Trends in atmospheric deposition fluxes of sulphur and nitrogen in Czech forests
Post on 25-Dec-2016
Embed Size (px)
at SciVerse ScienceDirect
Environmental Pollution 184 (2014) 668e675
Contents lists available
journal homepage: www.elsevier .com/locate/envpol
Trends in atmospheric deposition fluxes of sulphur and nitrogenin Czech forests
Iva Hunov*, Jana Maznov, Pavel KurfrstCzech Hydrometeorological Institute, Na Sabatce 17, 143 06 Prague 4 e Komorany, Czech Republic
a r t i c l e i n f o
Article history:Received 17 January 2013Received in revised form3 May 2013Accepted 7 May 2013
Keywords:Atmospheric depositionCzech RepublicNitrogenSulphurForests
* Corresponding author.E-mail address: email@example.com (I. Hunov).
0269-7491/$ e see front matter 2013 Elsevier Ltd.http://dx.doi.org/10.1016/j.envpol.2013.05.013
a b s t r a c t
We present the temporal trends and spatial changes of deposition of sulphur and nitrogen in Czech forestsbased on records from long-term monitoring. A statistically significant trend for sulphur was detected atmost of the sitesmeasuring forwet, dry, and total deposition fluxes and atmany of these the trendwas alsopresent for the period after 2000. The spatial pattern of the changes in sulphur deposition flux between1995 and 2011 shows the decrease over the entire forested area in a wide range of 18.1e0.2 g m2 year1
with the most pronounced improvement in formerly most impacted regions. Nitrogen still represents aconsiderable stress inmanyareas. The value of nitrogen deposition fluxof 1 gm2 year1 is exceeded over asignificant portion of the country. On an equivalent basis, the ion ratios of NO3
/SO42 and NH4/SO42 inprecipitation show significantly increasing trends in time similarly to those of pH.
2013 Elsevier Ltd. All rights reserved.
Atmospheric deposition has decreased in the recent two de-cades in Europe (EEA, 2011) substantially due to a combination ofseveral reasons as being:
More stringent legislation with consequent implementation ofeffective countermeasures at large sources and decreasingemission of sulphur dioxide (SO2) and to lesser extent of oxidesof nitrogen (NOx),
Restructuring of industry and economic activities after pro-found political and economic changes in the 1990s, and
Economic crises of recent years.
Nevertheless, the deposition values in many European regionsstill remain far from satisfactory (EEA, 2011). The long-term moni-toring of precipitation chemistry and ambient air pollution isessential for quantification of both wet and dry deposition andrevealing the time trends and spatial patterns under major envi-ronmental and climate change, and to link these with potentialenvironmental impacts (Skeffington and Hill, 2012). In contrast tomanyother Europeancountries, thenational network formonitoringof ambient air qualityandchemical compositionof precipitationover
All rights reserved.
the CR has operated for a fairly long time (Hunov, 2001). The reasonis that ambient air quality belonged to the most prominent envi-ronmental issues in the former Czechoslovakia and was negativelyperceived by the public (Moldan and Schnoor, 1992). It was consid-ered an important factor contributing essentially to forest dieback(Fanta, 1997), and also had the most negative consequences for hu-man health which manifested, for example, in life expectancydecrease (Bobak and Leon, 1992).
Sulphur dioxide (SO2) as the first pollutant started to bemeasured in ambient air in the CR in 1960s. The oldest records ofprecipitation chemistry date back to 1978. The major emissionsource of SO2 was the combustion of poor local lignite with highsulphur content. The daily mean concentrations of SO2 in the Czechpart of the so called Black Triangle reached very high values, e.g.1600 mg m3 in January 1982 (CHMI unpublished data). Annualmeans reached hundreds of mg m3 in the most polluted areas(Bridgman et al., 2002).
Severe air pollution, ranking the former communist Czecho-slovakia in the most polluted European countries (Moldan andSchnoor, 1992), has decreased substantially. Profound socio-economic changes in Central Europe in 1989 resulted in signifi-cant improvement of many environmental indicators, includingindustrial emissions. The trend in SO2, NOx and NH3 emissions fromthe Czech sources is shown in Fig. 1.
The aim of the paper is to present the temporal trends and spatialpatterns of sulphur andnitrogen depositionfluxes inCzechmountainforests for the entire period between measurement commencementDelta:1_given nameDelta:1_surnameDelta:1_given namemailto:firstname.lastname@example.org://crossmark.crossref.org/dialog/?doi=10.1016/j.envpol.2013.05.013&domain=pdfwww.sciencedirect.com/science/journal/02697491http://www.elsevier.com/locate/envpolhttp://dx.doi.org/10.1016/j.envpol.2013.05.013http://dx.doi.org/10.1016/j.envpol.2013.05.013http://dx.doi.org/10.1016/j.envpol.2013.05.013
Fig. 1. SO2, NOx and NH3 emissions from Czech emission sources, 1990e2011.
