Abstract

Sensor fouling affects continuous water quality monitoring. Experiments were performed with probes for 10 months with copper guards; PVC guards coated with paint mixed with copper powder; and a water jet mechanism. Previous studies focused on the use of one antifouling method, using biocides such as copper and silver and mechanical cleaning, such as air jets, mostly on single parameter probes. The present work looks at multi-parameter probes and compares different antifouling options. Additional periods of reliability were verified with all alternatives. For the conductivity parameter, an additional period of up to 29 days (48%) with copper guards was observed, 26 days (43%) with water jets, and 37 days (61%) with copper powder paint mix. For dissolved oxygen, an additional period of up to 23 days (38%) was observed with copper guards, 29 days (48%) with water jets, and 60 days (100%) with copper powder paint mix. For pH, monitoring was reliable for 90 days (100%) with copper guards, and 60 days (67%) with copper powder paint mix. For turbidity, an additional period of up to 7 days (12%) was observed with copper guards, 3 days (5%) with water jets, and 57 days (95%) with copper powder paint mix.

INTRODUCTION

Water quality affects human health and ecological balance. Maintaining water quality depends on monitoring, as reliable data are critical for environmental management and protection. Continuous monitoring is still a challenge, since the equipment requires constant calibration and has a high cost. In a reliable monitoring system, the equipment must provide continuous readings for the longest time possible. The results help to detect pollution and determine actions to maintain water quality (Rao et al. 2013). Water quality monitoring allows contaminants to be detected (Rocha 2009) and is usually part of a Water Resources Plan (Knapik 2009), which serves as a basis for planning and implementing water body control or recovery instruments (Coelho 2013).

Water quality monitoring is a long-term activity (Behmel et al. 2016). Technological advancements contribute to the development of cost-effective sensors that are used in long-term monitoring systems. These monitoring systems, when automated and in conjunction with the development of advanced computer programs, increase the ability to monitor water quality. Communication is via the internet, creating and storing increasing amounts of data (Glasgow et al. 2004; Flynn et al. 2010; Wong & Kerkez 2016). However, the remaining issue is long-term sensor reliability.

Water management depends on monitoring to detect changes in the environment, using continuous and real-time probes that remain in the water for prolonged periods, under the effect of fouling agents. With emphasis on the need for continuous monitoring to obtain data to be used by the committees and bodies responsible for water management, several works report periods of reliable data acquisition ranging from 3 to 60 days for multiparametric probes (Champely & Dolédec 1997; Aloui & Gueddari 2009; Bowes et al. 2011; Xu et al. 2011; Halliday et al. 2012; Albert et al. 2015; Holeck et al. 2015; Huser et al. 2016; Jiang et al. 2016; Romero et al. 2016; Tian et al. 2016; Staehr et al. 2017). Sudden changes in water characteristics, daily fluctuations and long-term trends make it necessary to clean the sensors frequently. Sensor fouling is the main factor that limits the continuous monitoring of water characteristics (Aisopou et al. 2012; Lawlor et al. 2012).

Biofouling affects sensor performance because it decreases sensitivity and interferes with readings in a short period of time. The accurate repetition of readings, as well as their resolution, can be affected, because the main characteristics of the detection will be changed. In the occurrence of biofouling, it is necessary to clean and recalibrate the sensors (Aisopou et al. 2012). The variety of parameters and the influence of biofouling on the reading of each parameter is still a challenge for detection in water quality monitoring (Bhardwaj & Gupta 2015).

Fouling makes it necessary to maintain the sensors. In order for performance to be unaffected and data reading reliable, a cleaning and maintenance routine must be established. Optical sensors are more affected, such as turbidity sensors. Dissolved oxygen and pH sensors are less affected since they are based on electrochemical principles. Conductivity sensors based on resistivity are even less affected. Temperature sensors based on thermoresistance are least affected. The geographic area and the environmental conditions can influence the type of biofouling, and should be taken into account when choosing the sensor maintenance approach (Lawlor et al. 2012).

Several authors (Evans et al. 1997; Ethier & Bedard 2007; Orrico et al. 2007; Kotamäki et al. 2009; Braga 2013; Tew et al. 2014) report monitoring carried out on rivers and bays with intervals between sensor maintenance and/or cleaning ranging from 36 hours to 2 months. The biofouling affects the different sensors in a diverse way. In optical sensors, such as the turbidity sensor, it affects the amount of transmitted light. In electrochemical sensors, such as some dissolved oxygen sensors, it creates a barrier between the membrane and the water.

In situ sensors are subjected to fouling, sediment accumulation, biofilms, and other debris, requiring methods and/or technology that make instrumentation resistant to fouling. A combination of antifouling techniques increases the quality of data in long-term implementations and results in lower maintenance costs (Bringhurst & Adams 2011; Blaen et al. 2016).

