Phosphorus dynamics in a shallow coastal lake system, Canterbury, New Zealand.
Thesis DisciplineWater Resource Management
Degree GrantorUniversity of Canterbury
Degree NameDoctor of Philosophy
Te Roto o Wairewa/Lake Forsyth (Wairewa) is a small (6.3 km²), polymictic, shallow (< 2 m), intermittently closed and open lake system on the south side of Banks Peninsula, New Zealand. The lake is hypertrophic and experiences regular algal blooms which are linked to high phosphorus (P) concentrations in the water column. This study investigates P dynamics in the lake and catchment, using field and laboratory investigations as well as geochemical modelling. The flux of P from the catchment to the lake was quantified, and the relative importance of internal P loading from the lake sediments determined. The important sedimentary P binding and release mechanisms were identified as well as the key environmental conditions controlling these mechanisms in the lake. A P budget was developed for the lake using a mass-balance approach over a fifteen month period (December 2012- March 2014). More than 7000 kg P.yr⁻¹ was transported to the lake from the catchment. This flux occurred primarily as particulate associated P (80 % of total) during large flood events, particularly from the Okana River subcatchment. The lack of a permanent opening to the sea resulted in the retention of 70 % of the external P load in the lake. This retained P was predominantly bound in the lake sediments. Mass balance calculations indicated that the flux of P from this sediment reservoir to the water column exceeded the external load during the summer months. Increases in total P (TP) concentrations during algal bloom events could only be explained by such internal loading. P binding and mobility in the lake sediments were investigated by sequential chemical extraction, pore water chemistry and geochemical modelling. Surface sediments were highly enriched in P with total sedimentary P concentrations of around 1500 mg P.kg⁻¹. With depth, the concentrations of the more mobile P fractions (organic-P and P associated with Fe and Al (hydr)oxides) decreased, indicating the release and upward flux of orthophosphate (PO₄⁻³) by organic decomposition and desorption from inorganic forms. In contrast the more refractory organic-P fraction and P associated with detrital apatite were immobile, and hence are progressively buried. The upward flux of released PO₄⁻³, which is immobilised again in the shallower, more oxic sediments where PO₄⁻³ adsorbs to phases such as Fe and Al (hydr)oxides, explains the highly P enriched layer at the sediment surface. Geochemical modelling indicated that in the surface sediments this cycle is affected by the rapid precipitation and dissolution dynamics of an amorphous Fe(III) (hydr)oxide such as ferrihydrite, and at depth by FeS₂ regulation of Fe availability. Secondary mineral phases such as MnHPO₄, strengite and vivianite may also exert some control over P binding in the sediments. Trends in pore water chemistry indicated significant seasonal changes in redox conditions, with more reduced sediments at the surface during winter. This was due to decomposition of the increased amounts of organic matter present in the surface sediments during winter. A more oxidised upper sediment profile with less organic content existed during the early summer growing season (prior to algal bloom formation). This was consistent with the increased pool of Fe-bound P in the upper sediments observed in summer, which was then available for release when redox boundaries migrated upward again, as observed during periods of algal bloom formation. The potential for release of P from the sediments, and the mechanisms which may promote this, were investigated by geochemical modelling and by laboratory experiments, which comprised a mixture of sediment core incubations and sediment slurry experiments. The potential for P release into an oxic, circum neutral-pH water column was negligible, likely due to the high ratio of reactive Fe to P in the sediments. Redox related P release from bed sediments began once the dissolved oxygen (DO) in the water column was completely depleted, and Eh was < 100 mV. Average release rates of 21 and 64 mg.m⁻².d⁻¹ were observed (Eh ≈ -200 and -300 mV respectively). pH related release was significant at pH > 9, and release rates from bed sediments of 3.3 and 11 mg.m⁻².d⁻¹ were observed (pH ≈ 9 and 10 respectively). These release rates have the potential to explain the increases in [TP] observed during algal blooms. pH-related release rates were significantly higher from resuspended sediment (78 and 100 mg.m⁻².d⁻¹ in pH ≈ 9 and 10 respectively) than from the bed sediments. Geochemical modelling indicated that increased salinity would increase pH-related P desorption, as well as decreasing the P binding capacity of the sediments, due to the reductive dissolution of Fe hydroxides and Fe immobilisation as FeS₂. Salinity gradients across the sediment-water interface resulting from ‘freshening’ events in the lake, resulted in P release of up to 13 mg.m⁻².d⁻¹. Environmental conditions which may control P binding and release in the lake sediments were investigated by the use of DO and pH data loggers as well as spot sampling in the lake. Although the water column was generally well mixed and oxygenated, bottom water anoxia did occur as ‘patchy’ spatially discrete and short lived (days, occasionally weeks) events. These occurred due to thermal stratification in macrophyte beds and the flux of organic material to the sediments from senescing macrophytes and algae. Salinity stratification may explain periods of variation in DO concentration associated with floods and lake openings. High salinity over longer time frames (4 months) was also correlated with high concentrations of chlorophyll-a in the lake. Macrophyte and algal photosynthesis produced water column pH > 9 and up to 10.5 for long periods of time, conditions compatible with significant sedimentary P release. Potential P release from observed DO and pH patterns was enough to explain observed increases in TP concentration associated with algal blooms. In this research a combination of field and laboratory work, as well as geochemical modelling was used to determine key processes associated with P dynamics in Wairewa. The high external P load, coupled with high P retention, indicated that the management of water quality in the lake must address the flux of P from the catchment, especially that associated with flood events in the Okana sub-catchment. At the same time, the large reservoir of P in the lake sediments, and the importance of internal P loading processes, means that a response to external P load reductions is likely to be slow. Hence management of internal P loading should be considered. The key roles of redox potential and water column pH, the ‘patchy’ and short lived nature of low DO events, and the significant roles of macrophyte beds and salinity in the internal cyling of P in the lake, are all important findings for the management of Wairewa and shallow coastal lakes in general.