THE FATE OF PHOSPHORUS IN WETLANDS A REVIEW Report No. 96/15 For the Queensland Department of Natural Resources THE FATE OF PHOSPHORUS IN WETLANDS A REVIEW ACTFR Report No. 96/15 Prepared by J. W. Faithful of the Australian Centre for Tropical Freshwater Research James Cook University of North Queensland, Townsville Q 4811 The Fate of Phosphorus in Wetlands - A Review, ACTFR Report No 96/15 TABLE OF CONTENTS 1. INTRODUCTION ...............................................................................................................................1 2. SCOPE OF THE REVIEW............................................................................................................1 3. WETLAND SYSTEMS.................................................................................................................2 4. TYPES OF ARTIFICIAL WETLANDS .......................................................................................3 5. MECHANISMS FOR PHOSPHORUS REMOVAL IN WETLANDS ........................................6 6. REVIEW OF LITERATURE RELATED TO PHOSPHORUS RETENTION BY NATURAL AND CONSTRUCTED WETLANDS ....................................................................13 6.1 Australian Research ...................................................................................................13 6.2 International Research ...............................................................................................20 6.3 Underlying Criticisms of the Reviewed Journal Articles and Proceedings Papers ...27 7. CONSTRUCTED WETLAND CONSIDERATIONS FOR THE REMOVAL OF PHOSPHORUS............................................................................................................................28 8. MANAGEMENT RECOMMENDATIONS ...............................................................................34 9. CONCLUSIONS..........................................................................................................................36 10. REFERENCES ............................................................................................................................38 APPENDIX Australian Centre for Tropical Freshwater Research Page i The Fate of Phosphorus in Wetlands - A Review, ACTFR Report No 96/15 1. INTRODUCTION Natural and artificial (created or constructed) wetland systems are economical and aesthetic alternatives to water treatment technology for the control of nutrients and pollutants in waters. Over the past twenty years they have been used effectively to decrease the concentration of various pollutants from wastewater, stormwater, mining discharge and agricultural runoff, particularly in Europe and North America. There is a myriad of literature which reports the various functions of wetlands and the effectiveness of certain wetland designs in removing and treating pollutants. Much of the information has been focused on temperate wetlands, i.e. in North America and Europe, but research is growing on a wider international basis. Australian research is increasing in this area and several key groups are gaining credibility within this widening scientific field. The Cooperative Research Centre for Freshwater Ecology (Albury and Mt. Waverley), CSIRO Division of Water Resources (Griffith) and the University of Western Sydney are currently making effective inroads into understanding the use and design of wetlands in Australia. Despite the numerous articles published on wetlands in recent years there is a notable gap in the literature regarding research on natural and constructed tropical wetlands and their efficiency in reducing pollutants. It is difficult to apply temperate wetland data to wetlands in the tropics because of large biological and chemical differences due to warmer climates (such as increased plant productivity, increased potential for plant diversity, soil matrix activity, etc.). It has been suggested by Barbier (1994) that tropical wetlands have a high economic value and a crucial role to play in development, and the current lack of scientific data on their ecological relationships and functions is a cause for concern, particularly in developing countries. The issue of wetland efficiency in pollutant removal is further confounded by the interpretation of some data presented in books and journal articles, as well as the inconsistency of units used to express loading rates, plant uptake rates, etc. Much of the data has been presented on an inflow/outflow basis without consideration to the processes involved within the wetlands (chemical, physical and biological), with the assumption that wetlands act as a ‘black box reactor system’ capable of removing all manner of pollutants effectively. In recent years there has been a push to ensure that scientists and engineers, at all levels, take these factors into consideration when deciding upon the use of wetlands for any wastewater treatment application. A major source of the recent literature presented in this review has been from international conferences which have brought scientists and engineers together in an attempt to determine the best criteria for wetland design. 2. SCOPE OF THE REVIEW This review examines the fate of phosphorus (P) in natural and constructed (artificial) wetlands by focussing on recent Australian and international research. There has been much debate regarding the long term success of wetlands in removing P to any extent, and the processes by which the P is removed. In general, the issues that are addressed in this review are : • aspects of P mobility in both natural and constructed wetlands; • the success of constructed wetlands in P removal; • design criteria for the removal of P in artificial wetlands, such as plant species composition, planting patterns, retention time, wetland depths, type of wetland (open surface or gravel bed), effect of wetting and drying cycles on P removal, wetland shape, substrate types, necessity for harvesting and other parameters; • management recommendations for wetlands specific to P removal; and • comments on the appropriateness of constructed wetlands for P removal especially in the long term. This review examines the importance of all mechanisms occurring in natural and constructed wetlands that effect P removal, including plant uptake, assimilation by micro-organisms and algae, attachment to sediments and soil particles, and precipitation. Special emphasis is given to long term removal systems. Australian Centre for Tropical Freshwater Research Page 1 The Fate of Phosphorus in Wetlands - A Review, ACTFR Report No 96/15 The success of artificial wetlands for P removal is also examined with primary attention given to raw effluent quality, climatic conditions, age of the wetland and design criteria. The level of monitoring carried out with respect to specific studies is also addressed. 3. WETLAND SYSTEMS The most productive ecosystems in the world are those dominated by aquatic macrophytes (Brix 1993), and these may be divided into separate groups according to domination by free floating, submerged or rooted emergent species. Their high productivity results from having high availability of light, nutrients, and water, and from the plant’s morphological and physiological ability to take advantage of this environment. High levels of activity also occur at the microbial level resulting in the decomposition of organic matter and other substances. For these reasons, aquatic ecosystems (in particular wetlands) have been considered as alternatives and/or supplements to a variety of water treatment and recycling processes (Bavor et al. 1995; Kadlec 1995; Wood 1995; Brix 1994b; Cullen 1989). The conditions generated by an environment which is constantly saturated or covered by water limit gaseous exchange between the air and the sediment and this may result in anoxic sediment. As a consequence decomposition and mineralisation rates are significantly reduced, prompting the accumulation of organic matter on the sediment surface. This mass of organic litter, and the macrophytic vegetation, provides a significant surface area for microbial growth and therefore provides the impetus for large amounts of organic matter and nutrient transformation. The water/sediment interaction and associated microbial activity is the driving force behind water purification processes and therefore a sink for nutrients in both constructed and natural wetlands. Another significant factor which determines the effectiveness of a wetland as a water treatment system is the amount of time that the water stays in contact with the wetland, and this is related to the size of the wetland and the amount of water it receives. Natural wetlands are capable of significantly improving the quality of water flowing through them but the extent to which this occurs depends on the predominant water treatment process(es) acting within the wetland, such as macrophyte assimilation and substrate adsorption. There is great difficulty predicting the effectiveness of one wetland area compared with another. Natural wetlands are of high conservation value as they are very susceptible to changes in structure (in terms of species composition) and effectiveness with prolonged exposure to pollutants. There are, of course, exceptions to the rule, and several studies have shown that natural wetlands can function effectively in reducing nutrient loading from effluent discharge in various areas over a long period of time. Constructed wetlands have been used in various parts of the world for the past twenty years with little consistency in design criteria. Data obtained for these wetlands have generally been variable, but on the whole indicate that gross pollutant indicators such as biochemical oxygen demand (BOD), suspended solids and bacterial matter, can be effectively removed from the inflowing water. These constructed wetlands have been used to ‘polish’ treated outflow from primary and secondary treatment plants. Their design presents the water inflow with a defined flow pattern over a specific substrate and vegetation type which, as opposed to natural wetlands, is site selected and sized specifically for a controlled hydraulic pathway and retention time. Some advantages of constructed wetlands are their low cost of construction and maintenance when compared to the costs of treatment plants, their low requirement for energy, their flexibility and the fact that they are low-technology based systems. The disadvantages include the requirement for large amounts of land, depending on their use, and seasonal variability in their effectiveness. The most common wetland system used in water treatment is the macrophyte-based system where a wetland is constructed with one or several shallow ponds, and one or more species of aquatic macrophyte. The water inflow is generally regulated surface flow or sub-surface flow. The pollutants are removed from the inflowing water by a combination of processes (chemical, physical and biological) within the wetland, such as sedimentation, precipitation, adsorption to soil particles, assimilation by plant tissue and microbial transformations. Macrophytes can enhance pollutant removal within the system by either assimilating them directly or by providing an environment for surface microbial attachment to transform Australian Centre for Tropical Freshwater Research Page 2 The Fate of Phosphorus in Wetlands - A Review, ACTFR Report No 96/15 and uptake pollutants. The rhizosphere of aquatic plants is also a primary site for pollutant uptake and transformation as it is a zone of oxygen transfer between the plant and sediment which is a requisite for sediment microbial activity and pollutant oxidation (Brix 1994a). Over the past few years specific designs of constructed wetlands have resulted from data generated by many experimental treatment facilities. Factors such as horizontal surface or subsurface flow, vertical flow gradient and retention time have all been considered depending on the use of the wetland. These will be discussed in further detail at a later stage in the review. Pollutants such as settleable and suspended solids, BOD, metals and pathogens are generally removed by sedimentation and filtration processes, some of which can be provided by primary mechanisms such as gross pollutant traps situated before the wetland inflow. Soluble organic matter and nutrients (total and dissolved inorganic) are generally removed by bacterial oxidation and/or anaerobic degradation. The major removal mechanism of nitrogenous material is nitrification and denitrification, with some direct plant assimilation, and most constructed wetlands seem to be fairly efficient in providing mechanisms for these treatment processes. The mechanisms controlling the fate of the aforementioned pollutants are well understood, but P removal in artificial wetlands is not. The major processes that govern the removal of P in wetland systems are plant assimilation and substrate adsorption and complexation. Additionally, precipitation reactions can occur under certain conditions when the inflowing water comes into contact with available aluminium, iron, calcium and other clay minerals in the sediment (Howard-Williams 1985). Sediment fixation of P can also occur, and is enhanced by alternating wet and dry conditions. Plant uptake may be another important process and this is significant when specific loading and the soil P concentration is low. There have also been reports that phosphate reduction to gaseous hydrogen phosphides may occur under anaerobic conditions. 4. TYPES OF CONSTRUCTED WETLANDS Aquatic Macrophyte Wastewater Treatment Systems The dominant form of macrophyte within the wetland classifies the wetland treatment system. There are free-floating macrophyte systems, rooted emergent macrophytic systems, submerged macrophyte systems and multi-stage systems, which are a combination of the preceding systems, and other kinds of low technology systems (oxidation ponds, sand filters, etc.). The free-floating macrophyte systems are generally limited to systems based on water hyacinth (Eichhornia crassipes) or duckweed (Lemna, Spirodella or Wolffia spp.) in which the plants are allowed to grow in shallow ponds or cells within treatment ponds. Much has been documented on the productivity and effectiveness of these systems, particularly the water hyacinth (Karpiscak et al. 1994, Aoyama and Nishizaki 1993, Reddy and Sutton 1984). Hyacinth is more effective in warmer climates making it an ideal system for tropical environments but, as it is a declared noxious weed in Australia, it cannot be used. Lemna spp. and other duckweeds are more manageable; however, they are a minute plant, and therefore require barrier systems or compartments to minimise movement on the water by wind to ensure water surface coverage. Duckweed has a wider geographical coverage and climatic tolerance and, being a smaller plant without an extensive root system, is mainly used to assimilate nutrients and provide anaerobic light reduced water columns for denitrification and precipitation. The design of a Lemna based treatment system (The Lemna System) has been patented in the United States and has been effective in the removal of BOD, suspended solids and total nutrients (Poole and Ngo 1993). Emergent aquatic macrophytes are the most dominant form of aquatic plant in natural wetlands and marsh systems. In general they grow within a water table ranging from 50cm below the soil surface to a water depth of 150cm or more. They produce aerial stems and leaves and possess an extensive root and rhizome system. The depth penetration and mass of the root system, and hence their exploitation of sediment, differs amongst macrophytic species. In constructed wetlands, the most commonly used rooted emergent macrophytes are the common reed (Phragmites australis), cattail or cumbungi (Typha latifolia, T. orientalis or T. domengensis) and bulrush (Schoenoplectus validus or S. mucronatus). All species are Australian Centre for Tropical Freshwater Research Page 3 The Fate of Phosphorus in Wetlands - A Review, ACTFR Report No 96/15 adapted to growing in water-logged anoxic sediments as they have large internal air spaces for the transportation of oxygen to the roots and rhizomes. The extensive lacunal systems possessed by these plants occupy a large percentage of the overall plant. Oxygen is transported to the roots and surrounding rhizomes by either direct diffusion and/or by convective air flow. Air diffusion by the roots and rhizomes to the rhizosphere creates an envelope of aerated soil in the otherwise anoxic soil which stimulates organic matter decomposition and the growth of nitrifying bacteria (Brix 1994a). There are three major types of treatment systems based on emergent macrophytes as described by Brix (1993) : 1. Surface flow systems These systems represent the oldest type of artificial wetland design. They can vary from open wetlands, such as the very large constructed wetlands sometimes utilised in the U.S. (thousands of hectares), to short constructed channels. The channel system is typically no more than 100m long, 3-5m wide and planted with a species of bulrush. The water treatment processes are favoured by the presence of submerged portions of stems and litter which serve as a suitable substrate for attached microbial growth. The bottom of the channels are generally sealed to prevent wastewater leakage. 2. Horizontal subsurface flow This design was pioneered in Germany by Seidel in the 1950s and developed further in the 1970s (Brix 1994b). The system typically consists of a bed planted with common reed and sealed by an impermeable membrane to prevent flow loss. The medium in the channel is soil and/or gravel of varying size fractions so that during the passage of flow through the system, water contact with the plant rhizosphere is promoted and organic matter can be decomposed microbiologically, nitrogenous compounds denitrified and P and metals fixed in the soil. Two important functions occur through this system as a result of the reeds : • oxygen is supplied to the heterotrophic organisms in the rhizosphere, and • hydraulic flow through the medium is increased and stabilised. As a result of the subsurface contact and adsorption activity, the amount of nutrient assimilated by the plants represents a small portion of the overall content in the initial wastewater flow. Recycling of the nutrients bound in the plant tissue occurs upon the senescence and decay of the plant. BOD and suspended solids are removed effectively via these systems but the removal of nitrogen and P varies greatly depending on the loading rate of the wastewater, type of substrate, and the type and composition of the wastewater. The flow rate is an important factor as high input resulting in surface flow has to be avoided as this prevents the wastewater coming into contact with the sediment and the rhizosphere. 3. Vertical subsurface flow Vertical subsurface flow systems allow for improved hydraulic conditions and water/rhizome contact. This design provides percolation flow with intermittent loading which improves soil oxygenation when compared to horizontal flow systems. During the loading period, air is forced out of the soil and during the percolation phase the surface soil dries out drawing air back into the soil pore spaces. This process therefore provides alternating oxidising/reducing conditions in the soil promoting alternating nitrification and denitrification reactions and P adsorption. Vertical flow, and more significantly, vertical upflow systems are currently being developed by the CRC for Freshwater Ecology in Albury and preliminary findings appear to indicate that these systems are promising as single-use, low load systems such as household treatment systems, particularly for P removal (Breen and Chick 1995; Chick and Mitchell 1995; Mitchell et al. 1995; Heritage et al. 1995). Australian Centre for Tropical Freshwater Research Page 4 The Fate of Phosphorus in Wetlands - A Review, ACTFR Report No 96/15 The submerged macrophyte system uses plants which have their photosynthetic tissue entirely submerged. The diversity of plants available for use is great and includes low-productivity oligotrophic water species (Lobelia dortmanna), commonly occurring species in freshwater systems (Potamogeton spp., Ceratophyllum spp. and Myriophyllum spp) and high productivity species that thrive in eutrophic waters (Elodea, Elodea canadensis; Hydrilla, Hydrilla verticillata). These plants can assimilate nutrients directly from the water but only grow well in oxygenated waters. Therefore, these systems are not suitable for receiving wastewater with a high loading of organic matter. Their primary use could be to polish treated water, whether derived from a secondary treatment system or a low pollutant effluent source. High photosynthetic activity in these systems accompanied by good light penetration and warm temperature depletes dissolved inorganic carbon in the water resulting in pH values greater than 7 and, therefore, ideal conditions for P precipitation and the volatilisation of ammonia. The associated high concentration of dissolved oxygen will also favour the mineralisation of organic matter. The nutrients assimilated are generally thought to be translocated to, and retained within, the rooting tissues and microflora attached to the plant. Senescence and decay of the plant means that the plant nutrient and detrital matter rarely leaves the littoral detritus and macrophyte-epiphytic complexes. In some species of macrophytes the translocation of nutrients to the below ground biomass is not efficient so that senescing leaves or fronds have a substantial nutrient composition which may become available to the water column upon decomposition. Multi-stage systems are the most favourable designs to consider effective removal of pollutants. They allow for a wider range of pollutant removal, as well as allowing for the effective removal of fractions of the same contaminant, i.e. dissolved inorganic, particulate, and organic forms. Multi-stage systems permit the user to work from the process point of view and so dictate the required removal mechanisms within the wetland design. An example could consist of : • a mechanical clarification step for primary treatment to remove coarse sediments, oil and grease, etc., in the form of a gross pollutant trap; • a floating or emergent macrophyte system for secondary treatment which incorporates subsurface flow; and • a floating, emergent or submergent macrophyte system for tertiary treatment which can also incorporate subsurface flow. The multi-stage system, therefore, attempts to mimic the conventional wastewater treatment systems by separating the treatment processes to optimise the performance of the wetlands in relation to the needs of the treatment system (Brix 1993). The system design will of course, be based on factors such as the wastewater characteristics, the treatment requirements, the climate and the amount of available land. Currently the design of artificial wetlands is based on the premise, or expectation, that the system will have a simultaneous pollutant removal capacity, i.e. aerobic BOD degradation, microbial denitrification and nitrification, and P fixation occurring within the same ‘reactor’. Understanding of the key processes within wetlands has been qualitatively documented to date but the quantitative data on the rates of these processes and the factors that influence them are poorly understood or ignored. Recent studies have attempted to improve understanding of the wetland processes (Mitsch et al. 1995, Reddy et al. 1995, Johengen and LaRock 1993, Mitsch and Reeder 1991, Howard-Williams 1985). One cannot assume that an outcome of efficient pollutant removal can be successfully achieved in all cases. So, where wetland systems are proven effective in removing certain pollutants they have to be examined and their processes fully understood so that further wetland systems can be developed on the successful design criteria. 