Heberling Matthew T, Price James I, Nietch Christopher T, Elovitz Michael, Smucker Nathan J, Schupp Donald A, Safwat Amr, Neyer Tim
US Environmental Protection Agency, Office of Research and Development, Cincinnati, OH 45268, USA.
University of Wisconsin - Milwaukee, School of Freshwater Sciences, Milwaukee, WI 53204, USA.
Water Resour Res. 2022 May 1;58(5):1-17. doi: 10.1029/2021wr031257.
We estimate a cost function for a water treatment plant in Ohio to assess the avoided-treatment costs resulting from improved source water quality. Regulations and source water concerns motivated the treatment plant to upgrade its treatment process by adding a granular activated carbon building in 2012. The cost function uses daily observations from 2013 to 2016; this allows us to compare the results to a cost function estimated for 2007-2011 for the same plant. Both models focus on understanding the relationship between treatment costs per 1,000 gallons (per 3.79 m3) of produced drinking water and predictor variables such as turbidity, pH, total organic carbon, deviations from target pool elevation, final production, and seasonal variables. Different from the 2007-2011 model, the 2013-2016 model includes a harmful algal bloom toxin variable. We find that the new treatment process leads to a different cost model than the one that covers 2007-2011. Both total organic carbon and algal toxin are important drivers for the 2013-2016 treatment costs. This reflects a significant increase in cyanobacteria cell densities capable of producing toxins in the source water between time periods. The 2013-2016 model also reveals that positive and negative shocks to treatment costs affect volatility, the changes in the variance of costs through time, differently. Positive shocks, or increased costs, lead to higher volatility compared to negative shocks, or decreased costs, of similar magnitude. After quantifying the changes in treatment costs due to changes in source water quality, we discuss how the study results inform policy-relevant decisions.
我们估算了俄亥俄州一家水处理厂的成本函数，以评估因原水水质改善而避免的处理成本。法规和原水问题促使该厂在2012年通过增加一座颗粒活性炭设施来升级其处理工艺。成本函数使用了2013年至2016年的每日观测数据；这使我们能够将结果与同一工厂2007 - 2011年估算的成本函数进行比较。两个模型都着重于理解每1000加仑（每3.79立方米）生产的饮用水的处理成本与诸如浊度、pH值、总有机碳、与目标水池水位的偏差、最终产量以及季节变量等预测变量之间的关系。与2007 - 2011年的模型不同，2013 - 2016年的模型纳入了有害藻华毒素变量。我们发现新的处理工艺导致了一个与涵盖2007 - 2011年的模型不同的成本模型。总有机碳和藻毒素都是2013 - 2016年处理成本的重要驱动因素。这反映出不同时期原水中能够产生毒素的蓝藻细胞密度显著增加。2013 - 2016年的模型还显示，处理成本的正向和负向冲击对波动性（成本方差随时间的变化）的影响不同。与幅度相似的负向冲击（成本降低）相比，正向冲击（成本增加）会导致更高的波动性。在量化了因原水水质变化而导致的处理成本变化之后，我们讨论了研究结果如何为与政策相关的决策提供信息。