There are technologies that reduce the impact of biofouling such as antifouling coatings, automatic wipers or compressed air injection (Lillycrop & Howell 1996; Blaen et al. 2016). To decrease maintenance frequency, several antifouling approaches have been tested. Historically, copper-based paints have been used for many years as antifouling agents for boats, as copper prevents cell division by interfering in cell membrane enzymes. As copper is corroded, the oxidized molecules are released into the water instead of remaining on the metal surface (Depner et al. 2015). These molecules are toxic to microorganisms, but not to humans. Copper considerably reduces fouling and interferes with the growth of microorganisms that favor biofouling (Manov et al. 2003; Flynn et al. 2010). In monitoring, copper guards were developed to investigate whether their use can prevent or delay the formation of biofilm or biofouling (Cowle et al. 2014)

There are several products used to prevent algae growth and biofouling, such as products based on TBT (tributyltin), alcohol sulfates and carbonates of ionic nature, but these are not fully effective. In addition, TBT is a toxic product and should not be used. Other anti-fouling agents, such as cayenne pepper, affect the readings of optical sensors with their coloration. The roughness that these products can create on the surface of the sensors can promote an environment favorable to the creation of biofilms. In northern Japan, in the Pacific Ocean, copper was tested as an antifouling agent and increased the reliable monitoring period from 10 to 60 days (Manov et al. 2003).

Several antifouling materials, mechanical cleaning devices and biocides have been evaluated to decrease sensor maintenance frequency. Biological growth decreases when biocides such as copper and silver are used, whereas mechanical cleaning means have limited success. Examples of long-term implementations may involve using a combination of antifouling techniques (Bringhurst & Adams 2011).

Some antifouling techniques using biocides to coat sensors are no longer usable, because they caused negative environmental impacts and were prohibited. New techniques and antifouling proposals have been developed, such as biomimetic techniques, which mimic natural mechanisms. Nano scales are being used because they benefit both sensor construction and resistance to biofouling. Sensor surfaces constructed with scales and hierarchically organized in varying ranges have been tested, decreasing roughness and thus hindering the deposition of biological materials that favor the growth of the biofilm. Another strategy using nano technology is to use nano structures of cuprous oxide (Cu2O), leading to a greater resistance to long-term biofouling, requiring less copper than in antifouling coatings (Zhuiykov 2012).

The aim of the present work is to evaluate strategies to improve the long-term reliability of water quality sensor readings using copper guards, copper paint, and a cleaning mechanism. Previous studies, discussed above, have looked at one antifouling strategy applied, in most cases, to single parameter probes. Even when multi parameter probes are investigated, parameter-specific results are not reported. Here, three strategies are compared when used with two models of multi parameter probes, and results are reported for each parameter monitored. Copper guards and copper paint coating were selected because they have been tested successfully before (Cowle et al. 2014) and copper paint coating is well established in naval applications. Among mechanical devices, water jets were chosen because they were considered the most practical mechanical cleaning alternative when compared with air injection and wipers.

METHODS

Two multi-parameter probe models were used. Model A probes measured pH, conductivity and dissolved oxygen. Model B probes measured temperature, pH, conductivity, dissolved oxygen, and turbidity. Three probes of each model were available, all used equipment. One probe of each model was always stored clean and used to acquire data to compare with the readings from the continuous monitoring sensors. The probes were calibrated following the procedure described in the manufacturer's manual. Turbidity sensors were optical, while DO and pH sensors were electrochemical. Conductivity sensors were based on resistivity, and temperature sensors based on thermoresistance.

Experiment one – baseline

In order to establish a baseline, experiment one tests consisted of continuous monitoring of tap water stored in translucent 20 L plastic containers over 31 days. Readings were performed weekly. Literature indicates that biofouling may occur from a few days (Braga 2013) to 30 days (Evans et al. 1997; Ethier & Bedard 2007; Kotamäki et al. 2009; Flynn et al. 2010; Burghina et al. 2014; Tew et al. 2014), with only one article reporting 60 days (Orrico et al. 2007).

Experiment two – polluted water

For the second experiment, water was collected from the Areiãozinho stream (25°7′59.9″S 49°13′43.0″W), and kept in two containers of 20 L each during the entire observation period. This stream receives domestic sewage discharges. Weekly readings were performed over 97 days, and four probes were used in continuous monitoring. The conditions were similar to a lentic environment, with the containers stored outdoors and subject to natural conditions of light and temperature.

Experiment three – copper

Experiment three involved tests with copper guards and PVC guards coated with paint mixed with copper powder. Polluted water was used, collected from the same water body described above. Exploratory tests to determine the mix ratio of copper powder and paint revealed that 100 g of powder in 100 mL of enamel paint were appropriate to prevent the coating from detaching from the PVC guard after application.

