Oxidation of Microcystis aeruginosa and Microcystins with Peracetic Acid.

Peracetic acid (PAA) shows potential for use in drinking water treatment as an alternative to prechlorination, such as for mussel control and disinfection by-product precursor destruction, though its impact as a preoxidant during cyanobacterial blooms remains underexplored. Here, Microcystis aeruginosa inactivation and microcystin-LR and -RR release and degradation using PAA were explored. The toxin degradation rates were found to be higher in alkaline conditions than in neutral and acidic conditions. However, all rates were significantly smaller than comparable rates when using free chlorine. The inactivation of M. aeruginosa cells using PAA was faster at acidic pH, showing immediate cell damage and subsequent cell death after 15-60 min of exposure to 10 mg/L PAA. In neutral and alkaline conditions, cell death occurred after a longer lag phase (3-6 h). During cell inactivation, microcystin-LR was released slowly, with <35% of the initial intracellular toxins measured in solution after 12 h of exposure to 10 mg/L PAA. Overall, PAA appears impractically slow for M. aeruginosa cell inactivation or microcystin-LR and -RR destruction in drinking water treatment, but this slow reactivity may also allow it to continue to be applied as a preoxidant for other purposes during cyanobacterial blooms without the risk of toxin release.

Using total adenosine triphosphate (tATP) measurements for cyanobacterial bloom monitoring and response assessment during algaecide treatments.

Total adenosine triphosphate (tATP) was investigated for its potential as a rapid indicator of cyanobacterial growth and algaecide effectiveness. tATP and other common bloom monitoring parameters were measured over the growth cycles of cyanobacteria and green algae in laboratory cultures and examined at a drinking water source during an active bloom. Strong correlations (R2>0.78) were observed between tATP and chlorophyll-a in cyanobacteria cultures. tATP offered greater sensitivity by increasing two orders of magnitude approximately 7 d before changes in chlorophyll-a or optical density were observed in Lyngbya sp. and Dolichospermum sp. cultures. Increases in tATP per cell coincided with the onset of exponential growth phases in lab cultures and increase in cell abundance in field samples, suggesting that ATP/cell is a sensitive indicator that may be used to identify the development of blooms. Bench-scale trials using samples harvested during a bloom showed that tATP exhibited a clear dose-response during copper sulfate (CuSO4) and hydrogen peroxide (H2O2) treatment compared to chlorophyll-a and cell counts, indicating that cellular production and storage of ATP decreases even when live and dead cells cannot be distinguished. During Copper (Cu) algaecide application at a reservoir used as a drinking water source, tATP and cell counts decreased following initial algaecide application; however, the bloom rebounded within 10 d showing that the Cu algaecide only has limited effectiveness. In this case, tATP was a sensitive indicator to bloom rebounding after algaecide treatments and correlated positively with cell counts (R2=0.7). These results support the use of tATP as a valuable complementary bloom monitoring tool for drinking water utilities to implement during the monitoring and treatment of cyanobacterial blooms.

Kinetics of chlorine and chloramine reactions in reverse osmosis permeate and their impact on radical formation during UV/chlorine advanced oxidation for potable reuse.

Knowledge of the speciation of chlorine and chloramines in reverse osmosis (RO) permeate is needed to estimate the performance (i.e., pollutant log reduction) of subsequent UV/chlorine advanced oxidation processes (AOPs). To accurately predict the speciation, a previously reported breakpoint chlorination kinetic model was experimentally validated for pH 5.5 and reaction times < 3 min and used to predict the kinetics of breakpoint chlorination in RO permeate. The predictions showed that eliminating chloramines by adding chlorine at a dose beyond the chlorine-to-nitrogen (Cl/N) breakpoint ratio is not practical due to the high breakpoint Cl/N ratio for RO permeate (∼3.0 molar ratio) and an estimated > 40 min reaction time. The conversion from monochloramine (NH2Cl) to dichloramine (NHCl2) is the major process involved, and either or both free chlorine and chloramines may be the major species present, depending on the Cl/N ratio. Model simulations showed that increasing the oxidant dose may not always enhance the performance of UV/chlor(am)ine in RO permeate, due to the need for a low free chlorine dose for optimal •OH exposure in RO permeate. Further UV/AOPs modelling showed that it is important to control the NH2Cl concentration to improve the UV/AOP performance in RO permeate, which may be achieved by extending the reaction time after chlorine is added or increasing the applied Cl/N ratio (e.g., increasing chlorine dose). However, these measures only enhance the pollutant percentage removal by about 5 % under the conditions modelled. A simulation tool was developed and is provided to predict the speciation of chlor(am)ine in RO permeate.

