Vol. 36 No. 4 2010 2

POLISH ACADEMY OF SCIENCES INSTITUTE OF ENVIRONMENTAL ENGINEERING COMMITTEE OF ENVIRONMENTAL ENGINEERING

ARCHIVES OF ENVIRONMENTAL PROTECTION

vol. 36 no. 4 2010 2

ARCHIVES OF ENVIRONMENTAL PROTECTION vol. 36 no. 4 2010

CONTENTS

The Occurrence of Trace Elements in Flue Gases from Circulating Fluidized Bed Boilers – Jan Konieczyński, Katarzyna Stec ...... 3 Structure of Rotifer Communities of Restored Small Peat-Bog Resevoirs of Poleski National Park – Andrzej Demetraki-Paleolog, Monika Tarkowska- Kukuryk ...... 21 Nitrogen Compounds in Well Water as a Factor of a Health Risk to the Maciejowice Commune Inhabitants – Jolanta Raczuk ...... 31 PAH Soil Concentrations in the Vicinity of Charcoal Kilns in Bieszczady – Ewa Lisowska ...... 41 Relations between Circulation and Winter Air Pollution in Polish Urban Areas – Jolanta Godłowska, Anna Monika Tomaszewska ...... 55

Ventilation Control based on the CO2 and Aerosol Concentration and the Perceived Air Quality Measurements – a Case Study – Bernard Połednik, Marzenna Dudzińska ...... 67 Sorption of Ibuprofen on Sediments from the Dobczyce (Southern Poland) Drinking Water Reservoir – Katarzyna Styszko, Katarzyna Sosonowska, Piotr Wojtanowicz, Janusz Gołaś, Jerzy Górecki, Mariusz Macherzyński . . 81 Organics (COD) Removal in dependence on loading of Volatile Fatty Acids (VFA) – Katarzyna Bernat, Irena Wojnowska-Baryła, Adriana Dobrzyńska . . . . 93 LCC Analysis of Rainwater Utilization System in Multi-Family Residential Buildings – Daniel Słyś, Tadeusz Bewszko ...... 107 DGGE-based Monitoring of Bacterial Diversity in Activated Sludge Dealing with Wastewater Contaminated by Organic Petroleum Compounds – Aleksandra Ziembińska, Sławomir Ciesielski, Jarosław Wiszniowski ...... 119

PL ISSN 0324-8461 3

ARCHIVES OF ENVIRONMENTAL PROTECTION vol. 36 no. 4 pp. 3 - 19 2010 PL ISSN 0324-8461

THE OCCURRENCE OF TRACE ELEMENTS IN FLUE GASES FROM CIRCULATING FLUIDIZED BED BOILERS

JAN KONIECZYŃSKI1, KATARZYNA STEC2

1 Institute of Environmental Engineering of the Polish Academy of Sciences M. Skłodowskiej-Curie str. 34, 41-819 Zabrze, Poland e-mail: [email protected] 2 Institute of Glass, Ceramics, Refractory and Construction Materials, Refractory Materials Branch Toszecka str. 99, 44-100 Gliwice, Poland e-mail: [email protected]

Keywords: Hazardous air pollutants, air contamination with dust, coal combustion, CFB boilers, grain size composition of flue ash, trace elements in the emitted dust, distribution in grain size fractions, emission factors.

Abstract: The emission of dust from coal fired furnaces introduces a lot of contamination into the environment, including dangerous metal compounds, which occur as trace elements in hard and brown coal. After the coal is burnt, they are contained in the grains of respirable dust, which creates health hazard. The results of investigations into the distribution of several trace elements in granular composition of ash emitted from CFB boilers used in coal-fired heat and power station are presented. The research material was taken by means of a cascade impactor, enabling a different granulometric fraction to be separated from a stream of dust that penetrated the electrofilter. The ICP-AES method (inductively coupled plasma atomic emission spectroscopy) was used to determine trace elements after prior mineralization of samples by microwave method. The Authors presented the results of measurements and analyses, determining the ranges of trace elements’ occurrence in dust, characterizing the distribution in PM1, PM2.5 and PM10 granulometric fractions and determining the emission factors.

INTRODUCTION

The achievements in the protection of air against contamination cannot make us blind to the fact that the hazard caused by the emission of relatively small amounts of substances which, due to particularly dangerous properties, create hazard to people’s health still exists. These include heavy metals. When introduced into the air from natural and anthropogenic sources, they are subject to wet or dry deposition. Their harmful effect on the biological functions in the human organism, into which they penetrate chiefly by respiratory tracts, but also via the alimentary system while participating in particular links of food chain, is enhanced by the heavy metals’ ability to move between particular components of the environment. The air emission is one of the main reasons for the presence of heavy metals in soil, waters and plants. These substances include trace elements and their com- 4 JAN KONIECZYŃSKI, KATARZYNA STEC pounds. In the USA, twelve of them have been identified as hazardous air pollutants: Sb, As, Be, Cd, Cu, Cr, Co, Pb, Mn, Hg, Ni and Se. These elements, most frequently in a form of different compounds, are emitted into the atmosphere from various types of furnaces and boilers which burn coal [11]. On a global scale a reduced emission of heavy metals is being observed [28], which chiefly results from the progress in flue gas dedusting and the use of cleaner technologies. In Poland, an inventory of contamination emission is taken, which serves the needs of national statistics and, as National Reference Center, serves European institutions such as EEA, EMER or IPCC. Table 1 presents data on the annual emission of heavy metals in Poland. It has to be added that the emission varies depending on the region – most sources of the discussed emissions are located in the Upper Silesian and Lower Silesian Provinces [10].

Table 1. Total emission of heavy metals in Poland according to the types of activity in 2006 [Mg]

Specification As Cr Zn Cd Cu Ni Pb Hg Professional powerhouses 2.8 3.5 26.1 0.2 8.9 7.3 10.7 8.3 and CHP plants District heat generating 1.4 1.9 55.1 2.2 0.7 7.5 13.6 0.8 plants Municipal heat 1.0 1.4 41.5 1.6 6.0 5.4 10.4 0.2 generating plants Housing and services 13.1 15.6 465.2 20.1 68.9 66.2 115.9 1.0 Combustion processes in boilers, turbines and 1.1 1.4 42.0 1.7 6.0 7.4 10.4 0.5 engines in the industry Production processes 1.0 9.3 171.4 2.6 18.3 6.8 90.1 1.4 Road transport 0.0 2.0 0.0 0.3 2.9 4.9 17.5 0.0 Waste burning 0.0 0.0 0.9 0.1 0.1 0.0 1.4 0.1 Total 46.3 46.8 1303.2 42.2 344.7 177.5 524.2 21.3

Combustion of fuels is one of the main sources of heavy metals’ emission, though the previously mentioned progress in flue gas dedusting, involving the use of electrofilters or bag filters, reduces the emission in large combustion plants. It also has to be emphasized that flue gas desulphurization installations (FDG) which are commonly usedin these plants ensure an effective second stage flue gas dedusting and contribute to a further drop in metal emission. The interest in air pollution due to coal burning is justified in such countries as Poland because of a particularly high – reaching 62% – contribution of this fuel in the current use of power resources [10]. The combustion of coal in different furnaces causes the emission of ca 270 000 Mg/a, which accounts for 61% of dust emission from all the sources in Poland, amounting to 443 000 Mg/a [10]. The above mentioned interest is justified especially that on a global scale the position of hard and brown coals is strong and has a rising tendency. This results from the analyses forecasting an increased use of coal – from 6.3 billion Mg in 2010 to 9.6 billion of Mg in 2030 [36]. THE OCCURRENCE OF TRACE ELEMENTS IN FLUE GASES FROM CIRCULATING FLUIDIZED... 5

TRACE ELEMENTS IN COAL, THE EMISSION OF TRACE ELEMENTS IN PROCESSES INVOLVING THE USE OF COAL

Trace elements originate from Carboniferous era plants which gave rise to coal beds. With regard to quantity, the concentration of trace elements in coals from various basins and beds is very differentiated. Recent investigations of coal from less known regions: China, North Korea, revealed a general similarity of all coals in the world: a considerable enrichment in ash of potentially dangerous elements and most of them being bound with inorganic matter [31, 44]. Their amount and composition is considerably influenced by the enrichment processes in later geological eras [35]. It has been found that some trace elements occur in coal beds in higher concentrations than those in the earth crust. The degree of trace elements’ accumulation in coal ashes is characterized by an enrichment coefficient expressed by the ratio of trace element concentration in coal ash to its average concentration in the earth curst. The content of trace elements in hard and brown coals has been thoroughly examined. In their earlier studies the Authors investigated the composition of coals used in Polish power industry [33] and coke plants [43]. Table 2 presents the ranges of concentrations of selected trace elements in the ash from 35 power coal samples from many Upper Silesian mines. The ranges have been reduced so as to present the analyses of 80% of ash samples included in the investigations.

Table 2. Ranges of concentrations of the selected trace elements in power ashes of Upper Silesian coals [ppm]

Element Content Ag 0–2 As 0–36 Ba 373–3737 Cd 0–9 Co 44–173 Cr 82–306 Cu 98–466 Mn 136–1586 Mo 4–16 Ni 40–364 Pb 50–200 Rb 74–119 Sb 0–3 Sn 0–4 Sr 659–2960 V 182–597 Zn 117–3242 6 JAN KONIECZYŃSKI, KATARZYNA STEC

After the coal is burnt, the mineral substance contained in it forms slag and flue ash. The latter is composed of different size grains, of the diameter ranging from submicrons to 100 mm, which are lifted from the furnace with a flue gas stream. The character of flue ash granulometric composition is bimodal [41]. The factors influencing the formations of flue ash particles are still the subject of investigations [3]. Flue ash is chiefly composed of: oxides of calcium, silicon, iron, aluminium, magnesium, sodium and potassium as well as numerous secondary or trace components. They include the compounds of chlorine and fluorine, as well as compounds of trace elements: antimony, arsenic, barium, beryllium, boron, bromine, chromium, tin, zinc, cadmium, cobalt, manganese, copper, molybdenum, nickel, lead, mercury, selenium, thallium, vanadium and others. In coal they are chiefly bounded with mineral matter, and partly with organic matter in a form of metal-organic compounds, organic acid salts and others. Depending on the behavior in the process of coal combustion, the elements which form mineral matter fall into the category of non-volatile (I), including aluminium, calcium, potassium, magnesium, silicon, iron, titanium, thorium and others; volatile elements which condense on ash particles (II), including, among others, barium, chromium, manganese, sodium, beryllium, cobalt, copper, nickel, phosphorus, vanadium, arsenic, cadmium, molybdenum, lead, antimony, thallium, zinc; and volatile elements which es- cape in a form of vapors (III) – these include: boron, bromine, chlorine, fluorine, mercury, iodine and selenium. The elements belonging to I group are those which do not evaporate during combustion, their concentration is similar in all solid products of combustion. The remaining elements evaporate in the boiler. Vapors of some of them, after the waste gas has been cooled, condensate mainly on the finest particles of flue ash, whose specific surface area is the biggest [23]. The elements which condensate inside the installation are included in group II. The elements which are emitted in a form of vapors belong to group III [35]. The composition and quantity of trace element compounds released into the atmosphere with flue gas depend on coal burnt [22]. The enrichment of fine particles through the condensation of trace components [8, 15–17, 25], which has been confirmed in many studies, may be limited by competitive reactions on the surface of bigger grains [32]. The flue ash produced in the combustion of biofuels is marked by differences in the manner of enriching the flue ash granular fractions [4, 5]. Despite the observed deviations, general tendencies in the distribution of trace elements in three streams have been determined, chiefly for solid waste (ash and slag), and partly for waste emitted into the air (particulate matter and gases or vapors) [20]. Investigations into the process of pyrolysis of Upper Silesian coke coals have provided information on the behavior of trace elements [43]. In the process of pyrolysis, trace elements were released with volatile products of pyrolysis. The degree of release was differentiated: the highest for mercury, cadmium and lead, medium – for selenium, arsenic, nickel and manganese, and the lowest – for beryllium.

FLUE GASES DEDUSTING, TOTAL AND FRACTIONAL EFFICIENCY

The characteristic feature of all dust collecting devices is fractional efficiency, specific for a given type of device, which drops with the decreasing size of dust grains. Even in the case of the highest technically possible total collection efficiency, the fractional collection efficiency of a dust fraction with grain size lower than 0.4 μm does not exceed 95%. THE OCCURRENCE OF TRACE ELEMENTS IN FLUE GASES FROM CIRCULATING FLUIDIZED... 7

The electrofilters achieving very high total collection efficiency display a minimum efficiency when collecting grains whose diameter ranges from 0.1 do 2 μm [40]. This results from the limited electric charging capacity in the case of dust particles of this size.

In consequence, the mass contribution of PM2.5 in the emitted dust is increased, even by 50%. The efficiency of capturing many trace elements (occurring in the solid phase) is close to the total efficiency of dust collection. Some elements (manganese, beryllium, chromium) are slightly enriched; a bit higher enrichment was observed in the case of antimony, lead, nickel and selenium, while the highest – in the case of arsenic and cadmium [35]. More trace elements were found in the flue dust collected in a bag filter than in the one collected by an electrofilter [9]. Although the collection efficiency has a major effect on the size of trace elements emission, in the case of arsenic the effect of waste gas temperature has a strong influence [12]. A relative drop in the collection efficiency of fine dust fractions is not so high in the case of bag filters. However, even when PTFE membranes are used, the efficiency of filtration drops as the size of grain decreases [1]. The common use of flue gas desulphurization installation has also contributed to the reduced dust emission. A flue gas desulphurization (FGD) installation, most frequently by wet lime method placed behind the electrofilter, is at the same time the second stage of dust collection, the general efficiency of which reaches 80%. Behind the FGD the mass contribution of submicron fraction increases to 50%; however, this concerns the residual flue dust emission, which accounts for maximum 0.05% of the total amount of ash produced in the coal combustion process. Nevertheless, trace elements deposited on fine dust particles penetrate and are released into the atmosphere air [27]. The emission rates for elements characterized by low volatility depend directly on the efficiency of flue gases dedusting and cleaning [39]. The evaluation of the impact exerted by power plants using hard or brown coal is greatly dependent on the emission of PM2.5 dust due to the presence of dangerous substances in dust grains: PAHs, dioxins and heavy metals, and because of this dust capability to penetrate into the human organism through respiratory tracts. The presence of respirable dust in the air poses a health hazard. It is estimated that in the USA 15 000 premature infants die for this reason every year [19]. As regards power and heat boilers, it may be said that the use of an effective electrofilter and wet desulphurization process reduces the threat caused by the emission of trace elements [26, 34]. As trace elements accumulate in the products of dust removal and flue gas desulphurization, more attention has been recently paid to the effects of the washing out of toxic elements, including trace elements from the transported [30], deposited and utilized ashes, both fresh and old, obtained at different temperatures of coal combustion [6]. These investigations help get to know the manner of binding trace elements to organic and mineral coal matter [7, 38]. Compared to hand-fired furnaces, furnaces equipped with mechanical stocker and pulverized coal boilers, the emission of pollutants from fluidized bed boilers is not thoroughly investigated, and these boilers have a great prospect of being widely used for the combustion of coal, sludge, muds and waste fuels in industry boilers, power plants or power and heat plants. Combustion in a fluidized bed has evolved, starting with the technology of atmospheric fluidized bed combustion (AFBC), through pressurized fluidized bed combustion (PFBC), and finishing with combustion in a multi-solid fluidized bed (MSFB). Also combustion of coal and waste fuels in a fluidized bed with internal 8 JAN KONIECZYŃSKI, KATARZYNA STEC circulation of granular material – Twin Interchanging Fluidized Bed, or with heat ex- changers placed in the fluidized bed – Internally CFB has been mastered. Combustion in a circulating fluidized bed with the participation of an alkaline additive, a relatively low temperature of combustion and the use of high-efficient dedusting devices allows the ecological requirements to be satisfactorily fulfilled [18]. There are only few papers on trace elements in fly ash from CFB boilers. The newest one concerns leaching characteristics of metals in fly ash [42].

SELECTION OF PLANTS, CHARACTERISTICS OF CFB BOILERS AND FUEL BURNT

The presented investigations into the emission of selected trace elements have been carried out in four heat plants equipped with circulating fluidized bed (CFB) boilers, where an additive of limestone has been used to remove sulphur dioxide from waste gas. These are modern, well maintained plants which work in a combined system and play an important role in heat production in the Silesian province, inhabited by ca 5 million people in the south-western part of Poland. Tables 3 and 4 present technical parameters of the examined CFB boilers and coal burnt. Plant I is Elektrociepłownia Tychy SA (EC Tychy), Plant II – Elektrociepłownia Chorzów Elcho Sp. z o.o. (EC Elcho), Plant III – Południowy Koncern Energetyczny SA Elektrownia Jaworzno III (Elektrownia II) and Plant IV – Południowy Koncern Ener- getyczny SA Elektrociepłownia Katowice (EC Katowice).

Table 3. Technical parameters of the tested boilers

Boiler parameters Plant I Plant II Plant III Plant IV CFB Aker Kvaer- 2 × CFB 2 × CFB CFB Type of boiler ner ASA Foster Wheeler Foster Wheeler Foster Wheeler

40 MWe, 135 MWe, Boiler power output 274 MWt 180 MWt 70 MWt 200 MWt Type of boiler CFB Cymic CFB Compact CFB Compact CFB Fuel consumption 25.0 Mg/h 56.1 Mg/h 33.7 Mg/h 68.3 Mg/h Efficiency 88.5% 90.7% 91% 92% Nominal steam 135 Mg/h 404 Mg/h 260 Mg/h 483 Mg/h efficiency: Year of start-up 2000 2003 1999 1999 Mean stream of flue 3 3 3 3 gas volume during 157730 m n/h 420340 m n/h 261610 m n/h 556650 m n/h investigations Mean concentration 3 3 3 3 of dust in the emitted 26.5 mg/m n 21.9 mg/m n 12.0 mg/m n 26.4 mg/m n flue gas Density of dust (trapped in the 2555 kg/m3 2646 kg/m3 2701 kg/m3 2757 kg/m3 electrofilter) THE OCCURRENCE OF TRACE ELEMENTS IN FLUE GASES FROM CIRCULATING FLUIDIZED... 9

Table 4. Parameters of fuel burnt in the tested boilers

Plant III Plant IV Fuel parameters Unit Plant I Plant II coal sludgea coal sludgea Calorific value kJ/kg 19915 19775 18942 9000 19454 9053 Humidity content % 10.51 14.90 20.90 42.00 13.80 40.70 Sulphur content % 0.92 1.22 1.26 0.80 1.17 0.73 Ash content % 24.78 14.90 15.67 28.00 19.70 26.10 Volatile matter % 28.91 25.08 23.26 no data 27.00 no data a a mixture of coal and post-flotation coal sludge, the contribution of sludge corresponded to 30% of the chemical energy introduced with the fuel into the boiler

METHODS

Measurement of the emitted dust granulometric composition and dust sampling The emitted dust granulometric composition was measured by determining the mass of dust collected in particular stages of cascade impactor, introduced into the flue gas chan- nel behind the device for flue gas removal. A six-stage impactor Andersen Mark III with a backup filter arresting the finest dust grains was applied. The masses of separated grain fractions (6 + 1) were calculated into mass contributions and size composition was determined. The cut-off diameter at 50% efficiency of separation depends on the stream, density and temperature of waste gas aspirated through the impactor. In order to average the results for each plant and to enable a comparison of the granulometric composition of dust measured in particular plants, the results of all measurements were referred to the same unified fraction size ranges. For the needs of this work, four such fraction intervals: < 1; 1–2.5; 2.5–10 and > 10 μm were adopted. The cumulative values and fractions were calculated by interpolation method for those particular ranges.

Determination of trace elements in dust Before determining the content of trace elements by ICP AES method, the samples of flue dust were digested in a one-station mineralizer with focused micro-wave energy. The samples of the tested ashes were placed in a vessel and 15 cm3 of concentrated suprapure

HNO3 was added [24]. After the reaction vessel was closed, the microwave system worked for 60 minutes at 600 W. The process temperature was 483 K (with an admissible 10% measurement error), the pressure in the vessel was 15 bar. When the digestion process was completed and the inside of the reaction vessel reached the ambient temperature, the vessel was opened and the samples placed on a quantitative microporous filter paper. The filtrate was collected into a measurement flask, after supplementing it with deionised water to reach the volume of 50 cm3. The analytes were determined by means of ICP AES with optical detection. The residue on the filter paper consisted of filter paper remains with deposited dust. During the sample analysis by ICP AES method a glass Mainhard atomizer was used and optimized measurement parameters were applied [14]. The glass Mainhard atomizer is the most commonly used type of concentric atomizer. In this atomizer a sample is placed right into the centre of an argon stream, which enables achieving high measurement stability and sensitivity of determination [13]. 10 JAN KONIECZYŃSKI, KATARZYNA STEC

Apart from the selection of appropriate atomizer, the major problem in the process of optimizing the parameters of emission spectrometer’s work is to establish the optimal velocity of carrier and spraying gas flow. These parameters should be set so as to obtain a maximum ratio of peak intensity of the determined element to the intensity of the sample background for each analyzed element [37]. The emission lines were selected according to the Polish Standard (PN-EN ISO 11885, 2001) [29]. The spectrometer was calibrated using the certified multi-element solutions produced by Merck company. Concentrations used for calibrations reached 1 ppm, 5 ppm, 10 ppm and 15 ppm.

RESULTS

Concentrations of the selected trace elements in dust In the process of measuring the granulometric composition of the emitted dust by means of a cascade impactor, 84 samples of dust were obtained for investigations into trace elements’ content. The measurements consisted of three series, each containing 7 granular fractions from each of the four tested plants. The mass of single samples was small, ranging from a fraction to several mg. In order to ensure the required concentration of analytes in the mineralization solution, the samples containing grains with approximate diameters – 1 μm, 1–2.5 μm and > 2.5 μm were mineralized together. The concentration of the examined trace elements in granular fractions of the emitted dust has a very wide range. Table 5 presents the full range of the measured concentrations [ppm] and the range limited to values found in 80% of the examined samples.

Table 5. Full and reduced range of the examined trace elements’ concentrations in dust samples

Full range Reduced range Element average max min min max Be 16 176 1 2 52 Cd 4 64 0 0 12 Co 8 82 0 0 24 Cr 2872 205432 0 8 1175 Cu 706 14627 0 0 1033 Hg 68 618 0 0 205 Mn 1834 116615 0 35 1518 Mo 30 748 0 0 98 Ni 1917 86271 0 9 1996 Pb 14495 514038 5 28 1213 Sb 435 14763 0 0 944 Se 251 2366 0 8 693 Tl 343 7555 0 2 838 V 243 3636 0 0 538 W 54 1993 0 0 114 Zn 1566 18462 0 34 2961 THE OCCURRENCE OF TRACE ELEMENTS IN FLUE GASES FROM CIRCULATING FLUIDIZED...11

Table 6 presents concentrations of selected trace elements in real granular fractions taken at the tested plants as well as the calculated enrichment coefficient (Rj) for particular elements, determined from the ratio of concentration in the fraction containing the smallest and in the fraction containing the largest grains, according to the following formula:

(1) where: th Cj min – concentration of j trace element in the fraction containing the smallest grains, th Cj max – concentration of j trace element in the fraction containing the largest grains.

The content and contributions of selected trace elements in dust fractions On the basis of the determined trace elements’ concentrations in particular dust fractions separated by means of an impactor and unified fractions, the content (Tab. 7) and cumulated mass contributions (Mj) of trace elements in unified PM1, PM2.5 and PM granular fractions contained in the emitted dust (Tab. 8) were determined. The presented values have been calculated in the following way:

Dji=Cji · fi (2) where: th th Dji – content of j trace element bound to the i fraction in the emitted dust [mg/g], th th Cji – concentration j of trace element in the i fraction [mg/g], fi – fraction contribution. n

D= Dji (3) Σi=1 where: D – content of jth trace element in the emitted dust [mg/g]. D M = ji (4) ji D

th th Mji – cumulated mass contribution of j trace element in the i fraction of the emitted dust.

Emission factors (EF) of selected trace elements The obtained results were used to calculate the emission factors for particular trace elements. To this end the emission factors of EFPM1, EFPM2.5 and EFPM in 1 kg of the emitted dust in relation to 1 Mg of the burned fuel (Tab. 9), determined as a result of measurements taken in the tested plants [18] were used. Basing on the dust emission factors and the content of the examined trace elements in relevant unified granular fractions of the dust, the emission factors EFj for particular trace elements were calculated, both for PM1,

PM2.5 and PM fractions. The obtained results have been given in Table 10. 12 JAN KONIECZYŃSKI, KATARZYNA STEC j R 1.11 0.77 0.19 1.16 0.95 0.51 0.58 0.86 0.18 0.44 0.62 0.82 2.26 0.41 8.49 1.66 12.42 51.75 > 2.90 175.96 279.47 248.42 178.03 399.54 786.66 289.82 2587.70 2088.79 3539.98 1.63 27.78 43.65 55.55 21.03 111.11 484.11 226.18 309.51 595.21 EC Katowice 1150.75 1190.43 1305.50 1357.09 1.46–2.90 4.88 3.74 14.37 31.61 57.47 < 1.46 212.62 135.04 488.46 175.27 264.34 238.48 488.46 1057.37 1465.37 j R 4.48 4.15 0.83 2.95 0.39 1.28 1.12 2.76 1.20 0.98 1.39 1.24 0.98 0.12 8.21 2.24 24.63 30.10 60.20 57.46 > 3.30 125.87 875.61 845.51 205.22 889.30 175.12 328.36 1340.78 5.57 6.13 5.57 78.03 44.59 50.16 178.36 278.69 236.89 275.90 613.12 1700.01 1479.84 2452.47 EC Jaworzno II 0.68–3.30 6.18 6.72 24.19 29.57 59.14 < 0.68 564.48 102.14 725.76 327.94 263.43 999.94 209.66 456.96 1666.57 j R 0.54 0.10 1.45 1.86 0.98 1.18 1.22 0.86 0.93 1.65 1.06 1.28 1.63 15.08 7.37 10.84 58.53 43.35 13.01 > 2.80 563.59 351.16 444.37 541.92 888.74 2297.72 2276.05 2009.42 3273.17 EC Elcho 8.11 7.10 4.05 93.26 133.80 391.27 545.34 587.91 273.68 180.43 1135.28 1885.37 1709.00 3040.92 1.40–2.80 6.30 20.17 55.48 21.18 < 1.40 302.61 239.56 345.48 524.52 504.34 882.60 3303.46 2443.55 3479.98 1462.60 j R 1.84 3.20 2.65 1.06 0.30 1.95 1.31 1.01 2.22 1.24 2.13 1.55 1.56 14.79 5.72 6.94 1.04 32.95 84.98 31.22 > 2.53 190.77 953.84 360.73 520.28 1546.96 1097.79 1855.66 1352.73 EC Tychy EC 4.26 3.10 11.24 15.50 93.01 213.16 108.52 507.70 821.62 891.38 1085.16 2344.73 1007.65 2867.93 1.26–2.53 7.48 15.39 48.38 15.39 < 1.26 351.87 105.56 382.66 461.83 105.56 813.71 2529.09 2137.63 1869.33 2880.96 Concentrations of selected trace elements in real granular fractions (μm) emitted from a CFB boiler the examined plants [ppm] and enrichment coefficient Table 6. Table Element Hg W Sb Cu Cr Mn Tl Se Mo Co Zn Be Pb Cd THE OCCURRENCE OF TRACE ELEMENTS IN FLUE GASES FROM CIRCULATING FLUIDIZED...13 6.7 6.5 PM 52.9 16.3 500.5 213.3 186.9 139.8 412.6 532.7 729.0 1952.4 1872.3 2887.9 2.5 7.4 0.6 5.3 11.9 15.7 34.8 60.5 93.4 PM 292.3 125.6 331.4 371.7 163.9 344.8 EC Katowice 2 17 8.3 7.4 4.7 1.1 0.5 6.1 9.2 0.2 0.1 1 36.8 17.0 51.0 PM 8.4 3.4 PM 57.6 42.3 33.7 41.6 202.4 163.3 674.3 216.4 407.1 1074.8 1044.1 1638.6 2.5 2.7 1.9 1.8 15.9 81.1 76.1 25.3 12.9 88.3 74.2 PM 468.0 427.7 179.5 709.3 EC Jaworzno 1 3.3 5.6 1.6 1.3 0.3 0.4 11.6 31.2 40.1 18.1 14.5 55.2 PM 25.2 92.0 7.1 PM 11.3 10.7 101.7 554.4 326.2 287.1 366.6 499.4 1059.6 2290.6 1960.0 1087.6 3213.1 2.5 6.8 4.0 5.3 83.5 89.8 95.7 PM 327.1 262.5 219.3 313.0 714.7 1117.1 1335.8 1840.0 EC ELCHO 1 9.4 3.4 1.1 3.6 85.8 51.5 40.8 58.8 89.3 PM 562.3 150.2 415.9 248.9 592.3 PM 61.2 8.45 4.93 80.44 12.37 54.85 234.13 428.06 740.23 1496.59 1340.38 1009.73 1848.13 2320.04 2.5 9.85 6.37 4.55 PM 49.85 164.8 68.46 23.97 822.25 993.75 296.97 447.56 1449.2 551.16 1828.46 EC Tychy EC 1 3.23 PM 1.57 3.23 10.15 73.82 22.15 530.6 22.15 80.28 96.89 392.18 448.47 170.72 604.43 Content of selected trace elements in µg which are present unified granular fractions contained 1 g dust emitted from a CFB boiler the examined plants Table 7. Table Be Tl Hg W Sb Se Mo Cu Cr Mn Co Pb Cd Zn 14 JAN KONIECZYŃSKI, KATARZYNA STEC

Table 8. Cumulated mass contributions of the selected trace elements to unified granular fractions of the dust emitted from the examined boilers

Element EC Tychy EC Elcho EC Jaworzno II EC Katowice

PM1 PM2.5 PM1 PM2.5 PM1 PM2.5 PM1 PM2.5 Zn 0.261 0.788 0.184 0.573 0.056 0.433 0.018 0.119 Cd 0.655 0.923 0.336 0.495 0.010 0.043 0.015 0.815 Pb 0.231 0.745 0.229 0.657 0.062 0.441 0.023 0.225 Co 0.186 0.754 0.155 0.563 0.088 0.559 0.030 0.090 Mn 0.243 0.784 0.212 0.570 0.053 0.410 0.020 0.199 Cr 0.096 0.443 0.179 0.627 0.067 0.343 0.017 0.622 Cu 0.188 0.694 0.160 0.598 0.027 0.131 0.015 0.304 Mo 0.597 0.906 0.605 0.698 0.039 0.205 0.008 0.141 Se 0.789 0.944 0.453 0.524 0.067 0.262 0.002 0.144 Sb 0.425 0.837 0.497 0.757 0.054 0.327 0.022 0.226 W 0.404 0.437 0.523 0.914 0.047 0.383 0.008 0.085 Hg 0.261 0.796 0.301 0.602 0.155 0.321 0.031 0.454 Tl 0.396 0.741 0.245 0.583 0.037 0.435 0.009 0.048 Be 0.275 0.851 0.039 0.090 0.132 0.598 0.025 0.324

Table 9. EFPM1, EFPM2.5 and EFPM dust emission factors for the examined CFB boilers

Plant Emission factors [kg/Mg]

EFPM EFPM1 EFPM2.5 EC Tychy 0.167 0.035 0.107 EC Elcho 0.164 0.028 0.095 EC Jaworzno II 0.093 0.005 0.029 EC Katowice 0.215 0.007 0.061

EFjPMi = EFPMi Dj (5) where: th th EFjPMi – emission factor of j trace element in the i granular fraction of dust [mg/g], [mg/Mg], th EFPMi – emission factor of the i dust granular fraction [kg/Mg] th th Dji – content of j trace element bound to the i fraction in the emitted dust [mg/g], [mg/kg].

DISCUSSION

With a particular regard to dangerous substances it might be said that the concentration of beryllium (in ppm) ranges from 2 to 52, for cadmium it reaches 12, cobalt – 24, chromium – from 8 to 1175, mercury – 205, lead – from 28 to 1213, antimony – 944, and selenium – from 8 to 693 (Tab. 5, reduced range) The measured concentrations’ level does not depart from the one established in previous investigations into the ashes of Up- per Silesian power coals [33]. An attempt to evaluate the degree of fine dust fractions’ THE OCCURRENCE OF TRACE ELEMENTS IN FLUE GASES FROM CIRCULATING FLUIDIZED...15

Table 10. Emission factors of selected trace elements in unified granular fractions in mg per 1 Mg of burnt coal from a CFB boiler in the examined plants

Element EC Tychy EC Elcho EC Jaworzno II EC Katowice

EFjPM! EFjPM2.5 EFjPM EFjPM! EFjPM2.5 EFjPM EFjPM2.5 EFjPM2.5 EFjPM2.5 EFjPM2.5 EFjPM2.5 EFjPM Zn 21.2 195.6 387.4 16.6 174.8 526.9 0.5 20.6 152.4 0.4 21.0 620.9 Cd 0.1 0.5 0.8 0.1 0.5 1.8 0.0 0.1 3.9 0.0 0.3 1.4 Pb 6.0 59.0 123.6 7.0 67.9 178.4 0.1 5.2 37.9 0.1 10.0 156.7 Co 0.1 0.7 1.4 0.0 0.4 1.2 0.0 0.1 0.3 0.0 0.0 1.4 Mn 15.7 155.1 308.6 11.6 106.1 321.4 0.3 12.4 97.1 0.3 22.7 402.5 Cr 3.4 47.9 168.6 2.5 29.7 81.9 0.1 2.2 20.1 0.1 20.2 114.4 Cu 2.8 31.8 71.5 1.6 20.8 60.1 0.1 2.6 62.7 0.0 7.7 88.7 Mo 0.8 2.6 9.2 4.2 24.9 47.1 0.0 0.4 3.1 0.0 0.7 30.1 Se 0.1 1.1 2.1 0.1 0.6 1.9 0.0 0.1 0.8 0.0 0.4 3.5 Sb 18.6 106.3 223.8 15.7 126.9 375.7 0.2 13.6 100.0 0.1 5.7 419.8 W 0.8 7.3 13.4 1.1 9.1 173.8 0.0 0.7 3.9 0.0 3.7 40.2 Hg 2.6 17.6 39.1 1.4 8.5 53.5 0.2 2.2 15.2 0.1 2.1 45.9 Tl 13.7 88.0 249.9 2.4 31.1 90.9 0.1 2.3 18.8 0.1 17.8 107.6 Be 0.4 5.3 10.2 0.3 7.9 16.7 0.0 0.5 5.4 0.0 1.0 11.4 enrichment did not bring clear-cut results. Contrary to pulverized-fuel boilers, no enrichment of fine dust grains can be observed (Tab. 6). For the examined elements belonging to the second group according to Swain’s classification [35], the enrichment coefficient

(Ri) was not found to be higher than one in all 4 tested plants. The value of 1 in 3 of the examined plants was exceeded in the case of beryllium, cadmium, manganese, lead and zinc. It should be emphasized that significant differences were observed in the process of coal combustion in PC and CFB boilers, i.e. the temperature of combustion (1600 K and 1100 K respectively), the time of fuel grain retention (seconds and minutes), dust grain reaction (neutral/acid and alkaline), dust grain porosity (low and high). In the process of CFB combustion, an intense abrasion of dust grain, and in consequence renewing of grain surface was observed. The flue dust from CFB does not contain glassy phases characteristic of high-temperature processes; but there is gypsum, calcium and quartz as well as iron and magnesium oxides [2]. Reactions causing the redistribution of trace elements may occur on the renewed surface of dust grains due to continuous collisions and abrasion [32]. Co-burning of coal and coal sludge used in plants III and IV may also contribute to the changes in the manner of trace elements’ accumulation, which was observed in the case of co-burning with biomass [4]. From the point of view of environmental threat, the size of the examined trace elements’ contribution in granular fractions contained in the emitted dust is more important. In the opinion of the Authors, some interesting information is contained in Table 8. The presented values of contributions result from the processing of data on the content of the examined elements in unified granular fractions. The general observation is the difference in contributions calculated for coal-fired plants and plants fired 16 JAN KONIECZYŃSKI, KATARZYNA STEC with coal with addition of coal sludges. This difference cannot be clearly interpreted due to the low number of investigated examples. However, it should be emphasized that coal sludges containing much bigger amounts of mineral matter are a source of trace elements which are bound to this mineral matter and occur in inorganic compounds. In coal-fired boilers the contribution of PM2.5 fraction in the emission of the examined trace elements ranges from 0.443 to 0.944, and the contribution of PM1 fraction ranges from 0.096 to

0.789. Particularly high is the contribution of PM1 fraction in the emission of selenium, antimony, tungsten and thallium, and PM2.5 fraction in the emission of cadmium, cobalt, copper, mercury, manganese, molybdenum, lead, antimony, selenium, thallium and zinc. It should be noted that dangerous elements introduced into the air, namely, cadmium, cobalt, chromium, mercury, lead, antimony and selenium are present mainly in the respirable fraction. The size of these elements’ contributions is influenced by the previously mentioned processes of dust grain enrichment with trace elements due to condensation of vapors of these elements’ compounds or vapors of a given element. The fractional efficiency of dedusting devices is also significant. Electrofilters which achieve avery high total dedusting efficiency show minimum efficiency when removing the grains with diameters ranging from 0.1 to 2 μm [40]. In consequence, as observed in the case of two coal-fired boilers, the mass fraction of PM2.5 in the emitted dust increases to 64% [18]. A relative drop in the efficiency of fine dust fraction extraction is not so high in the case of bag filters. For this reason more trace elements were found in the flue ash collected in the bag filter compared to flue ash captured in the electrofilter [9]. One might then expect -fur ther progress in the reduction of dust emission from CFB boilers by using bag filters. This will allow achieving a dedusting efficiency comparable to that obtained for PC boilers equipped with FGD system by wet lime method, which is the second stage of dedusting. This does not eliminate completely the emission of submicron fraction which is accompanied by some amounts of trace elements [27], but it greatly reduced the threat [26, 34]. In order to evaluate the scale of the risk for the population due to emission of dust containing dangerous substances, the emission factors for selected trace elements from the examined plants were calculated (Tab. 10). Similarly to the previous observation regarding the contributions in granular fractions, a significant difference in the values of emission factors for coal-fired boilers and boilers fired with coal and coal sludge may be observed, with a similar efficiency of dust removal from the four examined plants. In the first group of plants the emission factors for the selected elements containing PM1 range from a fraction to 21 mg/Mg, and for those containing PM2.5 – from a fraction to nearly 200 mg/Mg. In the second group of plants the emission factors for the selected elements with PM1 do not exceed a fraction of mg/Mg, and with PM2.5 they range from a fraction to nearly 23 mg/Mg. The emission factors of chromium, lead and antimony in the respirable dust reaching the respective values of 30, 68 and 127 mg/Mg should be considered high. The example of mercury, whose average TSP emission factor reaches 38 mg/Mg for all the plants, is noteworthy. It proves that the emission of mercury even from modern power plants is high. The examined plants located within the area of ca 1000 km2, burning ca 106 Mg of coal annually, emit ca 40 kg of mercury in a solid phase. As the degree of binding mercury to the solid phase in a fluidized bed boiler is estimated to reach ca 50% [21], in the considered case the total emission of mercury may be evaluated as 80–100 kg/a. The example of mercury is important considering the fact that in 2020 a 50% reduction of mercury emission is planned to be achieved, as compared to 2000. THE OCCURRENCE OF TRACE ELEMENTS IN FLUE GASES FROM CIRCULATING FLUIDIZED...17

The results of investigations may contribute to updating the inventory of trace elements’ emission in Poland and Europe by specifying the dust emission factors, taking into account the real granular composition of the emitted dust. They will also facilitate a more accurate evaluation of the influence that the power plants exerts on the environment, including the effects of washing out toxic trace elements from the ashes transported for utilization or storage [6, 30].