I. Hunov et al. / Environmental Pollution 184 (2014) 668e675 669
and 2011. The trends of sulphur and nitrogen deposition for valuesmeasured until 2001 at selected Czech rural sites operated by theCHMI have already been published (Hunov et al., 2004). Neverthe-less, generally high inter-annual variations in meteorology canconfound attempts to detect trends in deposition through measure-ment. Long term data series of wet deposition of some two decadesare therefore required to measure deposition trends (Matejko et al.,2009). For approximation of the most recent deposition trends,however, we also separately checked the data after 2000.
2.1. Atmospheric deposition of sulphur and nitrogen
For our analysis, we used the data from 15 Czech sites (Table 1) operated by theCzech Hydrometeorological Institute collected in a nation-wide ambient air qualitydatabase (Ostatnick, 2012). The sites are situated as far away as possible andpractical from local sources to meet a demand for regional representativeness. Totaldeposition was calculated as a sum of wet and dry deposition fluxes. The evolutionof the atmospheric deposition measurements over the Czech rural area, thebeginning of precipitation composition measurements at individual sites andaccessibility of ambient air pollution concentration for dry deposition flux estima-tion is summarized in Table 2.
Wet depositionwas calculated based on automated wet-only samples onweeklybasis, analysed by standard methods with comprehensive QA/QC procedures,described in detail by Hunov et al. (2011). Dry deposition was estimated using theinferential method, combining measurements and modelling. This method,
Table 1Stations used for the analysis ranked according to increasing altitude.
Site Characterization Altitude[m a.s.l.]
Ostrava-Poruba Suburban site 242Hradec Krlov Suburban site 276Praha-Libus Suburban site 301Kucharovice Agricultural area, south Moravia 334st n.L.-Kockov Suburban site 367Kocelovice Agricultural area,
Stredocesk hilly area519
Kosetice Agricultural area,Czech-Moravian Highlands
Svratouch Agricultural area,Zdrsk vrchy
Primda Cesk Les 740Cerven Nzk Jesenk 749Sous Jizersk hory Mts. 771Rudolice v Horch Mountain site, Krusn hoty Mts. 840Luisino dol Mountain site, Orlick hory Mts. 875Bl Krz Mountain site, Moravsko-Slezsk
Krkonose-Rchory Mountain site, Krkonose Mts. 1001
recommended by Fowler et al. (2009), is based on the assumption of a steady-staterelationship F c. vd, where the dry deposition flux (F) is a product of the meanconcentration (c) and of the dry deposition velocity (vd). Dry deposition velocitiesused were as follows: for SO2: vd 0.7 cm s1 for forest (0.35 cm s1 for areasoutside the forest), for NOx: vd 0.4 cm s1 for forest (0.1 cm s1 for areas outsidethe forest), for NH3: vd 0.66 cm s1 for forest (0.44 cm s1 for areas outside theforest). Deposition velocities for SO2, NOx and NH3 were calculated using theresistance analogy based on meteorological and land use data for the CR (Dvorakovaet al., 1995; Fiala et al., 2001). Annual mean concentrations of SO2, NOx and NH3wereused as recorded at the sites. If these were not available, we used an expert esti-mation based on the measurements from the neighbouring sites of the same typeaccording to EoI classification (EC, 1997). For SO2 direct measurements were avail-able for 12 out of 15 sites. For NOx, the availability of direct measurements was muchpoorer. Ambient NH3 concentrations were measured only at three sites in total overthe CR. After rejection of one site which represented an area highly impacted bychemical industry, we used the mean calculated from the recorded values of the tworemaining sites as a rough approximation for all sites in the CR. To summarize, totalsulphur deposition was calculated as a sum of dry SeSO2 and wet SeSO42 deposi-tion. Total nitrogen deposition was calculated as a sum of dry NeNOx, NeNH3, andwet NeNO3 and NeNH4 deposition.
In the dry deposition calculationwe did not account for particulate SeSO42, andNeNO3, neither for NeHNO3(g) due to the unavailability of measured data.Currently, particulate NeNO3 and gaseous HNO3 is not measured in the CR. Forparticulate SeSO42 the only data available are from three sites (two rural and onecity-background). According to these records and using the vd 0.25 cm s1 (Fialaet al., 2001) we estimated the annual mean particulate SeSO42 deposition to be0.18 g m2 year1. When compared to the total deposition, this portion canreasonably be considered negligible.
We also did not account for occult deposition as the data is not available.