Predicting chlorine demand by peracetic acid in drinking water treatment.

Peracetic acid (PAA) may be used in drinking water treatment for pre-oxidation and mussel control at the intake. PAA may exert a downstream chlorine demand, but full details of this reaction have not been reported. There are three possible mechanisms of this demand: (1) PAA may react directly with chlorine; (2) PAA exists in equilibrium with hydrogen peroxide, which is known to react with chlorine; and (3) as H2O2 reacts with chlorine, PAA will hydrolyze to form more H2O2 to re-establish PAA/H2O2 equilibrium, thereby serving as an indirect reservoir of chlorine demand. While the H2O2 reaction with chlorine is well known, the other mechanisms of possible PAA-induced chlorine demand have not previously been investigated. The observed molar stoichiometric ratio of PAA to free chlorine (n) for the presumed direct PAA + free chlorine reaction was determined to be approximately 2, and the corresponding observed reaction rate coefficients at pH 6, 7, 8, and 9 were 2.76, 3.14, 1.61, 10.1 M-n·s-1, respectively (at 25 °C). With these estimated values, a kinetic model was built to predict the chlorine demand by PAA. The results suggest that chlorine demand from PAA is likely to be negligible over the course of several days (e.g., < 20% chlorine loss) for most conditions except for high pH (e.g., >8) and high PAA:Cl2 molar ratios (e.g., >2:1).

On the increasing competitiveness of UV/Cl to UV/H2O2 advanced oxidation as the organic carbon concentration increases.

UV/Cl and UV/H2O2 are advanced oxidation processes (AOPs) used for drinking water treatment and water reuse. This work explored the hypothesis that UV/Cl becomes more competitive to UV/H2O2 at neutral-to-high pH as the concentration of total organic carbon (TOC) increases. Lab experiments and kinetic modelling were used to compare initial pseudo first-order contaminant decay rate coefficients between the AOPs at various pH and TOC conditions. The relative effect of increasing TOC concentrations on UV/Cl vs. UV/H2O2 depended on the pH, contaminant, and organic matter reactivity towards radicals. For example, while the reaction rate coefficients during both AOPs generally decreased with increasing TOC, the UV/Cl reaction rate coefficients for the solely •OH-reactive sucralose decreased 41-138% less than the UV/H2O2 coefficients as the TOC concentration was increased from 0 to 5 mg-C L-1. However, UV/Cl was more affected than UV/H2O2 when targeting caffeine (a contaminant reactive to chlorine radicals). The data were used to define TOC-pH conditions for which either AOP would be more energy-efficient, under a set of standard conditions. The results suggest that UV/Cl may be competitive to UV/H2O2 under a wider range of treatment scenarios than has been conventionally thought based on tests in pure water.

Is In-Service Granular Activated Carbon Biologically Active? An Evaluation of Alternative Experimental Methods to Distinguish Adsorption and Biodegradation in GAC.

In-service granular activated carbon (GAC) may transform into biological activated carbon (BAC) and remove contaminants through both adsorption and biodegradation, but it is difficult to determine its biodegradative capacity. One approach to understand the GAC biodegradative capacity is to compare the performance between unsterilized and sterilized GAC, but the sterilization methods may not ensure effective microbial inhibition and may affect adsorption. This study identified the 14C-glucose respiration rate as the best metric to evaluate the effectiveness of three sterilization methods: sodium azide addition, autoclaving, and γ irradiation. The sterilization protocols were refined, including continuously feeding 300 mg/L of sodium azide, three cycles of autoclaving, and 10-12 kGy of γ irradiation. Parallel minicolumn tests were conducted to identify sodium azide addition as the most broadly effective sterilization method with an insignificant effect on adsorption in most cases, except for the adsorption of anionic compounds under certain conditions. Nevertheless, this problem was solved by decreasing the azide dosage as long as it is still sufficient to provide effective microbial inhibition. This study helps to develop an approach that differentiates adsorption and biodegradation in GAC, which could be used by future studies to advance our understanding of BAC filtration.

