Introduction

Water scarcity has been increasing globally especial in semi-arid areas due to the high variability of climate (Uwizeyimana, et al., 2019), scarce and intermittent rainfall (Nthuni, et al., 2014; Sayl, et al., 2019), or unforeseeable periods of flood and drought (Adham, et al., 2018). Water availability is the basic need of human development and population growth (Zaki, et al., 2019). Also, farming and livestock activities depend only on the water ponds accumulated during wet seasons in the semi-arid areas (Shemsanga, et al., 2016). The shortage of water affects every aspect of socioeconomic life including health, agriculture, food security, and other technical aspects.

The semi-arid regions are characterized by the low frequency of rainfall events with high intensity, short periods, and highly temporal-spatial distribution (Camarasa-Belmonte and Soriano, 2014). During rainfall, a fraction of the precipitation evaporates, some infiltrate or intercepted by vegetation, and others become surface runoff (Vikneswaran and Razak, 2015). The amount of surface runoff generated depends on several factors including rainfall intensity, initial soil moisture condition, land use and land cover, soil texture, topography, and basin characteristics. Surface runoff harvesting is the collection, accumulation, and storing of the rainwater into an artificial or man-made reservoir for future reuse (De Winnaar, et al., 2007). Rainwater harvesting is considered to be a sustainable method to collect water for use in domestic and agriculture activities and industry (Sayl, et al., 2019; Matomela, et al., 2020). Several techniques such as deep tillage, contour farming, and ridging have been used for in situ surface runoff harvesting (Hatibu and Mahoo, 1999).

The increase in population and economic growth becomes the driving force for the expansion of urban built areas in the developing cities (Li, et al., 2015). The expansion of built areas with several infrastructures leads to change in water balance parameters and urban land uses (Ligtenberg, 2017). The infrastructures built including buildings, roads, airports, and other impervious structures such as parking and so forth. The generated surface runoff either runs away from residential or premises, trapped for use, lost in the atmosphere through evaporation, or runs in the river or streams. Also, a high amount of surface-runoff is generated in the urban areas are due to the low infiltration that is turning into flooding during the heavy storm (Du, et al., 2015).

The identification of surface runoff sites for rainwater collection was determined normally using the geographical information system (GIS) and remote sensing (RS) tools. The factors affecting rainwater harvesting were synthesized in the GIS and RS tools with different methods and multi-criteria techniques to identify the potential sites. Several factors used include rainfall intensity, runoff coefficient, soil texture, drainage pattern, topography, vegetation cover, and socioeconomic conditions as described in Wu et al. (2018) and Terêncio et al. (2017). The potential sites were identified using the hydrological modelling (HM) integrated with GIS/RS as described in De Winnaar et al. (2007), multi-criteria analysis (MCA) integrated with HM and GIS-RS described in Weerasinghe et al. (2011), and Elewa et al. (2012) and MCA integrated with GIS as described in Kadam, et al. (2012). The climate-flood model was developed to understand the annual runoff characteristics based on the climatic variables including rainfall depth, air temperature, sunshine, evaporation, humidity, wind speed cloud cover, soil temperature, solar radiation, and atmospheric pressure (Ekwueme & Agunwamba, 2020). The Soil and Water Assessment Tool (SWAT) model was used with inputs from geospatial datasets and precipitation data to simulate the watershed for enhancing water resource management (Guiamel and Lee, 2020). The selection of the method depends on the data availability, accuracy, and requirement of the specific area (Ammar, et al., 2016). Table 1 shows the methods used and parameters applied by different scholars to identify the potential sites for surface runoff harvesting. However, the combination of MCA and GIS has proved potential to logical, objective, and non-biased approach to the selection of the suitable area for surface runoff harvesting (Isioye, 2012).

Table 1 Potential sites identification based on different methods and parameters from several scholars
Full size table

Research goals, motivation, and paper organization

According to the population census of 2012, the number of people in the Dodoma urban was 410,956 (URT The United Republic of Tanzania, 2013). However, the population has steadily increased in recent years following the shifting of the national capital from Dar es Salaam to the Dodoma region in 2016. The increased population in Dodoma urban has been marked to increase water demand (Shemsanga, et al. 2018; SHLC (Centre for Sustainable Health and Learning Cities and Neighbourhoods, 2020) and changes in land use. The current water demand in Dodoma is 45 Million liters per day and the supply is only about 20 million liters per day (DUWASA, Dodoma Urban Water Supply and Sanitation Authority, 2019). The water authority of Dodoma urban depends on the groundwater sources both in deep and shallow wells (Shemsanga, et al. 2018). This source of water is limited to the beneficiaries’ water supply network or the situation becomes critical, especially during the drought period (WB (The World Bank, 2019).

