1. Course 2 Unit 5 Introduction to Constructed Wetlands
Part A – Overview and types of constructed wetlands Part B – Basic principles of wastewater treatment in constructed wetlands Part C – Design of constructed wetlands Part D – Operation and maintenance of constructed wetlands; examples for greywater treatment
Lecturer: Dr Diederik Rousseau
for info on Diederik see next slide

2. About Diederik Rousseau
Dr. Diederik Rousseau is now Programme Coordinator & Lecturer Environmental Sciences at University College West Flanders. Before that, he was lecturer in Environmental Engineering in the Department of Environmental Resources at UNESCO-IHE Delft from October 2005 – end He holds both an MSc (1999) and PhD (2005) degree from Ghent University in Belgium in Applied Biological Sciences - Environmental Technology.
His teaching activities mainly focused on bio-monitoring based on plankton and macro-invertebrate communities, and on natural systems for wastewater treatment. Research was oriented on model-based evaluation of the performance of constructed wetlands.

3. Overview and types of constructed wetlands
Course 2 Unit 5
C2U5 Part A
Note: You will see a number of different names used for the same system (this can be confusing)
Overview and types of constructed wetlands

4. Classification overview: soil filters
unplanted filters (without plants)
planted filters (with plants)
vertical flow
horizontal flow
vertical flow
horizontal flow
Also called: subsurface biofilters, percolation beds, infiltration beds or intermittent sand filters
= constructed wetlands

5. Prerequisites for being able to use constructed wetlands
Wastewater not too toxic for bacteria and plants
Sufficient incident light to allow photosynthesis
Temperature not too low*
Adequate quantities of nutrients to support growth
Detention time long enough
Organic loading not too high (expressed as g BOD/m2/day)
Enough space, because it is a low-rate system
* But they also work in cold-climate countries such as Sweden and Norway

6. Applications – types of wastewater
Course 2 Unit 5
Applications – types of wastewater
domestic
combined / conventional
sewage
municipal
separate: greywater
dairy
ecosan
cattle
agricultural
swine
poultry
mine drainage
coal, metal
winery, abattoir, fish, potato, vegetable, meat, cheese, sugar , milk productions
industrial
food processing
heavy industry
runoff
urban
highway
field crop
nursery
greenhouse
airport

7. Applications of constructed wetlands in the ecosan context
Constructed wetlands can be used for:
Greywater treatment
Faecal sludge treatment (less common)
Post-treatment after anaerobic treatment of blackwater

8. General strong points and disadvantages of constructed wetlands
Site location flexibility (compared to natural wetlands)
Simple operation and maintenance
Can be integrated attractively into landscaping
Disadvantages/challenges:
Mosquitoes (in Free Water Surface Systems)
Start-up problems
Space requirement
Variable performance possible
Designs still largely empirical (to date)

9. Classification of constructed wetlands
Based on water flow characteristics
(free water) surface flow (abbreviated as FWS or SF)
subsurface flow (abbreviated as SSF)
Based on plant species characteristics
floating plants (e.g. Lemna, Nymphaea)
submerged plants (e.g. Elodea)
emergent plants (e.g. Phragmites, Papyrus)

10. Different groups of macrophytes
Course 2 Unit 5
Different groups of macrophytes
Helophytes
Pleustophytes
Hydrophytes
Pleustophytes
(Source: Vymazal et al., 1998)

11. Types of plants used Cattail Reed Rush Bulrush Sedge And many others

12. Selection of macrophyte species
Most often used are emergent plants (reeds, rushes, sedges) and floating plants (water hyacinth, duckweed)
Recommended to use local, indigenous species and not to import exotic, possibly invasive species
Easy lab-scale growth tests can be performed to check whether or not the plants can survive and grow in the given wastewater
Plants should have high biomass production, an extensive root system and should be able to withstand shock loads and short dry periods

13. CONSTRUCTED WETLANDS Type 1 Type 3 Type 2 Type 4 Free floating plants
Floating leaved plants
Emergent plants
Submerged plants
Surface flow (FWS)
Sub-surface flow (SSF)
Type 1
Vertical flow (VSSF)
Horizontal flow (HSSF)
Upflow
Downflow
Type 2
Type 3
Any combination of the above systems is called a “hybrid” system
Type 4

14. Table 1: Rule of thumb area requirements for different wetlands
Type of CW
Design area requirement (m2/PE)
Type 1: FWS
5-10
Type 2: HSSF
3-5
Type 3: VSSF
2-3
Type 4: Hybrids (horizontal – vertical, or H-V)
2.5 – 3
(but note: better nitrogen removal (denitrification) than VSSF)
Important:
1 PE = 1 Person Equivalent = 60 gBOD/cap/d
= 120 L/cap/d (in the Netherlands)
Valid for mixed domestic wastewater → lower values for greywater!