5. MECHANISMS FOR PHOSPHORUS REMOVAL IN WETLANDS There are several processes for the removal of P from wastewater in a wetland environment, but some have only a limited capacity for removal, and once the P assimilating/adsorbing capacity is reached or exceeded no further removal will be able to occur. In general, the ultimate pathway of P reduction is its movement from the water column to the sediment, uptake by macrophytes (from the water column or the Australian Centre for Tropical Freshwater Research Page 5 The Fate of Phosphorus in Wetlands - A Review, ACTFR Report No 96/15 substrate), algae and epiphytes and finally incorporation by micro-organisms associated with the litter. It does appear from work reported by Kadlec (1994) that accumulative P removal processes essentially follow first order reactions, which means that the removal of P to new soils is proportional to the concentration of P in the surface waters, and to the surface area of the wetland. Time should be allowed for wetland equilibrium to be achieved and this will result in stable water quality performance. In Kadlec’s case this was achieved after two years although the plant species composition and other biotic components continued to adapt over a longer period of time. Forms of Phosphorus within Wetlands Systems When determining the role of P retention by wetlands, particularly wetland substrates, it is important to understand the forms of P in the system. In wastewater entering a wetland, particularly secondarily treated effluent from a sewage treatment plant (STP), the P component will be composed predominantly of filterable reactive P (FRP) - which will consist of a mixture of orthophosphates and a proportion of labile condensed phosphates (McKelvie et al. 1993). The total P content comprises both inorganic and organic particulate and filterable non-reactive P forms, which will include a significant proportion of colloidal-P complexes and other P containing compounds, such as polyphosphates, metaphosphates and dissolved organic P (inositols, phospholipids, nucleic acids, phosphoamides, phosphoproteins, sugar phosphates, aminophosphonic acids, organophosphorus pesticides and organic condensed phosphates). The P composition of other sources of wastewater, such as urban stormwater, agricultural runoff and mining leachate, will differ and will be dominated by a particular fraction(s). The forms found within the substrate, particularly after water-substrate contact, are often delineated into inorganic and organic pools of P. The major pools of inorganic P (Pi) are defined as loosely adsorbed P, iron and aluminium P, and calcium and magnesium bound P. These forms are not generally discrete entities as transformations between the forms occur continuously to maintain equilibrium conditions (Sharpely 1995). The inorganic forms are dominated by hydrous sesquioxides, amorphous and crystalline aluminium, and iron compounds in acidic, noncalcareous soils and by calcium compounds in alkaline calcareous substrates. The loosely adsorbed P is important for plant growth and controlling the P concentration of the overlying water column (Reddy et al. 1995). This fraction responds to external P loadings and, as expected, is a proportion of both constructed and natural wetlands. The P associated with oxyhydroxides is readily desorbed under most conditions, but the P associated with crystalline iron and aluminium is desorbed only under extended waterlogged conditions (i.e. anoxic conditions). Redox conditions, however, have no effect on the sorption of P by aluminium, only on the sorption by iron (moody pers. comm.). In wetlands that have high levels of iron and aluminium in the catchment watershed, this fraction of P is generally the dominant pool of Pi. The calcium and magnesium forms of P are generally unavailable to biological assimilation under natural conditions and are not the predominant form of Pi under low pH conditions (i.e. those commonly associated with marshlands or high organic soils). However, under anoxic conditions, the sediment pH will most likely be neutral to alkaline and calcium and magnesium forms of P may be the dominant form. The organic P fraction (Po) primarily consists of the forms of P associated with phospholipids, inositols and fulvic acids, and forms of humic acids. This form of P is generally biologically reactive and can be hydrolysed to bioavailable forms. Organic P can be mineralised by alternate wetting and drying cycles, changes in substrate pH and increased microbial activity. Therefore, the bioavailability and mobility of P in wetland substrates under aerobic conditions is greater than that found in dry aerobic soils. The residual forms of P not able to be extracted by standard chemical extraction processes are considered to be highly resistant and therefore biologically unavailable. Australian Centre for Tropical Freshwater Research Page 6 The Fate of Phosphorus in Wetlands - A Review, ACTFR Report No 96/15 Plant Assimilation The capacity of aquatic macrophytes to assimilate P is dependent on their growth rates, the ionic composition of the water, the water depth, the sediment characteristics, the oxygen transfer capability of the plants into the root-zone, biochemical and physico-chemical processes functioning at the root-watersediment interface, plant density per unit area, plant harvesting and climate (Reddy et al. 1995). The potential rate of P uptake is limited by the plants’ net productivity and the concentration of nutrients in the plant tissue. Phosphorus storage is similarly dependent on plant tissue concentrations, and also on the ultimate potential for biomass accumulation (i.e. maximum standing crop). In addition to the plant assimilation of P, its removal in wetlands containing macrophytes is also affected by a number of biological, physical and chemical processes functioning in the water, sediment and rhizosphere. As aquatic plants grow, there is an uptake of P into the plant cells which will continue until after the plant is fully grown. At the end of the growing season, aquatic plants such as reeds die back and the leaves and stalks eventually fall to the bed where they break down. If the bulk of nutrients have not been translocated to the roots and rhizomes as is the case with some macrophytes, P will eventually return back to the system. New growth of these macrophytes will result in the uptake of P again so that an equilibrium will eventually develop where the P take up by plant growth in a year will equal the P return by dead plant breakdown. Therefore, if harvesting is not planned within wetlands vegetated by these particular macrophytes or the removal of their senesced leaves, the plants will bring about no net P removal. Hocking (1989) reported that, in general, the take-up of P into the plant cells is relatively small. The P content for plants such as reeds (P. australis) ranged from 0.9 to 1.35 mg g-1 (dry weight) for stems, 1.0 to 1.7 for leaves, and 0.9 to 1.63 for whole shoots. There is, however, no data on the assimilation rates into the roots and rhizomes of the plants. Davies and Cottingham (1993) calculated, for example, that if a yield of 95 tons dry weight per year per hectare of a very vigorous stand of P. australis up to 5 m tall was established (using the highest P content of 1.7 mg g-1 provided by Hocking, which excludes root biomass assimilation rates) a yearly removal of less than 6% P by harvesting the reeds would be expected from a typical artificial wetland with a loading of primary settled sewage with a P content of approximately 8mg P L-1 delivered at a rate of 96 L m-2 d-1. Hence the use of macrophytes to remove P from wastewaters, even under optimal conditions, can be severely limited if they are expected to provide the main P removing process. The main advantage of emergent macrophytes, however, is that they provide a greater surface area for epiphytic colonisation, and an additional pathway for nutrient assimilation, and that they produce litter stimulating microbial nutrient uptake. Data are available for the herbaceous macrophytes that are generally used in constructed wetlands to aid developers to select which plants are capable of storing high levels of nutrients. The general above ground P concentration (in vegetation and leaves) in temperate environments is of the order 0.1 to 0.3% (Johnston 1991). There is a wealth of literature available for temperate regions and the following plants have been rated highly for P content for leaves and herbs : Alternantha philoxeroides, 2.7-5.3 g m-2 (P content) and 0.39% P (P concentration at peak standing crop); Glyceria grandis, 5.2-6.8 g m-2 and 0.130.21% P; Phragmites communis, 2.0-5.3 g m-2 and 0.18% P; Sagittaria lancifolia, 3.58 g m-2 and 0.050.58% P; Typha glauca 3.17-3.74 g m-2 and 0.53% P, Typha latifolia, 0.7-3.2 and 0.1-0.4% P; Eichhornia crassipes, 0.5-18.0 g m-2 and 0.14-0.8% P; and, Lemna minor, 0.1-3.3 g m-2 and 0.75% P. These values underestimate the total P content within the plant because the root biomass and leaf litter are not taken into consideration. It is important to note that some of the plants which are highly rated for phosphorus removal are classified as noxious weeds in Australia (i.e. A. philoxeroides and E. crassipes) and cannot be considered for use in constructed wetlands. Finlayson et al. (1984) provided some information on the biomass and P assimilation potential of several macrophytes in a tropical lake in Mount Isa, north-west Queensland. In the winter month of July, the biomass of Potamogeton crispus was estimated at ~6 kg dry weight m-2, Hydrilla verticillata, ~3 kg dry weight m-2 and Salvinia molesta 0.81 kg dry weight m-2 and although P was determined as a limiting nutrient, the annual average P content was 0.07 to 0.47% P as dry weight for S. molesta, 0.09 to 0.41% P for H. verticillata and 0.11 to 0.38% P for P. crispus. Australian Centre for Tropical Freshwater Research Page 7 The Fate of Phosphorus in Wetlands - A Review, ACTFR Report No 96/15 A 12 month study by Shardendu and Ambasht (1991) in a tropical lentic wetland, found that the mean annual biomass of four species of submergent macrophytes was highest for H. verticillata (54 g m-2) which was present through the year, and lowest for Aponogeton natans with 12 g m-2 (the other species were Potamogeton pectinatus ~42 g m-2 and P. crispus ~22 g m-2). Although the submergent species only represented 6% of the total wetland biomass their involvement in net productivity around the submergent zone of the wetland made the study of their nutrient assimilation an important investigation of the natural wetland system. The low biomass percentage of these macrophytes was generally due to the absence of the species in this zone for at least one of the seasons. The mean annual P concentration varied from 1.03 mg P g-1 for P. pectinatus and 1.08 mg P g-1 for H. verticillata which varied significantly seasonally showing higher concentrations in the early growth phases and lesser concentrations when plants matured in the rainy or late rainy months. Water column concentrations of P declined during the growing phases of the macrophytes. Various scientists have attempted to compare the efficiency of several commonly used aquatic macrophytes in reducing nutrients and other organic matter under different flow rates of sewage representing retention times of 1.5 to 3 days (100 to 200L d-1). Ansola et al. (1995) compared Phragmites australis, Typha angustifolia, Iris pseudacorus and Scirpus lacustris cultures grown in small glass fibre tanks which maintained a water depth of 20cm. They found that total P removal from decanted village sewer wastewater (averaging 19mg P L-1) ranged from 47 to 61% depending on the application rates and was similar for all macrophyte species, with the plants having a greater affinity for P at higher retention times. The control tanks, however, removed a high proportion of P (40-47%) which was most likely a function of substrate adsorption. Rooted macrophytes obtain nearly all their nutrients from the sediment, whereas floating plants assimilate nutrients directly from the water column. Eichhornia crassipes is often reported as an ideal free floating plant to be used effectively to strip nutrients from wetland or pond systems due to its high productivity and nutrient retention capacity (Reddy and Sutton 1984), but it is classified as a noxious weed in Australia which prohibits its use as a wetland macrophyte in Australia. The plant is more effective in warmer climates making it an ideal wetland plant for tropical environments but it possesses great potential for blocking irrigation channels and rivers, spreading waterborne diseases and restricting drainage. If the use of hyacinth is considered, management of hyacinth then becomes a major issue as well as the constant consideration of harvesting to ensure optimum conditions for continual pollutant removal. Data from Aoyama and Nishizaki (1993) support these findings stating that as long as there is adequate area and harvesting for the plant to colonise, utilisation of water hyacinth would be a low cost and easily managed effective water treatment system. Tripathi et al. (1991) also reported on the greater potential for P removal by hyacinth than Pistia stratiotes (water lettuce), Lemna minor and Salvinia rotundifolia during the summer wet season. Pistia and Salvinia are also classified as noxious weeds in Australia and would not be considered for use in a constructed wetland system. Lemna dominated in the dry cooler months (winter) but on an annual basis hyacinth had the highest nutrient removal capacity, suggesting that the ideal free floating macrophyte strategy for nutrient removal in tropical ponds would be to combine hyacinth and Lemna. Species of Chara have also been shown to absorb reactive phosphate over a wide range of concentrations in ponded situations (Kufel and Ozimek 1994) and their capacity to outcompete phytoplankton makes this plant a good choice as a potential emergent/submergent wetland species. It is debatable, however, whether Chara would achieve a biomass sufficient to make it an effective P assimilator (G. Lukacs pers. comm.). Johengen and LaRock (1993) showed that nutrient uptake by Lemna sp. is significant and is advantageous as a free floating macrophyte for a wetland pond system because it has the ability to restrict phytoplankton growth by competing for available nutrients and limiting light availability. From mesocosm experiments, Lemna exhibited high removal rates (62%) when compared to the sediment (49%) and the water column (21%), and was comparable in P assimilation rates to the emergent macrophyte Pontedaria sp. The storage potential of P by floating aquatic plants is short term, however, because of its rapid turnover, and if the plants are not harvested, the stored P can be rapidly released into the water column during decomposition of detrital tissue. On the other hand, emergent macrophytes have more supportive tissue and thus provide greater potential for P storage. Australian Centre for Tropical Freshwater Research Page 8 The Fate of Phosphorus in Wetlands - A Review, ACTFR Report No 96/15 The role of plants in P retention has by and large been either not considered or grossly underestimated until recently. Reeder (1994) reported that for a small 56ha marsh near Lake Erie (Ohio, U.S.A.) plankton accounted for 67% of the water column biotic P removal with an uptake rate of 10g P m-2 y-1 compared to an estimated rate of 0.1g P m-2 y-1 by Nelumbo lutea (an emergent broad leafed macrophyte) which covered 23% of the wetland. This estimate was much higher than that predicted by a model previously developed from studies in the same area which estimated that 10-30% of any P flowing into the wetland would be retained or transformed by algae, depending on flow conditions where it would either be transported out of the system by viable plankton cells or deposited to the sediment of the wetland (Mitsch and Reeder 1991). Similarly, a study by DeBusk et al. (1995) showed that a periphyton filtration system removed at least 36.9g P m-2 y-1 from sugar cane runoff containing less than 100µg P/L. If a wetland design includes phytoplankton or enhanced periphyton communities as a predominant P removal process, management of the system must then ensure that overgrowth or blooms are controlled so that the system can maintain efficiency. It is suggested that periphyton-based systems would be more effective removing P from wastewaters providing the inflows are low load and low in P concentration. Species composition of plankton and periphyton associated with macrophytes are strongly seasonally dependent, particularly in warmer climates where a general shift from cyanobacteria in the summer to diatoms in the winter is observed in most wetlands. Vymazal and Richardson (1995) list, for example, the large range of species present in a study area in the Florida Everglades, and periphyton P concentrations ranged from 117µg P g-1 (winter) to 454µg P g-1 (summer) for attached periphyton. The periphyton P concentrations dramatically declined when the periphyton were detached. These biomass estimates were considered low and reflective of non-enriched areas as periphyton P values exceeding 5000µg P g-1 have been reported for nutrient-enriched wetland sites. It was also shown that where periphyton mass increased to such an extent as to form thick mats, the P assimilation capacity, and hence biomass P was greatly reduced. Table A.1 in the Appendix lists commonly used plants in international constructed wetlands with plants considered suitable for use in constructed wetlands in Queensland, Australia, for the treatment of wastewater. The table lists where possible, a comparison of mean annual production, mean annual standing stock, phosphorus storage potential, phosphorus content and phosphorus uptake rates. The table has many gaps which indicate that, particularly for plants considered suitable for Australian use, further research is necessary to understand production potential and phosphorus storage ability of many of the wetland plants. There are some cases of notably low estimates for mean annual standing stock (e.g. data from Shardendu and Ambasht 1991) particularly when compared to other data, so it must be assumed that some variability exists due to a number of factors that may not have been considered such as substrate effect, the influence of periphyton, plant age, climatic conditions, analytical error, etc. It has also been assumed from the data where it was not dictated, that most of the results have been expressed on a dry weight basis. Bacterial Assimilation Bacterial assimilation is generally at the litter/detrital zone above the sediment. Detrital tissue is found in wetlands dominated by macrophytes and generated as a result of factors such as the aging of plants, macrophyte overcrowding, pest damage and disease. Bacterial decomposition of the litter is predominant in the detrital zone and the principal form of nutrient transformation. Bacteria can assimilate P into their cell structure; but again, as a steady state is reached within the bed, no net P removal may occur thereafter. The biomass and productivity potential of bacteria in wetland systems, particularly their ability to assimilate P, is largely unknown. Physical Settlement and Accretion Physical settlement and accretion is one of the most important mechanisms of nutrient removal which could be exploited for P removal in natural and constructed wetlands. As a consequence wetlands are assumed to act as a P sink since most of the P that is retained ends up in the sediment (Richardson 1985). Providing that water retention is adequately addressed and the depth of the settlement area and flow rate of the influent taken into consideration, the physical settling of P laden contaminants in a well oxygenated system will result in a major proportion of the P being removed from the inflowing wastewater. Australian Centre for Tropical Freshwater Research Page 9 The Fate of Phosphorus in Wetlands - A Review, ACTFR Report No 96/15 Phosphorus accumulation in wetland substrates experiencing nutrient enhancement has resulted in peat accretion rates that are proportional to the concentration of P within the water column above the underlying substrate (Craft and Richardson 1993). The peat soils within eutrophic sectors of the Everglades, Florida, U.S., have been shown by the previous authors to provide effective storage of enriched P for extensive periods of time while at the same time providing a substrate stimulating net primary production which serves to increase peat production. Kadlec (1994) supports this view stating that the wetland biogeochemical cycle can operate to accrete new soils and sediments which contain P and that the soil-building processes provide a more permanent storage of P. Chemical Transformations The manipulation of various forms of P by chemical means to a form that will become easier to remove has been considered by several scientists. For example, the conversion of soluble P to insoluble particulate forms by chemical means (i.e. primary wastewater dosing with flocculants or within-bed means for adsorption or exchange) or by increasing pH in a photosynthetically active system, which associated with high calcium levels could precipitate phosphorus as calcium-phosphate minerals (i.e. a clear water lagoon with phytoplankton and/or rooted submergent macrophytes), would be an efficient method of removal. Diaz et al. (in press) suggests that that calcium concentrations would need to be greater than 100mg L-1 and the water column pH greater than 8.0 for effective P removal from surface waters, but this process can often account for greater than 60% of FRP precipitation. The ability of the beds to remove suspended solids ensures the retention of the contaminant if insoluble P particles can be successfully produced. This idea has led to the concept that chemical dosing within the bed, at a point where most of the initial suspended solids had already separated out, would be an effective and economical means of P removal. This concept relies on the secondary step of substrate adsorption to remove the formed insoluble particulate form and is discussed below. In any case, the ability of the substrate to adsorb and retain P will be governed by the complex interactions of redox potential (and therefore organic carbon), pH, iron, aluminium and calcium concentration and the amount of background substrate P (Faulkner and Richardson 1989). The direction of P flux across the substrate-water interface is regulated by the P concentration gradients across the interface, pH of the water column, sorption/precipitation reactions at the substrate-water interface, uptake by algae and macrophytes, the physico-chemical properties of the substrate and the incidence of any bioturbation at the interface (Reddy et al. 1995). The substrate-water interface layer is usually oxidised and its thickness is dependent on oxygen diffusion potential and oxygen demand within the zone. This zone can therefore potentially function as a P sink by immobilising P into insoluble ferric or calcium phosphate, as well as uptake and storage of P into the bacterial biomass. It is often assumed that oxic or aerobic conditions completely prevent P release from the substrate but mass balance studies by Ryding and Forsberg (1977) have shown that the release of P can be substantial from sediments to well aerated waters. This is generally found when weakly buffered, low pH, low P concentration waters (i.e. rainwaters) come into contact with sediments containing high concentrations of natural P. Substrate Adsorption The ability of wetland substrates to retain P depends on the physico-chemical characteristics of the substrate (Reddy et al. 1995). If P is to be removed by means of the matrix material it must be contained within the bed by ion exchange, adsorption or chemical reaction in an inert insoluble form. As these factors have a finite capacity, P removal is expected to cease when the capacity is reached regardless of the solution P concentration. The amount of P that a substrate will absorb will be determined by the P concentration in the water column; the higher this concentration, the greater the absolute amount of P which can be removed. Clay-type media with their abundance of aluminium, iron and calcium and large surface area have the greatest potential to trap and hold P, but their very low hydraulic conductivity in a bed designed for subsurface flow results in most of the water travelling across the surface and not making contact with the bed proper, thereby precluding efficient P containment. Australian Centre for Tropical Freshwater Research Page 10 The Fate of Phosphorus in Wetlands - A Review, ACTFR Report No 96/15 Various laboratory experiments have shown that wetland soils can adsorb P. Richardson (1985) demonstrated that organic soils were less suitable for P removal and that greater adsorption could be achieved from substrates with a high amorphous iron and aluminium content, which is found in mineral based soils. The presence of these ions under anoxic conditions resulted in more P being adsorbed by substrates when the solution P was high (i.e. retained by oxides and hydroxyoxides of iron and aluminium, and calcium carbonates) but more substrate P being solubilised and released when the solution P concentration was low. This followed from a study by Hill and Sawhney (1981) which concluded that reducing anoxic conditions caused by flooding cultivated soils actually increased P mobility after prolonged P loading. They did show, however, that periods of resting water (i.e. standing water due to no inflow) regenerated the sorption sites increasing the potential for additional P adsorption. Gravel media, with their high conductivity, permit all of the water to flow within the bed but because of their impermeable nature have only a limited surface area for adsorption, ion exchange, and/or chemical reaction to take place. Once the active sites are utilised P removal ceases. Davies and Hart (1990) found that unplanted and planted 30 x 5m channels of reed beds (Phragmites australis), unharvested over three years, had virtually no P removal in a basaltic substrate. However, a channel with a low hydraulic conductivity sandy soil substrate had an excess of 20% removal where most of the water flowed across the bed surface. In the same study it was found that in the gravel bed channel most of the wastewater solids were removed in the first 10 to 15m of the 30m reed bed channel so dosing at the beginning of the final 10m section would be at a point where most of the void spaces in the gravel would be available for the collection of insoluble P particles. The presence of plant roots in the wetland substrate reduces the void