Salt content of reclaimed water from sanitation facilities in informal settlements and management options for sustainable agricultural irrigation

The salt content is the most important parameter when evaluating the suitability of water for agricultural irrigation (Pescod 1992; FAO 2003; Rhoades 2012). Salts need to be controlled to prevent buildup in soil. In the long run, sustainable agricultural land management of irrigated areas is only possible if salts are removed from the root zone. In areas where rainfall is high during at least some time of the year, infiltrating water is usually sufficient to leach salts from the soil. This might not be the case for areas where rainfall is low (Ayers & Westcot 1985; Letey 2000; Tanji & Wallender 2012). In arid and semi-arid regions, around 25% of irrigated areas are affected by soil salinization (FAO 2002). Economic losses on salt-affected soils range between 15 and 69% (Qadir et al. 2014).

Salinity is a measure for the content of dissolved salts in water. The exact salt content can only be determined by a complete chemical analysis. Since this procedure is very time consuming, salinity is often determined by drying and weighing of water samples or by use of a surrogate parameter such as electrical conductivity (EC), density, sound speed, or refractive index. When using a surrogate parameter, the empirical relationship between salinity and the chosen parameter has to be known (FAO 2003; Tchobanoglous et al. 2004; Eaton & Franson 2005). This study presents water quantities, EC, and total dissolved solids (TDS) in water flows from a project for sanitation and water reuse in North Namibia. Options for salinity management are assessed.

Namibia is the driest country in Sub-Saharan Africa. Total water demand exceeds supply (The World Bank 2009). In most areas of the central north, groundwater resources are not usable. Surface water bodies dry up after rainy seasons. Annual rainfall ranges between 350 mm in the east and 550 mm in the west. In the Outapi region, average precipitation is between 350 and 400 mm per year. More than two-thirds of the rain falls between January and March, only 4% from May to October. Precipitation may vary considerably in time (from year to year) and in space with a variation coefficient between 40 and 60%. This high variation, its limitation to a certain time of the year, and the lack of autochthonous water sources pose a risk for water-depending activities such as farming. For industrial, domestic, and agricultural water supply, the region depends on canal water supplied from the Angolan border river Kunene (Mendelsohn et al. 2000).

Namibia's population has grown from 1.8 million in 2001 to 2.1 million in 2011, half of them living in the central north. In the same period, the percentage of people living in urban areas has increased from 33 to 43% (Namibia Statistics Agency 2011). This development is a challenge for the provision of sanitation infrastructure and is reflected by the decreasing number of people with access to basic sanitation in the urban areas, from 82% in 2001 to 57% in 2010 (National Planning Commission Namibia 2004, 2013).

The Namibian water and sanitation challenges are representative for the region as a whole. Southern Africa is the most urbanized region of the continent. In 2010, 59% of the population lived in urban areas. This percentage is expected to increase to 78% by 2050 (Kayizzi-Mugerwa et al. 2014). Environmental constraints put additional pressure on water resources. Rising temperatures and lower rainfall will increase the number and severity of droughts. Altogether, these impacts are threatening food and livelihood security, water availability, and human health (UN-HABITAT 2014).

As a response to the water and sanitation challenges of the region, CuveWaters and the Outapi Town Council are implementing facilities for sanitation and water reuse in the city of Outapi in central northern Namibia. The infrastructure includes three different types of sanitation facilities, a vacuum sewer system for sewage conveyance, a combined anaerobic/aerobic wastewater treatment, a storage pond, and agricultural irrigation of crops for human consumption.

A communal washhouse in a very young informal area provides toilets, showers, hand washbasins, and laundry basins for the residents of the settlement as well as for people from a nearby market place. Thirty small cluster washhouses are shared by four families each in another area of the town. The cluster washhouses are equipped with indoor shower, toilet, and hand washbasin and an outdoor laundry washbasin. In a self-build neighborhood, households are individually connected to water pipes and sewers.

The collected wastewater undergoes the following treatment steps: anaerobic pretreatment (UASB reactors – upflow anaerobic sludge blanket), aerobic treatment (RBC – rotating biological contactors), lamella clarifiers, microsieve (15 μm), and UV disinfection. The reclaimed water is stored in a pond and then used to irrigate and fertilize crops that are sold on local markets. Operation of the sanitation facilities and the wastewater treatment plant started successively in 2013. In April 2014, the storage pond was included as last component of the water reuse scheme. It is planned to provide water and sanitation services for up to 1,500 persons.