Evaluating perfluorooctanoic acid (PFOA) and perfluorooctanesulfonic acid (PFOS) removal across granular activated carbon (GAC) filter-adsorbers in drinking water treatment plants.

Granular activated carbon (GAC) was harvested from six filter-adsorbers that are used for taste and odour control in three drinking water treatment plants in Ontario, Canada, and evaluated for the removal of perfluorooctanic acid (PFOA) and perfluorooctanesulfonic acid (PFOS) using minicolumn tests under different operational conditions. Parallel column tests were conducted using unsterilized GAC and sterilized GAC to distinguish adsorption from potential biodegradation of PFOA and PFOS across the GAC. It was observed that the GAC could achieve approximately 20% to 55% of PFOA and PFOS removal even after a long period of GAC operation (e.g., 6 years). There was no evidence of PFOA and PFOS biodegradation, so the removal in GAC can be attributed solely to adsorption under the conditions tested. However, in one location, there was evidence suggesting both removal and formation of PFOS and PFOA across the GAC, with the formation presumably due to the biotransformation of pre-existing precursors in the source water. Additionally, GAC service time and empty bed contact time (EBCT) were identified to be important factors that could affect the removal of PFOA and PFOS. Based on this information, an empirical model was proposed to predict PFOA and PFOS removal in GAC filter-adsorbers as a function of GAC service time and EBCT. This study provides useful information for utilities that have installed GAC for taste and odour control but may consider per- and polyfluoroalkyl substances (PFAS) removal as an additional voluntary objective or due to more stringent guidelines.

Evaluating the relative adsorption and biodegradation of 2-methylisoborneol and geosmin across granular activated carbon filter-adsorbers.

This study investigated the relative contributions of adsorption vs. biodegradation towards 2-methylisoborneol (MIB) and geosmin removal in the granular activated carbon (GAC) harvested from six filter-adsorbers in three drinking water treatment plants in the Great Lakes region. Column tests using azide-treated (sterilized) and untreated GAC in parallel were used to isolate the two effects. It was identified that substantial MIB and geosmin biodegradation in the GAC was occurring in one location, and that GAC in some cases had significant adsorption capacity after as much as 9 years of operation. Four alternative biological parameters (adenosine triphosphate, esterase activity, phosphatase activity, and 14C-glucose respiration rate) were measured to quantify the biological activity of the GAC, and 14C-glucose respiration rate was identified to be a potential indicator for GAC biodegradative capacity in terms of MIB, geosmin, and dissolved organic carbon. Several potential MIB and geosmin biodegradation products were also identified using non-targeted screening analysis. By using the new tools identified in this study, we can begin to better understand where adsorption vs. biodegradation may predominate under real-world conditions (e.g., different temperatures, influent concentrations, and empty bed contact time), leading ultimately to more cost-effective use of GAC.

Understanding adsorption and biodegradation in granular activated carbon for drinking water treatment: A critical review.

Drinking water treatment plants use granular activated carbon (GAC) to adsorb and remove trace organics, but the GAC has a limited lifetime in terms of adsorptive capacity and needs to be replaced before it is exhausted. Biological degradation of target contaminants can also occur in GAC filters, which might allow the GAC to remain in service longer than expected. However, GAC biofiltration remains poorly understood and unpredictable. To increase the understanding of adsorption and biodegradation in GAC, previous studies have conducted parallel column tests that use one column of GAC (potentially biologically active) to assess overall removal via both adsorption and biodegradation, and one column with either sterilized GAC or biological non-adsorbing media to assess adsorption or biodegradation alone. Mathematical models have also been established to give insight into the adsorption and biodegradation processes in GAC. In this review, the experimental and modeling approaches and results used to distinguish between the role of adsorption and biodegradation were summarized and critically discussed. We identified several limitations: (1) using biological non-adsorbing media in column tests might lead to non-representative extents of biodegradation; (2) sterilization methods may not effectively inhibit biological activity and may affect adsorption; (3) using virgin GAC coated with biofilm could overestimate adsorption; (4) potential biofilm detachment during column experiments could lead to biased results; (5) the parallel column test approach itself is not universally applicable; (6) competitive adsorption was neglected by previous models; (7) model formulations were based on virgin GAC only. To overcome these limitations, we proposed four new approaches: the use of gamma irradiation for sterilization, a novel minicolumn test, compound-specific isotope analysis to decipher the role of adsorption and biodegradation in situ, and a new model to simulate trace organic adsorption and biodegradation in a GAC filter .