The area is overpopulated and increases land use, especially the construction of public buildings and residences. Knowing that the land use was already given without showing the provision of routines for surface runoff (URT, 2019), this makes the paper to be valuable to the study area and provides information to the other developing cities for considering the identification of surface runoff routes during land use map preparation. Therefore, the motivation of this paper is to reduce the pressure of pumping groundwater resources in the study area, especially during the wet season and encourage the practice of rainwater harvesting in the urban built-up areas. The objective of this study is to identify the potential sites for surface runoff harvesting and estimate the available surface area in emerging urban cities for minimizing floods.

The research works identified the potential sites for surface runoff harvesting based on the general procedures of the analytic hierarchy process (AHP) under the GIS environment. The AHP method was analytically triggered with a weighted linear combination method to identify the very high suitability area when the slope is less than 5°, high drainage density, low lineament density, and higher runoff depth.

This paper is organized as follows: the introduction part consists of the assessment of water needs, land uses, and socioeconomic activities; the methodology part consists of the description of the study area, and identification of the six factors that affected the harvesting of surface runoff. The curve numbers were determined using hydrological soil groups, land use, and rainfall data, while the normalized weights were determined using runoff depth, drainage density, slope, and lineament density; the results and discussion part consists of the spatial distribution of runoff depth; then, potential sites were identified using the AHP method with normalized weights. Also, the surface areas for potential sites were calculated using the triangulated irregular network, and natural ponds were identified using Google Earth images; the conclusion part supports the harvesting practices of surface runoff in the semi-arid areas.

Description of the study area

The Dodoma urban is the capital city of Tanzania, located in the central part of Tanzania having an elevation between 925 and 1691 m above sea level with an average elevation of 1100 m as shown in Fig. 1. The high elevation area has been allocated in the North and South of the Dodoma urban. About 49% of the area is situated in an elevation ranging between 1074 and 1163 m. For more than 2 decades, three dams were built and used to store the surface runoff for domestic, livestock, and farming activities. These dams including Imagi, Mkalama, and Hombolo as shown in numbers 1, 2, and 3 at the geographical locations of − 6.2051, 35.7529; − 6.1992, 35.7727; and − 5.9469, 35.9679, respectively. Dodoma urban receives a single rainfall period between October and May ranging between 400 and 900 mm with an average total annual rainfall of 550 mm (Kassile, 2013). The average temperature of Dodoma urban is 24.7 °C with November being the hottest month in a year and July the month with the lowest average temperature.

Fig. 1
figure 1

Location of Dodoma urban with the spatial elevation map (Google Earth photo as of 09 March 2021)

Full size image

Materials and methods

Data types and sources

The identification of potential sites for surface runoff harvesting was evaluated using a combination of different data set as presented in Table 2.

Table 2 Data type, resolution, and uses from different literature sources
Full size table

Factors affecting the rainwater harvesting

The originality of this study is to use simple methods and freely available data to solve the problem related to planning the land use without thinking about the surface runoff routes, especially in the developing cities, where high knowledge and finance conditions are major barriers. Figure 2 shows the flow of needed parameters to develop the surface runoff that can be harvested or leave the place opens to avoid flooding or restricts water flows. Three parameters were used to estimate the runoff depth which is the main source of surface runoff harvesting, and then three physical parameters include drainage density, slope, and lineament density were applied to identify the percentage of surface runoff that can reach a particular time. The land use or cover type and soil groups were used to estimate the adjusted curve number. Then, the adjusted curve number was combined with rainfall using Soil Conservation Service (SCS) method to obtain the runoff depth. Again, runoff depth, drainage density, slope, and lineament density were combined using weighted linear combination (WLC) to identify the potential surface runoff sites.