15. Course 2 Unit 5
Example: simple calculation for area requirement (see Part C for more details)
Design for 50 people
Use Type 2 wetland (HSSF)
Take conservative figure of 5 m2/PE
Therefore, an area of 5 x 50 = 250 m2 is needed to treat the wastewater of these 50 people in a HSSF constructed wetland

16. Type 1: FWS (Free water surface flow)
Water flows on top of soil medium, water depth < 50 cm deep.
Mostly planted with sedges, reeds, rushes.
This is a very land intensive system (5-10 m2 per PE).

17. FWS - Example of typical lay-out (slide 1 of 3)
diversion weir
pre-settlement
Diversion weir
normal flow goes to CW
storm flow bypassed
Photo: Deinze, Belgium

18. FWS - Example of typical lay-out (slide 2 of 3)
diversion weir
pre-settlement
Presettlement pond, removes large part of suspended solids, enables easy access for desludging
Photo: Deurle, Belgium

19. FWS - Example of typical lay-out (slide 3 of 3)
Course 2 Unit 5
FWS - Example of typical lay-out (slide 3 of 3)
diversion weir
pre-settlement
Wetland basins, serpentine shape (high length-to-width ratio) promotes plug flow and avoids dead zones.
Photo: Kruishoutem, Belgium

20. Type 2: HSSF or vegetated submerged beds (horizontal subsurface flow)
Water flows inside a layer of sand, gravel or soil (60-80cm).
Most often emergent plants like cattails or reeds are used. Amount of land reduced (3-5 m2 per PE) compared to FWS CW.

21. gravel, sand or soil bed planted with emergent macrophytes
HSSF - Example of typical lay-out (slide 1 of 3)
septic tank
gravel, sand or soil bed planted with emergent macrophytes
Adequate pretreatment extremely important to avoid clogging (reduction of pore volume by accumulation of solids).

22. HSSF - Example of typical lay-out (slide 2 of 3)
The inlet zone is filled with small rocks or coarser gravel. Together with multiple vertical riser pipes, this ensures that the wastewater is distributed equally over the entire width and depth of the wetland.
Photo: Zemst, Belgium

23. HSSF - Example of typical lay-out (slide 3 of 3)
Course 2 Unit 5
HSSF - Example of typical lay-out (slide 3 of 3)
Inlet zone (small rocks)
Gravel bed (fine gravel)
Outlet zone (small rocks)

24. Type 3: VSSF or “infiltration beds” (vertical subsurface flow)
Water is pumped on the surface and then drains down through the filter layer which consists of coarse sand or fine gravel. Amount of land is minimal (2-3 m2 per PE).

25. VF - Example of single-household system
(slide 1 of 3; VF = vertical flow)
Construction of impermeable basins.
With parallel basins, one bed is loaded, the other one is resting. During the resting period, accumulated organic material can be degraded and oxygen can penetrate down the filter.
Photos provided by Peter Vandersnickt (at his house in Stabroek, Belgium)

26. VF - Example of single-household system (slide 2 of 3)
Basins have been filled with sand.
On top you can see the influent distribution system (tubes with openings at regular distances to ensure equal distribution of water over entire surface).

27. VF - Example of single-household system (slide 3 of 3)
Course 2 Unit 5
VF - Example of single-household system
(slide 3 of 3)
Basins have been planted with young reed plants.
Top layer of gravel to cover the influent distribution system.
Garden has been replanted.

