Wastewater management is a major challenge in developing countries due to rapid growth of population particularly in urban centres, which is not

matched with the expansion of wastewater treatment infrastructure. Because of inadequacies in wastewater treatment systems, water borne diseases such as cholera have persisted. On one hand, use of conventional wastewater treatment systems has been proved difficult due to high construction and operation costs, their dependence on fossil based energy. On the other hand Waste Stabilisation Ponds and Constructed Wetlands are considered to be cheap, effective wastewater treatment systems, because of their reliance on natural processes. Other advantages include of these systems are simplicity, low cost, low maintenance, low energy consumption, robustness, and sustainability. The study conducted in Tanzania in mid 90s, established that more than 70% of the WSPs, were not performing according to their designs. This was attributed to lack of appreciation by many designers of the complex, physical, biological and chemical processes within WSPs and CWs; lack of consistency in design, construction and operation aimed at optimal performance; lack of appropriate design tools and methodologies suitable for local conditions; and changing nature of rapidly-developed technology. Studies were initiated in 1995, to develop an understanding of the complex processes in these systems with the intention of developing designs that may be suitable in tropical conditions. The studies have culminated into two design manuals. [1] Following is a summary of the design criteria for the CW and WSP.

Design of Waste Stabilization Ponds

WSP systems comprise of anaerobic, facultative and MPs in series. Anaerobic and FPs are designed mainly for removal of Biochemical Oxygen Demand (BOD), and MPs for pathogen removal. MPs are only required when the effluent is to be used for unrestricted irrigation, thereby having to comply with the WHO guideline of >1000 faecal coliform bacteria/100 ml.

Design of an anaerobic pond (AP)

Anaerobic ponds (APs) are commonly 2 5 m deep and receive wastewater with high organic loads (i.e. 100 g BOD/m3.day, 3000 kg/ha.day for a depth of 3m). They normally do not contain dissolved oxygen or algae. BOD removal is achieved by sedimentation of solids, and subsequent anaerobic digestion in the resulting sludge. Anaerobic digestion is more intense at temperatures above 15oC. The anaerobic bacteria are usually sensitive to pH <6.2. A properly-designed AP will achieve about 60% at 20o C. The APs are designed on the basis of volumetric loading (v, g/m3/d), i.e.:

(1) Where Li is influent BOD (mg/l), Q is flow rate (m3/day), and Va is AP volume (m3). The recommended that the loading should be between 100 400 g/m3.day[2] The hydraulic retention time is calculated from: (2) A retention time less than one day should not be used for APs; if it occurs, a retention time of one day should be used, and the volume of the pond should be recalculated.

Design of a facultative pond (FP)

FPs (1-2 m deep) are of two types, primary FPs (receive raw wastewater), and secondary FPs (receive particle-free wastewater (usually from APs, septic tanks, primary FPs, and shallow sewerage systems). The oxidation of organic matter by aerobic bacteria is dominant. FPs are designed for BOD removal on the basis of a relatively low surface loading (100 400 kg BOD/ha.day). FPs can be designed on the basis of kinetic or empirical models.

Kinetic model
BOD removal is often approximated from first-order kinetics:

(3) Where L is the quantity of BOD (mg/l) at any time t, and k1 is first-order rate constant for BOD removal (day-1). The FPs assumes to be completely mixed reactors and BOD5 removal follows first-order kinetics. [3] The rational design equation is as follows:

(4) or

The mid-depth area of the pond is calculated as:

(5) or

(6) Where Q is volumetric flow rate (m3/day), D is pond depth (m), and A is the mid-depth area (m2). The rate (k1) is a gross measure of bacterial activity and, is strongly temperature-dependent according to Arrhenius equation: (7) Where is the Arrhenius constant (usually between 1.01 - 1.09). Note that the temperature should be taken as the mean temperature of the coldest month.

