Bacterial, viral and turbidity removal by intermittent slow sand filtration for household use in developing countries: Experimental investigation and modeling

TitleBacterial, viral and turbidity removal by intermittent slow sand filtration for household use in developing countries: Experimental investigation and modeling
Publication TypeJournal Article
Year of Publication2011
AuthorsJenkins, MW, Tiwari, SK, Darby, J
JournalWater Research
Pagination6227 - 6239
Date Published11/2011
Keywordsbacteria removal, biosand filter, drinking water treatment, head, ms2 bacteriophage, pause period, point-of-use, residence time, sand size, turbidity, virus removal
AbstractA two-factor three-block experimental design was developed to permit rigorous evaluation
and modeling of the main effects and interactions of sand size (d10 of 0.17 and 0.52 mm)
and hydraulic head (10, 20, and 30 cm) on removal of fecal coliform (FC) bacteria, MS2
bacteriophage virus, and turbidity, under two batch operating modes (‘long’ and ‘short’) in
intermittent slow sand filters (ISSFs). Long operation involved an overnight pause time
between feeding of two successive 20 L batches (16 h average batch residence time (RT)).
Short operation involved no pause between two 20 L batch feeds (5 h average batch RT).
Conditions tested were representative of those encountered in developing country field
settings. Over a ten week period, the 18 experimental filters were fed river water
augmented with wastewater (influent turbidity of 5.4e58.6 NTU) and maintained with the
wet harrowing method. Linear mixed modeling allowed systematic estimates of the
independent marginal effects of each independent variable on each performance outcome
of interest while controlling for the effects of variations in a batch’s actual residence time,
days since maintenance, and influent turbidity. This is the first study in which simultaneous
measurement of bacteria, viruses and turbidity removal at the batch level over an
extended duration has been undertaken with a large number of replicate units to permit
rigorous modeling of ISSF performance variability within and across a range of likely filter
design configurations and operating conditions.
On average, the experimental filters removed 1.40 log fecal coliform CFU (SD 0.40 log,
N ¼ 249), 0.54 log MS2 PFU (SD 0.42 log, N ¼ 245) and 89.0 percent turbidity (SD 6.9 percent,
N ¼ 263). Effluent turbidity averaged 1.24 NTU (SD 0.53 NTU, N ¼ 263) and always remained
below 3 NTU. Under the best performing design configuration and operating mode (fine
sand, 10 cm head, long operation, initial HLR of 0.01e0.03 m/h), mean 1.82 log removal of
bacteria (98.5%) and mean 0.94 log removal of MS2 viruses (88.5%) were achieved.
Results point to new recommendations regarding filter design, manufacture, and
operation for implementing ISSFs in local settings in developing countries. Sand size
emerged as a critical design factor on performance. A single layer of river sand used in this
investigation demonstrated removals comparable to those reported for 2 layers of crushed
sand. Pause time and increased residence time each emerged as highly beneficial for
improving removal performance on all four outcomes. A relatively large and significant
negative effect of influent turbidity on MS2 viral removal in the ISSF was measured in parallel with a much smaller weaker positive effect of influent turbidity on FC bacterial
removal. Disturbance of the schmutzdecke by wet harrowing showed no effect on virus
removal and a modest reductive effect on the bacterial and turbidity removal as measured
7 days or more after the disturbance. For existing coarse sand ISSFs, this research indicates
that a reduction in batch feed volume, effectively reducing the operating head and
increasing the pore:batch volume ratio, could improve their removal performance by
increasing batch residence time.
Short TitleWater Research
One Liner: 
BSF's work best at removing bacteria and viruses when they have lower loading heads, lower maximum flow rates, finer sand, and longer pause periods.
CAWST Analysis: 

This is an extensive and rigorous laboratory study of 18 filters over 10 weeks. It investigated the effects of sand size, hydraulic head and residence time (pause period) on the removal of bacteria, virus and turbidity.

