Two Decades Constructed Wetland Experience in Treating Municipal Effluent for Power Plant Cooling at the Shand Power Station, SaskPower Part VI: Multivariate (Quadratic) Function Model(s) for o-PO43--P Removal

Emmanuel K. Quagraine



This paper is sixth in publication series on 20-year period constructed wetland operating experience at SaskPower’s Shand Power Station and describes multivariate regression models foro-PO43--P (OP) removal efficiency during two main plant growth cycle periods: the plant growing season of May to September (R2= 0.84-0.95, p<0.001) and during the die-off period of October to November.For the growing season, the best most meaning model to describethe OP removal efficiency was a quadratic function as follows: –4.81(OP)2– 4.53T + 0.39OP*BOD + 4.25OP*pH + 58.88TAN – 6.91pH*TAN + 9.36; where p for each coefficient is <0.05. Aging effect of the CW on OP removal efficiency during the plant growing period was somehow also revealed but it was statistically weak (p = 0.15) and effect OP release of only ~0.45% per year. Of particular note, the results here suggest pH as an important variable during the plant growing season showing up in TAN, o-PO43-speciation and in some other ill-defined forms, whose proper control and optimization could lead to more efficient CW performance. HPO42- (and not thecommonly purported H2PO4-) was implicated as the dominant target species for biological uptake of OP, ashigh pH interactions with influent OP concentration favored OP removal within the operating pH range of 8.0 to 9.7. Unlike the growing season, the best model fits for the plant senescing period were non-polynomial,but multivariate functions. With exception of temperature, the derived regression coefficients (βis) and constant (intercept) were vastly different in magnitude and in some cases by the direction of influence; being especially marked by:net background release of OP and no significant effect by pH on OP removal.The consistent negative temperature effect shown for both periods is likely associated with microbial activities on detritus materials either as suspended solids in the overlying and/or in the sediments.

Keywords: Wetlands; temperature; phosphate-phosphorus; quadratic regression model; pH; BOD; TAN