Steady state critical loads of acidity for sulphur and nitrogen: a multi-receptor, multi-criterion approach

The Science of the Total Environment 288 Ž2002. 183᎐197

Steady state critical loads of acidity for sulphur and nitrogen: a multi-receptor, multi-criterion approach J. AherneU , E.P. Farrell Department of En¨ ironmental Resource Management, Uni¨ ersity College, Belfield, Dublin 4, Ireland Received 7 October 2000; accepted 24 June 2001

Abstract The critical load approach to emission reductions has become an important element of the United Nations Economic Commission for Europe ŽUNECE. Convention on Long-Range Transboundary Air Pollution ŽCLRTAP.. The steady state mass balance approach ŽSSMB. is the most widely used method in Europe for estimating critical loads, typically applied to forest ecosystems on mineral soils in conjunction with a base cation to aluminium ŽBc:Al. ratio using a default critical limit of 1.0. The ‘typical’ approach has been expanded in this paper by: inclusion of a broader range of natural and semi-natural receptor ecosystems; inclusion of multiple chemical criterion for each receptor; and an attempt to include organic soils Žpeatlands.. Critical loads of acidity were estimated for the Republic of Ireland using four receptor ecosystems Žconiferous forest, deciduous forest, natural grasslands and moors and heathlands . and seven chemical criteria. The dominant chemical criteria, in the determination of critical loads, were based on a critical pH limit for mineral soils or a critical pH shift in relation to pristine conditions for organic soils. Approximately 68 and 26% of the final distribution of critical loads were estimated using these criteria, respectively. The 5th-percentile critical loads of acidity were more sensitive than those previously estimated for Ireland due to the inclusion of organic soils. Furthermore, coniferous ecosystems had the lowest critical loads due to the high percentage occurring on organic soils, and the removal of base cations through harvesting. The results demonstrate that it may be more appropriate to use multiple criteria and receptors to ensure adequate protection of biological indicators. However, it is important that appropriate critical limits are chosen to protect the biological indicators and receptor ecosystems from long-term damage. 䊚 2002 Elsevier Science B.V. All rights reserved. Keywords: Weathering rate; Critical acid neutralising capacity leaching; SSMB; Ireland

U

Corresponding author. Environmental and Resource Studies, Trent University, 1600 West Bank Drive, Peterborough, Ontario, Canada K9J 7B8. Tel.: q1-705-748-1011 Žext.. 1348; fax: q1-705-748-1569. E-mail address: [email protected] ŽJ. Aherne.. 0048-9697r02r$ - see front matter 䊚 2002 Elsevier Science B.V. All rights reserved. PII: S 0 0 4 8 - 9 6 9 7 Ž 0 1 . 0 0 9 5 8 - 5

184

J. Aherne, E.P. Farrell r The Science of the Total En¨ ironment 288 (2002) 183᎐197

1. Introduction The critical load approach to emission reductions has become an important element of the United Nations Economic Commission for Europe ŽUNECE. Convention on Long-Range Transboundary Air Pollution ŽCLRTAP.. In 1994, critical loads of acidity were used in combination with a so-called ‘sulphur fraction’ ŽDowning et al., 1991. as inputs to the second Sulphur Protocol. More recently, European critical load maps were central to the development of the Gothenburg Protocol on acidification, eutrophication and ground level ozone Žadopted by the Executive Body of the UNECE CLRTAP in November 1999.. Under the CLRTAP, a critical load has been defined as ‘a quantitative estimate of exposure to one or more pollutants below which significant harmful effects on specified sensitive elements of the environment do not occur according to present knowledge’ ŽNilsson and Grennfelt, 1988.. The coupling of ecosystem response to deposition level is the central principle of the critical load approach. In order to apply the concept, four elements need to be considered: receptor; biological indicator; chemical criterion; and critical limit. For each receptor ecosystem under consideration, a representative biological indicator is chosen. A suitable chemical criterion, which is used to predict the risk of damage to the biological indicator, is selected. Finally, a critical chemical limit is assigned, which is the most unfavourable value that the chemical criterion may attain without long-term harmful effects to the receptor. The steady state mass balance ŽSSMB. approach is the most widely used method in Europe for estimating critical loads ŽPosch et al., 1999.. It is primarily used for estimating critical loads for forest ecosystems on mineral soils, but using appropriate chemical criteria and critical limits the SSMB can be applied to any ecosystem ŽWerner and Spranger, 1996.. However, this has generally not been carried out. The chemical criterion most widely used in conjunction with the SSMB is the base cation to aluminium ŽBc:Al. ratio with a default critical chemical limit of 1.0. From the 24 countries submitting data to the European critical

