Geoderma 228–229 (2014) 33–43
Contents lists available at ScienceDirect
Geoderma
j ourna l homepage: www.e lsev ie r .com/ locate /geoderma
Can treatmentwetlands be constructed on drained peatlands for efficientpurification of peat extraction runoff?
Heini Postila a,⁎, Jaakko Saukkoriipi b, Kaisa Heikkinen c, Satu Maaria Karjalainen c, Minna Kuoppala c,Hannu Marttila a, Bjørn Kløve a
a Water Resources and Environmental Engineering Laboratory, Department of Process and Environmental Engineering, University of Oulu, P.O. Box 4300, FI-90014 University of Oulu, Finlandb Pöyry Finland Oy, Environmental Studies, Tutkijantie 2A, P.O. Box 20, FI-90571 Oulu, Finlandc Finnish Environment Institute (SYKE), Freshwater Centre, P.O. Box 413, FI-90014 University of Oulu, Finland
⁎ Corresponding author. Tel.: +358 294 484503.E-mail address: [email protected] (H. Postila).
0016-7061/$ – see front matter © 2013 Elsevier B.V. All rihttp://dx.doi.org/10.1016/j.geoderma.2013.12.008
a b s t r a c t
a r t i c l e i n f oArticle history:Received 3 December 2012Received in revised form 11 November 2013Accepted 11 December 2013Available online 18 February 2014
Keywords:PhosphorusWetlandPeatVegetation
Peat extraction for energy purposes causes major changes in the aquatic and terrestrial environment. Accordingtonational strategies for extracting peat in Finland, newpeat extraction areas should be establishedonpreviouslydrained peatlands. On such areas it is difficult to find the natural, intact peatland required for treatmentwetlandsor so-called overlandflowareas,which are considered the best available technology for runoffwater purification.Therefore treatment wetlands must be constructed on drained peatland. It is known that drainage causes phys-ical, biogeochemical and hydraulic changes in the peat layer, as well as changes in vegetation. It is probable thatthese changes affect the purification efficiency ofwetland treatment systems inmanyways. This study evaluatedthe function and characteristics of drained peatland areas for purification of peat extraction runoff water. Studysites were established on 11 drainedwetlands and their purification efficiencywas evaluated. Detailedmeasure-ments of peat physical properties and hydraulic conductivity, as well as studies on vegetation, were also made inthe study areas. The results showed that wetlands constructed on drained peatland areas can purify peat extrac-tion runoff. However, leaching of phosphorus (P) and iron (Fe) was observed in some areas. Leaching is influ-enced e.g. by pH and the soil P pool. The chemical oxygen demand was also observed to increase in runoffwater from the wetland. The results indicated that low (Fe + aluminium (Al) + manganese (Mn))/P ratio(≤25) and quite high P content (N1200 mg/kg) in the surface peat characterised those areas where P leachingwas observed. The presence of a dense tree stand in a drained peatland area also seemed to indicate release ofnutrients from the area after its rewetting and use as a treatment wetland. Thus, potential nutrient releasefrom a drained peatland area intended for use as a treatment wetland can be assessed by studying the character-istics of the peatland area, especially the peatmineral content, and the vegetation, especially tree stand density inthe area. Using the findings obtained, a conceptual decision tree was drawn up in order to help to establish anddesign wetlands in previously drained peatland areas.
© 2013 Elsevier B.V. All rights reserved.
1. Introduction
Wetlands arewidely used globally as a purification step inwastewa-ter treatment, for example in agricultural (Dunne et al., 2005) andmunicipal systems (Kadlec and Wallace, 2009). Treatment wetlandsare designed to filter and treat runoff water pollutants based onsedimentation, physical filtration, geochemical processes and aerobicand anaerobic biological processes in the wetlands (Kadlec andWallace, 2009). Generally, numerous characteristics affect purification,such as residence time (Wörman and Kronnäs, 2005), soil properties(Heikkinen et al., 1995; Kadlec and Knight, 1996; Liikanen et al., 2004;Pant and Reddy, 2003), biological factors such as microbes andplants (Henze et al., 2002; Huttunen et al., 1996), climate (Kadlec,1999; Kuschk et al., 2003; Werker et al., 2002), redox dynamics
ghts reserved.
(Niedermeier and Robinson, 2007), pH (Grybos et al., 2009; Kadlecand Knight, 1996) and hydraulic load (Braskerud, 2002; Heikkinenet al., 2002).
In Finland, treatment wetland systems are commonly constructedon peatlands due to their wide availability and their suitability fortreating and polishing different types of wastewater. Due to the cool,humid climate and flat topography, peatland covers about one-third ofthe land area in Finland. Peat-based wetlands are used to treat runofffrom peat extraction areas (Heikkinen et al., 1995; Ronkanen andKløve, 2009) and in peatland forestry (Nieminen et al., 2005; Silvanet al., 2004), and also to polish municipal wastewater (Ronkanen andKløve, 2007) or mining effluents (Närhi et al., 2012; Päkkilä, 2008). Inpeat extraction, treatment wetlands are mainly called overland flowareas (OFAs, see Fig. 1) and they are established on undisturbedpeatlands. According to guidelines set by the Ministry of theEnvironment (2013), wetland slope (recommendation 1%), treatmentwetland area of the catchment area (recommendation N 3.8%), average
Fig. 1. Example of a water treatment wetland.
34 H. Postila et al. / Geoderma 228–229 (2014) 33–43
peat thickness (recommendation N 0.5 m) and average degree of humi-fication of surface peat (recommendation H1–H3) are parameterswhich should be analysed in OFAs, as they can affect their purificationefficiency. OFAs are considered to belong within best available tech-niques (BAT) in purifying runoff water from peat extraction sites(Heikkinen et al., 2009). However, the current national areal planningstrategy (Ministry of the Environment, 2007) directs the implementa-tion of new peat extraction sites to previously drained areas. In thesepeatland areas it is difficult to find the natural, intact peatland surfaceareas needed for OFA wetland treatment systems (Ihme, 1994; Ihmeet al., 1991), and treatment wetlands must thus be constructed ondrained peatland areas. Therefore, dozens of treatment wetlands havealready been established on drained peatland areaswithout any knowl-edge of their functionality or design parameters.
It is well known that peatland drainage changes the peat's physicalproperties (Burke, 1978; Holden et al., 2004; Vasander et al., 2003),such as porosity, hydraulic conductivity, degree of humification, watercontent and geochemical properties. Water pathways in peatlands arealtered by lowering the groundwater level and providing more rapidoutflow for surface runoff and rainfall (Holden, 2009). In treatmentwet-land areas, old drainage ditches can function as cut-off channels, thus
reducing retention time and impairing purification results. However,there is a wide range of variation in soil physical and hydraulic proper-ties and in vegetation cover in drained peatlands, which is often depen-dent on peat type and local characteristics (Kløve, 2000). Establishmentof a treatment wetland on a drained peatland area involves rewetting,which promotes further changes in soil, vegetation and hydraulicfeatures of the area (Haapalehto et al., 2011). Rewetting also changesthe aerobic and anaerobic conditions in peat layers. This may causeleaching of nutrients and geochemical elements, as has been noted atpeatland restoration sites (Koskinen et al., 2011; Vasander et al.,2003). All of these issues need to be clarified before treatmentwetlandscan safely be established on drained peatland areas. At present, little orno information is available on the major factors and mechanismsinfluencing purification processes in treatment wetlands establishedon drained peatland areas.