Evaluation of ultraviolet/peracetic acid to degrade M. aeruginosa and microcystins -LR and -RR.

The reactivity of peracetic acid (PAA) alone, and PAA exposed to ultraviolet radiation (UV), was investigated on Microcystis aeruginosa cells, and on microcystin-LR and -RR. Reaction rates between PAA and MC-LR (k = 3.46 M-1 s-1) and MC-RR (k = 2.67 M-1 s-1) were determined in an unbuffered acidic solution, and they are approximately 35-45 times lower than a previously reported reaction rate between MC-LR and chlorine at pH 6. Peracetic acid reacted with M. aeruginosa cells as a function of PAA and cell concentrations, with 10 mg/L PAA resulting in 1-log reduction of total MC-LR within 15 min. Advanced oxidation by UV/PAA readily degraded MC-LR and MC-RR, outperforming UV/H2O2 at pH 7.7 by > 50% on an equimolar basis. Indirect photolysis at this pH is due to •OH and organic radicals, as determined by trials in the presence of excess tert-butanol to scavenge •OH. The process is less effective when the pH departs from neutral conditions (5.9 or 10.6) due to the decreased effects of both radicals. These findings suggest that PAA alone might be a viable option for cyanobacteria and microcystins control in preoxidation applications and that UV/PAA is an effective process for degrading MC-LR and MC-RR at neutral pH.

Modeling the efficiency of UV at 254 nm for disinfecting the different layers within N95 respirators.

The study presented a Monte Carlo simulation of light transport in eight commonly used filtered facepiece respirators (FFRs) to assess the efficacy of UV at 254 nm for the inactivation of SARS-CoV-2. The results showed different fluence rates across the thickness of the eight different FFRs, implying that some FFR models may be more treatable than others, with the following order being (from most to least treatable): models 1512, 9105s, 1805, 9210, 1870+, 8210, 8110s and 1860, for single side illumination. The model predictions did not coincide well with some previously reported experimental data on virus inactivation when applied to FFR surfaces. The simulations predicted that FFRs should experience higher log reductions (>>6-log) than those observed experimentally (often limited to ~5-log). Possible explanations are virus shielding by aggregation or soiling, and a lack of the Monte Carlo simulations considering near-field scattering effects that can create small, localized regions of low UV photon probability on the surface of the fiber material. If the latter is the main cause in limiting practical UV viral decontamination, improvement might be achieved by exposing the FFR to UV isotropically from all directions, such as by varying the UV source to the FFR surface angle during treatment.

Machine learning for anomaly detection in cyanobacterial fluorescence signals.

Many drinking water utilities drawing from waters susceptible to harmful algal blooms (HABs) are implementing monitoring tools that can alert them to the onset of blooms. Some have invested in fluorescence-based online monitoring probes to measure phycocyanin, a pigment found in cyanobacteria, but it is not clear how to best use the data generated. Previous studies have focused on correlating phycocyanin fluorescence and cyanobacteria cell counts. However, not all utilities collect cell count data, making this method impossible to apply in some cases. Instead, this paper proposes a novel approach to determine when a utility needs to respond to a HAB based on machine learning by identifying anomalies in phycocyanin fluorescence data without the need for corresponding cell counts or biovolume. Four widespread and open source algorithms are evaluated on data collected at four buoys in Lake Erie from 2014 to 2019: local outlier factor (LOF), One-Class Support Vector Machine (SVM), elliptic envelope, and Isolation Forest (iForest). When trained on standardized historical data from 2014 to 2018 and tested on labelled 2019 data collected at each buoy, the One-Class SVM and elliptic envelope models both achieve a maximum average F1 score of 0.86 among the four datasets. Therefore, One-Class SVM and elliptic envelope are promising algorithms for detecting potential HABs using fluorescence data only.

Wavelength-dependent time-dose reciprocity and stress mechanism for UV-LED disinfection of Escherichia coli.