Experiment four – river

In this experiment, the probes were placed in a natural stream, viz. the Cotia River (25°41′40.8″S 49°12′41.6″W), in the Curitiba Metropolitan Region. Observations were conducted during 60 days. Two model B probes were installed, one with a copper guard and another with a water jet cleaning apparatus mounted on the PVC guard. Water jets were employed to remove waste from the sensor compartment. Readings were performed weekly, and compared with those obtained with a clean probe.

Water was injected into two regions of the sensor chamber as shown in Figure 1. The water leaving the pump is led to the sensors through a plastic hose that directs several jets through small orifices. The pump control, the water ejection period and the interval between the injections was controlled by Arduino. The activation of the two pumps is independent, with water ejection lasting 30 s at alternating positions, every 10 minutes.

Figure 1

Water jet cleaning mechanism; (a) schematic diagram; (b) actual setup; (c) plastic hose with small holes; (d) cleaning water jets.

Figure 1

Water jet cleaning mechanism; (a) schematic diagram; (b) actual setup; (c) plastic hose with small holes; (d) cleaning water jets.

A small protective cage was installed to protect the probes from debris. The Cotia river has an average flow velocity of 0.45 m·s−1 during dry periods, but this can reach 1.25 m·s−1 during rain events. The river bed is sandy, with depths near the monitoring station varying between 0.90 and 1.5 m. The water is usually turbid, with a Secchi depth of approximately 0.60 m. Figure 2 shows the monitoring location, while Figure 3 shows the probes after a 50-day immersion period.

Figure 2

Monitoring location: (a) continuously submerged probes; (b) data acquisition setup; (c) probes after 50-day continuous monitoring with copper guard; and (d) with water jet cleaning mechanism.

Figure 2

Monitoring location: (a) continuously submerged probes; (b) data acquisition setup; (c) probes after 50-day continuous monitoring with copper guard; and (d) with water jet cleaning mechanism.

Figure 3

Model A probes reading ratios (continuous monitoring/clean probe): (a) experiment one, (b) experiment two, (c) experiment three, PVC guard, (d) experiment three, copper guard.

Figure 3

Model A probes reading ratios (continuous monitoring/clean probe): (a) experiment one, (b) experiment two, (c) experiment three, PVC guard, (d) experiment three, copper guard.

RESULTS AND DISCUSSION

During tests, the clean probe was submerged for five minutes and the last reading recorded. Results are presented as ratios between readings from the continuously submerged probes and the clean probes.

Experiment one – baseline

This experiment was conducted to observe how fast sensor biofouling would occur in initially clean water. Model A probes in continuous and discrete (clean) monitoring (Figure 3(a)) provided similar readings for pH and temperature during the full, 1 month, observation period. The ratios for dissolved oxygen (DO) and conductivity, however, show differences starting at the end of the first week, likely indicating rapid fouling of these sensors. A similar situation was observed with model B probes. The abrupt drop in pH readings indicates a malfunction of the sensor, which was not used after this experiment.

The observations made at this stage indicate that although the probes are calibrated according to the manufacturer's recommendations, some sensors may have problems and should be carefully checked before their use for monitoring. Moreover, it was observed that DO sensors undergo rapid contamination, and their readings are compromised in less than a week of continuous immersion, both in initially clean water, and in polluted water as presented next.

Experiment two – polluted water

Changes in turbidity and biofouling on the probes were noticeable, as shown in Figure 4(a) and 4(b). Here, initial differences were observed in DO readings, while readings for the other parameters were close for all probes. A few days after the start, conductivity presented significant differences of reading. Figure 3(b) shows the ratios for Model A probes.

Figure 4

Experiment two: (a) murky water during observation period and (b) probe with biofouling in the body and sensors.

Figure 4

Experiment two: (a) murky water during observation period and (b) probe with biofouling in the body and sensors.

Model A probes provided similar readings for temperature over the entire monitoring period. For pH, the difference initially increased, then decreased and finally became compromised, indicating sensor fouling. Conductivity readings showed differences greater than the other parameters in the continuous monitoring period. Model B probes showed adequate sensor performance during continuous monitoring only for temperature.

Experiment three – copper

Copper guards

One continuous immersion probe of each model was placed in two separate containers, one container kept probes with PVC guards while the other had probes with copper guards. The probes with the copper guard had significantly less fouling than those with PVC guards.

Figure 3(c) shows the ratios between the readings for Model A probes (PVC guard and clean). Differences in temperature readings were under 5% during the entire observation period. pH readings showed an initial difference greater than 15%, but improved and remained within 5% until day 35, exceeding 10% in 11 days, and exceeding 15% in 55 days. Conductivity started with a 13% difference and showed a greater variation at 37 days, exceeding 15%, but improving in another 6 days and remaining below 15% until the end of the period. DO presented differences of more than 15% throughout the period.