load Co-ordination Centre for Effects ŽCCE., 14 use this criterion ŽPosch et al., 1999.. Ireland has a low, although rapidly increasing, coverage of forest ecosystems Ž7.5%, COFORD, 1994, 9.0% Anon, 2000. and large areas of organic soils Žpeatlands., approximately 17% of total land area ŽHammond, 1981.. This combination of forest ecosystems occurring on organic soils has not been widely considered in critical load determinations. In addition, Ireland is subjected to large inputs of marine-derived base cations due to its location on the western edge of Europe and the predominantly westerly air-flow originating over the Atlantic Ocean. The Bc:Al ratio is very sensitive to atmospheric deposition of marine salts, with high inputs producing high critical loads ŽReynolds, 2000; Aherne et al., 2001.. These factors Žlow forest cover, large areas of organic soils and high marine inputs. have led to the following modifications in the application of critical loads to the Republic of Ireland: inclusion of a broader range of natural and semi-natural receptor ecosystems; inclusion of multiple chemical criteria for each receptor; and an attempt to include organic soils Žcritical loads for peats.. The objective of this paper is to present a multi-receptor, multi-criterion approach for estimating critical loads of acidity. The approach has been used to calculate the maximum critical loads of sulphur and nitrogen for the Republic of Ireland, using four receptor ecosystems and seven chemical criteria across mineral and organic soils. The methodology is based on the hypothesis that a multi-criterion approach provides better protection for the biological indicator, and consequently the receptor ecosystem, against long-term harmful effects.

2. Methods 2.1. Maximum critical loads of sulphur and nitrogen Comprehensive descriptions of the methods for determining steady state critical loads of acidity have been presented by a number of authors ŽSverdrup and de Vries, 1994; Posch et al., 1995; Werner and Spranger, 1996.. In addition, Løkke

J. Aherne, E.P. Farrell r The Science of the Total En¨ ironment 288 (2002) 183᎐197

185

et al. Ž1996., Cresser Ž2000. and Hall et al. Ž2001a. critically review the limitations and uncertainties. Therefore, only a brief summary of the necessary equations is given here. Unless otherwise stated, long-term annual concentrations Žmol c ly1 . or fluxes Žmol c hay1 yeary1 . are required as inputs. The maximum critical load of sulphur, CL max ŽS., is the critical load for acidity assuming only sulphur contributes, i.e. when nitrogen deposition is zero. It is defined in order to take account of any buffering effects of base cation deposition and removal of base cations by vegetation growth:

that have a deleterious effect on the receptor ecosystem. The maximum critical load of nitrogen, CL max ŽN., is the critical load of acidity assuming only nitrogen contributes to acidification, i.e. when sulphur is zero ŽPosch et al., 1995; Werner and Spranger, 1996.. It is therefore equivalent conceptually to the maximum critical load for sulphur but also allows for nitrogen removal processes:

CL max Ž S. s CL Ž Ac act . q nmBc dep y Bc u

where Nu is net annual nitrogen uptake and Ni is annual nitrogen immobilisation, which are deposition independent and solely due to nitrogen removal processes in the soil. Determination of the maximum critical loads of acidity for sulphur and nitrogen, CL max ŽS. and CL max ŽN., requires estimation of the parameters in Eq. Ž1. ᎐Eq. Ž3. ŽBC w , ANC leŽcrit. , nmBc dep , Bc u , Nu and Ni ..

Ž1.

where nmBc dep is the total Žwet plus dry. annual deposition flux of non-marine base cations ŽCa2q, Mg 2q and Kq, note Bc excludes Naq ., Bc u is net annual uptake of base cations and CLŽAc act . is defined as the critical loads of Žactual. acidity ŽSverdrup and de Vries, 1994; Werner and Spranger, 1996.: CL Ž Ac act . s BC w y ANC leŽcrit.