From a water treatment point of view, it is important for treatmentwetlands to retain nutrients (P and nitrogen (N)) and suspended solids(SS) present in runoff water. Removal of P from runoff water is perhapsone of larger challenges in treatment wetlands established on drainedpeatland areas compared with OFAs. The purification results for P inOFAs or other types of treatment wetlands depend on different
35H. Postila et al. / Geoderma 228–229 (2014) 33–43
biological, chemical and physical processes (Kadlec and Knight, 1996).In OFAs, chemical adsorption by peat has been identified as the mainprocess in removing PO4–P from peat extraction water (Heikkinenet al., 1994; Heikkinen et al., 1995). In treatment wetlands in general,P is also known to be retained via sedimentation and filtration of partic-ulate matter. Some retention of dissolved organic Fe–P colloids has alsobeen detected in OFAs (Heikkinen and Ihme, 1995), probably as a resultof floc formation and sedimentation,which canmaintain their P remov-ing capacity (Ronkanen and Kløve, 2009). In treatment wetlands, P canalso be retained through uptake by plants andmicroorganisms. Howev-er, the role of plants as annual nutrient sinks is reported to be small inwastewater wetlands (Richardson and Nichols, 1985) and OFAs(Heikkinen et al., 1994; Huttunen et al., 1996).
The main aim of the present study was to evaluate the function andcharacteristics of drained peatland areas for water purification pur-poses. Our specific objectives were twofold: 1) To study possibilities topurify peat extraction runoff using wetlands established on drainedpeatland areas; and 2) to determine the characteristics of drained wet-lands and the processes leading to P removal. Our starting hypothesiswas that in some situations malfunctions can occur, leading e.g. toleaching of P from the treatment wetland. It was also assumed thatthe peatland areas could function well as treatment wetlands even ifthey are poorly drained and/or gradually self-restored. Furthermore,we aimed to develop a preliminary conceptualmodel to assist in design-ing treatment wetlands on previously drained peatland areas. Our hy-pothesis in this case was that peat soils rich in Fe, Al and calcium (Ca)can retain P.
2. Materials and methods
2.1. Study sites
The study sites comprised 11 treatment wetlands constructedon drained peatlands in the North Ostrobothnia, Lapland andCentral Ostrobothnia regions of Finland (Fig. 2). Mean annual
Fig. 2. Location of t
precipitation in these regions varies between 550 and 650 mm(1981–2010) and mean annual temperature is 1–3 °C. Mean annualevaporation is 250–350 mm and the length of thermal winter (meantemperature b 0 °C) is 5–6 months. The wetlands studied wereestablished during the period 1992–2010 (Table 1), which provided agood basis for evaluating purification efficiency over different time pe-riods. Wetland area comprised between 1.8 and 8.9% of the total catch-ment area. Wetland ditches were located partly in the flow directionand partly against it, depending on site.
2.2. Peat properties, mineral soil contact, vegetation andhydraulic conductivity
Samples were collected (2009–2011) from the peat surface (about0–10 cm depth) from 4 to 11 locations at each study site (cf. Table 2).The degree of humification of surface peat was quantified according tothe von Post scale (Hobbs, 1986). P, Fe, Al, Ca, magnesium (Mg) and Mncontents were analysed in peat samples according to the EPA3051A stan-dard using ICP-OES. Based on the mean (Fe + Al + Mn)/P ratio of sur-face peat and P content, wetlands were divided into two differentgroups, one with a high ratio (N45) and other with a low ratio (≤25),together with quite high P content (N1200 mg/kg). Differences betweenthese groups were analysed by the Median test. Loss of ignition (LOI)was determined for the Hankilaneva 1, Hankilaneva 2, Kapustaneva,Luomaneva, Savaloneva and Äijönneva peat samples according to theSFS 3008 standard. Correlation analysis was performed on the results ofthe surface peat analyses for all thesewetlands. The Spearman correlationcoefficient (rs) was used in the analysis in order to reduce errors dueto outliers and non-normal distributions. Different fractions of P inthe peat were determined for four treatment wetlands (Hankilaneva 1,Kapustaneva, Savaloneva and Äijönneva) in 3–5 samples each(cf. Table 3). The measured fractions were: 1) loosely bound and water-soluble P; 2) Fe- and Al-bound P; 3) Ca-bound P; and 4) organic Pincluding humic and fulvic acids. Sequential extraction was performedaccording to Golterman (1996). Soil analyses were performed in a
he study sites.
Table 1Basic information on the 11 treatment wetlands studied.
Wetland Wetland established Peat extraction started in thecatchment area of wetland
Treatment wetlandarea (ha)
Treatment wetland areaas % of catchment area
Slope (%) Average peatthickness (m)
Hankilaneva 1 1992 1994 8.9 8.9 0.6 2.1Hankilaneva 2 1992 1994 7.8 3.5 0.07 1–1.5Kapustaneva 2008 2008 6.9 4.6 0.3 1.4Luomaneva 1998 1992 3.2 2.8 0.7–0.9 2.6Savaloneva 2005 2009 6.1 7.4 0.06–0.4 0.9Äijönneva 2009 2010 5.8 5.6 0.25 1.2Iso-Lamminneva 2010 2011 1.6 4.1 0.07 2.5Itäsuo 1995 1979 12 6.3 0.23 1Kuljunneva 2009 2010 5.8 6.6 0.02 2Lumiaapa 2 1996 1977 9.5 5.1 0 0.5Pohjoinen Latvasuo 1994 1996 2.8 1.8 0.6 1.1
36 H. Postila et al. / Geoderma 228–229 (2014) 33–43
FINAS-accredited laboratory. Mineral soil contact via ditcheswas evaluat-ed either by measuring ditch depth and comparing it with the depth ofpeat layer or by field testing for a mineral soil bottom in the ditch.
Vegetation on the study wetlands was determined once during2009–2011 at 3–6 points in each wetland. Plant species were identifiedusing 2 m ∗ 2 m squares. Required trophic level of plants was consid-ered according to Eurola et al. (1994). Estimation of tree stands in treat-ment wetlands was made by visual assessment at the sites.
Hydraulic conductivity values in the peat profile were measured at4–12 points in each treatment wetland. Each point included 1–7 differ-entmeasuringdepths, usually at 10-cm intervals. The hydraulic conduc-tivitywasmeasured in the fieldwith a direct-push piezometer using thefallingheadmethod (Ronkanen andKløve, 2005). In somewetlands andsome points, measurement of hydraulic conductivity in the upper layerwas difficult as the piezometer continuously fell (sank) into the deeperlayers. This decreased the number of depthmeasurements at the points.However, there were generally equal numbers of measured depthswithin wetlands.
2.3. Water purification efficiency in treatment wetlands
To determine the purification efficiency of the treatment wetlands,water samples were taken from incoming and outgoing water andanalysed for pH (SFS 3021:1979), suspended solids (SS; SFS-EN872:2005), total nitrogen (total N; SFS-EN ISO 11905-1:1998), ammonianitrogen (NH4–N;Methods for the Examination ofWaters and AssociatedMaterials: Ammonia in Waters 1981or SFS 3032:1976 depending on thelaboratory), sum of nitrate and nitrite nitrogen (NO2 + 3–N; Methods forthe Examination of Waters and Associated Materials: Oxidised Nitrogenin Waters 1981 and EPA Method 353.1 (approved at 40 CFR Part 136,not approved at Part 141) or SFS-EN ISO 13395:1997), total phosphorus(total P; in-house method, which corresponds in principle to SFS-EN1189 Section 6 (1997) or SFS-EN ISO 6878:2004), dissolved total phos-phorus, phosphate phosphorus (PO4–P; EPA Method 365.1 or SFS-ENISO 6878:2004), iron (Fe; SFS 3047:1980 or SFS 3028:1976) and chemicaloxygen demand (CODMn; SFS 3036:1981). Dissolved total P was deter-mined by filtering through a Whatman GF/C (~1.2 μm) filter. Particulatetotal P was then calculated by deducting the total P content of the filtratefromtheP content of the correspondingunfiltered sample. PO4–P concen-tration was measured in unfiltered, conserved water samples, except forthe Kapustaneva site, where dissolved inorganic P was measured infiltered (0.45 μm) samples. Water analyses were performed in laborato-ries accredited by the Finnish Accreditation Service (FINAS) in 2009 and2010. Water quality data from previous years were obtained from thepeat extraction load monitoring programme.