Figure 3(d) shows the ratios between the readings for Model A probes (copper guard and clean). Temperature, again, showed good agreement throughout the monitoring period. pH differences became greater than 5% after the sixth week. DO readings, once again, showed large differences. Conductivity showed a 15% difference in the first week. The copper guard seems to have extended the period of continuous monitoring for pH.

Readings from model B probes with PVC guards showed smaller differences for temperature, pH, and conductivity, while DO and turbidity readings revealed larger differences between the continuous monitoring probe and the clean probe.

Readings from model B probe with a copper guard showed that sensor performance in the probe with a copper guard was superior to that of the probe with a PVC guard. Temperature, pH, and conductivity readings showed smaller differences over the period. Initial differences were observed for pH, which improved after the second week. Conductivity ratios stayed below 15% until day 53, while turbidity ratios were below this threshold until day 16. Temperature ratios were below 15% for the entire period.

Copper powder paint mix

The other copper-related strategy tested was the use of a copper powder paint mix coating on a probe's body and PVC guard. Figure 5(a) shows the coated probe at the end of a 77-day monitoring immersed in polluted water. Figure 5(b) shows the ratios of the readings obtained with this solution and with the clean probe. Few differences are observed. The parameter with the highest ratios was turbidity, after 57 days. Among the others, conductivity and DO showed greater differences throughout the period, while pH showed variations after 13 days, greater variations in 48 days, and finished the monitoring period without major differences.

Figure 5

Model B probes with PVC guard coated with copper paint (a) after a 77-day continuous monitoring period, (b) reading ratios (copper paint coating continuous monitoring/clean probe), experiment three.

Figure 5

Model B probes with PVC guard coated with copper paint (a) after a 77-day continuous monitoring period, (b) reading ratios (copper paint coating continuous monitoring/clean probe), experiment three.

Experiment four – river

In this experiment, only Model B probes were used, since they allow monitoring of more parameters. Two probes were used, one with a copper guard and one with a PVC guard and a water jet cleaning mechanism. The results are different from those observed in the previous experiments, since the monitored environment was lotic, allowing for less accumulation of nutrients and microorganisms on the probe surface. Thus, no biofouling was noticeable, but only the accumulation of clay and debris. Figure 2 shows the probes after 50 days in the river.

Figure 6(a) shows the ratios of the readings obtained with the probe with the copper guard and those obtained with the clean probe. Turbidity is not shown – it had a difference of approximately 12% in the first 10 days, increasing to more than 15% and then the sensor no longer worked. Temperature ratios improve after the first few days, but the initial difference is likely due to the sensor not having reached the external temperature at the time of the first reading. pH and conductivity show differences at the start and end the monitoring period above 15%. DO ratios fluctuate – in 10 days they reach 15%, decreasing after 14 days to a little more than 5%, continuing so for 28 days, and decreasing to less than 5% at the end of the monitoring period.

Figure 6

Model B probes reading ratios, experiment four: (a) copper guard continuous monitoring/clean probe, (b) cleaning mechanism continuous monitoring/clean probe.

Figure 6

Model B probes reading ratios, experiment four: (a) copper guard continuous monitoring/clean probe, (b) cleaning mechanism continuous monitoring/clean probe.

Figure 6(b) shows the ratios of the readings obtained with the probe with the cleaning mechanism and those obtained with the clean probe. Turbidity is not shown, since ratios were large and quickly extrapolated the axis scale reaching nearly 200. Comparing Figure 6(a) and 6(b), temperature measurements are similar, while for the other parameters the probe with the copper guard seems to perform slightly better.

Tables 1 and 2 summarize the results of all experiments performed, indicating the observed parameters and the respective periods of similar readings between the continuously immersed probes and the clean probes. Differences of 5%, 10% and 15% were considered to define the additional period of reliability provided by the solutions tested in experiments 3 and 4.

Table 1

Experiments, probe types, parameters and period of similar (maximum 15% difference) readings between continuous monitoring probes and clean probes

Experiment (duration, days)Probe model (medium)Parameters monitoredPeriod of similar readings (up to 15% difference)
1 (31) A (20 L tap water) Temperature Entire period 
pH Entire period 
Conductivity 14 days 
Dissolved oxygen 7 days 
1 (31) B (20 L tap water) Temperature Entire period 
pH Entire period 
Conductivity 7 days 
Dissolved oxygen 7 days 
Turbidity 14 days 
2 (97) A (polluted water 20 L) Temperature Entire period 
pH 55 days 
Conductivity 21 days 
Dissolved oxygen 21 days 
2 (97) B (polluted water 20 L) Temperature Entire period 
pH 55 days 
Conductivity 23 days 
Dissolved oxygen 36 days 
Turbidity 42 days 
Experiment (duration, days)Probe model (medium)Parameters monitoredPeriod of similar readings (up to 15% difference)
1 (31) A (20 L tap water) Temperature Entire period 
pH Entire period 
Conductivity 14 days 
Dissolved oxygen 7 days 
1 (31) B (20 L tap water) Temperature Entire period 
pH Entire period 
Conductivity 7 days 
Dissolved oxygen 7 days 
Turbidity 14 days 
2 (97) A (polluted water 20 L) Temperature Entire period 
pH 55 days 
Conductivity 21 days 
Dissolved oxygen 21 days 
2 (97) B (polluted water 20 L) Temperature Entire period 
pH 55 days 
Conductivity 23 days 
Dissolved oxygen 36 days 
Turbidity 42 days 
Table 2