Ž2.

where BC w is the annual release of base cations ŽCa2q, Mg 2q Kq and Naq, note BC includes Naq. from weathering and ANC leŽcrit. is the critical acid neutralising capacity ŽANC. leaching flux, which is the chemical criterion linking chemical changes to harmful effects. For acid soils ŽpH5.0., the critical ANC leaching can be defined in relation to critical concentrations of labile monomeric aluminium ions and hydrogen ions

CL max Ž N. s CL max Ž S. q Nu q Ni

2.2. Base cation weathering, BCw The base cation weathering rate is a crucial component in the calculation of critical load of actual acidity Eq. Ž2.. The Skokloster classification ŽNilsson and Grennfelt, 1988. is a crude method for estimating weathering ranges based on the mineralogy of the soil parent material. The classification is based on ranking, or allocating, soil types to five weathering ranges ŽTable 1.. An assessment of this empirical approach for Irish soils has been carried out by Aherne and Farrell

Table 1 Skokloster classification: mineralogical and petrological classification of soil material Class

Minerals controlling weathering

Usual parent material

Weathering rate Žmolc hay1 yeary1 .

1 2

Granite, quartzite Granite, gneiss

- 200 200᎐500

3

Quartz, K-feldspar Muscovite, plagioclase, biotite Ž- 5%. Biotite, amphibole Ž- 5%.

500᎐1000

4 5

Pyroxene, epidote, olivine Ž- 5%. Calcite, dolomite

Granodiorite, greywacke, shale, schist, gabbro Gabbro, basalt Limestones, marlstones

Source: after Nilsson and Grennfelt Ž1988..

Ž3.

1000᎐2000 ) 2000

186

J. Aherne, E.P. Farrell r The Science of the Total En¨ ironment 288 (2002) 183᎐197

Ž2000. using the general soil map of Ireland Žscale 1:575 000 . and its accompanying bulletin ŽGardiner and Radford, 1980.. A rough validation of the assigned weathering rates was carried out using the PROFILE model in combination with measured soil mineralogy for a small number of soil samples Ž n s 20, Aherne and Farrell, 2001.. PROFILE is a steady state hydrogeochemical model that allows soil solution composition and weathering rate of base cations to be calculated ŽWarfvinge and Sverdrup, 1992.. The model is based on a conceptual model of a soil profile, which is structured into different compartments to represent the natural vertical differences in soils. Detailed descriptions of PROFILE have been documented by Warfvinge and Sverdrup Ž1992. and Jonsson et al. Ž1995.. A ¨ simplified modelling approach, which relied heavily on default inputs, was used for the validation as soil and climate data were limited. Soil structure was simplified by amalgamating soil horizons into three layers, corresponding approximately to A, B and C horizons ŽAherne and Farrell, 2001.. Simplifying soil structure also reduced uncertainty in the selection of default input parameters. Separate model runs were carried out for each soil sample and the estimated weathering rates were scaled to a horizon depth of 50 cm, to correspond to the weathering estimates assigned using the Skokloster classification. Similar studies have been carried out by Langan et al. Ž1995, 1996.. 2.3. Critical acid neutralising capacity leaching, ANCl e(cr it) The basic premise of the SSMB critical loads of acidity approach is that the sensitivity of the biological indicator to aluminium and Žor. hydrogen concentrations can be quantified by specifying a critical limit for one or both concentrations. Aluminium criteria are generally considered less appropriate for organic soils as they contain negligible amounts of minerals. The various chemical criteria and critical limits for estimating soil solution ANC leŽcrit. have been discussed in detail by many authors ŽSverdrup et al., 1990 Hettelingh et al., 1991; Sverdrup and Warfvinge, 1993; Posch et

al., 1995; Werner and Spranger, 1996; Aherne et al., 2001.. A brief description of the different methods is presented here. The critical ANC leaching flux may be estimated using either a critical aluminium or hydrogen ion concentration Žor both. according to: ANC leŽcrit. sy Q= Žw Alx crit q w Hx crit .