CONCLUSIONS

Although the flue gas emitted from CFB boilers is removed by means of effective electrofilters, it contains dust grains, in which compounds of numerous trace elements, including elements considered as dangerous i.e. beryllium, cadmium, cobalt, chromium, mercury, lead, antimony and selenium, are present. The range of these elements’ concentrations does not differ from the one observed in the case of dust emitted from PC boilers. The examination of dust granular composition and the analysis of granular fractions, separated by means of a cascade impactor, revealed that the phenomenon of fine grains’ enrichment with some trace elements takes place also in CFB boilers. The investigations and analyses showed that a significant, or even a prevailing part of the emitted trace elements’ mass are released into the atmosphere with the respirable fraction of PM2.5 dust, which poses a potential threat to people’s health. The calculated emission factors for the investigated trace elements, expressed in mg per a ton of fuel burnt, are not very high, but they prove that modern CFB boilers, equipped with effective electrofilters cannot be omitted in the inventory of emission of trace elements, including heavy metals. The results of investigations will enable to accurately specify the size of global emission and will help evaluate the influence exerted by power plants on the environment and the selection of technical methods of reducing the harmful effect of these plants. Further investigations are needed to obtain more information on the behavior of trace elements contained in fuel in the process of combustion in a CFB boiler and their distribution in the products of combustion.

Acknowledgement In this work the Authors have used the results of investigations conducted within the scope of research project number 4 T10B O65 25 “Charakterystyka emisji pyłowych i ga- zowych substancji zanieczyszczających przy spalaniu węgla w paleniskach fluidalnych” (Characteristics of Emissions of Dust and Gaseous Substances Causing Contamination in the Process of Coal Burning in Fluidized Beds), financed by the Research Committee. The authors are indebted to Mr Bogusław Komosiński (M. Eng.) for his help in organ- izing the industrial measurements.

REFERENCES

[1] Bologa A., H.R. Paur, H. Seifert, K. Woletz: Novel wet electrostatic precipitation for sub-micron particles, Dust International Conference, Maastricht 2007. [2] Botha F.D.: Overview of the fluidized bed combustion process and material, in Proceedings of State Regulation of Coal Combustion By-Product Placement at Mine Sites, A Technical Interactive Forum, Harrisburg, PA, May 4–6, 2004. [3] Buhre B.J.P., J.T. Hinkley, B.P. Gupta, P.F. Nelson, T.F. Wall: Fine ash formation during combustion of pulverised coal-coal property impacts, Fuel, 85, 185–193 (2006). 18 JAN KONIECZYŃSKI, KATARZYNA STEC

[4] Camerani M.C., B.M. Steenari, R. Sharma, R. Beckett: Cd speciation in biomass fly ash after size separation by centrifugal split, Fuel, 81, 1739–1753 (2002). [5] Fernandez A., J.O.L.Wendt, N. Wolski, K.R.G. Hein, S. Wang, M.L. Witten: Inhalation health effects of fine particles from the co-combustion of coal and refuse derived fuel, Chemosphere, 51, 1129–1137 (2003). [6] Frieder E.S., G. Shelet, N. Kress: Chemical changes in different types of coal ash during prolonged, large scale, contact with seawater, Waste Management, 23, 125–134 (2003). [7] Gentzis T., F. Goodarzi: Chemical fractionation of trace elements in coal and coal ash, Energy Sources, 21, 233–256 (1999). [8] Giere R., L.E. Carleton, G.R. Lumpkin: Micro- and nanochemistry of fly ash from a coal-fired power plant, American Mineralogist, 88, 1853–1865 (2003). [9] Goodarzi F: Characteristic and composition of fly ash from Canadian coal-fired power plants, Fuel, 85, 1418–142 (2006). [10] GUS: Ochrona środowiska, Wydawnictwo GUS, Warszawa 2005. [11] Hazardous Air Pollutants – HAP’s, U.S. EPA: Study of hazardous air pollutant emissions from electric utility steam generating units –Final Report to Congress, EPA-453/R-98-004a (February 1998), (Utility Air Toxics Study), Exec. Summ., At ES-4 (1998). [12] Hughes I.S.C., R.F. Littlejohn: Emissions of trace elements from coal- fired industrial boilers, Interna- tional Journal of Energy Research, 12, 473–480 (2007). [13] Jarvis I., K.E. Jarvis: Inductively coupled plasma – atomic emission spectrometry in exploration geo- chemistry, Journal of Geochemical Exploration, 44, 139–200 (1992). [14] Jarvis I., K.E. Jarvis: Plasma spectrometry in earth sciences: techniques, applications and future trends, Chemical Geology, 95, 1–33 (1992). [15] Kaleta P., U. Tomza: Trace elements in size fractionated coal ash from two Polish power plants, Journal of Radioanalytical and Nuclear Chemistry, 121, 237–244 (1988). [16] Karayigit A.I., Y. Bulut, X. Querol, A. Alastuey, S. Vassilev: Mass balance of minor and trace elements in a coal-fired power plant, Energy Sources, 28, 1311–1320 (2006). [17] Karayigit A.I., Y. Bulut, X. Querol, A. Alastuey, S. Vassilev: Variation in fly ash composition from the Soma Power Plant, Turkey, Energy Sources, 27, 1473–1481 (2005). [18] Konieczyński J.: Emission of pollutants from circulating fluidized bed boilers, Works & Studies, Institute of Environmental Engineering of the Polish Academy of Sciences, Zabrze 2005 (in Polish). [19] Laden F., L. Neas, D. Dockery, J. Schwartz: Association of fine particulate matter from different sources with daily mortality in six US. cities, Environmental Health Perspectives, 108, 941–947 (2000). [20] Llorens J.F., J.L. Fernandez-Turiel, X. Querol: The fate of trace elements in a large coal-fired power plant, Environmental Geology, 40, 409–416 (2001). [21] Lopez-Anton M.A., J.M.D. Tascon, M.R. Martinez-Tarazona: Retention of mercury in activated carbons in coal combustion and gasification flue gases, Fuel Process. Technology, 53, 77–78 (2002). [22] Martinez-Tarazona M.R., Spears, D.A.: The fate of trace elements and bulk minerals in pulverized coal combustion in a power station, Fuel Process. Technology, 47, 79–92 (1996). [23] Mattigod S., D. Rai, J.E. Amonette: Concentrations and distribution of major and selected trace elements in size-density fractionated fly ashes, Biogeochemistry of Trace Elements in Coal and Coal Combustion Byproducts, Kluwer Academic, Plenum Publishers, 1999. [24] Matusiewicz H., R.E. Sturgeon: Present status of microwave sample dissolution and decomposition for elemental analysis program, Analytical Atomic Spectroscopy, 12, 21–39 (1989). [25] Mazur J., J. Konieczyński: Distribution of trace elements in granulometric fractions of fly ash emitted from power stations, Wydawnictwo Politechniki Śląskiej, Gliwice 2004 (in Polish). [26] Meij R., N. Winkel: The emissions and environmental impact of PM10 and trace elements from a modern coal- fired power plant equipped with ESP and wet FGD, Fuel Process Technology, 85, 641–656 (2004). [27] Nielsen M.T., H. Livbjerg, C.L. Fogh, J.N. Jensen, P. Simonsen: Formation and emission of fine particles from two coal –fired power plant, Combustion Science and Technology, 174, 79–113 (2002). [28] Pacyna J.M., E.G. Pacyna: An assessment of global and regional emissions of trace metals to the atmosphere from anthropogenic sources worldwide, Environmental Reviews, 12, 269–298 (2001). [29] PN-EN ISO 11885: Determination of 33 elements by inductively coupled plasma atomic emission spectroscopy, 2001. [30] Popovic A., D. Djordjevic, D.P. Polic: Trace and major element pollution originating from Coal Ash suspension and transport processes, Environment International, 26, 251–255 (2001). [31] Ren D., F. Zhao, Y. Wang, S. Yang: Distribution of minor and trace elements in Chinese coals, Interna- tional Journal of Coal Geology, 40, 109–118 (1999). THE OCCURRENCE OF TRACE ELEMENTS IN FLUE GASES FROM CIRCULATING FLUIDIZED...19

[32] Senior C.L., J.J. Helble, A.F. Sarofim: Emission or mercury, trace elements, and fine particles from sta- tionary combustion sources, Fuel Processing Technology, 65-66, 263–288 (2000). [33] Staisz J., A. Pasoń-Konieczyńska, J. Konieczyński: Preliminary estimation of trace elements emission during combustion of bituminous coal, Archives of Environmental Protection, 26, 1, 7–20 (2000). [34] Suarez A.E., J.M. Ondov: Ambient aerosol concentration of elements resolved by size and by source: Concentration of some cytokine-active metals from coal- and oil fired power plants, Energy Fuels, 16, 562–568 (2002). [35] Swaine D.J., F. Goodarzi: Environmental aspects of trace elements in coal, Kluwer Academic Publishers, The Netherlands, 1995. [36] The Future of Coal, An Interdisciplinary MIT Study, Massachusetts Institute of Technology, 2007. [37] Walsh J.N.: Handbook of inductively coupled plasma spectrometry, Blackie, Glasgow and London 1989. [38] Wang J., A. Sharma, A. Tomita: Determination of the modes of occurrence of trace elements in coal by leaching coal and coal ashes, Energy Fuels, 17, 29–37 (2003). [39] Weber G.F: A comprehensive assessment of toxic emissions from coal-fires power plans: Phase ,I U.S. Department of Energy, 1996. [40] White H.: Electrostatic precipitation of fly ash, Journal of the Air Pollution Control Association, 27, 15–21 (1977). [41] Yoo J., Y. Seo, T. Shinagawa: Particle –size distribution and heavy metal partitioning in emission gas from different coal – fired power plants, Environmental Engineering Science, 22, 272–279 (2005). [42] Yuan Ch-G.: Leaching characteristics of metals in fly ash from coal-fired power plant by sequential extraction procedure, Microchim. Acta, 165, 91–96 (2009). [43] Zajusz-Zubek E., J. Konieczyński: Dynamics of trace elements release in coal pyrolysis process, Fuel, 82, 1281–1290 (2003). [44] Zheng, J., B. Hu, R.B. Finkelman, B. Wang, M. Wang, S. Li, D. Wu: Concentration and distribution of sixty-one elements in coal from DPR Korea, Fuel, 85, 679–688 (2006).

Received: November 19, 2009; accepted: August 16, 2010.

WYSTĘPOWANIE METALI W SPALINACH Z KOTŁÓW Z CYRKULACYJNYM ZŁOŻEM FLUIDALNYM

Emisja pyłu ze źródeł energetycznych powoduje wprowadzenie do środowiska wielu zanieczyszczeń, w tym związków niebezpiecznych metali, występujących jako pierwiastki śladowe w węglu kamiennym i brunat- nym. Po spaleniu węgla znajdują się one w ziarnach pyłu respirabilnego, co stwarza zagrożenie dla zdrowia ludzi. Przedstawiono wyniki badań nad dystrybucją wybranych kilkunastu pierwiastków śladowych we frakcjach ziarnowych popiołu lotnego emitowanego z kotłów z cyrkulacyjnym złożem fluidalnym stosowanych w elektrociepłowniach opalanych węglem. Materiał badawczy został pobrany za pomocą impaktora kaska- dowego umożliwiającego wydzielenie ze strumienia spalin odpylonych w elektrofiltrze frakcji pyłu o różnej wielkości ziarna. Do oznaczenia pierwiastków śladowych wykorzystano metodę atomowej emisyjnej spektro- metrii o wzbudzeniu plazmowym ICP-AES po uprzedniej mineralizacji próbek metodą mikrofalową. Przed- stawiono wyniki pomiarów i analiz, określając zakresy występowania pierwiastków śladowych w popiele lot- nym, charakteryzując dystrybucję we frakcjach ziarnowych PM1, PM2,5 i PM10 i wyznaczając wskaźniki emisji. 20 21

ARCHIVES OF ENVIRONMENTAL PROTECTION vol. 36 no. 4 pp. 21 - 30 2010 PL ISSN 0324-8461

STRUCTURE OF ROTIFER COMMUNITIES OF RESTORED SMALL PEAT-BOG RESEVOIRS OF POLESKI NATIONAL PARK

ANDRZEJ DEMETRAKI-PALEOLOG*, MONIKA TARKOWSKA-KUKURYK

University of Life Sciences, Department of Hydrobiology Dobrzańskiego str. 37, 20-262 Lublin, Poland * Corresponding author e-mail: [email protected]

Keywords: Small water reservoirs, peatbogs, planktonic rotifers, Poleski National Park.

Abstract. Artificial small water reservoirs existing over peatbogs of Łeczyńsko-Włodawskie Lakeland have high ecological value. Planktonic rotifer communities of the reservoirs are characterized by high species diversity, presence of rare species as well as high abundance of periphytic and benthic-periphytic forms. Low nutrients concentrations of studied reservoirs confirmed low total rotifers density, presence of indicatory species and high ratio of algaevorous to detritivorous species. The analysis of domination structure of planktonic rotifers showed very low stability of their community and high vulnerability for changes of habitat conditions.

INTRODUCTION

Planktonic rotifers constitute an important link in trophic structure of water ecosystems. On the one hand, these organisms are consumers of bacteria, algae, protozoans and detritus, on the other, they become an important diet component of small invertebrates and juvenile fish [1, 11, 19]. The appearance of rotifer species and community structure depends on many environmental factors, such as temperature, oxygen, toxins, phospho- rous concentration and osmotic conditions [24]. Thus the composition and abundance of planktonic rotifers can be used to diagnose the water quality [12, 17, 21]. Small water reservoirs created as a result of human activity differ in relation to trophic conditions, surface area, depth and presence of aquatic vegetation [14, 18]. Spe- cific environmental conditions (nutrients concentrations, pH and water transparency) in such ecosystems strongly influence the presence of invertebrate communities [3, 27]. Artificial water bodies are typical for peat-bogs of Poleski National Park. They were created as a result of intensive meliorations during the 1970s and 1980s. Great trophic diversity of the reservoirs, the presence of basic, humic and eutrophic waters make them rare and valuable habitats to a variety of resident and migratory fish and waterfowl [23]. Some of reservoirs have been studied to recognize physical and chemical water properties, protozoans’ fauna and other water zoocenosis [16, 22, 23]. However, there is still lack of comparative studies of water zoocenosis, such as planktonic rotifers in different types of peat-bog reservoirs. 22 ANDRZEJ DEMETRAKI-PALEOLOG, MONIKA TARKOWSKA-KUKURYK

The aim of the study was to determine the characteristic of the Rotifera community structure of seven small restored water bodies (basic, eutrophic and humic) situated on peatbogs on the area of a national park. The results presented in the study constitute a part of the project entitled “Evaluation of the effects of restoration of water-peatbog ecosystems. Part 2.A. Evaluation of ecological status of water ecosystems based on abundance and species structure of zooplankton”.

STUDY AREA AND METHODS

The studies were performed on the area of Poleski National Park (the area of 9647.7 ha) created in 1990 [4]. This is the first national park in which peatbog ecosystems are protected [10]. Studies included seven small peat-bog reservoirs differed in surface area, water properties and species structure of vegetation (Tab. 1, Tab. 2). One reservoir (Moszne) classified as humic is surrounded by high moor; one reservoir (Jagodne) is eutrophic, situated on the area of transitional peat bog. The remaining five reservoirs are basic, located on the area of carbonate peat bog (Bagno Bubnów). Planktonic rotifers were studied in June and September of 2008. Rotifers were sampled by “Toń II” apparatus from the depth of 0 to 0.5 m. Each sample consisted of 10 cm3 of water. In both studied months and at each reservoir three samples were taken. Sampled water was sieved through the planktonic net of mesh size 25 µm and condensed to the constant volume of 100 cm3. The samples were preserved by Lugol`s liquid and next by 4% formaldehyde with glycerine. Collected rotifers were counted and identified under inverted microscope. A number of individuals was calculated per 1 dm3 of water. The normal distribution of collected data was verified using Shapiro-Wilk test. The significance of differences between rotifers densities in particular in peat-bog reservoirs was tested using non-parametric rang test of Kruskal-Wallis. All statistical analyses were made using SAS Programme. The similarity of rotifer communities was determined by calculating Jaccard index using cluster method performed by MVSP-3,1 Programme. The analysis of faunistic similarity of rotifer communities was determined using UPGMA method. Ad- ditionally the influence of dominating rotifer species on the similarity of rotifer communities was estimated using PCA analysis of MVSP-3,1 Programme. The analysis included: index of domination, sustainability of domination structure [Bielańska-Grajner 2005], index of Shannon-Wiener and classification of rotifer species into ecological groups [8, 19].

RESULTS AND DISCUSSION

The studied small water reservoirs were inhabited by 40 planktonic rotifer species (Tab. 3). The highest number of species 11-12 was noted in eutrophic peat bog reservoir Jagodne and one of basic peat bog reservoirs (BB2); the lowest number of species was observed in humic peat bog reservoir Moszne. In the remaining peat bog reservoirs, a total number of rotifer species ranged from 9 to 10 (Tab. 4). Species richness measured as value of Shannon-Wiener index was high, and showed little differences between basic reservoirs (values of H index = 2.4-2.8). In Moszne peat bog reservoir, the value of H index reached 1.9. The lowest Shannon index (H = 0.6) STRUCTURE OF ROTIFER COMMUNITIES OF RESTORED SMALL PEAT-BOG RESEVOIRS... 23 -3 3 N-NO mg dm 0.465 ±0.117 0.300 ±0.098 0.886 ±0.213 0.488 ±0.213 0.528 ±0.318 0.487 ±0.028 0.844 ±0.814 -3 4 N-NH mg dm 0.656 ±0.112 0.684 ±0.240 0.780 ±0.206 1.102 ±0.282 0.594 ±0.086 2.961 ±2.142 4.848 ±1.518 -3 4 P-PO mg dm 0.011 ±0.013 0.011 0.020 ±0.016 0.013 ±0.013 0.015 ±0.018 0.014 ±0.004 0.318 ±0.260 0.006 ±0.008 -3 Total P Total mg dm 0.034 ±0.011 0.038 ±0.022 0.054 ±0.066 0.021 ±0.017 0.015 ±0.006 0.684 ±0.592 0.095 ±0.066 -3 - - - mg dm 6.92 ±4.1 20.85 ±5.3 17.13 ±3.4 Chlorophyll-a 236.69 ±111.8 -3 dm 2 - oxygen 9.0 ±1.3 7.4 ±2.1 5.9 ±3.3 5.9 ±1.3 4.9 ±0.9 12.1 ±2.1 Dissolved mg O ) -1 (µS cm 88.3 ±9.5 520.5±26.2 306.0 ±15.5 390.5 ±38.9 176.6 ±21.8 245.5 ±21.9 488.0 ±111.7 Conductivity Conductivity pH 4.70 ±1.4 8.34 ±0.03 7.69 ±0.48 7.66 ±0.26 7.78 ±0.76 8.50 ±2.19 6.50 ±0.93 (m)* Max. depth 0.5-1.0 0.7-1.5 0.7-1.5 0.5-1.5 1.0-2.0 1.0-2.0 0.3-1.0 0.7 0.5 0.4 3.0 2.2 1.6 0.4 (ha) area Surface Table 1. Morphological, physical and chemical (± SD) characteristic of studied peat-bog reservoirs (mean values for months) Table (J) (M) 5 (BB5) 2 (BB2) 3 (BB3) 4 (BB4) 1 (BB1) Humic Moszne Peat-bog reservoir Eutrophic Jagodne Basic Bagno Bubnów Basic Bagno Bubnów Basic Bagno Bubnów Basic Bagno Bubnów Basic Bagno Bubnów * water level depended on precipitation depended level on * water 24 ANDRZEJ DEMETRAKI-PALEOLOG, MONIKA TARKOWSKA-KUKURYK

Table 2. Structure of vegetation of studied peat-bog reservoirs in Poleski National Park in 2008 (peat-bog reservoirs: J - eutrophic Jagodne, M - humic Moszne, BB1 - basic Bagno Bubnów 1, BB2 - basic Bagno Bubnów 2, BB3 - basic Bagno Bubnów 3, BB4 - basic bagno Bubnów 4, BB5 - basic bagno Bubnów 5) Peat-bog reservoirs M J BB1 BB2 BB3 BB4 BB5 Emergent macrophytes Cicuto-Caricetum pseudocyperi de Boer + Hottonia palustris L. + Phragmites australis (Cav.)Trin. ex Steud + + + Salicetum pentandro-cinerea (Almq.) + Typha latifolia L. + + + Floating-leaved macrophytes Aldrovanda vesiculosa L. + Nymphaea candida Presl. + Potamogeton natans L. + + + + Utricularia vulgaris L. + Submerged macrophytes Chara aculeolata Ktitz. + + Chara hispida L. + Myriophyllum spicatum L. + was noted in eutrophic peat bog reservoir Jagodne (Tab. 4). A similar low species diversity was observed in peat bog of Himalaya Mountains [25]. In peat bog reservoirs near Parczew as well as on the area of Wielkopolski National Park, number of rotifer species showed much higher values [13, 20]. Euplanktonic rotifer species, numerously represented in lakes of Łęczyńsko- Włodawskie Lakeland [6, 19] and rivers of Lublin region [5], occurred in a very little number, only 3 species in basic peat bog reservoirs (Tab. 4). Periphytic and benthic- periphytic species (from 6 to 11 dependently on the reservoir) reached higher numbers and constituted from 70% up to 92% of total rotifer species (Tab. 4). Relative abundances of different ecological groups were not a consequence of a type of peat-bog reservoir. Usually the lowest and the highest percentages of ecological groups were observed in basic reservoirs. In the three of studied peat-bog reservoirs there were noted rare species for Polish fauna: in basic reservoir BB5 – Lepadella Rottenbergi (Gosie), in basic reservoir BB1 – Trichocerca cavia (Gosse) and in eutrophic reservoir – Platyias p-atulus (Müll.). In the studied peat bog reservoirs indicatory species were presented. Two eutropho- bionts were noted in eutrophic reservoir - Anuraeopsis fissa Gosse (92% of total rotifers density) and Keratella cochlearis f. tecta (Gosse). The group of oligotrophobionts represented one species, Chromogaster ovalis (Berg.) in basic reservoir BB1. Two species typical for dystrophic waters, rare for Lubelszczyzna region, Macrochaetus subquadratus Perty and Microcodides chlajna Gosse were noted in basic peat-bog reservoir BB4 (Tab. 4). Based on food preferences, rotifer community can be divided into detritivorous, algaevorous, omnivorous and predatory species [9]. Detritivorous species are usually the STRUCTURE OF ROTIFER COMMUNITIES OF RESTORED SMALL PEAT-BOG RESEVOIRS... 25

Table 3. Species composition of planktonic rotifers of studied peat-bog reservoirs of Poleski National Park in 2008 (peat-bog reservoirs: J - eutrophic Jagodne, M - humic Moszne, BB1 - basic Bagno Bubnów 1, BB2 - basic Bagno Bubnów 2, BB3 - basic Bagno Bubnów 3, BB4 - basic bagno Bubnów 4, BB5 - basic bagno Bubnów 5) Peat-bog reservoirs M J BB1 BB2 BB3 BB4 BB5 1 Anuraeopsis fissa Gosse .+ 2 Ascomorpha ovalis (Berg.) .+ 3 Bdelloidea non. det. .+ .+ 4 Colurella adriatica Ehrb. .+ .+ .+ .+ 5 Colurella colurus (Ehrb.) .+ .+ 6 Colurella uncinata (Müller) .+ 7 Chromogaster ovalis (Berg.) .+ 8 Elosa spinifera Wiszn. .+ 9 Euchlanis dapidula Parise .+ 10 Keratella cochlearis f. tecta (Gosse) .+ 11 Lecane acus (Harr.) .+ 12 Lecane bulla (Gosse) .+ .+ .+ 13 Lecane closterocerca (Schm.) .+ .+ .+ 14 Lecane crenata Harr. .+ .+ 15 Lecane flexilis(Gosse) .+ 16 Lecane hamata (Stokes) .+ 17 Lecane ludwigii (Eckst.) .+ .+ .+ .+ 18 Lecane luna (Müll.) .+ 19 Lecane lunaris (Ehrb.) .+ .+ 20 Lecane opias (Herr. & Myers) .+ .+ .+ .+ 21 Lecane quadridentata (Ehrb.) .+ 22 Lecane stichaea (Harr.) .+ 23 Lepadella acuminata (Ehrb.) .+ 24 Lepadella cristata (Rouss.) .+ 25 Lepadella ovalis (Müll.) .+ .+ 26 Lepadella rhomboides (Gosse) .+ .+ 27 Lepadella rottenburgi (Lucas) .+ 28 Macrochaetus subqudratus Perty .+ 29 Microcodides chlaena Gosse .+ .+ 30 Mytilina mucronata (Müll.) .+ 31 Mytilina ventralis (Ehrb.) .+ 32 Platyias p-atulus (Müll.) .+ 33 Polyarthra vulgaris Carl. .+ .+ 34 Testudinella patina (And. et Shep.) .+ 35 Testudinella truncata (Gosse) .+ 36 Trichocerca bicrystata (Gosse) .+ 37 Trichocerca pusilla (Laut.) .+ .+ .+ .+ 38 Trichocerca rattus (Müll.) .+ .+ .+ 39 Trichocerca similis (Wierz.) .+ .+ .+ 40 Trichocerca tigris (Müll.) .+ .+ 26 ANDRZEJ DEMETRAKI-PALEOLOG, MONIKA TARKOWSKA-KUKURYK

Table 4. Ecological characteristic of rotifer communities of studied peat-bog reservoirs of Poleski National Park in 2008 (peat-bog reservoirs: J - eutrophic Jagodne, M - humic Moszne, BB1 - basic Bagno Bubnów 1, BB2 - basic Bagno Bubnów 2, BB3 - basic Bagno Bubnów 3, BB4 - basic bagno Bubnów 4, BB5 - basic bagno Bubnów 5). Densities marked by the same letters don’t differ significantly

Peat-bog reservoirs M J BB1 BB2 BB3 BB4 BB5 Number of euplanktonic species 0 1 1 1 3 3 3 Number of benthic-periphytic species 5 5 5 8 6 5 3 Number of periphytic species 1 4 3 3 1 2 4 Number of epibiontic species 1 1 0 0 0 0 0

Number of rare species 0 1 1 0 0 0 1

Number of indicatory species for 1 2 0 1 0 1 1 eutrophic waters Number of indicatory species for 0 0 1 0 0 0 0 oligotrophic waters Number of indicatory species for 0 0 0 0 0 2 1 dystrophic waters

Number of predatory species 0 0 0 0 0 0 0 Number of detritivorous species 0 0 1 0 1 1 1 Number of algaevorous species 0 3 1 2 2 3 2 number of omnivorous species 7 8 7 9 7 6 5

Total number of species 7 11 9 12 10 10 10 Shannon index 1,9 0,6 2,4 2,7 2,6 2,6 2,8 Density ind. dm-3 19a 407b 18a 20a 18a 18a 63c ± 4.56 ± 71.29 ± 5.11 ± 4.89 ± 6.22 ± 7.21 ± 13.81 most abundant group of planktonic rotifers in freshwater ecosystems [5, 7, 8, 9, 26]. Detritivorous rotifers were presented only in four basic reservoirs (Tab. 4). More numerously were presented algaevorous rotifers while the most abundant were omnivorous species. The dominance of algaevorous species is evidence of low water fertility and low availability of food resources [9]. Gliwicz [9] stated that under low concentrations of nutrients in water, small algae constitute the most available food for rotifers, because of their successful competition with other primary producers. Besides in waters of low nutrients concentration is usually observed little amounts of small detritus, which does not enhance the rapid development of detritivorous or omnivorous species. Low total densities of planktonic rotifers observed in the studied peat bog reservoirs confirm low fertility of these ecosystems. Its values ranged from 18 -19 ind. dm-3 (basic reservoirs BB1, BB3, humic reservoir M) up to 407 ind. dm-3 (eutrophic reservoir J) (Tab. 4). Low rotifers densities observed in humic and basic reservoirs showed similar values to those obtained by Sharma and Bhattariai [25] in small acidic peat bogs of Buthan, and by Radwan [20] in peat bog reservoirs near Parczew. Much higher densities (from 250 ind. dm-3 up to 400 ind. dm-3), similar to those of studied eutrophic reservoir, were noted in peat bog reservoirs near Turwia in Wielkopolski National Park [15]. The domination structure of planktonic rotifers showed very interesting results (Fig. 1). STRUCTURE OF ROTIFER COMMUNITIES OF RESTORED SMALL PEAT-BOG RESEVOIRS... 27

100% Lepadella ovalis 90% Polyarthra vulgaris Mytilina ventralis 80% Lecane opias Lecane lunaris 70% Lecane ludwigii Lecane flexilis 60% Lecane closterocerca Lecane bulla 50% Euchlanis dapidula Colurella uncinata 40% Colurella colurus 30% Colurella adriatica Bdelloidea non. det. 20% Anuraeopsis fissa Lecane stichaea 10% Trichocerca tigris Trichocerca rattus 0% Trichocerca similis Testudinella truncata M J BB1 BB2 BB3 BB4 BB5 peat-bog reservoirs Microcodides chlaena

Fig. 1. Relative abundances of particular rotifers species in studied peat-bog reservoirs of Poleski National Park in 2008 (peat-bog reservoirs: J - eutrophic Jagodne, M - humic Moszne, BB1 - basic Bagno Bubnów 1, BB2 - basic Bagno Bubnów 2, BB3 - basic Bagno Bubnów 3, BB4 - basic bagno Bubnów 4, BB5 - basic bagno Bubnów 5)

The group of dominants consisted of 21 rotifer species, which constituted above 50% of all species. The group of dominants was strongly influenced by faunistic differen- tiation; the group of dominants of each of the studied reservoir was represented by other species (Fig. 1). Bielańska-Grajner [2] classified rotifer communities as sustainable and non-sustainable domination structure. The authors assumed the community as sustainable when: 3 domination classes occurred (dominants, subdominants and recedents), at least 3 species represent dominants and none of them exceeds 45% of total density. Based on those criteria any of the studied peat-bog reservoir did not have a sustainable domination structure. The most non-sustainable domination structure was observed in eutrophic reservoir (J). The dominants were represented by one species – Anuraeopsis fissa Gosse, which amounted 92% of total density and recedents were not distinguished (Fig. 1). In the other peat-bog reservoirs, a number of dominants was higher and ranged from 3 to 4; recedents were not presented. In basic peat-bog reservoir (BB5), the share of dominating species (Colurella adriatica Ehrb.) exceeded 60% of total rotifer density. The cluster analysis of planktonic rotifer communities showed small faunistic similarity of studied peat-bog reservoirs. The values of Jaccard`s index ranged from 0.08 to 0.2 (Fig. 2). Nevertheless, faunistic analysis led to distinguishing three groups of peat-bog reservoirs. First group represented two basic reservoirs (BB2 and BB4) and humic peat-bog reservoir (M). Similarity indices ranged from 0.2 to 0.23 (Fig. 2). The second group included two basic reservoirs (BB1 and BB3) of lower faunistic similarity in comparison to the first group (Fig. 2). The third group represented basic peat-bog reservoir (BB5) and eutrophic reservoir (J). The Jaccard`s index for these reservoirs amounted to 0.15 (Fig. 2). 28 ANDRZEJ DEMETRAKI-PALEOLOG, MONIKA TARKOWSKA-KUKURYK

B. B3 1 B. B1

B. B5 2 J

B. B4

B. B2 3

0,04 0,2 0,36 0,52 0,68 0,84 1

Jaccard's Coefficient

Fig. 2. Structure of similarity of rotifers communities based on rotifers densities in studied peat-bog reservoirs of Poleski National Park in 2008 (peat-bog reservoirs: J - eutrophic Jagodne, M - humic Moszne, BB1 - basic Bagno Bubnów 1, BB2 - basic Bagno Bubnów 2, BB3 - basic Bagno Bubnów 3, BB4 - basic bagno Bubnów 4, BB5 - basic bagno Bubnów 5)

The PCA analysis of similarity of rotifer communities based on their densities confirmed the results obtained in cluster method and showed important faunistic differences among rotifer communities in the studied peat-bog reservoirs (Fig. 3).

PCA case scores 2.04 1.63 J 1.22 0.82 B. B5 B. B1 0.41 xis 2

A B. B3 -2.04 -1.63 -1.22 -0.82 -0.41 0.41 0.82 1.22 1.63 2.04 -0.41 -0.82 B. B4 -1.22 M. -1.63 B. B2 -2.04 Axis 1

Fig. 3. PCA analysis of rotifers communities based on rotifers densities in studied peat-bog reservoirs of Poleski National Park in 2008 (peat-bog reservoirs: J - eutrophic Jagodne, M - humic Moszne, BB1 - basic

Bagno Bubnów 1, BB2 - basic Bagno Bubnów 2, BB3 - basic Bagno Bubnów 3, BB4 - basic bagno Bubnów 4, BB5 - basic bagno Bubnów 5) STRUCTURE OF ROTIFER COMMUNITIES OF RESTORED SMALL PEAT-BOG RESEVOIRS... 29

Axis 1 explaines rotifers variability in 22%, Axis 2 in 19%. The analysis of similarity of rotifer communities based on domination structure showed approximate results, with an exception of basic reservoir BB4 and humic reservoir (M), where the PCA analysis showed high similarity.

CONCLUSIONS

The most optimal habitat conditions for planktonic rotifers were presented in eutrophic peat bog reservoir; there were noted 11 rotifer species and the highest density of 407 ind dm-3. However, the highest species diversity of rotifers (H index = 2.4-2.8) was observed in basic peat bog reservoirs. High biodiversity of the studied reservoirs has been confirmed by the presence of rare species, domination of periphytic taxa on benthic- periphytic and euplanktonic forms. Low number and densities of indicatory species for eutrophic waters, high ratio of algaevorous to detritivorous species, as well as absence of Keratella cochlearis f. tecta (Gosse), confirm low trophy of studied peat bog reservoirs. Such ecosystems, mostly basic and humic, are rarely presented in the area of Łęczyńsko- Włodawskie Lakeland. In the paper, rotifer communities of peat bog reservoirs showed high faunistic differences. The group of dominants included 21 rotifer species, different dependently on a type of reservoir. Domination structure of rotifer communities was non- sustainable, mostly due to the absence of recedents or high domination of one species. All these observations indicate low stability of rotifer communities for changes of environmental conditions of their habitat.

REFERENCES

[1] Arndt H.: Rotifers as predators on components of the microbial web (bacteria, heterotrophic flagellates, ciliates), Hydrobiologia, 255/256, 231-246 (1993). [2] Bielańska-Grajner I.: Wrotki (Rotifera) psammonowe zbiorników wodnych wybranych obszarów Polski, Wyd. Uniw. Śląski, 2005, pp. 1-114. [3] Błędzki A.L., A. Ellison: Diversity of Rotifer from north-eastern U.S.A. bogs with new species records from North America and New England, Hydrobiologia, 497, 53-63 (2003). [4] Chmielewski T.J.: W sprawie utworzenia Parku Narodowego na Pojezierzu Łęczyńsko-Włodawskim, Chrońmy Przyrodę Ojczystą, T. XLII z. 4. (1986). [5] Demetraki-Paleolog A.: Wrotki (Rotifera) planktonowe rzek Zachodniej Lubelszczyzny, WAR, Lublin 2007, pp. 1-123. [6] Demetraki-Paleolog A.: Planktonic Rotifers of four polimictic lakes of Łęczyńsko-Włodawskie Lakeland (Eastern Poland), Teka Komisji Ochrony i Kształtowania Środowiska Przyrodniczego 2009 (in press). [7] Ejsmont-Karabin J.: Long-term changes in the abundance and structure of the community of planktonic Rotifera in a humic lake as a result of liming, Ekol. Pol., 44(1-2), 39-51 (1996). [8] Ejsmont-Karabin J., S. Radwan, I. Bielańska-Grajner.: Monogononta – atlas gatunków. 32.B. (W:) S. Radwan (red.). Wrotki (Rotifera). Fauna słodkowodna Polski. 32. Polskie Towarzystwo Hydrobiologic- zne, Uniwersytet Łódzki. Oficyna Wydawnicza Tercja, Łódź: 2004, pp. 147-448. [9] Gliwicz M.: Status troficzny gatunków zooplanktonu słodkowodnego, Wiad. Ekolog., XX, 3, 197–203 (1974). [10] Harasimiuk M., Z. Michalczyk, M. Turczyńska (Red.): Jeziora łęczyńsko-włodawskie. Monografia przyrodnicza, Biblioteka Monitoringu Środowiska, UMCS, PIOŚ, Lublin 1998, pp. 1-210. [11] Hillbricht-Ilkowska A.: The influence of the fish population on the biocenosis of a pond using Rotatoria fauna as an illustration, Ekol. Pol. A, 12, 453-503 (1964). [12] Karabin A.: Pelagic zooplankton (Rotatoria + Crustacea) variation in the process of lake eutrophication. I. Structural and quantitative features, Ekol. Pol., 33, 567-616 (1985). [13] Klimaszyk P., N. Kuczyńska-Kippen: Peat-bog pool (Wielkopolski National Park) as a habitat of specific communities of zooplankton, Acta Agrophysica, 7(2) 375-381 (2006). 30 ANDRZEJ DEMETRAKI-PALEOLOG, MONIKA TARKOWSKA-KUKURYK

[14] Koc J., Z. Nowicki: Czynniki warunkujące właściwości chemiczne wód drobnych zbiorników wodnych w krajobrazie rolniczym, II Ogólnopolska konferencja Naukowa. Przyrodnicze i techniczne problemy ochrony i kształtowania środowiska rolniczego. Materiały konferencyjne 1997, pp. 91-97. [15] Kuczyńska-Kippen N, E. Arczyńska Chudy, H. Gołdyn: Comparison of the Rotifera and Crustacean community structure of two post-excavated peat pits near Turew, Wielkopolska Region Poland - Teka Commission of Protection and Formation of Natural Environment, 3, 96-103 (2006). [16] Mieczan T.: Relationships among ciliated protozoa and water chemistry in small peat-bog reservoirs (Łęczna Włodawa Lakeland, Eastern Poland), Oceanological and Hydrobiological Studies, 36, 77-86 (2007). [17] Paleolog A., S. Radwan, W. Kowalik, C. Kowalczyk, R. Stryjecki, W. Zwolski: Fauna wodna bezkręgowców Parku Krajobrazowego „Lasy Janowskie” [w:] Środowisko Przyrodnicze Parku Krajo- brazowego „Lasy Janowskie”, Wyd. UMCS Lublin 1997, pp. 117-133. [18] Puchalski W. Poeksploatacyjne zbiorniki wodne – wstęp do charakterystyki ekologicznej, Wiadomości Ekologiczne, 31, 3-24 (1985). [19] Radwan S.: Wrotki pelagiczne jezior Pojezierza Łęczyńsko-Włodawskiego. Studium faunistyczno-ekologiczne, Skrót rozprawy habilitacyjnej, AR. Ser. Rozpr. Hab.,8; 1-57, 1973. [20] Radwan S.: Wrotki (Rotatoria) torfianek okolic Parczewa, An. UMCS, Sec. C, Lublin, 29, 215-230 (1974). [21] Radwan S., B. Jarzynowa, W. Zwolski, K. Girsztowtt, C. Kowalczyk, W. Kowalik, A. Paleolog: Ekolog- iczna charakterystyka wód górnego i środkowego biegu rzeki Bystrzycy Lubelskiej i jej dopływów oraz Jeziora Zemborzyckiego, Roczn. Nauk. PZW, t 1, Warszawa 1988, pp. 123-156. [22] Radwan S., J. Sender: Kształtowanie się różnorodności biologicznej w obszarach wodno-błotnych Pojezi- erza Łęczyńsko-Włodawskiego [w:] S. Radwan (red.) Poleski Park narodowy, PTH, Wyd. UMCS Lublin 1996, pp. 45-57. [23] Radwan S., W. Kowalik, W. Wojciechowska, R. Kornijów, J. Sender, M. Kolejko: Ekosystemy wodne Poleskiego Parku Narodowego [w:] S. Radwan (red.). Poleski Park Narodowy. Monografia przyrodnicza. Wyd. MORPOL Lublin 2002, pp. 133-161. [24] Scheffer M.: Ecology of shallow lakes, Kluwer Academic Publishers. Dordrecht, Boston, London 2001, pp. 357. [25] Sharma B.K., S. Bhattarai: Hydrobiological analysis of a peat-bog with emphasis on its planktonic diversity and population dynamics in Bumdeling Wildlife Sanctuary, eastern Bhutan, Limnology, 6, 183-187 (2005). [26] Wallace R.L., C. Ricci: Rotifera [in:] S.D. Rundle, A.L. Robertson, J.M. Schmid-Araya (eds.), Freshwa- ter Meiofauna, Biology and Ecology, Backhuys, Pub. Leiden, Holandia, 15-44, 2002. [27] Wollmann K., R. Deneke, B. Nixdorf, G. Packroft: Dynamics of planktonic food webs in the mining lakes across a pH gradient (pH 2-4), Hydrobiologia, 433, 2-14 (2000).