28. Anoxic → denitrification Aerobic → nitrification
Type 4: Hybrid systems
Anoxic → denitrification
Aerobic → nitrification
Two-step constructed wetland consisting of a HSSF and VSSF flow bed.
Different conditions in both wetlands trigger different removal pathways → hybrid systems are usually more efficient (particularly for nitrogen removal)
The recycle flow is to improve denitrification in the horizontal (anoxic) first compartment; aerobic conditions prevail in the vertical compartment

29. Basic principles of wastewater treatment in constructed wetlands
C2U5 Part B
Basic principles of wastewater treatment in constructed wetlands

30. Compartments in wetlands
Sediment / gravel bed
Root zone / pore water
Litter / detritus
Water
Air
Plants
Roots
Bacteria growing in biofilms
→ Treatment is the result of complex interactions between all these compartments

31. Wastewater treatment mechanisms (slide 1 of 4)
BOD removal
particulate BOD by settling and filtration, then converted to soluble BOD by hydrolysis
soluble BOD due to degradation by attached microbial growth (biofilms on stems, roots, gravel particles etc)
(Note: possible greenhouse gas emissions due to anaerobic processes; negligible however when compared to other sources)
Suspended solids removal
removal occurs within few meters near inlet by settling and filtration
* See appendix at the end for further information on BOD

32. Wastewater treatment mechanisms (slide 2 of 4)
Nitrogen removal
nitrification/denitrification in biofilms
plant uptake
volatilization as ammonia (at pH > 8.5)
Phosphorus removal
retention in the soil (adsorption)
precipitation with Ca, Al and Fe

33. Wastewater treatment mechanisms (slide 3 of 4)
Course 2 Unit 5
Wastewater treatment mechanisms (slide 3 of 4)
Pathogen removal
predation by protozoa
sedimentation and/or filtration
absorption
die-off from unfavorable environmental conditions (UV light, pH and temperatures)
Heavy metal removal
precipitation and adsorption
plant uptake
Trace organics removal
adsorption by the organic matter and clay particles

34. Wastewater treatment mechanisms (slide 4 of 4)
Redox conditions
FWS generally aerobic in the upper water layers, anaerobic in the sediment
HSSF mostly anaerobic (greenhouse gas emissions!)
VF mostly aerobic due to intermittent loading (pores can refill with air in between two loads)
Aerobic patches around roots due to oxygen release
Upper layers of biofilms can be aerobic whereas deeper layers can be anoxic/anaerobic
In most CWs there is thus a mosaic of sites with different redox conditions which trigger different removal processes (this is one of the major differences and advantages compared to other technologies!)

35. Role of aquatic plants in free water surface (FWS) constructed wetlands
Nutrient uptake
Heavy metal accumulation in plant tissue
(Note: usually not a problem with domestic wastewater or grey water)
Habitat for wildlife
Aesthetics
Stems = mechanical filter + attachment of biofilm
Limitation of algal growth by providing shadow
Reduce water current velocity → increases settling

36. Role of aquatic plants in sub-surface flow (SSF) constructed wetlands
Nutrient uptake
Heavy metal accumulation in plant tissue
(Note: usually not a problem with domestic wastewater or grey water)
Habitat for wildlife
Aesthetics
Root system = mechanical filter + attachment of biofilm
Root system maintains hydraulic conductivity
Oxygen transfer (active and passive) → plants are transporting oxygen to their root zone to allow the roots to survive in anaerobic conditions. Part of this oxygen is available for microbial processes.

37. Design of constructed wetlands
C2U5 Part C
Design of constructed wetlands

38. Basic design question: how much area needed?
Wind
Land
Sun
Seeds
Soils
Plants
Microbes
Natural systems
(rule of thumb 2-10 m2/PE)
Since wetlands are low-rate systems which are completely depending on solar energy, they need a much larger surface area than conventional systems with electrical energy input (e.g. activated sludge plants).

39. Table 2: Design criteria for different types of constructed wetlands
Design parameter
FWS
(free water surface)
HSSF
(horizontal sub-surface flow)
VSSF (vertical sub-surface flow)
Data is for which wastewater type
Mixed domestic wastewater
greywater
Detention time (days)
5 - 14
2 - 7
N/A
Max. BOD loading rate (g/m2/day) 8
7.5
4-6
Water or substrate depth (m)
0.1 – 0.5
0.1 – 1.0
Hydraulic loading rate (mm/d)
7 - 60
2 - 30
Area requirement (ha/m³/day)
0.002 – 0.014
0.001 – 0.007
Aspect ratio – length/width
2:1 to 10:1
0.25:1 to 5:1
Mosquito control
Required
Not required
Harvest frequency (years)
3-5
Source: Wood (1995) for FWS and SSF; Ridderstolpe (2004) for VSSF

40. Conversions of different units
Course 2 Unit 5
Conversions of different units
BOD load:
1 kg/ha/d = 0.1 g/m2/d
(because 1 hectare = 10,000 m2)
Hydraulic load:
1 L/m2/d = 1 mm/d
(because 1 L = 10-3 m3, and 1 m = 1000 mm)