Ž4.

where Q is the annual precipitation surplus flux, wAlx crit and wHxcrit are the critical concentrations of labile monomeric aluminium and hydrogen. Sverdrup et al. Ž1990. proposed a critical limit of less than 2.0 mg ly1 for labile aluminium in forest soil solution. Ulrich Ž1987. described soil buffering systems and their pH ranges, he stated that at pH less than 4.2, Al 3q is displaced, enters solution and permanent charge is lost. Operationally, the relationship between wAlx and wHx can be described by the gibbsite equilibrium: w Alx s K gibb = w Hx 3 or w Hx s Žw Alx % K gibb . 1r3

Ž5.

where K gibb is the gibbsite equilibrium constant, the value of which depends on the soil. A widely used default value is 300 m6 moly2 or log K gibb c of 8.0, which represents the B horizon of a mineral soil with approximately 5% organic matter ŽHettelingh et al., 1991; Werner and Spranger, 1996.. With increasing organic matter K gibb decreases, a value of 9.5 m6 moly2 or log K gibb of c 6.5, is widely used for organic soils. In the current study, the spatial distribution of K gibb was defined by reclassifying the general soil map of Ireland into three classes: K gibb 9.5 m6 moly2 for c organic soils; K gibb 100 m6 moly2 for peaty podc zols and peaty gleys; and K gibb 300 m6 moly2 for c the remaining soils. Critical ANC leaching may be estimated using a critical Bc:Al molar ratio in soil solution according to: ANC leŽcrit. s y1.5=

Bc dep q Bc w y Bc u y Q 2r3 Ž Bc:Al. crit

J. Aherne, E.P. Farrell r The Science of the Total En¨ ironment 288 (2002) 183᎐197

= 1.5=

ž

Bc dep q Bc w y Bc u Ž Bc:Al. crit = K gibb

1r3

/ Ž6.

where Bc dep is the total Žwet plus dry. annual deposition flux of Žmarine plus non-marine. Bc, Bc w is the annual weathering flux of Bc and ŽBc:Al.crit is the critical Bc to aluminium ratio. Sverdrup and Warfvinge Ž1993. summarised limiting Bc:Al ratios for a variety of plant species. A default value of 1.0 mol moly1 has been suggested for coniferous forests ŽSverdrup and de Vries, 1994; Werner and Spranger, 1996.. For organic soils Žpeats. an alternative approach based on a base cation to hydrogen ŽBc:H. ratio has been proposed: ANC leŽcrit. s y0.5=

Bc dep q Bc w y Bc u Ž Bc:H . crit

Ž7.

where ŽBc:H.crit is the critical Bc to hydrogen ion ratio. The default value for Bc:H is 0.3= Bc:Al ratio Žmol moly1 . for deciduous trees and ground vegetation, and for spruce and pine it has been suggested to use the same value as the Bc:Al ratio ŽWerner and Spranger, 1996.. Critical ANC leaching may be estimated to prevent depletion of secondary aluminium phases and complexes as this may cause structural changes in soils ŽSverdrup and de Vries, 1994.: ANC leŽcrit. s yp= BC w y Q 2r3 =

ž

p= BC w K gibb

rium with the precipitation they have experienced over recent years. Thus, to define the critical load with respect to current peat conditions would result in values being fixed at current deposition rates ŽCresser et al., 1993; Smith et al., 1993.. The critical ANC leaching flux is estimated from pristine peat pH less a critical pH shift: ANC leŽcrit. sy Q= 10 y Ž pristine peat pH y pH crit .

Ž9.

where the pristine peat pH is estimated from pristine precipitation pH allowing for evapotranspiration and assuming pristine precipitation pH is 5.0 Žsee Smith et al., 1993., pH crit is the critical reduction in peat pH, which is taken to be 0.2 pH units ŽCresser et al., 1993; Smith et al., 1993.. The ANC leŽcrit. value used in the final critical loads of acidity distribution is the minimum ANC leŽcrit. based on seven different criteria using Eq. Ž4. ᎐Eq. Ž9. according to: ANC leŽcrit. s minimum Ž ANC 1leŽcrit. , 2 ANC leŽcrit. , . . . , ANC 7leŽcrit. .

Ž 10.

where 䢇

䢇

/

䢇

Ž8.