Empirical model
In this model the FP is designed on the basis of surface loading: (8) Where Li is the concentration of influent sewage (mg/l), and Af is the FP area (m2). The design values of s increases with temperature[4,5] (9) Once the surface loading has been selected, the area of the FP can be calculated using equation (8) and retention time can be calculated

(10)) Where D is the pond depth (usually 1.5m), and Qm is the mean flow (m3/day).

The mean flow is the mean of the influent and effluent flows (Qi and Qe), the latter being the former less net evaporation and seepage. Therefore, equation (10) becomes:

(11) If seepage is negligible, Qe is given by: (12) Where e is net evaporation rate (mm/day).

(13) A minimum retention time value of 5 days should be adopted for temperatures below 20o C, and 4 days for temperature above 20o C. This is to minimize hydraulic shortcircuiting, and to give algae sufficient time to multiply (i.e., to prevent algal washout).

Design of maturation ponds (MPs) for faecal coliform removal

This design assumes that faecal coliform removal is according to a first-order kinetic model in a completely-mixed reactor[6]

(14)

Where Ne and Ni are the number of faecal coliform/100 ml in the effluent and influent, kT is the first-order rate constant for faecal coliform removal (day-1). For a series of anaerobic, facultative and MPs Equation (14) becomes:

(15) Where the sub-scripts, a, f and m, refer to the anaerobic, facultative and MPs, and n is the number of MPs Equation (15) assumes that all the MPs are equally-sized (the most efficient configuration, [6] this may not be topographically possible (in which case, the last term of the denominator becomes:

(16) Marais [5] found that: (17) At this stage of design tm and n, are unknown, ta and tf are calculated, Ni is measured or estimated, Ne is set (e.g., 1000/100 ml for unrestricted irrigation), and kT is calculated from Equation (15). The best approach to solving Equation (17) is to calculate the values of tm corresponding to n = 1, 2, 3, etc., and then adopt the following rules to select the most appropriate combination of tm and n, namely: (a) tm tf and (b) tm tmmin Where tmmin is the minimum acceptable retention time in a MP. A value of 3 days at temperatures below 20oC, and values of 4-5 days are recommended. [6] The remaining pairs of tm and n, together with the pair and , where is the first value of n for which tm is less than tmmin, are then compared, and the one with the least product selected, since this will identify the least land area requirements. The BOD loading on the first MP must not be higher than that on the preceding FP. The maximum BOD loading in the first MP should be 75% of that on the preceding FP. It is not necessary to check the BOD loadings on subsequent MPs. The loading on the first MP is calculated on the assumption that 70% of unfiltered BOD has been removed in the preceding anaerobic and FPs (or 80% for temperatures above 20oC). Mara and Pearson [7] suggested that 90% cumulative removal in APs and FPs, and then 25% in each MP, for temperatures above 25oC (80% and 20%, respectively, for temperatures below 20oC), when the BOD is based on filtered BOD values. Thus:

(18) Since Qtm =Am1D:

(19) The MP area is then calculated as: Am =2Q1Qm/(2D+0.001etm) (20)

Design of a constructed wetland (CW)

Constructed wetlands (CWs) are planned systems designed and constructed to employ wetland vegetation to assist in treating wastewater particularly secondary and tertiary treatment of municipal, industrial wastewater. The system is particularly useful for treating septic tank effluent or grey water, landfill leachate and other wastes that require removal of high concentrations organic materials, suspended solids, nitrate, pathogens and other pollutants. CWs should not be used to treat raw sewage and, in industrial situations, the wastes may need to be pre-treated so that the biological elements of the wetlands can function effectively with the effluent. Subsurface flow (SSF) systems are most appropriate for treating primary wastewater, because there is no direct contact between the water column and the atmosphere, there is no opportunity for vermin to breed, and the system is safer from a public health perspective. The environment within the SSF bed is mostly either anoxic or anaerobic. Oxygen is supplied by the roots of the emergent plants and is used up in the Biofilm growing directly on the roots and rhizomes, being unlikely to penetrate very far into the water column itself. SSF systems are good for nitrate removal (denitrification), but not for ammonia oxidation (nitrification), since oxygen availability is the limiting step in nitrification. The environment in the Free Water Surface (FWS) systems is generally aerobic at, and near, the surface, tending toward anoxic conditions near the bottom sediment. The microbial film grows on all available plant surfaces, and is the main mechanism of pollutant removal. FWS usually exhibits more biodiversity than does SSF systems. Accurate hydrological budget is needed to properly design a CW. The water balance to a wetland can be calculated as:

(21) Where Qi is influent wastewater flow (volume/time), Qe is effluent wastewater flow (volume/time), P is the precipitation (volume/time), ET is evapotranspiration (volume/time), V is volume and t is time. Attached growth models are the most common approach used to design constructed wetland systems. Depending on the type of CW being designed, a completely mixed reactor (no concentration gradients within the CW) or Plug flow (no back mixing) may be used. The general relationship for first-order plug flow models is:

(22) Hydraulic residence time is represented as:

or Substituting for t:

(23)

or Where kT is calculated using equation (7).

(24)

Where Ce is effluent BOD5 (mg/l), Co is influent BOD5 (mg/l), kT = Temperature dependent first order reaction kinetics rate constant (day-1), t is hydraulic residence time (day), Q is average flow rate through the system (m3/d), h is depth of the submergence (m), is Porosity of the bed as a fraction and As is surface area of the systems, m2. Basing on First Order - Reaction Kinetics Plug Flow Approach, two research-based design methods exist: Reeds method for the design of constructed wetlands Kadlec and Knight design method

Reeds method for the design of constructed wetlands This method is is based on the first-order plug flow assumption for those pollutants that are removed primarily by biological processes (i.e. BOD, ammonia, nitrate).[8] Reed suggests separate equations for total suspended solids (TSS) and total phosphorus (TP). For the removal of pathogenic organisms in CWs, the same approach as that used for WSPs.

For removal of BOD, NH4 and NO3 in constructed wetlands:

Equations 22 24 apply.

For TSS removal:

In SSF wetlands: (25)

In FWS wetlands: For Pathogen Removal:

(26)

Reed argues that the mechanisms for pathogen removal are essentially the same in both waste stabilization ponds and constructed wetlands. Then (27)

For TP removal:
In both SSF and FWS wetlands: (28)

The cross-sectional area of the flow is then calculated as:

(29) And the bed width

(30) Where Ce is effluent faecal coliform concentration (number/100 mL), Co is influent faecal coliform concentration (number/100 mL), kT is temperature dependent rate constant (day1 ), is number of cells in series, t = Hydraulic residence time (days) The cross sectional area and bed width are determined on the basis of Darcys law: (31) Where Kp is first order phosphorous reaction rate (2.73 cm/day), Ac is cross-sectional area of wetland bed (d*W) perpendicular to the direction of flow (m2), d is the depth (m), Ks is the hydraulic conductivity of the medium (m3/m2.day), S is The slope of the bed or hydraulic gradient (as a fraction or decimal). Kadlec and Knight design method A first-order decay, plug flow model for all pollutants, including BOD, TSS, total phosphorous (TP), total nitrogen (TN), organic nitrogen (OrgN), ammonia nitrogen

(NH4-N), oxidised nitrogen (NOx-N), and faecal coliform (FC) is assumed.[9] Their model is based on areal rate constants, which are independent of temperature.

(32)

Where

(33)

(34) Alternatively

(35) Where As is treatment area of the wetland (m2), Ce is target effluent concentration (mg/l), Ci is target influent concentration (mg/l), C* is background pollutant concentration (mg/l), KT is first order aerial rate constant (m/yr), q is hydraulic loading rate (m/yr), Q is average flow rate through the wetland (m3/day). Kadlec and Knight [9] advocate the use of the global parameters they determined from plug flow analysis of performance data available to date on the North American Data Base (NADB) [10] in other systems. They suggest that specific parameters should be locally determined prior to investment in a full-scale system, in order to ensure suitability of design.