This study focuses on quantity and salinity of the water flows and how to manage salts. The paper starts by describing the figures used for project design and how TDS were quantified using literature data. Then, the methods for data collection after implementation are outlined. In the next section, results on water quantities, TDS, and EC are presented and compared; differences are discussed; and possible options for the control of TDS input to the agricultural area are presented. For both literature and monitoring data, the TDS input through water usage is quantified in total and per capita. The last section offers some general recommendations on planning and implementation of measures for salinity management when reusing water from communal washhouses and individual households in informal settlements.

During planning, the Outapi water reuse scheme, expected salt loads and concentrations were estimated using average values in excreta from physiological data (Lentner 1977) (Table 1). The infrastructure was designed for up to 1,500 users and a water use of 60 L/(capita × d). From this, EC of the wastewater was calculated using a conversion factor (Eaton & Franson 2005). EC of the tap water was measured in one grab sample.

The water demand of crops was modeled by Woltersdorf et al. (2014) using the FAO software CROPWAT 8.0 (Smith 1992). Data for evaporation and rainfall were taken from Mendelsohn et al. (2000). Effective rainfall – that part of the rainfall that can be utilized by the crop – was estimated according to Savva & Frenken (2002). Loss is caused by evaporation, infiltration below the root zone, and surface runoff (Savva & Frenken 2002).

The data presented were collected in Outapi between 01 April and 31 October 2014 (exceptions: values for rainfall and evaporation refer to data collected from 01 October 2012 to 15 October 2014). Water quantities are measured with conventional (household) water meters at the communal washhouse, the cluster washhouses, and the individually connected households. The water use was recorded weekdays (community washhouses) resp. on a weekly to monthly basis (cluster washhouses, individually connected households). Water quantities in the influent of the wastewater treatment plant are measured online with electromagnetic flow meters (Promag 10W40, Endress + Hauser Reinach, Switzerland). Precipitation and evaporation are continuously recorded by an electronic weather station (iMETOS, Pessl Instruments Weiz, Austria). EC was measured in grab samples of tap water (at irregular intervals) and the storage pond (once per week) (1970i, TetraCon 325, WTW Weilheim, Germany) and continuously in the effluent of the wastewater treatment plant (Condumax, Endress + Hauser Reinach, Switzerland). TDS were determined in grab samples taken 26–31 July 2014 and 11 October–7 November 2014. For TDS, the water was filtered through glass microfiber filters (Whatman 934-AH Buckinghamshire, UK) and dried at 105 °C in porcelain evaporation dishes. Mean volumes evaporated per assay were 18.3 (±6.4) L for tap water, 10.9 (±0.3) L for the effluent of the wastewater treatment plant, and 15.7 (±0.5) L for the storage pond.

Figure 1 shows the components of the water reuse scheme, water quantities, EC, TDS loads of the water, and the factors affecting increase and decrease. An average value for the relationship between TDS and EC is used to calculate TDS content from EC of the tap water. The average quotient for TDS/EC is 0.625 (Eaton & Franson 2005). Accordingly, the calculated TDS content for the measured EC of 75 μS/cm is 47 mg/L (0.625 × 75 μS/cm = 47 mg/L). This value is used for calculating the total TDS load in tap water. Correspondingly, EC of the wastewater is calculated taking into account TDS contained in excreta and tap water.

EC and TDS increase while water is used in the sanitation facilities (for toilet flushing, showering, laundry washing, etc.). No change in EC and TDS was assumed within the wastewater treatment plant as no salts, flocculants, or other chemicals are added. Owing to evaporation (at constant mass of TDS), EC increases in the storage pond. EC of the irrigation water is expected to be about 1,205 μS/cm. The available amount of water (tap water + rain (pond) − evaporation (pond) + effective rain (irrigation site) − drainage water) is sufficient for irrigating 1.4 ha. A cropping pattern achieving highest revenues on local markets was chosen for implementation (Woltersdorf et al. 2014). For this cropping pattern (including plants such as tomatoes, maize, and water melon), irrigation demand is 17,760 m³/ha and year according to Woltersdorf et al. (2014). More than 21 t of TDS will accumulate on the irrigated area every year if no measures for salinity management are implemented. In this case, TDS are removed by a drainage system.

The data presented here were collected in Outapi between 01 April and 31 October 2014. It is used to calculate annual water quantities and TDS loads (Figure 2).