䢇 䢇

where p is the stoichiometric ratio of Al to base cation weathering in primary minerals. Sverdrup and de Vries Ž1994. proposed a default value, based on typical North European soils, of 2 Ž1.5᎐3.0. mol c moly1 c . Finally, critical ANC leaching for organic soils Žpeats. may also be estimated in terms of a reduction in peat pH compared to pristine conditions. The concept of defining a critical load with respect to pristine conditions, rather than existing soil conditions, is necessary because most peats may be regarded as being more or less in equilib-

187

䢇

䢇

ANC 1leŽcrit. s critical wHx and critical wAlx ŽpH, Al.; ANC 2leŽcrit. s critical wHx, wAlx determined using the gibbsite relationship ŽpH, K gibb .; 3 ANC leŽcrit. s critical wAlx, wHx determined using gibbsite relationship ŽAl, K gibb .; 4 ANC leŽcrit. s critical Bc:Al ratio ŽBc:Al.; ANC 5leŽcrit. s critical Bc:H ratio ŽBc:H.; ANC 6leŽcrit. s critical Al mobilisation rate Žsoil stability .; and ANC 7leŽcrit. s critical pH shift ŽpH shift..

There is some overlap between the different ᎐3 , which are essenmethods, specifically ANC 1leŽcrit. tially the same. It is possible that all three criteria can result in convergence on the same value. However, in reality the interplay between critical limits, K gibb and precipitation surplus will produce different critical ANC leaching fluxes. In the current literature, there is a noticeable

188

J. Aherne, E.P. Farrell r The Science of the Total En¨ ironment 288 (2002) 183᎐197

lack of recommended criteria and limits for most receptors, except coniferous forests. Similarly, criteria and limits for organic soils are limited. The model input data and critical limits used in this study are given in Table 2. The chosen limits were based on default literature values. The modelling approach used did not concentrate on validation of the limits for receptors on mineral or organic soils, the objective was to present the methodology and its application to the Republic of Ireland. More detailed discussions of chemical criteria and critical limits are given by Aherne et al. Ž2001. and Hall et al. Ž2001a,b.. 2.4. Base cation deposition, Bc d e p The spatial distribution of base cation deposition flux was estimated from interpolated Žkriging. long-term annual bulk concentration measurements Žapprox. 20 points for the period 1985᎐1994. and interpolated long-term rainfall volumes Žapprox. 600 points, 30 year means.. Non-marine bulk deposition flux of base cations was estimated on the basis of their ratios to sodium in seawater. It was assumed that all sodium in rainwater came from sea-spray and that all seasalt components were transported and deposited in the same ratio as is found in seawater.

Filter factors, derived from throughfall measurements, were used to scale from bulk deposition to total Žwet plus dry. deposition ŽAherne and Farrell, 2001.. The filter factors account for increased deposition to receptor ecosystems due to canopy filtering or scavenging. The filter factor approach is based on the theory that there is no canopy exchange of sodium and therefore, all sodium originates from atmospheric deposition ŽUlrich, 1983.. Furthermore, it is assumed that base cation-bearing particles are of the same size as those carrying sodium ŽLovblad et al., 1992.. A ¨ filter factor of 2.0 was used for forests, 1.5 for heathlands ŽBobbink et al., 1992. and 1.0 for all other ecosystems. 2.5. Base cation uptake, Bc u Base cation uptake represents the net annual removal of base cations from the receptor ecosystem. It was assumed that commercial harvesting only occurs in coniferous forests. Moreover, it was assumed that Sitka spruce Ž Picea sitchensis ŽBong.. Carr. was the dominant tree species in commercial growth coniferous plantations, as it is the predominant species in 60% of public sector forests in Ireland ŽCOFORD, 1994.. For deciduous forests, natural grasslands, and moors and heathlands, it was assumed that the only net

Table 2 Model input parameters for multi-receptor, multi-criterion approach Parameter

Units

Value

Reference

wAlxcrit wHxcrit Bc:Al

molc my3 molc my3 mol moly1

0.2 0.063 ŽpH 4.2. 1.0c , 0.8d 1.5g , 0.8h

Bc:H

mol moly1

1.0c , 0.3d 0.5g , 0.3h

p pH shift Bcu Nu BcleŽmin. Bc filter Kgibb Ni

molc molc y1

2.0 0.2 390c , 45dgh 310c , 71dgh 0.002 2.0cd , 1.00g , 1.5h 9.5, 100, 300 145, 215