Experiments, probe types, parameters and period of similar readings between continuous monitoring probes and clean probes

Experiment (duration – days)Probe model (medium)Parameters monitoredPeriod of similar readings with clean probe (up to 15% difference)Additional period provided by the best solution in relation to the worst
The percentage indicates the difference between readings with the tested solution and the clean probe
5%10%15%
3 (90) A (polluted water 20 L) Temperature Entire period    
pH Entire period    
Conductivity 21 days (PVC)/46 days (copper) 20 days (copper) 95% 22 days (copper) 105% 25 days (copper) 119% 
Dissolved oxygen 16 days (PVC)/50 days (copper) 21 days (copper) 131% 25 days (copper) 156% 34 days (copper) 213% 
3 (90) B (polluted water 20 L) Temperature Entire period    
pH Entire period    
Conductivity 36 days (PVC)/65 days (copper) 14 days (copper) 39% 18 days (copper) 50% 29 days (copper) 81% 
Dissolved oxygen 30 days (PVC)/53 days (copper) 12 days (copper) 40% 19 days (copper) 63% 23 days (copper) 77% 
Turbidity 8 days (PVC)/15 days (copper) 4 days (copper) 50% 6 days (copper) 75% 7 days (copper) 88% 
3 (60) B (copper powder paint in polluted water 20 L) Temperature Entire period    
pH Entire period 28 days (paint) Entire period (paint) Entire period (paint) 
Conductivity 36 days (PVC)/37 days (paint) 28 days (paint) 78% 34 days (paint) 94% 37 days (paint) 103% 
Dissolved oxygen 30 days (PVC)/60 days (paint) 20 days (paint) 67% 46 days (paint) 153% 60 days (paint) 200% 
Turbidity 8 days (PVC)/57days (paint) 55 days (paint) 688% 56 days (paint) 700% 57 days (paint) 713% 
4 (60) B (river) Temperature Entire period    
pH 10 days (water jet)/16 days (copper) 2 days (copper) 20% 4 days (copper) 40% 6 days (copper) 40% 
Conductivity 60 days (water jet)/32 days (copper) 28 days (water jet) 88% 28 days (water jet) 88% 28 days (water jet) 88% 
Dissolved oxygen 40 days (water jet)/11 days (copper) 6 days (water jet) 55% 25 days (water jet) 227% 29 days (water jet) 264% 
Turbidity 11 days (water jet)/8 days (copper) 1 day (water jet) 13% 2 days (water jet) 25% 3 days (water jet) 38% 
Experiment (duration – days)Probe model (medium)Parameters monitoredPeriod of similar readings with clean probe (up to 15% difference)Additional period provided by the best solution in relation to the worst
The percentage indicates the difference between readings with the tested solution and the clean probe
5%10%15%
3 (90) A (polluted water 20 L) Temperature Entire period    
pH Entire period    
Conductivity 21 days (PVC)/46 days (copper) 20 days (copper) 95% 22 days (copper) 105% 25 days (copper) 119% 
Dissolved oxygen 16 days (PVC)/50 days (copper) 21 days (copper) 131% 25 days (copper) 156% 34 days (copper) 213% 
3 (90) B (polluted water 20 L) Temperature Entire period    
pH Entire period    
Conductivity 36 days (PVC)/65 days (copper) 14 days (copper) 39% 18 days (copper) 50% 29 days (copper) 81% 
Dissolved oxygen 30 days (PVC)/53 days (copper) 12 days (copper) 40% 19 days (copper) 63% 23 days (copper) 77% 
Turbidity 8 days (PVC)/15 days (copper) 4 days (copper) 50% 6 days (copper) 75% 7 days (copper) 88% 
3 (60) B (copper powder paint in polluted water 20 L) Temperature Entire period    
pH Entire period 28 days (paint) Entire period (paint) Entire period (paint) 
Conductivity 36 days (PVC)/37 days (paint) 28 days (paint) 78% 34 days (paint) 94% 37 days (paint) 103% 
Dissolved oxygen 30 days (PVC)/60 days (paint) 20 days (paint) 67% 46 days (paint) 153% 60 days (paint) 200% 
Turbidity 8 days (PVC)/57days (paint) 55 days (paint) 688% 56 days (paint) 700% 57 days (paint) 713% 
4 (60) B (river) Temperature Entire period    
pH 10 days (water jet)/16 days (copper) 2 days (copper) 20% 4 days (copper) 40% 6 days (copper) 40% 
Conductivity 60 days (water jet)/32 days (copper) 28 days (water jet) 88% 28 days (water jet) 88% 28 days (water jet) 88% 
Dissolved oxygen 40 days (water jet)/11 days (copper) 6 days (water jet) 55% 25 days (water jet) 227% 29 days (water jet) 264% 
Turbidity 11 days (water jet)/8 days (copper) 1 day (water jet) 13% 2 days (water jet) 25% 3 days (water jet) 38% 

