Abstract:

I-landers (Belgium) is confronted with reactive phosphorus concentrations in streams and
lakes which are three to four times higher than the 0.1 mg.I. ; environment limit set by the
Water f ramework Directive. Much of the excessive P input in surface walers is derived from
agriculture. Specifically, direct P input from artificial!} installed subsoil drainage pipes which
shun-circuits the buffering capacil} of the subsoil is suspected to be one of the major sources,
hi this stud}, we aim to develop simple and cheap fillers that can be directly installed in the
field to reduce P concentration from farm drain water. To achieve this, our specific objectives
included; (1) determining the link between physical size distribution, bulk densil} and initial
saturated hydraulic conductivity (Ksal) of tiller materials. (2) evaluating P removal eflkienc}
and durability of the filler materials. (3) determining the cumulative P sorption until
saturation point and subsequently test P desorption efficiency and (4) assessing the
performance of the fillers under a pilot field trial. I lie filter materials used were made of iron
coaled sand and mixtures of iron coaled sand w ith glauconite. We determined Ksal using the
constant head method and Darcy's equation, particle size distribution by sieving the materials
through sieves of 0-5 mm mesh size and bulk density by measuring dry mass per volume.
I lie method of Murph}' and Kiley was employed to analyse P using a spectrophotometer. We
found that Ksal is linked to particle size distribution and bulk density by the equation; Ksal
0.00-1 ¦ 0.004*1)20 :0.010*1)10 - 3.013*TXP-06*BD. We also found that the filter with
the highest P removal efficiency was Grobbendonk 90/10% with 98.84% P removal
efficiency but had a lower saturated sorption capacity of 538.9 mg.kg1. The most durable
filler was Grobbendonk 100% with P sorption capacil} at breakthrough point of 942.2 mg
P kg . We estimated that Grobbendonk 100% was able to realise 62% of its total sorption
capacil} in the laboratory experiment (942.2 mg P.kg ' out of 1500 mg P.kg1). Therefore, our
best filler in terms of both efficiency and durability was Grobbendonk 100%. Similarl}. our
field trial identified Grobbendonk 100% as the best filler with estimated sorption capacil} al
breakthrough point of 985 mg P.kg ‘. P removal efficiency of 74% and achiev ing (>5.7% of its
estimated saturated sorption capacity. Based on these findings, we suggested that the
experiment to determine the maximum sorption capacil} of the filters be concluded so that
we can compare lhe capacity of the fillers used in relation to their total capacil} al saturation.
W e also suggested that a comprehensive large scale field trial be conducted for best
performing filters to validate both the laboralorv and pilot trial results.