spaces available for water-substrate contact and hence the life of the substrate for P adsorption, but only to a limited degree as some of the P content of the water will be transformed and assimilated by the plant root system. Most substrates within the wetland beds will have a capacity to contain insoluble P and, if the wetland is designed with a chemical dosing treatment system prior to the wetland or incorporated into the wetland system, the wetland substrate would have a very long effective life. If, however, localised blocking of voids occurred a small amount of surface flow would not be a serious problem as it would flow in the plant litter layer and the litter would remove any phosphate particles until the flow re-entered the bed further down. When alum dosing is used in wastewater to remove P (as insoluble aluminium phosphate) at low levels of phosphate (10mg P L-1), there is competition from the formation of aluminium hydroxide and this necessitates a higher Al:P ratio than 1:1. Formation of the hydroxide lowers the pH significantly and reduces the formation of the aluminium phosphate because of its rapid increase in solubility as the pH decreases below 7.0. The addition of lime (calcium hydroxide) with the alum prevents the formation of acid conditions and thus improves the P removal. The issue of P reduction by dosing with lime and alum led Davies and Cottingham (1993) to establish an effective dosing mixture of 150 mg L-1 lime and 50 mg L-1 alum and found it to be economical and efficient for the removal of P within 30 x 5m gravel reed bed. Dosing directly to the bed presented mixing problems resulting in the wastage of alum by the formation of hydroxide instead of phosphate. The problem could be overcome by the use of a multi-stage system, a planted bed for the removal of normal solids, BOD and pollutants; a small pond for the dosing, mixing and sedimentation of the phosphate sludge; and a planted bed for final treatment and removal of any insoluble phosphate carried over form the dosing pond. Sediment or soil amendment within the channel bed by the addition or replacement of various substrates could be considered to improve the P adsorption potential of soils by utilising material such as sandy loam, industrial wastes or alkaline fly ash (Geohring et al. 1995; Cheung et al. 1994). Similarly, Mann (1994) suggested that by selecting specific P adsorbing substratum, such as industrial by-products, instead of relying on regional soils, P removal has the potential to be enhanced. Several potential problems exist with the use of external sources of material for wetland channel beds. One is their potential to have impacts on the groundwater environment by leaching and/or their potential toxicity to wetland plants. Another is on a cost/benefit basis where the cost to the developer to purchase and provide the soil supplement has to be measured against the requirement of the wetland for substrate amendment to improve phosphorus removal. Australian Centre for Tropical Freshwater Research Page 11 The Fate of Phosphorus in Wetlands - A Review, ACTFR Report No 96/15 Jones and Amador (1992) suggest that many techniques used to determine estimates of soil and sediment retention are mainly based on FRP, or orthophosphate, removal from water by soil or sediment columns under saturated flow. They argued that differences observed between the apparent kinetics and mechanisms of TP and FRP removal have ecological and environmental consequences as measurements of FRP removal are likely to underestimate the extent and rate to which P retention occurs. The rate of TP uptake was constant for a period of two days in peat soils from the Florida Everglades in saturated flow, while the rate of FRP uptake slowed and followed saturation kinetics. The majority of TP removed was within the 1-20µm range and it was shown that the primary mechanism for the removal appeared to involve abiotic hydrophobic and ionic interaction mechanisms rather than a filtering process. The removal of organic and inorganic P fractions from solution by physico-chemical processes is likely to describe the P dynamics in Everglades soils more adequately than data from sorption isotherms using only FRP. Much of the water within these marshes flowed vertically as well as horizontally through the soils so much of the water was in constant contact with the soil (density 0.3 g cm-3). Loss to the Atmosphere Loss of P to the atmosphere as phosphine was considered by a number of early workers and it is still debated to some extent (Devai et al. 1988, see also Brix 1993). Johnstone (1991), however, stated that even if phosphine was produced it would be adsorbed by the soil and quickly transformed and would not be liberated to the atmosphere. Australian Centre for Tropical Freshwater Research Page 12 The Fate of Phosphorus in Wetlands - A Review, ACTFR Report No 96/15 6. REVIEW OF LITERATURE RELATED TO PHOSPHORUS RETENTION BY NATURAL AND CONSTRUCTED WETLANDS 6.1 Australian Research The predominant wetlands research in Australia is with constructed wetlands and their application to improving water quality discharge for a number of industrial systems (i.e. sewerage treatment plants (STP), processing plants, etc.). Much of this work commenced in the 1970’s (Mitchell et al. 1995) and has been centred around south-eastern Australia. The use of constructed wetlands is, however, increasing throughout Queensland with strong Queensland Department of Natural Resources initiatives to integrate wetlands as a tertiary discharge treatment device for STPs and other applications in major regional areas. Research into the effectiveness of wetlands to reduce P export has been limited to the last few years although the data reported in many of the research publications and conference proceedings lists the final concentrations of P export as almost a matter of interest. The literature reviewed here has been selected on the basis of its reference to P reduction by natural and constructed wetlands. Sewerage Treatment Effluent Water Quality Control There are limited data and literature on P uptake and/or retention in natural wetlands. Finlayson et al. (1986) found that the P load from a natural wetland receiving secondarily treated sewage effluent, from the STP at Thredbo, N.S.W. in 1982, was reduced by 44% in the summer and 34% in the winter months. Patruno and Russell (1992) reported on a natural wetland in Yamba, N.S.W. which successfully ‘polished’ effluent from an STP which serviced approximately 4500 persons over the last twenty years. Phosphorus retention was on average 94%, despite the wetland not being in a pristine state. The wetland drained into a marine lagoon (Wooloweyah Lagoon) and concern has always been expressed as to the long term viability of the wetland as the loading is expected to increase due to a forecasted population of 8000 persons by the year 2000. Average P reduction was typically 5.5 mg TP L-1 to 0.34 mg TP L-1 at the outlet to the lagoon. Little information was given on the description of the wetland, a common oversight in many of the published articles on wetlands, so the predominant method of P removal cannot be deduced. Another natural wetland which has been successfully treating secondarily treated sewerage effluent is the Racecourse Wetland in Wyong, N.S.W. (Soukoup 1995). The 90ha wetland has been receiving on average 1.7ML effluent per day for twenty years from the STP which serves approximately 5000 people for most of the year. The dominant macrophytes were mainly submerged (Myriophyllum sp. and Potamogeton tricarinatus) and floating species (Ottelia ovalifolia, Azolla filiculoides, Lemna sp. and Eichhornia crassipes) interspersed through a variety of trees and reeds. Notably the wetland is a permanent lagoon which is covered by 0.3 to 1m water depth. Typical orders of reduction of P have been 10mg TP L-1 and 7.5mg FRP L-1 from the outlet of the oxidation ponds to around 50.0µg TP L-1 and 50µg FRP L-1 at the outlet of the wetland to the Wyong river. Some slight seasonal differences were noted in the retention/assimilation efficiency of P but were not considered significant. Constructed wetlands in Australia have been more variable in efficiency when it comes to removing P from effluent waters. Thomas et al. (1995) conducted a pilot study with a constructed wetland at the STP in Wodonga, Victoria, utilising a sub-surface inflow/outflow design with and without macrophytes (Schoenoplectus validus, Juncus ingens). Phosphorus removal was found to be approximately 13% mean removal efficiency for each of the cases studied, i.e. unvegetated, vegetated (mixed and single species), for several hydraulic loading rates and substrate sizings (gravel 10-20mm sizes) and through seasonal cycles. Studies in tropical regions of Australia are rare. Mockeridge (1995), however, reported at a National Conference on Wetlands for Water Quality Control at James Cook University, Townsville, on the use of wetlands to ‘polish’ secondarily treated effluent for re-use as a supply to the Town Common which is an established water bird habitat. Six wetland channels (15:1 length to width ratio) were constructed and planted with a variety of floating and emergent macrophytes (Typha orientalis, Lemna sp., Cyperus sp, Australian Centre for Tropical Freshwater Research Page 13 The Fate of Phosphorus in Wetlands - A Review, ACTFR Report No 96/15 Marsilea mutica (Nardoo), Phragmites sp. and Spikerush). Overall P reduction in the experimental channels was negligible over the duration of the trials and it was found to be difficult to maintain the planted vegetative communities without other macrophyte invasion or dominance by one or two species. Some initial reduction in P was noted in the discharge from the channels early in the trial and when two of the channels were put back on line but the capacity to assimilate P over any extended period by whatever means was reached very early in the trial. Mockeridge (1995) did note, however, that the roots and rhizomes of the macrophytes did have a higher P content than the macrophytes from the control channels. Vertical flow wetlands are an Australian development and the process has since been patented (Breen et al. 1989). These small scale systems utilise upflow or downflow hydraulic patterns and rely on plant uptake via the rootzone as one of the major nutrient removal mechanisms which supplements substrate adsorption (Breen 1990; Rogers et al. 1990). In the case of an upflow system, influent is generally introduced under positive pressure to the bottom of container housing the plant and substrate and the effluent collected at the top of the container via peripheral drainage tube (the downflow system is the reverse with the influent being applied to the top of the container). They are designed for use as small scale treatment wetlands with the focus on single use domestic treatment where loading is low and variable. There have been instances, however, where unplanted systems have been shown to be more efficient in removing P than planted systems utilising Typha orientalis, Schoenoplectus validus, Cyperus involucratus and Baumea articulatea (Heritage et al. 1995). The systems were loaded with 60L/day of approximately 4 to 8mg P L-1 wastewater with a retention time of 5 days. Some seasonal differences were noted over the thirteen month study whereby increases in outflow P concentrations where noted during periods of old shoot senescence in the spring, particularly with the systems planted with Schoenoplectus. The outflowing P was mainly found to be organic and particulate forms of P with little variation in FRP which was always low in concentration, particularly in the first three to four months of the study when absorption sites within the gravel substrate were more available. Early work at the University of Western Sydney involved the determination of nutrient budgets for wetland microcosms which again utilised the bucket approach and vertical upflow, i.e. a 10L bucket filled with a gravel substrate and planted with macrophytes. Different macrophytes were used (Phragmites australis, Schoenoplectus validus, Typha orientalis, Juncus ingens, Eleocharis sphacelata and Baumea articulata) and compared to non-vegetated controls within the experiment whereby primary settled sewage was applied for an approximate retention time of five days. The P load reduction was 99% (Breen 1990) which contrasts with the performance of a planted horizontal trench system at the Wodonga STP which received secondary treated effluent which only managed to reduce the P loading by 3%. In recent years there has been a large body of wetlands research published by the University of Western Sydney which have essentially reported the findings of applied research at several wetland treatment trials at STPs in Byron Bay and Richmond, N.S.W. Bavor and Andel (1994) reported on preliminary work in Byron Bay which involved discharging alum-treated effluent (dosed to reduce P concentrations in the effluent discharge to 1mg P L-1) to a treatment system comprising a number of upstream experimental scale units (rated at 120kL d-1), several small pilot scale systems and a 6ha broad acre wetland constructed with a number of treatment cells receiving 1ML d-1. The upstream experimental scale units comprised of four different units, each of which was monitored to determine efficiency in terms of P removal capacity. The first unit was an all gravel substrate which was divided into six sectors - one control, two Phragmites-planted, two Schoenoplectus-planted and one Typha-planted section. The preliminary results showed that the annual discharge from this unit was 0.15mg P L-1, representing a reduction in P loading by 85%. In preliminary results examining the capacity of individual species of macrophytes, Schoenoplectus and Typha performed better than Phragmites. Unit 2 was another horizontal gravel trench design planted with Phragmites followed by a grass planted open water meadow. The effluent from this unit averaged a final concentration of P of 0.09mg P L-1. The shallow open meadow section operated as a filtration system and enabled sediment adsorption of P. Unit 3 was an all-meadow, intermittent feed system, consisting of two grassed meadows, alternatively flooded to 5-10cm over two days and then drained over the next two days. P removal was not as good as the previous units for the first five months and the average concentration of P in the outflow has been close to or less than 0.5mg P L-1. Unit 4 was a shallow free water surface trench and results were not possible due to construction Australian Centre for Tropical Freshwater Research Page 14 The Fate of Phosphorus in Wetlands - A Review, ACTFR Report No 96/15 problems. The broad acre wetland has maintained low P levels (<0.5mg P L-1) with higher values associated with breakdowns at the STP. With respect to the P removal capacity over the first two years of operation an average of 80% reduction in P concentration, or 67% reduction in P loading has been achieved. Colleagues and students of John Bavor have investigated the P removal issue at the Richmond STP, N.S.W. Sakadevan et al. (1995) used a free water flow multi-trench system to remove P (and nitrogen) from secondary treated sewage under differing hydraulic loading conditions, P concentrations and retention times. The trenches were identical in design and size (30m x 5m) with a depth of 1m and were filled with clay to a depth of 30cm and topsoil to 30cm. Each trench was divided into seven segments and planted with three macrophytes (Phragmites sp., Schoenoplectus sp. and Triglochin sp.). It was found that the greatest P removal (62 and 63%) was observed in the trenches with the lowest hydraulic loading (2000L d-1) and the highest retention time (15 days). On average the outflow P concentrations were lower than their respective inflows but as is found when reviewing such work in many studies, some decreases in the outflow concentrations are small in respect to the inflow concentration, and may not be significant (e.g. one trench had an inflow of 8.3±0.5mg P L-1 and an outflow of 6.1±2.1mg P L-1). As expected, it was found that the substrates from the trenches receiving the highest concentrations of P in the inflow (~8mg P L-1) accumulated the most soil P. Rob Mann, a former student of John Bavor, has presented several papers at conferences over the last few years which report on the use of alternative and modified substrates in wetlands to improve P removal efficiency (Mann in press; Mann 1995; Mann and Bavor 1993; Mann 1990). Most of his work was conducted in field trials at Richmond with large trench systems (100m x 4m x 0.5m) along with laboratory trials testing the P assimilation capacity of various substrates. In trials that compared an open water gravel trench (T1) to a gravel trench planted with Typha sp. (T2), although 60 to 80% of P was removed by both trenches, it was reported that more P was removed by T1 in the first 18 months only because it contained more gravel, and that the capacity for P assimilation declined over the three years in both trenches (Mann 1990). Additionally, release slugs of P were noted in the open water stages of T2 indicating that P removal will be highly variable, thus agreeing with the findings of other authors (Steiner and Freeman 1990). At the end of the study, seasonal averages showed that the open water trench removed 40% of the P compared to 25% for the vegetated trench and that the substrate capacity for P adsorption declined after the first year of operation. Mann and Bavor (1993) reported the findings of a similar study but included an additional gravel based trench planted with Schoenoplectus validus. At best, only 40% of the inflow P could be removed by the trench wetland systems with the control trench behaving similarly to the planted trenches. As the substrate was again found to be the governing P removal factor, laboratory studies were used to investigate the efficiency of material (such as industrial wastes) which could be used as alternative substrates to enhance P removal by either substitution or modification of the existing substrate within constructed wetlands. Mann (1990) concluded that selection of a local material for use as a substrate within a constructed wetland should include an evaluation of the P assimilation capacity and its physical and chemical characteristics, such as aluminium and iron oxides content, porosity, hydraulic conductivity, particle size content and the calcium and magnesium content. Material such as blast furnace slag (160 to 420µg P g-1) and fly ash (260µg P g-1) were found to have much greater adsorption capacity when compared to regional gravel soils (25.8 to 47.5µg P L-1), and taking into account the potential for leachability and toxicity, are a strong candidate for substrate alternative and modifier for wetlands substrate where P removal is a requirement of the wetland. The other listed publications by Rob Mann (Mann in press, Mann 1995) are dedicated to management issues with regard to the design of constructed wetlands for use as P removal tools. Design criteria for sustainable operation were noted to be considered in relation to macrophyte harvesting requirements, substrate selection, manipulation of the system to allow for additional stages and whether multiple units are required for sustainability. The issue of monitoring was also considered which included the frequency of effluent and substrate sampling, as well as appropriate testing parameters which can be based upon the size of the treatment plants. Additionally, where wetlands are required for industrial effluent treatment, it was suggested that further consideration needs to be given to characterise the influent and the monitoring Australian Centre for Tropical Freshwater Research Page 15 The Fate of Phosphorus in Wetlands - A Review, ACTFR Report No 96/15 based on these parameters. For stormwater treatment, monitoring was recommended for high flow and flood events (during and following) where possible. Mann (1995) also introduces chemical dosing as an addition to the wetland treatment system and makes the point that where large P releases occur, an option of chemical dosing may need to be employed to reduce P concentrations prior to effluent discharge to the waterways. The assumption is that when treating secondary influent to a waterway where most of the P and nitrogen is in an oxidised form, chemical dosing should be applied before the influent reaches the constructed wetland systems to reduce the actual P load and thereby improve the sustainability of the wetland. Adcock et al. (1995) have taken a step towards addressing the process of P removal by attempting to determine the fate of P in a three year old constructed clay-based surface-flow wetland at Byron Bay (West Byron Bay STP), N.S.W. They found that the plants contained higher P concentrations at the inlet of the wetland and that the concentration declined with distance away from the inlet. Phosphorus concentrations in the sediment were high (due to high background levels) but the plants were responsible for the majority of P accumulation within this system (above ground biomass, roots and litter). The STP system utilised an aerated activated sludge system dosing with aluminium sulphate reducing the concentration of influent P to 1mg P L-1. The primary wetland (Unit 2a) was a subsurface flow wetland (25m x 50m) planted with Phragmites australis which received wastewater directly from the STP catch pond. It was then connected to Unit 2b by a series of V-notch weirs. Unit 2b (30m x 50m) had depths of 10-20cm and a retention time of two to three days. The macrophytic vegetation was dominated by Leersia hexandra, Urochloa mutica, with some Typha orientalis, Persicaria lapithofolia, Aster subulatus and Eichinochloa crus-galli which were all growing essentially hydroponically as there was no top soil and penetration into the clay base was limited. Phosphorus loads decreased through the wetland over the study period (inlet 400±130µg P L-1 and outlet 127±57µg P L-1) with the majority of assimilated P ending up in the plant stems, particularly near the inlet site. The plants accounted for 46% of the total P retained by the wetland while the standing water column P concentration accounted for less than 1% of the total P retained. The sediment P concentration remained fairly consistent through the wetland and indicated that water penetration did not exceed 2cm through the wetland. The high background P level made it difficult to assess the amount of P adsorption by the soil during the study. The CSIRO Division of Water Resources at Griffith has also provided an instrumental research team working on experimental wetland systems. Rogers et al. (1990) understood that sediment adsorption processes were instrumental for P uptake and became limiting when their capacity is reached. They utilised a vertical hydraulic flow system (both upflow and downflow) and provided a high porosity substrate which encouraged Schoenoplectus validus growth so that plant root contact was effectively promoting plant P uptake. Over the forty week study several treatments were applied which varied the application rates of P. During the first 29 weeks there was no significant difference in the performance of the up and down flow systems (98% TP, 8.0 to 0.7mg P L-1; and 93% FRP, 4.0 to 0.1mg P L-1 removal). When inflow rates where increased to 4L d-1 the upflow systems performance in removing FRP (85%) and TP (82%) declined. The downflow systems, however, maintained high removal efficiency with FRP from 10 to 0.2mg P L-1 and TP from 13 to 2mg P L-1. Unplanted control systems showed poor performances and their assimilation capacity reduced from 60% reduction at week 1 to <30% at week 37. It was found that the downflow systems were more efficient because of the high concentrated mass of plant root mass at the surface of the systems which not only acted as a baffle to reduce hydraulic flow but allowed for greater wastewater-plant contact. As a result 86% of the P load was found in the plant tissues which was maintained throughout seasonal cycles. These systems have been ideally devised to treat oneperson loadings at an equivalence of primary settled sewage and the results indicated that the potential effluent quality under these conditions would be acceptable. Breen and Chick (1995) compared horizontal and vertical flow systems with respect to root zone dynamics in nutrient retention with wastewater application. A comparison was made between 20L plastic experimental buckets planted with Schoenoplectus validus for the vertical flow systems and a 50m horizontal trench planted with Eleocharis sphacelata (pilot scale unit located at the Wodonga STP). Root densities were found to be the critical factor determining efficiency of nutrient retention in the two systems. The vertical flow systems operated more efficiently when the root densities were between 112 Australian Centre for Tropical Freshwater Research Page 16 The Fate of Phosphorus in Wetlands - A Review, ACTFR Report No 96/15 and 251g m-2 which took between 221 and 412 days to achieve. The horizontal flow system was disadvantaged by the less uniform vertical distribution of root mass which led to short circuiting of the flow, reducing flow contact with the roots (see also Bowmer 1987) and nutrient retention by both root assimilation and