41. Further design criteria for horizontal flow constructed wetlands
Source: Morel and Diener (2006), p Same type of information is available for the other types of soil filters / constructed wetlands in that publication

42. Source: Morel and Diener (2006), p. 34
Example photos of horizontal flow constructed wetlands for greywater treatment
Source: Morel and Diener (2006), p. 34

43. Example calculation based on BOD loading rate
You need to know the design flowrate (Q) and BOD concentration (measure existing greywater or make estimate), e.g ML/d and 200 mg/L BOD (see also Course 1 Unit 2 for greywater characteristics)
Now calculate the BOD load:
LBOD = Q · CBOD = 0.01 ML/d · 200 mg/L = 2 kg/d = 2000 g/d
Pick a design value from Table 2, say for VSSF constructed wetland: qdesign = 6 gBOD/ m2 /d
Then area required is:
A = LBOD / qdesign = 2000 g/d / 6 g/ m2 /d = 333 m2
So the required area for this design is 333 m2

44. Course 2 Unit 5
Table 3: Comparison of actual design parameters and costs for 107 constructed wetlands in Flanders, Belgium
Parameter
FWS
VSSF
HSSF
Combined
Design size (PE)
1 – 2000
4 – 2000
152 and 350
Area (m2/per PE) 7
3.8
5.9 and 3.7 5
Values from Table 1 (m2/per PE)
5 - 10
2-3
3-5
Investment cost (€/PE)
392
507
1636 and 879
919
Source: Rousseau et al. (2004)
Note: all these wetlands treat mixed domestic wastewater
The flowrates are in most cases not measured, so there is no information on the hydraulic or BOD load to these wetlands.

45. Table 4: Comparison of actual performance for 107 constructed wetlands in Flanders, Belgium (based on measured average concentrations)
Parameter
FWS
VSSF
HSSF
Combined
VSSF greywater
COD removal (%) 61 94 72 91
90 – 99 (BOD)
SS removal (%) 75 98 86
90-99
TN removal (%) 31 52 33 65 30
TP removal (%) 26 70 48
30 – 95
Source: Rousseau et al. (2004)
Note: all these wetlands in Flanders treat mixed domestic wastewater
The column on the right in green is for greywater treatment in vertical sub-surface flow wetlands (Ridderstolpe, 2004)

46. Comments on previous slide regarding performance
Comparisons of systems in Flanders regarding removal performance:
FWS: worst overall performance
VSSF: best performance for COD, SS and TP
Combined/hybrid: best performance for TN
But the data gives no indication on how much area is required to achieve this performance – the key parameter to compare would be load of BOD removed per m2 and d (in g/ m2 /d)
But to determine this value, one would have to know the flow rate
Unfortunately, the flow rate is rarely measured at small constructed wetlands

47. Rules of thumb for greywater treatment in VSSF CW
Water should percolate through the soil in an unsaturated flow
Design life: years
Removal efficiencies:
90-99% for SS and BOD
30-95% for phosphorus
30% for nitrogen
Greywater may need pre-treatment before wetland to remove suspended solids and excessive amounts of fat
Source: Ridderstolpe (2004)

48. Rules of thumb for mixed domestic wastewater treatment in VSSF CW
Water should percolate through the soil in an unsaturated flow (hydraulic loading rates of up to 800 mm/day have been achieved without clogging)
Intermittent loading needed (generally 2-4 times/ day)
Effluent quality (BOD removal + nitrification of domestic wastewater):
< 10 mg/L BOD
< 10 mg/L TSS
< 2 mg/L NH4-N
Design: 2-3 m2 / PE
Note: wastewater may need pre-treatment before wetland to remove suspended solids and excessive amounts of fat
Source: Cooper (2004)

49. Other considerations: water balance
Course 2 Unit 5
Other considerations: water balance
Influent wastewater flow
- effluent wastewater flow
+ precipitation
- evapo-transpiration
= change in water volume over time
Because of their large surface areas, constructed wetlands are very sensitive to precipitation and evapotranspiration!