Fig. 2
figure 2

Simplified flowchart showing the procedures for identifying potential sites for rainwater harvesting

Full size image

Runoff depth factor using SCS method

The runoff depth was estimated using the SCS method based on the land use/land cover, soil, and rainfall data. The SCS method was developed by the Natural Resources Conservation Service of the United States Department of Agriculture (USDA) in 1954 to estimate the runoff depth (Gajbhiye, 2015). The method originated from the assumption of the ratio of runoff depth to precipitation minus initial abstraction is equal to the ratio between the infiltration to soil potential maximum retention as expressed in Eq. (1) (Li, et al. 2015). The method considers that the initial abstraction had a proportional constant to soil potential maximum retention (Patil, et al. 2008). The method has been used to estimate the runoff depth in a wide climatic zone ranging from temperate humid zone to arid areas (Boughton, 1989).

$$\left.\begin{array}{r}\begin{array}{cc}Q=\frac{\left(P-\lambda S\right)^2}{\left(P-\lambda S\right)+S}&for\;p>\lambda S\end{array}\\\begin{array}{cc}Q=0&for\;p>\lambda S\end{array}\end{array}\right\}$$
(1)

where Q is the runoff depth, mm; P is the precipitation, mm; S is the soil potential maximum retention; and ʎ is the proportionality constant that is depending on the climate and the geology of the region under study (Patil, et al. 2008). The value of ʎ parameter ranges between 0.1 and 0.3 inclusively. Most of the reach works used an average value of ʎ = 0.2 (Al-Ghobari, et al., 2020). However, some of the researchers used a maximum value of ʎ = 0.3 (Satheeshkumar, et al. 2017). Comparison of different values of ʎ = 0.1, 0.2, and 0.3 was conducted by Farran and Elfeki (2020). The findings were relatively agreed with the other literature data for using ʎ = 0.2 and adopted in this study.

The soil potential maximum retention (S) was determined indirectly by using the curve number (CN) of hydrological soil cover. The value of CN depends on soil type, land use, and the antecedent moisture condition (AMC) (Satheeshkumar, et al. 2017). The relationship between the S and CN is given in Eq. (2).

$$\mathrm{S}=\frac{25400}{\mathrm{CN}}-254$$
(2)

The anteceded moisture condition (AMC) indicated the degree of wetness of the soil before the storm. This AMC has a great effect on the quantity of water to be retained or overflow on the surface of the earth (Rawat and Singh, 2017). To take care of the AMC, SCS has developed a manual to adjust the curve number based on the rainfall of 5 days preceding the storm event as grouped into three types (Farran and Elfeki (2020). Normally the calculation of the CN is first done by assuming AMC type two condition, i.e., 5 days antecedent rainfall of dormant season ranged between 12.7 and 27.9 or growing season ranged between 35.6 and 53.3 mm (Farran and Elfeki (2020). The adjusted \({\text{CN}}_{a}\) as expressed in Eq. (3) depends on the rainfall condition type one of less than 12.7 in the dormant season and less than 35.6 mm in the growing season that suits this study area.

$${\text{CN}}_{a}{=}\frac{CN}{2.2754}-{0.01275}{CN}$$
(3)

Soil texture factor

The FAO soil map of Africa acquired from global data sets that are available online was used under the GIS to clip the soil texture and soil composition (Table 3).

Table 3 Soil texture in the study area as taken from (FAO, 2003)
Full size table
Slope factor

The slope of different classes significantly affects the surface runoff and the infiltration amount (Garg, et al. 2013). A slope of less than 5° is recommended for surface runoff harvesting (Mugo and Odera, 2019). This implies that the selection of the suitability of an area for runoff harvest is a function of a slope. A slope in the study area was developed using the DEM obtained from USGS. The DEM in the study area clipped using the study area boundary shapefile and slope tool in the spatial analyst tool to calculate five classes of slope in degree as shown in Fig. 3. The DEM was filled to make sure all pixels contain values. The larger part of the study area had a percentage slope ranging between 0 and 1.7° of about 69% followed by 1.7 and 2.8° of about 14%. Typical slope classes from different literature sources were presented in Table 4.

Fig. 3
figure 3

Slope factor with five classes in degree

Full size image
Table 4 Typical slope classes from different literature sources
Full size table
Drainage density

Drainage density measures the distribution of streams in the study area to determine the amount of runoff losses due to infiltration. The higher the drainage density, the higher the potentiality for surface runoff harvesting (Sayl, et al. 2019). Typical drainage density classes from different literature sources are presented in Table 5. The stream order raster map generated in the ArcGIS was used in the line density analysis to estimate the magnitude per area within the radius of each pixel. Figure 4 shows the drainage density that varied between 26 and 104 m/m2.