Percent values indicate additional period of reliability achieved with respect to the shorter alternative (underlined).

Table 3 presents a comparison between periods of reliability reported in the literature and the periods reached during the experiments performed using the proposed solutions. Literature reports indicate periods of continuous monitoring of multiparameter probes, without differentiating between the monitored parameters. The period of reliability reported in most of the literature (Ethier & Bedard 2007; Flynn et al. 2010; Burghina et al. 2014; Tew et al. 2014) for multiparameter probes varies from 10 days to 1 month before needing maintenance. There is only one report of 2 months with reliable readings (Orrico et al. 2007). In the experiments performed during the present study, the multiparameter probes had a reliable period of up to 65 days. In relation to the turbidity probe, literature reports start from 36 hours (Braga 2013) and reach a month (Kotamäki et al. 2009). Here, turbidity readings were deemed reliable until between 8 and 57 days, depending on the solution tested. As for the monitoring periods reported in the literature, no anti-incrustant technique or mechanism was employed.

Table 3

Comparison between periods reported in the literature and periods with antifouling alternatives tested

Probe typeCleaning frequency reported in the literatureCopper guard probeWater injection probeProbe with copper powder paint
Multiparameter Lee River, Cork, Days, (Flynn et al. 2010);
Port of Dublin, 15 days in winter and 10 days in spring (Burghina et al. 2014);
Nanuan Bay to the south of Thailand, 2 weeks (Tew et al. 2014);
Fraser River Estuary, 1 month (Ethier & Bedard 2007);
Chesapeake and Yaquina Bays, 2 months (Orrico et al. 2007
From 11 to 65 days From 10 to 60 days From 37 to 60 days 
Turbidity Barigui River basin, 36 hours, (Braga 2013);
Rivers Swale and Trent, England, 1 week, (Evans et al. 1997);
Japan's North Pacific, 10 days (Manov et al. 2003);
Basin in southern Finland, 1 month in winter and 1 week in summer (Kotamäki et al. 2009
From 8 to 15 days
(Evans et al. 1997; Manov et al. 2003; Kotamäki et al. 2009; Braga 2013
9 days 57 days 
Probe typeCleaning frequency reported in the literatureCopper guard probeWater injection probeProbe with copper powder paint
Multiparameter Lee River, Cork, Days, (Flynn et al. 2010);
Port of Dublin, 15 days in winter and 10 days in spring (Burghina et al. 2014);
Nanuan Bay to the south of Thailand, 2 weeks (Tew et al. 2014);
Fraser River Estuary, 1 month (Ethier & Bedard 2007);
Chesapeake and Yaquina Bays, 2 months (Orrico et al. 2007
From 11 to 65 days From 10 to 60 days From 37 to 60 days 
Turbidity Barigui River basin, 36 hours, (Braga 2013);
Rivers Swale and Trent, England, 1 week, (Evans et al. 1997);
Japan's North Pacific, 10 days (Manov et al. 2003);
Basin in southern Finland, 1 month in winter and 1 week in summer (Kotamäki et al. 2009
From 8 to 15 days
(Evans et al. 1997; Manov et al. 2003; Kotamäki et al. 2009; Braga 2013
9 days 57 days 

CONCLUSIONS

In the scope of the present work, different mechanisms were tested that led to different results. The tested solutions indicated that the characteristics of the monitored environment should be observed. Copper and copper powder experiments led to better results in a lentic environment, while water injection experiments led to better results in a lotic environment.

Most literature reports of continuous water quality monitoring mention multiparameter probes, in general without reference to specific parameters. When a specific parameter is mentioned, it is usually turbidity. In the experiments reported here, the observations were specific by parameter, establishing specific reliability periods for each individual parameter.

The analysis of strategies to increase the reliability of continuous water quality monitoring showed that the use of unmodified probes leads to reliable monitoring periods similar to those reported in the literature, ranging between 8 and 36 days. With the use of a copper guard, the reliability period was extended in 34 days. The efficacy of this strategy was verified in a system similar to a lentic environment. In a lotic environment, the same efficacy was not observed, and a cleaning mechanism led to better results, extending the period of reliability in 29 days in relation to the copper guard solution.