Received: May 6, 2009; accepted: May 30, 2010.

STRUKTURA ZGRUPOWAŃ WROTKÓW TORFIANEK NA ZRENATURALIZO- WANYM OBSZARZE POLESKIEGO PARKU NARODOWEGO

Jednym z ciekawszych i mniej poznanych ekosystemów wodnych są różnego typu torfianki: węglanowe, hu- musowe i eutroficzne. Analiza wrotków planktonowych w siedmiu takich zbiornikach Pojezierza Łęczyńsko- Włodawskiego wykazała: dużą różnorodność gatunkową, obecność gatunków rzadkich, znaczny udział form peryfitonowych i bentosowo-peryfitonowych oraz duże różnice faunistyczne między wrotkami różnych torfianek. Mała liczebność wrotków, gatunki wskaźnikowe, znaczny udział gatunków roślinożernych w stosunku do detrytusożernych wskazują na małą żyzności większości torfianek. Mimo tych cech analiza struktury dominacji wykazała brak zrównoważenia zgrupowań wrotków i możliwość ich małej stabilności w przypadku jakichkol- wiek zmian środowiskowych. 31

ARCHIVES OF ENVIRONMENTAL PROTECTION vol. 36 no. 4 pp. 31 - 39 2010 PL ISSN 0324-8461

NITROGEN COMPOUNDS IN WELL WATER AS A FACTOR OF A HEALTH RISK TO THE MACIEJOWICE COMMUNE INHABITANTS (MAZOWIECKIE VOIVODESHIP)

Jolanta Raczuk

University of Podlasie, Department of Ecology and Environmental Protection, Prusa str. 12, 08-110 Siedlce, Poland Corresponding author e-mail: [email protected]

Keywords: Nitrogen compounds, health risk, well water.

Abstract: Well water quality monitoring was carried out in the Maciejowice Commune in the years 2005-2006. - Water was sampled from 20 dug and drilled wells five times. Chemical analyses involved determination of NO3 , - + 3- - 2+ 2+ NO2 , NH4 , PO4 , Cl , Ca , Mg , total hardness, pH and electrolytic conductivity. Health risk resulting from the presence of nitrates in water consumed by people was assessed in this paper. The obtained results indicated that 50% of examined waters did not meet the set standards. Ammonium ions and nitrates were the main ions contaminating drinking water. A negative correlation between nitrate, nitrite and ammonium ion concentration and well depth was found. People drinking water from 60% examined wells ingest excess quantities of nitrates (safety margin ADI: EDI < 1), which, in the case of long–term exposure, can be harmful, especially for infants and pregnant women. From among 20 analysed wells the water from 5 wells conformed to recommended standards. Its good quality was a result of the appropriate location of the wells within the household premises, appropriate depth and insulation and farmers’ complying with the Code of Good Agricultural Practice.

INTRODUCTION

Groundwater is the source of drinking water for many people around the world, especially in rural areas [14]. Groundwater pollution is one of the most common environmental problems nowadays [5]. Numerous contaminants can render groundwater unsuitable - for human consumption. Nitrate (NO3 ) is the most frequently introduced pollutant into groundwater systems [20, 22]. Natural levels of nitrate in groundwater depend on soil type and geology. Crop fertilizers, domestic sewage and animal waste are common sources of nitrate contamination in rural areas [20, 29]. Nitrogen compounds in these sources are oxidized in aerated soil to soluble nitrate. Nitrate is highly water-soluble and therefore tends to migrate from soil to groundwater [18]. In Poland, rural domestic wells are especially vulnerable to nitrate contamination [11, 12, 15, 16, 23, 25]. Depending on the country regions, from 30 to 85% of these well intakes contain water with an exceeded concentration of nitrates [2]. High level of nitrate in drinking water is most often associated with shallow dug wells of less than 15 m in depth [11, 16]. 32 Jolanta Raczuk

Nitrate can have several adverse effects upon human health, most notably methemoglobinemia, gastric cancer and non-Hoadgkin’s lymphoma [1, 4, 6, 7, 28]. The toxicology - of nitrate (NO3 ) to humans is mainly attributable to its reduction in the digestive system - to harmful nitrite (NO2 ) [1, 27]. Methemoglobin (MetHb) is formed when nitrite oxidizes the ferrous ion in hemoglobin (Hb) to the ferric form [4]. MetHb cannot bind oxygen, and the condition of methemoglobinemia is characterized by cyanosis, stupor, and celebral anoxia [4]. Under normal conditions, < 2% of the total Hb circulates as MetHb [4]. Raised levels of methemoglobin (greater than 10%) in infants under 4-6 months can produce cyanosis, referred to as “blue-baby syndrome” [1, 29]. In addition, nitrite can undergo ni- trosation reactions in the stomach with amines and amides to form N-nitroso compounds (NOC) [27]. Most of these compounds are potent animal carcinogens, inducing tumors at multiple organ sites [26]. Therefore, nitrate concentration is an important criterion of drinking water quality [3, 17, 29]. To protect consumers from the adverse effects associated with the high nitrate intake, the World Health Organization (WHO), the European Union (UE) and Poland have set standards to regulate the nitrate concentration in drinking water. The drinking water standard set by the WHO [29], the UE [3] and Poland [17] for nitrate is 50 mg∙dm-3. The objective of the study was to determine nitrogen compound concentrations in well water and to assess health hazards which are associated with the nitrate presence in the drinking water of household wells in the Maciejowice Commune, Mazowieckie Voivodeship.

MATERIAL AND METHODS

Study area and sampling collection Well water quality monitoring was carried out in the Maciejowice Commune in the years 2005-2006. The commune is located on the left bank of the River Wisła in the southern part of the Mazowieckie Voivodeship, and is a part of the Garwoliński administrative district. It is a predominantly agricultural commune where agricultural land makes up 50% and forested area 37% of the total area of the commune. The water supply system in the Maciejowice Commune is currently under construction. By 2006, mains water was supplied to as little as 3% of the commune’s households. As a result, the inhabitants predominantly utilize the water from shallow household wells. Studies included twenty wells being a source of water for both consumption and domestic purposes. Of the wells, sixteen were located in Maciejowice and four in Kochów; twelve wells were 3 to 50 m-deep drilled wells, and eight wells were dug wells whose depth ranged between 3.6 and 6 m; eleven wells were located in the high-density area (wells no 1-7, 9, 13, 14, 16), whereas the remaining wells were situated in the low-density area and surrounded by either cultivated fields (well no 10, 11, 17-20) or fallow land (well no 8, 12, 15). Water samples for testing were taken five times: in October 2005 and in April, May, June and October 2006.

Analytical procedures Water samples for physical and chemical analysis were collected in polyethylene bottles and taken to the laboratory. Analysis was carried out immediately. The pH and electrical conductivity EC20 values were measured immediately after collection with the help of a NITROGEN COMPOUNDS IN WELL WATER AS A FACTOR OF A HEALTH RISK TO... 33 portable pH-meter and conductivity - meter. Various colorimetric methods were used for - - + 3- the determination of NO3 , NO2 , NH4 , PO4 [8, 13]. The concentrations of nitrates and nitrites were determined by the modified Griess method [13]. This method is based on the reaction of diazotization of the primary aromatic amines followed by coupling of diazo- nium salts yielding azo dye. Nitrite concentration was determined with sulphonamide and N-(1-napthyl) ethylenediamine. The same method was used for determination of nitrates + after their reduction to nitrites. Amonium ion (NH4 ) was determined by the indophenol 3- blue method and PO4 was determined by the molybdenum blue method. Total hardness, Ca2+ and Mg2+ were determined by the complexometric titration with EDTA and chloride (Cl-) was determined by the argentometric method with silver nitrate [8].

Evaluation of a health risk Health hazards associated with nitrate-contaminated drinking water were assessed by comparing Acceptable Daily Intake (ADI) with Estimated Daily Intake (EDI). The ADI value set for nitrates by the Joint FAO/WHO Expert Committee on Food Additives [6] -1 -1 - - is 5 mg NaNO3·kg b.w. ·day , which, when converted into NO3 , is 3.65 mg·kg b.w. 1·day-1. Following Cadum, Szczerbiński et al. [24 ] report the ADI value for nitrates (V) -1 -1 consumed with drinking water to be 0.365 mg NO3·kgm.c. ·day , that is 10% of the total ADI value consumed. Estimated Daily Intake (EDI) was calculated according to the formula: EDI = F ∙ R, where: F – average daily water consumption – 2 dm3∙person-1∙day-1, R – concentration of nitrates in drinking water – mg·dm-3. Risk posed by nitrates in water was evaluated by determining a safety margin between ADI and EDI expressed as a value indicating how many times the daily intake of nitrates would have to increase to reach the ADI value, that is the value which, when exceeded, cannot be accepted as safe. The safety margin = ADI : EDI. The following ranges were accepted following Szczerbiński et al. [24] on the basis of the safety margin calculated for nitrates in well water: < 1 – below safety margin, 1–10 – low safety margin, 10–20 – medium safety margin, 20–30 – high safety margin, > 30 – very high safety margin. In the present work calculations were performed for an assumed human body weight of 70 kg.

Statistical analysis Kolmogorov-Smirnov test was first used to check the normality of the data. Because the data were non-normally distributed (p < 0.05), Spearman’s rank order correlation was utilized to study associations between variables. The differences between concentrations of nitrate, nitrite and ammonium ions coming from various times of water sample collected, were estimated by nonparametric Kruskal-Wallis test. A value of p < 0.05 was considered statistically significant. Nonparametric methods have been found to be more accurate for the analyses of environmental data, especially groundwater data [9]. Statistical analysis was conducted using Statistica 5.0 computer program. 34 Jolanta Raczuk

RESULTS AND DISCUSSION

Nitrate concentration in the well water tested ranged from 0.25 to 142.00 mg·dm-3 (Tab. 1). Throughout the studies the highest was the nitrate concentration in the water of well no 17 situated in Kochów. It was a very shallow well, 4.5 m in depth, located within the premises of a household whose wastewater was often disposed to the nearby meadow, despite the household possessing a septic tank. Nitrate concentration, ranging from 62.03 to 86.01 mg·dm-3, was detected in the water of wells no 2, 3 and 14. They were dug wells situated in a high-density area and belonged to the households equipped with leaky septic - -3 tanks. In general, the standard value of 50 mg NO3 ·dm [17] was exceeded in the water of - 25% examined wells. The lowest nitrate concentration, which did not exceed 10 mg NO3 ·dm-3, was found in the water of wells no 1, 9, 12, 16 and 18. The wells were drilled to the depth ranging from 7 to 50 m and situated within the household’s premises at the appropriate distance from the house and livestock buildings. The households were equipped with properly situated and non-leaking septic tanks.

- Table 1.Concentration of NO3 , estimated daily intake EDI and safety margin for nitrates taken with well water Concentration EDI Safety margin Locality and mg NO - dm-3 mg NO - ∙pers.-1∙day-1 ADI : EDI well number 3 3 range median median median Maciejowice 1 0.41-5.22 4.71 9.42 2.72 2 50.01-62.03 58.02 116.04 0.22 3 25.01-79.55 57.03 114.06 0.22 4 14.61-46.13 40.02 80.04 0.32 5 0.52-14.62 0.71 1.42 18.25 6 28.02-45.11 33.01 66.02 0.39 7 8.33-19.72 17.52 35.04 0.71 8 13.62-28.12 17.33 34.66 0.73 9 0.52-12.01 4.81 9.62 2.66 10 25.0-48.11 44.13 88.26 0.29 11 15.6-29.23 18.61 37.22 0.69 12 0.42-5.814 0.81 1.62 15.86 13 0.43-10.63 1.82 3.64 7.02 14 41.01-86.01 52.02 104.04 0.24 15 11.41-21.04 14.73 29.46 0.88 16 0.25-0.33 0.30 0.60 42.94 Kochów 17 39.02-142.00 107.00 214.00 0.12 18 0.41-1.91 0.92 1.84 14.04 19 2.71-16.62 15.31 30.62 0.84 20 0.50-30.34 1.62 3.24 7.93 ADI- Acceptable Daily Intake; EDI- Estimated Daily Intake NITROGEN COMPOUNDS IN WELL WATER AS A FACTOR OF A HEALTH RISK TO... 35

Statistical analysis revealed a negative correlation between nitrate concentration and well depth (Tab. 2). It indicates that shallow dug wells are most likely to be nitrate contaminated. Significant relationships between well depth and nitrate concentration were confirmed by other authors too, e.g. Hudak [10], Jaszczyński et.al.[11], Nas and Berktay [14], Ostrowska and Płodzik [15].

- - + Table 2. Spearman rank order coefficients sr for the relationships between concentrations of NO3 , NO2 , NH4 and some physical and chemical indexes of well water (n = 100)

Depth of Index NO - NH + PO 3- CaCO Ca2+ Cl- EC 2 4 4 3 20 well - NO3 0.598** ns 0.239* ns ns 0.519*** 0.375*** -0.458*** - NO2 ns 0.207* ns ns 0.443*** 0.374*** -0.356*** + NH4 ns ns 0.326** 0.283** 0.263** 0.402*** -0.379*** *significant at p < 0.05; **significant at p < 0.01; *** significant at p < 0.001; ns- not significant

Health risk assessment showed that consumption by people of water from 60% examined wells was associated with nitrate intake which exceeded ADI, and as a result was below the safety margin (ADI : EDI < 1) (Tab. 1). The water of well no 17, whose ADI : EDI ratio was 0.12, was characterized by nitrate concentration which exceeded ADI eightfold. Long-term exposure to drinking water with such parameters increases the risk of digestive tract cancer [7, 28]. Nitrate intakes with drinking water representing the low safety margin (1–10), medium safety margin (10–20) and high safety margin (> 30) were associated with 20, 15 and 5% examined wells, respectively. Nitrite concentrations in the water of wells sampled ranged from 0.00 to 9.80 -3 - mg·dm (Tab. 3). The water of well no 14 exceeded the standard value of 0.5 mg NO2 -3 - -3 ·dm [17] by as much as 20 times (9.80 mg NO2 ·dm ). In the close proximity of the well there was situated an organically-manured vegetable garden and a farmyard manure heap. Nitrogen is present in human and animal wastewater in organic form, which may then subsequently be mineralized to inorganic forms. The bacteria action on such organic matter results in its degradation and release of ammonia. The ammonia so produced may be oxidized to nitrite by bacteria such as Nitrosomonas, which can further be oxidized to nitrate by other bacteria such as Nitrobacter. These biologically mediated reactions are collectively reffered to as nitrification [21]. The nitrite ion contains nitrogen in a relatively unstable oxidation state. Chemical and bilogical processes can futher reduce nitrite to various compounds or oxidize it to nitrate. High nitrite concentrations may indicate that, among others, an intensive process of protein degradation is taking place. High dynamics of nitrate and nitrite concentrations in the examined well water are reflected in a statistically significant correlation between the ions concentrations (Tab. 2). There was also found a significant and negative correlation between nitrite concentration and well depth. Ammonium ion concentration in the investigated well waters ranged from < 0.10 to 20.00 mg·dm-3 (Tab. 3). The highest ammonium ion concentration, within the range of 8.50 to 20.00 mg·dm-3, was found in the water of wells no 13 and 14. The standard value + -3 was exceeded by 0.5 mg NH4 ·dm [17] and was detected in the water sampled from ten wells. Such a high ammonium ion concentration indicates that there existed a nearby source of organic contamination, that is livestock buildings, farmyard manure heaps, and leaking septic tanks. Similar ammonium sources were mentioned in the studies by Ku- 36 Jolanta Raczuk

Table 3. Range and median values of some physical and chemical indexes of well water

Index Range Median - -3 NO3 (mg ∙dm ) 0.25-142.00 15.53 - -3 NO2 (mg ∙dm ) 0.00-9.82 0.05 + -3 NH4 (mg ∙dm ) < 0.10-20.00 0.45 3- -3 PO4 (mg ∙dm ) 0.10-20.00 1.11 Cl- (mg ∙dm-3) 5-110 35.5 -3 CaCO3 (mg ∙dm ) 88-654 255 Ca2+ (mg ∙dm-3) 76.9-195.6 74.1 Mg2+ (mg ∙dm-3) 3.6-46.8 14.6 pH 5.8-7.3 6.6 -1 EC20 (μS∙cm ) 151-1500 549 czewski [12] and Szperliński [25]. A high ammonium ion concentration disqualifies the examined wells as sources of drinking water. There was found a negative correlation between ammonium ion concentration and well depth, which means that the shallower a well is the more leaking from natural manures and domestic waste occurs. Low concentrations of ammonium ions and nitrites accompanied by high nitrate concentrations were determined for well no 17, which indicates that the contaminants covered a long way and in the meatime were subjected to the oxidation process. Date of water sampling did not significantly differentiate nitrate, nitrite and ammonium ion concentrations, which was established by means of the nonparametric Kruskal- Wallis test (p > 0.05). The problem of poor quality of well waters does not pertain exclusively to Ma- ciejowice commune. Due to economic reasons or to a lack of pipelines, people in rural areas of our country often exploit dug and drilled wells which contain shallow ground waters. Numerous studies [11, 12, 15, 16, 23, 25] indicate that water from farm wells does not conform to the existing standards mainly because of high concentrations of nitrogen compounds. Concentrations of nitrites, ammonium ions and particularly of nitrates may exceed allowable standards several times. Phosphate concentrations in the examined water samples were within a broad range of 0.10 to 20.00 mg·dm-3 with a median value of 1.11 mg·dm-3 (Tab.3). Phosphorus infiltration into groundwater may be associated with soil fertilization and livestock production but it may be as well introduced into the soil with detergents [19]. The presence of marked phosphorus concentrations in water may be indicative of the declining ability of the soil to bind this element. Studies by Hudak [10] demonstrated that nitrates and chlorides are the most frequently infiltrating water contaminants. Chloride concentration in the water tested ranged from 5 to 110 mg·dm-3 with the median vale of 35.5 mg·dm-3 (Tab.3). There were no cases of the concentration exceeding the standard value of 250 mg·dm-3 [17]. Chlorides in well water come from livestock wastewater or domestic wastewater. Higher nitrate concentrations in the water of examined wells was accompanied by increased chloride concentrations. The relationship was confirmed by a statistically significant correlation between these indicators. Similar associations for well water were also reported by Jaszczyński et NITROGEN COMPOUNDS IN WELL WATER AS A FACTOR OF A HEALTH RISK TO... 37 al. [11]. Zahn and Grimm [30] reported that chloride concentration in water is an indicator of the presence of nitrates. The well water analysed was characterized by a varied degree of hardness; moder- ately hard and hard water constituted 35% in total. Water hardness depends on two cati- ons, that is calcium and magnesium. There are no Polish standards for calcium concentration, whereas magnesium concentration should lie within the range of 30–125 mg·dm-3 [17]. The element was deficient in the water sampled from 90% wells studied. It is probably connected with the geological make-up of the terrain and low magnesium content of the rocks. Calcium concentration in the water samples ranged from 76.9 to 195.6 mg·dm-3 with the median value of 74.1 mg·dm-3 (Tab. 3). The electrolytic conductivity in the water samples tested fell within the range of 151 to 1500 μS·cm-1 (Tab. 3) and at no occasion did it exceed the standard value of 2500 μS·cm-1 [17]. There was found a statistically significant correlation between the concentration of nitrogen compounds analysed and specific electrolytic conductivity, which indicates that the compounds influence the level of water mineralization. It should be stressed that the overall sanitation status of well water tested was not satisfactory, which indicates that further studies are clearly warranted on a larger scale. The studies should be accompanied by an educational program to make people using contaminated water aware of potential risks associated with the poor quality of the water.

CONCLUSIONS

1. Nitrogen compound concentrations in 50% of examined well waters did not meet the set standards. Ammonium ions and nitrates were the main ions contaminating drinking water. 2. People drinking water from 60% examined wells ingest excess quantities of nitrates (safety margin < 1), which, in the case of long-term exposure, can be harmful, in particular for infants and pregnant women. 3. Water sampled from five wells met the drinking water standards throughout the whole study period. Its good quality was a result of the appropriate location of the wells within the household premises, appropriate depth and insulation and farmers’ complying with the Code of Good Agricultural Practice. 4. Due to occurrence of non-standard concentrations of indicators affecting human health in well water as well as their marked dynamics, monitoring of well water quality ought to be carried out several times a year.

REFERENCES

[1] Bruning-Fann C.S., J.B. Kaneene: The effects of nitrate, / nitrite and N-nitroso compounds on human health: a revie, Vet. Hum. Toxicol., 35, 521-538 (1993). [2] Brzozowska A.: Toksykologia żywności, Wyd. SGGW, Warszawa 2004. [3] Council of the European Communities (CEC): Directive of 3 novembe 1998 relating to the quality of water intendent for human consumption /98/83/EC, Offic J. EEC 330, 32-54 (1998). [4] Fann A.M., C.C.Willhite, S.A. Book: Evaluation of the nitrate drinking water standard with reference to infant methemoglobinemia and potential reproductive toxicity, Regul. Toxicol. Pharmacol., 7(2), 135-148 (1987). [5] Fatta D., D. Naoum, M. Loizidou: Integrated environmental monitoring and simulation system for use as a management decision support tool in urban areas, J. Environ. Manage., 64(4), 333-343 (1998). 38 Jolanta Raczuk

[6] Food and Agiculture Organisation of the United Nations (FAO)/World Health Organization (WHO). Ni- trate (and potential endogenus formation of N- nitroso compounds), [in:] Safety evaluation of certain food additives and contaminants. Joint FAO/WHO Expert Committee on Food Additives.WHO Food Additives Ser. 50, Geneva 2003, http://www.inchem.org/documents/jecefa/jecmono/v50je06.htm. [7] Gulis G., M. Czompolyova, J.R. Cerchan.: An ecologic study of nitrate in municipal drinking water and cancer incidence in Trnava District, Slovakia, Environ. Res., 88A, 182-187 (2002). [8] Hermanowicz W., J. Dojlido, W. Dożańska, B. Koziorowski, J. Zerber: Fizyczno-chemiczne badanie wody i ścieków, Arkady, Warszawa 1999. [9] Hensel D. R.: Advantages of non parametric procedures for analisis of water quality data, Hydrol. Sci. J., 32, 179-190 (1987). [10] Hudak P.F.: Chloride and nitrate distributions in the Hicory Aquifer, central Texas, USA,. Environ. Int., 25(4), 393-401(1999). [11] Jaszczyński J., A. Sapek, S. Chrzanowski: Wskaźniki chemiczne wody do picia z ujęć własnych w gosp- odarstwach wiejskich w otulinie Biebrzańskiego Parku Narodowego, Woda Środ. Obsz. Wiej., 62(18), 129-142 (2006). [12] Kuczewski K.: Wpływ nieuporządkowanej gospodarki wodno-ściekowej na wsi na jakość wody w studniach kopanych, Zeszyty Nauk. AR we Wrocławiu, 314, 42-56 (1997). [13] Marczenko Z., M. Balcerzak: Separation, preconcentration and spectrophotometry in inorganic analysis, Elsevier, Amsterdam 2001. [14] Nas B., A. Berktay: Grounwater contamination by nitrates in the city of Konya, (Turkey): A GIS perspective, J. Environ. Manage., 79, 30-37 (2006). [15] Ostrowska E. B., M.A. Płodzik: Wpływ otoczenia zagrody wiejskiej na jakość wody w studniach przy- domowych, Wiad. IMUZ, 20(1), 19–27 (1999). [16] Raczuk J., K. Sarnowska: Jakość wód studni wiejskich w wybranych gminach województwa lubelskiego, Archives of Environmental Protection, 28(3), 63-75 (2002). [17] Rozporządzenie Ministra Zdrowia z dnia 29 marca 2007 r. w sprawie wymagań dotyczących jakości wody przeznaczonej do spożycia przez ludzi, Dz. U. 2007, nr. 61, poz. 417. [18] Sapek B.: Wymywanie azotanów oraz zakwaszenie gleby i wód gruntowych w efekcie działalności rolniczej, Mater. Inf. nr 30. Wydaw. IMUZ, Falenty, 1995. [19] Sapek A.: Rozpraszanie fosforu pochodzącego z rolnictwa i potencjalne zagrożenie dla środowiska, Zesz. Probl. Post. Nauk Rol., 478, 269-280 (2001). [20] Sapek A.: Agricultural activities as source of nitrates in groundwater, [in:] Nitrates in groundwater. L. Razowska-Jaworek, A. Sadurski (ed.), Leiden: Balkema Publ., 2004. [21] Shrimali M., K.P. Singh: New methods of nitrate removal from water, Environ. Pollut., 112, 351-359 (2001). [22] Splanding R.F., M. E. Exner: Occurense of nitrate in groundwater-a review, J. Environ. Qual., 22, 392- 402 (1993). [23] Skorbiłowicz M., E. Skorbiłowicz: Quality of well water in context of the content of nitrogen and phosphorus compounds in the upper Narew river valley, J. Elementol., 13(4), 625-635 (2008). [24] Sczerbiński R., J. Karczewski, J. Filon: Azotany (V) w wodzie do picia jako czynnik ryzyka zdrowotnego ludności województwa podlaskiego, Rocz. PZH, 57(1), 39-48 (2006). [25] Szperliński Z., J. Żabowski, K. Badowska- Olenderek, M. Olesiejuk-Kowalska: The quality of ground water in dug and drilled wells on the Łomianki commune area, Polish Ecolog. Stud., 13(3-4), 343-362 (1987). [26] Tricker A. R., R. Preussmann: Carcinogenic N- nitrosoamines in the diet: occurence, formation, mechanisms and carcinogenic potential, Mutat. Res., 259, 27- 89 (1991). [27] Walker R.: Nitrates, nitrites and N-nitroso compounds: a revive of the occurrence, in food and diet and the toxicological implication, Food Addid Contam., 7, 717-768 (1990). [28] Ward M.H., T. M. deKok, P. Levallois, J. Brender, G. Gulis, B.T. Nolan, J.VanDerslice: Workgroup report: drinking –water nitrate and health-recent findings and research needs, Environ. Health Persp., 113,107-114 (2005). [29] World Health Organisation (WHO). Guidelines for drinking water quality: second addendum to third edition. Vol.1. Recomendation. Geneva 2008. [30] Zahn M.T., W.D. Grimm: Nitrate and chloride loading as anthropologic indicators, Water Air Soil Pollu- tion, 68, 469-483 (1993).

Received: May 11, 2009; accepted: July 15, 2010. NITROGEN COMPOUNDS IN WELL WATER AS A FACTOR OF A HEALTH RISK TO... 39

ZWIĄZKI AZOTU W WODZIE STUDZIENNEJ JAKO CZYNNIK RYZYKA ZDROWOTNEGO MIESZKAŃCÓW GMINY MACIEJOWICE (WOJEWÓDZTWO MAZOWIECKIE)

W roku 2005-2006 na terenie gminy Maciejowice prowadzono monitoring wody studziennej. Badaniami objęto wodę pochodzącą z 20 studni kopanych i wierconych. Próby wody pobierano pięciokrotnie i oznaczono w nich - - + 3- - 2+ 2+ stężenie: NO3 , NO2 , NH4 , PO4 , Cl , Ca , Mg , twardość ogólną oraz pH i przewodność elektrolityczną. W pracy oszacowano ryzyko zdrowotne wynikające z obecności azotanów (V) w wodzie spożywanej przez ludność. Otrzymane wyniki wskazują, że stężenie związków azotu w 50% badanych wód studziennych nie odpowiada obowiązującym normom. Głównymi jonami zanieczyszczającymi wodę są jony amonowe i azotany (V). Wykazano ujemną korelację pomiędzy stężeniem azotanów (V), azotanów (III), jonów amonowych a głębokością studni. Oceniając ryzyko zdrowotne, stwierdzono, że ludność pijąca wodę z 60% badanych studni, pobiera azotany (V) w nadmiernej ilości (margines bezpieczeństwa ADI:EDI < 1), co w razie długotrwałości tego stanu może przynieść negatywne skutki zdrowotne, szczególnie dla niemowląt i kobiet w ciąży. Przez cały okres badań woda pochodząca z 5 studni spełniała zalecane normy. Dobra jakość tej wody wynikała z właściwej lokalizacji studni, a także przestrzegania przez rolników Kodeksu Dobrej Praktyki Rolniczej. 40 41

ARCHIVES OF ENVIRONMENTAL PROTECTION vol. 36 no. 4 pp. 41 - 54 2010 PL ISSN 0324-8461

PAH SOIL CONCENTRATIONS IN THE VICINITY OF CHARCOAL KILNS IN BIESZCZADY

EWA LISOWSKA

W. Szafer Institute of Botany, Polish Academy of Sciences Lubicz str. 46, 31- 512 Kraków

Keywords: PAH, Bieszczady, charcoal kilns, soil contamination.

Abstract: The aim of the study was to assess the degree of soil contamination by PAHs in the area of charcoal kiln basis, located in the East Carpathian Biosphere Reserve. The concentrations of PAH in soil samples derived from various sampling locations pointed to a strong or a very strong contamination of the ecosystem by these compounds (8,95 µg×g-1 –283,53 µg×g-1). PAH concentrations in the soil differed significantly between the sampling locations. Analysis of samples from different soil layers (to 30 cm) pointed to a threat of washing out into groundwater. The highest concentrations of PAH corresponded to soil samples collected near kilns (distance of 1.5 m), and were in the range of 17.81 µg×g-1 – 435.54 µg×g-1. PAH content in soil gradually decreased with increasing distance from the kilns to values < 1 µg×g-1. The analysis of the data from three sampling periods (June-August) pointed to higher concentrations of PAHs in soil collected in the middle of the burning season, what was probably due to their more intense emission and a relatively small amount of precipitation.

INTRODUCTION

Polycyclic aromatic hydrocarbons (PAHs) are classified as a group of semi-volatile organic compounds (SOCs), occurring both in the solid and gas phase. The widespread occurrence of polycyclic aromatic hydrocarbons is due to their structure and physical- chemical properties. Their molecules are composed of two or more benzene rings. Lower (two- and three-ring) homologues are more volatile and more soluble in water than four- ring hydrocarbons, which are better soluble in oils and organic solvents [16, 27]. These compounds show toxic properties, both for animals (they can be built in nucleotides of DNA and RNA), and for plants, because they are precursors for the synthesis of tropospheric ozone. Due to the mutagenic properties, PAHs have been classified by the Euro- pean Commission and the U.S. Environmental Protection Agency as major air pollutants, 16 of which were deemed most harmful to human health [6, 18]. Polycyclic aromatic hydrocarbons are formed as a result of natural and anthropogenic processes. Human activity introduces large amounts of these compounds into the environment. They are outcome of incomplete combustion of various types of fuel, such as wood, charcoal, oil, gas oils, and as a result of industrial activity. Significantly high concentrations of these compounds have been identified in samples from large urban ag- glomerations, where industry and traffic are the main source of pollution [7]. Factors that increase the intensity of PAH emission into the atmosphere are the incomplete combus- 42 EWA LISOWSKA tion of organic matter with low oxygen concentration and temperatures in the range of 650-900°C. Contamination of non-industrialized and agricultural areas may occur as a result of using fuel to heat homes and charcoal burning [6, 18, 37]. Almost 90% of the total amount of PAHs in the environment concentrates in soil [8]. Due to the sorption and accumulation properties of soil, PAHs may remain in it even for decades. PAH binding in soil depends on the amount of organic matter and increases along with its increasing content in soil [22, 34]. PAH low solubility in water limits migration to deeper soil layers and leaking into groundwater [22]. During the active burning season in the East Carpathian Biosphere Reserve, there are about 200 charcoal kilns, acting as the major point sources of PAH contamination in this area. The Reserve was recognized under UNESCO’s Man and Biosphere (MAB) Program in 1992. Bilateral Polish-Slovak Biosphere Reserve was extended to Ukrainian part in 1995 and presently covers a total area of 2,132.11 km2 [5]. Polish part (1089 km2) was divided into 3 zones: core and buffer zone – Bieszczady National Park, transitional zone – Ciśniańsko-Wetliński Landscape Park and Dolina Sanu Landscape Park. Local economic activities may occur only in the transitional zone, according to environmental protection regulations [20]. Reserve became one of the most important area of the Nature 2000 network in the European Union and includes the areas of special bird and habitat protection. The East Carpathian Biosphere Reserve has a significant value for conservation of biodiversity, natural ecosystems and genetic resources. It provides the ability of ecological and environmental research and creating conditions for popular and specialist education [12, 20]. For this region, with exception of some specific reports, there is no information on the level of organic pollutants in the environment. Neither industry nor road network is developed in the Reserve nor in the adjacent area, and the total territory is assumed to be unpolluted. Research conducted in 1997–1999 in the whole territory of Carpathians showed high concentrations of ground tropospheric ozone (approximately 100 mg m-3) in the Bieszczady region, both in the Polish and Ukrainian part [9]. It seems that residential heating and charcoal kilns may constitute a major source of pollution in the area. The degrading effect of charcoal production on soil was under the study in the East Carpath- ian in 1998 [10, 31]. It showed increasing intensity of soil metabolism after a long-term exposure to smoke. The purpose of this study is to assess the degree of soil contamination with PAHs and the risk from the charcoal kilns activity for the environment of the East Carpathian Biosphere Reserve.

MATERIALS AND METHODS

Area of research Area of research was located in the vicinity of 6 charcoal kiln bases, placed directly in the forest within the East Carpathian Biosphere Reserve (southern Poland) (Fig. 1). Six sampling locations spread up to 60 m from each charcoal kiln basis. The bases differed in their burning activity and the number of kilns. These were: 1 – Wola Michowa (seven active kilns for 10 years) 2 – Balnica (closed in 1998), 3 – Wola Michowa (four closed kilns, three – in 2008, one in 2006), 4 – Maniów (ten kilns operating since 2004), 5 – Szczerbanówka (three kilns closed in 2007), 6 – Habkowce (two kilns operating since PAH SOIL CONCENTRATIONS IN THE VICINITY OF CHARCOAL KILNS IN BIESZCZADY 43

2006). Samples representativeness was conditioned by similar physical and chemical soil properties in all sampling locations which are presented in Figure 1.

Fig. 1. Sampling locations 1–6 (1 – Wola Michowa, 2 – Balnica, 3 – Wola Michowa, 4 – Maniów, 5 – Szczerbanówka, 6 – Habkowce).

Locations 1–5 were in Komańcza Forest District, location 6 – in Cisna Forest Dis- trict. Sampling points were located in the respective distances to avoid overlapping of emission from individual kilns. The choice of kilns was conditioned by using the same technology. In the area of 5 charcoal kiln bases (1–5) portable ring kilns of the capacity of 15 m3 were used, however, in Szczerbanówka (6th charcoal kiln base) there was a closed stable kiln of the capacity of 20 m3. In 5 charcoal bases (1–5) beech wood (approx. 90%) with an admixture of grey alder, silver birch and aspen poplar were used for charcoal production. Only in the base in Habkowce (6) gray alder and sycamore maple were used. Charcoal is formed as a result of destructive distillation process of wood, running without air. The phase of wood carbonization lasts 24 hours and is characterized by emission of white, thick smoke from the kiln. During burning process the temperature inside the kiln is 400°C. Wood burning is a source of different gaseous and atmospheric particle bound pollutants (CO, CO2, CH4, H2, NO, NO2, O2), PAHs, heavy metals as well as liquid products (tar, methanol, water) and solids (charcoal, coke breeze) [15].

Sampling collection In each of the sampling location, 4 sampling points were set up, up to 60 m from the kilns in a downwind direction. The first sampling point was located 1.5 m from the kilns, the second – 20 m from the kilns, the third – 40 m from the kilns and the fourth – 60 m from the kilns. The total number of sampling points was 24 (4 sampling points for each of 6 sampling locations). One surface sample (0–5 cm) was collected at each sampling point, with two-month intervals during a season (June-October). Samples from each sampling 44 EWA LISOWSKA location were collected during three sampling periods, except those from the 6th location, where the first collection was not included. Additionally, soil samples were collected from the depths of 5, 15 and 30 cm at the distance of 1.5 m from the kilns, in three sampling locations (1, 4 and 5). Soil samples were transported and stored in closed polyester bags.

Analytical methods In dried, sieved fractions, concentration of 16 PAHs (from the U.S. EPA) and soil pH value were determined. Extraction of PAH from soil samples was carried out using the ASE-200 Dionex Company extractor, nitrogen shortened the time. The method validation was performed according to Dionex Application [3]. The extraction was carried out with a mixture of dichloromethane and acetone (1:1) at the temperature of 100°C and nitrogen pressure of 14 MPa. In order to eliminate fresh water from the analyzed samples, approximately 7.0 g of each soil sample was mixed with sodium sulfate [30]. Then the samples were further purified using membrane SPM (Semi Permeable Membranes), made from a semi-permeable polyethylene foil 80 μm thick. Purification using SPM membranes was based on the diffusion of analytes from the extract placed inside the membrane sleeve to the solvent located outside the membrane through the pore with the diameter of about 1 nm, retain- ing fat and other macromolecular impurities. N-hexane was used as the receiver solvent. The extract was concentrated in the air stream, and dissolved in 1 ml of acetonitrile, then injected into the HPLC system column. Isolation of PAH from water samples was carried out by shaking twice with n-hexane. In order to improve wettability of potential suspensions and glass surface, before extraction, 1% (v/v) methanol was added to samples [4]. The concentrated extracts after dissolution in acetonitrile were analyzed. Determination of extracted compounds was carried out by HPLC using the DX-500 Dionex Company chromatograph, with the VYDAC 201TP5415 column. The mobile phase was acetonitrile - water given with gradient of 50–100% acetonitrile in 45 min. at a flow of 1 ml/min. Quantitative determination of PAH (naphthalene, acenaphthylene, acenaphthene, fluorene, phenanthrene, pyrene, benzo/a/anthracene, chrysene, benzo/b/fluroanten, benzo/a/pyrene, benzo/k/fluoranthene, benzo/g, h, i/perylene, dibenz/a, h/anthracene, indeno/1,2,3-c, d/pyrene) was carried out using UV-VIS detector with wavelength 254 nm at 30°C. Chromatograph calibration and quantitative analysis control were carried out using the external standard method based on certified reference material CR104. The PAH concentration in soil samples was converted to dry weight equivalent. Soil pH was determined potentiometrically in 1 mol KCl×dm-3 solution in redistilled water [35].