Table 5 Typical drainage density classes from different literature sources
Full size table
Fig. 4
figure 4

Drainage density with five classes

Full size image

Lineament density

Lineament density was obtained using the total length of the observed lineaments divide by the total area of the basin as shown in Fig. 5. The lineament density indicates the suitability of an area for surface runoff harvest or groundwater recharge. High lineament density is suitable for groundwater recharge and the low lineament density is suitable for surface runoff harvest (Maina and Raude, 2016). Typical lineament density classes from different literature sources were presented in Table 6. The Landsat 8 operational land image was imported into Catalyst Professional software for the demarcation of lineaments as described in Mugo and Odera (2019). The obtained lineament under the GIS environment is used to define the lineament density.

Fig. 5
figure 5

Lineament factor with two classes

Full size image
Table 6 Typical lineament density classes from different literature sources
Full size table

Land use and land cover

The land use and land cover (LULC) data were loaded in the ArcGIS environment. The land use classes and corresponding coverage area were obtained as shown in Fig. 6. A total of eight land use classes were identified which are barren or sparsely vegetated, cropland or natural vegetated, deciduous broadleaf forest, grassland, savanna, open shrublands, and urban built up. The covered areas of cropland and urban built-up increased each year resulted to decrease other five LULC classes.

Fig. 6
figure 6

Land use and land cover from the year 2001 to 2019 using satellite image data

Full size image

Data analysis

Curve number identification

Six classes of land use/land cover were reclassified to fit the standard classes provided by HSG (Neilsen and Hjelmfelt, 1998) and integrated with the soil classes to define the curve number as shown in Table 7. The land covered with dense forest experienced less curve number compared with the bare or man-made land features. The adjusted CN values were obtained using Eq. 3 and spatially interpolated using the raster calculator to obtain Fig. 7. The CN ranged between 65 and 99.6 with an average of 71.5.

Table 7 Lookup table for land use/land cover and soil types with the corresponding curve number
Full size table
Fig. 7
figure 7

Spatial distribution of adjusted curve number

Full size image

Analytic hierarchy process for potential site identification

This study used the AHP method that was used frequently by scholars and ranked first followed by the other decision methods, according to Mardani et al. (2015). The features of runoff depth, drainage density, slope, and lineament density factors were assigned weights based on Saaty’s scale from 1 to 9. The value of 9 shows the highest influential and the value of 1 shows the lowest weight of influential (Saaty, 2004). Weights on the features of thematic layers were assigned based on the opinion of experts and tested using a consistency ratio (Al-Adamat, et al. 2010; Ammar et al. 2016). Pairwise comparison of the assigned matrix for factors was conducted using the AHP technique to obtain a matrix as shown in Table 8. The values of the matrix were based on the influence of one factor compared to another. For instance, a runoff depth influences to surface runoff compared to the slope is 3, but the slope influences as reciprocal to runoff, i.e., slope factor equal to 0.33. Then, normalized eigenvectors were obtained as 53%, 27%, 14%, and 6% for runoff depth, slope, drainage density, and lineament density, respectively.

Table 8 Pairwise matrix comparison of factors affecting potential sites identification
Full size table

The values of the matrix were tested using the consistency ratio (RC) to measure the acceptance level. The RC was computed using the consistency index and random consistency index (RCI) as expressed in Eq. 4. When the value of RC is between 0 and 0.1, the consistency is acceptable (Zhang, et al. 2014); otherwise, the subjective judgment is revised for all factors. The computed value of principal eigenvector (γmax) as the product of each element of eigenvector and reciprocal matrix sum of columns and RCI was equal to 0.9 for four factors.

$$\mathrm{RC}=\frac{{\upgamma }_{\mathrm{max}}-\mathrm{n}}{\left(\mathrm{n}-1\right)\mathrm{xRCI}}$$
(4)

The maximum principal eigenvector (γmax) was 4.158 as calculated using the assigned weights of four factors in the comparison pairwise matrix. For checking the probability of randomness of the generated matrix ratings, the consistency index is 0.052 with the formula adopted from Jha et al. (2014). Then, the consistency ratio (RC) was 5.8% that lied within the acceptance level to provide the results using the AHP technique. The four thematic layers/maps together with their corresponding normalized weights were integrated into the GIS environment to select the potential harvesting site for surface runoff. The WLC method was used to sum all four factors as equal to weights for each factor times the normalized eigenvectors to define each weighted class (WC) for potential site identification.

Results and discussion

This session covers the assessment of six factors affecting the surface runoff, variation of assigned scales to factors, and normalized weights, and then potential site map was developed and overlaid to the natural streams.

Assessment of factors for potential site identification

Assessment of rainfall factor

The measured rainfall data from 1981 to 2011 as adopted from Hamisi (2013) is shown in Fig. 8. The mean value of total annual rainfall was 550 mm at the percentage of time equal to or exceeded 50%.

