Predicting the long-term performance of low-cost phosphorus-sorbing materials using small-scale adsorption column experiments

Anthropogenic phosphorus (P) and nitrogen (N) inputs to aquatic environments have rapidly increased in magnitude since the middle of the twentieth century, and nutrient pollution has become the primary threat to global surface water quality. Because P is most often the nutrient limiting primary production in freshwater ecosystems, the control of P losses from anthropogenic sources is of paramount importance to the prevention of cultural eutrophication, i.e. the nutrient enrichment of water bodies as a result of human activities. Left unchecked, eutrophication causes optimal conditions for sudden, dramatic increases in populations of algae and many species of noxious aquatic weeds. The long-term effects of this are marked reductions of aquatic biodiversity, disruption of natural biogeochemical balances, and an accelerated ageing of water bodies resulting in their premature succession from clear water, to swamp/marsh, and finally to dry land. More obvious short-term effects are increased incidences of harmful algal blooms and associated fish kills resulting from low dissolved oxygen levels, i.e. hypoxic conditions, caused by the microbial decay of organic matter originating from the overgrowth of plants and algae. Sources of P losses can be broadly divided into two main categories: point sources and nonpoint sources. The former represents easily identifiable, discreet sources of P such as outlets from municipal wastewater treatment plants, while the latter represents diffuse sources of P which are spatially and temporally heterogeneous. Advances in wastewater treatment technologies, coupled with increasingly strict legislative controls in recent decades, have lessened the threat posed by point source P losses, and nonpoint source P losses have become the dominant contributor to nutrient pollution in many regions, including Ireland, where nonpoint sources account for approximately 57% of P inputs to surface waters. Mitigation of nonpoint P losses is currently achieved through the implementation of source reduction strategies, such as matching fertilizer additions with crop deficiencies, and source interception strategies, such as trapping and retaining P-laden particles in runoff with the use of buffer zones. These best management practices (BMPs) can often fail to produce desired results, and sometimes their implementation is not practicable as a result of unfavourable onsite conditions. An example of such a case is the loss of dissolved P from clearfelling on blanket peat forestry plantations. These P losses are unamenable to mitigation with standard BMPs such as the use of buffer zones, and source control is not feasible either, as these P losses primarily originate from harvesting residues utilised to form brash mats. These mats provide support to heavy harvesting machinery, and their use is therefore an integral part of the process of clearfelling on peat soils. One solution in cases such as these may be to use low-cost P sorbing materials, and there has been much research in recent years attempting to identify and characterize novel sorbents for use in the treatment of various wastewaters. Such investigations more often than not use batch adsorption tests as a means of assessing the potential of such media, and, although these are an essential first step, there has been an overreliance on such methods which, when performed alone, provide insufficient information to predict media longevity. The purpose of this study was to develop a simple, economical, and rapid method by which to assess the potential in-field longevity of low-cost P sorbing materials, and then apply this method to identify media which might be suitable for use in pilot-scale in-filed filters to prevent P losses associated with peatland forestry harvesting. After an initial screening process, using batch adsorption methods to identify potentially suitable materials, small-scale column adsorption tests were carried out using four media: aluminium water treatment residual (Al-WTR), ferric water treatment residual (Fe-WTR), and two grades of crushed concrete. Long-term, large-scale adsorption column studies were also performed using these media (excluding Fe-WTR, which performed poorly in the small-scale tests), and it was demonstrated that the behaviour of these large-scale filters could be predicted accurately based on the performances of small-scale columns subjected to equivalent loadings. A mathematical modelling approach was then developed which would enable researchers to use the results of the small-scale column tests to predict the propagation of saturation and pore concentration fronts within large-scale filters. With these modelling techniques, the performance of large-scale filters (of any size) subjected to any loading could be predicted using results from a number of small-scale column tests to determine the necessary model coefficients. This significantly improved the utility of small-scale tests, as it enabled performance predictions to be made for any possible large-scale filter configuration using results from small-scale column tests. The experimental and modelling methodologies developed over the course of this study were applied to assess the suitability of crushed concrete and Al-WTR for the treatment of peatland forestry runoff. The results of these investigations suggest that Al-WTR could successfully be used to remove P from this runoff, and there were no issues which would preclude its usage in a peatland forestry environment; the media showed no significant release of metals (total cumulative releases of nickel, copper, and aluminium were 0.16, 5.5, and 13.63 µg g-1 filter media, respectively) and the pH of effluent from Al-WTR filters was well within the EPA recommended environmental quality standard (EQS) of 6 to 9 (average effluent pH was 7.31±0.36). Crushed concrete was even more effective than Al-WTR at removing P from forestry runoff, due to a higher P adsorption affinity at low concentrations. However, the high pH of effluent from filters containing this media indicated that significant pre-washing (≥ 240 bed volumes) of the filter media would be required to bring the filter effluent below the recommended EQS of 9, were it to be used in a forestry context. Crushed concrete also released much greater quantities of Cu and Al (31.90 and 96.17 µg g-1 filter media, respectively), further highlighting the need for washing of the filter media prior to use. For these reasons, it was decided that Al-WTR shows greater promise, and it was concluded that its use for the treatment of peatland forestry runoff should be studied further in pilot-scale in-field filters. Al-WTR was also found to be effective in removing P from dairy soiled water, a wastewater which had a P concentration that was approximately 30 times that of forestry runoff; this suggests that Al-WTR holds great promise for the removal of P from point sources as well as nonpoint sources, and further research should be carried out on this topic.

Funder

Irish Research Council

Rights

Attribution-NonCommercial-NoDerivs 3.0 Ireland

cbn

Collections

University of Galway Theses (PhD Theses)