Statistical methods PAH concentrations [µg×g-1] in soil samples from 6 sampling locations were used in statistical calculations. For each location, the following descriptive statistical parameters were calculated: arithmetic mean, standard deviation, and coefficient of variation. The data were statistically evaluated using one and two factor analysis of variance without repetition test (ANOVA, MANOVA) and post hoc NIR, accepting the likelihood of error p = 0.05. ANOVA was performed for 5 sampling locations (1–5). The sixth location was excluded due to the presence of an additional source of emission. The data used in the statistical analysis was transformed into a natural logarithm. PAH SOIL CONCENTRATIONS IN THE VICINITY OF CHARCOAL KILNS IN BIESZCZADY 45

RESULTS

PAHs in soil Mean concentrations of total PAH in soil from six locations ranged from 8.96 µg×g-1 (location 3) to 283.53 µg×g-1 (location 6) (Table 1). Mean concentrations of light and heavy fractions of PAH and B(a)P were in the range: 1.13 µg×g-1-22.76 µg×g-1 (2–3 PAHs), 6.84 µg×g-1-260.77 µg×g-1 (4–6 PAHs), 0.01 µg×g-1-0.49 µg×g-1 (B(a)P) (Table 1). The standard deviation and coefficient of variation were higher for mean values of PAH at locations with operating kilns (1, 4 and 6). The standard deviation did not exceed twice the arithmetic mean, which indicates normal distribution of data (Table 1).

Table 1. Mean concentrations of the total PAH [µg×g-1] in soils from all sampling points in each sampling location (1–6. see Fig. 1) in the vicinity of charcoal kiln basis in the Bieszczady (for three sampling periods).

Location Standard Coefficient of n Mean Maximum Minimum no deviation variation [%] 1-3 Sampling period 1 12 111.46 709.84 2.36 206.34 183.48 2 11 15.30 45.74 0.52 14.13 92.36 Total 3 12 8.95 26.31 0.25 8.17 91.27 PAH 4 12 184.96 1072.92 1.45 313.39 169.44 5 12 10.27 30.10 0.08 9.75 94.96 6 8 283.53 835.92 8.76 363.06 128.05 1 12 18.96 116.38 0.04 35.03 184.74 2 11 1.13 4.92 0.05 1.73. 153.50 3 12 2.12 8.49 0.03 3.00 141.89 2–3 PAH 4 12 17.29 80.22 0.54 25.68 148.57 5 12 3.08 14.10 0.00 4.11 133.45 6 8 22.76 97.02 0.16 36.71 161.30 1 12 93.49 650.80 2.29 183.34 196.10 2 11 14.17 41.53 0.39 13.49 95.25 3 12 6.84 19.51 0.09 5.70 83.30 4–6 PAH 4 12 167.67 1025.43 0.64 297.32 177.32 5 12 7.18 25.81 0.08 2.51 121.56 6 8 260.77 738.90 6.72 331.57 127.14 1 12 0.09 0.50 0.00 0.15 180.00 2 11 0.01 0.05 0.00 0.02 129.98 3 12 0.02 0.13 0.00 0.04 160.47 B/a/P 4 12 0.49 5.14 0.02 1.47 298.57 5 12 0.01 0.07 0.00 8.73 168.48 6 8 0.15 0.67 0.01 0.24 163.64 1 12 5.12 6.86 3.57 1.15 21.41 2 11 5.00 6.49 3.48 0.90 19.83 3 12 7.22 7.49 6.33 0.40 9.46 pH 4 12 5.98 3.90 7.78 1.37 24.15 5 12 5.24 7.64 3.42 1.69 32.52 6 8 7.02 7.52 5.59 0.64 10.61 46 EWA LISOWSKA

Two-factor ANOVA showed highly significant (p ≤ 0,001) effect of time and place of samples collection on the total PAH and 4–6 ring PAH concentration in soil with no significant interaction between both factors. 2–3 ring PAH concentration was only slightly dependant on the location of sampling points (p ≤ 0.05). Concentration of PAH in soil from different locations was related to the period of kilns activity and number of kilns. Concentration of PAH and 4–6 PAH (four-, five- and six ring polycyclic aromatic hydrocarbons) was significantly higher in soil samples from the locations with active kilns (1 and 4 location) comparing with other locations (p ≤ 0.001), whereas concentration of 2–3 PAH (two- and three ring polycyclic aromatic hydrocarbons) was slightly higher at the 1st and the 4th location (p ≤ 0.05). The highest concentrations of total PAH and their non-volatile fraction were determined at the 6th and the 4th location, while the highest concentrations of volatile hydrocarbons were at the 6th and the 1st location (Table 1). The percentage of non-volatile hydrocarbons was 82% of identified hydrocarbons in examined soil samples, and indicated higher values at two operating locations – 1 and 4 (Fig. 2).

Fig. 2. The percentage of 2–3 ring PAH (A) and 4–6 ring PAH (B) fractions in soils at sampling locations (1–6, see Fig. 1) in the vicinity of charcoal kiln basis in the Bieszczady (for three sampling periods).

Concentrations of 16 individual polycyclic aromatic hydrocarbons from the six sampling locations are presented in Table 3 (for explanation of abbreviations see Annex). Among the identified PAHs the highest values were reached by dibenz/a,h/antracen – 194.23 µg×g-1 and benzo/k/fluroanten – 168.95 µg×g-1 (Table 2). The analysis performed for samples taken from the soil profile showed that individual PAH concentrations generally decreased steadily with depth in soil profile. The mean concentrations of PAH in soil at a depth of 15 cm were about 1/3, and at a depth of 30 cm almost 2/3 lower than in the surface layer (Fig. 3). PAH SOIL CONCENTRATIONS IN THE VICINITY OF CHARCOAL KILNS IN BIESZCZADY 47

Table 2. Mean concentrations of 16 PAHs [µg×g-1] in soils from 6 sampling locations (1–6, see Fig. 1) in the vicinity of charcoal kiln basis in the Bieszczady (for three sampling periods).

1-3 Sampling period Mean SD ٭WWA 1 2 3 4 5 6 Nap 0.12 0.34 0.07 4.30 1.46 3.18 5.37 Acy 0.26 0.17 0.59 0.31 0.25 2.41 2.19 Ace 0.34 0.11 0.04 5.25 0.10 2.73 6.91 Fle 0.22 0.26 0.05 1.18 0.12 6.26 4.57 Ph 0.26 0.07 0.14 0.37 0.34 3.91 2.58 An 2.54 0.12 1.22 6.29 0.81 15.50 19.28 Fla 3.44 1.17 0.07 0.42 0.06 13.77 8.06 B/a/A 0.33 0.19 0.06 2.06 0.09 1.57 3.42 Py 0.84 0.01 0.02 0.53 0.02 0.15 0.82 B/a/P 0.09 0.02 0.03 0.27 0.03 0.65 0.44 Chr 0.13 0.02 0.01 0.10 0.03 0.08 0.97 B/b/F 0.08 0.01 0.01 0.05 0.01 0.07 0.10 B/g,h,i/P 5.02 6.35 2.06 168.95 3.70 45.49 181.09 B/k/F 1.45 0.21 0.04 0.04 0.02 2.02 1.46 D/a,h/A 10.50 4.68 3.10 73.10 2.03 194.23 112.87 Ind 5.61 0.69 1.44 5.90 1.20 2.73 5.32 * Nap–Naphthalene, Acy–Acenphthylene, Ace–Acenaphthene, Fle–Fluorene, Ph–Phenanthrene, An-Anthra- cene, Fla–Fluoranthene, Py–Pyrene, B/a/A-Benzo/a/anthracene, Chr–Chrysene, B/b/F–Benzo/b/fluoranthene, B/a/P–Benzo/a/pyrene, B/k/F–Benzo/k/fluoranten, Ind–Indeno/1,2,3-c,d/pyrene, D/a,h/A-Dibenz/a,h/anthracene, B/g,h,i/P–Benzo/g,h,i/perylene

Fig. 3. Participation of individual PAHs in soil profile (1: 0–5cm; 2: 5–15cm; 3: 15–30cm); mean values for three locations (1, 4 and 5 see Fig. 1). 48 EWA LISOWSKA

Concentrations of total PAH, 2–3 ring PAH, 4–6 ring PAH were significantly correlated with each other (p < 0.01; according to Bonferronii’s correction p < 0.0016). Total PAHs were much more highly correlated with 4–6 ring PAHs ( = 0.99) than with 2–3 ring PAHs (r values from 0.70 to 0.73). Concentrations of PAH slightly correlated with soil pH values (p < 0.01; according to Bonferronii’s correction p < 0.0016). The relationships between pH and total PAHs and 4–6 ring PAHs (r values from 0.40 to 0.41) were a little stronger than relationship pH/2–3 ring PAHs (r = 0.38).

The variation in the PAH concentrations during the growing season The concentration of PAH and their non-volatile fractions in the soil showed large seasonal variations (p ≤ 0.001), being significantly higher in soils collected in July (p ≤ 0.001) and October (p ≤ 0.05) (Fig. 4). In July the highest values of total PAH and 4–6 PAH were observed for location 1 and 4, while in October for location 6. 2–3 PAH concentration showed no significant seasonal variation (Fig. 4).

The effect of distance from source pollution on PAHs content in soil One-factor ANOVA showed a significant effect of sampling points location within a single charcoal base on PAH concentration in soil. The effect was stronger (p ≤ 0.001) in the case of 2-3 ring compounds than for total and 4–6 rings (p ≤ 0.05). In each sampling location, the highest concentrations of PAH were found in the soils collected closest to kilns. Along with increasing distance from the burning point, PAH concentrations decreased (Fig. 5). The mean concentrations of 2–3 ring PAH in soils collected the farthest from the kilns (60 m) were significantly lower than at the other sampling points (1.5 m, 20 m and 40 m) (p ≤ 0.001). The mean concentrations of total PAH and 4–6 PAH in soil samples collected the closest to the kilns (1.5 m) were significantly higher than means in the soils collected 40 and 60 meters away from the kilns (p ≤ 0.05).

DISCUSSION

In the East Carpathian Biosphere Reserve, the absence of industrial activity and heavy traffic point to residential heating and charcoal kilns as the main sources of PAH emission. As a result of wood burning at homes larger quantities of some hydrocarbons (Ph, Fla, Py, B/a/P) are emitted to the atmosphere [11]. During the warm period, when the test was performed, home heating emission was significantly reduced, however, at the end of the spring, charcoal burning season began, which ended in the autumn, together with the increased precipitation [18]. According to Polish Regulation [25] the limit value for the content of 9 PAHs in the upper layer of soils from the protected areas under the Nature Conservation Act is 1 µg×g-1, while the content of individual compounds should not exceed 0.01 µg×g-1 (with exception of B/a/P with limit concentration of 0.02 µg×g-1) [25]. In Poland, the mean content of PAH in uncontaminated soils varies within the limit of 0.3–0.6 µg×g-1 [19, 33]. In highly contaminated areas (affected by industry or transport) PAHs concentrations may reach hundreds of µg×g-1 [2, 33]. Research conducted in the Bieszczady Mountains showed very high concentrations of polycyclic aromatic hydrocarbons. The mean concentrations of PAHs in soils at the PAH SOIL CONCENTRATIONS IN THE VICINITY OF CHARCOAL KILNS IN BIESZCZADY 49

Fig. 4. Variability of total PAH (A), 2-3 PAH (B)and 4-6 PAH (C) [µg×g- 1] concentration in soils from 6 sampling locations (1-6, see Fig. 1) in the vicinity of charcoal kiln basis in the Bieszczady (for three sampling periods). 50 EWA LISOWSKA

Fig. 5. Variations in concentration of the total PAHs [µg×g-1] in soils collected from 1.5 to 60 m from the kilns located in the vicinity of charcoal kiln basis in the Bieszczady (means for three sampling periods). different sampling locations varied from 8.95 µg×g-1 (location 3) to 294.79 µg×g-1 (location 6), and at individual points they even reached the value of 1500 µg×g-1 in soil. PAH concentration from different locations was dependent on the length of each charcoal base activity, number of kilns and burning intensity. At the locations with active kilns (1, 4, 6), these values were from a few to several times higher than those from inactive locations (3, 5). The highest mean concentrations of PAH were determined at the 6th location, where there were 2 active kilns. These values were higher than in other active charcoal bases with more kilns (location 1 and 4), what may be due to an imposition of pollutants from other, close existing charcoal bases. A similar trend was observed for the volatile and non- volatile fraction, and B/a/P. As a result of wood combustion more non-volatile PAHs are emitted into the atmosphere comparing emission from coal and fuel burning process. Coal-burning on an industrial scale (e.g., coke furnace) may emit up to 90% of individual volatile compounds, e.g. naphthalene, while the wood burning at homes emits about 10% of this hydrocarbon, and the percentage of non-volatile fraction increases several times (Fla, Py, B/a/P B/g,h,i/P) [11]. The process of wood burning at a reduced amount of oxygen can also increase the amount of emitted 4–6 PAHs up to 80% [37]. The prevailing percentage of non-volatile PAH fractions in relation to volatile fractions was found throughout the whole research area. This proportion varied slightly between closed and active charcoal bases, where the participation of 4–6 PAHs was even greater. The reason of the difference was probably more intensive 4–6 PAHs emission than 2–3 PAHs emission. Part of emitted PAH volatile fractions might undergo photochemical oxidation process [21], reducing their concentration in precipitation. In addition, during the vegetative season, green plants contribute significantly to the retention of pollutants and can remove up to 44% ambient air pollutants [26]. PAH SOIL CONCENTRATIONS IN THE VICINITY OF CHARCOAL KILNS IN BIESZCZADY 51

Processes, which PAHs undergo in the soil, depend on both the physicochemical properties of soil, and individual compounds, whose properties presented by Malisze- wska-Kordybach [17] have a greater impact on PAH transformations. High PAH concentrations in soils rich in organic matter may be related to high accumulation and slow desorption processes. PAH binding in soil depends on the amount of organic matter, especially its lipophilic organic fraction. Soils with a high content of organic matter contain usually more PAHs than soils with low humus content [17]. Acid brown soils are typical of Bieszczady Mountains and the average organic-C content ranges from 4 to 5% [35]. Studies on charcoal kilns emission showed a higher carbon content in soils near the kilns and, therefore, high levels of PAH was the result of both their binding intensity in soil and continuous, intense emission [31, 34, 36]. Besides, 4–6 ring PAHs show a greater sorption capacity for organic matter than the volatile fractions, which can easier evaporate from the soil surface [22, 26]. Creating bonds between PAHs and humus is also dependent on the presence of cata- lysts (chemicals, enzymes), and sustainability of these bonds may increase with increasing hydrolytic acidity value of the soil [34]. Researches conducted in charcoal kiln bases, relic charcoal hearths and surrounding sites indicated higher pH values in soil [10, 31, 36]. The studied soils were characterized mostly as slightly acidic, which had additional effect on PAH binding by organic matter (Tab. 1) and creation of unfavorable environment for soil microorganisms, which could contribute to PAH biodegradation [13, 23]. Soil physical factors such as humidity and temperature, have a direct effect on PAH metabolism in soils. Favorable soil humidity enhances biodegradation of these compounds by soil microorganisms [36]. However, high soil humidity inhibits PAH adsorption by organic matter and mineral fractions, due to PAH interaction with water molecules [23]. Low temperature favors the sorption of PAH by soil colloids and influences reduction of solubility, decreasing the activity of soil microorganisms [23]. High soil temperature causes evaporation of PAH volatile fractions from soil, even adsorbed permanently on soil particles [23]. The highest PAH concentration in soil from two active charcoal bases (location 1 and 4) was found in the second period (July), and the lowest in the first period (June). These differences are likely due to burning intensity, and PAHs accumulation during the following months. Due to heavy rainfall occurring in the middle of the season, they can contribute significantly to the PAHs migration into the soil profile, where these compounds may be permanently bound, thus increasing their concentration in the soil. High temperatures in the beginning of the season may have contributed to PAH evaporation from the soil, and intense sunshine induced photochemical oxidation process, what confirmed the lower value of volatile fractions in this period. The PAH concentrations in soil at the depth of 30 cm were more than twice lower than in the surface layer (Fig. 3). However, these values were still high, posing a threat to plants and surface waters. PAH accumulation occurs mainly in the surface soil layer up to 20 cm, although some quantities are also present at the depth of 2 m (e.g., Ph) [28]. This is related to the presence of organic matter (PAH binding) in the surface soil layer and the low solubility of PAH in water, which significantly slows down the movement into the soil profile. Some PAHs may migrate to deeper soil layers and the process depends on the presence of carriers in the soil, such as: soluble organic matter, inorganic colloidal particles (clays, iron oxides), soil microorganisms [23]. As a result of PAH emissions into the atmosphere, some of them can be transported for up to several hundred meters away [1]. PAHs with low molecular weight can be 52 EWA LISOWSKA transported for long distances, unlike five- and higher ring PAHs, which deposit close to the source of emissions. This may explain high concentrations of those compounds in studied soils (Fig. 2). Humidity decreases the amount of PAH in the atmosphere; rain and snow cause their removal from the atmosphere to ground [18]. In the studied soils, PAH concentration decreases with increasing distance from the source of emission, but a significant decrease was visible at the distance of 40 m. At a distance of 60 m the concentration of PAH was almost ten times lower than at the sampling point closest to the kilns. Unfortunately, all along the transect PAH concentrations were very high in comparison to unpolluted areas. In all locations the concentrations of PAHs were at least 10 times higher than the Polish guide values [25] and in some of the sampling points they exceeded the limit values of over two orders of magnitude. Research conducted in the area of relic charcoal hearths existing in the 19th and 20th centuries showed minimal lasting effect on soil properties (higher pH, carbon and calcium content). Forest density and biodiversity differ on hearths comparing adjacent area [13, 36]. Studies on existing kilns reveal significant alterations in forest soil chemistry (higher pH, carbon and calcium content, cation exchange capacity with lower P content) pointed to local soil and air pollution, changes in local vegetation (lower density, plants degradation), threat of leaking into groundwater [10, 31]. Although high carbon content reduces PAH biotoxicity in soil, extremely high PAH concentrations, comparable with those in studied soils, may have toxic effect on earthworms and most of the soil microorganisms [8, 34].

CONCLUSIONS

The results of the study pointed to very high PAH concentrations in soil in the vicinity of charcoal kiln bases in the Bieszczady area. This may create a serious ecological risk to East Carpathian Biosphere Reserve. Concentrations differed significantly between the sampling locations and reached higher values at locations with the longest and the most intense burning activity. PAH concentrations gradually decreased along with increasing distance from the kilns. Analysis of the data from three sampling periods (June-August) pointed to a higher PAH concentration in samples collected in the middle of the burning season.

REFERENCES

[1] Aamot E., E. Steinnes, R. Schmid: Polycyclic aromatic hydrocarbons in Norwegian forest soils: Impact of long-range atmospheric transport, Environ. Pollut., 92, 275-280 (1996). [2] Adamczewska M., J. Siepak, H. Gramowska: Studies of levels of polycyclic aromatic hydrocarbons in soils subjected to anthropopressure in the city of Poznań, Pol. J. Environ. Studies, 9(4), 305-321 (2000a). [3] Application Note 313, DIONEX: Extraction of PAHs from environmental samples by Accelerated Sol- vent Extraction (ASE), 1997. [4] Bartulewicz J., J. Gawłowski, E. Bartulewicz: Pobieranie i przygotowanie prób do oznaczania związków organicznych metodami chromatografii, Państwowa Inspekcja Ochrony Środowiska Biblioteka Monitor- ingu Środowiska, Warszawa 1997, 13-27. [5] Biosphere Reserve Information. Poland-Slovakia-Ukraine East Carpathian. The MAB programme, www. unesco.org, 26.02.2007. [6] Bojakowska I., G. Sokołowska: Polycyclic aromatic hydrocarbons in materials of burned peatlands, Pol. J. Environ. Studies, 12, 401-408 (2003). [7] Borkowska M., A. Dutkiewicz, A. Strusiński: Wpływ transportu samochodowego na zawartość benzeno/a/ pirenu w glebie i roślinach, Roczniki PZH, T. XXXIII, 4, 297-302 (1982). PAH SOIL CONCENTRATIONS IN THE VICINITY OF CHARCOAL KILNS IN BIESZCZADY 53

[8] Brandys J.: Wielopierścieniowe węglowodory aromatyczne (WWA). Chemiczne substancje toksyczne w środowisku. Komisja Nauk Medycznych – Oddz. PAN w Krakowie, 163-176 (1990). [9] Bytnerowicz A., B. Godzik, W. Frączek, K. Grodzińska, M. Krywult, O. Badea, P. Barančok, O. Blum, M. Černy, S. Godzik, B. Mankovska, W. Manning, P. Moravčik, R. Musselman, J. Oszlanyi, D. Postelnicu, J. Szdźuj, M. Varšavova, M. Zota: Distribution of ozone and other air pollutants in forests of the Carpathian Mountains in central Europe, Environ. Pollut., 116, 3-25 (2002). [10] Fisher Z., P. Bieńkowski: Some remarks about the effect of smoke from charcoal kilns on soil degradation, Environ. Monitoring & Assessment, 58(3), 349-358 (1999). [11] Gawlik M.B., M. Bilek: Możliwości obniżenia emisji wielopierścieniowych węglowodorów aromatyc- znych ze źródeł antropogenicznych, Medycyna Środowiskowa, 9(1), 79-82 (2006). [12] Główny Urząd Statystyczny. Ochrona Środowiska 2009. Informacje i opracowania statystyczne. Wersja elektroniczna. Warszawa 2009. [13] Hart J.L., S.L.Van de Gevel, D.F. Mann, W.K. Clatterbuck: Legacy of charcoaling in a Western Highland Rim Forest in Tennessee, The American Midland Naturalist, 159(1), 238-250 (2008). [14] Kluska M.: Soil contamination with the polycyclic aromatic hydrocarbons in the vicinity of the Ring Road in Siedlce City, Pol. J. Environ. Studies, 12(3), 309-313 (2003). [15] Laurow Z.: Pozyskiwanie drewna i podstawowe wiadomości o jego przerobie, Wyd. SGGW, Warszawa 1999. [16] Magnus F.: Toxic substances in environment, Environ. Sci. Technol., [in:] A. Wiley: Interscience Series of Texts and Monographs, 279-281 (1994). [17] Maliszewska-Kordybach B.: Zależność między właściwościami gleb i zawartością w nich WWA; na przykładzie gleb z terenów użytków rolnych w województwie lubelskim, Archiwum Ochrony Środowiska, 24 (3), 79-91 (1998b). [18] Maliszewska-Kordybach B.: Sources, concentrations, fate and effect of polycyclic aromatic hydrocarbons (PAHs) in the Environment. Part A: PAHs in air, Pol. J. Environ. Studies, 8, 131-136 (1999). [19] Maliszewska-Kordybach B., B. Smreczek, A. Klimowicz-Pawlas: Concentrations, sources, and spatial distribution of individual polycyclic aromatic hydrocarbons (PAHs) in agricultural soils in the Eastern part of the EU. Poland as a case study, Sci. Total Environ., 104(12), 3746-3753 (2009). [20] Michalik S.: Karpaty Wschodnie. Międzynarodowy Rezerwat Przyrody, Wyd. 2 – Olszanica, BOSZ, 10- 14, 1997. [21] Niu J., J. Chen, D. Martens, X. Quan, F. Yang, A. Kettrup, K. W. Schramm: Photolysis of polycyclic aromatic hydrocarbons absorbed on spruce [Picea abies (L.) Karst.] needles under sunlight irradiation, Environ. Pollut., 123, 39-45 (2003). [22] Nelson A.H. (ed.): PAHs and related compounds, The Handbook of Environ. Chem., 3, Part I, Springer- Verlag Berlin Heidelberg, 56-65 (1998). [23] Oleszczuk P.: Zanieczyszczenia organiczne w glebach użyźnianych osadami ściekowymi. Część II. Losy zanieczyszczeń w glebie, Ecological Chemistry and Engineering, 14(S2), 185-198 (2007). [24] Reid B.J., K.C. Jones, K.T. Sample: Bioavailability of persistent organic pollutants in soils and sediments - a perspective of mechanisms, consequences and assessment, Environ. Pollut., 108, 103-112 (2000). [25] Rozporządzenie Ministra Środowiska z 9 września w sprawie standardów jakości gleby oraz standardów jakości ziemi (Dz.U. rok 2002, nr 165, poz. 1359). [26] Simonich S.L., R.A. Hites: Importance of vegetation in removing polycyclic aromatic hydrocarbons from the atmosphere, Nature, 370, 49-51 (1994a). [27] Sims R.C., M.R. Overcash: Fate of polynuclear aromatic compounds (PNAs) in soil-plant system, Resi- due Rev., 88, 1-68 (1983). [28] Stanković D., B. Krstić, N. Nikoli: Effect of traffic on the soil contamination with polycyclic aromatic hydrocarbons (PAHs), Biotechnol. & Biotechnol., 22(2), 736-741 (2008). [29] Steiner C., K.C. Das, M.Garcia, B. Förster, W. Zech: Charcoal and smoke extract stimulate the soil microbiological community in a highly weathered xanthic Ferralsol, Pedobiologia, 51(5-6), 359-366 (2008). [30] Technical Note 208, DIONEX: Methods optimization in Accelerated Solvent Extraction (AEE), 1997. [31] Uvarov A.V.: Effect of smoke emission from a charcoal kiln on the functioning of forest soil system: A microcosm study, Environ. Monitoring & Assessment, 60(3), 337-357 (2000). [32] Van Jaarsveld J.A., W.A.J. Van Pul, F.A. De Leeuw: Modeling transport and deposition of persistent organic pollutants in the European region, Atmos. Environ., 31, 1011 (1997). [33] Wcisło E.: Soil contamination with polycyclic aromatic hydrocarbons (PAHs) in Poland – a Review, Pol. J. Environ. Studies, 7(5), 267-272 (1998). [34] Weissenfels W.D., H.J. Klewer, J. Langhoff: Adsorption of polycyclic aromatic hydrocarbons (PAHs) by soil particles: Influence on biodegradability and biotoxicity, Applied Microbiology and Biotechnology, 36(7), 689-696 (1992). 54 EWA LISOWSKA

[35] Woźniak L., S. Dziedzic: Organic matter and some element contents in soil profile of meadows in the mountain region of Bieszczady – Poland. “Organic matter and element interactions, Austrian – Polish Workshop, ALVA Mitteilungen Heft 3(2005). [36] Young M.J., J.E. Johnson, M.D. Abrams: Vegetative and edaphic characteristics on relic charcoal hearths in the Appalachian mountains, Plant Ecology, 125(1), 43-50 (1996). [37] Zou L.Y., W. Zhang, S. Atkinson: The characterization of polycyclic aromatic hydrocarbons emissions from burning of different firewood species in Australia, Environ. Pollut., 124, 283–289 (2003).

Received: October 10, 2009; accepted: May 20, 2010

STĘŻENIA WWA W GLEBIE W SĄSIEDZTWIE WYPALARNI WĘGLA DRZEWNEGO W BIESZCZADACH

W niniejszej pracy zostały przedstawione dane dotyczące zanieczyszczenia gleb przez wielopierścieniowe węglowodory aromatyczne (WWA) w okolicy wypalarni węgla drzewnego, usytuowanych na terenie Rezer- watu Biosfery Karpaty Wschodnie. Celem badań była ocena stopnia zanieczyszczenia gleb przez WWA oraz zasięgu emisji zanieczyszczeń z poszczególnych wypalarni. Stężenia WWA w glebach pochodzących z posz- czególnych stanowisk badawczych wskazywały na silne lub bardzo silne zanieczyszczenie ekosystemu tymi związkami (8,95 µg×g-1 – 283,53 µg×g-1). Stężenia WWA w glebie różniły się istotnie pomiędzy badanymi stanowiskami. Analiza próbek pochodzących z warstw gruntowych (do 30 cm) wskazywała na zagrożenie przedostawania się części z tych zanieczyszczeń do wód gruntowych. Najwyższe stężenia WWA pochodziły z próbek gleby pobranych najbliżej wypalarni (1,5 m) i mieściły się w przedziale 17,81 µg×g-1 – 435,54 µg×g-1. Zawartość WWA w glebie malała stopniowo wraz z odległością od retort (do 60 m), osiągając jednak nadal wysokie wartości dla dwóch najintensywniej działających wypalani (4 i 6): 0,44 µg×g-1 – 69,3 µg×g-1. Analiza danych z trzech poborów (czerwiec – październik) wskazuje na wyższe stężenia WWA w glebie w środku sezonu wypałowego, co wynika prawdopodobnie z najintensy- wniejszej ich emisji oraz stosunkowo małej ilości opadów atmosferycznych. 55

ARCHIVES OF ENVIRONMENTAL PROTECTION vol. 36 no. 4 pp. 55 - 66 2010 PL ISSN 0324-8461

RELATIONS BETWEEN CIRCULATION AND WINTER AIR POLLUTION IN POLISH URBAN AREAS

JOLANTA GODŁOWSKA*, ANNA MONIKA TOMASZEWSKA

Institute of Meteorology and Water Management, Krakow Branch, Poland Department of Monitoring and Modelling Air Pollutions Piotra Borowego str. 14, 30-215 Kraków * Corresponding author e-mail:[email protected]

Keywords: Urban air pollution, circulation, urban boundary layer, sodar, ground inversion, elevated inversion.

Abstract: We determined the performance of different Circulation Type Classifications (CTCs) to stratify air pollutants concentrations in Polish cities in winter. Our analysis is based on 15 CTCs calculated by COST 733 as well as on 5 manual universally used manual weather type classifications. For this purpose we compared and tested the explained variation (EV) and within-type standard deviation (WSD) methods. Finally, EV method has been chosen for evaluating classifications for daily values of SO2, NO2, PM10 and CO as well as vertical dispersion conditions obtained from SODAR data. We also presented the methodology of choosing smog episode days based on 90-percentile values. For the winter smog episodes data from Krakow different classifications have been compared using Gini coefficient method. The best results for separate air pollution data series as well as for smog episode days were obtained for Hess-Brezowski Großwetterlagen classification (HBGWL). Moreover, good results were obtained for the based on principal component analysis PCACA classification, Pol- ish Niedzwiedz TCN21, modified Polish Litynski LITTc, modified Lamb LWT2, and three modified HBGWL (GWTC26, OGWL, OGWLSLP) classifications. The same classifications except for HBGWL are good for SODAR data. For the best CTCs, the differences between various classes are visible, however a big scattering is still observed. Main urban air pollution problems arise in situations when flow with Southerly component is observed. Correlations between air pollution data and SODAR data (calculated for marginal means obtained for different classes) confirm a negative role of both low height of the ground-based inversion and long duration of the low-level elevated inversion in urban areas.

INTRODUCTION

Urban areas are characterised by both large heterogeneity of emission and spatial variability of roughness. Turbulence induced by merged interactions of temperature, wind and friction causes some homogenization of the air. However, physical, chemical and meteorological parameters characterizing air conditions change inside cities [6, 24]. Nev- ertheless, a similar day by day variability of different pollutants measured in different places inside a city is observed. It shows that there is some outside factor which influences this common changeability. Air pollution concentration depends on both emission and dispersion [8]. As emission usually does not change quickly with time (except for daily variability), weather conditions connected with dispersion play the crucial role in changing pollutants concentration [9]. Nowadays, means of transport and heating sys- 56 JOLANTA GODŁOWSKA, ANNA MONIKA TOMASZEWSKA tems based on coal burning are the main sources of winter air pollution in Polish urban areas. As these sources of emission are situated near the ground level, mainly in urban canopy layer or just above it, meteorological conditions such as wind speed and thermal structure of urban boundary layer, play a key role in displacing pollutants outside the city. The ground - based inversion of temperature causes the accumulation of pollutants in low-lying layer. Moreover, the area accessible for vertical mixing becomes smaller as a result of the blocking caused by the elevated inversions. This reflects the behaviour of air pollution concentrations that are the highest when low-lying elevated inversion can be found [11]. A connection between synoptic situation and air quality in Upper Silesia and Kraków was originally examined using Niedzwiedz classification [21, 22]. Later research showed that the increase of air pollution concentrations in Krakow is observed for Litynski SWA and OOO classes [10]. For these classes frequent low-lying elevated inversions are also seen. The connection between high PM10 concentration and circulation classified by Niedzwiedz was also confirmed in Katowice agglomeration [2]. In this paper, the usefulness of different circulation type classifications for winter urban air pollution is examined.

DATA AND METHODS

A key goal of this paper is to compare the performance of different Circulation Type Classifications (CTCs) to stratify concentrations of air pollution in winter. The analysis is based on catalogues of atmospheric circulation types that have recently been made available within the COST 733 action of the European Science Foundation (www.cost733.eu). We use the release 1.1 and 1.2 of circulation type classifications catalogue [4]. Compari- son between CTCs with different number of classes is difficult, so we analyse only the subset of collection with about the same number of classes (about 27). Finally, 15 “objective”, i.e. computer-assisted methods of classification developed using the ECMWF ERA-40 dataset and 5 manual, universally used, CTCs are used. The catalogues used are listed in Table 1. For a thorough description of CTCs see [23, 12]. The classification procedures used within COST 733 action are applied on Pan-European scale (large domain) as well as on scale of a few countries (ten sub-domains). In this paper objective CTCs from domain D07 (3°E-26°E, 43°N-58°N) are used (Figure 1). Additionally, Pol- ish Niedzwiedz’s (TCN21 – 21 classes) classification [19, 20] is analysed. The analysis has been made using daily mean of SO2, NO2 and PM10 as well as maximum daily 8 - hour mean of CO from Warsaw, Krakow, Upper Silesia, Lodz and Wroclaw. Research has been conducted for winter air pollution data (December, Janu- ary, February) from 1997 to 2002 with the exception of Krakow, where data from 1994 to 2002 have been analysed. Air pollution data come from the AIRBASE database [1]. Meta - information from AIRBASE database concerning stations (type of area, type and altitude of station) and amount of missing data was used. Moreover, on the basis of the SODAR data from Krakow, the relation between boundary layer conditions and circulation was examined. The SODAR data come from December, January and February of between 1994 and 1999. Different parameters characterising thermal structure and variability of Urban Boundary Layer were determined. For each day a mean height of convection (C), ground-based inversion (GI) and elevated inversion (EI) were calculated. Moreover, we specified the number of hours when each of these boundary layer phenom- ena was observed. RELATIONS BETWEEN CIRCULATION AND WINTER AIR POLLUTION IN POLISH URBAN... 57

Previously, the relationship between the atmospheric circulation and the surface environment was determined mainly on the basis of the Pearson correlation coefficient, main square error, and root main square error [27, 5]. In this paper, two new methods suggested by COST733 action such as Explained Variation (EV) and Within – type Stan- dard Deviation (WSD) [15] are tested and used to establish separability and within-type variability characterizing different classifications. Apart from these methods, the Gini coefficient method [7] is used to determine which one of CTCs is the best for determination of smog days. Usage of EV (EV = 1-WSS/TSS) is based on the fact that for a grouped random variable, the sum of squares of deviation from the overall mean (TSS) can be divided into two parts: the sum of squares based on the within-group variability (WSS) and the sum of squares based on the between-group variability (BSS) [3]. The best CTC for stratify- ing air pollution data ought to have the lowest WSS. This gives rise to conclusion that the best CTC is connected with the highest values of EV. Within – type Standard Devia- tion measures average within-type standard deviation after partitioning data into k classes (where SDi is the standard deviation in i class). For this method, the problem of comparing different CTCs correctly will arise, if we have different number of cases in different classes. Higher separability and smaller within-type variability are connected with lower WSD values. Both methods were tested on the basis of COST 733 classification catalogue data (cost733cat-1.1) and winter air pollution data (1999 – 2002) from Upper Silesia. The relationship between the number of classes and evaluation of different CTCs is observed for both parameters (Fig. 1 – the left and middle).

Fig.1. Relation between number of classes N for different classifications and EV (the left), WSD (in the middle) and nWSD (the right)

Moreover, WSD values depend on the kind of pollutants. It is very inconvenient and made us normalize WSD by dividing it by standard deviation calculated for all data (nWSD – Fig. 1 – the right). The best classification chosen by nWSD is ESLPC30, while EV favours LWT2 and HBGWL (Fig. 2). The comparison between expected marginal means of SO2 for ESLPC30, LWT2 and HBGWL led us to choose EV parameter for evaluation of CTCs as classification with higher separability is preferred by EV (Fig. 3). 58 JOLANTA GODŁOWSKA, ANNA MONIKA TOMASZEWSKA

Fig. 2. Comparison between EV and nWSD values calculated for CTCs from catalogue 1.1 and daily SO2 in Gliwice

Fig. 3. Expected marginal means of SO2 for ESLPC30 classification (chosen by WSD and nWSD methods - on the left) as well as for LWT2 and HBGWL classifications (chosen by EV method – the middle and the right).Confidence interval 95% is marked

The third method allows to observe how different CTCs separate days when air pollution thresholds are exceeded for many pollutants and stations. In this case evaluation methods proposed by COST733 cannot be used and as a result we applied the Gini coefficient method [7] based on the Lorenz curve [18]. In this method only two numbers for each class characterise every classification – a number of days meeting our criteria (e.g. smog days) and total number of days for each class. In order to calculate Gini coefficient

G for a classification, the probabilityp i = mi/ni of occurrence days with some characteristic (e.g. high pollution concentration, large precipitation, fog) for each class ought to be calculated and finally sorted according to risingp i. Then

for

where ni is a total number of days for class i (after sorting), mi is a number of days meeting our criteria for class i , N is a total number of days for all classes, M is a total number of days meeting our criteria for all classes and L is a number of classes. RELATIONS BETWEEN CIRCULATION AND WINTER AIR POLLUTION IN POLISH URBAN... 59

RESULTS

Daily values of SO2, PM10, NO2 and CO in different Polish cities In order to determine separability of pollutants’ concentrations by circulation types we used EV values calculated for different CTCs from catalogue 1.2. A similar EV variability (Table 1 - top) according to CTCs for the same pollutant measured in different places is observed. However, the mean EV level is neither the same for each pollutant nor for each city. Probably it is because of monitoring stations’ localisation and irregular distribution of emission sources. A low level of EV values is observed for both NO2 and CO, which have mainly their origin in traffic. The largest EV values for these pollutants are achieved by Polish Niedzwiedz TCN21 classification in Krakow. This is the best classification for this city. EV values for SO2 in Gliwice (Upper Silesia) and EV values for PM10 in Wro- claw are greater than EV values in other cities. Each classifications has been ranked by number of EV > 0.3 and number of data series ranking high on the EV list (classifications which are first or second on the list EV for each data series). German, universally used manual classification the “Hess-Brezowski Großwetterlagen” HBGWL [12] is a good classification for most cities and pollutants. Classification calculated on the base of Principal Component Analysis PCACAC27 and Polish classification TCN21 also rank high. Apart from these classifications good results are obtained for modified (according COST 733 arrangements) Litynski classification LITTc [17], objective version of Lamb-Weather types [16] LWT2 and for three objectivized versions of the Hess and Brezowsky Großwetterlagen (GWTC26, OGWL, OGWLSLP) [14]. For the best classifications we drew the Box-Whiskers plots of pollution concentration (SO2 Gliwice - HBGWL, PM10 Wroclaw – PCACA, NO2 Krakow - TCN21 and CO Krakow – LWT2) for different classes (Fig. 4). Differences between various classes are visible, however big scattering is still observed. Nevertheless, the Analysis of Variance conducted for all data series points out that for most classifications the marginal means for different classes vary at a significant level p < 0.01.

