| dc.description.abstract | Woodchip filters have received attention in recent years for their ability to sustain denitrification activity across multiyear time frames. However, in most freshwater aquatic ecosystems, phosphorus (P) rather than nitrogen (N) is the nutrient considered most responsible for eutrophication. Previous studies have indicated that woodchip filters have limited ability to remove dissolved P, but P export in agricultural runoff is often dominated by particulate P (PP). Woodchip media, because of its high porosity, permeability, surface roughness and plate-like structure of the particles, could be effective for physical filtration of particulate phosphorus. In this study, woodchip filter systems were tested for treatment of PP in agriculturally impacted surface waters at five sites in southern Ontario.
A woodchip filter system installed near Bradford, ON was used to treat highly turbid root vegetable wash water from a local farm and focused on the treatment of total suspended solids (TSS) and PP. The full-scale treatment system consisted of a sedimentation tank (12.3 m3) followed by the woodchip filter (16.1 m3) and had two stages of testing. In the initial stage, the filter media consisted of woodchips with a layer of sawdust, and in the second stage, the media contained woodchips only. The full-sale treatment system was sampled from November 2014 to March 2016 and proved effective for TSS and PP removal during both treatment stages, averaging overall removal of 99% and 91%, respectively, in the first stage, and 96% and 77%, respectively, in the second stage. During the operation of the full-scale treatment system, the sludge within the sedimentation tank was regularly monitored and was removed on two occasions. Also during this time, sludge accumulation within the top layer of woodchips required replacement of the top layer on one occasion, September 2015.
A woodchip filter was installed near Barrie, ON to treat particulate P in an agricultural drainage ditch adjacent to fields where row crops are grown. In this case the filter consisted of 20 m3 of woodchips trenched in to the bottom of the stream (stream-bed filter). Stream flow was induced through the filter by placement of a gravel riffle at its downstream end. This filter was monitored from December 2014 to March 2016 and proved effective for P removal in the stream water, which had low to moderate turbidity, averaging total P removal of 58%, the majority of which was PP. Nitrate removal in the filter was modest, averaging only 1 mg/L NO3--N, because the filter was operated at relatively high flow rates (average hydraulic retention time of 0.4 days) such that denitrification activity was incomplete.
A woodchip filter was installed near Keswick, ON to remove TSS and associated PP, as well as NO3--N, from a tile drain at a sod farm. The filter consisted of 36 m3 of woodchips trenched into the subsurface near the drain outlet and was monitored intermittently from May 2014 to March 2016. Overall, geochemical parameters were not substantially changed during treatment in the filter. This was primarily because TSS, total P and NO3--N concentrations were relatively low at this site, averaging 20 mg/L, 24 µg/L and 3.0 mg/L, respectively, such that the woodchip filter had little opportunity to further diminish these already low values. Secondly, the tile drain unexpectedly remained dry throughout the summer and early fall months and the filter experienced freezing problem during winter. Consequently, achieving desired flow rates through the filter was problematic. Results indicated that TSS and PP values were too low and therefore this site was not well-suited for the implementation of this type of woodchip filter.
In a previous study (van Driel, 2006), a woodchip filter was installed near St. Marys, ON, in 2002, to treat NO3--N from an agricultural drainage tile, adjacent to a field where row crops are grown. Although extensive monitoring of the media longevity for NO3--N removal has been undertaken, little attention has been paid to P removal associated with this filter. During this study, in addition to NO3--N removal, the P removal capacity of the filter was monitored from May 2014 to March 2016, at which time the filter was 12-14 years old. Monitoring revealed that TP in the filter effluent actually increased, although the significance was low, from a mean of 29 ug/L to 83 ug/L, and this increase was dominated by SRP and not PP. There was little correlation between TP and SRP removal with hydraulic retention time (HRT), however, the relationships between these two parameters and HRT were significant. In contrast, NO3--N removal in the filter remained significant, decreasing from a mean of 7.2 mg/L in the influent to 2.3 mg/L in effluent. Monitoring at this site provides evidence that wood particle filters have the potential to leach low levels of dissolved phosphorus during long term operation. This could be the result of slow leaching of P associated with the sediment retained within the filter. Also, the correlation of higher SRP values with increased Fe concentrations, suggests that reductive dissolution of ferric iron solids and subsequent release of sorbed P also plays a role in the observed SRP increases. This observed occurrence of SRP leaching at this site has important implications for the long term management strategies for such filters.
In a previous study (Robertson and Merkley, 2009) a woodchip filter, of the same design as the Big Bay stream-bed filter, was installed at the headwaters of Avon River, near Stratford, ON. The filter was designed to treat NO3--N in a drainage ditch, adjacent to fields where row crops are grown.. The filter has been monitored extensively in previous studies for NO3--N removal, but P removal received little attention. During the current study, both the NO3--N and P removal capacities of the filter were monitored during the period May 2014 to November 2015. During this monitoring period there was a consistent problem with inadequate flow rates (< 4 L/min) through the filter and secondly, there was an observation of very dissimilar chloride values in the filter influent and effluent. This indicated that the woodchip media had likely become substantially impermeable due to sediment accumulation in the pore space, such that the filter effluent was apparently dominated by incoming groundwater flow, rather than flow originating from the stream. Consequently, the filter is no longer functioning as designed.
This study has demonstrated that woodchip filters are a cost effective and low maintenance method for the removal of particulate P from agricultural waters under appropriate conditions. The study also demonstrates design options, flow conditions and maintenance requirements for the newly installed and older (> 10 years) filters for effective nutrient removal. | en |