Ultraviolet (UV) disinfection efficiency by low-pressure (LP) mercury lamp depends on the UV fluence (dose): the product of incident irradiance (fluence rate) and exposure time, with correction factors. Time-dose reciprocity may not always apply, as higher UV-LP inactivation of E. coli was obtained at a higher irradiance over shorter exposure time, for the same UV fluence. Disinfection by UV LEDs is limited by low radiant flux compared to mercury LP lamps. Our goal was to determine the UV-LED time-dose reciprocity of E. coli for four different central LED wavelengths (265, 275, 285 and 295 nm) under different fluence rates. Inactivation kinetics determined at UV-LED265 was not affected by the fluence rate or exposure time for a given UV fluence. In contrast, UV-LED275, UV-LED285, and UV-LED295 led to higher inactivation at low fluence rate coupled to high exposure time, for the same UV fluence. The intracellular damage mechanisms for each LED central wavelength were determined by using the bioreporters RecA as an indicator of bacterial DNA damage and SoxS as an indicator of oxidative stress. For 265 nm, higher DNA damage was observed, whereas for 285 and 295 nm, higher oxidative stress (possibly due to reactive oxygen species [ROS] damage) was observed. ROS inactivation of E. coli was predicted to be more effective when keeping the ROS concentration low but allowing longer exposure, for a given UV fluence.

Light propagation within N95 filtered face respirators: A simulation study for UVC decontamination.

This study presents numerical simulations of UVC light propagation through seven different filtered face respirators (FFR) to determine their suitability for Ultraviolet germicidal inactivation (UVGI). UV propagation was modeled using the FullMonte program for two external light illuminations. The optical properties of the dominant three layers were determined using the inverse adding doubling method. The resulting fluence rate volume histograms and the lowest fluence rate recorded in the modeled volume, sometimes in the nW cm-2 , provide feedback on a respirator's suitability for UVGI and the required exposure time for a given light source. While UVGI can present an economical approach to extend an FFR's useable lifetime, it requires careful optimization of the illumination setup and selection of appropriate respirators.

Evaluation of phosphorus removal from a lake by two drinking water treatment plants.

The impact of drinking water treatment plants on phosphorous in a lake has never been previously reported. In this mass balance study, phosphorus removal by a conventional plant and a membrane plant on Lake Simcoe was monitored. Approximately 16 kg of phosphorus per year were removed from the lake by the membrane plant, representing 72% of the influent phosphorous load to the plant. The membrane plant did not practice coagulation, so approximately two-thirds of the removal was via circulation of the treated water to the municipal wastewater treatment plant where phosphorous was removed. The remaining third was removed by the membranes. The conventional plant removed approximately 10 kg of phosphorus per year, representing 92% of the influent phosphorus loading. In this plant, polyaluminum chloride coagulation and subsequent sludge removal were responsible for approximately two-thirds of the phosphorous removal, with the remainder removed via circulation of the treated water to the municipal wastewater treatment plant.

External Standard Calibration Method To Measure the Hydroxyl Radical Scavenging Capacity of Water Samples.

The hydroxyl radical (•OH) scavenging capacity is a useful parameter for the design and operation of an advanced oxidation process (AOP) in water treatment. The scavenging capacity may change with time, and it would be useful to continuously measure this change to be able to optimize AOP doses. In this study, we first reviewed current methods for scavenging capacity measurement to identify strengths and weaknesses of each method. This information helped guide the design of an external calibration method to allow straightforward laboratory and field measurement of •OH scavenging capacity. The method used low-pressure UV/H2O2 as the •OH generation system, methylene blue (MB) as the probe compound, and isopropyl alcohol (IPA) as the standard. By monitoring, offline, the color decay of MB in a series of IPA solutions with different scavenging capacity, a calibration curve was established between the color decay rate and the scavenging capacity. The measured color decay in real water samples can then be used with this external calibration to estimate their scavenging capacity. Work was undertaken to ensure that the process would be robust under a wide range of water quality conditions. Parallel tests using this method compared with the benchmark methods confirmed its robustness and accuracy.

Chlorine is preferred over bisulfite for H2O2 quenching following UV-AOP drinking water treatment.