substrate absorption. The overall performance of the systems with regard to P removal showed that the experimental buckets reduced approximately 90% of total P whereas the trench actually released P (planted : winter; top 92.7%, bottom 84.6%; summer; top 93.0%, bottom 88.3%; unplanted : winter; top 78.8%, bottom 80.5%; summer; top 29.1%, bottom 35.7%; trench : winter -1.2%; summer 14.8%). This compares to a vertical downflow process, reported in a poster by Rogers et al. (1990), which assumed that 80% of P could be removed by the plants directly before the remaining P would be able to be adsorbed to the substrate. The efficiency of a series of pilot vertical flow wetlands located at Coffs Harbour, N.S.W. to treat primary settled sewage was reported by Chick and Mitchell (1995). The study was based on monitoring the inflow and outflow water quality on a weekly basis over a year (Jan 1991 to Feb 1992) and found that, in general, the removal of nutrients was lower than expected despite excellent removal of BOD, suspended solids and faecal coliforms. The overall treatment system was designed to treat 130KL d-1 whereby several vertical flow wetlands (VFW systems 1-4) dealt with 10KL d-1, and VFWs 5 and 6 sustained 85KL d-1 (VFW 5 had a modified design). The VFW were 12m2 at the top reducing to 10m2 at the base, filled with a gravel substrate and planted with emergent aquatic macrophytes with a theoretical retention time of 5 days. The VFWs removed about 37% of the P concentration and load within the system but it was shown that during extended retention times the removal capacity increased to 69% with 10 days retention. The longer retention periods favoured plant uptake, adsorption and transformation, and microorganism assimilation processes. The concentration of inflow was highly variable (2 to 12 mg P L-1) whereas the wetland outflow remained within 2 to 5mg P L-1, which is still higher than most Australian regulatory bodies permit for discharge to receiving waters. Like other systems which remove some of the P loads, the avenue is then open for other options to be implemented to apply additional forms of P removal and disposal, i.e. chemical precipitation. A limitation of this system is that the flow was directed upwards which has been shown in laboratory experiments by Rogers et al. (1990) to have a lower capacity to reduce P than downflow vertical wetland systems. The Environmental Sciences Faculty at Griffith University and the Queensland Department of Primary Industries has had access to ten pilot wetlands in Queensland which have provided the researchers with data based on different wetland configuration and application, and the effectiveness of different macrophytes types and species for a variety of different climatic. This work is particularly important as it involves the monitoring of wetlands constructed in different climatic conditions (the Queensland climate extends from arid to sub-tropical to tropical conditions) and therefore provides information on warm climate wetlands which has not been available in Australia to date. The research describes the wetlands in use throughout Queensland with several case studies already been reported at national wetland conferences (Bolton and Greenway 1995, see also Mockeridge 1995 - described earlier) presenting preliminary results of various established wetlands (in some instances over 2 years old) used to ‘polish’ wastewater effluent (Townsville, Ingham and Blackall). Overall, the wetlands reported showed greater than 40% reductions in BOD, TSS, and TN, with TP reduction at less than 10%. In all cases, as might be expected in the warmer conditions, the wetlands were quickly colonised with additional species of macrophytes. A potential problem in warmer climate constructed wetlands is increased density of plant biomass which will most likely cause a decrease in influent detention times but additionally the increased productivity of the plants might increase the potential for nutrient uptake. Phosphorus removal variability reported for some of the aforementioned wetland systems may in fact be due to the increased productivity because despite greater P uptake and accumulation in the plants (Greenway and Simpson 1996) increased litter production and microbial release would be expected in these conditions. These highly productive systems evident in tropical wetlands might suggest that macrophyte harvesting is an advantageous option to be considered in tropical constructed wetland design. In a follow-up study of a greater variety of wetlands, macrophytes in constructed wetlands were found to maintain higher productivity indicating a tolerance of nutrient enriched waters and were found to have higher nutrient bioaccumulation than similar species grown in natural control wetlands (Greenway in press). The maximum concentrations of P accumulated (uptake rates were not established) (c.f. other P Australian Centre for Tropical Freshwater Research Page 17 The Fate of Phosphorus in Wetlands - A Review, ACTFR Report No 96/15 contents in Table A.1 of the Appendix) were very high and this suggested that potential for nutrient removal utilising a variety of submerged and free and attached floating macrophyte species was very good. Pot trials using several species of Melaleuca trees under different conditions of effluent application were conducted by Bolton and Greenway (1995) to determine their effectiveness in accumulating P. Melaleuca commonly grows in seasonally inundated conditions and is often found fringing naturally occurring wetlands in Queensland. Bolton and Greenway found that all species (M. quinquenervia, M. alternifolia and M. leucadendra) accumulated high concentrations of P in the leaves with M. alternifolia accumulating twice the levels of the other two species. It was also found that P was also excluded by the Melaleuca in high P conditions as higher levels of P were found in senescent leaves, when alternatively in lower P treatments P recycling from leaves was evident due to very low concentrations of P in senescent leaves. These findings, and the generally tolerant nature of Melaleuca spp. in a variety of adverse environmental conditions, suggested that they may have very effective potential for use in wetlands. Water Quality Control of Rural Runoff Work conducted by Max Finlayson and co-workers in the 1980s looked at the use of wetlands to improve the quality of effluent discharge from rural industries such as piggeries, chicken farms, abattoirs, etc. Finlayson et al. (1987) reported on a pilot scale linear trench system to treat effluent from a piggery which was vegetated with Typha spp. but was relatively unsuccessful due to the concentration of the effluent and the associated problems with very eutrophic standing water, such as odour, algal and mosquito problems. Finlayson and Chick (1983) showed that three trench systems with gravel substrates separately containing Typha spp., Phragmites australis and Schoenoplectus validus slightly reduced phosphorus from effluent originating from a chicken abattoir. Similar odour, algal and mosquito problems were encountered in the trenches as the systems were initially designed to allow free standing water. Changes in design and inflow rate allowed for water levels to be maintained below the surface of the substrate. Effluent from these systems was clear and nutrient levels were markedly reduced. The effluent from a rendering plant was extensively investigated by Bowmer (1987) utilising gravel based trench systems vegetated with Typha spp., Schoenoplectus validus and Eleocharis sphacelata. She found that by comparing the vegetated trench with an unvegetated control trench, the vegetated system was prone to short circuiting, prompting preferential flow along the base and the sides of the trench. Approximately 40% of phosphorus was retained in the summer months which was presumed to be a result of substrate absorption. Irrigation Drainage Water Quality Control Natural wetlands have been used to control the quality of water that is discharged from agricultural drainage. En-route wetlands, or in-line wetlands, are characterised as wetlands with a single input and output and Cottingham (1995) has reviewed the effectiveness of four en-route wetlands in the Goulbourn Irrigation Area, N.S.W., for removing nutrients from irrigation drainage. The study area concerned is described in more detail in an establishment report (Milestone Report No. 1) to LWRRDC (Cottingham et al. 1994). The study was in an early stage and the data presented suggested that nutrient levels, particularly P, were higher in the leaving waters than the influent. It was suggested that as the wetlands were not ‘constructed’ for nutrient reduction, unregulated discharge of irrigation discharge would result in overloading and therefore reduced retention efficiency. As found in a similar study by Raisin and Mitchell (1995) in the Kiewa river catchment, there appeared to be a seasonal influence in wetlands nutrient removal efficiency of agricultural runoff where net retention was found in summer and net release in the winter months. A conclusion drawn by both sets of researchers was that where wetlands have to be used to efficiently remove nutrients from irrigation drainage, the processes involved in nutrient Australian Centre for Tropical Freshwater Research Page 18 The Fate of Phosphorus in Wetlands - A Review, ACTFR Report No 96/15 removal/retention have to be understood so that constructed wetlands can take into account the design features necessary to provide these processes. Wetlands in the highest area of P export from agricultural areas in the Peel-Harvey catchment, W.A. (the Meredith drain sub-catchment) were examined by Chambers et al. (1993) to determine their potential as nutrient sinks. It was found that in the three wetlands studied which were dominated by the sedge Lepidosperma gitudinale and surrounded by fringing Melaleuca sp. (1, in a reserve and undisturbed; 2, in farmland and undisturbed; and 3, in a farmland and disturbed), the surface sediments were the largest store of P and that the vegetation mediated the P pathway and its transformations, and through the formation of litter, its contribution to the sediment. The undisturbed wetlands had a higher organic composition in the soil (peat), higher vegetation to water ratios and much lower P concentrations in the water. Bucket experiments with Lepidosperma gitudinale and its peat substrate showed that after a three month establishment period, the P reduction from a source containing 9.9mg P L-1, which was continually applied via recirculation, was 67% in the first week and 98.8% after 40 days. Similar experiments were conducted varying the substrate, P application concentration and flow rate which showed that these were important considerations when attempting to design efficient systems if interested in P retention. Lepidosperma after three months establishment in either sand or peat achieved P reduction of 92 and 91%, respectively, with the application of 19.7mg P L-1 at an application rate of 42.0mg m-2 d-1 at 2.4L m2 -1 d over 59 to 161 days, which was a greater reduction than achieved for the unvegetated sand or peat (59% and 50%, respectively); and better than that achieved for higher application rates of 11.3mg P L-1 at an application rate of 120.7mg m-2 d-1 at 12.0m-2 d-1 over 13 to 45 days (48 and 60% for Lepidosperma in sand and peat, respectively) and 19.6mg P L-1 at an application rate of 248.3mg m-2 d-1 at 12.0L m-2 d-1 over 48 to 55 days (63 and 72% for Lepidosperma in sand and peat, respectively). When the substrate volume was increased from buckets to larger scale containers it was shown that the substrates were not overloaded and able to retain higher loads of P from 10 to 200 mg m-2 d-1 over a 68 day period. The conclusions stated, however, that although wetlands, natural and constructed, had the capacity for nutrient uptake from inflowing water, they may be of limited value in climates generating seasonal rainfall which produce high volumes of runoff in short periods of time. The greater volume of runoff may overload the wetland and disturb the vegetation which may in turn release nutrients to the receiving waters. The design of constructed wetlands therefore needs to accommodate the higher flow rates in these climates, including tropical regions where wet season rainfall can generate large volumes of runoff, and provide mitigation to ensure that flow rate is kept constant. Stormwater Water Quality Control The Blacktown City Council, N.S.W., implemented a 75ha stormwater control channelled wetland at Plumpton Park which was designed with a preceding gross pollutant trap (GPT) and 28 species of indigenous rooted emergent macrophytes. On a mass loading basis P reduction was estimated at 27% over an annual cycle which was retained by the wetland, with little influence in P reduction by the GPT (Hunter and Claus 1995). The order of P decrease through the wetland was 0.01-1.97mg P L-1 at the source to 0.01 to 0.18mg P L-1 at the outlet (averaged for wet and dry conditions). The GPT was found to be an integral part of the wetland design, not only because of its role in removing coarse litter and sediment, but that it reduced inflow velocities thereby reducing the potential for scouring and sedimentation in the initial stages of the wetland. Extended studies are required to determine the long term effectiveness of the system. 6.2 International Research Much of the international research on constructed and natural wetlands with respect to nutrient retention and/or removal has been carried out in Europe and North America, though some work is reported from Asia and New Zealand. This section of the review will discuss aspects of this work that relate to P removal, P dynamics and warmer climates. Europe Australian Centre for Tropical Freshwater Research Page 19 The Fate of Phosphorus in Wetlands - A Review, ACTFR Report No 96/15 European countries have utilised constructed wetlands to a varied degree over the past ten to twenty years. The more commonly used systems are essentially horizontal subsurface flow systems because they are a cheap alternative to secondary treatment, particularly for sewage effluent. Stringent legal rules have now led to the development of highly efficient vertical flow systems as well as combined systems (i.e. constructed wetlands in association with conventional biological plants) (Haberl et al. 1995). Vymazal (1995) presents an inventory of the constructed wetlands in the Czech republic which have a high capacity for reducing BOD, COD and TSS for a variety of treatment areas and population but limited potential for P removal. In general, the average P reduction capacity for European wetlands is 47.1% (n=338, Borner 1992, in Haberl 1995). A pilot scale system in Slovenia has reported a P reduction of 97% from greywater (25.7 to 0.7mg P L-1) and 71% from secondary sewage (1.0 to 0.3mg P L-1) during the first months of use (Urbanc-Bercic and Bulc 1995). This system relied on vertical flow through an initial bed system followed by a horizontal subsurface flow trench to optimise nitrification and denitrification processes. The initial beds were loaded with 30mm d-1 of secondarily treated sewage and greywater and were made up of two compartments of differing substrate sizes (A, coarse sand and gravel; B, fine sand and gravel) which drained into the horizontal trench at a rate of 1.5 mm d-1. The finer grained sediment in the vertical flow system (VFS) was more efficient in removing P from the greywater (95%, compared to 28% for the coarse grained substrate) but little difference was noted between the substrate grain size for the secondary sewage (A, 36% and B, 43%). The reduction of non-point pollution from agricultural runoff is not an easy task and often relies on the knowledge of pollutant flux concentrations and the transfer mechanisms. Dorioz and Ferhi (1993) examined the processes involved in a 14ha experimental watershed which flowed into a 3ha wetland with a mean outflow of approximately 6L s-1 and a maximum flow during storm events of 1000L s-1 within the Redon River watershed, France. The mass balance equation showed that the wetland received measured P export coefficients of 0.6kg ha-1 y-1 of which half the P export was FRP. The P fluxes from the watershed were very variable and essentially transferred during stormflows. These flows intercepted by the wetland were shown to decrease P loads by 75% (via decantation and precipitation) which when extrapolated to the whole Redon watershed (33km2), accounted for 50% of the P load. The buffer zones generated by the wetlands therefore played an important role in the improvement of water quality from agricultural areas particularly to receiving waters. The United Kingdom Water Industry have been using Phragmites reed bed treatment systems for the past ten years and research into their use and effectiveness has produced 200-300 operational wetlands which have finally become appropriate and efficient systems for secondary and tertiary treatment and nitrification for English villages (Cooper and Green 1995). Their focus is towards a mainstream treatment process for the removal of BOD, TSS, ammonia and organic nitrogen with little consideration is given for P removal potential. This treatment is achieved via horizontal flow, vertical flow wetland systems or a combination of both. Green and Upton (1995) present more specific data on the performance of secondary, tertiary and stormwater Phragmites reed bed wetland systems under the control of Severn Trent Water, Birmingham, which is a privatised regional water authority serving a population of approximately 8 million. The privatisation has brought much more emphasis on the legal conformity to discharge requirements which has meant that the wetland design criteria has had to be successful for over a hundred sites. The Yorkshire constructed wetlands have been operating since the mid 1980s, although there were examples cited from 1903 (Earby) and 1964 (Highroyd) (Hiley 1995). These were mainly surface flow, subsurface flow (soil and gravel) and raft lagoon (floating algal mats or dense macrophytes) wetland systems with inflow volumes varying from 0.02 to 1.3m3 m-2 d-1. The P removal capacity was minimal and the P concentrations, relative to the load, were highly variable and could not be correlated with the type of wetland used. The surface flow wetlands achieved high effluent standards but generally lower load removal than subsurface flow systems. North America Australian Centre for Tropical Freshwater Research Page 20 The Fate of Phosphorus in Wetlands - A Review, ACTFR Report No 96/15 In North America there are a number of wetland studies running throughout the United States and Canada which vary from large open natural wetland treatment systems for agricultural and stormwater runoff to small horizontal flow systems for sewage treatment. A study by Gehrels and Mulamoottil (1989) of the transformation and export of P from an 18ha natural wetland (predominantly Typha sp. marsh) in Ontario, Canada, found that total P imports were nearly double the exports and orthophosphate export was 22% greater than imports. Phosphorus budgets indicated that surface P loadings to the wetland were 2.0kg ha-1 (TP) and 0.4kg ha-1 (FRP) which accounted for 75% (TP) and 68% (FRP) entering the wetland from an 130ha agriculturally dominated catchment. This led to the assumption that internal processes within the wetland transformed the sediment bound P to available P which was then seasonally discharged in fall with higher water levels promoting anaerobic conditions and at the end of the growing season when leaching occurred from decaying vegetation. Very high retention rates were observed in the spring and 95% of this retained P fraction was bound to the substrate. These findings showed that seasonal fluxes in environmental conditions are limiting to the overall efficiency of wetlands and must be considered in the overall design criteria if constructed wetlands are to be of use to water treatment managers. A cold climate study was conducted on a small wetland treating fertiliser leachate from a 10km2 athletic field prior to drainage into Lake Tahoe, U.S.A. (Reuter et al. 1992). The 660m2 (33 x 20m) wetland was excavated to 1m depth and filled with sub-6mm gravel substrate and vegetated with Typha latifolia. Inflow to the wetland was non-regulated and a function of climate (i.e. fall precipitation, winter snow, snow melt and dry summers) with an estimated runoff volume of 873m3 for 1988 study period and temperature range of -3 to 21°C. The wetland system had limited ability to remove FRP from the influent with an annual removal capacity of -28% (concentration) and -41% (load) meaning that the system acted as an exporter of FRP. During the summer months the system assimilated FRP, removing 14% but dropped to -31% during the winter. The authors suggested that the assimilation in summer was mainly due to plant uptake, but the poor annual retention was a result of the contractors using unwashed gravel to fill the wetland bed during construction. Particulate P removal was more effective with an average annual reduction of 47% (concentration) and 44% (load), dominated by the summer removal (68%). The wetland was therefore a sink for the particulate P, unlike FRP which remained in solution and was easily transported out of the wetland during periods of water flow. A compounding factor appeared to be that the Typha had not established sufficiently for it to effectively reduce FRP, but then again higher production rates would be expected in less mature growing plants. The study was preliminary and should have been allowed to establish for a longer period of time for an accurate assessment of the P removal capacity of this wetland system. A practical design for a constructed wetland to improve the quality of agricultural runoff is shown in Higgins et al. (1993) which treats runoff from a non-point source prior to discharge into Long Lake, Maine, U.S.A. The design utilised an inlet flume which directed inflow to a sedimentation basin (5 x 46m) after which it was evenly dispersed by a level lip spreader (0.6 x 15m and 0.46m deep) over a grass filter strip (38 x 46m) used to detect the presence of pesticides and herbicides (6% slope). The overflow, and subsurface flow was then directed to the 980m2 wetland (2% slope), planted with Typha latifolia which buffered the flow into a detention pond (15 x 46m and 2m deep). The wetland-pond system had a significant role in decreasing the impact of desorbed FRP. Agricultural soils above the treatment system had a low pH (therefore an increased ability to adsorb P) but when they were eroded during increased surface flow from storm event runoff and entered surface waters (which may have higher pH), the soils ability to hold P decreased. The pond maintained a pH greater than 7.0 which allowed for P assimilation by the wetland and pond biological community. This may have been due to strong algal productivity and a predominance of calcium and magnesium bicarbonate composition in the water as the pond sediment was found to have 12% of the influent TP recorded over the study period. Outflow was then directed through an outlet flume to a 46m long vegetated swale that drained to receiving waters before flowing into the lake. Annual removal efficiencies for TP between 1990 and 1991 were 82 to 92%, although seasonal removal variability was high (in spring the outflow exported more P than imported - mainly due to outflow exceeding inflow due to a higher groundwater table). It was concluded that, for the amount of land required and the relative construction costs, the system worked effectively because most of the P was associated with the solids. Australian Centre for Tropical Freshwater Research Page 21 The Fate of Phosphorus in Wetlands - A Review, ACTFR Report No 96/15 An upland wetland wastewater system was used to test the effectiveness of wastewater treatment for a single family home in North Carolina by House et al. (1994). It consisted of a mound which provided an aerobic environment to effectively reduce P concentrations by precipitation as calcium and magnesium phosphates which was then followed by a wetland cell. Two emergent macrophyte species were compared, Typha augustifolia and Phragmites australis, with P. australis proving to be more effective. The evaluation over 15 months showed that 86% of the P could be removed by the mound due to its marl component, and that the remaining P reduction by P. australis was due to the higher above ground biomass contributing to a greater accretion of P as organic matter. This indicated that substrate adsorption and P precipitation within the mound component was most effective at the removal of the P while the wetlands contributed little. The Des Plaines River Wetlands Demonstration Project near Chicago, Illinois, has been used to assist in evaluating the use of wetlands to improve the turbid river water quality of the Des Plaines River which drains an urban and agriculturally impacted catchment, and assess the wetland function. Four wetlands have been used, varying from 2 to 3.5 ha dominated by Typha spp., with a maximum depth of 1.5m (Hey et al. 1994). The data showed that regardless of the relatively low concentrations of nutrients, the constructed wetlands were effective in reducing the nutrient