2.2. Temporal trends
Different methods are used for detecting trends in atmospheric deposition(Marchetto et al., 2013). We analysed the data for temporal trends using the ManneKendall non-parametric test, recommended by the WMO for this kind of data andused in similar studies (Salmi et al., 2002; Sicard et al., 2007; Bashir et al., 2008). Thistest is used to identify the monotonous trend in data-series which does not showany seasonal variation or cycles. Atmospheric chemistry data, however, usually havedistinct seasonal variability and the ManneKendall test is thus applied to annualvalues to overcome the problem of seasonality and autocorrelation. Non-parametricmethods generally have the advantage of being insensitive to outliers, missingvalues are allowed, and data do not need to be normally distributed (Gilbert, 1987;Salmi et al., 2002), which was appropriate for the analysed data set. For calculationwe used software produced by the Finnish Meteorological Institute (Mtt et al.,2002). We analysed both nitrogen and sulphur deposition with respect to contri-bution of wet and dry deposition fluxes, and additionally the wet deposition of H.Apart from deposition fluxes, we also checked the ratios of major ions in precipi-tation e nitrate to sulphate, and ammonia to sulphate e on an equivalent basis. Alltrends were detected for the entire period of measurement which differed amongindividual sites (see Table 2).
2.3. Spatial patterns
A fine resolution maps were developed to study the spatial patterns of atmo-spheric deposition. For construction of maps of total deposition we used all sites
Table 2Beginning of precipitation composition measurements at individual sites andaccessibility of ambient air pollution concentrations for dry deposition fluxestimation.
Site H S/SO42 S/SO2 N/NO3 N/NH4 N/NOx
Wet Wet Dry Wet Wet Dry
Ostrava-Poruba 1995 1996 1996 1995 1995 1995Hradec Krlov 1983 1983 1983 1983 1983 1984Praha-Libus 1986 1985 1992 1986 1988 1988Kucharovice 2002 2002 2002 2002 2002 2002st n.L.-Kockov 1991 1980 1991 1991 1991 1995Kocelovice 1997 1997 1997 1997 1997 eKosetice 1990 1990 1990 1989 1990 1993Svratouch 1978 1978 1985 1978 1989 1994Primda 1990 1989 1991 1990 1990 1994Cerven 1995 1995 1995 1995 1995 1997Sous 1996 1996 1993 1996 1996 1993Rudolice v Horch 1996 1996 1996 1996 1996 1996Luisino dol 2003 2003 2003 2003 2003 2003Bl Krz 1989 1989 1989 1989 1989 1994Krkonose-Rchory 2002 2002 1995 2002 2002 1995
I. Hunov et al. / Environmental Pollution 184 (2014) 668e675670
classified as background according to the EoI classification, EC 1997 (Fig. 2). Thoughthere were more measuring sites at our disposal for the CR, we excluded localitiesclassified as urban, industrial and traffic. Their spatial representativeness is fairlysmall and so they were not considered in our spatial analysis.
The spatial pattern of wet-only deposition was derived from the data of pre-cipitation composition from sites operated by the CHMI, the Czech Geological Sur-vey, the Forestry and Game Management Institute and T.G. Masaryk Water ResearchInstitute. For improving the interpolation in border regions we benefited from dataprovided by the German LfULG and Polish WIOS and IMGW. To be specific: inaddition to 15 measuring sites operated by the CHMI (used also for temporal trendanalysis as described above), we used 28 other rural sites (22 for the Czech area, and6 for the Polish and German border regions), resulting in 43 sites in total forinterpolation. Maps were prepared using the Inverse Distance Weighted (IDW)spatial interpolation technique (e.g. Issaks and Srivastava, 1989). Important infor-mation on precipitation totals was gained from a map based on 750 precipitationgauging sites recorded by standard gauges by the CHMI. Due to the high density ofthe precipitation monitoring network, this map was prepared using universal linearkriging taking into account the dependence of precipitation totals on altitude (Tolaszet al., 2007).
The fields of annual concentrations of SO2 and NOx, essential for estimation ofdry deposition, were constructed using IDW interpolation of values from themeasuring sites in combinationwith the Gaussian dispersionmodel SYMOS 97 (EEA,2010) which adheres to the referencemodelling method of the CR. SYMOS 97 annualconcentrations were calculated using detailed emission inventories for the CR and
Fig. 2. Station network for precipitation composi
relevant meteorological data (Ostatnick, 2012). The maps were derived usingassimilation of measured and modelled data with subsequent interpolation ofresiduals.
The final deposition maps were produced by adding sulphur and nitrogen wetand dry deposition flux maps having a fine spatial resolution of 1 1 km. Themethod was described earlier in detail by Hunov et al. (2011). A digital map ofCzech forests produced from the European digital Land Use map (CORINE LandCover 2000, http://etc-lusi.eionet.europa.eu/CLC2000) was used. All maps wereprepared with ArcGIS Geostatistical Analyst (Johnston et al., 2001). The maps ofdifferences in deposition fluxes were prepared using the tools for map algebra. Weinvestigated the differences between the current situation and 1995 because for thisyear the first spatial patterns, both for sulphur and nitrogen atmospheric depositionfluxes, were possible to prepare. Before 1995 the measuring sites were sparse andtheir results did not allow derivation of the maps due to an insufficient number ofmeasuring points in space.
Table 3 summarizes the temporal trend...