Evaluation of storage strategies and techniques for the hydrochemical analysis of nutrients in Atlantic Ocean coastal water samples

Filtered and unfiltered samples of coastal seawater from the Atlantic Ocean (Muizenberg and Llandudno beaches, Cape Town) representing a non upwelling region and an upwelling region respectively were stored at -20 deg C. Deviations in the concentrations of silicate, inorganic phosphate, nitrite, nitrate and urea from the initial concentrations determined on collection of the seawater samples were monitored at intervals over a twelve week storage period. Chemical analysis of filtered and non-filtered water samples on the day of collection demonstrated that filtration affected the concentration of nutrients in the samples. In three of the eight filtered seawater samples the filtrate contained a lower concentration of nutrients than the corresponding non-filtered seawater samples. Sample t-tests indicated that these differences were not significant. The study shows that no single storage strategy for the determination of all nutrients in seawater can be recommended and that the filtering/storage strategy used should be adapted depending on the nature of the nutrient to be analysed. The mean inorganic phosphate concentrations present on sampling at Llandudno and Muizenberg beach ranged from approximately 0.4 μmol to 0.7 μmol L -1. Filtered frozen samples for inorganic phosphate determinations stored well for twelve weeks with minor deviations in concentration from the initial concentration obtained at week 0 being recorded. This strategy showed less variability over the twelve week trial than shown with non-filtered samples and is the recommended strategy for the preservation of inorganic phosphate. Ideally samples should be analysed within the first 2 weeks of storage. The mean dissolved silicate concentrations present on sampling at Llandudno and Muizenberg beach ranged from approximately 9 μ mol to 14 μmol L -1. Neither of the two storage strategies explored during this study proved ideal as fluctuations in the level of dissolved silicate were noted using both storage regimes. Where immediate analysis is not possible the strategy which provided the least variability in this study i.e. storage of non-filtered samples at -20°C is recommended. The mean nitrite concentrations present on sampling at Llandudno and Muizenberg beach ranged from approximately 0.4 μmol to 0.7 μmol L -1. Ideally samples (filtered and non-filtered) should be analysed within the first 2 weeks of storage as further delays in analysis resulted in fluctuations in the nitrite concentration. The mean dissolved urea concentrations present in non-filtered water on sampling at Llandudno was 0.2 μmol L -1 with 1.5 μmol L-1 present in the sample taken at Muizenberg beach. The major feature of the preservation study on variations in the concentration of urea in stored water over time is the relative stability of samples stored frozen for periods for periods of up to 8 weeks. A second study on ten Atlantic Ocean coastal water samples collected at sites on the Cape Peninsula compared analytical techniques used to determine silicate and dissolved nitrate concentrations in seawater. For inorganic silicate analysis a modification of the manual method of Grasshoff (1976) and an automated flow injection analysis (FIA) method were used. The FIA method yielded concentrations that were on average two times higher in seven out of the ten analyses. For the determination of dissolved nitrate concentrations two manual methods (differing in the method of reduction of nitrate to nitrite) were used: shaking the sample with cadmium pellets and passing the sample through a cadmium column. In both cases the effect of using Tris and ammonium chloride (NH4 Cl) buffers was determined. In addition nitrate concentrations were determined using FIA. Dissolved nitrate concentrations determined by the FIA and the cadmium column method (using both Trisand NH4 Cl buffers) differed significantly whereas those obtained using the shaking method were frequently low or negative. The shaking method used to determine nitrate concentrations is not recommended