Evaluating the necessity and usefulness of water filters for domestic tap water

My above project has finally concluded after months of experimental work by my project team.

Due to the popularity of water filters in the market these past years, I thought it will be interesting to find out more about them, especially if they are necessary and useful.

Filter 1












Figure: Home-made gravity filter using Doulton Supersterasyl filter candle (the white thingy in the middle) - ceramic (diatomaceous earth) filter impregnated with silver and incorporating a granular carbon core

Filter 2

Figure: Pressure filter train utilising standard 10-in filter housing. Order of filtration from input to output: pleated 5-um, twisted yarn 1-um, granular activated carbon (GAC), ion-exchange, fluoride reduction, pleated 0.5-um, pleated 0.35-um

Tap water was sampled from 4 different tertiary institutions well distributed spatially in Singapore. It was then allowed to run through the 2 types of filters. The raw tap water, filtered water from filter 1 and filtered water from filter 2 were tested for the following parameters - turbidity, pH, conductivity, nitrate, fluoride, total hardness, chloride, sulphate, phosphate, iron, manganese, copper, aluminium, total chlorine, aerobic bacteria count, lead. Most of the above parameters are similar to the routine testing done by PUB.

Figure: Portable water quality with typical PUB values and WHO guidelines

Tap Water Results
From the above table, you may have noticed that most of the PUB's values are well below the limits under WHO guidelines. Well folks, good news! Most of our results lie within PUB's values so PUB is not kidding in providing quality tap water. (However, WHO guidelines are not omnipotent as you may have read in my previous post - water filters in Singapore.)

For the astute reader, you may have noticed the word "most". For 2 institutions, copper content in tap water is at 0.21mg/L and 0.15mg/L, respectively - much higher than the PUB value of 0.05mg/L. In case you start jumping to the wrong conclusion, don't forget that the values above only apply to water coming of PUB's waterworks. The water still has to travel through miles of piping through the distribution network and store at various locations along the way before ending up at your tap. Lots of things can go wrong - leaking lines, corroded piping, poorly maintained tanks etc. Older buildings probably have corroding copper pipes which jack up the copper content in tap water.

The other parameter of interest is fluoride. Surprisingly, 2 of the institutions have tap water containing only 0.1mg/L fluoride while the other 2 just meet the minimum of 0.4mg/L as given by PUB. Officially, fluoride is added (yes, it is not a contaminant but is intentionally added) into water to fight tooth decay. Are we getting enough fluoride in our water for it to do its job? On the other side of the story, critics have fought against fluoridation at any level as we are already getting enough fluoride in our diet and taking in more can compromise our health - see fluorosis.

The next big thing in water is chlorine. Relax... our water is not overchlorinated. Instead, the tap water results are in the range of 0.05 - 0.15mg/L. Different sources cite different values for adequate chlorination at the end user, ranging from 0.2 - 0.5mg/L or even more. Our results are rather low so the question is whether our tap water is adequately disinfected. As you would have guessed, there is another side of the story - consuming chlorinated water has been linked to several cancers and heart disease. It is also inked to the production of trihalomethanes (THM) and haloacetic acids (HAA) in water, both of which are likely carcinogens. Hence, a low level of chlorine in tap water may actually be good news.

2 institutions have very high levels of bacteria but don't be alarmed. We are measuring aerobic bacteria and not an indicator bacteria like faecal coliform or Enterococcus. Of course, the older the distribution network and the poorer maintained it is, you are likely to see more bugs in your tap water. Actually, the purpose of measuring aerobic bacteria is to gauge the effectiveness of filtration by comparing the values before and after filtration.


Filtration Results
Frankly, because our tap water is so clean, the filters we tested could not perform as well as they could have or were certified for. Certification e.g. NSF/ANSI standard 42 or 53 requires a certain challenge inflow concentration. If the outflow concentration falls by a certain percentage, say 90%, the device has passed the test. But if the inflow concentration is already very low, a 90% reduction is not likely to be achieved.

The gravity filter appears significantly more capable at removing chlorine (25%), iron (92%), manganese (34%) and aluminium (55%) while the pressure filter seems better at reducing nitrate (42%), fluoride (58%), hardness (35%), chloride (40%), copper (75%) and bacteria (> 94%).

