean monthly (1992–2009) Piche measurements to other stations (Figure 9
b). Evapotranspiration loss was found to be high in the northern and northwestern part of the basin (Figure 10
a). The highest evapotranspiration record was observed from March to May.
Mean annual rainfall deficit was computed by deducting mean annual potential evapotranspiration from the mean annual rainfall. The equal interval method was used to classify the mean annual rainfall deficit into suitability classes (Figure 10
a). Permanently unsuitable land (constraint), water body, wetlands, urban/built up areas and protected areas were taken away from factors’ suitability class. Figure 10
b shows the constraint map where ‘0’ and ‘1’ denote constraints and various suitability classes, respectively.
Overall, suitability assessment of individual factors assists with partially characterizing the potential of Lake Tana Basin for irrigation. Lake Tana accounts for about 20% of the basin, which is considered as permanently unsuitable, constraint, (N2) for irrigation practice for all criteria. About 72% of the basin is suitable for irrigation practices when considering topography; S1 (9%), S2 (34%), and S3 (29%), whereas the remaining 8% is currently unsuitable (S4). The easternmost and northernmost part of the basin has a steeper slope (i.e., less suitable for irrigation practice). About 77% of the land use is suitable for irrigation, S1 (56%), S2 (11%), and S3 (10%), whereas S4 is about 3%. Cultivated land is the dominant land use in the basin (~56%). The soil capability index results showed that the entire basin is suitable (except the constraint) for irrigation when soil characteristics are considered: S1 (51%), S2 (27%), and S3 (2%). Proximity criteria (road, urban and river) were classified using an equal ranging technique, and the majority of the basin falls under suitability classes (S4 is less than 1%). Groundwater depth and sanity were included for suitability assessment for irrigation from groundwater. About 22% of the basin is suitable for irrigation when groundwater depth is considered: S1 (13%), S2 (2%), and S3 (7%), whereas the remaining 58% is S4. The majority of the irrigable land (based on groundwater depth) is found in the central and southern part of the basin. The entire basin is suitable for irrigation (except the constraint) when groundwater salinity is considered.
3.2. Factors Weight and Assessment of Irrigable Land
The pairwise weighting method was used to calculate the weights of each factor. The average of expert opinions was used for comparison of factors one-to-one. The eigenvector was calculated from a pairwise matrix based on Podvezko [74
] as the n
-th root of the product of values in rows. The cumulative eigenvector was used to calculate the final weights of each factor. The weight of factors was computed separately for rivers, rainfall (Table 4
) and groundwater sources (Table 5
) after normalizing the eigenvector with the cumulative value. In Table 4
, river proximity was replaced with mean annual rainfall depth when computing the weight for irrigation suitability from rainwater harvesting sources while the remaining six parameters were the same.
The consistency of pairwise matrix was checked, CR = 0.08 and 0.005, for surface and groundwater sources, respectively, and found to be trustworthy (CR ≤ 0.2) based on Chen et al. [75
]. Based on the results of factors weight, proximity to rivers, rainfall, and groundwater depth were found to be the most influential factors from the surface, rainwater harvest, and groundwater sources, respectively. Slope and soil capability index were found to be the second and the third influential factors from surface water (Table 4
), whereas groundwater salinity and slope were the second and the third influential factors for irrigation from groundwater (Table 5
The weights were distributed to individual factors’ suitability classes based on an equal interval ranging technique, and the factors were combined using weighted sum overlay. A total of seven criteria layers for surface water and rainwater harvesting sources and eight criteria for groundwater source were combined to produce the single suitability index. The value of the suitability index ranges from 0 (permanently unsuitable land) to 100 (most suitable land). Suitability threshold is a value that needs to be exceeded to determine potentially irrigable land. At 85% threshold, about 345,000 ha of land is suitable for irrigation from surface water sources (Figure 11
a). The results depicted about 3% increment on the irrigable land when compared to Worqlul et al. [14
]. The reason was that seven additional rivers, twenty-eight more urban centers and more refined road networks were considered in this study. Further suitability classes were established based on an equal interval approach (Figure 11
b). The majority of the irrigable land, about 83% were in the three major river catchments: Gilgel Abbay (~54%), Gumara (~13%) and Ribb (~18%). If an 80% suitability threshold was used, the irrigable land would be increased to about 594,000 ha.