  • Two sand sizes were evaluated; fine sand - with effective size (ES) of 0.17 mm and uniformity coefficient (UC) of 2.4; and coarse sand with ES of 0.52 mm and UC of 2.1. [CAWST recommends ES of 0.15 to 0.20 mm and UC < 2.5].
  • Three hydraulic heads (initial height of water above top of sand); 10 cm, 20 cm, 30 cm. [CAWST version 10 biosand filter (BSF) design has a hydraulic head of 17 cm]
  • The filters were operated under two residence times (RT); Long RT (overnight pause period) and Short RT (no pause period).  This refers to the contact time that each batch of water had with the sand.

Bacteria, virus and turbidity were measured weekly for each filter.  After maintenance (cleaning) was carried out measurements were suspended for 7 days to allow the biolayer to re-establish. Natural river sand was used in the study. Influent water was river water augmented with 5% wastewater (for bacteria) and spiked with the MS2 virus. The influent water turbidity varied from 5 to 59 NTU.
Key Findings:

  1. When all the results were combined, the removal of bacteria was 96%, the virus removal was 71% and the turbidity removal was 89% with the effluent turbidity consistently under 3 NTU. [These results include those measurements taken from filters with coarse sand, no pause period and 30 cm head so they do not represent the full potential of the BSF.]
  2. Sand size was found to be the most critical design factor in this study. Fine sand resulted in significantly better performance for bacteria, virus and turbidity removal compared to coarse sand. [Emphasizes the need for consistent quality in preparing the sand and placing/ packing the sand into each filter.]
  3. Operating the filter with a long pause period between batches improved performance for bacteria, virus and turbidity removal.  This was especially true of virus removal.  It was found that each additional hour of residence time (up to 27 hours) was found to be beneficial.  [Water used for drinking would preferably be the water that was filtered with the greatest pause period - normally overnight.]
  4. Reduced hydraulic head (to minimize shear forces on the biofilm during the initial stage of the batch run) demonstrated improved removal but not as significantly as sand size or pause period.
  5. It was found that higher influent turbidity had a significantly negative effect on virus removal, but a (much smaller) positive effect on bacteria removal. [This is the first study to identify this negative effect on virus removal.]
  6. Based on measurements taken 7 days or more after maintenance was carried out (cleaning the filter by swirl-and-dump or ‘wet harrowing’ ), no effect on virus removal and only a modest effect on bacteria and turbidity removal was found.  [The lack of effect on virus is believed to be because viruses are removed deep in the filtration sand, not in the biolayer where bacteria and turbidity are mostly removed.]

A single layer of manually processed river sand appeared to produce comparable performance as two layers of crushed manufactured sand.


in above experiment ineed to know same details of sand layer thikness  and arrangment in the filter
about the result it is enhancement our intersting in this method of wastewater treatment specially in my country as ardid area and in the end the question.    
are there possibilty to applied in large scal ?

This study found no difference in removal rates between river sand and the "manufactured"  sand.  I've also seen from a paper by Duke??? that after 12 months all sands have same removal rates.   Yet the common line from Manz and CAWST is  that river sand is the least preferred source because of algae contamination, grain shape etc.  I'd like to see further discussion or see more articles on this. 

I understand that the reason river sand was considered 'least preferred' was to explain the cause of (occasional) negative removals from the biosand filter. In some water tests it was found that the bacteria count of the water coming out of the filter was higher than the water going into the filter. Dr. Manz felt that this situation could happen if the sand itself provides the nutrients for bacteria to multiply and possibly become unattached from the sand surface and re-enter the water. The issue is that if surface water contacts sand it may leave vegetable matter attached to the sand grains which would provide nutrients to bacteria even if the sand itself contained no bacteria (i.e. was sterilized). This 'negative removal' normally stops after a period of time (weeks or sometimes months) if it occurs at all. As you mention, Dr. Bill Duke found that the results were much the same for river sand, sea shore sand and crushed rock sand after 12 months. I direct you towards an excerpt of a presentation made to the Learning Exchange at CAWST in 2011 showing the results from Duke and other information ( As you mention, the research by Jenkins et al. also found no difference. CAWST believes it is better if crushed sand is used if it is available as this avoids a potential problem, but many implementers have used river sand (or river-deposited sand) very successfully and often this is the only sand that is available.

Derek Baker