However, sulphate does not appear to be significantly reduced by either filter (< 7%).


Final thoughts
Are water filters necessary?

If you trust WHO guidelines and PUB seems to be doing a good job at providing quality water consistently, then no, you probably don't need them.

But if you feel that for whatever reason, WHO guidelines (or PUB typical values) are not adequate at safeguarding your health, by all means get a filter but know that to reduce the contaminant level even further, a filter probably won't cut it. You probably will need reverse osmosis, distillation or air-to-water units to give you the next level of water purity. In any case, it is good to be sure of exactly what you are concerned about e.g. chlorine, fluoride. Better still, send samples of your tap water to a commercial lab for analysis of that particular concern. AND submit for analysis another sample after going through your preferred treatment. You should then be able to have a good idea if your treatment unit is working up to your expectations.

Of course, you may suspect that there is something wrong with the distribution network. My advice is the same - be clear of your concern, analyse your water and get a treatment unit only if necessary, preferrably with a follow-up analysis of your treated water.

Water quality monitoring course coming up: 1 Jun 2011

The next run of my water quality monitoring (WQM) course will be held on 1 Jun 2011. Formerly only offered to MOE teachers and staff, it is now open to members of the public too. (This came about because of the various requests from non-MOE staff and nature lovers who want to participate in the course as well.)

WQM involves checking the health of waterways on a regular basis. Whether in natural streams or urban canals, most of the water ends up in our reservoirs which provide our drinking water. Hence, WQM plays a vital role in safeguarding our water supplies.

If you are interested, you can find out more about the fascinating world of WQM in this course. Teachers and MOE staff, as well as members of the public can sign up here. For more details on the course, click

Course title: My forest is DYING and the role of water quality

Date: 1 June 2011
Time: 0900 - 1700
Venue: Singapore Polytechnic and field location
No. of participants: 20

More details can be found here. Write-up of previous runs may be found here - 1, 2. (My other WQM courses for students can be found here - 1, 2)

The things that nobody told you about Pasir Ris

The water quality off Pasir Ris Park has piqued my interest since it has been in the limelight for failing the Enterococcus test for primary contact activities (e.g. swimming) for 3 years in a row (2008 - 2010) and the mass fish kill in Dec 09/ Jan 10 (see previous posts). I also found out that further back in Dec 06/ Jan 07 (1, 2, 3), another mass die-off has been observed and the reason was suggested to be a sharp drop in salinity (Update: In her blog, Ria Tan questioned whether Chek Jawa will survive the incessant rain in Jan 2011). Therefore in 2010, I decided to get my team to test out the water quality off Pasir Ris at 2 patches of seagrasses - 1 close to Sg Loyang and the other close to Sg Tampines.

Figure: Patch of seagrass near Sg Loyang

Figure: Another patch of seagrasses near Sg Tampines

Fast forward to the present, and we have some interesting results to share, not just on Pasir Ris but also on the other 2 locations we monitored. Evidently, Enterococcus is not the only thing we have to worry about.

Data were obtained from sampling seawater and sediments in three selected seagrass areas, namely Chek Jawa (CJ), Pasir Ris (PR) Park and Pulau Semakau (PS) from June to December 2010. Sampling dates and times corresponded to tide levels of 0 to 0.5m for safety reasons and ease of seagrass observation. Our team made 6 visits to Chek Jawa (2 sites), 6 visits to Pasir Ris Park (2 sites) and 2 visits to Pulau Semakau (3 sites) during the monitoring period. (The smaller number of trips to Pulau Semakau was due to logistical constraints hence the team could only head there together with Team Seagrass on their routine seagrass monitoring sessions.)


Our on-site tests include air temperature, water temperature, pH, dissolved oxygen (DO), electrical conductivity (EC), turbidity and nutrients in water - ammonia (NH3), nitrate (NO3-) and phosphate (PO43-). Laboratory testing of water samples involves Escherichia coli (E. coli), chemical oxygen demand (COD), total alkalinity, total hardness, calcium and magnesium hardness, metal ions (aluminium, copper, iron and lead) and chlorophyll a. Sediment samples were also brought back to the laboratory to analyse for nutrient content - ammonia (NH3), nitrate (NO3-) and phosphate (PO43-).