loads and clarifying the inflowing water. Inflow total P concentrations were significantly reduced at the outfalls of each wetland (155µg P L-1 to 2 48µg P L-1, accounting for a 52-99% reduction of P) with the major removal mechanism assumed to be a settling of sediment particles and biological uptake. Improved efficiency within the second year of operation was attributed to increased plant uptake and carp control efforts. In a concurrent study by Christensen (1994), the role of changing flow conditions in the wetlands on P retention was modelled and it determined that increasing inflow rates resulted in increased P retention but a decreased retention efficiency. A subsequent follow-up paper by Mitsch et al. (1995) showed that average P concentrations decreased from a three year average of 176µg P L-1 by 64-92% (11 - 40µg P L-1) in low flow wetlands and by 53-90% (12 - 57µg P L-1) in high flow wetlands and that intensive sampling showed that retention in the low flow wetlands was better than high flow wetlands. Clear gradients of P removal were shown in the high flow systems whereas the low flow wetlands exhibited near homogenous concentrations of P within the system. The paper supported sedimentation as the predominant removal mechanism with some capacity for macrophyte uptake and a lesser involvement by periphyton and microbial communities. The constructed wetlands were shown to retain 0.5 to 3.0 g m-2 y-1 which agreed with the values suggested by the model developed by Christensen (1994). Two-stage lagoon systems are commonly used in the dairy industry and Chen et al. (1995) presented the preliminary results of a project at Louisiana University using several types of aquatic plants as a tertiary treatment to improve wastewater generated by the dairy feeding lot washings. The aquatic plant systems consisted of a pond with black willow (Salix niger, selected because of its rapid growth in wet conditions) and duckweed (Spirodela sp.), and a pond with water hyacinths (Eichhornia crassipes, selected because of its rapid growth rates in summer months). Hydraulic loading to the ponds varied from 0 to 107 m3 ha-1 d-1 and they were found to be very effective for TSS, BOD and faecal coliforms. Phosphorus removal was not significant (i.e. non-apparent) in either pond. It was suggested that harvesting the plants might have improved the P removal, although given the high concentration of the lagoon effluent (20-125mg P L-1) the system was probably superloaded. Alternatives such as dosing should have been considered to reduce the P loading prior to the wastewater inundating the macrophyte systems to make them more effective. A great deal of research is currently being conducted in the Florida region, particularly with emphasis on the bioremediation of P from agricultural wastewater and runoff. Pioneering work in central Florida was conducted by Dolan et al. (1981) on P dynamics within a freshwater marsh dominated by Sagittaria lancifolia, Pontedaria cordata, Panicum spp. and Hibiscus sp. receiving secondarily treated effluent (approximately 38g P m-2). The results showed that over the initial study year on 2km2 plots constructed within 32ha of marshlands, 97% of the P was removed by the wetlands, little P was assimilated by the groundwater and there was a much higher net production in the plant biomass. The effluent discharged from the plots remained close to the background marsh and receiving water quality. The substrate stored approximately 69.2% (assuming the validity of the soil uptake calculation), the below ground plant biomass stored 23.2% and the dead above ground plant material accounted for 5.2% of the P mass. The Australian Centre for Tropical Freshwater Research Page 22 The Fate of Phosphorus in Wetlands - A Review, ACTFR Report No 96/15 substrate was typically acidic (pH 4.5) and consisted of approximately 1.5m of highly organic peat soil containing sand and clay which was generally covered by water to a maximum depth of 1m during the year. There was no reported flush or release of P during any seasonal stage of the study. The marsh increased its P retention capacity by virtue of the fact that peat production continued because of the higher plant net production brought about by the constant effluent application. Peatlands are noted for their P retention capacity as a function of the assimilation capacity of both the organic soil and the litter zone. In another Florida based study, DeBusk et al. (1995) utilised two very different methods to ‘polish’ the wastewater: i) a nutrient source for the culture of the wetland macrophyte Pontedaria cordata receiving dairy wastewater with P in excess of 1.7mg P L-1, and ii) periphyton filtration of sugar cane runoff containing less than 100µg P/L. The uptake of P in the P. cordata systems varied from 0.2 to 66.8 mg P m-2 d-1 depending on the age of the plants and season (growth was very reduced in young seedlings and in winter conditions). On the other hand the periphyton system was not limited by the cooler months averaged 101 mg P m-2 d-1 during the winter months (December to May). Gale et al. (1994) utilised an experimental wetland site in Orlando, Florida, to determine the P retention capacity of wetland soils for the treatment of treated wastewater and the likely mechanisms of P removal. During a 21 day retention period, the constructed wetlands (sandy, low organic matter soils) retained 5266% of the added P, compared to the 46-77% retained by the natural wetland (high organic matter soils) from a 8.5mL d-1 inflow which was relatively low in nutrients, i.e. TP values around 0.1mg P L-1 (the control was a natural wetland which only received rainfall and runoff inputs). The organic P pools and the iron-aluminium-bound fractions controlled the P chemistry in both wetlands and anaerobic conditions increased P solubility and availability. The removal followed first order kinetics so that the rate of removal was proportional to its concentration in the inflowing water. The results did not take into account the uptake by plants so the total P removal was expected to be underestimated. Artificial wetlands are considered very valuable for lake protection and restoration in the U.S. (Cooke et al. 1993). Their capacity to reduce inflow rates and improve water quality of the inflowing waters makes them a necessary link within watersheds which incorporate either lakes or reservoirs within their catchment. Unfortunately, the amount of wetland alteration and removal over the last fifty years has resulted in many lakes and reservoirs located in developed regions to develop eutrophic conditions. For example, Holland et al. (1995) stated that over a ten year period approximately 40% of the wetlands identified by the National Wetlands Inventory (NWI) within the metropolitan area of Portland, Oregon, had been lost to human activity or were missing due to drought. Lowe et al. (1992) proposed an interesting idea which will hopefully culminate in the successful restoration of Lake Apopke (Florida) which is hypereutrophic and suffers from algal problems all year round. As the problem is associated with agricultural P runoff, of which the predominant form is particulate P, they proposed to recirculate lake water through constructed wetlands. A combination of this proposed system which will effectively filter out the P and the restriction of P application to the watershed area is predicted to allow the lake to return to normal within sixty years. The use of wetlands to improve inflowing water, however, has not always been successful as reported for Lake McCarrons, Minnesota (Oberts and Osgood 1991). Vincent (1994) reported on the efficiency of a wetland that was built on the fringe of a lake at an artificial beach to treat recreational water in the City of Montreal, Canada. The wetland system occupied 20km2 and consisted of four ponds of varying depth with over 100,000 macrophytes of different species and treated water that was pumped from the swimming area to series of ponds before being pumped back to the lake (Pond 1, 1m depth : Eichhornia crassipes, Hydrocharis morsus-ranae; Pond 2, 0.5m depth : Iris versicolor, Phragmites communis, Scirpus acutus, Typha latifolia; Pond 3, 0.8m depth : Alisma triviale, Elodea canadensis, Myriophyllum spicatum, Pontederia cordata; Pond 4, 1m depth : Elodea canadensis, Potamogeton pectinatus, Vallisneria americana). The annual organic P loading reduced by 82% in 1990 over the course of the ponds, whereas only 38% of FRP and 15% of TP were reduced and the reduction was more significant after Pond 2 where the highest plant density was noted. This reduction was not as marked in 1991 (Organic P, 57%; FRP, 26% and TP 0%) particularly when TP loads in all of the ponds were significantly higher than the inflow (assumed to be due to bank instability in Ponds 1 and 3 and the Australian Centre for Tropical Freshwater Research Page 23 The Fate of Phosphorus in Wetlands - A Review, ACTFR Report No 96/15 presence of large fish). The authors suggested that the greater reduction of organic P was probably associated with epiphytic algal colonies growing on the submerged plants. D’Angelo and Reddy (1994a) reported on the efficiency of a constructed marshland (from previously drained agricultural land) in treating recycled nutrient rich water from the hypereutrophic Lake Apopke, central Florida (see also Lowe et al. 1992 - mentioned previously). Lake water was recycled through the wetlands with a 3-12 day retention period and this eventually resulted in a floc sediment layer forming on the native peat soil surface. The soil-water flux over the first 13 months resulted in high concentrations of water column FRP (31mg P L-1) which reduced to 0.8mg P L-1 after ten months. Microbial degradation and nutrient regeneration from settled labile organic matter supported the nutrient flux to the water column and after the thirteen month study period accounted for 65% of the FRP within the water column. Floc sedimentation was due to the mineralisation of the settled organic matter under anaerobic conditions between the floc sediment and peat soil layers. The microbial process which mediated the decomposition process was supported by the supply of electron acceptors (D’Angelo and Reddy 1994b). The recently deposited, organic rich soils maintained a strong demand for electron acceptors which is presumably due to the microbial catabolic process. As aerobic and anaerobic decomposition by fermentation, SO42reduction and methanogenesis accounted for P and nitrogen resuspension as soluble products (rates of 3.3 to 14mg N L-1 d-1, and 0.5 to 0.6mg P L-1 d-1), these processes therefore depended on the availability of electron acceptors within the wetlands. The nutrients that were made available by this process were then translocated to the water column by diffusive and advective transport. Work in New York state examined specialised substrates to add to wetland sediments to increase the removal capacity of P (Geohring et al. 1995). Various wetland substrates were investigated on the basis of their potential P assimilation capacity, some of which were the waste products of industrial processes, i.e. wollastonite tailings (calcium metasilicate mixed with ferrous metasilicate), iron ore tailings and paper mill waste coal fly ash. The other substrates examined were garnet (ferrous metasilicate), sand with oxidised iron and sand oxidised with aluminium. Wollastonite was shown to be the best adsorbent of P adsorbing 5mg P per gram of substrate in a 5 to 1 solution to substrate mixture and was more effective than sand with iron oxide over a greater pH range. Ninety percent of the influent P was retained in the wollastonite substrate as it was passed through 1m long columns over a retention time of 12 hours. The coal fly ash was a disappointment as it only retained 20% of the P and this was interpreted as being a result of not being stabilised with lime before use. The sand treated with aluminium showed poor retention showing only 21% P removal at pH 4.2, and 0% at pH>6.0. This was found to be due to the aluminium oxide source not being an activated form and hence not very soluble. Sand amended with iron oxide exhibited the highest removal of P (97%) below pH<7.0 buts its effectiveness dramatically decreased at higher pH (in fact P was again released at the high pH values). The examination of grain size and P retention showed that finer grain sizes, particularly the fine wollastonite, increased the capacity for P adsorption particularly at pH 7.0. Asia Recent Asian studies follow the trends of the north American and European regions as to the steady increase in utilising wetland processes for domestic sewage treatment systems. Hosomi et al. (1994) reported on a four year study of a 1224m2 natural wetland on non-cropping farm land (former rice paddy) vegetated with emergent macrophytes (Phragmites and Typha spp.) which treated domestic sewage from a small residential area in Japan. The results were expressed on a seasonal basis as the authors were concerned that microbial activity would decrease in the winter months thereby reducing the effectiveness of the wetland. The average influent flow rate was 38.9m3 d-1 over the four year study period with TP reducing from 0.97mg P L-1 to 0.24mg P L-1 through the wetland system. Despite seasonal variation noted for many study parameters (e.g. BOD, COD, TOC and TN) no clear variation was noted with TP. The wetland removed on average 77% of P from the influent over the study period. Adsorption to the substrate constituted the predominant removal process (24 mg m-2 d-1) of which 8.2 mg m-2 d-1 was assimilated by the plants. The wetland purification rate for P, even in winter, was cited as 0.023 g m-2 d-1. China has practised sewage irrigation since the late 1960’s over millions of hectares and as a result crops, soils and groundwater have been polluted to varying degrees (Xianfa and Chuncai 1995). Since the mid- Australian Centre for Tropical Freshwater Research Page 24 The Fate of Phosphorus in Wetlands - A Review, ACTFR Report No 96/15 1980s many resources have been utilised in upgrading sewage land treatment systems with agricultural application sewage. The predominant wetland system employed in northern China appears to be the reed bed system but some of the examples cited in this paper refer to a combination of rapid infiltration, slow infiltration, overland flow and surface flow wetlands depending on the resources available. The three examples of sewage treatment wetlands and one of a papermill effluent treatment wetland are of reed bed designed to treat approximately 120 to 1800m3 d-1 of municipal sewage (with preliminary settling, a hydraulic load of 1.0 to 6.0 cm d-1 and a retention time of between 1.5 to 10 days) and showed a P removal capacity of 55 to 86%. The findings of studies of these wetland systems have allowed Chinese researchers to develop an empirical model for the design of wetlands. The parameters stipulate a retention time of 7 to 10 days, a hydraulic load of 2 to 20cm d-1, a water depth of <10cm in summer and >30cm in winter, an organic load of 15 to 120kg BOD ha-1 d-1 (less for non-preliminary treated sewage), a rectangular shape with a length to breadth ratio >10 with various water distribution facilities and vegetation dominated by Typha and Phragmites spp. A point worth noting is that most of these wetlands are followed by an aquatic plant pond and fish ponds which have an enormous economic value in that they provide water in water-poor regions as well as fish, waterfowl and other food sources. The effluent quality from these wetlands was found to meet the national standards of fish culture water quality guidelines. The authors’ concern is the problem of availability of area required to implement these systems, although they argue that such integrated systems would prove to be economical. Yang et al. (1995) reported on a four year research program analysing the nutrient and waste removal efficiency of a large (8400m2) wetland in Bainikeng, Shenzen, a subtropical region of China, receiving 3100m3 d-1 of municipal wastewater. The constructed wetland was a subsurface flow system with a series of Phragmites communis beds (loading rate 95.4cm d-1, gravel 80cm deep and 1512m2), attached to two beds separately containing Phragmites communis and Cyperus malaccensis (gravel 100cm deep and 1739m2) which were then attached to three parallel oxygenation ponds (water depth 150cm and 1710m2) containing Nelumbo nutifera and Eichhornia crassipes (which drain finally into two gravel Cyperus malaccensis and one gravel Lepironia articata bed (loading rate 100.7cm d-1, gravel 100cm deep and 2850m2). The removal capacity for P within the subsurface beds was 30.6% for the study period and relatively poor when compared with BOD, COD and TSS removal rates, but this increased to 41% after the oxygenation ponds although there was a marked improvement in wetland removal capacity over the study period (1991, 20% to 1993, 39%). It was concluded that the vegetated beds were therefore quite important in removing the gross pollutants, while the ponds were equally important in nutrient removal. Artificial wastewater with a concentration of up to 40mg P/L was passed through a 10 x 70cm column of mangrove soil from Hong Kong containing 0.028% P (and 10.3mg L-1 FRP) and found to produce an effluent P concentration of less than 0.2mg L-1 consistently over a 55 day period (Tam and Wong 1994) for a variety of wastewater concentrations (10 to 40mg P L-1). Wastewater (100mL) was applied daily (equivalent to 1.3cm d-1) at a rate of 40 to 50mL h-1 and the column conditions mimicked tidal inundation, i.e. an eight hour wet/dry cycle with alternate seawater/wastewater applications. FRP was found to be adsorbed quickly within the first 4cm of the column for each treatment. In a comparative study of domestic raw wastewater treatment in Thailand between a facultative pond and a water spinach pond (Ipomea aquatica) it was found that hydraulic retention times of greater than eight days were required to achieve any significant P removal (Ipomea; 8 days 32.9%, 16 days 90%: facultative pond; 20.1 and 21.4% respectively) with an influent P concentration averaging 1.9mg L-1 (Karnchanawong and Sanjitt 1995). This contrasted with the facultative ponds ability to remove nitrogen which was much better than that by the spinach ponds. The Ipomea tended to die off quickly after high TSS inflow due to a clogging of the root biomass and it was suggested that harvesting would be necessary to ensure continued growth, P uptake and to reduce the potential for accumulating P in the bottom of the ponds. There is an increasing environmental problem in India which generates an estimated 8642 x 106m3 of wastewater per annum of which only 23% is treated at primary level (Juwarkar et al. 1995). The remaining 77% is apparently discharged directly into surface waters or onto the land. Results of a simple study of a constructed wetland in India at Bhubaneshwar consisting of the macrophytes Phragmites carca and Typha latifolia (cement pipes having 0.1256m2 and 0.8m deep filled with 30% soil and 70% sand) Australian Centre for Tropical Freshwater Research Page 25 The Fate of Phosphorus in Wetlands - A Review, ACTFR Report No 96/15 found that between 23 and 48% P was removed from an application rate of 5cm d-1 with a mean inflow P concentration of 14.9mg P L-1. The Phragmites was the preferred plant due to its ability to establish quickly and grow more profusely. New Zealand Wetland research has been conducted in New Zealand over the past few years with an emphasis on treating rural runoff such as dairy farm wastewaters. Tanner (1994) utilised wetlands at the Ruakura Research Farm in Hamilton, to assess the performance of four horizontal channels (9.5 x 2m filled to 0.4m with pre-washed alluvial rhyolitic gravel 10-30mm diameter, 36% effective pore space) and two upflow pilot-scale (1.5m diameter filled to 1m with the same substrate) wetlands planted with Schoenoplectus validus over a twenty month period. The removal of TP was correlated to retention time with the horizontal flow beds showing better results than those produced by the upflow systems, particularly when the loading rates were increased due to the longer holding times and wastewater production during the spring/summer periods. The TP retention varied markedly through both flow systems (-50 to 90% reduction), except in the longer retention period of the horizontal flow (7 days) but showed a gradual decline in effectiveness from 90 to 40% reduction capacity during the study period. It is important to note that a comparison between the two systems was difficult to make due to the difference in substrate contact volume and relative retention periods, but when these factors were corrected for, the results from the upflow system were similar to some of the tested parameters under horizontal flow conditions, i.e. TSS and BOD. Tanner et al. (1995) reported on the removal capacity of alluvial rhyolitic gravel wetlands (19m2) planted with Schoenoplectus validus over varying retention times for the same wetland facility at Hamilton mentioned above. The average TP inflow concentrations were 11.2g m-3 during the study period and resulted in mean loading rates between 0.17 to 0.8g m-2 d-1 of which the FRP comprised 80% of the influent. The rate of P mass removal over the twenty month study period increased from 36 to 74% in the planted wetland as the retention time increased from 2 to 7 days which exceeded the retention capacity for the unplanted wetlands (12 to 37% respectively). The mean annual removal rates of the planted wetlands were 0.13 to 0.32g m-2 d-1 which increased gradually with the mass loading rates. The net storage by the plants accounted for 3 to 60% of the P removal. A natural wetland system in Waitangi forest, Northland, New Zealand, receives sewage effluent after primary treatment in an oxidation pond in the resort town of Paihia which had an approximate population 2000 in winter rising to 8000 in mid-summer. Studies of the character and effectiveness of the wetlands are found in Cooke 1992, Cooke et al. 1992, Cooke 1994, and Nguyen et al. 1995. The wetland area subjected to the sewage inflow was a small proportion of the total area of the natural wetland and as such was subject to different hydrologic conditions and dominating macrophytes (permanently flooded with Typha orientalis and Eleocharis sphacelata compared to the natural wetland areas which have a summer drawdown and were dominated by Baumea-Isachne). The chemistry of the two wetland waters was distinctly different as the natural wetland waters were more pristine and oligotrophic, and had a lower pH. Mass transport studies reported that 30 to 70% of the influent P was removed from inflow waters which were highly loaded with P (approximately 34g P m-2 y-1). Phosphorus was shown to decrease markedly after the sewage discharge waters encountered the natural wetland flow primarily because of the iron content in the flow (present as a soluble complex or colloidal state with humic material). The iron adsorbed and/or flocced out the P as a Fe-P-humic complex due to the increase in pH, and because the sewage-impacted wetland waters were mixed with those originating from a much larger area of natural wetland. As the iron input was relatively unlimited it implied that the P removal was sustainable. The authors suggested that if P removal was the main objective of a constructed or natural wetland treatment system, then a high loading to a small area which also receives drainage from mineralised natural wetlands may be a more cost-effective and sustainable design criterion than the more general approach of lightly loading a large, hydrologically-isolated wetland system with a finite capacity to retain P. The potential for release of P is a constant problem to be considered regardless of wetland design and ultimately relies on factors such as the redox potential of the iron compounds, the concentration of P in the surface sediment and upon the sustainability of iron in the natural wetland waters. Australian Centre for Tropical Freshwater Research Page 26 The Fate of Phosphorus in Wetlands - A Review, ACTFR Report No 96/15 6.3 Critical Evaluation of the Reviewed Journal Articles and Proceedings Papers During the course of the review process it was sometimes difficult to interpret data presented in the articles, particularly proceedings papers. There are often a large number of variables that are not defined which makes it very difficult to draw conclusions. Some examples of these oversights are as follows : • Wetland substrate composition, both chemical and physical, were often not fully described; • Macrophyte densities were often not reported and turnover rates were seldom measured. Standing stock differences between climatic regimes can be difficult to interpret and can be enormous, e.g. in some cases established wetlands containing Phragmites can vary from 600g m-2 to 15000g m-2 (see Table A.1 in the Appendix). These differences may at least be partly due to climate but are more likely the result of other undefined variables. Also in some instances it was not clear whether estimates referred to above ground or whole plant biomass; • Productivity rates also seemed to vary by an order of magnitude for given species of macrophytes between various studies (see Table A.1 in the Appendix) so it is not surprising that there seems to be so much conflict regarding on the role of macrophytes. There was only a few occasions where periphyton, plankton and microbial assimilation were taken into account when determining the total P assimilation; • Seasonal growth varies enormously in temperate climates and reported growth rates in these conditions seemed to adequately address the effect of season upon growth rate. However, there is little data relating to the role season plays in tropical conditions, i.e. what effect does extended dry periods or extended wet periods actually have on macrophytes in tropical climatic conditions?; • It was not always clearly defined whether P values were referring to concentration or mass; and • The P composition of effluent was often not sufficiently described to discriminate potential differences between, for example, agricultural runoff from different soils and cropping methods, stormwater, and primary and secondary treated sewerage effluent produced by different types of aerobic and anaerobic treatment processes. This refers to differences in the relative proportions of various chemical forms of P such as organic particulate, dissolved organic, dissolved inorganic, particulate (sorbed) inorganic, etc. The composition of P is expected to be quite variable and therefore the requirements for final treatment and removal by wetland processes are most likely to vary accordingly. As a consequence, it is very difficult to produce a simple optimum design for wetlands which will effectively remove P and does not waste resources by incorporating unnecessary steps or excessive land area. Given the available data and the increasing research into wetlands (especially in the tropics) it should become easier for managers to implement designs based on a knowledge of general biological and chemical principles. 