Smog days in Krakow Because of irregularly situated emission sources (industrial, traffic, unorganized) research concerning relation between air pollution and boundary layer meteorology is difficult. The analysis based on air pollution measured only in one site might be distorted by the influence of isolated sources. Moreover, the analysis based only on days when air pollution norms are exceeded omits pollutants without norms. To solve these problems, the analysis based on exceeding 90-percentile (calculated for each month of the year and for each pollutant separately) is carried out. Days when 90-percentile is exceeded for at least 50% different air pollutants as well as for at least 50% sites inside agglomeration are chosen and called “smog days”. The Gini coefficients, calculated on the basis of air pollution data from Krakow, show that the best classifications are HBGWL, PCACAC27, LWT2 and TCN21 (Table 1 - the middle). They are the same CTCs as chosen by EV parameter calculated for each data series separately. This fact points out the consistency of both methods.

60 JOLANTA GODŁOWSKA, ANNA MONIKA TOMASZEWSKA TCN21 1 0.11 0.11 0.23 0.30 0.12 0.51 0.35 0.14 0.18 0.12 0.29 0.34 0.30 0.26 0.13 0.33 0.17 0.18 0.17 0.16 0.22 0.20 0.33 4 (5) 2 (0) places

nd POLISH PERRET 0

and 2

st 0.11 0.13 0.14 0.15 0.37 0.23 0.09 0.14 0.17 0.16 0.29 0.18 0.17 0.15 0.14 0.13 0.15 0.15 0.15 0.13 0.15 0.18 0.24 0 (0) 0 (1) OGWLSLP 0

0.11 0.14 0.15 0.09 0.47 0.28 0.16 0.18 0.18 0.13 0.34 0.22 0.22 0.21 0.17 0.15 0.20 0.15 0.26 0.21 0.21 0.22 0.33 2 (5) 0 (1) OGWL 0 2(1) 0.11 0.15 0.18 0.12 0.47 0.25 0.13 0.23 0.16 0.19 0.34 0.24 0.24 0.17 0.15 0.15 0.21 0.14 0.24 0.20 0.21 0.26 0.32 0 (0)

SUBJECTIVE HBGWL 1

0.24 0.21 0.16 0.51 0.29 0.16 0.32 0.17 0.21 0.14 0.38 0.18 0.27 0.28 0.24 0.12 0.13 0.28 0.35 0.31 0.25 0.28 0.48 0 (2) 5 (11) GWTC26 0

0.11 0.18 0.21 0.12 0.42 0.31 0.11 0.10 0.13 0.10 0.23 0.34 0.29 0.19 0.15 0.10 0.25 0.16 0.13 0.12 0.13 0.27 0.27 1 (2) 1 (0) LITTC 0

0.19 0.21 0.12 0.47 0.34 0.15 0.21 0.14 0.24 0.09 0.40 0.31 0.23 0.18 0.17 0.24 0.16 0.15 0.21 0.18 0.21 0.23 0.31 2 (2) 2 (1) LWT2 1 0.11 0.19 0.24 0.14 0.52 0.33 0.15 0.23 0.24 0.10 0.41 0.31 0.25 0.15 0.15 0.24 0.19 0.23 0.18 0.16 0.17 0.21 0.31 2 (4) 2 (1)

THRESHOLD WLKC28 0

0.2 0.11 0.11 0.11 0.19 0.22 0.08 0.48 0.22 0.19 0.23 0.13 0.40 0.22 0.20 0.15 0.19 0.09 0.17 0.15 0.16 0.19 0.31 2 (0) 0 (0) P27C27 0

0.16 0.19 0.12 0.41 0.34 0.14 0.12 0.17 0.20 0.12 0.35 0.28 0.23 0.15 0.12 0.23 0.14 0.17 0.15 0.15 0.16 0.26 0.29 1 (0) 1 (0) PCA TPCAC27 0

0.1 0.11 0.17 0.20 0.12 0.49 0.26 0.09 0.07 0.17 0.20 0.34 0.27 0.22 0.15 0.22 0.13 0.21 0.18 0.12 0.18 0.18 0.28 1 (0) 0 (0) LUNDC27 0

0.11 0.14 0.22 0.14 0.46 0.31 0.13 0.06 0.19 0.21 0.30 0.26 0.21 0.15 0.09 0.23 0.16 0.22 0.22 0.16 0.17 0.22 0.28 1 (0) 1 (0) ESLPC27 0

0.11 0.14 0.15 0.44 0.27 0.13 0.08 0.18 0.17 0.14 0.30 0.26 0.23 0.14 0.14 0.15 0.16 0.14 0.19 0.19 0.19 0.21 0.28 1 (0) 0 (0) LEADER KHC27 0

0.17 0.22 0.11 0.48 0.05 0.21 0.12 0.07 0.17 0.21 0.08 0.27 0.23 0.24 0.14 0.09 0.25 0.14 0.14 0.16 0.15 0.18 0.31 1 (0) 0 (0) CKMEANSC27 0

0.21 0.49 0.16 0.36 0.17 0.17 0.21 0.19 0.18 0.31 0.31 0.21 0.14 0.13 0.20 0.17 0.22 0.23 0.19 0.17 0.22 0.30 0.19 2 (0) 2 (5) NNWC27 0

0.11 0.11 0.42 0.19 0.09 0.08 0.14 0.19 0.11 0.07 0.23 0.27 0.20 0.08 0.14 0.13 0.10 0.10 0.07 0.12 0.13 0.18 0.16 0 (0) 0 (0) PCACAC27 1

0.23 0.52 0.41 0.13 0.18 0.12 0.24 0.24 0.16 0.44 0.33 0.28 0.17 0.17 0.23 0.19 0.19 0.21 0.16 0.21 0.27 0.34 0.21 2 (7) 2 (4) PETISCOC27 0

0.11 0.20 0.43 0.24 0.10 0.14 0.11 0.08 0.15 0.17 0.12 0.25 0.19 0.18 0.08 0.20 0.10 0.15 0.12 0.12 0.17 0.25 0.14 0 (0) 0 (0) CLUSTER SANDRAC27 0

0.11 0.11 0.20 0.46 0.37 0.16 0.16 0.21 0.10 0.34 0.30 0.24 0.15 0.14 0.19 0.18 0.22 0.18 0.18 0.19 0.21 0.27 0.20 1 (0) 2 (1) SANDRASC27 0 0.11 0.15 0.47 0.24 0.12 0.12 0.21 0.09 0.21 0.09 0.27 0.26 0.19 0.20 0.13 0.12 0.14 0.22 0.15 0.22 0.24 0.30 0.23 1 (1) 0 (0)

for each data series are printed in bold. EV greater than 0.3 as well Gini coefficients 0.5 are marked for each data series are printed in bold. EV STATIONS Łódź Łódź Gliwice Gliwice Kraków Gliwice Kraków Kraków Kraków Kraków Wrocław Wrocław Wrocław Warszawa Warszawa Convectin Convectin Warszawa Warszawa Ground Inv. Ground Inv. Elev.Inversion Elev.Inversion EV EV EV EV EV EV CO SO2 NO2 Mean Smog Days Gini Rank PM10 Air Pollution EV Rank Air Pollution EV Height of: GINI coef. METHOD in Krakow (in the middle) and EV of some boundary layer thermal structure characteristics (at the bottom) calculated for different CTCs. Correlations in 1 of some boundary layer thermal structure characteristics (at the bottom) calculated for different in Krakow (in the middle) and EV Smog Days Persistance: Table 1. Comparison of EV values for SO2, PM10, NO2 and CO measured in different Polish urban areas (at the top) as well Gini coefficients for smog days occurrence values for SO2, PM10, NO2 and CO measured in different 1. Comparison of EV Table Vertical Dispersion EV Rank Dispersion EV Vertical RELATIONS BETWEEN CIRCULATION AND WINTER AIR POLLUTION IN POLISH URBAN... 61

Fig. 4. Median, first and third quartile, maximum and minimum air pollution concentrations calculated for different classes PCACAC27 (PM10), HBGWL (SO2), LWT2 (CO) and TCN21 (NO2)

Boundary layer thermal structure in Krakow As it is widely known, low wind speed (horizontal dispersion) is an important factor in creating the smog episodes. The role of the thermal structure of the atmospheric boundary layer (vertical dispersion conditions) has not been so well-known, because the necessary data have been hardly available. The Boundary Layer in Krakow has been continuously monitored since the beginning of the last decade of the 20th century with use of vertical sounding sodar SAMOS-4C (the improved version of earlier applied sodars) [25, 26]. This remote-sensing device emits sound waves of the precisely chosen parameters. Some part of the acoustic signal is backscattered by the thermal heterogeneity of the atmosphere. The received part of the backscattered signal is analyzed and presented in the graphic form – the sodarogram. To determine the power of the relation between SODAR data and CTCs, the values of EV are used (Table 1 – the bottom). The biggest EV values for all classifications are observed for ground-based inversion height and the number of hours of elevated inversion presence. These parameters play a crucial role in air pollution concentration in Krakow [11, 10]. The best results (EV = 0.41 for GI height and EV = 0.33 for EI presence) are obtained for PCACAC27 classification. LWT2, and LITTC classifications also rank high. 62 JOLANTA GODŁOWSKA, ANNA MONIKA TOMASZEWSKA

DISCUSSION

Taking all the obtained results into consideration it seems that winter air pollution as well as vertical dispersion conditions are connected with circulation. The best CTCs in diver- sifying air pollution conditions in urban areas are based mainly on earlier used methods. Only PCACA classification has no connection with old, climatological research. The best results for separate air pollution data series and for smog episode days are obtained for Hess-Brezowski Großwetterlagen classification (HBGWL). Moreover, good results are obtained for principal component analysis based on PCACA, Polish Niedzwiedz TCN21, modified Polish Litynski LITTc, modified Lamb LWT2, and three modified HBGWL (GWTC26, OGWL, OGWLSLP) classifications. The same classifications apart from HBGWL are good for air pollution and dispersion. The power of this relation is not very big. Only for a few CTCs, EV values exceed 0.3, mainly for SO2 and PM10 data as well as for both ground inversion height and elevated inversion presence. The greatest values of Gini parameter do not exceed 0.6. The biggest SO2 marginal mean is obtained for 23rd class of HBGWL (HNFZ - high over Northern Sea - Fennoscandia, cyclonic). A similar behaviour of SO2 and PM10 for different cities is observed. Composite plot prepared by WG2 COST 733 shows that this class is characterised by the lack of precipitation, high level pressure and a very low temperature accompanied with meridional SE flow over the territory of Poland (Fig. 5). Both LWT2 and TCN21 classifications prefer two classes: Anti-cyclonic South- Westerly (6th class for both CTCs) and Cyclonic Southerly (22nd class for LWT2 and 15th class for TCN21). Moreover, high air pollution concentration is observed for both 14th (Unbiased South-Westerly) and 21st (Cyclonic South-Easterly) LWT2 classes and for 5th (Anti-cyclonic Southerly), 6th (Anti-cyclonic South-Westerly), and 16th (Cyclonic South-Westerly) TCN21 classes. Finally we can conclude that main urban air pollution problems arise when the flow with Southerly component is observed. Correlations between expected marginal means of air pollution and marginal means of both GI height and EI presence, calculated only for CTCs with EV (air pollution) > 0.3

Table 2. Correlations between marginal means of air pollution concentrations and marginal means of chosen SODAR parameters (calculated for different classes CTCs). Only CTCs with EV > 0.3 for air pollution data are shown. Correlations at significant level p < 0.05 are shown in bold

ELEVATED INVERSION GROUND INVERSION Method CTC’s PRESENCE HEIGHT CO PM10 SO2 NO2 CO PM10 SO2 NO2 SANDRAC27 0.67 0.82 0.75 0.55 -0.14 -0.45 -0.50 -0.17 CLUSTER PCACAC27 0.64 0.79 0.71 0.53 -0.29 -0.54 -0.69 -0.10 CKMEANSC27 0.56 0.76 0.68 0.52 -0.11 -0.39 -0.59 -0.02 LEADER LUNDC27 -0.18 -0.47 -0.53 -0.53 PCA P27C27 -0.15 -0.44 -0.52 -0.10 LWT2 0.77 0.90 0.82 0.58 -0.27 -0.40 -0.52 -0.10 THRESHOLD LITTC 0.66 0.87 0.74 0.55 -0.27 -0.55 -0.69 -0.12 GWTC26 -0.21 -0.35 -0.40 -0.40 POLISH TCN21 0.77 0.82 0.79 0.70 -0.16 -0.32 -0.42 -0.05 RELATIONS BETWEEN CIRCULATION AND WINTER AIR POLLUTION IN POLISH URBAN... 63

Fig. 5. Composite plot prepared by WG2 COST 733 (average situation in December, January and February in 1957-2002) for 23rd class of HBGWL (HNFZ class - high over Northern Sea - Fennoscandia, cyclonic)

(Table 2) suggest that air pollution in Krakow (especially for PM10 and SO2) depends on boundary layer conditions. Negative correlations between pollutants’ concentration and GI height confirm that shallow ground-based inversions in urban areas are connected with harmful air pollution conditions (Fig. 6). Additionally, positive correlations between pollutants’ concentrations and the number of hours with EI presence show that long duration of low-level elevated inversion causes the escalation of air pollution problems in urban areas (Fig. 7). 64 JOLANTA GODŁOWSKA, ANNA MONIKA TOMASZEWSKA

Fig 6. Comparison between marginal means of SO2 (top) or PM10 (bottom) and marginal means of ground- based inversion height calculated for LITTc classification on the basis of data from Krakow

Fig. 7. Comparison between marginal means of SO2 (top left), CO (top right), NO2 (bottom right) PM10 (bottom left) and marginal means of elevated inversion presence calculated for LWT2 (SO2 and PM10) and TCN21 (NO2 and CO) classifications on the basis of data from Krakow RELATIONS BETWEEN CIRCULATION AND WINTER AIR POLLUTION IN POLISH URBAN... 65

Acknowledgements This work was supported by the Ministry of Science and Higher Education under the Project No. 147/COS/2006/01.

REFERENCES

[1] AIRBASE – European Air Pollution database: http://air-climate.eionet.europa.eu/databases/EuroAirnet/ [2] Blazek Z., L. Ośródka, K. Korbek, M. Wojtylak: Concentration of suspended dust PM10 in Katowice agglomeration and Ostrava-Karvino agglomeration versus the meteorological conditions, Wiadomości IMGW, XXIII(XLIV), 3, 53-69 (2000). [3] Brandt S.: Statistical and computational methods in data analysis (Ed 3), Springer Verlag, New York 1997. [4] Catalogue of the circulation type classifications of COST733 Action – http://geo21.geo.uni-augsburg.de/ cost733wiki. [5] Comrie, A.C. and B. Yarnal,: Relationships between synoptic-scale atmospheric circulation and ozone concentrations in metropolitan Pittsburgh, Pennsylvania, Atmos. Environ., B, 26B, 3, 301-312 (1992). [6] Fisher B., S. Joffre, J. Kukkonen, M. Pringer, M. Rotach, M. Schatzmann (ed.): Meteorology Applied to Urban Air Pollution Problems – Final Report COST Action 715, Demetra Ltd Publishers, Bulgaria 2005. [7] Gini, Corrado: Measurement of Inequality and Incomes, The Economic Journal 31: 124–126 (1921). [8] Godłowska J., A.M. Tomaszewska: Imisja wybranych zanieczyszczeń powietrza a procesy w warstwie granicznej - analiza statystyczna, Wiadomości IMGW, XXIII(XLIV), 3, 47-52 (2000). [9] Godłowska J.: The particulate matter PM10 air pollution in Cracow. Wiadomości IMGW, XXVII(XLVIII), 1, 79-90 (2004). [10] Godłowska J.: Wpływ cyrkulacji atmosfery (wg Lityńskiego) na występowanie podwyższonych stężeń zanieczyszczeń i niekorzystnych warunków wentylacji w Krakowie, [w:] Ekstrema pogodowe w Polsce – ob- serwacje, pomiary; Instytut Meteorologii i Gospodarki Wodnej, Seria: Monografie, Warszawa2008, 82-98. [11] Godłowska J., A. M. Tomaszewska, M. Hajto: Związek wysokości imisji zanieczyszczeń w Krakowie z warunkami w miejskiej warstwie granicznej, określonymi na podstawie danych sodarowych, [in:] Klimat i bioklimat miast; Wyd. Uniw. Łódz., Łódź 2008, 455-465. [12] Hess P., H. Brezowsky: Katalog der Grosswetterlagen Europas, Ber. dt. Wetterdienst in der US-Zone, 33, 5, 1-39 (1952). [13] Huth R., C. Beck, A. Philipp, M. Demuzere, Z. Ustrnul, M. Cahynová, J. Kyselý, O.E. Tveito: Clas- sifications of atmospheric circulation patterns: recent advances and applications, [in:] L. Gimeno, R. Garcia Herrera, R.M. Trigo (Eds.): Trends and Directions in Climate Research, Ann. NY Acad. Sci., 1146, 105–152, Wiley-Blackwell (2008). [14] James P.M.: An objective classification for Hess and Brezowsky Grosswetterlagen over Europe. Theor. Appl. Climatology., 88, 17-42 (2007). [15] Kalkstein L.S., G. Tan, J. A. Skindlow: An Evaluation of Three Clustering Procedures for Use in Synoptic Climatological Classification,Journal of Climate and Applied Meteorology, 26, 717–730 (1987) [16] Lamb H.: British isles weather types and a register of the daily sequence of circulation paterns, 1861- 1971. Geophysical Memoirs, 116, 1-85 (1972). [17] Lityński J., Liczbowa klasyfikacja typów cyrkulacji i typów pogody dla Polski. Prace Państwowego Insty- tutu Hydrologiczno-Meteorologicznego, 97, Warszawa 1969. [18] Lorenz M. O.: Methods of measuring the concentration of wealth, Publications of the American Statistical Association, 9, 209–219 (1905). [19] Niedźwiedź T.: Kalendarz sytuacji synoptycznych dla dorzecza górnej Wisły (1951-1985), Zesz. Nauk. UJ, Prace Geogr., 71, 37-86 (1988). [20] Niedźwiedź T.: Calendar of Circulation Types for Southern Poland 1873-2009. Computer File Available at Department of Climatology, University of Silesia. http://klimat.wnoz.us.edu.pl (2010). [21] Niedźwiedź T., Z. Ustrnul: Influence of synoptic situations on the occurrence of weather types favorable to concentration or dispersion o fair pollution above the Upper Silesia Industrial Region, Wiadomości IMGW, 12 (1-2), 31-39 (1989). (in Polish, summary in English). [22] Niedźwiedź T., Z. Olecki: Influence of synoptic situations on the air pollution in Krakow. Zeszyty Nau- kowe UJ, Prace Geograficzne, 96, 55-67 (1994). (in Polish, summary in English). [23] Philipp A., J. Bártholy, C. Beck, M. Erpicum, P. Esteban, X. Fettweis, R. Huth, P.M. James, S. Jourdain, F. Kreienkamp, T. Krennert, S. Lykoudis, S. Michalides, K. Pianko, P. Post, D. Rassilla Álvarez, A. Spekat, 66 JOLANTA GODŁOWSKA, ANNA MONIKA TOMASZEWSKA

F.S. Tymvios, in press. COST733CAT – a database of weather and circulation type classifications. Phys- ics and Chemistry of the Earth (2010). [24] Piringer M., Joffre S. (ed.): The Urban Surface Energy Budget and Mixing Height in European Cities: Data, Models and Challenges for Urban Meteorology and Air Quality. Final Raport of Working Group 2 of COST-715 Action. Demetra Ltd Publishers, Bulgaria (2005). [25] Walczewski J.: Application of Sodar in Urban Air-Quality Monitoring Systems. Acoustic Remote Sensing Applications. S.P.Singal, Springer Verb. - Narosa Publ. H., New York-New Delhi, 385-394 (1997). [26] Walczewski J.: Application of Acoustic Remote Sensing in Air Quality Monitoring. Proc. Fourth Interna- tional Symposium and Exhibition on Environmental Contamination in Central and Eastern Europe War- saw’98, September 1998, Warsaw, Poland. CD-ROM, Florida State University, Institute for International Cooperative Environmental Research (1999). [27] Yarnal, B.: Synoptic Climatology in Environmental Analysis. Belhaven Press. London, UK. (1993).

Received: March 8, 2010; accepted: July 15, 2010.

ZWIĄZEK POMIĘDZY CYRKULACJĄ ATMOSFERY I ZANIECZYSZCZENIAMI POWIETRZA W OBSZARACH ZURBANIZOWANYCH NA TERENIE POLSKI

Określono ocenę przydatności różnych klasyfikacji typów cyrkulacji (CTCs) do różnicowania stężeń SO2, NO2, PM10 i CO w zimie w polskich obszarach zurbanizowanych. Analiza bazuje na 15 nowych klasyfikacjach obliczonych w ramach Akcji COST 733 oraz pięciu powszechnie stosowanych klasyfikacjach histo- rycznych. Porównano i przetestowano trzy metody oceny jakości klasyfikacji: EV - wyjaśnianej wariancji, WSD - wewnątrzklasowego odchylenia standardowego i metodę Giniego. Ostatecznie ocenę jakości CTCs dla zróżnicowania dobowych wartości stężeń zanieczyszczeń a także warunków dyspersji pionowej przeprowadzo- no opierając się na metodzie EV, zaś ocenę przydatności klasyfikacji do prognozowania wystąpienia epizodów smogowych w Krakowie wykonano stosując metodę Giniego. Zaprezentowano także metodologię wyboru dni z epizodami smogowymi opartą na wartości 90 percentyla stężeń. Najlepsze rezultaty, dla pojedynczych serii danych a także dla epizodów smogowych, otrzymano dla klasyfikacji Hess-Berezowski (HBGWL). Ponadto dobre rezultaty uzyskano dla opartej na analizie składowych głównych klasyfikacji PCACA, klasyfikacji Niedźwiedzia, zmodyfikowanych w ramach Akcji COST733 klasyfikacjach Lityńskiego LITTc, Lamba LWT2 i HBGWL (GWTC26, OGWL, OGWLSLP). Te same klasyfi- kacje, z wyjątkiem HBGWL i jej modyfikacji, są dobre dla różnicowania danych sodarowych. Dla najlepszych klasyfikacji różnice średnich stężeń dla różnych klas są widoczne, jakkolwiek duży rozrzut wewnątrz klas jest ciągle obserwowany. Najwyższe stężenia zanieczyszczeń obserwowane są dla adwekcji mas powietrznych z kierunków SE, S i SW. Korelacje pomiędzy wielkością imisji i danymi sodarowymi, obliczone dla średnich brzegowych uzyskanych dla różnych klas, potwierdziły negatywną rolę zarówno niskiej wysokości przygrun- towej inwersji, jak i długotrwałego utrzymywania się niskich inwersji wzniesionych. 67

ARCHIVES OF ENVIRONMENTAL PROTECTION vol. 36 no. 4 pp. 67 - 80 2010 PL ISSN 0324-8461

VENTILATION CONTROL BASED ON THE CO2 AND AEROSOL CONCENTRATION AND THE PERCEIVED AIR QUALITY MEASUREMENTS – A CASE STUDY

Bernard Połednik*, Marzenna Dudzińska

Faculty of Environmental Engineering, Lublin University of Technology, Nadbystrzycka str. 40B, 20-618 Lublin, Poland * Corresponding author e-mail: [email protected]

Keywords: Aerosol particle number concentration, CO2 concentration, perceived air quality, acceptability of indoor air quality, ventilation control strategy.

Abstract: One of the concepts of the ventilation rate control in buildings with dense and unpredictable occupan- cies is based on the CO2 measurements. There are many limitations regarding the validity of CO2 measurement inputs as suitable to the ventilation rate control. Verifying research has been conducted in an air-conditioned auditorium, in the real conditions at altered ventilation air thermal parameters and variable occupancy. The CO2 and the number concentrations of the fine and coarse aerosol particles (> 0.3 μm) and bioaerosol particles (bacteria and staphylococci) as well as the indoor air thermal parameters were measured in the individual sectors of the occupied area. The sensory assessments and instrumental determinations of the acceptability of indoor air quality (ACC) were also performed. The ventilation control strategy based, apart from the CO2 measurements, on the continuous monitoring of the perceived air quality (PAQ) in the auditorium sectors has been suggested. The PAQ monitoring could be accomplished by aerosol concentration measurements and the ACC instrumental determinations. This strategy should ensure a desired PAQ in sectors which benefit the occupants’ comfort, health and productivity as well as energy savings not only in the case of its implementation in the considered auditorium.

INTRODUCTION

A positive assessment of the indoor air quality is necessary for the occupants to feel comfortable in indoor environments. Unfortunately, despite the efforts of the designers and contractors, the number of buildings in which the indoor air quality is negatively assessed is constantly increasing [30, 32]. It concerns a wide spectrum of buildings including residential homes, offices, recreation facilities and schools [7, 17, 20, 22]. The majority of the objections are connected with the indoor air pollutants perceived by the sense of smell. Aerosol particles, both organic, nonorganic and biological, are those contaminants, which significantly affect the perceived indoor air quality (PAQ) [32, 8, 34]. Room occupants and their activities are approved to be one of the main sources of such widely apprehended indoor generated aerosols [1, 4, 12, 16, 18, 19]. Heating, ventilation and air-conditioning systems are also of significance. It is due to the fact, that the concentration level of the aerosol particles is dependent on the indoor air thermal parameters and 68 Bernard Połednik, Marzenna Dudzińska the correlated infiltration-exfiltration processes in the building [2, 13]. Air-conditioning systems themselves could be a source of aerosols (bioaerosols) [14, 35]. Taking into account the room occupants, their presence, apart from generating aerosols, is connected with bioeffluent emissions and CO2 concentration changes indoors. Since people produce a predictable amount of CO2 as a result of respiration, the CO2 level could be treated as an indirect measure of room occupancy [9, 15, 21]. The measurements of the CO2 concentration in the indoor air are often used to monitor the performance of the ventilation systems. These measurements are also crucial to adjust the outdoor airflow rate as regards the con- cept of the CO2-based demand control ventilation (DCV) [9, 21]. However, according to the literature data, the maintenance of the CO2 set point levels does not always guarantee an acceptable indoor air quality [9, 23, 36]. The determination of the influence of the number of people, their activity and the duration of their stay on the aerosol particle and CO2 (bioeffluents) concentrations is important when proper PAQ has to be assured in buildings with variable occupancy. It is essential as the available methods of controlling the ventilation rate do not always ensure the desired PAQ. The implementation of the CO2-based DCV strategy which takes into account only CO2 levels and overrides other important indoor air pollutants and relevant relations may carry such risks. Much data on the indoor air thermal parameters, pollution levels, the indoor air quality and human sensation in relation to the occupation can be found in the literature. It comes from the studies of different indoor environments with different types of occupancy. Different methods of ventilation control are also presented. Only few of them present the idea of measuring the air parameters deciding about the PAQ in every occupied zone of the indoor environment and to control the ventilation on the basis of such measurements. However, there are no works which describe the ventilation control on the basis of a direct instrumental determination of the PAQ indoors. The present study aims to verify on the example of an auditorium the relations between the indoor air parameters and the occupancy and to indicate that in addition to the CO2 measurements, zonal monitoring of aerosols and continuous instrumental determinations of the PAQ could be important while implementing an effective ventilation control in indoor environments with variable occupancy.

METHODS

The experiments were carried out in a new air-conditioned auditorium at the Lublin Uni- versity of Technology (LUT). The auditorium with seats for 186 people is located on the first floor of the LUT building and has a floor area of approximately 300 m2 and a volume of approximately 1200 m3 (average dimensions – 20x15x4 m). The auditorium has 4 double pane air tight windows (2.6 x 3 m), which overlook towards the south-west and 4 windows of the same size which overlooksthe north-east. The mounted window-blinds diminish the inflow of the solar rays into the auditorium and prevent the inside from ex- cessive heating. All surfaces in the auditorium are finished with low-polluting materials. Apart from the occupants, the audio-visual and electronic equipment (2 projectors, speak- ers, a visualizer and a computer) and also the air-conditioning system constitute pollution sources which exacerbate the indoor air quality in the auditorium. The system filters are replaced every three months and were changed one week prior to the measurements. The VENTILATION CONTROL BASED ON THE CO2 AND AEROSOL CONCENTRATION... 69 auditorium is ventilated with a total supply of 1.6 m3/s of conditioned air (50% recycled). The ventilation air of set thermal parameters is delivered by 12 steadily emplaced inlets. The ventilation air temperature was set from 18 to 26oC and the relative humidity from 30 to 50%. The given thermal parameters of the supplied air were maintained at the same level throughout the each whole measurement day. When the auditorium was not occupied (in the evenings and at night, i.e. between 8 p.m. and 8 a.m. the air-conditioning system was operating in the Standby Mode and did not deliver the ventilation air when the indoor air temperature ranged ±2oC around the set point. The plan view of the auditorium with occupation area and marked measurement sectors is shown in Figure 1.

Fig. 1. Scheme of the auditorium with a non occupied area A2 and occupied area A1 with the measurement

sectors: 1,2,3,4,5,6,7,8,9 – temperature, relative humidity and air velocity measurements; 1,3,5,7,9 – CO2 measurements; 4,5,6 – radiation temperature measurements; 5 - aerosol and bioaerosol measurements.

The occupation area in the auditorium was divided into two parts: lower A1 with seats for 116 students and upper A2 with seats for 70 students. During the experiments, the students took seats only in the area A1. An approximately equal number of students (from 3 to 10 depending on the overall number of students present) was in each measurement sector. The students were not allowed to change their seats during the lecture. The auditorium occupancy usually changed every second hour according to a traditional 70 Bernard Połednik, Marzenna Dudzińska academic system of 2-h lectures. Separate lectures were divided by 15-minute breaks during which the students moved in and out of the room. The auditorium was irregularly occupied from Monday through Friday, usually from 8 a.m. to 8 p.m. according to the schedule of the LUT courses. Nine measurement sectors were allocated within the area A1. Sensors for measuring the air thermal parameters and CO2 concentrations were located in the middle of each sector, approximately at the height of the students’ heads (1.2 m). The measurements of the indoor air thermal parameters and the CO2 concentrations were performed using a multisensor system Almemo 5690-2M (Ahlborn, Germany). Aerosol and bioaerosol concentrations were measured in the middle of the area A1 (in sector 5). Aerosol particle number concentrations were determined by means of a laser particle counter ROYCO 243A, equipped with iso-Diluter D50 (Pacific Scientific Instruments, USA). The samples were collected at 60-second intervals with a 15-minute delay time. The counter categorized the collected particles into four size ranges: 0.3-0.5, 0.5–5, 5–10 and >10 μm. Signal to noise ratio was 1.6:1 at maximum sensitivity (0.3 μm). Bacteria concentrations were periodically determined before, during, and after lectures according to the Standard [27]. Air sampling was performed using settling plates [28]. Petri dishes containing a solid nutrient medium (Nutrient Agar for a total number of bacteria and Chapman Agar for staphylococci) were left open to air for 15 minutes. Microbes carried by inert particles fell onto the surface of the nutrient and after 48h of in- cubation at 37°C the grown colonies were counted. The results were presented as colony forming units per cubic meter of air (CFU/m3). The acceptability of the indoor air quality (ACC) which is considered as an indicator of the PAQ was determined on the basis of the measurements of the indoor air thermal parameters and with the application of the Weber-Fechner psychophysical low and the patent pending dependencies [5, 6]. The ACC was determined from the following formulae: ACC = a lnt + b ln RH + c (1) where t is the indoor air temperature [°C], RH is relative humidity [-] and a, b, c are empirical constants, which for conditions in the examined auditorium were assumed as follows: a = -2.1508, b = -0.0727 and c = 7.0404. These values of the ACC were determined in the auditorium sectors continuously (in every 10 seconds). The ACC for the whole auditorium was also continuously determined as well as the average ACC values calculated for 15-minute intervals. The ACC was also sensory assessed by the students participating in the experiments. Immediately upon entering the auditorium and at the end of the lecture, they were asked to fill out a questionnaire, which contained the vertical continuous scale for rating the air acceptability [33]. The scale was coded as follows: 1 – clearly acceptable, 0 – just acceptable/ just not acceptable, -1 – clearly not acceptable. The mean ACC of the subjects’ votes in the given sector were used to describe the PAQ in that sector. The average values of the sensory assessed ACC for all sectors at the beginning and at the end of the lectures were calculated. Only these average values of the sensory assessed ACC were considered for this study. The subjects were untrained LUT students; male and female aged between 21 and 24, either smoking or non-smoking, having lectures in the auditorium according to the LUT course schedule. The assumption was that the students were to feel thermally comfortable; therefore, prior to each lecture they were informed about the experimental VENTILATION CONTROL BASED ON THE CO2 AND AEROSOL CONCENTRATION... 71 procedures and they were allowed to modify their clothing to feel comfortable. How- ever, the students were not informed about the experimental conditions as regards the air quality and the air thermal conditions experienced during a given lecture. They were not examined medically and were not questioned about their health condition, e.g. chronic diseases, allergies or past medical conditions. The experiments were conducted during several days in the springtime 2008.

RESULTS AND DISCUSSION

The first objective consisted in verifying how the aerosol, bioaerosol and CO2 concentrations as well as the assessed and instrumentally determined PAQ change with the altered indoor air thermal parameters and variable occupancy in the examined auditorium. The graphs in Figure 2 present the time series of the concentration of the measured aerosol size fractions, the indoor air temperature and relative humidity in the auditorium. Finer aerosol particle (0.3–0.5 μm and 0.5–5 μm) number concentrations changed almost in accordance with the changes of the indoor air thermal parameters, namely they generally increased with the increase of the relative humidity and the decrease of the indoor air temperature. The reason could be the shift of the finer particle size distributions associated with water uptake or release when these particles are exposed to changing humidity conditions [25]. The opposite tendency was demonstrated by coarser aerosol particle (5–10 μm and >10 μm) number concentrations. In this case a certain increase of the concentrations was observed with the increase of the air temperature (decrease of the relative humidity). Such changes could be relevant to interdependence between coarser particle concentration, activity of occupants and indoor air temperature [4, 15]. The calculated correlation coefficients and levels of significance are presented inTable 1.

Table 1. Correlation coefficients between the temperature (T), relative humidity (RH), number of students

(SN), CO2 concentration (CO2), bacteria concentration (CB), staphylococci concentration (CS), acceptability of the air quality (ACC) and concentration of the measured particle size fractions

Particle size fraction [μm] 0.3 – 0.5 0.5 – 5 5 – 10 > 10 > 5 T -0.196*** -0.266*** 0.213*** 0.209*** 0.217*** RH 0.216*** 0.168*** 0.043 0.041 0.043 SN 0.034 -0.001 0.573*** 0.565*** 0.634***

CO2 -0.045 -0.062 0.769*** 0.728*** 0.775***

CB 0.240 0.222 0.611** 0.499* 0.599**

CS 0.305 0.280 0.546** 0.461* 0.561** ACC 0.167*** 0.225*** -0.209*** -0.208*** -0.214*** * P < 0.05; ** P < 0.01; *** P < 0.001

It must be noted that the indoor air temperature in the occupied auditorium evidently increased above the set values of the input air temperature. In turn, the sole presence of the students favored coarser aerosol particles. It is confirmed by the graphs in Figure 3, which show the changes of the concentration of the measured particle fractions and the variations of the number of students present in the auditorium. 72 Bernard Połednik, Marzenna Dudzińska

Fig. 2. Time series of aerosol particle number concentration, the indoor air temperature and relative humidity in the auditorium. VENTILATION CONTROL BASED ON THE CO2 AND AEROSOL CONCENTRATION... 73

Fig. 3. The changes of the aerosol particle concentration and the number of students in the auditorium. 74 Bernard Połednik, Marzenna Dudzińska

It can be seen that the coarser aerosol particles were mainly measured when the students were present in the auditorium. The statistically significant positive correlations were found between the number of students and concentrations of these particles (Table 1). Braniš et al., Ferro et al. [4, 12] and other researchers reported analogous results as far as the trend is concerned. However, such high correlation was not obtained in the conducted measurements (r ≈ 0.60). It could result from the fact that coarser particles were registered generally at the beginning of the lectures when the students were taking seats and in the end when they were leaving the auditorium. At that time, there was an increased air turbulence which resulted in the resuspension of the particles deposited on the indoor surfaces as well as in more intensive emissions from the handled materials [14, 18, 19]. During the lectures or when no lectures were conducted, either small amounts of coarser particles were registered or they were not registered at all. According to the literature data, both aerosol particles [3, 31] and interdependent indoor air thermal parameters [10, 11] influence the PAQ. The above is also verified by the obtained results in the examined auditorium. The graphs in Figure 4a,b present the changes of the PAQ which could be equivalently displayed by the changes of the instrumentally determined ACC and the concentrations of the fine (0.3–5 μm) and coarse (> 5 μm) aerosol particle fractions. It can be clearly seen from the graphs that the changes of the ACC in the auditorium trace the variations of the fine aerosol particle concentrations. In the case of coarse aerosol particles, this relation is reversed (the ACC decreases when the coarse particle concentrations increase). Indoor air thermal parameter changes should not be disregarded because, as it has been mentioned before, they influence the aerosol concentrations and the ACC. According to the study of Fang et al. [10] the ACC is in inverse to the air thermodynamic properties – the ACC linearly decreases with the increase of the air enthalpy. However, taking into consideration the fact that the main cause of the coarse particles increase is the presence of the students in the auditorium, one can assume that they are responsible for the deterioration of the PAQ. The graphs in Figure 4c which present the time series of ACC and the number of present students may confirm it. The variable occupancy in the considered auditorium can, in turn, be closely tracked by the CO2 concentration changes. This is due to the fact that the students who perform a specific activity in the auditorium exhale CO2 at a predictable level. The observed changes of the CO2 concentration and simultaneous changes of the PAQ shown in Figure 4d are consistent with the variations of the number of the present students. The changes of the PAQ are displayed by the according changes of both the instrumentally determined and sensory assessed ACC values [5]. These two ways of the ACC determination in the sectors of the occupied area in the examined auditorium gave similar results. The mean differences were within the error of the ACC sensory assessments which amounted to 0.15. The graphs in Figure 5a,b,c,d il- lustrate respectively the relationships between the CO2, the coarse aerosol particle and the bacteria (staphylococci) mean concentrations as well as the instrumentally determined/ sensory assessed ACC and the number of students present in the auditorium. The values of the indoor air parameters constitute the averages of their measurements in 15 minute time intervals for the whole auditorium. Only the values of the sensory assessed ACC come from the assessments performed at the beginning and at the end of the lectures. In the case of microorganisms, the data concerning their concentration levels was used only as an additional confirmation of the influence of the number of students on the PAQ in the examined auditorium. The determined bacteria (staphylococci) VENTILATION CONTROL BASED ON THE CO2 AND AEROSOL CONCENTRATION... 75

Fig. 4. The changes of the acceptability of air quality, the concentration of fine and coarse aerosol particles,

the number of students and the CO2 concentration in the auditorium. 76 Bernard Połednik, Marzenna Dudzińska

Fig. 5. The dependence of the CO2, the coarse aerosol particle, the bacteria concentrations and the acceptability of air quality on the number of students in the auditorium. VENTILATION CONTROL BASED ON THE CO2 AND AEROSOL CONCENTRATION... 77 concentrations were not significantly different from the data presented in the literature [3, 34]. Linear regressions are calculated for the individual relations presented in Figure 5 and the 95% prediction intervals are marked. It follows from the results that despite of the operation of the air-conditioning system and altered indoor thermal conditions in the auditorium, the increase of the number of students results in the increase of the CO2, the coarse aerosol particle and the determined microorganism concentrations and at the same time in the deterioration of the PAQ. The correlation coefficients and levels of significance are presented in Table 2.