Summary of results

The values below were obtained by averaging all the results for each location. For ease of reading, only the notable water quality parameters are highlighted.

* ND = not detected



DO (mg/L) 10.5 (CJ), 8.79 (PS), 5.85 (PR)
DO is significantly lower at Pasir Ris. While considered acceptable compared to waters elsewhere in Singapore, these DO levels were relatively low for a seagrass area which is supposed to photosynthesise and produce oxygen in the day.


EC (mS/cm) 43.9 (CJ), 46.0 (PS), 42.7 (PR)
The average EC appeared consistently healthy for all 3 areas except for Chek Jawa in November 2010 for which the EC plummeted to 24.67 mS/cm at site 2.



Turbidity (NTU) 30.6 (CJ), 18.9 (PS), 262 (PR)
The water at Pasir Ris is significantly more turbid. Turbidity measures the loss in transparency in water due to the presence of suspended particles such as phytoplankton and silt. Fishes start to experience stress when exposed to a turbidity of 10-100 NTU for several hours.


E. coli (CFU/ml) 17.3 (CJ), 2.17 (PS), 23.4 (PR)
Though E. coli is a different bacterium from Enteroccus, it too indicates the presence of faecal contamination, just like Enteroccus.

Pasir Ris and Chek Jawa are significantly higher in E. coli count compared to Pulau Semakau.

E. coli has the highest count at 76 CFU (colony forming unit)/mL in November for Pasir Ris site 2. United States Environmental Protection Agency (USEPA) standards provide a maximum of 10 CFU/mL faecal coliforms for primary contact activities e.g. swimming, and a maximum of 50 CFU/mL for secondary contact activities e.g. boating.


COD (mg/L) 135 (CJ), 136 (PS), 132 (PR)
Surprisingly, the organic matter content in all 3 areas was rather similar and compares well with values obtained in our coastal areas in 2009.


Iron (mg/L) 0.998 (CJ), 0.314 (PS), 1.34 (PR)
Though not considered a threat to marine life, Pasir Ris has the highest iron concentration.


Aluminium (mg/L) 0.612 (CJ), 0.0148 (PS), 1.64 (PR)
Pasir Ris has significantly higher concentration of aluminium. The toxicity of aluminium depends to a large extent on the water pH, with the most toxic aluminium compound forming at pH 5.2-5.8. At pH 7.0, a concentration as low as 0.52 mg/L was found to reduce the growth of rainbow trout. In another example, a mass kill of maraena and peled fry occurred at 0.3 mg/L at a pH of 7.0-7.5. Seawater contains an average of 0.01 mg/L aluminium.


Lead (mg/L) 0.181 (CJ), 0.0428 (PS), ND* (PR)
Superb! Pasir Ris water has no lead detected!


Chlorophyll a (µg/L) No data (CJ), No data (PS), 7.59 (PR)
Though not shown in the averaged result above, chlorophyll a concentrations peaked at 25.1 µg/L at site 2 and 19.2 µg/L at site 1 for Pasir Ris in December. This placed the water within eutrophic condition as given under the guidelines for USA (10-500 µg/L). Eutrophic water has excessive nutrients which support an abundance of algae releasing chlorophyll a.

Nutrient in Water

Ammonia (mg/L-N) 0.174 (CJ), 0.178 (PS), 0.345 (PR)
Pasir Ris has significantly higher concentration of ammonia. Ammonia is toxic by itself. It can also contribute to eutrophication.


Nitrate (mg/L-N) 0.103 (CJ), 0.0875 (PS), 0.242 (PR)
Pasir Ris has significantly higher concentration of nitrate. Nitrate can contribute to eutrophication. Concentrations of 0.1-0.75 mg/L nitrate + ammonia N in fresh waters or even less in saline waters have been enough for a bloom.


Phosphate (mg/L-P) ND* (CJ), ND* (PS), 0.103 (PR)
Pasir Ris has significantly higher concentration of phosphate-phosphorus. Like nitrogen, phosphorus can contribute to eutrophication. A level of 0.005-0.5 mg/L P is already enough to cause phytoplankton blooms.