7. CONSTRUCTED PHOSPHORUS WETLAND CONSIDERATIONS FOR THE REMOVAL OF The P removal capabilities of free surface water emergent macrophyte bed wetlands appears to be somewhat unreliable and is probably influenced by subtle variables that have not yet been fully taken into consideration. It is not possible to discount the possibility that this situation will be improved in the near future but at this time it appears that this type of wetland alone is not the best design for P removal. High diversity natural wetlands appear to be capable of achieving net P reductions and it is possible that with improved understanding of the precise mechanisms involved, constructed wetlands that function similarly may soon be developed. However, at this time the most consistent removal success is being accomplished using alternative or combined approaches and some of these are described below. The following considerations are made on the assumption that discharge data has been fully established such as the hydraulic potential, P loading, available land area, target effluent quality, etc. Australian Centre for Tropical Freshwater Research Page 27 The Fate of Phosphorus in Wetlands - A Review, ACTFR Report No 96/15 Taking into account the variability of results described in the literature, success in removing or assimilating P can be accomplished by utilising several different processes. The fundamental requirements for achieving success in specific wetland construction for this purpose is a complete understanding of the wetland function, allowing for a suitable establishment time and allowing for the ‘self-designing’ capacity of nature (Mitsch and Wilson 1996). Phosphorus can be removed from both the water column and sediments, particularly during the growing season, and on a short term basis, when it can be taken up and stored by wetland vegetation. Interactions within the sediment occur slowly in wetland systems, and P can be readily immobilised by calcium, aluminium and iron by adsorption and precipitation reactions. Fine mineral soils usually have higher concentrations of these ions and therefore have a greater capacity to retain P than organic soils. Options for Wetlands When considering the design of the wetland whose primary function is to retain and transform nutrients, the wetland must be capable of physically detaining the nutrients (Marble 1992). This is best achieved by regulating the inflowing water velocity so that sediments and their adsorbed nutrients settle out. Various types of wetlands and pre-wetland systems may be proposed for the effective removal of P and each are considered below. In all cases, the purpose of the wetland needs to be taken into account as well as the characteristics of the wastewater influent (type, P loading potential and volume). Several wetland systems that can be considered are following : Gross Pollutant Traps This is a very simple component that can be applied immediately before a wetland system or incorporated into the preliminary stage of a wetland. They can be particularly useful if the influent is high in particulate matter (inorganic or high in BOD). The traps reduce flow rates and this subsequently allows time for particulate matter to settle out. Suspended matter or particles are often associated with contaminants, such as trace metals and phosphorus, so by taking steps to reduce the concentration of suspended solids from the inflow, a proportion of the contaminant load should be readily extracted. The only draw back is that the traps may require regular maintenance to clean out the settled particulate matter, and any resulting algal growth, so that they maintain effectiveness and do not promote contaminant pulses, particularly if the traps are deep enough to promote stratification during hot conditions. The design has to be deep enough to reduce flow and promote settlement, and the margins protected or baffled to prevent wind induced turbulence which may re-suspend some of the finer material. Inflow should be regulated to minimise the effects of occasional flood or high volume discharges. As mentioned earlier, stratification could be potentially problematic, especially in tropical conditions, so shading the traps may reduce surface temperature increases as well as reduce the potential for algal growth. Shading will, however, increase the likelihood of anaerobic conditions which would potentially produce malodorous conditions (unless floating macrophytes were used) so a suitable alternative would be to introduce the influent at various points along the bottom of the trap enhancing circulation. Sedimentation would then be restricted to various regions of downwelling. Clear Water Lagoons Clear water lagoon incorporating rooted submerged macrophytes should eliminate a major proportion of FRP from inflowing waters. Apart from the focus of direct plant assimilation of FRP, the additional chemical precipitation of P is expected due to photosynthetic activity of the submerged macrophytes and other associated vegetation (i.e. phytoplankton, algae and periphyton). This process will reduce the level of dissolved carbon dioxide in the water raising the pH levels. Phytoplankton, periphyton and some submerged macrophytes, can utilise bicarbonate as an alternate source of inorganic carbon at these higher pH conditions thereby maintaining the effectiveness of P reduction in these lagoons. The high pH levels associated with high concentrations of calcium (which may be naturally high) will also precipitate P as Australian Centre for Tropical Freshwater Research Page 28 The Fate of Phosphorus in Wetlands - A Review, ACTFR Report No 96/15 calcium-phosphate minerals. Dosing application of calcium and aluminium based compounds can be considered as an additive option in waters naturally low in these minerals to precipitate residual FRP. Investigations of the potential for utilising clear water lagoons with submerged macrophytes in any climatic condition have been limited to date. Gumbricht (1993), however, has suggested that the use of these systems should be more applicable for use as a tertiary treatment step, ‘polishing’ effluent or treating eutrophied natural waters, as pre-treated waters will be less likely to impact upon a more sensitive system (i.e. influent with low turbidity and nutrient loads). Submersed macrophytes species such as Cladophera spp. Enteromorpha spp., Potamogeton spp., Ceratophyllum spp. Myriophyllum spp., Elodea spp. and Egeria sp. are commonly used species and have good potential for P uptake (see Table A.1 in the Appendix). Values for P uptake by these species range from 5.9mg P m-2 d-1 (Ceratophyllum) to 410mg P m-2 d-1 (Egeria densa) and all have the potential to generate large biomass. Productivity rates are likely to increase markedly in warmer tropical conditions so P removal potential is likely to be very significant. It is envisaged that maintenance of these lagoon systems should be minimal once they are fully established and providing influent rates and P loading (as well as the loading of other nutrients) are constant. The ideal use of this system would be to locate it downstream of more rigorous treatment systems so that the predominant form of P is FRP which is present in concentrations between 0.1 and 1mg P L-1. Locating the lagoon downstream will also reduce the potential for problematic inflows if any of the previous systems fail causing loading rates to dramatically increase for any period of time. In the case of treating secondary effluent from a sewage treatment facility, dosing with a flocculation agent to precipitate the majority of FRP (FRP should be the predominant form of P), especially during the establishment stages of the lagoon should be effective in stripping P from the inflow. In warmer climates it is possible that invading species will be a problem (both submersed and free floating species). Free floating species can be very productive and have the capacity to shade submersed species, and any other associated periphyton and phytoplankton, which will reduce the effectiveness of the lagoon, particularly if the lagoon is well established. Depending on the design of the lagoon, harvesting of the free floating species should alleviate this problem. Invading submersed species will be more difficult to eliminate but monitoring the system should allow managers to determine if invading species are supplementing P reduction. The occurrence of algal blooms is a potential problem particularly if the treatment system is relying on regulated inflow. Spikes of high nutrient inflow, due to a system malfunction above the lagoon, is a possibility and this is most likely to lead to blooms of cyanobacteria. Regular monitoring the lagoon system and inflow quality will assist in the prediction of high concentration pulses which should allow managers to shut down the inflow for a period giving the lagoon time to assimilate any increase in load or to flush out the lagoon with an alternate water supply diluting the algal concentrations and nutrient spike. It is recommended that prior to use the lagoon is well established to allow the manager to make these decisions without any detrimental effect on the lagoon system. Banked Grassland Overflow Controlled flow over a banked grassland is a possibility particularly in constant influent treatment systems to further reduce flow and utilise grassland/sediment uptake of dissolved nutrients. This approach was termed a grass filter strip by Higgins et al. (1993) and proved very effective. Flow can be regulated to allow for substrate drying so that particulate and organic P fractions present in the wastewater can be mineralised and therefore become available to the grasses. This can be achieved by having a valve switch at a constant inflow point to split flow to two grassland systems so that as flow to one is shut down for the drying cycle, flow to the other begins for the wetting cycle. Alternatively inflow can be cyclical so that inflow ceases and resumes after set periods of time. The inflow is directed across a ‘level lip spreader’ which is essentially a narrow trench filled with crushed rock the purpose of which is to distribute the wastewater evenly over the width of the grassland bank which is set on a 6° down angle. This form of application essentially reduces channelisation and the erosion of the bank causing a short circuiting of the overland flow. Higgins et al. (1993) utilised drainage Australian Centre for Tropical Freshwater Research Page 29 The Fate of Phosphorus in Wetlands - A Review, ACTFR Report No 96/15 ports underneath the filter strips to promote infiltration of the overflow thereby improving the efficiency of nutrient uptake. The effectiveness of this system will be based on the size of the filter strip, the rate of inflow and loading concentrations, the type of grass selected and the character of the substrate selected. Maintenance should be minimal providing invading grass species are monitored. Periodic harvesting by ‘mowing or bailing’ the strip is recommended (or if the strip is large enough periodic grazing by cattle, etc.) and should be effective at maintaining a monoculture. The selection of grasses would need to take into account a tolerance of wetting and drying cycles and ideally a disposition to ponded pasture conditions. Channelisation is a potential problem with this system particularly in large run off events. The placement of the inlet relative to the location of the lip spreader and grassed bank and inflow control will be essential to minimise the potential for channelisation and ensure sheet flow which is critical for adequate performance of a grass filter (Higgins et al. 1993). Horizontal Subsurface Flow Systems A wetland system allowing for a sub-surface flow input to the wetland dominated by Typha or Phragmites species is recommended mainly due to the potential for mean annual production and standing stock (see Table A.1 in the Appendix). Whenever possible, the wetland substrate can be amended to contain a grain size of predominantly 10 to 30mm with well mixed additives to the gravel base, such as sandy loam or coal fly ash, to ensure adequate and constant sub-surface flow ensuring stable hydraulic conductivity and the supplementary adsorption of P. An adequate depth of substrate should be provided (nominally 0.5m) to promote root and rhizome growth and sufficient substrate/water contact. It is suggested that the width and length of the wetland is designed to accommodate the intended inflow volume and loading of the wastewater, and if necessary allow for harvesting. In the warmer climates high production rates of Typha and Phragmites should prevent any colonisation by invading species but a drawback to the higher production, and depending on the level of nutrient loading, is that these species could become very dense over a short period of time potentially causing flow barriers throughout the wetland. The selection of this form of channelled wetland over surface flow systems, and these particular emergent macrophyte species, for effective phosphorus removal is due to their potential for large above and below ground biomass which ensures adequate phosphorus storage and sufficient below ground surface area for bacterial growth (also a filter for solids) and oxidised micro-environments to promote organic matter decomposition. These plants, particularly Typha spp., are also hardy and the nutrient content in their senescing leaves are readily translocated meaning that the litter will not be a significant source of nutrients (Greenway pers. comm.). Cyperus involucratus may also be chosen as a suitable emergent macrophyte because it grows in dryland and flooded conditions (Hocking 1985) and therefore may be useful where inflow is erratic. Its root system is not as aggressive as Typha or Phragmites and this may make it suitable for planting in channelled systems which do not have an extensive substrate depth. Maintenance of this particular horizontal flow system is not expected to be high once the wetland is established. High productivity rates may require that harvesting is considered and this may improve P removal efficiency. However, in these systems where there is a significant soil/sediment component it is envisaged that harvesting the macrophytes may reduce aerenchymatic oxygen transport and microbial activity and decrease soil conductivity by inducing sediment compaction (Gumbricht 1993). It is therefore suggested that for these wetland systems harvesting is not to be considered a significant requirement for maximum P removal. Plant Composition The plant composition of a wetland designed to assimilate and retain as much P as possible will be a plant or combination of plants that are able to grow quickly and store P efficiently in the region where the wetland is required. Wetland plants function in several ways to reduce P within the water column or Australian Centre for Tropical Freshwater Research Page 30 The Fate of Phosphorus in Wetlands - A Review, ACTFR Report No 96/15 sediment. The stems and leaves of emergent and multi-stemmed plants offer resistance to inflow promoting further settling. Seasonal assimilation from the sediment and water column by biological growth processes also occur which is served best by emergent and aquatic bed vegetation. In Queensland, plants that are capable of best achieving the objective over the majority of climatic conditions are Typha, Schoenoplectus, Triglochin or Phragmites spp. and are used fairly commonly. In the wetland systems described above, the submergent macrophytes for the clear water lagoon could be Myriophyllum spp., Ceratophyllum demersum, Najas tenuifolia, Hydrilla verticillata, Vallisneria spp., Potamogeton spp., Egeria densa or Elodea spp. The introduction of Melalueca quinquinervia around the outer portions of the wetland may also serve to assimilate nutrients and maintain subsurface water uptake thereby minimising the potential for impacts on the localised groundwater. Overall, the selection of macrophyte species should be done on a local basis only, and following a local survey. Some species that may be considered ideal may easily become weeds and difficult to control (see Mockeridge 1995). This includes Typha and Cyperus involucratus as they have the propensity to become nuisance species, as noted in irrigation channels and drains, and within natural wetlands (G. Lukacs pers. comm.). The submerged macrophyte species will require special selection, and careful management, as some are considered noxious weeds (e.g. Elodea spp.). An important consideration when selecting emergent macrophytes is to determine whether the plant or plants have the capability to translocate nutrients from senescing fronds or aboveground biomass. Bald (pers. comm.) stated that Triglochin procerum retains nutrients in the senescing fronds which are then released back into the water column after decomposition. Harvesting the senescing material, or harvesting the plant on a seasonal basis to prevent wide scale die-off, is then an important consideration in macrophyte selection to maintain year round low water column nutrient concentrations but is difficult to carry this out without disturbing the bottom sediment. Planting Patterns Whole submersed plants should be translocated to the clear water lagoon system and it is expected that the establishment of the macrophytes will take a period of time depending on the season, and the quality of water they are planted in. The rooted submerged macrophytes could be planted initially with 0.5m2 and monitored throughout their development. The planting pattern of emergent macrophytes within the wetland systems described above will be such that water contact with the substrate and hence root biomass will be constant. Therefore an establishment period of at least three to four months is necessary for the plants which will need to be planted initially within every 0.5 m2 along extensive stands completely spanning the wetland to achieve the maximum growth potential. The planting pattern will vary depending on the climatic regime but generally, the more dense the vegetation, the greater the ability to remove and assimilate phosphorus. As a rule of thumb, Marble (1992) stated that for emergent macrophytes the water depth should not exceed 50% of the plant height. There is potential to increase the vegetation diversity by planting different forms of macrophytes within the lagoon and the wetland to ensure the greatest potential for nutrient cycling processes. Retention Time The longer the retention time, the better a wetland will be in reducing the P load, providing P uptake is working efficiently. The retention time for the clear water lagoon could be as low as one or two days especially for lagoons that will be receiving influent that has been pre-treated with a flocculating agent to reduce the P load (to less than 1 mg P L-1). This will be entirely dependant on high standing stocks of submersed macrophytes and clear water, the chemistry of the lagoon water and influent loading rates. Australian Centre for Tropical Freshwater Research Page 31 The Fate of Phosphorus in Wetlands - A Review, ACTFR Report No 96/15 A retention time of two weeks is recommended for the sub-surface flow wetland as the minimum requirement for total P removal from the wastewater, the actual timing varying with wetland size. Crites (1994) recommended that significant P removal will require a long detention time (15 to 25 days) and will be enhanced with lower P loading rates (<0.7g P m2 d-1). Australian Centre for Tropical Freshwater Research Page 32 The Fate of Phosphorus in Wetlands - A Review, ACTFR Report No 96/15 Wetland Depths It is suggested that the depth of the gross pollutant trap would be 1m to allow for immediate settling of coarse organic and inorganic matter. The water flow over the grasslands would not be expected to exceed 1cm depth at any one time and regulation of storm flow would be expected to maintain the sheet flow conditions. The level lip spreader would act as a settlement trap to some degree and could effectively operate at approximately 0.5m deep. The depth of the submerged macrophyte pond should be 2m providing an extensive retention time and a large medium for macrophyte growth, including plankton and periphyton communities. Fish and other aquatic wildlife would be introduced to the pond but depending on their density may in fact become a significant source of nutrients. The horizontal sub-surface flow wetland water depth would be expected to be a maximum of 0.5m. This shallow water depth would increase frictional resistance and reduce water flow velocity. The dense annual macrophytic vegetation would be optimal to assist flow resistance. The lagoon and wetland area are systems that would remain constantly saturated. As mentioned earlier the grassed filter strip is designed to be split into two systems so that inflow to each grassland/substrate can be alternated by a flow switch control. This would allow cycling between wet and dry hydroperiods which are important to mineralise organic forms of P. Wetting and Drying Cycles Wetting and drying cycles would be utilised for the grassland filter strip treatment (although it could be a possible design option for the horizontal flow wetland) to enhance mineralisation of P fractions immobilised by the substrate for plant uptake, and to allow time for harvesting or grazing, if necessary. Willett (1982) has shown that flooding and drying has a strong effect on increasing the soil sorption capacity by immobilising added P, especially when soil was rich in organic carbon and reducible iron. The cycle period is expected to be on a two day rotational basis which is within the minimum flooding period suggested by Willet (1982), but is dependant on grass and soil type, and climate. A potential drawback for this type of treatment is the gradual reduction in the availability of soil P for uptake by the grasses, particularly if there is a finite amount of reducible iron in the substrate. Wetland Shape The shape of the horizontal flow wetland has been traditionally rectangular with a length to breadth ratio of approximately 3 or 4 to 1 and a slope of around 2°. Variations to this rectangular shape are sometimes necessary due to land availabity and particularly to avoid channelled shortcircuiting flow ensuring water contact with all of the wetland. It is also recommended that the horizontal flow wetland has a varied depth profile with shallow and deep portions throughout the course of the wetland allowing the water to spread over a wide area and mimic as close as possible natural wetland/marsh structure. The size of the wetland should reflect the nutrient loading of the water entering the system. The multi-depth profile will be provided with vegetative barriers and will assist in reducing wind effects and resuspension problems. The gradient of the inflow prior to wetland entry should be gradual to allow for slow inundation of inflowing waters. The outlet, or series of outlets, at the end of the wetland would be constricted to promote wetland saturation and therefore increased potential for sedimentation, adsorption, biological transformation and nutrient retention. Australian Centre for Tropical Freshwater Research Page 33 The Fate of Phosphorus in Wetlands - A Review, ACTFR Report No 96/15 Substrate Types The substrate best suited for P removal is primarily alluvial, alfisols, ferric, clay or other fine soils. If the natural substrate is to be used, it should be comprehensively analysed for soil type, pH and P retention capacity prior to the wetland construction. It is important that the P retention capacity is determined under both oxic and anoxic conditions because of the effects of redox potential on P sorption. Where the P adsorbing capacity is found to fall below 80% (remembering that the grassland and wetland substrate will be mixed with a gravel base to improve hydraulic conductivity), substrate amendment with a material capable of adequately adsorbing P should be considered. Marble (1992) suggested that where possible fine mineral sediments or soils containing high levels of aluminium and iron should be utilised. The long term P removal capacity of acidic wetland soils would be directly related to the extractable content of aluminium . The depth of the substrate should be at least 0.5m. The substrate should be placed within the wetland area through general excavation of the terrain to the shape required. A sealed surface should then be put in place to prevent groundwater seepage into and wetland water out of the wetland. Flooding Regulated flow will reduce the potential for flooding through the wetland but in the event of increased inflow (i.e. during the instance of plant mechanical breakdown or heavy storm rainfall) it is envisaged that the design of the wetland should allow for periodic inundations of the system. The gentle slope grasslands may be affected by storm inflow with the development of channels within the slopes. If the grass is kept densely vegetated, flow short-circuiting should be avoided. Necessity for Harvesting Harvesting should not be considered a necessity unless emergent or floating macrophyte invasion of the pond is apparent or the macrophyte community within the wetland is so dense that wetland flow is impeded. The grassland slope will need to be harvested (by mowing, bailing or grazing) and this should be conducted prior to, and throughout, the growing season when necessary. 