Table 2. Correlation coefficients between the CO2 concentration (CO2), bacteria concentration (CB),

staphylococci concentration (CS), air acceptability (ACC) and the temperature (T), relative humidity (RH), number of students (SN)

T RH SN

CO2 0.380*** 0.060 0.832***

CB -0.047 0.355 0.502*

CS -0.276 0.511* 0.557** ACC -0.983*** -0.096 -0.319*** * P < 0.05; ** P < 0.01; *** P < 0.001

The best association was obtained between the CO2 concentration and the number of students. In this case the correlation coefficient r = 0.832; the significance level p < 0.001. The statistically significant relations between the coarse aerosol particle and the microorganism concentrations as well as the ACC and the number of students occur despite the substantial scattering of the experimental points around the best-fit lines. It can be assumed that all the presented relations could be more accurate if, which is planned in future studies, results from more experiments with a larger range of number of students present in the auditorium (to its maximum capacity of 186 students) were considered in determining those relations and the altered indoor conditions where taken into account. The next aim of the paper was to present a suggestion of improving the control of the existing auditorium ventilation system and the indication of the further direction of the research. The performed measurements showed a dependency between the number of students present in the auditorium and the parameters determining the PAQ. Pursuant to the literature data [9, 36] and the performed measurements, the CO2 level may be treated as an indirect measure of the occupancy. Therefore, the CO2 monitoring seems to be a ra- tional basis for controlling the ventilation in the auditorium. However, various performed research demonstrated that CO2-based demand control ventilation strategy (DCV) may save energy but it does not always ensure the desired PAQ [9, 21, 36]. Energy savings come from controlling the ventilation on the basis of the actual occupancy indicated by the indoor CO2 level. The PAQ is, in turn, affected by the type of pollution sources in indoor spaces. Under certain conditions, CO2-based DCV strategy may be applied in rooms in which people are the main contaminant source. If the contaminants associated with the building are dominant, this DCV strategy may lead to an unsatisfactory PAQ. Moreover, it may be effective only in steady state conditions [9, 36]. In the case of the examined auditorium, there are constant variations of the number and activities of the students. 78 Bernard Połednik, Marzenna Dudzińska

Considerable variations of the activities occurring at the beginning and at the end of the classes (the students entering and leaving the auditorium) result in noticeable changes in the PAQ. Those PAQ changes are not reflected in the changes of the CO2 content and would not be taken into consideration in the discussed conventional DCV strategy. Simi- larly, the changes of the PAQ in sectors differentiated in terms of the number and activity of the students and also in terms of the local thermal loads would be considered only to a small extent or would not be considered at all. In consequence, an identical amount of air with the same physicochemical parameters would be supplied to all ventilated sectors in the auditorium. It would result in an over-ventilation in some sectors and an under ventilation in others [24, 37]. The differences in PAQ which arise in those sectors may be leveled with the use of the DCV strategy for multi-zone ventilation [9, 29]. This strategy could supply the varied amount of air of appropriate parameters according to the zonal needs. However, for such ventilation control strategy to be effective in the considered auditorium, it would have to ensure thermal comfort and PAQ in the individual sectors while maintaining ventilation requirements [26] and minimizing energy consumption.

Considering PAQ, this strategy would have to be based not only on the CO2 monitoring but also on the results of the indoor air other significant parameter measurements and relations which influence the PAQ in the sectors. The local variations in the number and activities of students could be taken into account by aerosol concentration measurements and continuous instrumental determinations of the ACC in the auditorium sectors. If such ventilation control strategy was accomplished it should ensure the comfort, health and productivity of the students. It should also bring financial benefits related to a more economical energy use as a result of improving the air-conditioning system overall performance through, e.g. the elimination of the over-ventilation and optimizing the thermal parameters of the supplied air to the sectors. Further, more detailed studies are necessary in order to implement this ventilation control strategy to the examined auditorium and to other rooms of similar type and purpose.

CONCLUSIONS

The coarse aerosol, bioaerosol particle and the CO2 concentrations and also the PAQ in the air-conditioned auditorium with the altered indoor air thermal parameters are significantly dependant on the presence and activities of the students. The effective ventilation control strategy should take into consideration, apart from the CO2 level, other significant parameters of the indoor air and relations which influence the PAQ in the occupied area. In order to take into account the local variations in the number and activities of students, continuous monitoring of the PAQ in the individual sectors has been suggested. The PAQ continuous monitoring could be accomplished by aerosol concentration measurements and the ACC instrumental determinations. The practical implementation of the ventilation control strategy based on continuous instrumental determinations of the PAQ in the sectors should improve the occupants’ comfort, health and productivity. It should also bring financial benefits connected with a more economical energy use.

Acknowledgements This research has been supported by grant No. 4084/T02/2007/32 from the Ministry of Science and Higher Education, Poland. VENTILATION CONTROL BASED ON THE CO2 AND AEROSOL CONCENTRATION... 79

REFERENCES

[1] Afshari A., U. Matson, L.E. Ekberg: Characterization of indoor sources of fine and ultrafine particles: a study conducted in a full-scale chamber, Indoor Air, 15, 2, 141-150 (2005). [2] ANSI/ASHRAE 62-2007: Ventilation for Acceptable Indoor Air Quality, Atlanta 2007. [3] Blondeau P., V. Iordache, O. Poupard, D. Genin, F. Allard: Relationship between outdoor and indoor air quality in eight French schools, Indoor Air, 15, 1, 2-12 (2005). [4] Braniš M., P. Řezáčová, M. Domasová: The effect of outdoor air and indoor human activity on mass

concentrations of PM10, PM2.5 and PM1 in a classroom, Environmental Research, 99, 2, 143-149 (2005). [5] Burek R., B. Połednik, A. Raczkowski: Study of the relationship between the perceived air quality and the specific enthalpy of air polluted by people, Archives of Environmental Protection, 2, 21-26 (2006). [6] Burek R., B. Polednik, A. Raczkowski: Method of the perceived air quality and sensory pollution measurement, Polish Patent Office Bulletin,1, 862, 23 (2007). [7] Daisey J.M., W.J. Angell, M.G. Apte: Indoor air quality, ventilation and health symptoms in schools: An analysis of existing information, Indoor Air, 13, 53-64 (2003). [8] Davidson C.I., R.F. Phalen, P.A. Solomon Airborne particulate matter and human health: A review. Aero- sol Science and Technology, 39, 737 (2005).

[9] Dougan D.S., L. Damlano: CO2-Based Demand Control Ventilation. Do risks Outweigh Potential Re- wards?, ASHRAE Journal, 46, 47-54 (2004). [10] Fang L., G. Clausen, P.O. Fanger: Impact of temperature and humidity on the perception of indoor air quality, Indoor Air, 8, 80-90 (1998). [11] Fang L., G. Clausen, P.O. Fanger: Impact of temperature and humidity on perception of indoor air quality during immediate and longer whole-body exposures, Indoor Air, 8, 276-284 (1998). [12] Ferro A.R., R.J. Kopperud, L.M. Hildemann: Source strengths for indoor human activities that resuspend particulate matter, Environ. Sci. Technol., 38, 1759-1764 (2004). [13] Franck U., O. Herbrth, B. Wehner, A. Wiedensohler, M. Manjarrez: How do the indoor size distributions of airborne submicron and ultrafine particles in the absence of significant indoor sources depend on outdoor distributions?, Indoor Air, 13, 174 (2003). [14] Guo H., L. Morawska, C.He, C. Gilbert D.: Impact of ventilation scenario on air exchange rates and on indoor particle number concentrations in an air-conditioned classroom, Atmospheric Environment, 42, 4, 757-768 (2008). [15] Heudorf U., V. Neitzert, J. Spark: Particulate matter and carbon dioxide in classrooms - The impact of cleaning and ventilation, International Journal of Hygiene and Environmental Health, 212, 45 (2009). [16] Holmberg S., Q. Chen: Air flow and particle control with different ventilation systems in a classroom, Indoor Air, 13, 2, 200-204 (2003). [17] Hoskins J.A.: Health effects due to indoor air pollution. Indoor and Built Environment 12, 427, (2003). [18] Hussein T., T. Glytsos, J. Ondráček, P. Dohányosová, V. Ždímal, K. Hämeri, M. Lazaridis, J. Smolik, M. Kulmala: Particle size characterization and emission rates during indoor activities in a house, Atmos- pheric Environment, 40, 4285-4307 (2006). [19] Jamriska M., L. Morawska, D.S. Ensor: Control strategies for sub-micrometer particles indoors: model study of air filtration and ventilation, Indoor Air, 13, 2, 96-105 (2003). [20] Jo W.K., Y.J. Seo: Indoor and outdoor bioaerosol levels at recreation facilities, elementary schools, and homes, Chemosphere, 61, 11, 1570-79 (2005).

[21] Lawrence T.M.: Selecting CO2 criteria for Outdoor Air Monitoring, ASHRAE Journal, 46, 18-27 (2004). [22] Lee S.C., H. Guo, W.M. Li, L.Y. Chan: Inter-comparison of air pollutant concentrations in different indoor environments in Hong Kong. Atmospheric Environment, 36, 1929 (2002). [23] Mossolly M., K. Ghali, K. Ghaddar: Optimal control strategy for a multi-zone air conditioning system using a genetic algorithm, Energy, 34, 1, 58-66 (2009).

[24] Nassif N., S. Kajl, R. Sabourin: Ventilation control strategy using the supply CO2 concentration setpoint, HVAC&R Res. 11, 2, 239- 262 (2005). [25] Nazaroff W.W.: Indoor particle dynamics, Indoor Air, 14, 175-183 (2004). [26] PN-83/B-03430/Az 3:2000: Wentylacja w budynkach mieszkalnych, zamieszkania zbiorowego i użyte- czności publicznej. Wymagania. [27] PN-89/Z-04111/02: Ochrona czystości powietrza. Badania mikrobiologiczne. Oznaczanie liczby bakterii w powietrzu atmosferycznym (imisja) przy pobieraniu próbek metodą aspiracyjną i sedymentacyjną. [28] PN-89/Z-04008/08: Ochrona czystości powietrza. Pobieranie próbek. Pobieranie próbek powietrza at- mosferycznego (imisja) do badań mikrobiologicznych metodą aspiracyjną i sedymentacyjną. 80 Bernard Połednik, Marzenna Dudzińska

[29] Polednik B., M. Dudzinska, A. Raczkowski: The influence of occupation on aerosol and CO2 concentration in classroom. Indoor Air 2008, The 11th International Conference on Indoor Air Quality and Climate, Paper ID: 546, Copenhagen 2008. [30] Report No 22, EUR 19529EN, Risk assessment in relation to indoor air quality, 2000. [31] Schneider T., J. Sundell, W. Bischof, M. Bohgard, J.W. Cherrie, P.A. Clausen, S. Dreborg, J. Kildesø, S.K. Kjærgaard, M. Løvik, P. Pasanen, K. Skyberg: Airborne particles in the indoor environment. A European interdisciplinary review of scientific evidence on associations between exposure to particles in buildings and health effects, ‘EUROPART’, Indoor Air, 13, 1, 38-48 (2003). [32] US EPA Technical report 600/P-95/001aF, Air quality criteria for particulate matter, 1996. [33] Wargocki P.: Sensory pollution sources in buildings, Indoor Air, 14, 82-91 (2004). [34] WHO regional publications, European Series No. 31 1999. Indoor air quality: biological contaminants, Copenhagen 1999. [35] Wong L.T, K.W. Mui, W.Y. Chan: An energy impact assessment of ventilation for indoor airborne bacteria exposure risk in air-conditioned offices, Building and Environment, 43, 1939 (2008). [36] Xu X.H, S.W. Wang: An Adaptive Demand-controlled Ventilation Strategy with Zone Temperature Reset for Multi-zone Air-conditioning Systems, Indoor and Built Environment, 16, 5, 426-437 (2007). [37] Xu X.H., S.W. Wang, , Z.W. Sun, F. Xiao: A model-based optimal ventilation control strategy of multi- zone VAV air-conditioning systems, Applied Thermal Engineering, 29, 91-104 (2009).

Received: March 11, 2010; accepted: October 19, 2010.

Sterowanie wentylacją oparte na pomiarach CO2, koncentracji aerozoli i wyczuwalnej jakości powietrza – studium przypadku

Z koncepcją sterowania wentylacją w pomieszczeniach ze zmienną liczbą użytkowników, opartą na monitorow- aniu CO2, związanych jest szereg ograniczeń. Weryfikacyjne badania zostały przeprowadzone w rzeczywistych warunkach klimatyzowanej auli, przy zmienianych parametrach termicznych powietrza wentylacyjnego i zmiennej liczbie obecnych studentów. W strefie przebywania ludzi mierzone było stężenie CO2 i ilościowe koncentracje drobnych i grubych cząstek aerozolowych (> 0,3 μm) oraz bioaerozolowych (bakterii i gronkow- ców), a w poszczególnych sektorach tej strefy mierzone były w sposób ciągły parametry termiczne powietrza wewnętrznego. Na podstawie instrumentalnych pomiarów oraz sensorycznych ocen określana była również akceptowalność jakości powietrza (ACC). Zasugerowana została strategia sterowania wentylacją, która oprócz pomiarów CO2 wykorzystuje ciągły monitoring wyczuwalnej jakości powietrza (PAQ) w sektorach auli. Moni- toring PAQ mógłby być realizowany na podstawie pomiarów koncentracji aerozoli i instrumentalnie określanej ACC. Strategia ta zapewni pożądaną PAQ w każdym sektorze, co powinno korzystnie wpłynąć na komfort, zdrowie i produktywność użytkowników, a poprzez usprawnienie działania systemu klimatyzacji strategia ta powinna przyczynić się do oszczędności energii nie tylko w przypadku zastosowania w rozpatrywanej auli. 81

ARCHIVES OF ENVIRONMENTAL PROTECTION vol. 36 no. 4 pp. 81 - 91 2010 PL ISSN 0324-8461

SORPTION OF IBUPROFEN ON SEDIMENTS FROM THE DOB- CZYCE (SOUTHERN POLAND) DRINKING WATER RESERVOIR

Katarzyna Styszko1*, Katarzyna Sosnowska1, Piotr Wojtanowicz2, Janusz Gołaś1, Jerzy Górecki1, Mariusz Macherzyński1

1 Department of Environmental Sciences in Energy Research, AGH University of Science and Technology, Mickiewicza ave. 30, 30-059 Krakow, Poland 2 Department of Management and Environmental Protection, AGH University of Science and Technology, Mickiewicza ave. 30, 30-059 Krakow, Poland * Corresponding author e-mail: [email protected]

Keywords: Sorption, sediments, pharmaceuticals.

Abstract: This work determined the solid-water distribution coefficient Kd, the Freundlich constant KF and the organic carbon normalized coefficient OCK of ibuprofen in natural, aquifer sediments. They are characterized as silt sediments with different clay and sand fraction contents varied in specific surface areas. Content of organic carbon and pH are on the same level. For determining sorption coefficients values of ibuprofen in sediments, its concentration was measured in the aqueous and calculated in the solid phase. Batch tests were conducted following OECD Guideline 106. The resulting Kd values ranged between 1.14 and 2.29 L/kg, KF between 0.25 and 5.48 and KOC between 1.22 and 2.53 for ibuprofen in sediments S1 and S2, respectively. These experiments proved that the presence of clay minerals beside organic carbon and pH might be relevant in sorption of ibuprofen in sediments. A comparison of experimentally determined KOC with modelled KOC calculated on the base of octanol-water partitioning coefficient KOW shows that the prediction of sorption behaviour cannot be based only on KOW. This is probably due to the fact that these approaches well describe hydrophobic interactions, but fail to predict sorption of polar and ionic compounds.

INTRODUCTION

Sediments contaminated with organic and inorganic compounds resulting from human activities are recognized as a world-wide problem. Certain amounts of pharmaceuticals and their residues, among other contaminants, enter the aquatic environment and eventu- ally may occur in drinking water. Because of their environmental persistence they may avoid degradation in sewage and drinking-water treatment plants [7, 2, 3]. Ibuprofen is a non-steroidal anti-inflammatory, analgesic and antipyretic drug, used widely. In Germa- ny, diclofenac and ibuprofen, consumed in quantities of ~75 and 180 tons per year respectively [8] have been recognized as important contaminants in the water cycle. They have been found in sewage and surface water samples, for example, in Greece [15], Canada [31], Finland [16], Spain [19], Slovenia [14], United Kingdom [2] and Switzerland [26]. There has been a study of the occurrence of ibuprofen and another 13 pharmaceutical compounds in UK estuarine surface water [28]. In Poland, in 2000 the volume of ibuprofen sold reached 58 t [10]. 82 K. Styszko, K. Sosnowska, P. Wojtanowicz, J. Gołaś, J. Górecki, M. Macherzyński

The introduction of drugs into the environment is a function of the combination of several factors: the quantity manufactured; the dosage (amount frequency and duration); the excretion efficiency of the parent compound and metabolites; the adsorption/ desorption on soil; and the metabolic decomposition in sewage treatment [9]. They are consequently introduced into the surface waters with the effluents and are present in the receiving waters at concentrations in the ng – μg/L range [26, 27, 30]. Ternes et al. predicted that two processes are responsible for reduction, namely sorption and biodegradation [27]. At present, there has been continuing research carried out on conditions of the degradation of selected pharmaceuticals in wastewater treatment processes [3, 10, 27, 12, 4, 29, 33]. In long-term investigations of sewage and surface water, studies assessing the possible eco-toxicological effects of detected drug residues are very important, because pharmaceutical compounds found in the aquatic environment usually occur as mixtures, not as single contaminants [26, 6, 21, 17]. Sorption is extremely important because it may dramatically affect the fate and impact of chemicals in the environment. The distribution of a chemical between soil and aqueous phases is a complex process depending on a number of different factors: the chemical nature of the substance, the characteristics of the soil (organic carbon content, clay content, soil texture, pH) and climatic factors (rainfall, temperature, sunlight and wind) [18]. The sorption behaviour of organic contaminants can be reasonably predicted by Kd values determined in batch experiments. Low Kd values for ibuprofen reported for primary, and secondary sludge [27, 29] and obtained for digested sludge [5] indicated that in this case, sorption does not play a significant role for the removal of ibuprofen in wastewater treatment plants. According to Urase and Kikuta [29], sorption of acidic pharmaceutical compounds in activated sludge is increased with the decrease in pH. The results of leaching experiments for three types of soils indicated that the leaching potential found could be rated as low for ibuprofen [17]. It means that ibuprofen once applied onto the soil surface, will remain in the upper soil surface and may either be bound by soil particles or may be transformed in the soil [17]. Sorption of polar substances such as pharmaceuticals in different solid matrixes has been of an important concern in the last years and several studies are available in literature dealing with these components in activated sludge [27, 29, 1] and in digested sludge [5]. The solid-water distribution coefficient Kd, the Freundlich constant KF and the organic carbon normalized coefficient OCK for ibuprofen, diclofenac and carbamazepine in natural sandy sediments have been reported in Scheytt [22] studies. The sorption of ibuprofen and other pharmaceuticals with relatively high potential ecological risk and high consumption by river sediment samples has been described in Yamamoto et al. studies [32]. The aim of this study was to determine the sorption behaviour of ibuprofen to natural, aquifer sediments collected from the Dobczyce drinking water reservoir (Poland). The role of clay fraction in sorption of ibuprofen in sediments has been discussed. Total and organic carbon content, particle size distribution and specific surface area of sediments affecting ibuprofen sorption have been evaluated. The Kd, KF and KOC for ibuprofen in natural sediments have been determined. The experimentally determined KOC values were compared with the modelled KOC values calculated on the basis of the KOW (octanol/ water partition coefficient) values. The assessment of sorption of ibuprofen in sediments is of importance for the quality of drinking water in surface reservoir. SORPTION OF IBUPROFEN ON SEDIMENTS FROM THE DOBCZYCE... 83

EXPERIMENTAL

Chemicals and materials Ibuprofen of analytical grade (99% purity) was purchased from Dr. Ehrenstorfer GmbH

(Augsburg, Germany). It is a weak carboxylic acid with a pKa value of 4.52 [20]. Vapour pressure was 2.47 x 10-2 [Pa] and solubility 21 mg/L. A stock standard solution of 1000 µg/mL was prepared in methanol. Working standard solutions, at different concentrations, were prepared by appropriate dilution of the stock solution. All solutions were stored at 40C. Acetonitrile and methanol were HPLC-gradient grade quality from POCH (Gliwice, Poland). Acetone, ethyl acetate (both HPLC grade), potassium dihydrogen phosphate and calcium chloride of analytical grade were obtained from POCH (Gliwice, Poland). In all experiments deionised water (< 0.07 S/cm) from HLP5 pure water system (Hydrolab, Po- land) was used. Chromabond EASY polar modified polystyrene-divinylbenzene, (60 mg; 3mL) solid-phase extraction (SPE) cartridges were purchased from Macherey-Nagel (Düren, Germany).

Sediments The sediments were collected in 2005 in the Dobczyce drinking-water reservoir (South- ern Poland) which supplies over 50% of the tap water to the city of Krakow and other smaller towns around. It is situated to the south of the city of Krakow on the Raba river at 60 km from the source, 270 m above sea level 125 mln m3 and has an area covering 970 ha. The main contaminants originate from agriculture and municipal activities with smaller industrial contributions. Samples of sediments were air-dried, sieved through a 2 mm sieve and stored at 40C. Two types of sediments (S1) and (S2) from inlet and outlet of the reservoir were used in the experiments. Characteristics of the sediments are shown in Table 1.

Table 1. Properties of sediments

Characteristics Sediment S1 Sediment S2 Content of particles % (> 2 mm) Clay 7.3 17.8 (2 – 63 mm) Silt 81.6 79.3 (63 – 2000 mm) Sand 11.1 2.9 Specific surface area 2m /g 12.8 31.9 Total carbon g/kg 18.1 16.6

Fraction of organic carbon fOC kg/kg 0.015 0.016

pH (CaCl2) 7.7 7.6

Particle size distributions of sediments samples were analysed using a Coulter LS 230 Laser Granulometer (0.04 mm – 2000 mm). Morphology was analysed by Scanning Electron Microscopy (JEOL 5400 (EDS) with a Link ISIS 300 analyser, Oxford Instru- ments). The microscopic images (Fig. 1) confirmed previously obtained particle size distributions for each sampling point. From the mineralogical point of view the analysed sediments consist mainly of quartz and alumino-silicates [11]. 84 K. Styszko, K. Sosnowska, P. Wojtanowicz, J. Gołaś, J. Górecki, M. Macherzyński

Fig. 1. Scanning electron microscopic images for sediments samples S1(a) and S2(b) SORPTION OF IBUPROFEN ON SEDIMENTS FROM THE DOBCZYCE... 85

Total carbon and organic carbon were analysed using an ELTRA CS500 instrument. Specific surface areas of sediments were analysed using the BET method (Beckman

Coulter SA3100 Analyser). Sediment pH was measured in a solution of 0.01 M CaCl2.

Batch experiments Batch experiments were conducted as specified by OECD Guideline 106. All experiments were done at room temperature. Ibuprofen sorption for 10 µg/L, 20 µg/L, 40 µg/L, 80 µg/L initial concentrations was analysed. Ten grams of sediments were mixed, in dark glass bottles, with 50 mL of 0.01 M CaCl2 solution, spiked with appropriate working standard solutions and shaken for 24 h using a horizontal shaker at 50 rpm. The sediment suspensions were separated by centrifuging at 2500 rpm for 5 minutes. The aqueous phase was adjusted to pH2 and extracted by solid phase extraction. The concentration of ibuprofen bound to the sediments was measured indirectly as the difference between the concentration at the beginning and the concentration in the solution after equilibration.

Control samples with only the test substance in 0.01M CaCl2 solution (no sediment) were conducted, in the same steps as the test system, in order to check the stability of the test substance and its possible adsorption on the surface of the test vessels. A blank run per sediment with the same amount of sediment and total volume of 50 mL 0.01 M CaCl2 solution (without test substance) were conducted for both sediment samples.

Chemical analysis The analytical procedure for the sediment extracts is shown in Fig. 2.

Fig. 2. An analytical procedure for sediment extracts analysis 86 K. Styszko, K. Sosnowska, P. Wojtanowicz, J. Gołaś, J. Górecki, M. Macherzyński

Chromatographic analysis was performed on a Varian HPLC system with pump 9012, UV-Vis detector 9050 and auto-sampler 9100. Separations were carried out using a LiChrospher® 100 RP-18 (125 mm x 4 mm i.d. 5 µm) cartridge column (Merck, Darmstadt, Germany) protected by a LiChrospher® 100 RP-18 (4 mm x 4 mm id. 5 µm) guard column (Merck). The limit of detection (LOD) 1 µg/L and the limit of quantifica- tion (LOQ) 3 µg/L was established with the use of standard solutions. Depending on the sample concentration (10 – 80 µg/L), the analytical recoveries ranged from 80% to 95% with an R.S.D. between 4.5% to 13%.

Sorption coefficients

The distribution coefficient dK (Nernst partitioning) is the ratio of the concentration of the substance in the solid phase (cs, µg/kg) to the concentration of the substance in aqueous solution (cw, µg/L) at equilibrium.

Kd = cs/cw (1)

Experimentally determined isotherms can commonly be fitted to a relationship of the form:

1/n cs = KF cw (2)

This equation is known as the Freundlich isotherm [23].

The organic carbon normalized sorption coefficient OCK related to the sorption coefficient to the organic carbon content (fraction of organic carbon OCf ) of the sample:

KOC = KF / fOC (3)

RESULTS AND DISCUSSION

Sediment S1 coming from the initial settling trap for suspension, in the first part of reservoir is coarse material with higher content of sand fraction. In the deeper part of reservoir, close to the dum, is fine material, mainly consisting of clay and silt. The content of organic carbon and pH in both tested sediments are the same, so better sorption of ibuprofen on sediment S2 might be caused by higher amount of clay minerals in this sediment and conjugate with its higher specific surface area. Sorption coefficients for sediment S2 are significantly higher than those obtained for sediment S1 (Tab. 2).

Table 2. Distribution coefficients dK , KOC and KF of ibuprofen in sediments K K Sample d F logK 1/n r2 N L /kg mg1-1/nL1/n/kg OC S1 1.14 0.25 1.22 1.44 0.90 8 S2 2.29 5.48 2.53 0.78 0.89 12 N – number of samples SORPTION OF IBUPROFEN ON SEDIMENTS FROM THE DOBCZYCE... 87

In the case of sediment S1 sorption of ibuprofen should be considered mainly on organic matter. In the case of sediment S2 besides sorption on organic carbon, reaction of carboxylic group of ibuprofen with metals (Fe, Al) contained on the surface of the solid (clay minerals) should be considered. In this case a hydroxyl bound to a metal in the solid is displaced by the organic sorbate. Distribution coefficients Kd and organic carbon normalised sorption coefficient KOC for each tested sediment were on the similar level. However, most sorption coefficients values were of the same order of magnitude, but remarkably lower for sediment S1. Ibuprofen sorption coefficients determined for natural sandy sediments determined by Scheytt [22] Kd varied from 0.18 to 1.69 L/kg, KF from 0.21 to 0.83 and KOC from 2.14 to 2.21 for direct and indirect method, respectively. Sediment used by Scheytt [22] in batch experiment was characterized by higher sorption of ibuprofen although lower content of organic carbon (fOC 0.002–0.0013) in comparison to sediments used in our experiment, which might indicate that other sediment properties rather than organic carbon are crucial for sorption of ibuprofen. In this case pH might play main role in sorption of ibuprofen. The sediment used in batch experiment by Scheytt [22] was characterized by lower pH (4.8–6.7) than sediments tested in the current study. In those tested sediments with a pH of 4.8, the carboxyl group of ibuprofen should be at least partly protonated leading to a better sorption onto organic matter because of distinct lipophilicity in its indissociated form. The higher sorption coefficients for sediments with a higher organic carbon content (fOC 0.00075, 0.0087, 0.017) were obtained in Yamamoto study [32]. The sorption data were fitted to the logarithmic transform of the Freundlich equation (Fig. 3).

KF values were found to be 0.25 and 5.48 for sediment S1 and S2, respectively. Ac- cording to Schwarzenbach a value of n < 1 in the Freundlich equation reflects the situation in which at higher sorbate concentrations, it becomes more and more difficult to sorb additional molecules. Specific binding sites become filled or remaining sites are less at- tractive to the sorbate molecules [11]. The n > 1 in this case describes a situation in which previously sorbed molecules lead to the modification of the surface which favours further sorption. The n values determined from 1/n in the Freundlich equations for sediment S1 is

0.7 and 1.3 for sediment S2. There were significant differences between the dK values and the Freundlich constants, mostly due to the non-linearity of the isotherm. Organic carbon normalised sorption coefficients KOC have been calculated and shown to be very helpful in reducing the variability of results.

KOC values are often estimated based on the KOW (octanol/water partition coefficient) 1 values. Comparison of KOC calculated on the basis of correlation equations of Karickhoff 2 (log KOW = 3.5 [31], log KOC = 3.29) [13] and Sontheimer (log KOW = 3.5 [31], log KOC 3 = 2,89) [24] and Scheytt (log KOW = 2.48, log KOC = 2,27) [21] with experimental data showed that calculated values did not match the experimental data with the exception of the Scheytt [21] value, but only for sediment S2. The reason may be that ibuprofen is a carboxylic acid. There are negatively charged ions present, so ionic character plays a major role for sorption of ibuprofen at the tested pH (7.6; 7.7). The presence of polar func- tional group which might interact only with special parts of organic matter and minerals

1 [26] (log KOC = 1.01 log KOW - 0.21)

2 [28] (log KOC = 0.807 log KOW + 0.068) 3 Value based on equation by Karickhoff 88 K. Styszko, K. Sosnowska, P. Wojtanowicz, J. Gołaś, J. Górecki, M. Macherzyński

Fig. 3. Linear form of Freundlich isotherm for ibuprofen for sediment S1 and S2 SORPTION OF IBUPROFEN ON SEDIMENTS FROM THE DOBCZYCE... 89

and hence the KOW and KOC values are inappropriate to simulate pharmaceutical sorption properties in sediment samples.

CONCLUSION

A comparison sorption coefficients of ibuprofen reported in the literature for natural sandy sediments with values of sorption coefficients obtained in presented experiments showed that pH should be considered as one of crucial properties of sediments. Further- more, the presence of clay minerals might play a significant role in sorption of ibuprofen in sediments. In both sediments tested with pH above 7.5, the carboxyl group of ibuprofen occurs in its charged forms so ionic character plays a major role for sorption. The results as those reported in the literature show that the prediction of sorption behaviour cannot be based only on KOW. This is probably due to the fact that these approaches well describe hydrophobic interactions, but fail to predict sorption of polar and ionic compounds.

Acknowledgement This work was supported by the AGH University grant no 11.11.210.199 and by the Kościuszko Foundation, American Center for the Polish Culture with the funds provided by Alfred Jurzykowski Foundation.

REFERENCES

[1] Andersen H.R., M. Hansen, J. Kjolholt, F. Stuer-Lauridsen, T. Ternes, B. Halling-Sorensen: Assessment of the importance of sorption for steroid estrogens removal during activated sludge treatment, Chemos- phere, 61, 139-146 (2005). [2] Ashton D., M. Hamilton, K.V. Thomas: Investigating the environmental transport of human pharmaceuticals to streams in the United Kingdom, Sci. Total Environ., 333, 167-184 (2004). [3] Beausse J.: Selected drugs in solid matrices: a review of environmental determination, occurrence and properties of principal substances, TrAC, Trends Anal. Chem., 23, 753-761 (2004). [4] Carballa M., F. Omil, J.M. Lema, M. Llompart, C. Garcia-Jares, I. Rodriguez, M. Gomez, T. Ternes: Behavior of pharmaceuticals, cosmetics and hormones in a sewage treatment plant, Water Res., 38, 2918- 2926 (2004). [5] Carballa M., G. Fink, F. Omil, J.M. Lema, T. Ternes: Determination of the solid–water distribution coefficient (Kd) for pharmaceuticals, estrogens and musk fragrances in digested sludge, Water Res., 42, 287- 295 (2008). [6] Cleuvers M.: Mixture toxicity of the anti-inflammatory drugs diclofenac, ibuprofen, naproxen, and acetyl- salicylic acid, Ecotox. Environ. Safe., 59, 309-315 (2004). [7] Daughton Ch.C., T.A. Ternes: Pharmaceuticals and Personal Care Products in the Environment: Agents of Subtle Change?, Environ. Health Persp., 107 (Supplement 6) (1999). [8] Daughton Ch.C. and T. Jones-Lepp, eds. Pharmaceuticals and Personal Care Products in the Environ- ment, Scientific and Regulatory Issues, American Chemical Society Symposium Series 791, Washington, DC.: ACS/Oxford University Press (2001). [9] Diaz-Cruz M.S., M.J. Lopez de Alda, D. Barcelo: Environmental behavior and analysis of veterinary and human drugs in soils, sediments and sludge, TrAC, Trends Anal. Chem., 22, 340-351 (2003). [10] Felis E., K. Miksch, J. Surmacz-Górska, T. Ternes: Presence of pharmaceutics in wastewater from waste water treatment plant “Zabrze - Śródmieście” in Poland, Arch. Environ. Prot., 31, 49 (2005). [11] Golas J., B. Kubica, W. Reczynski, W.M. Kwiatek, M. Jakubowska, M. Skiba, M. Stobiński, E.M. Dutk- iewicz, G. Posmyk, K.W. Jones, M. Olko, J. Gorecki: Preliminary studies of sediments from the Dobczyce drinking water reservoir, Pol. J. Environ. Stud., 14, 577-584 (2005). [12] Joss A., S. Żabczyński, A. Goebla, B. Hoffmann, D. Loffler, Ch. S. McArdell, T.A. Ternes, A. Thomsen, H. Siegrist: Biological degradation of pharmaceuticals in municipal wastewater treatment: Proposing a classification scheme, Water Res., 40, 1686-1696 (2006). 90 K. Styszko, K. Sosnowska, P. Wojtanowicz, J. Gołaś, J. Górecki, M. Macherzyński

[13] Karickhoff S.W., D.S. Brown, T.A. Scott: Sorption of hydrophobic pollutants on natural sediments, Water Res., 13, 241-248 (1979). [14] Kosjek T., E. Heath, A. Krbavcic: Determination of non-steroidal anti-inflammatory drug (NSAIDs) residues in water samples, Environ. Int., 31, 679-685 (2005). [15] Koutsouba V., T. Heberer, B. Fuhrmann, K. Schmidt-Baumler, D. Tsipi, A. Hiski: Determination of polar pharmaceuticals in sewage water of Greece by gas chromatography–mass spectrometry, Chemosphere, 51, 69-75 (2003). [16] Lindqvista N., T. Tuhkanenb, L. Kronberg: Occurrence of acidic pharmaceuticals in raw and treated sewages and in receiving waters, Water Res., 39, 2219-2228 (2005). [17] Oppel J., G. Broll, D. Loffler, M. Meller, J. Rombke, T. Ternes: Leaching behaviour of pharmaceuticals in soil-testing-systems: a part of an environmental risk assessment for groundwater protection, Sci. Total Environ., 328, 265-273 (2004). [18] Organization for Economic Cooperation and Development, OECD Guideline for the testing of chemicals – 106, Adsorption – Desorption Using a Batch Equilibrium Method (2000). [19] Rodrıguez I., J.B. Quintana, J. Carpinteiro, A.M. Carro, R.A. Lorenzo, R. Cela: Determination of acidic drugs in sewage water by gas chromatography–mass spectrometry as tert.-butyldimethylsilyl derivatives, J. Chromatogr. A, 985, 265-274 (2003). [20] Roses C.R.M., E. Bosch: A comparison between different approaches to estimate the aqueous pKa values of several non-steroidal anti-inflammatory drugs, Anal. Chim. Acta, 338, 127-134 (1997). [21] Scheytt T., P. Mersmann, R. Lindstadt, T. Heberer: 1-Octanol/Water Partition Coefficients of 5 Pharma- ceuticals from Human Medical Care: Carbamazepine, Clofibric Acid, Diclofenac, Ibuprofen, and Propy- phenazone, Water Air Soil Poll., 165, 3-11 (2005). [22] Scheytt T., P. Mersmann, R. Lindstädt, T. Heberer: Determination of sorption coefficients of pharmaceuti- cally active substances carbamazepine, diclofenac, and ibuprofen, in sandy sediments, Chemosphere, 60, 245-253 (2005). [23] Schwarzenbach R.P., P.M. Gschwend, D.M. Imboden: Environmental Organic Chemistry, Jonh Wiley and Sons, Inc., New York 1993. [24] Sontheimer H., P. Cornel, M. Seym: Untersuchungen zur Sorption von aliphatischen Chlorkohlenwas- serstoffen durch Böden aus Grundwasserleitern, Veröffentlichungen des Bereichs und des Lehrstuhls für Wasserchemie der Universität Karlsruhe, Heft, 22 (1983). [25] Stuer-Lauridsen F., M. Birkved, L.P. Hansen, H.C. Holten Lützhøft, B. Halling-Sørensen: Environmental risk assessment of human pharmaceuticals in Denmark after normal therapeutic use, Chemosphere, 40, 783-793 (2000). [26] Tauxe-Wuersch A., L.F. De Alencastro, D. Grandjean, J. Tarradellas: Occurrence of several acidic drugs in sewage treatment plants in Switzerland and risk assessment, Water Res., 39, 1761-1772 (2005). [27] Ternes T.A., N. Herrmann, M. Bonerz, T. Knacker, H. Siegrist: A rapid method to measure the solid–water distribution coefficient (Kd) for pharmaceuticals and musk fragrances in sewage sludge, Water Res., 38, 4075-4084 (2004). [28] Thomas K.V., M.J. Hilton: The occurrence of selected human pharmaceutical compounds in UK estuar- ies, Mar. Pollut. Bull., 49, 436-444 (2004). [29] Urase T., T. Kikuta: Separate estimation of adsorption and degradation of pharmaceutical substances and estrogens in the activated sludge process, Water Res., 39, 1289-1300 (2005). [30] Wezel van A. P., T. Jager: Comparison of two screening level risk assessment approaches for six disinfect- ants and pharmaceuticals, Chemosphere, 47, 1113-1128 (2002). [31] Xiu-Sheng Miao, B.G. Koenig, C.D. Metcalfe: Analysis of acidic drugs in the effluents of sewage treatment plants using liquid chromatography–electrospray ionization tandem mass spectrometry, J. Chroma- togr. A., 952, 139-147 (2002). [32] Yamamoto H., Y. Nakamura, S. Moriguchi, Y. Nakamura, Y. Honda, I. Tamura, Y. Hirata, A. Hayashi, J. Sekizawa: Persistence and partitioning of eight selected pharmaceuticals in the aquatic environment: Laboratory photolysis, biodegradation, and sorption experiments, Water Res., 43, 351-362 (2009). [33] Yu J.T., E.J. Bouwer, M. Coelhan: Occurrence and biodegradability studies of selected pharmaceuticals and personal care products in sewage effluent, Agr. Water Manage., 86, 72-80 (2006).