Water quantities are recorded for each cluster washhouse, for the communal washhouse, and the individually connected households. In case a water meter could not be read (e.g., if a cluster unit was occupied or there was nobody at home at the individual households), values were interpolated from previous or later data. During the survey period, the total amount of tap water used in the sanitation facilities (recorded by 102 water meters) was 7,551 m³. The amount of untreated wastewater registered online at the wastewater treatment plant was 7,314 m³. The gap of 234 m³ between the tap water use and the influent of the wastewater treatment plant is very low and shows high consistency of the data. Projected to 1 year, the total water amount available for irrigation is 12,533 m³. This is only 38% of the value used for planning (32,850 m³/a).

TDS is 41 mg/L in tap water, 437 mg/L in the effluent of the wastewater treatment plant, and 641 mg/L in the storage pond (Table 2). TDS of tap water was determined six times ranging from 36 to 47 mg/L. TDS ranged from 436 to 440 mg/L (n = 3) in the effluent of the wastewater treatment plant and from 620 to 655 mg/L (n = 3) in the storage pond. Since salts are not removed or added during wastewater treatment, the TDS load of the water only increases during its use in the sanitation facilities (+4.1 t/a). TDS concentrations change in the sanitation facilities (+396 mg/L) and in the storage pond due to evaporation (+204 mg/L).

EC is 52 μS/cm in tap water and 570 μS/cm in the effluent of the wastewater treatment plant. In the storage pond, the measured EC increased from 450 μS/cm in April 2014 to 805 μS/cm in October 2014 and might have reached steady state. Since this is not yet proven, a calculated EC is used in Figure 2. Considering the EC in the effluent of the wastewater treatment plant and the pond's water balance, the final EC is expected to reach 767 μS/cm. It is assumed that the TDS input of rainwater does not significantly contribute to total TDS loads.

The TDS/EC ratio was calculated with the EC of the samples used for TDS determination. It is 0.79 in tap water, 0.64 in the effluent of the wastewater treatment plant, and 0.81 in the storage pond. The values obtained for tap water and the storage pond are high but still within the typical range (Lloyd & Heathcote 1985; Eaton & Franson 2005; Tanji & Wallender 2012).

Irrigable area is 0.6 ha (available for irrigation = 12,533–3,717 + 598 + 1,552 m³/a, irrigation demand = 17,600 m³/(ha × a)). If no drainage is applied, the total TDS load in the irrigation water is 4.6 t/a for an irrigable area of 0.6 ha or 7.7 t/ha.

In the Shack Dwellers settlement, it was planned to connect 66 households in total. Even though the number of connected households rapidly increased from 16 households on 01 January 2014 to 35 households on 01 August 2014, this number has not increased since then. The average household size is five persons (Kramm 2014). During the survey period, the average number of connected households was 29. For the annual water balance, water amounts generated by 35 households (175 users) are taken into account.

For planning, the total number of users of the cluster washhouses was estimated using data obtained in situation assessment workshops in 2010 (DRFN et al. 2010). The survey in April 2014 showed that the household structure is different. Whereas the average number of persons per household was assumed seven in 2010, it is currently only three (Kramm & Deffner 2014).

The communal washhouse is designed to provide toilets, showers, and laundry washing facilities for up to 250 users (calculated with multiple usages per person per day). During the survey period, 199 usages were registered each day. If an individual needs to use the (toilet-) facility three (–five) times per day to cover all sanitation and hygiene needs, this corresponds to an equivalent of (40–) 66 users. However, it is more probable that the washhouse is used only once a day and that the sanitation facilities are not used to urinate. Thus, one can only roughly estimate population equivalents from usages/visitors per day.

For all facilities, the average water consumption per capita (or per capita equivalent as in the case of the communal washhouse) is 64 L/(capita × d). This is very close to the average value assumed for planning (60 L/(capita × d)). The water consumption is 27 L/(capita × d) in the cluster washhouses and 47 L/(capita × d) in the individually connected households. In the communal washhouse, the water use is much higher: 96 L/usage or 288–479 L/(capita equivalent × d); c.f. Table 3.

The per capita TDS load is 20–21 g/(capita equivalent × d) instead of 37 g/(capita × d). This is the average value calculated from the total TDS load added in the sanitation facilities and the total number of users as given in Table 3. For the washhouse, 40–66 per capita equivalents were assumed. Since EC and TDS are measured in the mixed wastewater from all sanitation facilities, it is not possible to differentiate water quality or TDS loads from each type of sanitation facility.

Altogether, the collected data show lower concentrations of TDS, lower EC, lower total TDS loads, and lower per capita TDS loads compared to literature or planning data. The per capita water quantity roughly meets the expected value, however, varies among the different types of sanitation facilities.