8. MANAGEMENT RECOMMENDATIONS The main management issue is the monitoring of inflow volume to the wetland treatment system. Constant flow would be required to the pond or clear water lagoon which has to be maintained throughout the system. Diurnal and hydraulic loads will, however, vary cyclically so a balancing storage would be needed to achieve this. A balancing storage may not be feasible for a variety of reasons (e.g. limited space, etc.) so the clear water lagoon may have to be designed so that the lagoon could accommodate expected cyclic variations. Discharge from the source would also be expected to vary in contaminant concentration during the year for each STP facility but it would be anticipated that the system should be able to deal with gradual variations of up to 100% for nitrogen and phosphorus loadings. The monitoring of water quality of wastewater inflow and the outflows of each of the treatment stages would be necessary and an adaptation of the monitoring schedule described by Tchobanoglous (1993) is recommended. The schedule is detailed in the Table 1. Australian Centre for Tropical Freshwater Research Page 34 The Fate of Phosphorus in Wetlands - A Review, ACTFR Report No 96/15 Table 1 Summary of suggested monitoring parameters for constructed wetlands Parameter Water Quality Dissolved Oxygen Diurnal Dissolved Oxygen Temperature Conductivity pH BOD TSS Particle Size Distribution Nutrients Chlorophyll a Metals Bacteria (total and faecal coliforms) Trace Organics Biotoxicity Sediments Redox Potential Salinity pH Nutrients Wetland Phase (Pre- or During Construction, or Ongoing) Location (In, Out, Along the profile or Selected locations) Sampling Frequency (Weekly, Monthly, Quarterly, Semi-annually or Annually) O O P,D,O P,D,O P,D,O P,D,O P,D,O P,D,O P,D,O O P,D,O P,D,O I,O,A S I,O,A I,O I,O I,O,A I,O,A I,O,A I,O,A S I,O,A I,O W Q1 W W W W W W W W Q M P,D,O P,D,O I,O,A I,O S S P,D,O P,D P,D,O P,D,O I,O,A I,O,A I,O,A I,O,A Q Q Q Q O O O P,D,O P,D,O S S S S S Q Q Q Q Q O O O I,O,A S S C A S Biota Plankton Invertebrates Fish Birds Mosquitoes Wetland Development Flow Rate Flow Rate Distribution Water Surface Elevations Note : 1 - The frequency of these measurements will vary according to climatic regime. Australian Centre for Tropical Freshwater Research Page 35 The Fate of Phosphorus in Wetlands - A Review, ACTFR Report No 96/15 9. CONCLUSIONS Wetlands are sensitive to nutrient inputs and, although nitrogen is normally lost from the system, phosphorus inputs usually accumulate within the system. Phosphorus enriched waters interact with the vegetation, aquatic biota and wetland substrate. The assimilation of phosphorus into aquatic vegetation is usually short term and decomposition of the detrital matter is usually rapid which may release the P back to the water column. The undecomposed organic P within the litter zone is an integral part of the substrate P pool. Depending on the physico-chemical conditions within the wetland, it may act as a sink or a source which always questions the potential of a wetland to retain P. The pre-requisite of producing a wetland which is effective at removing P from inflowing water is an understanding of the substratewater P interaction and the factors regulating P retention and or release. There is clear evidence from the review that high diversity natural wetlands are capable of effecting long term removal of water column P (albeit that loading rates are relatively low and residence times are high) but limited diversity constructed wetlands, particularly emergent macrophyte monocultures, can prove to be generally unpredictable and/or unreliable. It is therefore envisaged that the system that mimics the natural wetlands systems (with some modification) to the best degree will find the most efficient P removing capability. Phosphorus adsorption is governed by the availability of sorption sites within the substrate and by the presence of amorphous and crystalline forms of iron, aluminium and organic matter. Aluminium is not involved in oxidation-reduction reactions and is therefore not affected by oxic or anoxic conditions but rather is regulated by pH and the presence of organic matter. However, there is a often a strong correlation between redox potental and pH in sediments it is likely that P reactions with aluminium may occur indirectly. Phosphorus interactions with iron can be significant at the substrate-water interface where ferric ions in the oxidised zone can function as a sink for phosphorus diffusing from the water column and from the underlying anaerobic substrate layers. During water flow, suspended particles can sequester available phosphorus from the water column and, because of their large surface area and concentrations of iron and aluminium oxides, will have a high capacity for sorption of FRP. Wetland substrates can be modified if not already capable of retaining phosphorus. Their capacity for retention can also be enhanced by the process of substrate accretion due to the mineral particle deposition during stream flow and the incorporation into the organic pool. The design of a wetland system which will efficiently remove P from inflowing wastewater is dependent on a number of factors such as the consistency of wastewater inflow and the effectiveness of the wetlands themselves, which rely on careful management. The lagoon system is the process that will require the greatest management time but, once established, should be self sustaining. The recommended approach to wetland design for effective P removal is to ensure an adequate development time, and to gradually introduce contaminant loading. The design considerations presented in this review are based on general biological principals that would be best suited to remove P from wastewaters but will require some trialing and review before final design criteria can be developed for each individual case. A combination of these systems will be more expensive than the simple single channel design system that has been incorporated into treatment plants to date, but the overall principles should be applicable for a variety of different climatic conditions. An advantage of selecting macrophytes which are native to a region are that they should be naturally productive within the constructed wetland, capable of storing P, and therefore more beneficial for nutrient uptake. It is very difficult to design a simple system that will optimally assimilate P from any given source of wastewater for a variety of different climatic conditions which exist throughout Queensland. Not enough information is currently available to accurately predict the effectiveness of some of the options considered in this review but they at least provide a starting point for researchers to develop design criteria for each region and each wastewater type. Research is beginning to move in this direction, particularly with respect to Queensland and its tropical areas (by the Queensland Department of Natural Resources, James Cook University of North Queensland, Griffith University and the University of Queensland) so gradually site specific practices should be developed ensuring that where required P is effectively removed form wastewaters prior to their reuse or discharge. Australian Centre for Tropical Freshwater Research Page 36 The Fate of Phosphorus in Wetlands - A Review, ACTFR Report No 96/15 On a cost return basis it is hard to quantify whether P removal by wetland design will be a more realistic option as opposed to chemical precipitation, or pollutant filtration or trapping of source wastewaters, particularly in the long term. However, the literature supports the concept of successful P removal, and the potential for increased effectiveness of P assimilation by macrophytes and substrate in warmer climates, suggest that wetland processes are a valid alternative for P treatment. The best option is a configuration which takes into account several systems in line each having a different process of P removal or assimilation, much like the multi-stage approach alluded to by Brix (1993). More importantly, however, and repeating what has been mentioned before, is that the approach required for any wetland design is going to be based on the source of wastewater, the location of the wetland and the rate and loading which is then applied to the system. Buchberger and Shaw (1995) support the application of model based design approaches to determine the ideal construction criteria for wetlands for wastewater treatment. This scheme should be of some assistance as it can be used to synthesise sub-models which can adequately describe variable loadings, atmospheric moisture and energy fluxes, contaminant fate and transport, and effluent release and recycle rules. As an understanding of fundamental wetland processes, particularly nutrient transformation and transport mechanisms improves, the model can be updated. The flexibility that a wetland model can offer is that the wetland does not have to be built to compare its effectiveness due to different release and recycle loads, and to determine how a wetland’s performance is affected by different treatment mechanisms, wetland operating depths and cell aspect ratios. It is doubtful that P transformation processes are sufficieently well understood for accurate models to be developed at this time but the collection of the emperical data that would be needed is a worthwhile goal. As our knowledge of the fate of P improves, Australian data can be compiled for each wetland system relevant to climatic, substrate, macrophyte type and diversity, etc., which should assist in the design of the most efficient wetland system for P removal. Australian Centre for Tropical Freshwater Research Page 37 The Fate of Phosphorus in Wetlands - A Review, ACTFR Report No 96/15 10. BIBLIOGRAPHY Ansola, G., Fernandez, C. and E. de Luis (1995) Removal of organic matter and nutrients from urban wastewater by using an experimental emergent aquatic macrophyte system. Ecol. Eng. 5: 1319. Aoyama, I. and H. Nishizaki (1993) Uptake of nitrogen and phosphate, and water purification by water hyacinth Eichhornia crassipes (mart.) solms. Wat. Sci. Tech. 28(7): 47-53. Adcock P.W., Ryan, G.L. and P.L. Osborne (1995) Nutrient partitioning in a clay-based surface flow wetland. Wat. Sci. Tech. 32(3): 203-209. Barbier, E.B. (1994) Valuing environmental functions : tropical wetlands. Land Econ. 70(2): 155173. Bavor, H.J. and E.F. Andel (1994) Nutrient removal and disinfection performance in the Byron Bay constructed wetland system. Wat. Sci. Tech. 29(4): 201-208. Bavor, H.J., Roser, D.J. and P.W. Adcock (1995) Challenges for the development of advanced constructed wetlands technology. Wat. Sci. Tech. 32(3): 13-20. Breen, P.F. and A.J. Chick (1995) Rootzone dynamics in constructed wetlands receiving wastewater: a comparison of vertical and horizontal flow systems. Wat. Sci. Tech. 32(3): 281-290. Brix, H. (1993) Wastwater treatment in constructed wetlands: system design, removal processes and treatment performance. In: Constructed Wetlands for Water Quality Improvement. Moshiri, G.A. (Ed). Lewis Publishers, Boca Raton. Pp. 9-22. Brix, H. (1994a) Functions of macrophytes in constructed wetlands. Wat. Sci. Tech. 29(4): 71-78. Brix, H. (1994b) Use of constructed wetlands in water pollution control: Historical development, present status, and future perspectives. Wat. Sci. Tech. 30(8): 209-223. Bolton, K.G.E. and M. Greenway (1995) Growth characteristics and leaf phosphorus concentrations of three Melaleuca species sand cultured indifferent effluent concentrations. n: Proceedings of the National Conference on Wetlands for Water Quality Control at James Cook University of North Queensland. Queensland Department of Primary Industries, Brisbane. Pp. 361-369. Bowmer, K.H. (1987) Nutrient removal from effluent by an artificial wetland: influence of rhizosphere aeration and preferential flow studied using bromide and dye tracers. Wat Res. 21(5): 591-599. Buchberger, S.G. and G.B. Shaw (1995) An approach toward rational design of constructed wetlands for wastewater treatment. Ecol. Eng. 4: 249-275. Chambers, P.A., Prepas, E.E., Bothwell, M.L. and H.R. Hamilton (1989) Roots versus shoots in nutrient uptake by aquatic macrophytes in flowing waters. Can. J. Fish. Aquat. Sci. 46: 435-439. Chambers, J.M., Wrigley, T.J. and A.J. McComb (1993) The potential use of wetlands to reduce P export from agricultural catchments. Fert. Res. 36: 157-164. Australian Centre for Tropical Freshwater Research Page 38 The Fate of Phosphorus in Wetlands - A Review, ACTFR Report No 96/15 Chen, S., Rahman, M., Chabreck, R.H. Jenny, B.F. and R.F. Malone (1995) Constructed wetlands using black will, duckweed and water hyacinth for upgrading dairy lagoon effluent. In: Proceedings from the Versatility of Wetlands in the Agricultural Landscape, Tampa, Florida, September 17-20, 1995. Campbell, K.L. (Ed). American Society of Agricultural Engineers. Pp. 273-282. Cheung, K.C., Venkitachalam, T.H. and W.D. Scott (1994) Selecting soil amendment materials for the removal of P. Wat. Sci. Tech. 30(6): 247-256. Chick A.J. and D.S. Mitchell (1995) A pilot study of vertical flow wetlands at Coffs Harbour, New South Wales, Australia. Wat. Sci. Tech. 32(3): 103-109. Christensen, N., Mitsch, W.J. and S.E. Jorgensen (1994) A first generation ecosystem model of the Des Plaines River experimental wetlands. Ecol. Eng. 3: 495-521. Cooke, J.G. (1992) Phosphorus removal processes in a wetland after a decade of receiving a sewage effluent. J. Environ. Qual. 21: 733-739. Cooke, J.G. (1994) Nutrient transformations in a natural wetland receiving sewage effluent and the implications for waste treatment. Wat. Sci. Tech. 29(4): 209-217. Cooke, J.G., Stub, L. and N. Mora (1992) Fractionation of P in the sediment of a wetland after a decade of receiving sewage effluent. J. Environ. Qual. 21: 726-732. Cooke G.D., Welch, E.B., Peterson, SA. and P.R. Newroth (1993) Restoration and Management of Lakes and Reservoirs. 2nd Ed. Lewis Publishers, Boca Raton. 548pp. Cooper, P. and B. Green (1995) Reed bed treatment systems for sewage treatment in the United Kingdom - the first 10 years’ experience. Wat. Sci. Tech. 32(3): 317-327. Cottingham, P.D. (1995) Nutrient removal from irrigation drainage using wetlands. In: Nutrient Management in Irrigated Agriculture - Research and Implementation. pp. 109-116. Cottingham P.D., Clark, T., Atkins, B.P., Lloyd, L.N. and J.E. Harrison (1994) A review of the effect of nutrients in irrigation drainage on wetland processes. A Water Ecoscience report to LWRRDC as part of the Establishment Report, Milestone No. 1., for the Evaluation of En Route Wetland Systems for Nutrient Removal from Irrigation Drainage. Craft, C.B. and C.J. Richardson (1993) Peat accretion and P accumulation along a eutrophication gradient in the northern Everglades. Biogeochem. 22: 133-156. Crites, R.W. (1994) Design criteria and practice for constructed wetlands. Wat. Sci. Tech. 29(4): 1-6. Cullen, P. (1989) The use of wetlands to filter nutrients. In: Proceedings of the Recycled Water Seminar, Dubbo, September 1989. D’Angelo, E.M. and K.R. Reddy (1994) Diagenesis of organic matter in a wetland receiving hypereutrophic lake water : I. Distribution of dissolved nutrients in the soil and water column. J. Environ. Qual. 23: 928-936. D’Angelo, E.M. and K.R. Reddy (1994) Diagenesis of organic matter in a wetland receiving hypereutrophic lake water : II. Role of inorganic electron acceptors in nutrient release. J. Environ. Qual. 23: 928-936. Australian Centre for Tropical Freshwater Research Page 39 The Fate of Phosphorus in Wetlands - A Review, ACTFR Report No 96/15 Davies, T.H. and P.D. Cottingham (1993) Phosphorus removal from wastewater in a constructed wetland. In: Constructed Wetlands for Water Quality Improvement. Moshiri, G.A.(Ed). Lewis Publishers, Boca Raton. Pp. 315-320. Davies, T.H. and B.T. Hart (1990) Reed bed treatment of wastewaters in a pilot scale facility. In: Constructed Wetlands in Water Pollution Control. Cooper, P.F. and B.C. Findlater (Eds). Permagon Press, Oxford. Pp. 517-520. DeBusk, T.A., Peterson, J.E. and K.R. Jensen (1995) Phosphorus removal from agricultural runoff: an assessment of macrophyte and periphyton-based treatment systems. In: Proceedings from the Versatility of Wetlands in the Agricultural Landscape, Tampa, Florida, September 17-20, 1995. Campbell, K.L. (Ed). American Society of Agricultural Engineers. Pp. 619-626. DeBusk, T.A. and K.R. Reddy (1987) Wastewater treatment using floating aquatic macrophytes: Contaminant removal processes and management strategies. In: Aquatic Plants for Water Treatment and Resource Recovery. Reddy, K.R. and W.H. Smith (Eds). Magnolia Publishing Inc. Orlando, Florida. Pp. 643-656. DeBusk, T.A. and J.H. Ryther (1987) Biomass production and yields of aquatic plants. In: Aquatic Plants for Water Treatment and Resource Recovery. Reddy, K.R. and W.H. Smith (Eds). Magnolia Publishing Inc. Orlando, Florida. Pp. 579-598. Denny, P. (1985) The structure and functioning of African euhydrophyte communities. The floating leaved and submerged vegetation. In: The Ecology and Management of African Wetland Vegetation. Denny, P. (Ed). Dr. W. Junk Publishers, Dordrecht. Pp. 125-152. Devai, I.L., Felfoldy, I.W. and S. Plosz (1988) Detection of phosphine: new aspects of the phosphine cycle in the hydrosphere. Nature. 333: 343-345. Diaz, O.A., Reddy, K.R. and P.A. Moore (in press) Solubility of inorganic P in stream water as influenced by pH and Ca concentrations. Water Res. Dolan, T.J., Bayley, S.E., Zoltek, J. and A.J. Hermann (1981) Phosphorus dynamics of a Florida freshwater marsh receiving treated wastewater. Journal of Applied Ecology. 18: 205-219. Dorioz, J.M. and A. Ferhi (1994) Non-point pollution and management of agricultural areas: P and nitrogen transfer in an agricultural watershed. Wat. Res. 28(2): 395-410. Eighmy, T.T., Jahnke, L.S. and P.L. Bishop (1987) Productivity and photosynthetic characteristics of Elodea nuttallii grown in aquatic treatment systems. In: Aquatic Plants for Water Treatment and Resource Recovery. Reddy, K.R. and W.H. Smith (Eds). Magnolia Publishing Inc. Orlando, Florida. Pp. 453-461. Faulkner, S.P. and C.J. Richardson (1989) Physical and chemical characteristics of freshwater wetland soils. In: Constructed Wetlands for Wastewater Treatment. D.A. Hammer (Ed.). Lewis Publishers, Chelsea, MI, U.S.A. Pp. 41-72. Finlayson, C.M. and A.J. Chick (1983) Testing the potential of aquatic plants to treat abattoir effluent. Wat. Res. 17: 415-422. Finlayson, C.M., Cullen, P., Mitchell, D.S. and A.J. Chick (1986) An assessment of a natural wetland receiving sewage effluent. Aust. J. Ecol. 11: 33-47. Finlayson, C.M., Chick, A.J., van Oertzen, I. and D.S. Mitchell (1987) Treatment of piggery effluent by an aquatic plant filter. Biol. Wastes. 19: 179-196. Australian Centre for Tropical Freshwater Research Page 40 The Fate of Phosphorus in Wetlands - A Review, ACTFR Report No 96/15 Finlayson, C.M., Farrell, T.P. and D.J. Griffiths (1984) Studies of the hydrobiology of a tropical lake in north-western Queensland. III. Growth, chemical composition and potential for harvesting of the aquatic vegetation. Aust. J. Mar. Freshw. Res. 35: 525-536. Gale, P.M., Reddy, K.R. and D.A. Graetz (1994) Phosphorus retention by wetland soils used for treated wastewater disposal. J. Environ. Qual. 23: 370-377. Geohring, L.D., Steehuis, T.S., Corrigan, N., Ries, M., Cohen, M., Cabral, K., Stas, R., De, R. and J.H. Peverly (1995) Specialised substrates for P removal in constructed wetlands. In: Proceedings from the Versatility of Wetlands in the Agricultural Landscape, Tampa, Florida, September 17-20, 1995. Campbell, K.L. (Ed). American Society of Agricultural Engineers. Pp. 607-618. Gehrels, J. and G. Mulamoottil (1989) The transformation and export of P from wetlands. Hydrol. Proc. 3: 365-370. Green, M.B. and J. Upton (1995) Constructed reed beds: appropriate technology for small communities. Wat. Sci. Tech. 32(3): 339-348. Greenway, M. (in press) Nutrient bioaccumulation in wetland plants receiving municipal effluent in constructed wetlands in tropical Australia. In: Proceedings of the 5th International Conference on Wetland Systems for Water Pollution Control. Vienna, Austria. September 1996. Greenway, M. and J.S. Simpson (1996) Artificial wetlands for wastewater treatment, water reuse and wildlife in Queensland, Australia. Wat. Sci. Tech. 33(10-11): 221-229. Gumbricht, T. (1993) Nutrient removal processes in freshwater submersed macrophyte systems. Ecol. Eng. 2: 1-30. Haberl, R., Perfler, R. and H. Mayer (1995) Constructed wetlands in Europe. Wat. Sci. Tech. 32(3): 305-315. Heritage, A., Pistillo, P., Sharma, K.P. and I.R. Lantzke (1995) Treatment of primary-settled urban sewage in pilot-scale vertical flow wetland filters: comparison of four emergent macrophyte species over a 12 month period. Wat. Sci. Tech. 32(3): 295-304. Hey, D.L., Kenimer, A.L. and K.R. Barrett (1994) Water quality improvement by four experimental wetlands. Ecol. Eng. 3: 381-397. Higgins, M.J., Rock, C.A., Bouchard, R. and B. Wengrezynek (1993) Controlling agricultural runoff by use of constructed wetlands. In: Constructed Wetlands for Water Quality Improvement. Moshiri, G.A (Ed). Lewis Publishers, Boca Raton. Pp. 359-367. Hiley, P.D. (1995) The reality of sewage treatment using wetlands. Wat. Sci. Tech. 32(3): 329-338. Hocking P.J. (1985) Responses of Cyperus involucratus Rottb. to nitrogen and phosphorus, with reference to wastewater reclamation. Water Res. 19(11): 1379-1386. Hocking P.J. (1989a) Seasonal dynamics of production, and nutrient accumulation and cycling by Phragmites australis (Cav.) Trin. ex Stuedel in a nutrient-enriched swamp in inland Australia. I. Whole Plants. Aust. J. Mar. Fresh. Res. 40: 421-444. Hocking P.J. (1989b) Seasonal dynamics of production, and nutrient accumulation and cycling by Phragmites australis (Cav.) Trin. ex Stuedel in a nutrient-enriched swamp in inland Australia. II. Individual shoots. Aust. J. Mar. Fresh. Res. 40: 445-453. Australian Centre for Tropical Freshwater Research Page 41 The Fate of Phosphorus in Wetlands - A Review, ACTFR Report No 96/15 Holland, C.C. Honea, J., Gwin, S.E. and M.E. Kentula (1995) Wetland degradation and loss in the rapidly urbanising area of Portland, Oregon. Wetlands. 15(4): 336-345. Hosomi, M., Murakami, A. and R. Sudo (1994) A four-year mass balance for a natural wetland system receiving domestic wastewater. Wat. Sci. Tech. 30(8): 235-244. House, C.H., Broome, S.W. and M.T. Hoover (1994) Treatment of nitrogen and P by a constructed upland-wetland wastewater treatment system. Wat. Sci. Tech. 29(4): 177-184. Howard-Williams, C. (1985) Cycling and retention of nitrogen and P in wetlands: a theoretical and applied perspective. Fresh. Biol. Spec. Rev. 15: 391-431. Howard-Williams, C. and J.J. Gaudet (1985) The structure and functioning of African swamps. In: The Ecology and Management of African Wetland Vegetation. Denny, P. (Ed). Dr. W. Junk Publishers, Dordrecht. Pp. 153-176. Hunter, G. and E. Claus (1995) Preliminary water quality results from a constructed wetland at Plumpton Park, Blacktown, N.S.W. In: Proceedings of the National Conference on Wetlands for Water Quality Control at James Cook University of North Queensland. Queensland Department of Primary Industries, Brisbane. Pp. 265-274. Johengen, T.H. and P.A. LaRock (1993) Quantifying nutrient removal processes within a constructed wetland designed to treat urban stormwater runoff. Ecological Engineering. 