Sampling, sample pre-treatment and transportation were per-formed according to the guidelines of the laboratory quality manual,based on the SFS-EN ISO/IEC 17025 standard. Water samples weretaken in summer and in some cases also in autumn on the same dayfrom incoming and outgoing water often once per month or every
secondweek. However, the exactmonitoring periods and sampling fre-quencies varied between years and wetlands (Table 4).
Water purification efficiency R was determined by calculating themean concentrations in incoming and outgoing water for each wetlandand year, and calculated using the equation:
R ¼ Cin−Cout
Cin� 100% ð1Þ
where Cin is mean concentration in incoming water and Cout mean con-centration in outgoing water.
3. Results
3.1. Peat properties, vegetation and hydraulic conductivity
There was great variation in the physical and chemical properties ofthe peat, as well as in the vegetation characteristics, between the 11treatment wetlands. The average degree of humification in surfacepeat varied from H3 to H7 according to the von Post scale (Table 2).The mean P content of surface peat was low (b800 mg/kg dry matter)in the Kapustaneva, Itäsuo, Lumiaapa 2 and Pohjoinen Latvasuo wet-lands. In the Hankilaneva 1, Hankilaneva 2, Iso-Lamminneva andPohjoinen Latvasuowetlands, themean (Fe + Al + Mn)/P ratio of sur-face peat was high (Table 2). The P fractionation results showed that Pwas mainly organic (Org.-P) in all four wetlands studied (Table 3). Thesecond most important P pool was Ca-bound P at all other study sitesexcept Kapustaneva, where the second highest P pool was Fe- and Al-bound P (Table 3). The proportion of loosely bound P was negligible atall four sites.
According to the results of surface peat analysis for the Hankilaneva1, Hankilaneva 2, Kapustaneva, Luomaneva, Savaloneva and Äijönnevawetlands, Al had a negative correlation with loss of ignition (LOI) inpeat (rs = −0.612). The corresponding correlation between Fe andLOI was also negative (rs = −0.76), indicating as expected that thehigher the LOI, the lower the mineral soil content. However, increasedmineral soil content did not increase the peat P content (Fig. 3). Abouthalf of the wetlands, mineral soil contact was visible due to the ditchespenetrating beyond the peat layer (Table 2).
The trophic level of vegetation in the study areas varied betweenombro-oligotrophic and mesotrophic (Table 2). Kapustaneva, Itäsuoand Kuljunneva had the lowest trophic levels and Hankilaneva 1 andLuomaneva the highest. The tree stands in the wetlands were mainlyformed by Pinus sylvestris and Betula spp. There were only a few treesin five of the treatment wetlands, but considerable tree densities infour wetlands (Table 2). The Kapustaneva and Iso-Lamminneva wet-lands had dense tree cover in some areas and sparse tree stands inother areas.
In some wetlands the hydraulic conductivity of the peat was high atall depths investigated (down to 60–70 cm), whereas in others it de-creased rapidly with depth (Fig. 4). The median hydraulic conductivity
Table 2Vegetation and peat properties of the 11 wetlands studied (DM = dry matter).
Wetland Trophic level indicatedby vegetation
Tree standa Average degreeof humificationof surface peat
Mineral content of surface peat
Mineral soil contact viawetlands ditch
LOI (%) P (mg/kg DM) (Fe + Al + Mn)/P (Ca + Mg)/P nb
Hankilaneva 1 Mesotrophic Sparse/no trees No H6 80 1509 57 5 8Hankilaneva 2 Oligo-mesotrophic Sparse/no trees Maybe no H7 83 1166 48 8 7Kapustaneva Ombro-oligotrophic Sparse/dense No H4 93 782 10 12 6c
Luomaneva Mesotrophic Dense Yes H5 85 2075 18 6 4Savaloneva Oligo-mesotrophic Dense Maybe yes (at least via
distribution ditches)H5 88 1861 25 1 11
Äijönneva Oligo-mesotrophic Dense No (except Äijönkanavaof which water is dischargedout of the wetland)
H5 91 1714 8 3 8
Iso-Lamminneva Oligo-(mesotrophic) Dense/sparse No H4 914 70 8 7Itäsuo Oligotrophic Sparse/ no trees Yes H3 796 14 24 9Kuljunneva Oligotrophic Dense No H5 1294 16 5 8Lumiaapa 2 Oligo-mesotrophic No trees Yes H4 465 16 18 8Pohjoinen Latvasuo Oligo-mesotrophic Very sparse/no trees Yes H5 751 50 17 8
a Where twodifferent characteristics, e.g. sparse/dense, arementioned, thewetland area has two different types of vegetation, i.e. areaswith only sparse tree cover and areaswithmanytrees
b Number of soil samplesc LOI specified only for five samples
37H. Postila et al. / Geoderma 228–229 (2014) 33–43
decreased below 10 or 20 cm in the Kapustaneva, Luomaneva,Savaloneva andÄijönnevawetlands and below40 cm in the Itäsuowet-land. At Iso-Lamminneva, the median hydraulic conductivity decreasedbetween 20 cm and 30 cm, but increased in the deeper peat layers.
3.2. Water purification efficiency
High variability in purification efficiency was observed between thestudy sites and annual variationswere also observed (Table 4). Thewet-lands retained inorganic N andmainly also total N and SS. The retentionof total, particulate and inorganic P and Fe showed some scattering;some sites gave good purification results, while leaching of P and Fewas observed in other wetlands. The treatment wetlands studiedseemed to retain particulate P more efficiently than dissolved P(Table 4). The results also indicate that the average chemical oxygen de-mand (CODMn) values mostly increased within the wetlands whencomparing their input and output concentrations.
Leaching of P and N immediately after wetland construction was aproblem in the wetlands Hankilaneva 2, Äijönneva and Iso-Lamminneva and some of these wetlands leached P constantly duringthewhole observation period. Theweakest purification resultswere ob-served in the Savaloneva wetland, which leached P, Fe, organic matter,total N and SS throughout the observation period. The incoming waterquality varied greatly between wetlands and years.
4. Discussion
4.1. Water purification efficiency
Some of the wetlands released P at least during the first years afterthe wetland was established (Table 4). This has also been observedafter rewetting and restoration of drained peatlands (Koskinen et al.,
Table 3Content of four different P poolsmeasured in surface peat samples from four different treatmen
Study site Loosely bound andwater-soluble P
Fe + Al–P
mg/kg % mg/kg %
Hankilaneva 1 3 0.2 3 0.2Kapustaneva 21 2.2 316 33Savaloneva 3 0.2 125 9Äijönneva 11 0.7 154 10
a Calculated as the sum of the P-fractions determined
2011; Nieminen et al., 2005) or other peatland areas (Kjaergaard et al.,2012), and has been attributed to changes in oxygen and biogeochemi-cal conditions after a rise in the groundwater level in peat profiles (seeSections 4.2–4.3). This first flushing effect also explains the greatestdifference between years, which was detected at Hankilaneva 2,where approximately 200% leaching of total P and PO4–P was observedin 1992, but 70% removal of these elements in 2010 (Table 4). Besides atHankilaneva 2, leaching of total P and PO4–P was mainly observed atLuomaneva, Savaloneva, Äijönneva and Kuljunneva, where dense treestands were present (Tables 2 and 4). On the other hand, theKapustaneva, Itäsuo, Lumiaapa 2 and Pohjoinen Latvasuo wetlands,which had sparse tree stands, all retained total P even in the first or sec-ond year after establishment of thewetland. In OFAs constructed on un-drained peatland areas (Heikkinen et al., 2002; Kløve et al., 2012;Tuukkanen et al., 2012), around a 40–55% reduction in total and inor-ganic P has been observed. The present results indicate that P retentionalso occurs in treatmentwetlands constructed on drained peatlands andthat the purification results can be nearly as good as in OFAs (Table 4).The average inflow concentration (80 μg/l total P and 26 μg/l PO4–P)in the treatment wetlands studied was also at approximately the samelevel as in OFAs studied previously (89 μg/l total P and 30 μg/l PO4–P)(Kløve et al., 2012; Tuukkanen et al., 2012) (Table 4).