Received: March 11, 2010; accepted: July 15, 2010. SORPTION OF IBUPROFEN ON SEDIMENTS FROM THE DOBCZYCE... 91

SORPCJA IBUPROFENU NA SEDYMENTACH POCHODZĄCYCH ZE ZBIORNIKA WODY PITNEJ W DOBCZYCACH

W niniejszej pracy oszacowano stopień sorpcji ibuprofenu na sedymentach pobranych ze zbiornika wody pitnej w Dobczycach poprzez wyznaczenie współczynników podziału Kd, ciało stałe-roztwór wodny, stałych Fre- undlicha KF oraz współczynników adsorpcji znormalizowanych względem węgla organicznego KOC. Proces sorpcji ibuprofenu badano na dwóch sedymentach (S1, S2) o zbliżonej zawartości węgla organicznego oraz pH, zróżnicowanych natomiast pod względem składu granulometrycznego i wielkości powierzchni właściwej. Analizy wykonano opierając się na wytycznych normy OECD nr 106. Eksperymentalne wartości współczynników charakteryzujących sorpcję ibuprofenu dla sedymentów S1 i S2 wynoszą odpowiednio: Kd

~ 1.14 i 2.29 L/kg, KF ~ 0.25 i 5.48 oraz KOC ~ 1.22 i 2.53. Wykazano, iż zawartość frakcji gliniastej obok zawartości węgla organicznego i pH materiału sorpcyjnego może mieć istotny wpływ na procesy sorpcji ibuprofenu. Ponadto porównano modelową wartość współczynnika KOC obliczoną na podstawie współczynnika podziału oktanol-woda KOW z wartością eksperymentalną KOC. Wykazano, iż zdolności sorpcyjne związków nie mogą być prognozowane tylko w oparciu o wartość KOW, gdyż współczynnik ten nie opisuje sorpcji związków polarnych i jonowych. 92 93

ARCHIVES OF ENVIRONMENTAL PROTECTION vol. 36 no. 4 pp. 93 - 105 2010 PL ISSN 0324-8461

ORGANICS (COD) REMOVAL IN DEPENDENCE ON LOADING OF VOLATILE FATTY ACIDS (VFA)

KATARZYNA BERNAT*, IRENA WOJNOWSKA-BARYŁA, ADRIANA DOBRZYŃSKA

University of Warmia and Mazury in Olsztyn, Department of Environmental Biotechnology, Sloneczna str. 45G,10-709 Olsztyn, Poland, phone +48 89 523 41 18, fax +48 89 523 41 31 * Corresponding author [email protected]; [email protected]

Keywords: VFA loading, activated sludge, poly-β-hydroxybutyrate storage, biomass growth, cell respiration, denitrification.

Abstract: The aim of the study was to estimate the influence of volatile fatty acids (VFA) loading on the contribution of the biomass growth, cell respiration, denitrification and poly-β-hydroxybutyrate (PHB) accumulation involved in COD removal by activated sludge. Kinetics of PHB production, PHB and COD consumption were determined. Experimental series were carried out in sequencing batch reactor. The amount of air entering

SBR was maintained at the stable set-point of 2 mg O2/L, oxygen depletion phase occurred in initial hours of the reaction time. SBR was fed with the mixture of municipal wastewater and supernatant from the digesters.

Feast period of the external organic substrate availability (f1) and famine period of little organics availability • (f2) were determined. With VFA loading (rVFA) increase from 0.029 to 0.052 g VFA/g VSS d in the feast period, the effectiveness of COD removal depended on the use of organics for denitrification and internal PHB storage.

PHB content in activated sludge increased from 0.2 to 0.35 Cmol/Cmol. In f1 biomass growth and cell respiration in COD removal decreased from 21 to 14% and from 12 to 5%, respectively. In the famine period the remaining organics were removed due to biomass growth and cell respiration, denitrification and internal PHB storage was not observed.

INTRODUCTION

The study of Barnard et al. [2] showed that organic compounds removal by activated sludge under aerobic conditions is the result of cell respiration, biomass growth and poly- β-hydroxybutyrate accumulation. Gujer and Henze [9] reported that under steady-state conditions the main processes in activated sludge are oxidation and biomass synthesis. These two processes assure 70% removal of organic carbon compounds in soluble and particular form presented in wastewater. Under transient conditions biomass growth becomes unbalanced and intracellular storage of organic polymers is an important adaptive mechanism to the specific conditions. Transient conditions are typical for the configuration where a substrate gradient occurs or where biomass experiences alternately high and low substrate concentrations. Under unsteady conditions in the presence of easily accessible and biodegradable organics, microorganisms able to organic substrates accumulation are promoted among oth- 94 KATARZYNA BERNAT, IRENA WOJNOWSKA-BARYŁA, ADRIANA DOBRZYŃSKA ers and dominated in activated sludge [18]. The biomass adapts to new conditions by increasing the growth rate and by rapid storage of the available substrate in the form of poly-β-hydroxybutyrate. Reserved substances can be used as the energy source essential for biomass synthesis and denitrification. Most of heterotrophic microorganisms are able to perform oxygen and nitrogen respiration, is lower in comparison with utilising another carbon source, i.e. methanol or acetate. However, intracellularly accumulated poly-β- hydroxybutyrate can be successfully used in denitrification under aerobic conditions [3]. Denitrification accounted for the removal of a significant fraction of the influent COD (between 15 and 20%) in the enhanced culture continuous flow system [1]. Types of carbon source and carbon to nitrogen ratio (COD/N ratio) affect the biological reduction of nitrate and nitrite. Kulikowska and Dudek [10] examined sugar-industry waste (molasses) as an organic carbon source for denitrification to determine process efficiency and kinetics. Denitrification rate at COD/N ratio of 6.0 and 5.0 was higher in the reactor with hydrolyzed molasses in comparison with SBR, where untreated molasses was a carbon source. Denitrification efficiency above 98%, irrespective of organic carbon source (untreated molasses, hydrolyzed molasses) was obtained at COD/N ratio. Several investigators have reported that short-chain volatile fatty acids (VFAs) produced through fermentation of primary sludge can be effective as a carbon source for denitrification. Organic short-chained acids as acetate or butyrate are generated in methanogenic fermentation of sewage sludge. According to Lim et al. [11] acetic acid is the main product of acidogenic phase of anaerobic digestion of the primary and the excess sludge and it comprises about 90% of volatile fatty acids. Oleszkiewicz and Barnard [13] reported that limitation of fermentation on the acidogenic step, and maintaining the redox potential above -350 mV are the conditions that promote the high volatile fatty acids concentration in supernatant. The sludge digester supernatant coming from the anaerobic fermentation step is usually fed into the wastewater treatment plant together with wastewater influent in order to increase availability of easily biodegradable organic compounds and promote nitrogen and phosphorus removal. In the supernatant ammonium, the nitrogen concentra- 3 tion ranges from 268 to 1000 g N-NH4/m . High ammonium concentration in anaerobic sludge digester supernatant influences COD/N ratio in wastewater. Therefore, the acces- sibility of easily biodegradable organic compounds for intracellular storage is changing. In this work, the influence of volatile fatty acids loading on the processes performance and kinetics of organic substrate removal was investigated. This is of significant importance on modeling and design of activated sludge treatment plant. The biological removal of organic matter from municipal and industrial wastewater can be accomplished by various process configurations. One of the systems that has demonstrated a good potential in biological processes is sequencing batch reactor (SBR) that offers various advantages, including minimal space requirements, easiness of management and possibility of modification during trial phases through on-line control of the treatment strategy. It allowed the development of techniques and operation strategies able to optimize the treatment plants both in terms of removal efficiencies and costs. ORGANICS (COD) REMOVAL IN DEPENDENCE ON LOADING OF VOLATILE FATTY ACIDS (VFA) 95

METHODS

Experimental design The experiment was carried out in a sequencing batch reactor Bioflo 3000 type, with a working volume of 5 L. The Bioflo 3000 included a temperature probe, stirring control, pH and DO control. The temperature of the reactor was maintained at 20°C using water jacket, the pH was maintained at 7.0. The reactor was equipped with a controlled air supply system. The gas flow rate was controlled by a thermal mass flow controller (TMFC). The constant rate of air entering the sequencing batch reactor was automatically adjusted to a stable set-point (set value was 2 mg O2/L without oxygen consumption by the microorganisms), but it meant that DO concentration in the reactor changed accordingly with the oxygen demand in activated sludge). The dissolved oxygen concentration in the reactor was measured online as a percentage of air saturation. The DO electrode showed changes in oxygen concentration in wastewater during the reaction time. The initial oxygen depletion phase in wastewater was a result of oxygen use by activated sludge. The oxygen depletion phase, despite constant air supply, enhanced PHB accumulation.

Fig. 1. Operating cycle of the BIOFLO 3000

The operating cycle of the bioreactor is shown in Figure 1. Reactor volumetric exchange rate (n) was 0.5 1/d, which means that a half of the treated wastewater was left in the reactor. Four experimental series were performed. SBR was fed with the mixture of 96 KATARZYNA BERNAT, IRENA WOJNOWSKA-BARYŁA, ADRIANA DOBRZYŃSKA municipal wastewater and supernatant from the digesters collected from the municipal WWTP. Supernatant was used as a source of easily biodegradable organic compounds. The content of supernatant from the digesters in the mixture with municipal wastewater increased from 125 to 500 ml (5-20% v/v). However, the high ammonium concentration in supernatant caused the VFA/TN ratio decrease in the mixture. Table 1 presents the characteristic of the substrate feed.

Table 1. Characteristic of the SBR influent

Parameters Series 1 Series 2 Series 3 Series 4 Supernatant content in wastewater [%] 5 10 15 20

Ammonium nitrogen [mg N-NH4/L] 54.72 77.5 110.1 148.6

Organic nitrogen [mg Norg/L] 29.96 35.34 57.54 56.6

VFA [mg CH3COOH/L] 80.95 94.3 128.03 135.25

rVFA [g VFA/g VSS·d] 0.03 0.034 0.046 0.052

COD consisted mainly of easily biodegradable compounds expressed as volatile fatty acids. Along with the increase of supernatant content in the mixture with wastewater, VFA loading (rVFA) increased from 0.03 to 0.052 g VFA/g·d. Activated sludge from a conventional nutrient removal wastewater treatment plant was used as inoculum. The sludge concentration was maintained at around 3 g VSS/l with the age of about 35 d.

Analytical methods In every series the adaptation period lasted about 30 days and was considered complete when the range of changes of particular parameters in the effluent (COD, TKN, N-NH4,

N-NO3, N-NO2) within 7 days’ time did not exceed 5-10%. After activated sludge adaptation to the experimental conditions, the cultivation of the biomass was conducted for about 4 weeks. During this time chemical analyses and PHB measurements (g/g VSS,

Cmol/Cmol) were carried out. Finally, at the end of each series, kinetic analysis, concerning organic substrate transformation, was conducted. When effluent parameters were established and deemed to have reached a steady state, the research was carried out to determine PHB production rate, PHB consumption rate, and the COD removal rate. In the working cycle of the reactor, periodic sampling and measurements of COD and nitrogen compounds were conducted. It was assumed that PHB synthesis proceeds with zero-order rate, PHB degradation with first-order rate in PHB concentration and that the active biomass concentration was constant for each series. COD consumption was divided in feast and famine period and was found to follow zero-order kinetics in both periods. The reaction rate constants were determined based on the experimental data by linear (zero-order) and non-linear (first-order) regression. In order to evaluate the fit of the model to the experimental points, the coefficient φ2 was used. If the coefficient 2φ is closer to zero the fit is better. Respirometric activity of activated sludge was determined with a respirometer Oxi- Top® Control (WTW Wissenschaftlich-Technische Werksträtten Gmbh, D-82326 Weil- heim, Germany). Respirometric measurements were done three times for each series, and the average values were calculated. At the beginning of the reaction time, when the ORGANICS (COD) REMOVAL IN DEPENDENCE ON LOADING OF VOLATILE FATTY ACIDS (VFA) 97 subsequent reactor cycle started, a sample was directly taken from the SBR. The substrate conditions in the respirometric vessels were identical as in the reactor (VSS, organic and nitrogen loading rate). In order to determine the oxygen uptake for organics and endogenous respiration, an autotrophic nitrification inhibitor (allylthiourea) was added. Allylthiourea was shown to be a strong, selective inhibitor of ammonia and nitrite oxidation without affecting other activity [8]. The oxygen uptake for endogenous respiration was determined in a measuring vessel with activated sludge washed twice with distilled water. Wastewater was replaced by distilled water. The system, operating at temperature of 20ºC, registers changes of the pressure in measuring vessel as a result of oxygen uptake by microorganisms. Changes of the pressure were automatically converted to the oxygen uptake expressed as mg O2/L [PN-EN ISO 9408:2005]. Respirometric calculations are presented by Bernat and Wojnowska-Baryła [3].

Chemical analyses Daily measurements of the effluent of the reactor included: chemical oxygen demand (COD), total Kjeldahl nitrogen, ammonia nitrogen, nitrite, and nitrate. The activated sludge was analysed for total suspended solids (TSS) and volatile suspended solids (VSS). PHB content in activated sludge was controlled. The analyses were performed according to APHA (1992). PHB measurements in activated sludge cells were determined with chloroform and sulphuric acid, as described by Gerhardt [7].

Calculation methods In order to make a balance of organic compounds in the reactor during the reaction time, the concentrations of organics used for the biomass growth, internal accumulation, cell respiration, and for denitrification were required. All the calculations to make a balance were according to Bernat and Wojnowska-Baryła [3].

The PHB fraction of active biomass (ƒPHB) in Cmol per total Cmol can be calculated according to Third et al. (2003), as:

ƒPHB(Cmol/Cmol)

Calculation by this method considers a slightly different composition and molecular weight of PHB (CH1.5O0.5, 21.5 g/Cmol) and net biomass (CH1.8O0.5N0.2, 24.6 g/Cmol).

Net PHB-free biomass (Xnet) was calculated by subtracting the PHB content of the cells from the apparent biomass concentration obtained from dry weight (DW).

RESULTS

The work presented was performed in a sequencing batch reactor Bioflo 3000 with controlled air supply rate (constant set value was 2 mg O2/L). DO concentration in the reactor changed ambiently with oxygen uptake by activated sludge, because of oxygen requirements changing with time. During the filling period and some hours of the reaction time DO concentration was nearly zero. Correspondingly, the substrate uptake by the organisms in the reactor had changed with time. At the end of feeding, the substrate concentra- 98 KATARZYNA BERNAT, IRENA WOJNOWSKA-BARYŁA, ADRIANA DOBRZYŃSKA tion in the reactor reached its highest point, and then started to decrease along with the reaction time. Upon the curves of organic compounds removal and dissolved oxygen profiles, the feast and famine periods during the reaction time in series 1-4 were appointed. In the feast period, a high rate of COD removal and intensive oxygen consumption by activated sludge was observed. During the period of higher substrate concentration there was a high oxygen uptake by activated sludge and the oxygen requirements surpassed the amount of available oxygen. Therefore, the DO concentration in the reactor diminished to a level below detection. The duration time with DO concentration below detection level was termed oxygen depletion time. Famine period was characterized by low COD removal rate and an increase in dissolved oxygen concentration in wastewater (Fig. 2).

Fig. 2. Changes in DO and COD concetration durings the reaction time a)series 1, b) series 2, c) series 3, d) series 4; (the rates of COD remoral in relation to the reaction period are shown in the table)

The length of the feast period at 5% content of the supernatant in the mixture with wastewater was 1 h. Along with the increase of the contribution of the supernatant to 20% of the content in the mixture, the duration of the feast period extended to 4 h. Cor- respondingly, oxygen depletion phase extended from 4 h to 7 h at volatile fatty acids loading changing from 0.029 to 0.052 g VFA/g VSS•d, respectively. In the feast period in series 1 at 0.029 g VFA/g VSS•d more than 40% of totally removed COD diminished. In series 2-4 at 0.034-0.052 g VFA/g VSS•d it was more than 60% of COD removed in this period. The specific COD consumption rate in the reactor, both inthe feast and famine periods, proceeded according to zero-order kinetics. Along with the ORGANICS (COD) REMOVAL IN DEPENDENCE ON LOADING OF VOLATILE FATTY ACIDS (VFA) 99 increase of the length of the feast period the specific COD consumption rate was reduced from 46 do 38 mg COD/L•h. There were no significant differences in the specific COD consumption rate in the famine period, the values oscillated around 3 mg COD/L•h (Fig. 2). Under constant aeration conditions with the oxygen depletion phase in the initial hours of the reaction time intracellular poly-β-hydroxybutyrate accumulation of the organic fraction from wastewater in biomass cells was observed. Fig. 3 shows poly-β- hydroxybutyrate content in activated sludge cells (ƒPHB) at t = 0 h depending on the volatile fatty acids loading. The lowest PHB concentration in the biomass – 0.18 g/g VSS.

(0.2 Cmol/Cmol) was noted at rVFA on the level 0.029 g VFA/g VSS•d. The higher contribution of the supernatant in the mixture with wastewater (more easily biodegradable compounds were supported) correlated with higher PHB content in the biomass cells. At rVFA of 0.052 g VFA/gVSS•d in series 4 there was almost 2-fold higher PHB concentration in activated sludge cells (0.32 g/g VSS; 0.35 Cmol/Cmol), in comparison with series 1.

Fig. 3. Poly – B – hydroxybutyrate (PHB) froction in biomass cells depending on the rolatile fatty acids loading rate in series 1-4

Poly-β-hydroxybutyrate content in the biomass cells at the beginning and at the end of the reaction time for each series was on the same level. Accumulation process of the reserved substances started in the feast period (f1), directly after feeding time, when the COD concentration was the highest. It was noted that with the increase of volatile fatty acids loading rate the time of PHB accumulation extended from 1 to 4 h. PHB accumulation took place under constantly aerated conditions, when oxygen depletion phase occurred. PHB production rate proceeded according to zero-order reaction (Fig. 4). At rVFA ranging from 0.029 to 0.046 g VFA/g VSS•d poly-β-hydroxybutyrate accumulation rate was stable on the level of about 16 mg COD/L•h. PHB production rate was lower in series -1 -1 4 and equalled 13.7 mg COD L h at rFVA of 0.052 g VFA/g VSS•d. It was calculated that for poly-β-hydroxybutyrate synthesis activated sludge used 16.6, 32, 48.3 and 54 mg COD/L in series 1-4, respectively. 100 KATARZYNA BERNAT, IRENA WOJNOWSKA-BARYŁA, ADRIANA DOBRZYŃSKA

Fig. 4. Variation of poly – β – hydroxybutyrate content in microbiol cells in fealt and famine periods depending on rolatile fatty acids loading rate: a) series 1, b) series 2, c) series 3, d) series 4, (table below the pictures presents constants rates (k) and the rates of PHB accumulation and degredation) When the moment of the limitation of COD removal started, in the famine period

(f2), PHB consumption began. In the table below (Fig. 4) the values of constant rate (k) and the initial rate of the PHB degradation in the famine period are presented. Poly-β- hydroxybutyrate consumption proceeded according to first-order kinetics. The initial rate of PHB degradation depended on the PHB content in the biomass cells. The increase of the polymer degradation rate from 2.40 to 15.12 mg COD/L•h corresponded to the PHB content in activated sludge cells (increasing from 0.18 to 0.32 g COD/g VSS). Under constant aeration with oxygen depletion phase in wastewater in the reactor in feast period, all the processes that take part in carbon removal (biomass growth, cell respiration, denitrification and intracellular accumulation by activated sludge) were observed. However, in famine period only biomass growth and cell respiration were the processes responsible for COD removal from wastewater. The contribution of the processes in feast and famine periods was changed depending on volatile fatty acids loading rate in series 1-4.

In feast period, an increase of rVFA from 0.029 to 0.052 g VFA/g VSS•d caused the drop in the contribution of the biomass synthesis from 21 to 14% and cell respiration form 12 to 5% in COD removal. Feast period favored denitrification and PHB accumulation as the processes responsible for carbon removal. The contribution of these processes in COD removal increased from 66 to 81%. ORGANICS (COD) REMOVAL IN DEPENDENCE ON LOADING OF VOLATILE FATTY ACIDS (VFA) 101

Fig. 5. The contribution of particular processes in COD removal depending on rolatile fatty acids loading rate in the feast (a) and famine period (b)

In the famine period, poly-β-hydroxybutyrate consumption took place. There was lack of COD consumption for denitrification. The only two processes responsible for COD removal contributed to 20-30% as for biomass synthesis and 80-70% as for cell respiration at the increasing volatile fatty acid loading from 0.029 to 0.052 g VFA/g VSS•d (Fig. 5).

DISCUSSION

Under constant air supply on the level of 2 mg/L organic carbon compound removal took place in feast (f1) and famine (f2) periods. The rate of COD removal shoved the border between the periods. The length of the feast period increased from 1 to 4 h along with the rise of volatile fatty acid loading from 0.029 to 0.052 g VFA/g VSS•d. Beun et al. [5] suggested feast and famine periods under dynamic conditions in activated sludge in a sequencing batch reactor. These authors considered as the beginning of the famine periods the moment when all external substrate (acetate) was consumed. During this time the rate of oxygen uptake by activated sludge was low. Kinetics of COD removal from the mixture of supernatant and municipal wastewater in feast and famine periods, showed in presented paper, can be described with a zero-order equation. It confirmed the results obtained by Beun et al. [5] who found the drop of COD as a linear model. However, in contrary to their suggestion that feast period lasted to the moment of total acetate exhaustion, our results showed fast and slow COD consumption, respectively in feast and famine periods. In feast period f1 the initial rate of COD consumption was 13-15-fold higher that in famine period. Along with the increase of the length of feast period and volatile fatty acid loading the specific COD consumption rate was reduced from 46 do 38 mg COD/L•h. In famine period there was no difference in the specific COD consumption rate (about 3 mg COD/L•h). In the presented study, the rise of volatile fatty acids loading from 0.029 to 0.052 g VFA/g VSS•d caused the drop of the effectiveness of organic carbon compounds removal. Dockhorn et al. [6] observed a decrease of COD removal when easily biodegradable fraction, expressed as VFA, increased. 102 KATARZYNA BERNAT, IRENA WOJNOWSKA-BARYŁA, ADRIANA DOBRZYŃSKA

Storage of PHB occurred to manage the surplus of external substrate. Van Aalst-van Leeuwen et al. [16] found out that both storage and an increase of the biomass growth occurred to manage the surplus of the external substrate at low SRT (0.5 day). Beun et al. [5] showed that at SRT of 9.5 and 19.1 days the specific growth rate was kept more or less constant during one SBR cycle. In this case, mainly storage of PHB occurred to manage the surplus of external substrate, thereby balancing the growth at constant rate. In presented experiment, PHB synthesis occurred in all series only in feast period. It was favoured by oxygen depletion. At SRT of 30-40 days the contribution of the biomass growth and cell respiration in COD removal in feast period decreased. However, in famine period both biomass growth and cell respiration were the only processes responsible for COD removal. It means that in the famine phase there was no denitrification and PHB synthesis. The poly-β-hydroxybutyrate accumulation and denitrification are involved in COD removal and these processes influence cell respiration. When in the environment there are more than one of the electron acceptors, other than oxygen, the contribution of cell respiration in organics removal diminishes. In the famine period of our study there was no use of COD for denitrification because this process was not observed. The process that dominated in organics removal was cell respiration with oxygen as an electron accep- tor. The contribution of cell respiration in COD removal in famine period was 70-80%. Van Niel et al. [17] examined the behaviour of Thiosphaera pantotropha regarding the NADH overflow in the presence of substantial concentrations of the substrate. Continu- ously grown acetate-limited cultures were exposed for short periods to excess acetate in batch. Acetate appeared to be converted mainly to poly-β-hydroxybutyrate (57% wt/wt). Respiration measurements showed that only 29% of the total acetate taken up was oxidized. The remainder (14%) was used for biomass synthesis. After acetate was completely taken up, the cellular PHB content was 42% of the dry weight. In contrary, Van Niel et al. [17] presented that aerobic culture Acinetobacter calcoaceticus oxidized up to 80% of acetate to carbon dioxide and water. In our study, cellular PHB content in activated sludge depended on volatile fatty acid loading. Correspondingly, with the increase of rVFA in the reactor the increase of PHB content in activated sludge from 18 to 32% of dry weight (0.2 to 0.35 Cmol/Cmol) was observed. In feast period, COD present in the mixture of supernatant and wastewater was mainly converted to PHB and was used for denitrification. The contribution of the biomass synthesis decreased from 21 to 14% and cell respiration from 12 to 5%. In the famine period, poly-β-hydroxybutyrate consumption took place and there was no COD consumption for denitrification. The only two processes responsible for COD removal in famine period contributed to 20-30% as for biomass synthesis and 80-70% as for cell respiration. Beun et al. [5] reported that, apparently, the storage mechanism of PHB is energeti- cally efficient. Stimulating the PHB pathway for growth will only marginally decrease sludge production. Authors showed that the reduction in the net biomass yield, with growth via the intermediate PHB, was 4-10% lower compared to direct use of acetate for growth. Independently of substrate conditions our study confirmed the correlation between the rate of PHB consumption and the content of the polymer in microbial cells. Consump- tion of PHB in general could be described kinetically with an nth-order equation. The best fit for all data sets was obtained for reaction order n = 1.3 and reaction rate constant ORGANICS (COD) REMOVAL IN DEPENDENCE ON LOADING OF VOLATILE FATTY ACIDS (VFA) 103 k = 0.44 [5]. Murnleitner et al. [12] examined reaction order of 2, 1, 2/3, and 1/2. The best results were achieved with a 2/3 order for PHB consumption. The kinetics of the enzymatic degradation reactions of the naturally produced polyesters poly-β-hydroxybutyrate (PHB) and poly-β-hydroxybutyrate-co-β-hydroxyvalerate (PHBV) were studied by Tim- mins et al. [15]. Kinetic analysis has revealed that the observed degradation behavior was inconsistent with classical Michaelis-Menten enzymatic kinetics from zero order to first order. Our study showed that PHB consumption proceeded with first-order equation and the rate of that process depended on the poly-β-hydroxybutyrate content in cells of activated sludge (ƒPHB). The higher was the contribution of the supernatant in the mixture with wastewater (more easily biodegradable compounds was supported) the higher was PHB content in the biomass cells. PHB degradation rate increased from 2.40 to 15.12 mg COD/L•h. Buen et al. [4] also revealed that the rate of PHB consumption mainly depended on its content in dry weight. Apparently, PHB degradation at low ƒPHB values occurs at a lower rate than at high ƒPHB values.

CONCLUSION

The main findings from the study are as follows: – Under constant aeration conditions the feast and famine periods during the reaction time were shown. In the feast period, a high rate of COD removal and an intensive oxygen consumption by activated sludge was observed. The famine period was characterized by a low COD removal rate and an increase of dissolved oxygen concentration in wastewater. – The specific COD consumption rate in the reactor, both in the feast and famine

periods, proceeded according to zero-order kinetics. With the increase of rVFA from 0.029 to 0.052 g VFA/g VSS•d the specific COD consumption rate in the feast period was reduced from 46 do 38 mg COD/L•h. In the famine period the values oscillated around 3 mg COD/L•h.

– The lowest PHB concentration in the biomass – 0.18 g/gVSS (0.2 Cmol/Cmol) was

noted at rVFA on the level of 0.029 g VFA/g VSS•d. At rVFA 0.052 g VFA/g VSS•d there was almost 2-fold higher PHB concentration in activated sludge cells – 0.32 g/g VSS

(0.35 Cmol/Cmol). – PHB synthesis proceeds with zero-order rate and its degradation with first-order

kinetic equation. At rVFA ranging from 0.029 to 0.046 g VFA/g VSS•d poly-β- hydroxybutyrate accumulation rate was stable on the level of about 16 mg COD/L•h.

Lower PHB production – 13.7 mg COD/L•h, was observed at rFVA of 0.052 g VFA/g VSS•d. With higher PHB content in activated sludge cells correlated higher rate of its degradation. Polymer consumption rate increased from 2.40 to 15.12 mg COD/L•h. – Of the total COD uptake in feast period, from 21 to 14% was used for the biomass synthesis and from 12 to 5% for cell respiration. The contribution of denitrification and PHB accumulation in organic carbon removal increased from 66 to 81%. In the famine period the only two processes responsible for COD removal contributed to 20-30% as for biomass synthesis and 80-70% as for cell respiration at the increasing volatile fatty acids loading. 104 KATARZYNA BERNAT, IRENA WOJNOWSKA-BARYŁA, ADRIANA DOBRZYŃSKA

REFERENCES

[1] Baraker P.S., P.L. Dold: Denitrification behavior in biological excess phosphorus removal activated sludge systems, Water Res., 30, 769–780 (1996). [2] Barnard J.L.: Designing of wastewater treatment plants with activated sludge in: Designing philosophy and WWTP operation, Krakow 2000, 13–60 (in polish). [3] Bernat K., I. Wojnowska-Baryła: Carbon source in aerobic denitrification, Biochem. Eng. J., 36, 116–122 (2007). [4] Beun J.J., K. Dirks, M.C.M. Van Loosdrecht, J.J. Heijnen: Poly-b-hydroxybutyrte metabolism in dynami- cally fed mixed microbial cultures, Water Res., 36, 1167–1180 (2002). [5] Beun J.J., F. Paletta, M.C.M. Van Loosdrecht, J.J. Heijnen. Stoichiometry and kinetics of poly-β- hydroxybutyrate metabolism in aerobic, slow growing activated sludge cultures, Biotechnol. Bioeng., 67, 379–389 (2000). [6] Dochorn T., N. Dichtl, R. Kayser: Comparative investigations on COD-removal in sequencing batch reactor and continuous flow plants, Water Sci. Technol., 43, 45–52 (2001). [7] Gerhardt P., R.G.E. Murray, W.A. Wood, N.R. Krieg: Methods for general and molecular bacteriology, American Society for Microbiology, Washington 1994. [8] Ginestet P., J-M. Audic, V. Urbain, J-C. Block: Estimation of nitrifying bacterial activities by measuring oxygen uptake in the presence of metabolic inhibitors allylthiourea and azide, Appl. Environ. Microbiol., 64, 2266–2268 (1998). [9] Gujer W., M. Henze: Activated sludge modelling and simulation, Water Sci. Technol., 23, 1011–1023 (1991). [10] Kulikowska D., K. Dudek: Molasses as a carbon source for denitrification, Arch. Environ. Prot., 36, 35 - 45 (2010). [11] Lim S.J., D.W Choi., W.G. Lee, S. Kwon, H.N. Chang: Volatile fatty production from food wastes and its applications to biological nutrient removal, Bioprocess Engin., 22, 543–545 (2000). [12] Murnleitner E., T. Kuba, M.C.M. Van Loosdrecht, J.J. Heijnen: An integrated metabolic model for the aerobic and denitrifying biological phosphorus removal, Biotechnol. Bioeng., 54, 434–450 (1997). [13] Oleszkiewicz, J.A., I.L. Barnard: Acid fermentation of the primary sludge for the improvement of phosphorus removal, Krakow 1997, 86–105 (in polish). [14] Third K.A., N. Burnett, R. Cord-Ruwisch: Simultaneous nitrification and denitrification using stored substrate (PHB) as the electron donor in an SBR, Biotech. Bioen., 83, 706–720 (2003). [15] Timmins M.R., W.R. Lenz, R.C. Fuller: Heterogeneous kinetics of the enzymatic degradation of poly(β- hydroxyalkanoates), Polymer., 38, 551–562 (1997). [16] Van Aalst-van Leeuwen M.A., M.A. Pot, M.C.M. Van Loosdrecht, J.J. Heijnen: Kinetic modeling of poly(β-hydroxybutyrate) production and consumption by Paracoccus pantotrophus under dynamic substrate supply, Biotechnol. Bioeng., 55, 773–782 (1997). [17] Van Niel E.W.J., L.A. Robertson, J.G. Kuenen: Rapid short-term poly-β-hydroxybutyrate production by Tiosphaera pantotropha in the presence of excess acetate, Enzyme Microb. Tech., 17, 977–982 (1995). [18] Wilderer P.A., R.L. Irvine, M.C. Goronszy: Sequencing batch reactor technology, Sci. Tech. Rep., 10, 34–48 (2001).

Received: April 4, 2010; accepted: July 15, 2010.

USUWANIE ZWIĄZKÓW ORGANICZNYCH (CHZT) W ZALEŻNOŚCI OD OBCIĄŻENIA OSADU CZYNNEGO ŁADUNKIEM LOTNYCH KWASÓW TŁUSZCZOWYCH (LKT)

W pracy określono udział usuwania związków organicznych (ChZT) ze ścieków w procesach przyrostu biomasy, oddychania komórkowego, denitryfikacji oraz syntezy kwasu poli-b-hydroksymasłowego (PHB) w zależności od obciążenia osadu czynnego ładunkiem lotnych kwasów tłuszczowych w ściekach. Ponadto przeanalizowano kinetykę magazynowania oraz degradacji kwasu poli-b-hydroksymasłowego. Eksperyment prowadzono w reaktorze sekwencyjnym SBR. Ilość powietrza doprowadzanego do reaktora była automatycznie ustawiona na poziomie 2 mg O2/L, w początkowych godzinach fazy reakcji notowano fazę wyczerpywania tlenu w reaktorze. SBR był zasilany ściekami komunalnymi z udziałem wód nadosadowych z komór fermentacyjnych. ORGANICS (COD) REMOVAL IN DEPENDENCE ON LOADING OF VOLATILE FATTY ACIDS (VFA) 105

Wydzielono fazę żywieniową (f1), w której notowano dużą dostępność związków organicznych oraz fazę głodową (f2) z niewielką dostępnością substratów organicznych. Wraz ze wzrostem obciążenia osadu czynnego ładunkiem lotnych kwasów tłuszczowych (rLKT) z 0,029 do 0,052 g LKT/g s.m.o.•d w fazie żywieniowej efektywność usuwania ChZT zależała od ich zużycia na denitryfikację i syntezę PHB. Zawartość PHB zmaga- zynowanego w osadzie czynnym wzrastała z 0,2 do 0,35 Cmol/Cmol. W f1 obserwowano spadek udziału syntezy biomasy z 21 do 14% oraz utleniania komórkowego z 12 do 5% w usuwaniu ChZT ze ścieków. W fazie głodowej pozostałe związki organiczne były wykorzystywane jedynie na przyrost biomasy oraz oddychanie komórkowe, denitryfikacja oraz syntez PHB nie zachodziły. 106 107

ARCHIVES OF ENVIRONMENTAL PROTECTION vol. 36 no. 4 pp. 107 - 118 2010 PL ISSN 0324-8461

LCC ANALYSIS OF RAINWATER UTILIZATION SYSTEM IN MULTI-FAMILY RESIDENTIAL BUILDINGS

Daniel Słyś1, Tadeusz Bewszko2

1 Rzeszów University of Technology, Department of Infrastructure and Sustainable Development, Powstańców Warszawy ave. 6, 35-082 Rzeszów, Poland, e-mail: [email protected], phone: +48 178651784, Fax. +48 178651172

2 Rzeszów University of Technology, Department of Power Electronics and Electrical Engineering, Pola str. 2, 35-959 Rze- szów, Poland, e-mail: [email protected], phone: +48 17 865 1977, Fax. +48 17 852 44 07

Keywords: Rainwater utilization system, simulation, LCC analysis, scenario analysis.

Abstract: The paper deals with examination of financial profitability of the introduction of rainwater utilization system (RWUS) in multi-family residential buildings. The aim of the work was to build a simulation model of such system and make an LCC analysis of some options of rainwater utilization system. The proposed concep- tion of a new method of selecting the most cost-effective option of RWUS includes: building of simulation model of such system, making the LCC analysis and using a scenario analysis for supporting decision making process with uncertainty. This new method has been applied to a dwelling house in Poland. The results obtained from the analysis demonstrate the unprofitability of the introduction of RWUS in multi-family residential buildings for the adopted location in Poland. The presented method can be used by individual designers and managers to decide on the selection of the most appropriate water supplying option for a specific location.

INTRODUCTION

Poland, despite its location in central part of Europe within the area of mesothermal climate zone, is a country with one of the most unfavorable hydrologic balances on the whole European continent. The annual average run-off of surface waters in lakes and rivers from the area of Poland amounted in the years 1951–2000 to 54 km3, or about 62 km3 with inflow from outside of the country boundaries taken into account. Expressed as a per capita figure this makes about 1600 m3 per year. In other European countries, per capita surface water resources are almost three times as high and amount to 4600 m3 per year as an average. Poland’s water resources are also characterized by a significant seasonal variance and the unevenness of territorial distribution. The most favorable situation exists in the mountainous regions in the southern part of the country (the Carpathian and Sudeten Mountains) and, in the north-east, in the Masuria Lake District area, while central regions of Poland and Silesia have the most unfavorable hydrologic balance. Utilization of atmospheric precipitation as a lower-quality water source is rather rare in Poland. While rainwater collection systems can be found in single-family houses where they are used mainly for irrigation purposes, commercial and multi-family residential buildings are equipped with such facilities very sporadically. 108 Daniel Słyś, Tadeusz Bewszko

As opposed to Poland, rainwater utilization systems (RWUS) are frequently used in other European countries in both small and large buildings. The literature of the subject includes studies on a possibility of using rainwater instead of municipal water in commercial buildings [5], university campuses [2], sport stadiums [15] and individual residential houses [6, 12], as well as studies concerning certain country regions [7, 8]. In on of the studies [13], financial results were presented concerning implementation of rainwater harvesting systems in small dwelling buildings and collection of rainwater from the ground in Namibia (Africa). In view or social and economic conditions, the analysis was focused on very simple rainwater utilization systems. Results of these analyses are especially favorable for systems with rainwater collection from roof surfaces. In the case of rainwater collection from land surface, economical analysis revealed necessity for search of additional financial support from national government in implementation of such systems. The problem of financial unprofitability of RWUS based on specific design solutions is not limited to African countries, but concerns also certain European countries, including Poland. The capital investment return period depends on many factors, of which the most important are: municipal water price, number of system users, rainwater collection and utilization system design. It follows from data published in the literature that the investment outlay return period can range from 6 up to even 210 years [10] in the case of rainwater being used for external purposes only, or about 30 years if it is utilized in a building’s internal systems. The results are confirmed by studies carried out by Brechbühl [3, 4] and the author [11]. Payback period as an investment’s economical effectiveness indicator is an imper- fect (static) tool as it does not take into account the money value evolution in time. For that reason, a definitely more favorable new investment economical analysis method consists in determination of the Life Cycle Cost (LCC). The paper presents results of LCC analysis performed for RWUS applied in a newly designed multi-family dwelling house. According to the Life Cycle Cost methodology, the calculations were performed for the whole undertaking’s existence cycle, taking into account both investment outlays and annual operation and maintenance costs. Calcula- tions were performed for different capacities of rainwater storage tanks and for a building without such system. In the calculations, a newly developed RWUS simulation model [12] was used as well as meteorological data for the selected location.