50. Other considerations: slopes and liners
Bottom slope:
0.5%or less for FWS systems
2%or less for HSSF systems
Liners (clay or plastic):
when groundwater contamination or water conservation is a concern (depends on local soil characteristics)

51. Other considerations: filling medium
The selection of the filling medium of a subsurface flow constructed wetland is based on:
hydraulic conductivity
→ high enough to allow easy water flow
local availability
→ reduced transport costs
phosphate sorption capacity*
→ the more P-binding sites available (depending on Fe, Al and Ca content), the longer and the more P can be adsorbed
Sand has better P-sorption capacity but lower hydraulic conductivity than gravel → higher clogging risk
* Less important for greywater

52. Design summary
No significant difference between design for conventional wastewater and greywater
If design is based on BOD load then check BOD concentration of greywater in question
If design is based on a per person pollutant load then remember that the per person load is lower for a greywater treatment system since faeces and urine are treated separately
CWs for greywater are in general slightly smaller than CWs for domestic wastewater (for same number of people)

53. O&m of constructed wetlands and examples for greywater treatment
C2U5 Part D
O&m of constructed wetlands and examples for greywater treatment

54. Basic maintenance
‘natural’, low-tech systems  require low but still adequate maintenance
Vymazal et al. (1998) recommends checking larger systems (> 500 PE) on a daily basis, including:
pretreatment units
inlet structures
outlet structures
If maintenance is ignored:
uneven flow distribution
local overloading
deterioration of treatment efficiency in the long term

55. Lack of maintenance (or wrong design): consequences (example for a vertical sub-surface flow CW)
This wetland is overloaded (no pre-treatment); sludge is accumulating very quickly on the surface (normally sludge should only accumulate at a much slower rate and be removed about once in 10 years)
Excessive sludge accumulation threatens to block the influent distribution system and the pore spaces

56. Should plants from a constructed wetland be harvested or not?
Advantages of harvesting:
net nutrient export from the system
prevention of thick layers of dead plant material with stagnant water in FWS which are ideal pest breeding places
Advantages of not harvesting:
creation of an isolating layer of dead plant material
provision of a detritus layer that can adsorb trace metals
provision of a carbon source for denitrification
no alteration of the ecological functioning of wetlands
Recommendation: harvest every 2-3 years (in winter when applicable)

57. (applies only to FWS CWs)
The mosquito problem
(applies only to FWS CWs)
The problem: mosquitos can bread in free water flow constructed wetlands
There is less mosquito breeding if the biodiversity and complexity of food web is high
Solutions:
No above-ground flow (use SSF wetlands instead)
Pretreatment to reduce organic loading rate
Temporary drying of the beds will eradicate larvae
Lower water depths  higher stream velocity
Insecticides  not sustainable, very expensive
Open, unplanted water areas support growth of predators (fish, invertebrates)
Biological control agents can be added: Bacillus thuringiensis, Gambusia
Gambusia fish

58. Ex 1: Culemborg, NL (slide 1 of 3)
Course 2 Unit 5
Ex 1: Culemborg, NL (slide 1 of 3)
Vertical flow constructed wetland for greywater treatment at EVA Lanxmeer (ecological residential area in Culemborg, The Netherlands).

59. Ex 1 (cont’d): Schematic (slide 2 of 3)
Distribution pipes
Lining of basin
Reed plants
Sand filter
Pump for grey-water
Drainage
floor
Sampling
point
Gravel + shells
Treats domestic wastewater from houses, offices and schools
Pre-settling tank before constructed wetland
Vertical flow over reed beds (helophytes) and bed of sand and gravel
Micro-organisms in the sand bed and on the roots of the plants degrade the wastewater
Discharge of treated effluent to surface water
Not much performance data available for this wetland

60.  Effluent discharged in surface water
Ex 1 (cont’d) – Performance data (slide 3 of 3)
Pollutant concentrations in effluent samples taken at Culemborg on 27 January 2006 (winter conditions)
 Effluent discharged in surface water
Source of this data: provided by Sander Booms, Municipality Culemborg

61. A system for one household of 7 people
Ex 2 – Kathmandu, Nepal: Vertical flow constructed wetland for single household (slide 1 of 2)
A system for one household of 7 people
Source: Morel and Diener (2006), p. 66

62. Source: Morel and Diener (2006), p. 67
Course 2 Unit 5
Ex 2 – Example Kathmandu, Nepal: Performance data (slide 2 of 2)
Source: Morel and Diener (2006), p. 67

63. Research outlook for CWs in general
The real challenge is to open up the ‘black box’ that constructed wetlands still are:
Process identification (rising evidence that for instance ANAMMOX* might take place)
 use of microbial marker techniques etc
Process quantification
 e.g. root oxygen release, sulphate reduction
Process modelling
Comprehensive mathematical models that represent the complex network of interacting processes
Can be used to optimize design and operation
* ANAMMOX = Anaerobic ammonia oxidation