Drinking water treatment using UV/H2O2 advanced oxidation typically results in residual H2O2 that requires quenching to minimize its interference with downstream processes. Chemical quenching using chlorine or bisulfite are options, but there is some uncertainty in the literature about the kinetics of the bisulfite reaction, with some reports quoting the reaction as fast, and others as slow. Part of the contradictory information may be due to interference in H2O2 analysis by bisulfite. An analytical method was developed to avoid this interference, in which monochloramine first selectively quenched bisulfite, and then H2O2 was measured spectrometrically using titanium(IV) oxysulfate for color development. The confirmatory experiments suggested that the bisulfite reaction with H2O2 is actually relatively slow, with a half-life in the order of hours to days depending on the pH and the reagent concentrations. As a result, within the typical pH range of drinking water treatment (e.g., 6-9), chlorine is preferred over bisulfite as the H2O2 quenching agent on the basis of reaction kinetics. However, a decrease in pH will lead to an increase in the bisulfite-H2O2 reaction rate along with a decrease in the Cl2-H2O2 reaction rate, such that at pH < 5.7 bisulfite is the faster reagent. Both bisulfite and chlorine were observed to react with H2O2 following a stoichiometric ratio of 1:1 in the natural water matrix tested.

Full-scale comparison of UV/H2O2 and UV/Cl2 advanced oxidation: The degradation of micropollutant surrogates and the formation of disinfection byproducts.

The photolysis of chlorine by UV light leads to the formation of the hydroxyl radicals (OH) as well as reactive chlorine species (RCS) that can be effective as advanced oxidation processes (AOPs) for water treatment. Much of the research to date has been done at laboratory- or bench-scale. This study reports results from a model that demonstrates that the relative effectiveness of the UV/Cl2 AOP compared to the more traditional UV/H2O2 AOP is a function of optical path length. As such, the relative effectiveness of the two treatment options evaluated at small scale may not reflect the relative performance at full-scale, making results previously obtained at small-scale potentially less scalable. This study therefore compares the performance of UV/Cl2 to UV/H2O2 at a full-scale water treatment plant, using sucralose and caffeine as spiked surrogates for contaminants that are reactive solely to OH radicals, and to both OH and RCS, respectively. pH was varied between 6.5 and 8.0. The results demonstrated that when using a medium pressure UV lamp, UV/Cl2 might lead to approximately twice the production of OH radicals as UV/H2O2 at pH 6.5 when using the same molar oxidant concentration, but adding chlorine to the UV reactor at pH 8.0 had a negligible impact on OH radical concentration in comparison to UV alone. The study also confirmed previous small-scale results that RCS can be a major contributor to UV/Cl2 treatment for compounds such as caffeine that are susceptible to RCS, with UV/Cl2 effective at both pH 6.5 and 8.0 for such compounds. Disinfection byproducts were monitored, with adsorbable organohalide (AOX) formation increasing by approximately 10 µg-Cl/L due to chlorine photolysis, but only at pH 6.5 and not at pH 8.0. This implies that UV/Cl2 might increase AOX mostly due to reaction between OH and organic precursors to make them more reactive with chlorine, and not due to RCS. The formation of specific DBPs of current or emerging regulatory interest was minimal under all conditions, except for chlorate. Chlorate yields were in the order of 6-18% of the photolysed chlorine.

Cyanotoxins and Cyanobacteria Cell Accumulations in Drinking Water Treatment Plants with a Low Risk of Bloom Formation at the Source.

Toxic cyanobacteria have been shown to accumulate in drinking water treatment plants that are susceptible to algal blooms. However, the risk for plants that do not experience algal blooms, but that receive a low influx of cells, is not well known. This study determined the extent of cell accumulation and presence of cyanotoxins across the treatment trains of four plants in the Great Lakes region. Samples were collected for microscopic enumeration and enzyme-linked immunosorbent assay (ELISA) measurements for microcystins, anatoxin-a, saxitoxin, cylindrospermopsin, and ß-methylamino-L-alanine (BMAA). Low cell influxes (under 1000 cells/mL) resulted in significant cell accumulations (over 1 × 105 cells/mL) in clarifier sludge and filter backwash samples. Microcystins peaked at 7.2 µg/L in one clarifier sludge sample, exceeding the raw water concentration by a factor of 12. Anatoxin-a was detected in the finished drinking water of one plant at 0.6 µg/L. BMAA may have been detected in three finished water samples, though inconsistencies among the BMAA ELISAs call these results into question. In summary, the results show that plants receiving a low influx of cells can be at risk of toxic cyanobacterial accumulation, and therefore, the absence of a bloom at the source does not indicate the absence of risk.