The reduction in SS in the treatment wetlands was generally asgood as in the OFAs (50–80%) (Heikkinen et al., 2002; Kløve et al.,2012; Tuukkanen et al., 2012). The average SS concentration ininflow was 20 mg/l in the 11 treatment wetlands constructed ondrained peatlands studied here (Table 4), which was very similar tothat in the OFAs (21 mg/l) (Kløve et al., 2012; Tuukkanen et al.,2012). The reduction, in particulate total P, correlates with thereduction of SS (rs = 0.76). The inorganic N reduction by the treat-ment wetlands studied was also at the same level as in OFAsestablished on undisturbed peatland areas, but their total N
t wetlands. Numbers of replicates (n) between the P pools at different sites are also shown.
Ca–P Org.-P Tot Pa n
mg/kg % mg/kg % mg/kg
246 17 1225 83 1477 5123 13 508 52 968 3333 25 891 66 1352 3308 19 1134 71 1607 5
Table 4Water purification efficiency in the 11 wetlands studied. A negative value indicates leaching of the substance from the wetland.
Wetland and
established
in year
Monito
ring
period
Sampling
frequency
pHin pHa nb Total
SSin
(mg/l)
Total
SS
(%)
nb Total
Nin
(µg/l)
Total
N
(%)
nb Inorg.
Nin
(µg/l)
Inorg.
N
(%)c
nb Total
Pin
(µg/l)
Total
P
(%)
nb Part.e
Tot.
Pin
(µg/l)
Part.e
Tot.
P
(%)
nb PO4–
Pin
(µg/l)
PO4–
P
(%)
nb CODMn
in
(mg/l)
CODMn
(%)
nb Fein
(µg/l)
Fe
(%)
nb
Hankilaneva
1 1992
20.5.–
28.6.10
About
every two
weeks
6,7 1 16 75 4 780 19 4 216 64 4 94 82 4 31 83 4 75 90 4 16 –2 3 11397 76 4
Hankilaneva
2 1992
19.5–
25.8.92
Two times
per month
(not in
July)
6,6 5 16 79 5 1734 –10 5 700 93d 5 65 –199 5 n.a. n.a. 28 –203 5 36 –248 5 n.a. n.a.
8.6.–
2.9.09
About once
per month
7,0 0 4 16 59 4 1117 26 4 238 50 3 78 47 4 n.a. n.a. 44 48 3 18 –37 4 10067 19 3
20.5.–
7.9.10
About once
per month
7,1 5 18 73 6 1116 40 6 377 90 4 82 67 6 85 94 1 56 77 4 22 3 6 10475 42 4
Kapustane–
va 2008
18.5. –
7.9.09
About
every two
weeks
6,6 8 10 75 8 2014 24 8 691 85 8 100 37 8 34 48 3 20 66 8 55 –31 8 3296 57 8
20.5–
6.9.10
About
every two
weeks
6,5 8 7 38 8 2599 27 8 1196 57 8 95 37 8 33 61 8 21 49 7 61 –27 8 3438 43 8
Luomaneva
1998
10.8.–
20.9.99
About
every two
weeks
n.a. n.a. 18 16 4 1990 6 4 400 53d 2 98 –138 4 n.a. n.a. n.a. n.a. 31 –102 4 4050 –119 2
23.7.–
2.9.09
About
every two
weeks
6,9 4 22 63 4 1910 32 4 286 42 4 118 –1 4 51 42 4 26 –188 4 40 10 4 3050 –12 4
20.5–
7.9.10
About
every two
weeks
6,8 9 25 75 9 1712 36 9 431 61 4 125 24 9 71 62 8 32 –108 9 44 22 9 3178 –13 9
Savaloneva
2005
13.5.–
16.9.09
Weekly 6,0 11 20 –134 17 3023 –39 17 1506 7 8 62 –201 17 35 –134 11 7 –908 16 53 –43 10 6454 –262 14
11.5.–
8.9.10
About once
per month
6,4 3 13 –112 5 3322 –26 5 2136 12 3 55 –204 5 24 –304 4 10 –614 5 50 –54 5 5220 –146 5
Äijönneva
2009
22.7–
15.9.09
About
every two
weeks
6,6 5 11 44 5 2128 –45 5 960 71 5 68 –269 5 23 –102 4 20 –565 5 37 –226 5 4660 –67 5
18.5.–
8.9.10
About
every two
weeks
6,6 9 17 54 9 2651 11 9 1099 59 9 126 –42 9 52 15 9 50 –75 9 47 –51 9 4778 3 9
Iso–
Lamminne–
va 2010
12.7–
4.10.10
About once
per mo nth
6,6 4 13 4 4 2445 –27 4 1225 21d
4 47 –2 4 n.a n.a. n.a. n.a. 30 –136 4 6325 –61 4
Itäsuo 1995 week
23–38
1995
Weekly n.a. n.a. 158 97 16 4262 66 16 n.a. n.a. 226 45 16 n.a. n.a. n.a. n.a. 97 27 16 n.a. n.a.
Kuljunneva
2009
15.9–
13.09
About once
per month
6,0 2 11 69 2 2450 7 2 970 69d 2 44 –98 2 n.a. n.a. n.a. n.a. 25 –269 2 n.a. n.a.
Lumiaapa 2
1996
10.6–
15.9.98
1–3 times
per month
n.a. n.a. 72 88 10 6316 20 10 2307 11d 10 84 50 10 n.a. n.a. 12 68 10 53 10 10 2370 38 10
7.6.–
14.9.99
About
every two
weeks
6,9 9 12 73 9 3985 24 9 1978 39d 9 45 8 9 n.a. n.a. 7 35 9 31 –33 9 2333 45 9
29.5.–
11.9.00
About
every two
weeks
7,3 9 12 74 9 2740 32 9 1772 62 9 35 13 9 n.a. n.a. 8 25 9 19 –32 9 3100 60 9
23.6.–
20.8.08
About
every two
weeks
7,0 5 23 96 5 2620 71 5 985 97 2 32 42 5 n.a. n.a. 17 65 2 21 12 5 5300 92 2
19.5–
14.7.09
About once
per month
7,0 3 13 81 3 2090 57 3 975 98 3 37 57 3 n.a. n.a. 14 45 3 25 11 3 4300 89 3
Pohjoinen
Latvasuo
10.6.–
1.9.96
About once
per month
n.a. n.a. 7 13 4 1095 29 4 415 85d 4 57 36 4 n.a. n.a. 25 56 4 19 –7 4 n.a. n.a.
1994 9.6.–
1.9.97
About once
per month
n.a. n.a. 5 15 4 1125 20 4 306 87d 4 55
80
29 4 n.a. n.a. 29 38 4 15 –15 4 n.a. n.a.
Average 20 262400 960 44 37 5200
a↑ = pH lower in incoming water, ↓ = pH lower in outgoing water, 0 = same pH in incoming and outgoing water.bNumber of sampling times.cInorganic N = NH4–N + NO2 + 3–N.dInorganic N = only NH4–N.eParticulate.
38 H. Postila et al. / Geoderma 228–229 (2014) 33–43
purification efficiency was partly slightly lower (about 30–55%) thanthat in OFAs (Heikkinen et al., 2002; Kløve et al., 2012; Tuukkanenet al., 2012). The inflow concentration of N at OFA sites was 2 mg/l(Kløve et al., 2012; Tuukkanen et al., 2012), i.e. slightly lower thanat the 11 sites studied here (2.4 mg/l) (Table 4).