On the other hand, about 720,000 ha of land is irrigable using an 85% threshold from rainwater harvest considering average annual rainfall (Figure 12
a). Further suitability classes were established based on an equal ranging approach (Figure 12
b). The potential of the rainwater harvesting system depends on types, size, and location of structures. Biazin et al. [76
] indicated the feasibility of macro-catchment rainwater harvesting in the long term despite the fact that the initial investment is high. Traditional open ponds, cisterns, earthen dams, and sand dams are among macro-catchment rainwater harvesting techniques. However, the average size of the farmer’s plot is small, which is about 0.5 ha per household [77
], and the use of rainwater harvesting structures could consume a substantial amount of land. In situ and micro-catchment rainwater harvesting could be feasible for smallholder farmers in the region. In situ rainwater harvesting includes mulching, ridging, furrowing and pot hoeing and conservation tillage, whereas micro-basins, contouring, and terracing are among micro-catchment rainwater harvesting techniques.
About 176,000 ha of the land is suitable for irrigation from groundwater considering 85% threshold (Figure 13
a). If the 80% threshold were used, the irrigable land would be increased to about 277,000 ha. Irrigable land at various threshold levels is shown in Figure 13
b. Groundwater depth was found to be the limiting factor for irrigation from a groundwater source.
3.3. Surface and Groundwater Potential
Analysis of mean monthly rainfall and crop water requirement was made for all river catchments. The consistency of monthly crop water requirement values was checked comparing calculated low-flow potential with [14
]. The result depicted that the rainfall is sufficient to meet crop water need from June to September for Gilgel Abbay, Gumara and Ribb river catchments while other river catchments do not need irrigation for July and August. Crop water requirement within Lake Tana Basin was high from January through March. Low-flow (Q90
) potential of rivers was analyzed using conventional irrigation techniques. Q90
potential of all rivers within the basin ranges from 3133 ha in April to 4142 ha in December. The Gilgel Abbay river has the highest Q90
potential (1866 ha) followed by Gumara river (728 ha). About 91% of the Q90
potential in the basin came from the four major rivers (Gilgel Abbay, Gumara, Ribb, and Megech); the other rivers can irrigate about 285 ha. Q90
potential of rivers can only satisfy about 1–1.2% of the irrigable land. The potential of the rivers is limited if only surface water source is used without storage structures. If seven irrigation dams along the major rivers within the basin are considered, the combined potential would be 72,666 ha (Figure 14
). Awulachew et al. [1
] estimated the runoff generation from 12 river basins for Ethiopia to be about 125 billion m3
per year. Most of this potential is lost due to limited infrastructure to make adequate use of surface water [78
]. This indicates the potential of irrigation dams and other water harvesting structures in maximizing potentially irrigable lands from surface water.
Crop water requirements in the regions and groundwater well yield were used to compute the potential of groundwater using conventional irrigation techniques. The potential of groundwater ranges from about 3873 ha in March to 4898 ha in November. About 1440 ha of land in Lake Tana Basin can be irrigated using either surface or groundwater sources (Figure 14
). There are various constraints and opportunities to expand irrigated agriculture in the region from surface and groundwater sources. The cost of development and operation, cost of irrigation equipment, inadequate capacity and community involvement, irrigation input costs and lack of comprehensive understandings of water resources are among the major constraints for irrigation [5
], whereas availability of abundant water resources, climate and land suitability, the reliability of water resources, inexpensive labor, adaptation of promising approach and technologies are some of several opportunities to expand irrigated agriculture [5
3.4. Applicability of Low-Cost Water-Lifting Technologies
The applicability of low-cost water-lifting technologies was identified comparing the operating depth and pumping capacity of the technologies with depth to groundwater and borehole yield in the region, respectively. The result showed that bucket pumps and direct-action hand pumps were found applicable on about 225,000 ha of land in the basin (Figure A1
a,b, respectively). Bucket pumps use the simple system, easy to repair, have lower water yield and lower operation cost [70
]. Maintenance of the direct-action hand pump is relatively simple. However, the cost of the system is fairly high [70
]. Deep wells’ diaphragm pumps and deep wells’ piston hand pumps were found to be applicable on about 240,000 ha and 135,000 ha of lands in the basin, respectively (Figure A1
c,d respectively). Deep wells’ diaphragm pump can lift water from deeper depth to higher elevation [81
]. Rope pumps and rope and bucket pumps were found applicable for on about 390,000 ha and 240,000 ha of lands in the basin (Figure A1
e,f, respectively). Rope pumps were found the most applicable water-lifting technology as compared to other technologies considered in this study. Suction plunger hand pumps were found applicable on about 195,000 ha of land in the basin (Figure A1
g). These water-lifting technologies can be used by smallholder farmers and helps to expand irrigated agriculture using groundwater sources.