Fig. 8
figure 8

Total annual rainfall from 1981 to 2011 at the Dodoma region

Full size image

Assessment of land use/land cover factor

Figure 9 shows the five classes of land use/land cover including agriculture activities covered the area of 58% followed by forest and other natural areas of about 33.6%. The other land use/land cover consists of artificial surfaces based on the constructions of buildings, roads, railway, and airport of about 5.9%; wetlands and inland water of about 2.4%; and bare land of about 0.1%.

Fig. 9
figure 9

Land use/land cover of Dodoma urban with six classes

Full size image

Assessment of soil factor

Two main soil textures were identified which are sandy-clay-loam and sandy-loam as shown in Fig. 10. The land is dominated by a sand-clay-loam soil texture of about 92.1%. The soil texture was used to define the HSG in the study area. The study area was covered with the HSG group C that is associated with the following features: (i) soil having moderately high potential runoff when thoroughly wet; (ii) contain 20–40% clay and 50% sand; (iii) soil texture: loam, silt-loam, sand-clay-loam, clay-loam, and silt–clay or sand clay; (iv) saturated hydraulic conductivity between 1 and 10 µm per second.

Fig. 10
figure 10

Classification of soil textures in Dodoma urban

Full size image

Assessment of runoff depth factor

The assessment of runoff depth was conducted using the average amount of rainfall, land use/land cover classes, and soil textures. The runoff depth was developed using Eq. (1) after estimating the curve numbers as shown in Fig. 11. The mean total annual rainfall of 550 mm at the 50% of time equal or exceeded and ʎ = 0.2 were used in this study. The highest runoff depth was generated in the area around the city center and the wetland area. Using natural breaks classification, the runoff depth ranged between 422 and 557 mm with an average of 451 mm. About 91% of the area had a runoff depth ranged between 422 and 480 mm.

Fig. 11
figure 11

Runoff depth map with the five classes

Full size image

Variation of assigned weights for factors affecting rainwater harvesting

The runoff depth, drainage density, slope, and lineament density maps were assigned weight as described in Table 9. Ideally, each map is supposed to have five classes, but the values for each map determined the number of classes; further descriptions are obtained in “Factors affecting the rainwater harvesting” section and “Assessment of runoff depth factor” section. This study indicated that the runoff depth had only three classes, drainage density for five classes, slope for five classes, and lineament density of two classes as shown in Fig. 12. A scale of 5 indicated the most influential and 1 indicated the least influential. Then, lineament density had more area-based influence with the covered area of 91.2% followed by slope factor with the covered area of 69%.

Table 9 Assigned weights for four factors with verbal judgement
Full size table
Fig. 12
figure 12

Spatial distribution of assigned weights and covered areas for the (a) runoff depth (b) drainage density (c) slope and (d) lineament density

Full size image

Normalized weights for factors affecting potential sites

Figure 13 shows the four normalized weight values for four factors arranged by adjacent intervals in a scale to indicate the spatial distribution. Three normalized weight classes of runoff depth ranged between 159 and 265% with an average of 212% and a standard deviation of 43.3%. Five normalized weight classes of percentage slope ranged between 27 and 135% with an average of 81% and a standard deviation of 38.2%. Five normalized weight classes of drainage density ranged between 14 and 70% with an average of 42% and a standard deviation of 19.8%. Two normalized weight classes of lineament density ranged between 24 and 30% with an average of 27% and a standard deviation of 3%.

Fig. 13
figure 13

Stepwise maps of normalized weight for drainage density, lineament density, percentage slope, and surface runoff depth

Full size image

Table 10 shows the order of the normalized weight and covered area for four factors. Based on the large percentage of covered areas, the runoff depth with the low normalized weight of 159% had a covered area of 92%; slope with the high normalized weight of 135% had the covered area of 69%; lineament density of high normalized weight of 30% had the covered area of 69%; and drainage density with the low normalized weight of 14% had the covered area of 47%.

Table 10 Normalized weights and percentage covered areas for four variables
Full size table

Potential surface runoff site identification

The four normalized weight maps of runoff depth, percentage slope, lineament density, and drainage density were synthesized with the WLC method using a raster calculator as shown in Fig. 14. Five weighted classes were obtained using natural breaks techniques to classify the potential sites. Most of the area in the north and south constituted a low percentage of a potential site for harvesting. The suitable area for surface runoff harvesting is placed in the area with a slope of fewer than 2.8° and drainage density greater than 52. The large areas are suitably located in the middle of the study area. Sensitivity analysis was performed to identify the influential variables to the suitability of the RWH site when its weight is relatively changed by a small amount (Sayl, et al. 2020).