Controlling organic loading

This is essential to control organic loading since it is a carbon source for denitrifying bacteria and to prevent overloading of the oxygen transfer ability of the emergent plants in the wetland system. If the carbon source is not available for denitrification, then lower overall nitrogen removal will result. Heavy organic loading, especially if not evenly distributed, will cause plant die off and odours. [11] BOD organic loading is expressed as follows:

(36)

Where OLR is BOD Organic Loading Rate Rates (kg/ha.d), Q is a verage flow rate through the system (m3/d), Ci is Influent BOD5 (mg/l), As is Surface area of the systems (m2). The BOD loading for SSFCW should not exceed 133 kg/ha.day. [12]

Controlling hydraulic loading

Hydraulic loading rate (HLR) refers to the loading on a water volume per unit area basis. It It is as determined as:

(37) Where HLR is hydraulic Loading Rate (cm/day), Q is average flow rate through the system (m3/d), As is surface area of the systems (m2). Hydraulic Loading Rate should not exceed 5cm/day. However Tanzanian experience reveals that Hydraulic loading of up to 20cm/day provide sufficient wastewater treatment.

Determine the Bed slope and Retention time of wastewater in the system
While bed slope influence water velocity in a treatment wetland thereby providing sufficient hydraulic gradient through the wetland, treatment performance in CWs is a function of detention time, among other factors. On one hand shorter detention times do not provide adequate time for pollutant degradation to occur. Longer detention times on the other hand can lead to stagnant, anaerobic conditions. Bed slope is determined as follows:

(38) Detention Time is determined as follows:

(39)

DAFTAR PUSTAKA Kayombo, S., Mbwette, T.S.A., Katima, J.H.Y., Ladegaard, N. and Jrgensen, S. E (2005), Waste Stabilization Ponds and Constructed Wetlands Design Manual, WSP and CW Research Project, College of Engineering and Technology, University of Dar es Salaam. Meiring, P.G.J.; Shipin, O.V.; Rose, P.D. (1998). PETRO system: A low-tech approach to the removal of waste-water organics. Water SA (ISSN 0378-4738), vol. 24, no. 4, October 1998, pp. 347-354 3. Marais G., Shaw V. (1961) A rational theory for the design of sewage stabilization ponds in central and south Africa. Transactions of the South African Institution of Civil Engineers Mara D. (2003) Design manual for waste stabilisation ponds in the United Kingdom, University of Leeds, School of Civil Engineering: Leeds, United Kingdom. Marais G.v.R. (1974) Faecal bacteria kinetics in stabilization ponds. Journal of Environmental Engineering (ASCE) 115(1): Mara, D. D. and Pearson, H. (1986). Artificial Freshwater Environment:Waste Stabilisation Ponds. In: Biotechnology (Rehm and Reeds, eds.). VCH Verlagsgesellschaft, Weinheim, Germany Reed, S. C., Middle brooks, E. J., Crites, R. W (1995), Natural Systems for Waste Management and Treatment, 2nd ed. McGraw-Hill, New York. 7. Kadlec, R. H and Knight, R. L (1996), Treatment Wetlands. CRC Lewis Publisher, Boca Raton, Florida, U. S. A 8. Knight R.L., Ruble R., Kadlec R.H., Reed S.C. (1993) Wetlands for wastewater treatment performance database. In: Constructed Wetlands for Water Quality Improvement, Moshiri G.A. (ed.) Lewis Publishers: Boca Raton, Florida, pp. 3558. 9. United States Environmental Protection Agency (1988). Design Manual: Constructed Wetlands and Aquatic Plant Systems for Municipal Wastewater Treatment, US EPA 625/1-88/022, September 1988. Metcalf and Eddy, Inc., 1991, Wastewater Treatment, Disposal and Reuse. Third edition, Revised by G. Tchobanoglous and F. L. Burton McGraw-Hill: New York. ISBN 0-07041690-7.