The lower total water quantity and lower total TDS load compared to planning data can be explained by the lower number of persons using the sanitation system during the survey period. Since the water use per capita is as expected, lower EC and lower TDS concentrations reflect incomplete excreta collection in the sanitation facilities. Social monitoring showed that 16% of the household members assigned to the cluster washhouses did not use the toilets even though residents are not charged for toilet use (Kramm & Deffner 2014). Users of the communal washhouse are charged for each visit. Here, the percentage of uncaptured excreta is assumed to be even higher. Even in the individually connected households of the Shack Dwellers settlement some excreta are lost. All users reported to pour their laundry water onto the plot area (Kramm 2014). Since dermal loss of NaCl and nitrogen can reach 60% in hot climates (Consolazio et al. 1963; IOM 2005), this user behavior might also contribute to lower excreta collection rates. It was observed (but not quantified) that open urination is still practiced widely, thus only a (small) fraction of urine per person is collected and urine separation is unintendedly practiced.

Without further promotion, the final number of users does not seem to be met in the near future. Owing to the incorrect estimation of users for the cluster units during planning, there is only limited potential to increase the number of users here. In the Shack Dwellers settlement, there are still a number of households that could be connected to the water supply and sewer system. However, the households have not yet applied for these services at the Outapi Town Council. In the community washhouse, a much higher number of users could be handled. Altogether, incentives are necessary to motivate residents to use the washhouses and to connect to the system. Measures could include the provision of general information to clarify the services provided and the conditions for usage (such as tariff details and connection procedures). Sanitation marketing is urgently required to create a higher demand for the services.

Literature data show that around 11 t TDS would accumulate on one irrigated hectare each year if no measures for salt removal are implemented (Figure 1). Monitored TDS loads for the Outapi water reuse scheme are lower but still substantial (ca. 9 t TDS/(ha × a), c.f. Figure 2). Strategies for soil salinity control in agriculture usually include measures on the agricultural fields, only (Pescod 1992; FAO 2003). In the Outapi case, additional measures prior to irrigation are considered. Possible measures are summarized as follows:

• - Salt removal prior to wastewater treatment: source separation of urine

• - Salt removal during wastewater treatment: reverse osmosis membrane filtration, ion exchangers, and electrodialysis

• - Removal of salts from soil: improvement of drainage, additional leaching, and salts contained in harvested crops

• - Reduction of salt input on fields: optimization of the irrigation system (e.g., drip irrigation instead of sprinkler irrigation), irrigation scheduling, evaporation reduction, fertilizer management, and blending of the irrigation water

• - Measures regarding crops and soil: cultivation of salt resistant crops, adjustment of planting procedures, and chemical treatment of soil

Source separation of urine is one measure to reduce salt input on agricultural fields. Urine contains 90% of the common salt, 80% of the nitrogen, and 50% of the phosphorus excreted by humans (DWA 2008; Powles 2013). Consequently, source separation of urine has a high potential for salt removal. In average, 80% of excreted urine is collectable (Johansson 2000).

In Outapi, urine diverting toilets are only feasible in the washhouses, but cannot be prescribed for the individually connected households. Considering the design data, only 64% of urine could be collected per year (80% urine collection rate from 250 + 840 users in the communal and cluster washhouses, 1,354 users in total, c.f. Table 3). This corresponds to 9.3 t (or 43%) of the 20.2 t TDS input per year. Considering monitoring data, 53–67% of urine would be collectable (80% of 40 to 66 + 318 capita equivalents in the communal and cluster washhouses, 539–559 capita equivalents in total) corresponding to 1.6 t/a (or 35%) of TDS added during water use.

To reduce salt input to the field, the collected urine would have to be disposed of or used outside the system boundaries. Nutrients contained would also be lost. Since no recipients for reasonable use of the urine outside the system boundaries were identified during planning, urine separation was not pursued for the Outapi sanitation concept. Technical considerations also played a role (e.g., need for double piping, further processing, storage, and transport of urine).

Measures during wastewater treatment, such as desalination with ion exchangers or reverse osmosis membrane filtration, could remove 90% of the TDS load. The TDS removal potential of electrodialysis is 50% (Tchobanoglous et al. 2004). Salt input could be reduced to 1.2 t/ha (literature data) or 0.9 t/ha (monitoring data). Still, accumulation of salts would be considerable in the long term. Besides, these options appeared to be too expensive in terms of operation, too energy-intensive and would increase the complexity of wastewater treatment. Therefore, these options were not selected for Outapi.