Total N leaching was detected in some of the wetlands studied im-mediately after establishment and rewetting, a phenomenon that isalso known to occur in the restoration of drained peatland areas(Koskinen et al., 2011). In drained peat, aeration leads to the oxidationof N to NO3–N (Kløve, 2001). Upon rewetting, therefore, large amountsof NO3 can be leached. In our study the retention of inorganic N was
usually good, so the leaching is probably due e.g. to organic N leachingwith organic matter. The N leaching was also mainly organic from re-stored drained peatland areas (Koskinen et al., 2011).
Leaching of Fe wasmainly associated with leaching of P in the treat-ment wetlands studied (Table 4). CODMn measurements indicatedleaching of organic matter (Table 4), probably mainly in the form ofdissolved organic matter. All these results indicate leaching of organicFe–P colloids (Heikkinen and Ihme, 1995) in the study areas. The rea-sons for this leaching are not yet precisely known, but one reasonmight be chemical reduction of Fe3+ to Fe2+ in anoxic peat layers ofthe wetland. Increased leaching of organic matter has also been
y = -0.005x + 94.438R² = 0.0548
0
10
20
30
40
50
60
70
80
90
100
0 500 1000 1500 2000 2500 3000 3500
LO
I (%
)
Total P (mg/kg)
Fig. 3. Relationship between loss of ignition (LOI, %) and total elemental phosphorus (P) content (mg/kg) in surface peat samples.
39H. Postila et al. / Geoderma 228–229 (2014) 33–43
observed after restoration or rewetting of drained peatland areas(Koskinen et al., 2011). Poor organic matter reduction is common intreatment wetlands constructed on peatland areas, where humicsubstances are constantly released to runoff waters by peat decomposi-tion. In OFAs, 20% organic matter reductions have been reported(Heikkinen et al., 2002). Riparian buffer zones established on peatsoils, used to purify peatland forestry runoff, often also release organicmatter or can retain only a few per cent of the total amount in theincoming water (Nieminen et al., 2005).
High hydraulic conductivity values indicate the depth of peat layerwhere the water flows in the wetland and where contact can beassumed between the runoffwater and the peat. In OFAs, high hydraulicconductivity has been observed until about 60 cm depth, whichcoincides with the typical acrotelm depth in pristine mires (Ronkanenand Kløve, 2005). In some of the treatment wetlands studied here, thehydraulic conductivity was observed to decrease strongly already at10 or 20 cm (Fig. 4). This is most likely due to a change in the acrotelmpeat properties and soil structure after drainage, by using circa60–100 cm deep ditches. The acrotelm disappears and the topsoillayer changes considerably after drainage and results in compaction ofthe topsoil layer caused by i) loss of buoyancy leading to rapid consoli-dation and soil compaction, and ii) oxidation, CO2 losses and slower soillosses. The surface vegetation also changes. Our results indicate that asthe high conductivity layer (acrotelm) disappears, the high conductivitylayer is also reduced, to 10–20 cm. Moreover, the observed values aresimilar to hydraulic conductivity values measured in forestry drainedpeatlands (Marttila and Kløve, 2010). This indicates that treatmentwetlands established on drained peatland areasmay often have a small-er active flow field in peat layers than undrained peatlands, and thusless space for purification processes to take place. However, it shouldbe noted that in half of the studied wetlands, especially in those thathad been in use for a long time, the peat hydraulic conductivity washigh even in the deeper layers (at least down to 40–70 cm) so thatthe potential flow depth could also be deeper (Fig. 4). The reason forthis could be that these older sites have recovered from drainage orwere initially constructed on peat soils that had been less influencedby drainage.
It should be noted that the water purification efficiency of the wet-lands studied here was affected not only by wetland properties, butalso e.g. by the inflow concentrations, the age of treatment wetlandsand climate. When the incoming water concentrations were very low,near the natural background concentrations, the reduction in the wet-lands was also low. Low incoming water concentrations were observede.g. at Hankilaneva 1, where the average N concentration in inflowwas780 μg/l in 2010 or at Lumiaapa 2, where the average PO4–P concentra-tion in inflow was 7 μg/l in 1999. If input concentrations are average orhigher, e.g. Total P in Hankilaneva 2 (inflow concentration 82 μg/l, puri-fication efficiency 67%) and Luomaneva (inflow concentration 125 μg/l,purification efficiency 24%) in 2010, then the purification efficiency isaffected more by wetland properties than by inflow concentrations. It
is possible that purification efficiency decreases after many years,when e.g. PO4–P sorption capacity decreases (Heikkinen et al., 2002),but this is not yet visible in our results. Furthermore, climate (mainlythe variable amount of precipitation and runoff between years) can af-fect the purification efficiency and result in between-year variations(Heikkinen et al., 2002).
Sampling date and frequency also affect the results of purificationefficiency between years, because in some years samples might betaken duringpeakflow situations and in other years during low flow sit-uations. The denser the sampling, the more reliable the results are. Inwetlands withmineral soil contact (Table 2), some of the water enteredthe groundwater below the peatland. This was observed at Luomanevaand Savaloneva, where the outgoing water volume was only a part ofthe incoming water volume. Mineral soil contact might also changethe adsorption properties, but the details are not well known. It wasfound that the values of wetland slope, treatment wetland area of thecatchment area, average peat thickness and average degree of humifica-tion of surface peat varied betweenwetlands, but this did not have a sig-nificant impact on the purification results (Tables 1, 2 and 4).
It should be noted that besides retaining nutrients, use of treatmentwetlands can also have a greenhouse gas effect on the environment(Mander et al., 2011; Søvik et al., 2006). Liikanen et al. (2006) reportedhigher emissions of methane (CH4) and nitrous oxide (N2O) in OFAsthan in natural peatlands. The CH4 increase is due to more anaerobicconditions after waterlogging. Increased input of N to OFAs comparedwith pristine peatlands can result in more N2O being released per unitarea (Liikanen et al., 2006; Mander et al., 2011; Søvik et al., 2006).
4.2. P in peatlands and factors affecting leaching or retention of P
Inorganic P compounds in acidic soils are commonly associatedwiththe oxides of trivalent metals like Fe and Al. On the other hand, in alka-line soils the adsorption of P is regulated mainly by Ca compounds. Or-ganic P forms are generally associated with living organisms, e.g.microbes, algae, vegetation, detritus and soil organic matter, and canbe divided into easily decomposable and slowly decomposable P(Reddy and DeLaune, 2008). Metals such as Fe and Al are also knownto be associated with organic matter in humus-rich soils (Giesler et al.,2005). Organically bound metals are suitable surface sites for inorganicdissolved P. Organic Fe–P colloids and particulate P in runoff waterscan also be precipitated to, or filtered in, wetlands. The binding of P topeat is strongly dependent on the physical conditions such as pH andredox conditions. Soil P in peatlands can be divided into: 1) looselyadsorbed and pore water-soluble P; 2) the redox-sensitive fraction of Pthat is bound to easily reducible metal oxides, e.g. Fe(III) oxides; 3) Pbound to oxides of non-reducible metals, e.g. P in non-reducible Fe(III)and Al; 4) P bound to Ca-containing minerals, such as apatite, mainlyin alkaline conditions; 5) mobile organic P bound mainly to metal-humate complexes; and 6) immobile organic P. The main P forms
40 H. Postila et al. / Geoderma 228–229 (2014) 33–43
studied here were loosely bound P, Fe + Al–P, Ca–P and Org.-P. The im-mobile Org.-P and redox-sensitive P fractions were not determined.