2: 347-366. Johnstone, C.A. (1991) Sediment and nutrient retention by freshwater wetlands : effects on surface water quality. Crit. Rev. Env. Control. 21(5,6): 491-565. Jones, R.D. and J.A. Amador (1992) Removal of total P and phosphate by peat soils of the Florida everglades. Can. J. Fish. Aquat. Sci. 49: 577-583. Juwarkar, A.S., Oke, B., Juwarkar, A. and S.M. Patnaik (1995) Domestic wastewater treatment through constructed wetland in India. Wat. Sci. Tech. 32(3): 291-294. Kadlec, R.H. (1994) Phosphorus uptake in Florida marshes. Wat. Sci. Tech. 30(8): 225-234. Kadlec, R.H. (1995) Overview: surface flow constructed wetlands. Wat. Sci. Tech. 32(3): 1-12. Karnchanawong, S. and J. Sanjitt (1995) Comparative study of domestic wastewater treatment efficiencies between facultative pond and water spinach pond. Wat. Sci. Tech. 32(3): 263-270. Kufel, L. and T. Ozimek (1994) Can Chara control P cycling in Lake Luknajno (Poland)?. Hydrobiologia. 275/276: 277-283. Lakshman, G. (1987) Ecotechnological opportunities for aquatic plants - A survey of utilisation options. In: Aquatic Plants for Water Treatment and Resource Recovery. Reddy, K.R. and W.H. Smith (Eds). Magnolia Publishing Inc. Orlando, Florida. Pp. 49-68. Lowe, E.F., Battoe, L.E., Stites, D.L. and M.F. Coveney (1992) Particulate P removal via wetland filtration: an examination of potential for hypereutrophic lake restoration. Environ. Management. 16(1): 67-74. Mann, R.A. (1990) Phosphorus removal by constructed wetlands : substratum adsorption. In: Constructed Wetlands in Water Pollution Control. Cooper, P.F. and B.C. Findlater (Eds). Permagon Press, Oxford. Pp. 97-105. Australian Centre for Tropical Freshwater Research Page 42 The Fate of Phosphorus in Wetlands - A Review, ACTFR Report No 96/15 Mann, R.A. (1995) The use of constructed wetlands for P removal. In: Proceedings of the National Conference on Wetlands for Water Quality Control at James Cook University of North Queensland. Queensland Department of Primary Industries, Brisbane. Pp. 121-128. Mann, R.A. (in press) Phosphorus removal by constructed wetlands. In: Proceedings of the Conference of Downstream Effects of Land Use, University of Central Queensland, Rockhampton, April 1995. Eyles, E.A., Hunter, H. and G.E. Raymont (Eds). Department of Primary Industries, Qld. Mann, R.A. and H.J. Bavor (1993) Phosphorus removal in constructed wetlands using gravel and industrial waste substrata. Wat. Sci. Tech. 27(1): 107-113. Marble, A.D. (1992) A Guide to Wetland Functional Design. Lewis Publishers, Boca Raton. 222pp. McKelvie, I.D., Shan, Y., Benson, R.L. and B.T. Hart (1993) New methods for P analysis. In: Proceedings of the National Workshop on Phosphorus in Australian Freshwaters. LWRRDC Occasional Paper No. 03/93, Canberra. Mitchell, D.S., Chick, A.J. and G.W. Raisin (1995) The use of wetlands for water pollution control in Australia: an ecological perspective. Wat. Sci. Tech. 32(3): 365-373. Mitsch, W.J., Cronk, J.K., Wu, X., Nairn, R.W. and D.L. Hey (1995) Phosphorus retention in constructed freshwater riparian marshes. Ecol. Applic. 5(3): 830-845. Mitsch, W.J. and B.C. Reeder (1991) Modelling retention of a freshwater coastal wetland : estimating the roles of primary productivity, sedimentation, resuspension and hydrology. Ecol. Modelling. 54: 151-187. Mitsch, W.J. and R.F. Wilson (1996) Improving the success of wetland creation and restoration with know-how, time and self design. Ecol. Applic. 6(1): 77-83. Mockeridge, P. (1995) A pilot study of the capacity of wetlands to provide effluent polishing, nutrient removal, and disinfection of sewerage effluent which has been treated to a secondary standard. In: Proceedings of the National Conference on Wetlands for Water Quality Control at James Cook University of North Queensland. Queensland Department of Primary Industries, Brisbane. pp. 395-404. Nguyen, L.M., Cooke, J.G. and G.B. McBride (1995) Phosphorus retention and release characteristics of sewage-impacted wetlands sediments receiving reactants from natural wetlands. In: Proceedings of the National Conference on Wetlands for Water Quality Control at James Cook University of North Queensland. Queensland Department of Primary Industries, Brisbane. pp. 71-89. Oberts, G.L. and R.A. Osgood (1991) Water-quality effectiveness of a detention/wetland treatment system and its effect on an urban lake. Environmental Management. 15(1): 131-138. Patruno, J. and J. Russell (1992) Natural wetland polishing effluent discharging to Wooloweyah Lagoon. . In: Proceedings from Wetlands Downunder, International Specialist Conference on Wetlands Systems in Water Pollution Control, Nov. 30 - Dec. 3, 1992. pp. 34.1-34.9. Poole W.D. and V. Ngo (1992) Lemna systems for water treatment in different climates. In: Proceedings from Wetlands Downunder, International Specialist Conference on Wetlands Systems in Water Pollution Control, Nov. 30 - Dec. 3, 1992. pp. 30.1-30.10. Raisin, G.W. and D.S. Mitchell (1995) The use of wetlands for the control of non-point source pollution. Wat. Sci. Tech. 32(3): 177-186. Australian Centre for Tropical Freshwater Research Page 43 The Fate of Phosphorus in Wetlands - A Review, ACTFR Report No 96/15 Reddy, K.R., Diaz, O.A., Scinto, L.J. and M. Agami (1995) Phosphorus dynamics in selected wetlands and streams of the Lake Okeechobee basin. Ecol. Eng. 5: 183-207. Reddy, K.R. and W.F. DeBusk (1987) Nutrient storage capabilities of aquatic and wetland plants. In: Aquatic Plants for Water Treatment and Resource Recovery. Reddy, K.R. and W.H. Smith (Eds). Magnolia Publishing Inc. Orlando, Florida. Pp. 337-357. Reddy, K.R. and D.L. Sutton (1984) Water hyacinths for water quality improvement and biomass production. J. Environ. Qual. 13: 1-13. Reeder, B.C. (1994) Estimating the role of autotrophs in non-point source P retention in a Laurentian Great Lakes coastal wetland. Ecol. Eng. 3: 161-169. Reuter, J.E., Djohan, T. and C.R. Goldman (1992) The use of wetlands for nutrient removal from surface runoff in a cold climate region of California - results from a newly constructed wetland at Lake Tahoe. Journal of Environmental Management. 36: 35-53. Richardson, C.J. (1985) Mechanisms controlling the retention capacity of freshwater wetlands. Science 228: 1424-1427. Rogers, K.H., Breen, P.F. and A.J. Chick (1990) Hydraulics, root distribution and P removal in experimental wetland systems. In: Constructed Wetlands in Water Pollution Control. Cooper, P.F. and B.C. Findlater (Eds). Permagon Press, Oxford. Pp. 587-590. Ryding, S.O. and C. Forsberg (1977) Sediments as a nutrient source in shallow polluted lakes. In: Interactions Between Sediments and Freshwater. Goltermann, H. (Ed.). Dr. W. Junk. The Hague. Pp. 227-235. Sakadevan, K., Ryan, G., Roser, D., Starrett, J., Bavor, J. and P. Osborne (1995) Phosphorus and nitrogen budgets for five experimental constructed wetland systems. In: Proceedings of the National Conference on Wetlands for Water Quality Control at James Cook University of North Queensland. Queensland Department of Primary Industries, Brisbane. Pp. 101-109. Shardendu and R.S. Ambasht (1991) Relationship of nutrients in water with biomass and nutrient accumulation of submerged macrophytes of a tropical wetland. New Phytol. 117: 493-500. Sharpley, A.N. (1995) Soil P dynamics: agronomic and environmental impacts. Ecol. Eng. 5: 261279. Soukoup, A., Williams, R.J., Cattell, F.C.R. and M.H. Krogh (1994) The function of a coastal wetland as an efficient remover of nutrients from sewage effluent: a case study. Wat. Sci. Tech. 29(4): 295-304. Steiner, G.S. and R.J. Freeman (1989) Configuration and substrate design considerations for constructed wetlands wastewater treatment. In: Constructed Wetlands in Wastewater Treatment. Hammer, D.A. (Ed). Lewis Publishers, Inc, Massachusetts. Pp. 363-377. Tam, N.F.Y. and Y.S. Wong (1994) Nutrient and heavy metal retention in mangrove sediment receiving wastewater. Wat. Sci. Tech. 29(4): 193-200. Tanner, C.C. (1994) Treatment of dairy farm wastewaters in horizontal and up-flow gravel-bed constructed wetlands. Wat. Sci. Tech. 29(4): 85-93. Tanner, C.C., Clayton, J.S. and M.P. Upsdell (1995) Effect of loading rate and planting on treatment of dairy farm wastewaters in constructed wetlands II. Removal of nitrogen and P. Wat. Res. 29(1): 27-34. Australian Centre for Tropical Freshwater Research Page 44 The Fate of Phosphorus in Wetlands - A Review, ACTFR Report No 96/15 Tchobanoglous, G. (1993) Constructed wetlands and aquatic plant systems: research, design, operational and monitoring issues. In: Constructed wetlands for water quality improvement. Moshiri, G.A.(Ed). Lewis Publishers, Boca Raton. Pp. 23-34. Thomas, P.R., Glover, P. and T. Kalaroopan (1995) An evaluation of pollutant removal from secondary treated sewage effluent using a constructed wetland system. Wat. Sci. Tech. 32(3): 87-93. Tripathi, B.D., Srivastava, J. and K. Misra (1991) Nitrogen and Phosphorus removal-capacity of four chosen aquatic macrophytes in tropical freshwater ponds. Environmental Conservation. 18(2): 143-147. Urbanc-Bercic, O. and T. Bulc (1995) Integrated constructed wetlands for small communities. Wat. Sci. Tech. 32(3): 41-47. van Oorschot, R., Tabart, K.J. and M. Sharpin (1995) Recent experience in wetland polishing of an industrial-domestic effluent mixture at Georgetown (Tasmania). In: Proceedings of the National Conference on Wetlands for Water Quality Control. Queensland Department of Primary Industries, Brisbane. pp. 275-283. Vincent, G. (1994) Use of artificial wetlands for the treatment of recreational wastewater. Wat. Sci. Tech. 29(4): 67-70. Vymazal, J. (1995) Constructed wetlands for wastewater treatment in the Czech republic - state of the art. Wat. Sci. Tech. 32(3): 357-364. Vymazal, J. and C.J. Richardson (1995) Species composition, biomass, and nutrient content of periphyton in the Florida Everglades. J. Phycol. 31: 343-354. Willett, I.R. (1982) Phosphorus availability in soils subjected to short periods of flooding and drying. Aust. J. Soil Res. 20: 131-138. Wood, A. (1995) Constructed wetlands in water pollution control: fundamentals to their understanding. Wat. Sci. Tech. 32(3): 21-29. Xianfa, L. and J. Chuncai (1995) Constructed wetland systems for water pollution control in North China. Wat. Sci. Tech. 32(3): 349-356. Yang, Y., Zhencheng, X., Kangping, H., Junsan, W. and W. Guizhi (1995) Removal efficiency of the constructed wetlands wastewater treatment system at Bainikeng, Shenzhen. Wat. Sci. Tech. 32(3): 31-40. Australian Centre for Tropical Freshwater Research Page 45 APPENDIX Table A.1 A list showing commonly used plants in international constructed wetlands and plants considered suitable for the use in Constructed Wetlands in Queensland used for the treatment of waste water (plant list adapted from updated Appendix A of the Interim Guidelines on Planning, Design and Management of Artificial Wetlands in Queensland, DPI (1995)). SCIENTIFIC NAME COMMON NAME, HABIT MEAN ANNUAL PRODUCTION g dry wt/m2/d MEAN ANNUAL STANDING STOCK g dry wt/m2 0.08 (Shardendu and Ambasht 1991) 6 - 27, mean 12 (Shardendu and Ambasht 1991) PHOSPHORUS STORAGE POTENTIAL mg P/m2 PHOSPHORUS CONTENT mg P /g dry wt PHOSPHORUS UPTAKE RATES mg P/m2/d SUBMERGED PLANTS Aponogeton natans* Blyxa spp. Ceratophyllum spp. C. demersum Chara spp. Egeria densa Coontail Hornwort Stonewort Dense waterweed Najas tenuifolia Nitella spp. Potamogeton perfioliatus P. pectinatus P. crispus P. scweinfurthii* Triglochin procera Utricularia gibba subsp. exoleta Vallisneria gigantea Waterweed Eriocaulon Hydrilla Water nymph Stonewort Clasped pondweed Sago pondweed Curly pondweed 5.9 (in Brix 1994) 6 13.3, max 14 (Greenway in press) 2.5 (in DeBusk and Ryther 1987) 850 (Denny 1985) 38 - 48 (Reddy and DeBusk 1987) 202 - 410 (Reddy and DeBusk 1987) 20 (in Gumbricht 1993) 1 3.6 - 12.9 (in Gumbricht 1993) Elodea spp. E. densa* E. nuttallii* Eriocaulon setaceum Hydrilla verticillata 700 - 1000 (Lakshman 1987) 800 (Denny 1985) 1300 - 4700 (Lakshman 1987) 4.6 - 23 (in Gumbricht 1993) 6 2.8 - 12.9 (in DeBusk and Ryther 1987) 0.0 - 4.5 (Eighmy et al. 1987) 11.0 - 23.0 (Eighmy et al. 1987) - dry wt 4.2 - 10.4 (in DeBusk and Ryther 1987) 0.13 (Shardendu and Ambasht 1991) 40 - 2970 (Finlayson et al. 1980) 25 - 71, mean 54 (Shardendu and Ambasht 1991) 0.9 - 4.1 (Finlayson et al. 1980) 1.1 (Shardendu and Ambasht 1991) 0.19 (Shardendu and Ambasht 1991) 6.8 (Denny 1985) 25 - 69, mean 42 (Shardendu and Ambasht 1991) 1200 (Denny 1985) 0.10 (Shardendu and Ambasht 1991) 0.14 (Denny 1985) 60 - 6410 (Finlayson et al. 1980) 5 - 35, mean 22 (Shardendu and Ambasht 1991) 40 -70 (in Denny 1995) 60 (Denny 1985) 1.0 (Shardendu and Ambasht 1991) 1.0 (in Denny 1985) 4.0 (Greenway in press) 1.8 - 3.9 (Chambers et al. 1989) 1.1 - 3.8 (Finlayson et al. 1980) 3.3 - 11.0 (Denny 1985) 2 23.4 (in Brix 1994) Water ribbons Yellow bladderwort Ribbon weed FLOATING PLANTS Azolla spp. A. pinnata Eichhornia crassipes* Azolla Water fern Water hyacinth* 2.9 - 7.9 (in DeBusk and Ryther 1987) 10 - 33 (Reddy and DeBusk 1987) 128 - 135 (Reddy and DeBusk 1987) 1 25 - 2200 (Lakshman 1987) 2000 - 2400 (Reddy and DeBusk 1987) 460 - 21200 (Lakshman 1987) 2 6 10.0 (DeBusk and Reddy 1987) 24.2 - 64.4 (DeBusk and Ryther 1987) 6 6 - 18 (Reddy and DeBusk 1987) 1.4 - 12 (Reddy and DeBusk 1987) 1.4 - 12 (in Gumbricht 1993) 50 - 240 (in DeBusk and Ryther 1987) 49 - 243 (Reddy and DeBusk 1987) 1 SCIENTIFIC NAME COMMON NAME, HABIT Hydrocotyle spp.* H. umbellata* Lemna spp. Pistia spp. P. stratiotes* Penny wort Duckweed Water lettuce Ricciocarpus natans Salvinia spp. Liverwort Water fern S. molesta* S. rotundifolia* MEAN ANNUAL PRODUCTION g dry wt/m2/d MEAN ANNUAL STANDING STOCK g dry wt/m2 PHOSPHORUS STORAGE POTENTIAL mg P/m2 16.4 -30.1(Reddy and DeBusk 1987) 16.4 - 30.1 (in Gumbricht 1993) 4.1 - 12.1 (Denny 1985) 8.2 - 16.4 (in Gumbricht 1993) 2000 - 2400 (in Gumbricht 1993) 5.7 (Aoyama and Nishizaki 1993) 1.4 - 8.0 (Johnston 1991) 700 - 1100 (in Gumbricht 1993) 10.3 - 29.7 (in DeBusk and Ryther 1987) 8.2 - 16.4 (Reddy and DeBusk 1987) 700 - 1100 (Reddy and DeBusk 1987) 2.0 - 13 (in Gumbricht 1993) 4.3 (Reddy et al. 1995) 2.0 - 12.5 (Reddy and DeBusk 1987) 2.3 - 7.5 (Reddy and DeBusk 1987) PHOSPHORUS CONTENT mg P /g dry wt 252 - 371 (Reddy and DeBusk 1987) 96 - 308 (Reddy and DeBusk 1987) 93.7 (in Brix 1994) 2 80 - 90 (DeBusk and Reddy 1987) 81 - 86 (Reddy and DeBusk 1987) 240 - 265 (Reddy and DeBusk 1987) 36 - 211 (Reddy and DeBusk 1987) 18 - 87 (Reddy and DeBusk 1987) 205 - 234 (Reddy and DeBusk 1987) 32 - 110 (Reddy and DeBusk 1987) 1 3.8 - 12.0 (in DeBusk and Ryther 1987) 1.6 - 7.1 (Reddy and DeBusk 1987) 1.6 - 7.1 (in Gumbricht 1993) 130 (Reddy and DeBusk 1987) 30 - 9000 (Lakshman 1987) 130 - 350 (in Gumbricht 1993) 13.7 - 21.9 (in Gumbricht 1993) 14.2 - 40.0 (in DeBusk and Ryther 1987) 13.7 - 21.9 (Reddy and DeBusk 1987) 600 - 1100 (in Gumbricht 1993) 600 - 1050 (Reddy and DeBusk 1987) 460 - 14600 (Lakshman 1987) 0.1 - 1.6 (Reddy and DeBusk 1987) 6 2 - 5.7 (Reddy and DeBusk 1987) 6 4000 - 6700 (Lakshman 1987) 200 - 300 (in Gumbricht 1993) Water fern 2.5 - 12.3 (Reddy and DeBusk 1987) 2.5 - 12.3 (in Gumbricht 1993) 9.4 (Finlayson et al. 1982) Water fern 6.4 - 13.9 (in DeBusk and Ryther 1987) 240 - 320 (Reddy and DeBusk 1987) 4.0 - 15 (Reddy and DeBusk 1987) 2 - 20 (in Gumbricht 1993) 10.9, max 18.4 (Greenway in press) 7.5 (Johnston 1991) 1.5 - 12 (in Gumbricht 1993) 1.5 - 11.5 (Reddy and DeBusk 1987) 7.3 (Greenway in press) 2 1 70 - 220 (DeBusk and Reddy 1987) 72 - 218 (Reddy and DeBusk 1987) 202 - 297 (Reddy and DeBusk 1987) 82 - 301 (Reddy and DeBusk 1987) 23.4 (in Brix 1994) 2 1 2 2 - 9 (in Gumbricht 1993) 6 0.4 - 2.4 (Reddy and DeBusk 1987) 1.4 - 4.7 (Finlayson et al. 1980) 5.7 (Greenway in press) 1.8 - 9.0 (Reddy and DeBusk 1987) 2.1 - 7.4 (Johnston 1991) 3.4 - 5.9 (in DeBusk and Ryther 1987) Spirodela polyrhiza* PHOSPHORUS UPTAKE RATES mg P/m2/d 32 - 105 (Reddy and DeBusk 1987) 203 - 217 (Reddy and DeBusk 1987) 25.2 - 123 (Reddy and DeBusk 1987) 34 (Reddy and DeBusk 1987) 139 - 248 (Reddy and DeBusk 1987) 1 2 1 2 Wolffia spp. Duckweed EMERGENT PLANTS Alternantha philoxeroides* Baumea teretifolia B. articulata B. rubiginosa Blechnum camfieldii B. indicum Bolboschoenus fluviatilis Brachiaria mutica Caldesia oligococca 3.9 (Johnston 1991) Sedge Sedge Sedge Fern Swamp water fern, Creeping fern Sedge Para grass Caldesia 2.4 - 3.7, max 8.7 (Greenway in press) SCIENTIFIC NAME Carex spp. C. fascicularis Colocasia esculenta* Cyperus spp. C. platystylis C. difformis C. involucratus* C. papyrus COMMON NAME, HABIT Sedge Tassel sedge, Sedge Wetland taro MEAN ANNUAL PRODUCTION g dry wt/m2/d MEAN ANNUAL STANDING STOCK g dry wt/m2 370 - 3400 (Lakshman 1987) PHOSPHORUS STORAGE POTENTIAL mg P/m2 PHOSPHORUS CONTENT mg P /g dry wt PHOSPHORUS UPTAKE RATES mg P/m2/d 1.0 - 4.5 (Johnston 1991) 5 5.2 (in DeBusk and Ryther 1987) 2.5 - 12 (Reddy and DeBusk 1987) 0.2 - 6.3 (in Reddy and DeBusk 1987) 3 Sedge Rice sedge, Sedge 2200 Papyrus, Sedge (Hocking 1985) - above ground 8.8 (Hocking 1985) - leaves and shoots 7.0 (Greenway in press) 5.2 - 26.5 (Hocking 1985) 22.9 (in DeBusk and Ryther 1987) 13.2 - 39.2 (Howard-Williams and Gaudet 15.2 (in Brix 1994) 1985) 34.2 (Denny 1985) C. unioloides Damasonium minus Eleocharis spp. Sedge Starfruit Spike rush 7.0 (Reddy and DeBusk 1987) 1.0 - 3.0 (Reddy and DeBusk 1987) 1.3 (Johnston 1991) E. phillipinensis E. ochrostachys Fuirena umbellata Hibbertia salisifolia Ipomea aquatica* I. diamantinensis* Isolepis inundata Emergent, Herbaceous Emergent, Shrub Water spinach 8.3 (in DeBusk and Ryther 1987) max 9.5 (Greenway in press) max 9.9 (Greenway in press) 14.6 (in DeBusk and Ryther 1987) 0.5 - 3.0 (Johnston 1991) Juncus effusus* Lepidosperma longitudinale Lepironia articulata Ludwigia peploides subsp. montevidensis Marsilea mutica M. drummondii Monochoria cyanea Myriophyllum spp. M. variifolium M. verrucosum Napar advena* Nelumbo lutea* N. nucifera Nymphaea spp. Nymphoides crenata Swamp club rush, Sedge Pithy swordsedge, Sedge Sedge Water primrose Nardoo Nardoo Monochoria Watermilfoil 4.1 - 5.4, max 9.9 (Greenway in press) 10 7.0, max 9.8 (Greenway in press) 1.4 - 24.7 (in Gumbricht 1993) 120 - 1150 (Lakshman 1987) 50 - 900 (in Gumbricht 1993) 40 - 50 (Denny 1985) 6 7.7, max 13 (Greenway in press) 13 - 27 (in Gumbricht 1993) 120 - 330 (in Gumbricht 1993) Red water milfoil 5.7 (in DeBusk and Ryther 1987) 10 - 160 (Reeder 1994) Lotus Water lily Wavy marshwort 50 - 128 (Denny 1985) 2.2 - 6.1, mean 3.4 (Reeder 1994) 0.6 - 0.9 actual, 27 - 41 theoretical (Reeder 1994) SCIENTIFIC NAME N. indica Ottelia ovalifolia Panicum sp.* Persicaria decipiens Philydrum lanuginosum Polygonum spp.* Pontedaria spp.* P. cordata* Potamogeton spp. P. tricarinatus P. javanicus Rorippa nasturtiumaquaticum Sagittaria spp. Sagittaria graminea Schoenoplectus litoralis S. mucronatus S. validus Schoenus brevifolius Sphaerolobium vimineum Triglohin striata Typha spp. COMMON NAME, HABIT MEAN ANNUAL PRODUCTION g dry wt/m2/d MEAN ANNUAL STANDING STOCK g dry wt/m2 PHOSPHORUS STORAGE POTENTIAL mg P/m2 Water snowflake Swamp lily PHOSPHORUS CONTENT mg P /g dry wt 8.4, max 16.6 (Greenway in press) 0.9 - 2.9 (Reddy et al. 1995) Slender knotweed Frogsmouth 0.1 - 24.1 0.8 - 2.3 (Reddy et al. 1995) 0.5 - 0.9 (Reddy et al. 1995) 2.3 - 7.4 (DeBusk et al. 1995) (DeBusk et al. 1995) Pondweed Floating pondweed Pondweed Water cress 200 - 930 (Lakshman 1987) 0.2 - 66.8 (DeBusk et al. 1995) 17 - 23 (Kirby and Albers in press) 6 0.5 - 4.5 (Johnston 1991) Sagittaria Sedge Sedge Sedge streaked arrow grass Cattail T. orientalis Cumbungi T. domingensis T. latifolia* Cumbungi 0.8 - 4.0 (Greenway in press) 1800 - 3600 (Tanner et al. 1995) 2.2 - 16.7 (Reddy and DeBusk 1987) 2.2 - 16.7 (in Gumbricht 1993) 4.4 - 8.2 (Howard-Williams and Gaudet 1985) 12.3 - 19.2 (Denny 1985) 970 - 10140 (Lakshman 1987) 430 - 2250 (in Gumbricht 1993) 700 - 3600 (Hosomi et al. 1994) 5 4.5 - 37.5 (Reddy and DeBusk 1987) 690 - 1690 (Cooke 1992) 52.6 (in DeBusk and Ryther 1987) 40.0, mean 17.7 (in DeBusk and Ryther 1987) 11.8 (in Johnston 1991) 6.0 - 10.7 (Tanner et al. 1995) 10 0.5 - 4.2 (Reddy and DeBusk 1987) 1 - 4 (in Gumbricht 1993) 1.6 - 2.4 (Hosomi et al. 1994) 3.2 (Greenway in press) 6.5 - 9 (Breen 1990) 2.0 - 4.4 (Cooke 1992) 1.5 - 3.2, max 7.2 (Greenway in press) 0.5 - 5.3 (Johnston 1991) 21 - 110 (Reddy and DeBusk 1987) 8.2 (Hosomi et al. 1994) 1.7 - 17 (Mitsch et al. 1995) 10 3.6 (in Johnston 1991) 48.0 (in Brix 1994) Villarsia exaltata Herbaceous Plants Cyperus procerus C. trnervis C. pilosus C. exaltatus C. haspan C. lucidus Eleocharis acuta* E. sphacelata PHOSPHORUS UPTAKE RATES mg P/m2/d Sedge Sedge Sedge Tall flat sedge, Sedge Sedge Sedge 1.2 - 4.5 (Greenway in press) 1.3 - 3.9, max 9.4 (Greenway in press) 10 Tall spikerush 10 SCIENTIFIC NAME E. spiralis E. cylindrostachys E. dulcis E. equisetina Fuirena ciliaris Juncus spp. J. prismatocarpus J. polyanthemos J. continuus Leptocarpus tenax Rumex brownii Schoenus apogon Smithia sensitiva COMMON NAME, HABIT Emergent Rush MEAN ANNUAL PRODUCTION g dry wt/m2/d MEAN ANNUAL STANDING STOCK g dry wt/m2 PHOSPHORUS STORAGE POTENTIAL mg P/m2 PHOSPHORUS CONTENT 14.6 (Reddy and DeBusk 1987) 14.5 (in Gumbricht 1993) 9700 - 14000 (Lakshman 1987) 220 (in Gumbricht 1993) 4.0 (Reddy and DeBusk 1987) 2 (Reddy and DeBusk 1987) 2 (in Gumbricht 1993) 0.1 - 0.7 (Reddy et al. 1995) 5 mg P /g dry wt PHOSPHORUS UPTAKE RATES mg P/m2/d 30.1 (Reddy and DeBusk 1987) Branching rush, Sedge Sedge Sedge Sedge Swamp dock Fluke bogrush Grasses Diplachne fusca Echinochloa inundata E. colona E. crus-galli Hymenachne acutigluma* Oryza meridionalis* Paspalum distichum Pennisetum alopecuroides Phragmites spp.* P. australis P. communis* Pseudoraphis spinescens Xanthorrhoea fulva Brown beetle grass Marsh millet Awnless barnyard grass Barnyard grass Water couch Swamp foxtail Reed Common reed Common reed Spiny mudgrass Swamp grass tree Shrubs Acrostichum speciosum Baeckea stenophylla B. diosmifolia Banksia robur 3.0 (Finlayson 1991) 1.8 (Finlayson 1991) 3.4 (Greenway in press) 5.7 (Finlayson 1991) 1.4 (Finlayson 1991) Mangrove fern, Fern Weeping baeckea Fringed baekea Broad leaved banksia 5 5 2.7 - 16.4 (Reddy and DeBusk 1987) 2.7 - 16.4 (in Gumbricht 1993) 12.3 - 19.2 (Denny 1985) 600 - 3700 (in Gumbricht 1993) 1500 - 3000 (Hosomi et al. 1994) 25.9 (in DeBusk and Ryther 1987) 26.1 (Hocking 1989) 9960 (14945) (Hocking 1989) 2.0 - 3.0 (Reddy and DeBusk 1987) 2 - 3 (in Gumbricht 1993) 1.7 - leaves (Hocking 1989b) 1.7 - 3.8 (Hosomi et al. 1994) 6.1 - 9.4 (Hocking 1989a) 1.4 - 2.0 (Greenway in press) 1.8 (Johnston 1991) 1.3 (Finlayson 1991) 4 1 820 - 12290 (Lakshman 1987) 5.2 (Finlayson 1991) 1.4 - 5.3 (Reddy and DeBusk 1987) 5 5 9.6 (Reddy and DeBusk 1987) 8.2 (Hosomi et al. 1994) 12.2 - 45 (Hocking 1989a) 45.7 (in Brix 1994) SCIENTIFIC NAME Boronia falcifolia B. parviflora Callistemon pachyphyllus Comesperma defoliatum Gahnia sieberiana Hibbertia salicifolia Leptospermum liversidgei L. semibaccatum L. juniperinum Melastoma affine Pultenaea paleacea Todea barbara COMMON NAME, HABIT MEAN ANNUAL PRODUCTION g dry wt/m2/d MEAN ANNUAL STANDING STOCK g dry wt/m2 PHOSPHORUS STORAGE POTENTIAL mg P/m2 PHOSPHORUS CONTENT mg P /g dry wt PHOSPHORUS UPTAKE RATES mg P/m2/d Swamp boronia Wallum bottlebrush Red fruited sawsedge Emergent shrub Blue tongue Southern king fern, Fern Viminaria juncea Trees Callistemon viminalis Eucalyptus robusta Glochidion ferdinandi Hibiscus diversifolius Lophostemon suaveolens Melaleuca alternifolia* M. quinquenervia Swamp messmate Cheese tree Swamp hibiscus Swamp mahogany 11 - leaves (Bolton and Greenway 1995) 1.5 - litter (Greenway, in press) Paperbarked tea tree 34 - 1061 Above ground 1460 - 4580 Below ground Combination of : Pontedaria, Saggitaria, Panicum and Hibiscus (in Dolan et al. 1981) 7 5 - 4.5 Above ground 1.3 - 4.6 Below ground (in Dolan et al. 1981) 0.6 - 29.5 (in Dolan et al. 1981) 7 PERIPHYTON Periphyton and Phytoplankton Phytoplankton 10 - 35, mean 21.2 (DeBusk et al. 1995) 3.7 (DeBusk et al. 1995) 0.6 - 0.9 (in Mitsch et al. 1995) 79 - 101 (DeBusk et al. 1995) 27.4 (Reeder 1994) 8 7 Notes : * - indicates a plant not selected by the QDPI as an appropriate wetland plant species. 1 - calculated using the growth rate in the linear phase of the plant’s growth and the average tissue phosphorus content. This represents phosphorus removal due to the plant alone. 2 - calculated using the phosphorus concentrations in the water and represents phosphorus removal due to the plant uptake and nutrient transformations. 3 - values represent predominantly above ground storages. 4 - Concentration includes above ground and below ground biomass (roots, young and mature rhizomes, and live, young and dead shoots) 5 - Represents above and below ground biomass production. 6 - Represents above ground biomass production only. 7 - Range of values between a control plot and a plot receiving high effluent waters. 8 - 79 mg P/m2/d was based on algal biomass only, the 101 mg P/m2/d included phosphorus adsorption to water borne particles. 9 - Whole plant production by the minimum-maximum method with the corrected estimate value in parentheses. 10 - Range includes nutrient accumulation in selected native Australian wetland species from constructed and natural wetlands (lowest phosphorus values in the constructed wetland species).
© Copyright 2026 Paperzz