In addition to the copper guard, an alternative was tested using a copper powder – enamel paint mix. The efficacy of this alternative in a lentic environment was verified, leading to reliable results between 28 and 60 days.

In practical applications, a broader strategy is suggested, with the association between copper guard or copper-paint coating, and a water jet cleaning mechanism to improve the long-term reliability of sensor readings.

REFERENCES

REFERENCES
Albert
S.
Fisher
P. L.
Gibbes
B.
Grinham
A.
2015
Corals persisting in naturally turbid waters adjacent to a pristine catchment in Solomon Islands
.
Marine Pollution Bulletin
94
(
1
),
299
306
.
Aloui
B. Z.
Gueddari
M.
2009
Long-term water quality monitoring of the Sejnane reservoir in northeast Tunisia
.
Bulletin of Engineering Geology and the Environment
68
,
307
316
.
Behmel
S.
Damour
M.
Ludwig
R.
Rodriguez
M. J.
2016
Water quality monitoring strategies – a review and future perspectives
.
Science of The Total Environment
571
,
1312
1329
.
Bhardwaj
J.
Gupta
K. K.
2015
A Review of Emerging Trends on Water Quality Measurement Sensors
. In:
India 2015 International Conference on Technologies for Sustainable Development (ICTSD-2015)
,
Feb. 04–06
,
Mumbai, India
.
Blaen
P. J.
Khamis
K.
Lloyd
C. E. M.
Bradley
C.
Hannah
D.
Krause
S.
2016
Real-time monitoring of nutrients and dissolved organic matter in rivers: capturing event dynamics, technological opportunities and future directions
.
Science of The Total Environment
569–570
,
647
660
.
Bowes
M. J.
Smith
J. T.
Neal
C.
Leach
D. V.
Scarlett
P. M.
Wickham
H. D.
Harman
S. A.
Armstrong
L. K.
Davy-Bowker
J.
Haft
M.
Davies
C. E.
2011
Changes in water quality of the River Frome (UK) from 1965 to 2009: is phosphorus mitigation finally working?
Science of The Total Environment
409
(
18
),
3418
3430
.
Braga
S. M.
2013
A new Approach for Integration Between Quantity and Quality of Water for Assessment of Diffuse Pollution
.
Dissertation
,
Universidade Federal do Paraná, Setor de Tecnologia, Curso de Pós-Graduação em Engenharia de Recursos Hídricos e Ambiental
,
Curitiba
,
196 f . il
.
Bringhurst
B.
Adams
J.
2011
Innovative Sensor Design for Prevention of bio-Fouling
, pp.
1
8
.
doi:10.23919/OCEANS.2011.6106931
.
Burghina
C. B.
Sullivan
T.
Chapman
J.
Regan
F.
2014
Continuous high-frequency monitoring of estuarine water quality as a decision support tool: a Dublin Port case study
.
Environment Monitoring and Assessment
186
,
5561
5580
.
Coelho
M.
2013
Strategy for Monitoring Water Quality for Water Resources Management in Urban Basins
.
Thesis
,
Universidade Federal do Paraná, Setor de Tecnologia, Programa de Pós-graduação em Recursos Hídricos e Ambiental
,
Curitiba, Brazil
.
Cowle
M. W.
Babatunde
A. O.
Rauen
W. B.
Evans B.
B. N.
Barton
A. F.
2014
Biofilm development in water distribution and drainage systems: dynamics and implications for hydraulic efficiency
.
Environmental Technology Reviews
3
(
1
),
31
47
.
Depner
R. F. R.
Depner
R. A.
Lucca
V.
Lovato
M.
2015
O cobre como superfície de contato antimicrobiana e sua potencial aplicação na medicina veterinária
.
Veterinária E Zootecnia
22
(
4
),
532
543
.
Ethier
A.
Bedard
J.
2007
Development of a Real-Time Water Quality Buoy for the Fraser River Estuary
. In:
Proceedings, International Symposium on OCEAN
, pp.
1
6
.
doi:10.1109/OCEANS.2007.4449424
.
Evans
J. G.
Wass
P. D.
Hodgson
P.
1997
Integrated continuous water quality monitoring for the LOIS river syndromme
.
Science of The Total Environment
194
,
111
118
.
Flynn
B. O.
Regan
F.
Lawlor
A.
Wallace
J.
Torres
J.
O'mathuna
C.
2010
Experiences and recommendations in deploying a real-time, water quality monitoring system
.
Measurement Science and Technology
21
(
12
).
Glasgow
H. B.
Burkholder
J. M.
Reed
R. E.
Lewitus
A. J.
Kleinman
J. E.
2004