According to our P fractionation data, the P was mainly bound to Feand Al humates in peat (Org.-P) in all four wetlands where these Psorption properties were measured (Table 3). It should also be notedthat P can be a structural part of organicmaterial, as in immobile organicP. The mineral soil content was not correlated with the P content(Fig. 3), indicating the importance of organically bound Fe and Al in Padsorption (Giesler et al., 2005). The proportion of Ca-bound P was ap-proximately 17% in Hankilaneva 1, 25% in Savaloneva and 19% inÄijönneva (Table 3). Adsorption of P to Ca is most active in alkaline con-ditions (pH ≥ 7) (Reddy and DeLaune, 2008; Søvik and Kløve, 2005).Hence acidic runoff water is known to increase leaching of P from theCa-bound P pool (Nieminen and Penttilä, 2004; Reddy and DeLaune,2008). The outgoing runoff water was acidic (pH ≈ 5.8) at Savalonevaand Äijönneva in the monitoring period in summer 2009 and 2010,whereas runoff water was neutral (pH 7.2) at Hankilaneva 1. A charac-teristic in common for the first two sites was the observed leaching oftotal and inorganic P during the monitoring period (Table 4), whilethe Hankilaneva 1 treatment wetland seemed to retain both forms ofP in the early summer of 2010. The acidic nature of runoff water couldbe onepossible explanation for the detected desorption and enrichmentof P in outgoing runoff water at Savaloneva and Äijönneva. The inorgan-ic oxides (Fe + Al–P) and organically bound fractions of Al and Fe (in-cluded in the Org.-P fraction) are commonly the main fractionsbearing the responsibility for P adsorption in acidic conditions, whichis the case in Kapustaneva (pH ≈ 6.5 in incoming and pH ≈ 5.7 in out-goingwater, (Fe + Al–P) = 316 mg/kg). The strong change in pH dur-ing passage through the wetland reflects the acidic conditions in thetreatment structure. The Fe + Al-bound P pool was low at theHankilaneva, Savaloneva and Äijönneva sites, indicating thatleaching of P can be a problem also at Hankilaneva 1 if the acidityof the runoff water increases.
4.3. Wetland characteristics indicating P retention capacity oftreatment wetland
The results indicate that it is possible at least on a rough level to eval-uate prior risks of P leaching from the studied treatment wetland typeson the basis of tree stands in the peatland area (see 4.1). Leaching of Pseemed to be highest when the treatment wetland area was coveredwith a dense tree stand (Tables 2 and 4). Dense coverage with trees inawetlandmay indicate that either drainage has been at least partly suc-cessful, increasing the growth of trees in the area, or that these areasmay have been originally covered with dense tree stands before drain-age activities. In the case of wetlands without trees or with only sparse
0
10
20
30
40
50
60
70
801.0E-08 1.0E-07 1.0E-06 1.0E-05
Dep
th (
cm)
Hydraulic conductivity (
Fig. 4. Median hydraulic conductivity of
tree stands, drainage has probably not had such a strong influence onthe hydrology, vegetation and physical characteristics of the peatlandarea. These mainly treeless areas in our study were almost exclusivelycovered with wetland species such as Carex and Sphagnum, indicatingsmall abiotic changes. According to Haapalehto et al. (2011), changesin the species composition of drained peatland areas reflect changes intheir abiotic factors. Thus the findings in the present study indicatethat the best performing treatment wetlands for peat extraction runoffwater purification, especially in terms of P, appear to be those thathave similar characteristics (e.g. as regards vegetation) to pristinepeatlands.
If the drained peatland area is treeless and exclusively wetlandspecies are present, then water management or use of the area as atreatment wetland does not significantly alter the peatland ecosystemin the area, unlike treatment wetlands which have a considerable num-ber of forest species in the area. At sites with forest stands, the risingwater table may cause death of forest species, which are then graduallyreplaced by wetland species. Extensive cover of forest species indicatesa highly increased aerobic soil environment in comparisonwith originalpeatland environments. More anaerobic conditions in soil increase Feand P mobilisation. The death of forest species and release of nutrients,especially from their roots, is also possible. Koskinen et al. (2011)observed that after restoration of a nutrient-rich, minerotrophic anddensely tree-covered drained peatland area, leaching of inorganic Nwas stronger than from ombrotrophic and less densely tree-coveredarea. Drained forested areas have also been observed to leach P whenused as riparian buffer zones at forest drainage sites (Nieminen et al.,2005).
The trophic level of treatment wetlands, indicated by their vegetationcharacteristics, did not correlatewithnutrient retentionor leaching in thisstudy. Nieminen and Penttilä (2004) observed that changes in differenttrophic levels of the drainage site changed the binding of P and the ratioof detected P pools. A very high proportion of easily soluble (easily plantavailable) P in peat has been observed at oligotrophic sites, while highertrophic level wetlands have more Ca- and Fe-bound P than oligotrophicones (Nieminen and Penttilä, 2004). In the present study, the site withthe lowest trophic level (Kapustaneva) contained larger amounts ofloosely bound P, but also more Fe + Al–P than the other three sites.
The element ratio (Fe + Al + Mn)/P in surface peat samples wasassessed as a possible tool to predict the suitability of a peatland areafor water purification. The hypothesis was that a peatland area withhigh P content and low (Fe + Al + Mn)/P ratio in the surface peat isnot a suitable site for a treatment wetland. The results for most of thewetlands studied here indicated that this hypothesis is correct: low(Fe + Al + Mn)/P ratio (≤25) and quite high P content (N1200 mg/kg)were detected mainly in those treatment wetlands (Luomaneva,
1.0E-04 1.0E-03
m/s)
Hankilaneva 1
Hankilaneva 2
Kapustaneva
Luomaneva
Savaloneva
Äijönneva
Iso-Lamminneva
Itäsuo
Kuljunneva
Lumiaapa 2
PohjoinenLatvasuo
the peat profile in the 11 wetlands.
Is there pristine condition, good area for overland flow purposes available?
Yes No
Compliance with the design instructions of overland flow area
Search for suitable drained peatland area
Is the peat layer thick enough so that the ditches do not penetrate the peat layer?
NoYes
Preferred choice Secondary choice
Vegetation survey
Sparse tree stand/no trees Dense tree stand
Preferred choice Higher risk of nutrient (especially phosphorus) leaching
Analysis of the mineral content of surface peat
P content is low and there is at least some Fe, Al, Mn, Ca or Mg
which may retain phosphorus or
if there is lot of P, but the (Fe+Al+Mn)/P ratio is high
High amount P and (Fe+Al+Mn)/P ratio is low
Probability of retaining nutrients (P) in runoff
Higher risk of phosphorus leaching
Use other areas if possible Wetland planning for this area
Fig. 5. Decision tree analysis for evaluating the suitability of previously drained peatland areas for use as treatment wetlands.
41H. Postila et al. / Geoderma 228–229 (2014) 33–43
Savaloneva, Äijönneva and Kuljunneva)where leaching of Pwas observed(Tables 2 and 4). High (Fe + Al + Mn)/P ratio (N45) seemed to indicategood or at least moderate P removal efficiency in cases where the treat-ment wetland had been established and rewetted less recently, as in theHankilaneva 1, Hankilaneva 2 and Pohjoinen Latvasuo wetlands. Basedon the Median test, the groups differed significantly (p = 0.029) fromeach other. At Kapustaneva, Itäsuo and Lumiaapa 2, the P content andthe (Fe + Al + Mn)/P ratio in the surface peat samples were quite low(P content b 800 mg/kg and (Fe + Al + Mn)/P ratio b 25). Despite this,the detected removal efficiencies for different fractions of P were quitegood. This is perhaps attributable to the small amount of P in the peat.Itäsuo and Lumiaapa 2 also had high (Ca + Mg)/P ratios and, at leastLumiaapa 2, quite neutral water (pH ≈ 7) (Table 4). This indicatesthat Ca could provide P adsorption capacity at these sites and couldexplain the higher P removal efficiencies. However, more data from a
larger group of treatment wetlands in different parts of Finland areneeded in order to allow final conclusions to be drawn on the impact of(Fe + Al + Mn)/P ratio and perhaps in some situations (Ca + Mg)/Pratio on P retention.
