The temporal and spatial variability of organic matter and its effect on membrane filtration

Abstract

The overall goal of this research was to investigate spatial and temporal variations in the concentration and characteristics of organic matter (OM) in the upper Mississippi River and the effects of these variations on the treatability of the water, specifically membrane filtration. OM is ubiquitous in surface waters as a result of natural and anthropogenic inputs. Although not directly harmful to human health or water quality, OM challenges drinking water utilities by increasing chemical consumption, serving as a precursor for disinfection byproducts, and fouling filtration membranes as seen through permeability or transmembrane pressure-normalized flux decline. It is therefore critical to understand how the concentration and composition of OM changes spatially and temporally in a watershed. Temporal OM variability in the upper Mississippi River was assessed by near-real time monitoring of dissolved organic carbon (DOC) concentration and ultraviolet absorbance at 254 and 280 nm for approximately one year. Spatial OM variability was assessed by collecting water samples along a 648-mile stretch of the upper Mississippi River from its headwaters at Lake Itasca to La Crescent, Minnesota and analyzing the samples for OM concentration and composition. In addition, monthly sampling of eight upper Mississippi River tributaries with varying dominant watershed land uses was performed for nine months to simultaneously assess both spatial and temporal OM variability. The water samples from these campaigns were filtered through four different ultrafiltration membranes to assess how the various combinations of membrane characteristics, such as hydrophobicity and pore size, and OM characteristics would affect membrane fouling. Near real-time monitoring of the Mississippi River, the water source for the 70 MGD Columbia Heights Membrane Filtration Plant (Columbia Heights, Minnesota), over an 11-month period showed that the permeability of hydrophilic ultrafiltration membranes with a pore size of 0.02 µm was negatively correlated with river flow (p < 0.01) and raw water specific ultraviolet absorbance (SUVA) (p < 0.01), but not raw water dissolved organic carbon (DOC) concentration (p = 0.865). From a spatial aspect, the concentration (as measured by TOC and DOC) and character (as observed through SUVA and fluorescence spectroscopy) of OM were relatively constant along the length of the upper Mississippi River during a late summer to early fall sampling period despite inputs from tributaries rich in organic matter and wastewater effluent. The estimated DOC export from the upper Mississippi River was 0.0088 kg/d. Overlaying sampling points on a digital elevation model and land use map showed that concentrations of TOC and DOC were negatively correlated with the percent of agricultural land and positively correlated with percent shrubland area. Stirred cell membrane trials on seasonal waters collected from the upper Mississippi River before and after lime softening showed that pretreatment significantly reduces membrane fouling (12.0 - 47.5%) by preferentially removing larger hydrophobic OM. Although performance varied, the most hydrophilic membrane tested (contact angle 22.6°) exhibited the lowest permeability decline of all of the membranes for all water samples before and after fouling. When compared to organic matter standards from International Humic Substances Society, the natural, unfractionated waters fouled the membranes more, possibly as a result of increased interactions between the hydrophobic and hydrophilic fractions. Additional stirred cell membrane trials on seasonal waters collected from tributaries of the upper Mississippi River with varying land uses indicated that both membrane hydrophobicity and seasonal factors such as snow melt and storm events and land use were important in determining overall membrane performance. The results of this research encourage extensive near real-time monitoring and membrane pilot studies as well as pretreatment and large-scale watershed assessments for utilities considering membrane filtration systems for surface water treatment. In addition, this work could aid in the formulation of policies and regulations governing land use, land development, stormwater management, and wastewater discharges.