Fig. 14
figure 14

Potential sites identification for RWH harvesting with the normalized weights of lineament density, slope, drainage density and (a) with all three classes of runoff depth, (b) with runoff depth of normalized weight of 159%, (c) with runoff depth of normalized weight of 212%, and (d) with runoff depth of normalized weight of 265%

Full size image

The sensitivity analysis was conducted by taking one class of normalized weight for runoff depth variable at a time and generated the potential sites for surface runoff harvesting while using the same values of lineament density, slope, and drainage density. The potential sites for surface runoff harvesting were increased with an increased covered area of runoff depth. Although, the very high suitability of potential sites were identified with a high class of normalized runoff depth. About 5.5% of the covered area was very high suitability sites, followed by 76.4% of the covered area as high suitability as shown in Table 11.

Table 11 Covered area with different suitability classes
Full size table

The potential site map was exported and overlaid to the Google Earth map (accessed on March 2021) indicated the number of ponds and major streams intersected with the area of very high and high suitability for harvesting the surface runoff. Figure 15 shows the main streams with the flow directed from Ntyuka and Nala to the Hombolo dam through Makutopora, Mkalama, Nzuguni, Swaswa, and Ng’ong’ona streams and also the main ponds including Nzige, Matumbulu, Ntyuka, Mkalama, Nala, Kigongwe, Makutopora, and Hombolo.

Fig. 15
figure 15

Existing intermittent streams and ponds near the suitable areas for surface runoff harvesting in Dodoma urban

Full size image

Surface runoff network and surface area estimation

The surface runoff network and surface area were estimated using triangulated irregular network (TIN) technique. The water surface with the triangular facets used the WLC class interval map as presented in Fig. 16. Each triangle indicates the surface runoff direction and slope gradients for surface runoff potential sites. The percentage of shape area and shape length of each triangle was summarized in Fig. 17. The values of WLC decreased with the slope for different triangles. The flatter of the ground surface consists of small slope values with the less surface runoff; but, the steeper the ground surface consists of the higher slope values that had the high surface runoff. Less than 400 triangles had a slope of greater than 10% and shape area less than 0.04%. The higher slope values are used to define the triangles of highest steepness to collect water in the surface runoff potential sites.

Fig. 16
figure 16

Surface runoff triangulated network with the WLC class intervals  

Full size image
Fig. 17
figure 17

Variation of triangle polygons with the slope, shape area, and shape length

Full size image

Conclusion

This study was conducted to identify potential sites for surface runoff harvesting for socioeconomic activities in the Dodoma urban. The Dodoma urban was selected for investigation due to single water source dependence and rapid growth with the increasing water needs for constructions and agriculture activities. Dodoma urban is dominated by agricultural land (58%) and forestry area (33%) with a sand-clay-loam soil texture. Clay content ranging between 18 and 33% makes Dodoma urban suitable for surface runoff harvesting.

In this study, runoff depth, drainage density, lineament density, and the slope were analyzed using AHP in an ArcGIS environment to identify a suitable location for surface runoff harvesting. The runoff depth was estimated using the SCS method based on the establishing curve numbers using land use, soil, and rainfall data. Runoff depth was the most influential factor in the selection of the site followed by slope, drainage density, and finally the lineament density. About 5.5% of the area showed a very high suitability site and 76.4% high suitability site for surface runoff harvesting. The remaining land portion of about 18.1% showed low to medium suitability for surface runoff harvesting, and most of these places correspond to the region with a high slope. The natural ponds were identified within the study area and corresponding to the identified potential sites for harvesting the surface runoff.

This study addresses a vital supplemental source in the areas of intermittent streams and scarce or unpredictable rainfall intensity. The identified potential map for surface runoff could be harvested in small and medium storage facilities. But historical data is used to classify the land use, rainfall, and soil for runoff depth prediction. The local anomalous features were neglected due to thelow resolution of the DEM map when calculated the lineament density, slope, and drainage density; should the current and future changes have a linear corresponding from the past, the suitable areas for surface-runoff harvesting would be significantly different. Subsidy individuals or water authorities could fasten the implementation of surface runoff harvesting. Further studies should set criteria to determine the specific catchments, estimate the quantity of surface runoff harvesting, and conduct economic analysis.