4.4. Evaluation of the suitability of a drained peat area for waterpurification purposes
Based on the results obtained, a conceptual decision tree was drawnup to help in designing and establishing treatmentwetlands on drainedpeatland areas. In the decision tree, different factors influencing thewater purification efficiency of a treatment wetland (especially in thecase of P) can be taken into account (Fig. 5). The tree can then be usedas a preliminary assessment to evaluate how suitable the drainedpeatland area potentially is for runoff treatment purposes in peat
42 H. Postila et al. / Geoderma 228–229 (2014) 33–43
extraction areas. However, in the tree it is assumed that drainedpeatland areas are selected for wetland purposes only if pristinepeatland areas are not available. The first step in evaluating a drainedpeatland area should be to assess the role of the mineral soil layer. Ifditches in the drained peatland area are cut into the mineral soil, thereis a risk that some of the purified water might be lost to mineral soilgroundwater during use of the area as a treatment wetland. As thegroundwater impacts are not known, these sites should be avoided ormeasures taken to reduce such uncontrolled loss of water. Secondly, atree stand survey should be performed. Sites with significant amountsof tree stands should be avoided. Thirdly, surface peat samples shouldbe collected and analysed for the most important metals affecting P re-tention (Al, Fe, Mn, Ca, Mg), as well as for their total elemental contentof P. By using this decision tree model, it is possible to construct a treat-ment wetland with a high possibility for good treatment efficiency, es-pecially in terms of P retention. However, it should be noted that thismodel cannot guarantee a certain level of purification efficiency, be-cause efficiency depends also on various local, hydrological and geo-chemical factors.
5. Conclusions
This study shows that treatment wetlands constructed on drainedpeatland areas can be used to remove nutrients and SS from runoff wa-ters originating from peat extraction sites. However, there were largevariations in purification efficiency between the study sites, and somewetlands even released P and Fe. Increased chemical oxygen demandvalues in the runoff water flowing through the wetlands were also ob-served. These results indicate that drainage may strongly change theproperties of a peatland area, e.g. peat hydraulic conductivity, activerunoff layer in the peat profile and vegetation such as tree stands grow-ing in the area. Our findings indicate that the best performing treatmentwetlands for purification of peat extraction water appear to be thosedrained peatland areas that have similar characteristics (e.g. mire vege-tation) to overland flow areas constructed on pristine peatlands.
The P retention capacity of wetlands seemed to be affected by vari-ous factors and mechanisms, by the geochemical properties of thepeat (available metal oxides, metal humates and P pools), and by thevegetation (presence of tree stands). Some wetland sites leach P imme-diately after rewetting and construction, but after the initial P losses, theP retention capacity can be recovered. The retention of P dependspartly on local conditions, which need to be further considered.Based on the prerequisites identified, such as sparse tree stand, high(Fe + Al + Mn)/P ratio (N45) or low P content and at least some Fe,Al, Mn, Ca, Mg in surface peat, we drew up a conceptual decision treeto guide the construction of treatment wetlands with high possibilitiesfor good treatment efficiency. However, this decision support treemust be refined and validated by collecting new data from variousother treatment wetlands in the future.
Acknowledgements
The studies weremostly conducted in the “Round-year treatmentof runoff from peat production areas (TuKos)” project. This projectwas funded mainly by the European Regional Development Fundand the Centre for Economic Development, Transport and the Envi-ronment for North Ostrobothnia. Other funding bodies and partnersin project were Vapo Oy, Turveruukki Oy, Kuopion energia Oy,Suomen turvetuottajat ry (Finland's peat producers), Jyväskylänenergia Oy, City of Oulu, Council of Oulu Region, and Regional Coun-cil of Central Finland. The studies were also funded by the FinnishCultural Foundation, K.H Renlund Foundation, Maa-ja vesitekniikantuki ry and Sven Hallin Foundation. We thank all funders. Wewould also like to thank Juha Riihimäki, Mika Visuri, Annika Vilmiand Raimo Ihme from the Finnish Environment Institute (SYKE)and Noora Laurila, Sari Kantonen, Virve Kupiainen, Jussi Härkönen,
Tuomo Reinikka, Tuomo Pitkänen and Paavo Vehkomäki from theUniversity of Oulu for their valuable assistance in the studies. Inaddition, we would like to thank the staff at the Finnish EnvironmentInstitute Laboratory for their laboratory services.
References
Braskerud, B.C., 2002. Factors affecting phosphorus retention in small constructed wet-lands treating agricultural non-point source pollution. Ecol. Eng. 19, 41–61.
Burke, W., 1978. long-term effects of drainage and land use on some physical propertiesof blanket peat. Ir. J. Agric. Res. 17, 315–322.
Dunne, E.J., Reddy, K.R., Carton, O.T., 2005. NutrientManagement in AgriculturalWatershed—
A Wetlands Solution. Wageningen Academic Publishers, The Netherlands.Eurola, S., Huttunen, A., Kukko-oja, K., 1994. Suokasvillisuusopas. Oulanka Reports 13 (in
Finnish).Giesler, R., Andersson, T., Lövgren, L., Persson, P., 2005. Phosphate sorption in aluminum-
and iron-rich humus soils. Soil Sci. Soc. Am. J. 69, 77–86.Golterman, H.L., 1996. Fractionation of sediment phosphate with chelating com-
pounds: a simplification, and comparison with other methods. Hydrobiologia335, 87–95.
Grybos, M., Davranche, M., Gruau, G., Petitjean, P., Pédrot, M., 2009. Increasing pH drivesorganic matter solubilization from wetland soils under reducing conditions.Geoderma 154, 13–19.
Haapalehto, T.O., Vasander, H., Jauhiainen, S., Tahvanainen, T., Kotiaho, J.S., 2011. The ef-fects of peatland restoration on water-table depth, elemental concentrations, andvegetation: 10 years of changes. Restor. Ecol. 19, 587–598.
Heikkinen, K., Ihme, R., 1995. Retention of organic Fe–P-colloids from peat mining waterin an overland flow wetland treatment system in northern Finland. Arch. Hydrobiol.134, 547–560.
Heikkinen, K., Ihme, R., Lakso, E., 1994. Processes contributing to the retention of nutri-ents, organic matter and iron in an overland flow wetland treatment system. Publ.Water Environ. Adm. A 193 (in Finnish with English summary).
Heikkinen, K., Ihme, R., Osma, A.-M., Hartikainen, H., 1995. Phosphate removal by peatfrom peat mining drainage water during overland flow wetland treatment.J. Environ. Qual. 24, 597–602.
Heikkinen, K., Savolainen, M., Ihme, R., Lakso, E., 2002. The use of overland flow wet-land treatment system for the purification of runoff water from peat miningareas — the service life of the system. In: Mander, U., Jenssen, P.D. (Eds.), NaturalWetlands for Wastewater Treatment in Cold Climates. WIT Press, Southampton,Boston, pp. 29–45.
Heikkinen, K., Karjalainen, S.M., Ihme, R., 2009. The effects of peat extraction on water-ways and water protection. Vesitalous 50, 50 (abstract in English).
Henze, M., Harremoës, P., la Cour Jansen, J., Arvin, E., 2002. Wastewater Treatment: Bio-logical and Chemical Processes, third ed. Springer, Berlin.
Hobbs, N.B., 1986.Miremorphology and the properties and behaviour of some British andforeign peats. Q. J. Eng. Geol. Hydrogeol. 19, 7–80.
Holden, J., 2009. Upland hydrology. In: Bonn, A., Allott, T., Hubacek, K., Stewart, J. (Eds.),Drivers of Change in Upland Environments. Routledge, Abingdon and New York,pp. 113–134.
Holden, J., Chapman, P.J., Labadz, J.C., 2004. Artificial drainage of peatlands: hydro-logical and hydrochemical process and wetland restoration. Prog. Phys. Geogr.28, 95–123.