PROBLEM FORMULATION

From the point of view of a newly designed building user, one of the most important factors affecting the decision on possible implementation of rainwater utilization system (RWUS) in such building consists in possibility of obtaining some annual saving in municipal water purchase cost. The important questions that must be answered by a decision-maker are: (i) should a RWUS be applied and (ii) what should be the optimum capacity of tank used to col- lect rainwater. In fact, the bigger the tank, the higher investment outlay for construction of the system will be, but also annual volume of rainwater used and thus the amount of purchased municipal water will be reduced (resulting in actual savings in annual house maintenance costs). Taking also into account an increasing level of air pollution [14], in LCC ANALYSIS OF RAINWATER UTILIZATION SYSTEM IN MULTI-FAMILY... 109 some cases RWUS should have the possibility of cleaning rainwater not only through mechanical filtration but also by the use of low-pressure membrane systems and disinfection. As the subject of analysis was a newly constructed building, the LCC methodology was used as an analytic tool (analysis covering the whole undertaking’s existence period taking into account both investment outlays and annual operation and maintenance costs). Calculations were performed for several variants corresponding to different rainwater tank capacities and for a variant without tank (the latter being the most frequently used in Poland). A schematic diagram of the most important elements of rainwater utilization system in a multi-family building being the subject of the study is presented in Fig. 1.

Fig. 1. A diagram of distribution of the most important elements of rainwater utilization system together with water flows 1 – tank, 2 – emergency overflow, 3 – filter, 4 – rainwater delivery from the roof, 5 – pump, 6 – water supply to receiving spots, 7 – control, 8 – emergency supply of tap water, 9 – auxiliary tank, Vw – water volume in storage tank, Vr – volume of rainwater inflow to storage tank,Vo – volume of rainwater discharge to sewage system, Vd – water demand for specific purpose, Vu – volume of rainwater outflow from storage tank to equipment using it,Vs – volume of tap water supplied to the system, Vt – capacity of the storage tank

MODEL SPECIFICATION

The RWUS simulation model was constructed in accordance with rules concerning decision-making-oriented mathematical modeling [9] and consists of two sub-models.

In the first RWUS simulation sub-model, the annual average rainwater volumeVR av utilized in a multi-family residential building is calculated on the grounds of meteorologi- 110 Daniel Słyś, Tadeusz Bewszko cal data concerning precipitation. That is a static model with the averaging period of 24 hours. The result of application of that sub-model in the form of the auxiliary variable

VRav represents a base for further calculations aimed at determination of Life Cycle Cost (LCC) cost and carried out with the use of the second RWUS sub-model. The result of calculations performed in that sub-model consists, in turn, in an output LCC variable determined for each of the adopted investment variants and for the assumed investment realization period.

Assumptions During the model realization, a number of assumptions were adopted, including the following important factors: – the storage tank capacity is fixed, Vt = constant; – the largest volume of rainwater accumulated in storage tank, Vr, equals the capacity of the storage tank, Vt; – demand for water in the house is satisfied primarily by water accumulated in the storage tank, and only then by water from the water-supply system; – any excess of rainwater, i.e. quantities exceeding the capacity of storage tank, Vt, is drained to sewage systems or to other rainwater-using equipment; – demand for water depends on the number of inhabitants, average water requirements for specific purpose, as well as the time of the year, Vs = constant; – the simulation model does not take into account the effect of wind direction and strength, as well as that of air temperature and humidity; – the scale of rainwater flow depends on the type of roofing, the roof surface area, roof inclination and the type of precipitation (rain, snow, etc.); – the capacity of the storage tank is higher than the daily demand for water by the respective sanitary system, Vt > Vd; – the model does not take into account the phenomenon of snow sublimation; – because of the small size of the systems, a time shift between the precipitation itself and the rainwater inflow to storage tank was not considered; – the precipitation has a random character, and its quantity is the parameter that char- acterizes it in the developed simulation model; – LCC analysis period of T = 30 years was adopted (service life of rainwater tank and water distribution piping); – based on 10-year archival data, the annual average volume of rainwater used in sanitary system was calculated as well as the annual average of municipal water that must be purchased.

Decision variables – inputs of the model

The decision variables in the model are denoted as xk – meaning implementation of a

RWUS with storage tank with Vtk volume (k ∈ K, K = {0, 1, 2, ..., m}, m ∈ N). Therefore, m + 1 investment variants are analyzed (m of them differing with tank volume Vtk and a variant without tank – Vt0).

Parameters of the model The input parameters of the simulation model are as specified below: – precipitation level over a time interval, p, mm; LCC ANALYSIS OF RAINWATER UTILIZATION SYSTEM IN MULTI-FAMILY... 111

– average demand for water for specific purpose in time interval per inhabitant, Vd, m3/day (24hrs); – number of inhabitants in the building, M, number of inhabitants; – roof surface, F, m2; – average coefficient of rainwater flow from the roof surface area,Y, -; 3 – storage tank capacity, Vtk, m ; – rainwater volume in the storage tank, Vw, m3; – level of storage tank filled with rainwater, h, m; – volume of rainwater inflow in a time interval,Vr , m3; – volume of rainwater outflow from the tank through overflow in a time interval,Vo , m3; – purchase price for 1 m3 of municipal water, Ctw, EUR; – fee for discharge of 1 m3 wastewater, Cs, EUR; – purchase price for 1 kWh of electric power (according to household tariff), Ce, EUR; – discount rate r for analyses in fixed and variable prices, –; – capital investment for each of k RWUS variants (taking into account the cost of rainwater tank and additional system supplying water from the tank to all toilets in the building, and the cost of a simple mechanical cleaning rainwater system);

INVk, EUR.

Auxiliary variables of the model In the analytic model of a decision situation it is also possible to define some auxiliary variables. Although their values are not important for the decision-maker, they facilitate the task of formulation of the model as a whole. This subsection contains auxiliary variables of both the first and the second model.

Auxiliary variables of the 1st sub-model The mode of functioning of the system according to its model is described by a series of conditions which determine the course of processes of rainwater inflow, their accumulation and outflow to sanitary systems and to sewage system. – The process of rainwater filling and accumulation in the tank is defined by the following conditions:

if Vwi + Vpi+1 > Vt, then Vki+1 = Vt, i = 1, 2, …, n;

if Vwi + Vpi+1 ≤ Vt, then Vki+1 = Vwi + Vpi+1, i = 1, 2, …, n. – Rainwater drawing (consumption), from storage tank, by the sanitary system, is characterized by these two conditions:

if Va i – Vdi < 0 , then Vwi = 0 and Vui = Va i, i = 1, 2, …, n;

if Va i – Vdi ≥ 0 , then Vwi = Va i – Vdi and Vui = Vdi, i = 1, 2, …, n. – Tap water consumption by sanitary systems is described by these conditions:

if Va i > Vdi , then Vsi = 0, i = 1, 2, …, n;

if Va i ≤ Vdi , then Vsi = Vdi – Va i, i = 1, 2, …, n. – The process of rainwater outflow (discharge) from storage tank to sewage system is defined by these conditions:

if Va i + Vpi ≤ Vt , then Vo i = 0, i = 1, 2, …, n;

if Va i + Vpi > Vt , then Vo i = Vwi + Vri – Vt, i = 1, 2, …, n. where: 112 Daniel Słyś, Tadeusz Bewszko

Va i – volume of rainwater kept in the tank prior to the water being drawn (consumption) by the system in the i-th time interval, m3; -th 3 Vdi – volume of total water consumption for specific purpose in thei time interval, m ; -th 3 Vki – volume of rainwater kept in the tank at the end of the i time interval, m ;

Vo i – volume of rainwater drained off the system to sewage system, or to ‘infiltration’ equipment, in the i-th time interval, m3; -th 3 Vpi – volume of inflowing rainwater in thei time interval, m ; -th 3 Vri – volume of rainwater supplied to the tank in the i time interval, m ; -th 3 Vsi – volume of tap water supplied to the system in the i time interval, m ;

Vwi – volume of rainwater kept in the tank after water drawing (consumption) by the system in the i-th time interval, m3.

– The amount of municipal water consumed during a year, for k-th investment variant, Vtwk, m3: (1)

-th 3 where VRyk – annual volume of rainwater used in RWUS for k investment variant, m / year: (2)

where Vujk – daily volume of rainwater outflow from the storage tank to the equipment for -th 3 k investment variant, m . Dependence of Vujk on the decision variable xk cannot be given by means of an analytical formula but its value is recursively calculated in the framework of the first sub-model. -th – Annual average of rainwater volume used in RWUS for k investment variant, VRavk, m3/year:

(3) where Tt – period of the used meteorological data on precipitation, years.

Auxiliary variables of the 2nd sub-model

– Annual cost of operation of the system supplying water to toilets, OMCk. In the analysis, costs related to purchase of municipal water, discharge of wastewater and electric power supplying the pumps were taken into account. For k-th investment variant, the cost will be calculated as follows:

(4) where Cp – cost related to transport of 1 m3 rainwater from the tank to users, EUR/m3:

(5) where: Ce – purchase price for 1 kWh of electric power according to tariff applicable to households, EUR/kWh. LCC ANALYSIS OF RAINWATER UTILIZATION SYSTEM IN MULTI-FAMILY... 113 g – acceleration of gravity, m/s2; Δh – average elevation of sanitary facilities with respect to minimum rainwater level in the tank, m; η – efficiency of the motor-pump system, –; ρ – water density, kg/m3.

Outcome variable – output of the model The Life Cycle Cost – LCC – is the total cost of a system over its entire lifespan (the investment cost of RWUS and discounted annual cost of using the RWUS during T years time). The LCC for k-th investment variant is:

(6)

Data used for the model The simulation studies concerning operation of rainwater utilization system were carried out for a multi-family building with the following parameters: – number of floors: 5; – number of stairwells: 4; – number of occupants: 200; – average daily water consumption for flushing toilets: 35 dm3. Simulation calculations were carried out with the use of archival data concerning twenty-four hour precipitation for a period of 10 years observed in the town Rzeszów of located in south-eastern part of Poland. Average annual precipitation in the analyzed period amounted to 612 mm and was slightly lower than the average for the years 1890– 1931, when precipitation was 642 mm, and for the period 1971–2000, when its average value was 629 mm. Other parameters of the simulation model adopted for calculations: 3 3 3 3 – storage tank capacity (6 options): Vt0 = 0 m , Vt1 = 5 m , Vt2 = 10 m , Vt3 = 15 m , Vt4 3 3 = 20 m , Vt5 = 30 m ; – purchase price for 1 m3 municipal water, Ctw = 0.94 EUR; – price for discharging 1 m3 of wastewater, Cs = 0.83 EUR; – purchase price for 1 kWh of electric power (according to tariff for households), Ce = 0.1159 EUR; – discount rate for analyses in fixed (r = 0.05) and variable prices (r = 0.08);

– capital investment for RWUS, INVk: INV0 = 0 EUR, INV1 = 8,080 EUR, INV2 = 8,750

EUR, INV3 = 9,500 EUR, INV4 = 10,250 EUR, INV5 = 11,700 EUR; – LCC analysis period, T = 30 years.

RESULTS OF SIMULATION

The LCC obtained for the analyzed investment is presented in Table 1. The lowest LCC was obtained for the variant without RWUS. It should be, therefore, concluded that implementation of such system in a multi-family residential building is not justified economically for the adopted location in Poland. Lower operation costs related to the reduced use of tap water in buildings with RWUS within the 30-year period 114 Daniel Słyś, Tadeusz Bewszko

Table 1. The LCC for the analyzed investment variants

Variant k Vtk INVk LCCk [EUR] [m3] [EUR] [EUR] 0 0 0 46405 1 5 8080 50880 2 10 8750 50460 3 15 9500 50671 4 20 10250 51084 5 30 11700 52118 covered by the analysis do not compensate high investment outlays related to piping and tank cost.

COPING WITH UNCERTAINTY OF THE MODEL

The whole decision-making support process is aimed at provision of assistance to a decision-maker in taking the best possible decision on the grounds of data from the past, the present and those obtained from forecasts concerning their values in the future. The preceding sections presented the creation process of the mathematical model of the RWUS-related decision problem and results of the performed simulation. Therefore, any decision taken on the grounds of that information will be a decision making use of data concerning both the past and the present. Selection of a solution satisfying the decision-maker on the grounds of additional information, related to forecasted variations of selected model parameters in the future, will permit the decision-maker to take into account the risk related to possible changes and make a better decision. In this paper, the scenario analysis [1] was applied in order to determine the effect of adopted set of model parameters representing the analyzed decision situation on the model’s output variable values. In the scenario analysis it was assumed that in the future (the period adopted for the scenario analysis was 30 years) the following model parameters can be subject to changes: – purchase price for 1 m3 of municipal water Ctw, – discharge fee for 1 m3 of wastewater Cs.

Scenarios for changes in values of selected parameters of the mathematical model In the performed analysis, the following scenarios concerning variations of selected mode parameters were taken into account: (i) the most probable scenario (obtained on the grounds of forecasting), (ii) optimistic scenario (assuming that both mode parameters will vary in a way favourable for the decision-maker – municipal water and wastewater discharge rates will increase slowly), (iii) pessimistic scenario (municipal water and wastewater prices will increase at a fast rate). Detailed values of the adopted parameters are presented in Figs. 2 and 3.

LCC ANALYSIS OF RAINWATER UTILIZATION SYSTEM IN MULTI-FAMILY... 115

Fig. 2. Archival data from 1997–2009 and municipal water price forecast for the years 2010–2039

Fig. 3. Archival data from 1997–2009 and wastewater discharge fee forecast for the years 2010–2039

Results of scenario analysis For each investment variant, a scenario analysis was performed corresponding to the three adopted water and wastewater price variation scenarios. The results of the analysis are presented in Fig. 4. 116 Daniel Słyś, Tadeusz Bewszko

Fig. 4. LCC analysis results for adopted variation scenarios of selected model parameters

From among all forecasts, the lowest LCC was obtained for the variant without rainwater utilization system (zero rainwater tank capacity). It should be therefore concluded that, on the grounds of financial analysis carried out in variable prices, the implementation of RWUS for the analyzed multi-family residential house is not justified for the adopted location in Poland.

CONCLUSIONS

The paper presents results of the Life Cycle Cost (LCC) analysis for rainwater utilization systems in multi-family residential buildings. In the beginning, the decision problem was formulated; then a simulation model for operation of such system was presented. Calculations were carried out for a selected multi-family residential building and actual meteorological data on precipitation. The presentation also included results of scenario analysis supporting the decision-making process in uncertainty conditions. The results obtained from analyses performed for both fixed and variable prices demonstrate explicitly the unprofitability of introduction of RWUS in multi-family residential buildings for the adopted location in Poland. The performed analysis entitles us to formulate a number of general conclusions. – The current municipal water prices and fees for wastewater discharge (compared to prices in Western Europe) significantly affect unprofitability of RWUS implementation for multi-family residential buildings. – A significant increase of municipal water price and wastewater discharge fees adopted in the forecasted scenarios do not result in satisfactory increase of profitability of RWUS implementation in buildings of that type. LCC ANALYSIS OF RAINWATER UTILIZATION SYSTEM IN MULTI-FAMILY... 117

– Investment outlays incurred for implementation of RWUS in Western Europe and Poland are similar, while savings resulting from replacement of municipal water with rainwater are significantly lower in Poland. – The use of RWUS in residential buildings in Poland is currently unprofitable, therefore, implementation of such systems must be based on other (non-financial) decision criteria. The presented LCC method can be used by individual designers and managers to decide on selection of appropriate water supplying option for a specific location. The paper also sets a good example of the full modeling cycle, including model specification, analysis and dealing with uncertainty.

REFERENCES

[1] Andrews C.J.: Evaluating Risk Management Strategies in Resource Planning, IEEE Transaction on Pow- er Systems, 10(1), 420-426 (1995). [2] Appan A.: A dual-mode system for harnessing roofwater for non-potable uses, Urban Water 1(4), 317-321 (1999). [3] Brechbühl B.: Planung und Betrieb von Regenwassernutzungsanlagen. Amt für technische Anlagen und Lufthygiene des Kantons Zürich, 1995. [4] Brechbühl B.: Verminderung der Wasser- und Abwasserkosten – Einsatz und Planung von Regenwasser- nutzungsanlagen Hochbauamt des Kantons Zürich, 1999. [5] Chilton J.C., G.G. Maidment, D. Marriott, A. Francis, G. Tobias.: Case study of a rainwater recovery system in a commercial building with a large roof, Urban Water, 1(4), 345-354 (1999). [6] Fewkes A.: The use of rainwater for WC flushing: the field testing of a collection system, Building and Environment, 34(6), 765-772 (1999). [7] Ghisi E., S. Mengotti de Oliviera: Potential for potable water savings by combining the use of rainwater and greywater in houses in southern Brazil, Building and Environment, 42(7), 1731-1742 (2007). [8] Ghisi E., A. Montibeller, R.W. Schmidt: Potential for potable water savings by using rainwater: An analysis over 62 city in southern Brazil, Building and Environment, 41(2), 204-210, (2006). [9] Makowski M.: Advances in Modeling Methodology for Supporting Environmental Policy-Making, Ar- chives of Environmental Protection, 36, 1, 129-143 (2010). A structured modeling technology, European Journal of Operational Research, 166(3), 615-648 (2005). [10] Mustow S., R. Grey, T. Smerdon, C. Pinney, R. Waggett: Water Conservation: Implications of Using Re- cycled Greywater and Stored Rainwater in the UK. BSRIA Building Services Research and Information Association, Bracknell 1997. [11] Słyś D.: The economic result of system application for utilizing precipitation waters in family housing construction (in Polish), Instal, 6, 66-69 (2005). [12] Słyś D.: Potential of Rainwater Utilization in Residential Housing in Poland, Water and Environment Journal, 23(4), 318-325 (2009). [13] Sturm M., M. Zimmermann, K. Schütz, W. Urban, H. Hartung: Rainwater harvesting as an alternative water resource in rural sites in central northern Namibia, Physics and Chemistry of the Earth, 34, 776- 785 (2009). [14] Wyrwa A.: Towards an Integrated Assessment of Environmental and Human Health Impact of the Energy Sector in Poland, Archives of Environmental Protection, 36, 1, 41-48 (2010). [15] Zaizen M., T. Urakawa, Y. Matsumoto, H. Takai: The collection of rainwater from dome stadiums in Japan, Urban Water, 1 (4), 355-359 (1999).

Received: April 17, 2010; accepted: October 20, 2010.

ANALIZA LCC SYSTEMU ZAGOSPODAROWANIA WÓD DESZCZOWYCH W WIELORODZINNYCH BUDYNKACH MIESZKALNYCH

Artykuł dotyczy analizy rentowności finansowej systemu wykorzystania wody deszczowej (RWUS) w wielorodzinnych budynkach mieszkalnych. Celem pracy było zbudowanie modelu symulacyjnego oraz przeprow- 118 Daniel Słyś, Tadeusz Bewszko adzenie analizy LCC różnych wariantów systemu wykorzystania wody deszczowej. Zaproponowano sposób wyboru najbardziej opłacalnego wariantu systemu RWUS na podstawie analizy w pełnym cyklu istnienia przedsięwzięcia LCC i analizy scenariuszowej, która wspiera proces podejmowania decyzji w warunkach niepewności. Metoda ta została zastosowana dla istniejącego obiektu mieszkalnego. Otrzymane wyniki dla przypadku studyjnego wykazały niską opłacalność wprowadzenia systemu RWUS w wielorodzinnych budynkach mieszkalnych o podobnej charakterystyce w warunkach polskich. Przedstawiona metoda może być wykorzystywana przez projektantów przy podejmowaniu decyzji o wyborze najbardziej odpowiednich wari- antów dostawy wody do budynków dla określonych celów. 119

ARCHIVES OF ENVIRONMENTAL PROTECTION vol. 36 no. 4 pp. 119 - 125 2010 PL ISSN 0324-8461

DGGE-BASED MONITORING OF BACTERIAL DIVERSITY IN ACTIVATED SLUDGE DEALING WITH WASTEWATER CONTAMINATED BY ORGANIC PETROLEUM COMPOUNDS

Aleksandra Ziembińska*1, Sławomir Ciesielski2, Jarosław Wiszniowski1

1 The Silesian University of Technology, Environmental Biotechnology Department, Akademicka str. 2, 44-100, Gliwice, Poland 2 University of Warmia and Mazury in Olsztyn, Department of Environmental Biotechnology, Słoneczna str. 45G, 10-719, Olsztyn, Poland * Corresponding author e-mail: [email protected]

Keywords: Bacterial diversity, DGGE, MBR, POCs, 16S rRNA gene.

Abstract: Polycyclic aromatic hydrocarbons (PAHs) belong to the group of recalcitrants that on reaching wastewater can irreversibly inhibit some sensitive biological processes in activated sludge such as nitrification. This situation leads to wastewater treatment failure due to the influence of these substances on bacteria responsible for important biochemical processes. Observation of the changes in bacterial diversity using molecular tools, such as denaturing gradient gel electrophoresis (DGGE), could be the first step in finding a way of preventing wastewater treatment failure. The aim of this experiment was to monitor bacterial biodiversity in a membrane bioreactor (MBR) dealing with synthetic wastewater contaminated with high concentration of petroleum organic compounds (POCs) and to study the influence of POCs contamination on bacterial changeability in activated sludge. COD removal in investigated membrane bioreactors was at a level of 93%. The organics removal efficiency was not affected by the maximal tested dose of petroleum contamination (1000 µl POCs/l of wastewater) and the MBRs wastewater treatment performance was undisturbed. DGGE analysis revealed that the biodiversity fluctuated slightly in control MBR, while in experimental MBR the biodiversity index decreased drastically after adding the highest experimental concentration of POCs. These results suggest that concentrations of POCs at levels from 50 µl/l to 500 µl/l stimulate biodiversity growth, while the concentration 1000 µl POCs/l of wastewater seems to inhibit the most sensitive processes in wastewater treatment by influencing the bacterial biocenosis.

INTRODUCTION

It is common knowledge that the increasing environmental pollution caused by the inappropriate treatment of some industrial wastewater has a dramatic influence on wastewater composition. There is a wide group of chemicals which on reaching a wastewater treatment plant may inhibit sensitive biological processes such as nitrification. The consequence of such a situation is the treatment failure. Polycyclic aromatic hydrocarbons (PAHs) belong to this group. Some of them can be mutagenic or potentially carcinogenic, while all of them can be dangerous to human health and aquatic life [1, 7, 120 Aleksandra Ziembińska, Sławomir Ciesielski, Jarosław Wiszniowski

14]. PAHs are present in petroleum products as petroleum organic compounds (POCs) and this is the reason why the dramatic development of industries such as petrochemistry, transportation and car exploitation in recent years may be one of the main sources of these substances in WWTPs (wastewater treatment plants). For the past several years membrane bioreactors (MBRs) come to be considered an excellent solution for wastewater treatment for several reasons: their smaller size, which is also suitable for lab-scale experiments, and in most cases an effluent of much better quality than conventional systems in terms of organic matter, suspended solids, and nutrients [4, 8]. The classic microbiological approach fails in the studies of activated sludge microbial diversity mainly because more than 90% of microorganisms present in bioreactors are uncultivable [11]. The only solutions for such experiments are molecular biology tools. One of the most useful methods to study bacterial diversity in complex communities is denaturing gradient gel electrophoresis (DGGE). DGGE allows for the electropho- retic separation of PCR products amplified on DNA derived from a microbial mixture (e.g. activated sludge) [2, 16]. In this study the 16S rRNA coding gene was used due to the fact that it is known to be universal bacterial molecular marker [9]. The diversity of bacteria in complex microbial communities can be estimated on the basis of DNA fingerprints obtained in DGGE, where the Shannon diversity index [H] as an estimation of species richness is calculated [13]. The aim of the study was to monitor the bacterial diversity changes in lab-scale membrane bioreactor (MBR) dealing with wastewater containing a high level of petroleum organic compounds (POCs) using denaturing gradient gel electrophoresis. An in- vestigation of the influence of the POCs on bacterial changeability in activated sludge in an MBR dealing with artificial municipal wastewater was also undertaken.

MATERIALS AND METHODS

Experiment conditions The experiment was carried out for 150 days. It was performed in two lab-scale membrane bioreactors each with a volume of 10.5 l supplied with a Kubota A-4 size membrane. The bioreactors were operated aerobically and inoculated with activated sludge

Table 1. The composition of the synthetic medium for MBRs Component MBR A MBR B Dry meat extract (mg/l) 80 80

CH3COONa (mg/l) 700–1300 700–1300 Yeast extract (mg/l) 10 10

NH4Cl (mg/l) 200–250 200–250

K2HPO4 (mg/l) 27 27

KH2PO4 (mg/l) 10 10

MgSO4×7H2O (mg/l) 15 15 Tween 80 (μl/l) 5-15 5-15 P-30 fraction (μl/l) 0 0–1000 DGGE-BASED MONITORING OF BACTERIAL DIVERSITY IN ACTIVATED SLUDGE... 121 derived from a municipal WWTP in Zabrze (Poland). Bioreactor A (control) and B (containing an increasing concentration of POCs) were fed with a synthetic medium (Tab. 1). The concentration of petroleum organic compounds in the form of a P-30 distillate of crude oil (Oil Refinery, Poland) gradually increased during the experiment in bioreactor B at a rate of 50–1000 µl POCs/l of wastewater. POCs were added to the feeding medium of bioreactor B at 4-week intervals as a mixture emulsified with Tween-80 after the 60 day acclimation period. Nitrogen compound concentrations were determined colorimetrically: ammonia with a Nessler reagent, nitrite with an alfanafty- loamine reagent and nitrate with a dimethylphenol reagent; COD was measured as described previously [15].

Activated sludge samples analysis Activated sludge samples (volume of 50 ml) were collected at 2-week intervals and stored at -20°C until the DNA extraction. Total bacterial DNA was isolated using a Fast DNA Spin Kit for Soil (MP Biomedicals) according to manufacturer’s instructions. PCR reaction was performed using standard primers: 338f-GC (5’ CGC CCG CCG CGC GCG GCG GGC GGG GCG GGG GCA CGG GGG GCC TAC GGG AGG CAG CAG 3’) and 518r (5’ ATT ACC GCG GCT GCT GG 3’) amplifying 180 bp of bacterial 16S rRNA gene [12] as described previously [18]. PCR products were separated using a Dcode Universal Mutation Detection System (BioRad) in polyacrylamide gel (8%, 37:1 acrylamide-bisacrylamide, Fluka) with a 30- 60% gradient of denaturant (urea) prepared according to the manufacturer’s instructions. The gel was run for 9 h at 60 V in a 1×TAE buffer (Tris, acetic acid, EDTA, pH = 8.0) at a constant temperature of 60°C. The gel was stained with SYBR Gold (1:10 000, Invitro- gen) in MiliQ water for 30 min and distained in MiliQ water for 40 min, and then visual- ized under UV light and photographed using the Gel Doc System (BioRad). Densitometric analysis of DGGE fingerprints was performed using Quantity One 1-D Software (BioRad) and the Shannon diversity index [H] was calculated for all of the samples [13].

RESULTS AND DISCUSSION

COD removal in bioreactors A and B (Fig. 1A, C) during the total length of experiment was at a level of 93% and 99%, respectively. The organics removal efficiency was not affected by the maximal tested dose of petroleum contamination reaching 1000 µl POCs/l of wastewater and both the control and experimental MBRs wastewater treatment performance was undisturbed. The ammonia removal in control bioreactor A revealed no disturbances during the experiment (Fig. 1B), while bioreactor B was influenced by petroleum contaminants at the end of the experiment (Fig. 1D). An inhibition of nitrification was observed when contamination was at a + level reaching 1000 µl POCs/l of wastewater. The results of COD and NH4 removal in the activated sludge sampling times are shown in Fig. 1A, B for bioreactor A and in Fig. 1C, D for bioreactor B. The results of the Shannon biodiversity index measurements are shown in Fig. 2A for bioreactor A, and in Fig. 2B for bioreactor B. 122 Aleksandra Ziembińska, Sławomir Ciesielski, Jarosław Wiszniowski

+ Fig. 1. COD and NH4 removal for bioreactor MBRA (A, B) and MBRB (C, D); 0 day sampling time refers to inoculum activated sludge.

Fig. 2. Shannon diversity index changes [H] for bacteria in MBRA (A) and MBRB (B) during the total length of the experiment; 0 day sampling time refers to inoculum activated sludge; numbers 1-4 indicate increasing doses of POC contamination in wastewater: 50 µl/l, 200 µl/l, 500 µl/l and 1000 µl/l, respectively.

DGGE analysis revealed that the biodiversity of the total number of bacteria in the seeding activated sludge samples (inoculum) for both bioreactors was at the same level. During the total length of the experiment the biodiversity fluctuated slightly in bioreactor A, while in bioreactor B the biodiversity index decreased drastically at the end of the DGGE-BASED MONITORING OF BACTERIAL DIVERSITY IN ACTIVATED SLUDGE... 123 experiment after adding the highest experimental concentration of POCs (1000 µl POCs/l of wastewater). DGGE fingerprint analysis showed that the addition of increasing doses of POCs to wastewater did not affect the total bacterial diversity of the activated sludge until the addition of the highest concentration of POCs. The biodiversity decreased within the first two weeks of the acclimation period in both MBRs – 0.5 for bioreactor A and 0.3 for bioreactor B (Fig. 2, sampling time 0-15th days). Such a situation is natural in the case of adaptation to the new environment. A drastic change in the bioreactor’s volume from a high volume of the WWTP bioreactor to a smaller one – 10.5 l resulted in lowering the level of biodiversity. A smaller bioreactor possesses fewer ecological niches for potential inhabitants than a bigger one. Moreover, the activated sludge in the MBR system was only operated under aerated conditions (dissolved oxygen level above 2.0 mg/l), while in the WWTP bioreactors underwent variable: anaerobic-anoxic and aerobic conditions. For MBR B a slight decrease in biodiversity appeared after adding the first dose of POCs – 50 µl/l in the 60th day of the experiment. The POCs dosage was the stress factor in the system, so the decrease in biodiversity was also caused by the adaptation to the new conditions. For the period when increasing doses of POCs were added to MBR B (60th–150th days of the experiment) no significant changes in the level of biodiversity were observed. This situation can be explained in two ways. Firstly, the level of POCs harmfulness for the activated sludge in this system was at the point of 1000 µl/l of wastewater so the microorganisms did not respond to lower doses of the chemicals until reaching the threshold of the negative influence of the POCs. Secondly, there is also possibility that the high concentrations of colloidal and soluble organic particles, such as extracellular polymeric substances (EPS) [3] and surfactants [5], in the activated sludge flocs protect bacteria against the harmful influence of POCs by forming soluble complexes [17] or storing POCs until the level of recalcitrants reached the threshold that was damaging for activated sludge microorganisms. It cannot be excluded that the POC contaminants could have a stronger influence on the biocenosis’ changeability, as well as wastewater treatment performance, that may appear after a longer time of exposition exceeding the length of this experiment. From the results obtained in this study, it can be concluded that concentrations of POCs at levels from 50 µl/l to 500 µl/l seem to stimulate biodiversity growth. It had been proven before that aromatic compounds can serve as a carbon source for several bacterial strains in soil [6] where the degradation of POCs is connected with solubility in water. In WWTP the fraction that it is possible for bacteria to use seems to be higher. So it cannot be excluded that petroleum organic compounds could be an additional carbon source for activated sludge bacteria in this range of concentrations, while a dose of 1000 µl/l of wastewater seems to be the concentration that inhibits the most sensitive processes in wastewater treatment. The results mentioned above can suggest that nitrification may be the first process to be disturbed by a high level of POCs in wastewater treatment. It is possible that nitrifiers would be the first group of microorganisms removed from the system, which would cause a decrease in the biodiversity. This statement requires further research because it has been stated that nitrification is more sensitive to temperature fluctuation than to the influence of recalcitrants due to the presence of flocculating compounds [10].

Acknowledgements This study was financially supported by the Polish Ministry of Science and Higher Educa- tion, grant no: N N 523 411 935. 124 Aleksandra Ziembińska, Sławomir Ciesielski, Jarosław Wiszniowski

References

[1] Bertilsson S., A. Widenfalk: Photochemical degradation of PAHs in freshwaters and their impact on bacterial growth-influence of water chemistry, Hydrobiologia, 469, 23-32 (2002). [2] Chong C. W., A. G. Y. Tan, R. C. S. Wong, M. J. Riddle, I. K. P. Tan: DGGE fingerprinting of bacteria in soils from eight ecologically different sites around Casey Station, Antarctica, Polar Biology, 32, 853-860 (2009). [3] Comte, S., G. Guibaud, M. Baudu: Biosorption properties of extracellular polymeric substances (EPS) resulting from activated sludge according to their type: soluble or bound, Process Biochemistry, 41, 815- 823 (2006). [4] Galil, N., C. Sheindorf, N. Stahl, A. Tenenbaum, A. Y. Levinsky: Membrane bioreactors for final treatment of wastewater, Water Science and Technology, 48/8, 103-110 (2003). [5] Harrison, E.Z., S. R. Oakes, M. Hysell, A. Hay: Organic chemicals in sewage sludges, The Science of the Total Environment, 367, 481-497 (2006). [6] Johnsen A., L. Y. Wick, H. Harms: Principles of microbial PAH-degradation in soil, Environmental Pol- lution, 133, 71-84 (2005). [7] Klejnowski K., B. Kozielska, A. Krasa, W. Rogula-Kozłowska: Polycyclic Aromatic Hydrocarbons in PM1, PM2.5, PM10 and TSP in the Upper Silesian Agglomeration, Poland, Archives of Environmental Protection, 36/2, 65-73 (2010). [8] Lerner M., N. Stahl, N. I. Galil: Comparative study of MBR and activated sludge in the treatment of paper mill wastewater, Water Science and Technology, 55/6, 23-29 (2007). [9] Ludwig W., K.-H. Schleifer: Phylogeny of bacteria beyond the 16S rRNA standard, ASM News, 65, 752- 757 (1999). [10] Miksch K., J. Sikora, J. Surmacz-Górska: Kilka faktów i mitów o nitryfikacji, Forum Exploatatora, 3/48, 112-116 (2010). [11] Muyzer G.: Genetic fingerprinting of microbiota communities-present status and future perspective, [in:] Methods of Microbial Community Analysis, Microbial Biosystems: New Frontiers, Proceedings of the 8th International Symposium on Microbial Ecology, Bell C.R., M. Brylinsky, P. Johnson-Green (ed.), Atlantic Canada Society for Microbial Ecology, Halifax, Canada (1999). [12] Muyzer G., E. C. De Waal, A. G. Uitierlnden: Profiling of complex microbial populations by denaturing gradient gel electrophoresis analysis of polymerase chain reaction-amplified genes coding for 16S rRNA, Applied and Environmental Microbiology, 59, 695-700 (1993). [13] Opelt K., Ch. Berg, S. Schönmann, L. Eberl, G. Berg: High specificity but contrasting biodiversity of Sphagnum-associated bacterial and plant communities in bog ecosystems independent of the geographi- cal region, The ISME Journal, 1, 502-516 (2007) [14] Świetlik R., D. Kowalczyk, J. Dojlido: Influence of selected physicochemical factors on the degradation of PAHs in water, Polish Journal of Environmental Studies, 11/2, 165-169 (2002). [15] Wiszniowski J., A. Ziembińska, S. Ciesielski: Membrane biological reactor (MBR) for treatment of wastewater contaminated by petroleum organic compounds, [in:] Proceedings of a Polish-Swedish-Ukrainian seminar, E. Plaza, E. Levlin (ed.), Stockholm, Sweden, September 23-25, Report no. 19 (2009). http:// www.lwr.kth.se/forskningsprojekt/Polishproject/SUP/index.htm. [16] Wu L., Y. Yu, T. Zhang, W. Feng, X. Xiang Zhang, W. Li: PCR-DGGE fingerprinting analysis of plankton communities and its relationship to lake trophic status, International Reviews of Hydrobiology, 94/5, 528-541 (2009). [17] Zheng X.-J., J.-F. Blais, G. Mercier, M. Bergeron, P. Drogui: PAH removal from spiked municipal wastewater sewage sludge using biological, chemical and electrochemical treatments, Chemosphere, 68, 1143- 1152 (2007). [18] Ziembińska A., S. Ciesielski, K. Miksch: Ammonia oxidizing bacteria community in activated sludge monitored by denaturing gradient gel electrophoresis (DGGE), Journal of General and Applied Microbi- ology, 55, 373-380 (2009).

Received: Jule 7, 2010; accepted: August 31, 2010. DGGE-BASED MONITORING OF BACTERIAL DIVERSITY IN ACTIVATED SLUDGE... 125

MONITORING RÓŻNORODNOŚCI BAKTERYJNEJ W OSADZIE CZYNNYM OCZYSZCZAJĄCYM ŚCIEKI ZANIECZYSZCZONE ORGANICZNYMI ZWIĄZKAMI ROPOPOCHODNYMI Z UŻYCIEM DGGE

Substancje ropopochodne należą do grupy zanieczyszczeń, które, docierając do oczyszczalni ścieków, mogą zaburzać wrażliwe procesy biochemiczne, takie jak nitryfikacja. Taka sytuacja prowadzi do problemów z oczy- szczaniem ścieków, ponieważ substancje te mogą wpływać na bakterie odpowiedzialne za podstawowe procesy biochemiczne. Obserwacja zmian różnorodności bakteryjnej w osadzie czynnym za pomocą narzędzi bio- logii molekularnej, takich jak elektroforeza w gradiencie denaturacji (DGGE), może być pierwszym kroki- em w opracowaniu metody zapobiegania zaburzeniom procesu oczyszczania. Celem tej pracy był monitoring różnorodności bakteryjnej w osadzie czynnym bioreaktora membranowego usuwającego ścieki zanieczyszczone związkami ropopochodnymi i określenie wpływu tych związków na zmienność bakteryjną badanej biocenozy. Usunięcie związków organicznych wyrażonych wskaźnikiem zanieczyszczeń ChZT zarówno w kontrolnym, jak i eksperymentalnym bioreaktorze membranowym, wynosiło powyżej 93%. Najwyższe zas- tosowane stężenie związku zanieczyszczającego (frakcja P-30, próżniowy destylat ropy naftowej) wynosiło 1000 µl/l ścieków i nie miało wpływu na efektywność oczyszczania ścieków. Analiza wzorów prążkowych DGGE wykazała, że bioróżnorodność bakteryjna w bioreaktorze kontrolnym zmieniała się nieznacznie w trak- cie trwania eksperymentu, podczas gdy w bioreaktorze eksperymentalnym spadła drastycznie po dodaniu dawki zanieczyszczającej o najwyższym stężeniu. Wyniki te sugerują, że stężenia modelowych związków ropopochod- nych w zakresie 50 µl/l to 500 µl/l stymulują wzrost bioróżnorodności, podczas gdy zastosowanie zanieczyszczenia w stężeniu 1000 µl/l ścieków prawdopodobnie hamuje najbardziej wrażliwe procesy oczyszczania ścieków poprzez wpływ na biocenozę bakteryjną. 126