64. References
Cooper P. (2005). The performance of vertical flow constructed wetland systems with special reference to the significance of oxygen transfer and hydraulic loading rates. Proceedings IWA Conference on Constructed wetlands, Avignon, France.
Morel A. and Diener S. (2006) Greywater Management in Low and Middle-Income Countries, Review of different treatment systems for households or neighbourhoods. Swiss Federal Institute of Aquatic Science and Technology (Eawag). Dübendorf, Switzerland. Download from (also placed under Extra Reading)
Ridderstolpe, P. (2004) Introduction to greywater management, Stockholm Environment Institute, Sweden, Report Download from (also placed under Extra Reading)
Rousseau D.P.L., P.A. Vanrolleghem and N. De Pauw (2004). Constructed wetlands in Flanders: a performance analysis. Ecological Engineering, 23(3), Also placed under Extra Reading
Vymazal, J., H. Brix, P.F. Cooper, B. Green and R. Haberl (Eds) (1998). Constructed wetlands for wastewater treatment in Europe. Backhuys Publishers, Leiden, 366 p.
Wood, A. (1995). Constructed wetlands in water pollution control: fundamentals to their understanding. Water Science and Technology, 32(3),

65. Further reading if you are really interested in constructed wetlands
Kadlec, R.H. and R.L. Knight (1996). Treatment wetlands. CRC Press, Boca Raton, FL, USA. – A new edition will come out in 2007 or 2008

66. Useful websites on constructed wetlands in general (not just for ecosan)
US EPA: Constructed wetlands for wastewater treatment and wildlife habitat
Constructed Wetlands Association

67. Appendix: information about BOD and COD
Course 2 Unit 5
Appendix: information about BOD and COD

68. Biochemical oxygen demand (BOD)
The BOD test is a bioassay in which the rate (and extent) of the aerobic degradation of organic matter is assessed in terms of the amount of oxygen consumed in its degradation.
Unit is mgO2/L, or in short mg/L
It is used to determine:
The organic strength of the wastewater
The approximate amount of oxygen required to biologically stabilize the organics in wastewater
The size (capacity) of wastewater treatment facilities (based on BOD load in kg/d)
Efficiency of some treatment processes (e.g. % BOD removal)
Compliance with wastewater discharge permits (effluent BOD concentration)

69. Principles of BOD test Simplified reaction:
Microorganisms + Organic matter + O2 
More microorganisms + CO2+ H2O + Residual organic matter
To ensure the presence of excess dissolved oxygen (DO) throughout the test, a DO depletion of no more than 70% is allowed for a valid test. Water at 20C (the standard temperature of the BOD test) contains only about 9 mg/L DO. Therefore, samples for BOD measurement must usually be diluted.
Oxygen is supplied by saturating the sample with air.
Microorganisms are supplied by seeding the dilution water with an appropriate inoculum (usually sewage, treated sewage or, in some cases, microorganisms acclimatised to the particular substrate of interest).

70. BOD5 test in practice To the BOD bottle add:
Diluted sample (usually diluted 50 times or more for wastewater!)
Innoculum (bacteria)
Oxygen to saturate to ~9 mg/L
Measure how much oxygen is left after 5 days (must still be > 2.5 mg/L else dilute more)
The less oxygen is left, the higher the BOD, therefore the more organic pollution
The BOD test is time consuming and has a low precision

71. Chemical oxygen demand (COD)
Expressed as amount of oxygen required for chemical oxidation of organic matter by a strong oxidant (permanganate or dichromate) in acid solution.
Unit is mgO2/L, or in short mg/L
Consumption of permanganate or dichromate is converted into an equivalent oxygen demand (amount of oxygen which will be consumed if the oxidation had taken place by using oxygen)

72. COD test in practice
To perform a test, users have to add a wastewater sample to a cuvette (with reagent) and leave it in a heater for 2 hours. At the end of this period the intensity of colour in the solution is directly related to the COD value in the sample, and can be measured with a spectrophotometer.
Cuvettes with reagent
Spectrophotometer

73. Relationship of COD to BOD
The COD concentration is always higher than the BOD concentration for a given sample because:
many organic substances which are difficult to oxidise biologically can be oxidised chemically
inorganic substrates that are oxidised by the dichromate increase the apparent organic content (COD value) of the sample
certain organic substances may be toxic to the microorganisms used in the BOD test