Cultivated crops are sold directly on site or on local markets. Salts contained in the crops would partially re-enter the sanitation system and partially leave the system. Data on the quantification of salt uptake by agronomic crops are scarce in literature. Richards (1954) and Ayers & Westcot (1985) exclude this topic. FAO (2003) recommends yearly or periodical cultivation of salt harvesting crops, such as sudax, barley, Bermuda grass, and sorghum, to reduce salinity buildup in soil, but does not give detailed information.

Following an optimistic calculation, a salt content of 5% is assumed in the dry mass of crops (Lyerly & Longenecker 1957). Up to 2 t TDS/(ha × a) are contained in, e.g., harvested tomatoes (65 t/ha per harvest, three harvests per year, 80% water content) and up to 1.2 t TDS/(ha × a) in harvested maize cobs (9 t/ha, three harvests per year, 10% water content) (FAO 2013). Thus, the salt removal potential per irrigated hectare would be 18% for the literature study and 25% for the monitoring data. This is in accordance with an overview given in Heuperman et al. (2002) who conclude that salt removal by crops is only significant for irrigation water with relatively low TDS content.

In regions with alternating wet and dry seasons (as in North Namibia), leaching during wet season can contribute to TDS removal. Mechilia (2002) concludes that 500 mm precipitation leads to salt removal in the upper soil layer (0–125 cm), and more than 600 mm are needed for leaching to a depth of 200 cm. In the case of the Sacramento-San Joaqin Delta in California, 400 mm of rainfall was sufficient for salt leaching (Ayers & Westcot 1985).

Precipitation in Outapi is 375 mm in average, but varies considerably. Since 1940, roughly 40% of the recorded years had precipitation above 500 mm (Mendelsohn et al. 2000). Leaching during rainy season combined with high crop production might be sufficient to remove salts from the root zone. Since literature data suggested much higher TDS loads during planning the Outapi case, salt management is carried out via regular drainage and leaching of the fields. A drainage system has been implemented discharging drainage water to a basin where the water is evaporated. As an additional measure, blending of the irrigation water with tap water is possible if required.

Salts in irrigation water and soil have to be controlled to allow sustainable irrigation. For water reuse, this is even more important than for irrigation schemes using conventional water sources, since salinity levels are higher in reclaimed water. During its use, EC of the water increases by 400–1,000 μS/cm (water use of 60 L/(capita × d)). Water with an EC above 750 μS/cm has a high salinity hazard (Richards 1954) and might cause yield loss to some crops such as peppers, maize, and fruit trees (Ayers & Westcot 1985). Even if TDS concentrations in tap water are very low and excreta collection is incomplete, 5–10 t of TDS will accumulate on the agricultural area per hectare and year.

Salt management measures prior to the agricultural use, such as urine separation or reverse osmosis membrane filtration, can reduce TDS loads; however, salts will still accumulate on the fields in the long term.

Salt uptake in crops is only substantial to the salt balance if the TDS content of the irrigation water is relatively low. When using low-salinity water for irrigation, salt removal with crops and leaching during wet season might be sufficient for salt control.

However, in arid and semi-arid regions excessive water use is not desired. If feasible, options not requiring additional water input, such as leaching during rainy season, salt removal with crops, water-saving irrigation techniques, or irrigation scheduling, should be considered for salinity control.

In the case presented, TDS loads are about 20 g/(capita × d) and therefore less than reported in literature. This is contributed to lower excreta collection rates in the washhouses but also to the small size of the catchment area. Residents might spend some of their time in other parts of the town and use other facilities for sanitation needs. Accessibility of the sanitation facilities plays a very important role for their acceptance and utilization and thus the amounts and concentration of the wastewater that has to be treated. High tariffs, unclear payment procedures, poor management, long walking distances, and unfavorable opening times lead to low utilization rates. This means that for optimal planning, the tariff system as well as specifications on operation and management of the communal sanitation facilities needs to be more or less clear before the design of the facilities for wastewater collection, wastewater treatment, and irrigation starts. For planning of further projects aiming at reclaiming water from communal sanitation facilities, it is recommended to use lower values for collected excreta. The per capita TDS load and water quantities presented in this work are specific results for the operation mode in Outapi, however, can serve as a basis for estimating water quantities for comparably managed communal washhouses. Management of communal sanitation facilities and social marketing measures are key points in achieving proposed utilization rates after implementation of the infrastructure.