Huttunen, A., Heikkinen, K., Ihme, R., 1996. Nutrient retention in the vegetation of anoverland flow treatment system in northern Finland. Aquat. Bot. 55, 61–73.
Ihme, R., 1994. Use of the overland flow wetland treatment system for the purification ofrunoff water from peat mining areas. VTT Publications 798 (in Finnish with Englishsummary).
Ihme, R., Heikkinen, K., Lakso, E., 1991. The use of overland flow for the purification ofrunoff water from peat mining areas. Publ. Water Environ. Res. Inst. 9, 3–24.
Kadlec, R.H., 1999. Chemical, physical and biological cycles in treatment wetlands. WaterSci. Technol. 40, 37–43.
Kadlec, R.H., Knight, R.L., 1996. Treatment Wetlands. Lewis Publisher, New York.Kadlec, R.H., Wallace, S.D., 2009. Treatment Wetlands, second ed. CRC Press, Boca Raton, FL.Kjaergaard, C., Heiberg, L., Jensen, H.S., Hansen, H.C.B., 2012. Phosphorus mobilization in
rewetted peat and sand at variable flow rate and redox regimes. Geoderma173–174, 311–321.
Kløve, B., 2000. Effect of peat harvesting on peat hydraulic properties and runoff genera-tion. Suo, Peat and Mires 51, 121–129.
Kløve, B., 2001. Characteristics of nitrogen and phosphorus loads in peat mining waste-water. Water Res. 35, 2353–2362.
Kløve, B., Saukkoriipi, J., Tuukkanen, T., Heiderscheidt, E., Heikkinen, K., Marttila, H., Ihme,R., Depre, L., Karppinen, A., 2012. Pre-estimation of loading onwater bodies from peatproduction and new pollution control methods (TuVeKu). The Finnish Environment35/2012 (in Finnish with English summary).
Koskinen, M., Sallantaus, T., Vasander, H., 2011. Post-restoration development of organiccarbon and nutrient leaching from two ecohydrologically different peatland sites.Ecol. Eng. 37, 1008–1016.
Kuschk, P., Wießner, A., Kappelmeyer, U., Weißbrodt, E., Kästner, M., Stottmeister, U.,2003. Annual cycle of nitrogen removal by a pilot-scale subsurface horizontal flowin a constructed wetland under moderate climate. Water Res. 37, 4236–4242.
Liikanen, A., Puustinen, M., Koskiaho, J., Väisänen, T., Martikainen, P., Hartikainen, H.,2004. Phosphorus removal in a wetland constructed on former arable land.J. Environ. Qual. 33, 1124–1132.
43H. Postila et al. / Geoderma 228–229 (2014) 33–43
Liikanen, A., Huttunen, J.T., Karjalainen, S.M., Heikkinen, K., Väisänen, T.S., Nykänen, H.,Martikainen, P.J., 2006. Temporal and seasonal changes in greenhouse gas emissionsfrom a constructedwetland purifying peatmining runoff waters. Ecol. Eng. 26, 241–251.
Mander, Ü., Maddison, M., Soosaar, K., Karabelnik, K., 2011. The impact of intermittent hy-drology and fluctuating water table on greenhouse gas emissions from subsurfaceflow constructed wetlands for wastewater treatment. Wetlands 31, 1023–1032.
Marttila, H., Kløve, B., 2010. Managing runoff, water quality and erosion in peatland for-estry by peak runoff control. Ecol. Eng. 36, 900–911.
Ministry of the Environment, 2007. Finnish Government decision-in-principle on WaterProtection Policy Outlines to 2015. The Finnish Environment 10/2007 (in Finnishwith English summary).
Ministry of the Environment, 2013. Guidelines for environmental protection in peat min-ing. Environmental Administration Guidelines 2/2013 (in Finnish with Englishsummary).
Niedermeier, A., Robinson, J.S., 2007. Hydrological controls on soil redox dynamics in apeat-based, restored wetland. Geoderma 137, 318–326.
Nieminen, M., Penttilä, T., 2004. Inorganic and organic phosphorus fractions in peat fromdrained mires in northern Finland. Silva Fenn. 38, 243–251.
Nieminen, M., Ahti, E., Nousiainen, H., Joensuu, S., Vuollekoski, M., 2005. Does the use ofriparian buffer zones in forest drainage sites to reduce the transport of solids simul-taneously increase the export of solutes? Boreal Environ. Res. 10, 191–201.
Närhi, P., Räisänen,M.L., Sutinen, M.-L., Sutinen, R., 2012. Effect of tailings onwetland veg-etation in Rautuvaara, a former iron–copper mining area in northern Finland.J. Geochem. Explor. 116–117, 60–65.
Pant, H.K., Reddy, K.R., 2003. Potential internal loading of phosphorus in a wetland con-structed in agricultural land. Water Res. 37, 965–972.
Päkkilä, J., 2008. Remediation of arsenic and nitrogen from mine drainage water in anoverflow wetland in the Suurikuusikko gold mine. University of Oulu, Departmentof Process and Environmental Engineering, Water Resources and Environmental En-gineering, Oulu, Master thesis (in Finnish with English summary).
Reddy, K.R., DeLaune, R.D., 2008. Biogeochemistry of Wetlands: Science and Applications.CRC Press, Boca Raton, FL, Taylor & Francis Group.
Richardson, C.J., Nichols, D.S., 1985. Ecological analysis of wastewater managementcriteria in wetland ecosystems. In: Godfrey, P.J., Kaynor, E.R., Pelczarski, S.,Benforado, J. (Eds.), Ecological Considerations in Wetlands Treatment of MunicipalWastewaters. Van Nostrand Reinhold, New York, pp. 351–391.
Ronkanen, A.-K., Kløve, B., 2005. Hydraulic soil properties of peatlands treating municipalwastewater and peat harvesting runoff. Suo, Peat and Mires 56, 43–56.
Ronkanen, A.-K., Kløve, B., 2007. Use of stabile isotopes and tracers to detect preferentialflow patterns in a peatland treating municipal wastewater. J. Hydrol. 347, 418–429.
Ronkanen, A.-K., Kløve, B., 2009. Long-term phosphorus and nitrogen removal processes andpreferential flow paths in Northern constructed peatlands. Ecol. Eng. 35, 843–855.
Silvan, N., Vasander, H., Laine, J., 2004. Vegetation is the main factor in nutrient retentionin a constructed wetland buffer. Plant Soil 258, 179–187.
Søvik, A.K., Kløve, B., 2005. Phosphorus retention processes in shell sand filter systemstreating municipal wastewater. Ecol. Eng. 25, 168–182.
Søvik, A.K., Augustin, J., Heikkinen, K., Huttunen, J.T., Necki, J.M., Karjalainen, S.M., Kløve,B., Liikanen, A., Mander, Ü., Puustinen, M., Teiter, S., Wachniew, P., 2006. Emissionof the greenhouse gases N2O and CH4 from constructed wetlands in Europe.J. Environ. Qual. 35, 2360–2373.
Tuukkanen, T., Kløve, B., Marttila, H., Heikkinen, K., Karppinen, A., 2012. Processes and sitecharacteristics controlling nutrient and sediment runoff loads from peat harvestingsites. Extended abstract No 312, 14th International Peat Congress.
Vasander, H., Tuittila, E.-S., Lode, E., Lundin, L., Ilomets, M., Sallantaus, T., Heikkilä, R.,Pitkänen, M.-L., Laine, J., 2003. Status and restoration of peatlands in northernEurope. Wetlands Ecol. Manag. 11, 51–63.
Werker, A.G., Dougherty, J.M., McHenry, J.L., Van Loon, W.A., 2002. Treatment variabilityfor wetland wastewater treatment design in cold climates. Ecol. Eng. 19, 1–11.
Wörman, A., Kronnäs, V., 2005. Effect of pond shape and vegetation heterogeneity on flowand treatment performance of constructed wetlands. J. Hydrol. 301, 123–138.