Real-time remote monitoring of water quality: a review of current applications, and advancements in sensor, telemetry, and computing technologies
.
Journal of Experimental Marine Biology and Ecology
300
(
1–2
),
409
448
.
Halliday
S. J.
Wade
A. J.
Skeffington
R. A.
Neal
C.
Reynolds
B.
Rowland
P.
Neal
M.
Norris
D.
2012
An analysis of long-term trends, seasonality and short-term dynamics in water quality data from Plynlimon, Wales
.
Science of The Total Environment
434
,
186
200
.
Holeck
K.
Rudstam
L. G.
Watkins
J. M.
Luckey
F.
Lantry
J. R.
Lantry
B. F.
Trometer
B.
Koops
M.
Johnson
T.
2015
Lake Ontario water quality during the 2003 and 2008 intensive field years and comparison with long-term trends
.
Aquatic Ecosystem Health & Management
18
,
7
17
.
Jiang
B.
Chen
J.
Luo
Q.
Lai
J.
Xu
H.
Wang
Y.
Yu
K.
2016
Long-term changes in water quality and eutrophication of China's Liujiang River
.
Polish Journal of Environmental Studies
25
(
3
),
1033
1043
.
Knapik
H. G.
2009
Reflections on Monitoring, Modeling and Calibration in Water Resources Management: Case Study of Water Quality in the Upper Iguaçu Basin
.
Thesis
,
Universidade Federal do Paraná, Setor de Tecnologia, Curso de Pós-Graduação em Engenharia de Recursos Hídricos e Ambiental
,
Curitiba
.
197 f. :il
.
Kotamäki
N.
Thessler
S.
Koskiaho
J.
Hannukkala
A. O.
Huitu
H.
Huttula
T.
Havento
J.
Järvenpää
M.
2009
Wireless in-situ sensor network for agriculture and water monitoring on a river basin scale in Southern Finland: evaluation from a data user's perspective
.
Sensors
9
,
2862
2883
.
Lawlor
A.
Torres
J.
Brendan O'Flynn
B.
Wallace
J.
Regan
F.
2012
DEPLOY: a long term deployment of a water quality sensor monitoring system
.
Sensor Review
32
(
1
),
29
38
.
Lillycrop
L. S.
Howell
G. L.
1996
In-situ Long-Term Deployment of Water Quality Sensors Adversely Affected by Biological Fouling
.
US Army Engineer Waterways Experiment Station Coastal Engineering Research Center
,
3909 Halls Ferry Rd Vicksburg, MS, 39180, USA
.
Orrico
M. C.
Moore
C.
Romankol
D.
Derr
A.
Bernard
A. H.
Janzen
C.
Larson
N.
Murphy
D.
Johnson
J.
Bauman
J.
2007
WQM: A New Integrated Water Quality Monitoring Package for Long-Term In-Situ Observation of Physical and Biogeochemical Parameters
. In:
OCEANS 2007
,
Vancouver, BC
, pp.
1
9
.
doi:10.1109/OCEANS.2007.4449418
.
Rao
A. S.
Marshall
S.
Gubbi
J.
Palaniswami
M.
Sinnott
R.
Pettigrovet
V.
2013
Design of low-cost autonomous water quality monitoring system 2013
. In:
International Conference on Advances in Computing, Communications and Informatics (ICACCI)
,
Mysore
, pp.
14
19
.
doi:10.1109/ICACCI.2013.6637139
.
Rocha
Z. M. D.
2009
Micro Autonomous Laboratories to Monitor Water Quality Parameters
.
Tese (Doutorado)
,
Escola Politécnica da Universidade de São Paulo. Departamento de Engenharia de Sistemas Eletrônicos
,
São Paulo
, p.
171
.
Romero
E.
Gendre
R. L.
Garnier
J.
Billen
G.
Fisson
C.
Silvestre
M.
Riou
P.
2016
Long-term water quality in the lower Seine: lessons learned over 4 decades of monitoring
.
Environmental Science & Policy
58
,
141
154
.
Tew
K. S.
Leu
M. Y.
Wang
J. T.
Chang
C. M.
Chen
C. C.
Meng
P. J.
2014
A continuous, real-time water quality monitoring system for the coral reef ecosystems of Nanwan Bay, Southern Taiwan
.
Marine Pollution Bulletin
85
(
2
),
641
647
.
Tian
S.
Youssef
M. A.
Richards
R. P.
Liu
J.
Baker
D. B.
Liu
Y.
2016
Different seasonality of nitrate export from an agricultural watershed and an urbanized watershed in Midwestern USA
.
Journal of Hydrology
541
,
1375
1384
.
Wong
B. P.
Kerkez
B.
2016
Real-time environmental sensor data: an application to water quality using web services
.
Environmental Modelling & Software
84
,
505
517
.
Xu
J.
Lee
J. H. W.
Yin
K.
Liu
H.
Harrison
P. J.
2011
Environmental response to sewage treatment strategies: Hong Kong's experience in long term water quality monitoring
.
Marine Pollution Bulletin
62
(
11
),
2275
2287
.