Sverdrup et al. Ž1990. Ulrich Ž1987. Werner and Spranger Ž1996. Sverdrup and Warfvinge Ž1993. Werner and Spranger Ž1996. Sverdrup and Warfvinge Ž1993. Werner and Spranger Ž1996. Cresser et al. Ž1993. Aherne and Farrell Ž2001. Aherne and Farrell Ž2001. Sverdrup and de Vries Ž1994. Aherne and Farrell Ž2001. Werner and Spranger Ž1996. Posch et al. Ž1995.

molc hay1 molc hay1 molc my3 m6 molcy2 molc hay1

Abbre¨ iations: c, coniferous forests; d, deciduous forests; g, natural grasslands; h, moors and heathlands.

J. Aherne, E.P. Farrell r The Science of the Total En¨ ironment 288 (2002) 183᎐197

removal of base cations was due to animal grazing. Base cation uptake was estimated according to: Bc u s minimum Ž Bc avail , Bc net .

Ž 11.

Bc avail is the physiological available supply of base cations estimated according to: Bc avail s Ž Bc w q nmBc dep y Bc leŽmin . .

Ž 12.

where Bc leŽmin. is the minimum annual base cation leaching flux, which corresponds to those base cations that escape uptake due to physiological limitations ŽSverdrup and de Vries, 1994.. Bc net is the net annual uptake of base cations estimated from growth data or uptake from grazing. For forest soils, this corresponds to base cations that are permanently removed in harvesting, if stems and branches are removed during harvesting then this should be taken into account. Base cation uptake was estimated according to de Vries Ž1991.: Bc net s YC= ␳ = Bc conc

Ž 13.

where YC is yield class or annual growth rate, ␳ is the density of stem wood and Bc conc is the content of base cations in stems. Net annual base cation uptake was estimated using a yield class of 16 m3 hay1 yeary1 , a wood density of 390 kg my3 and stem contents of calcium 0.056%, magnesium 0.021% and potassium 0.067% Župtake estimated at 390 mol c hay1 yeary1 .. The chosen yield class is the average yield class for Sitka spruce in Ireland ŽCOFORD, 1994. and the stem contents are taken form Sitka spruce in Wales ŽEmmett and Reynolds, 1996.. For deciduous forests, natural grasslands, and moors and heathlands a Bc u of 45 mol c hay1 yeary1 was assigned to account for uptake by grazing. In the current study there was limited spatial variation in base cation uptake. Only two values, one for coniferous forests and the other for grazing, were used. However, in regions where the physiological supply is less than the estimated or assigned uptake, then uptake is estimated according to Eq. Ž12..

189

2.6. Nitrogen uptake, Nu Net annual nitrogen uptake for coniferous forests was estimated using the same method as for base cation uptake. Based on a yield class of 16 m3 hay1 yeary1 , wood density of 390 kg my3 and stem content of 0.05%, uptake was estimated at 310 mol c hay1 yeary1 . For deciduous forests, natural grasslands, and moors and heathlands an uptake of 71 mol c hay1 yeary1 was set to account for grazing. Again, the spatial distribution was limited to two values, except in regions where base cation uptake was limited by physiological supply. In this situation, nitrogen uptake was scaled in proportion to base cation uptake. 2.7. Nitrogen immobilisation, Ni Hornung et al. Ž1995. suggested the range 70᎐215 mol c hay1 yeary1 for annual nitrogen immobilisation in coniferous and deciduous forests depending on the climate range warm᎐cold. In the current study, nitrogen immobilisation was approximated according to Downing et al. Ž1991. histosols and podsols were set to 215 mol c , hay1 yeary1 and all remaining soils to 145 mol c hay1 yeary1 . This approach is similar to that used in other European countries ŽPosch et al., 1995.. The spatial distribution of Ni was defined by reclassifying the general soil map of Ireland to two classes. 2.8. Stages in the multi-criterion, multi-receptor approach The stages in the modelling approach are presented in Fig. 1. Stage 1 involves the acquisition and set-up of the spatial databases and the selection of receptor critical limits and parameters. Critical loads were estimated for four receptor ecosystems: coniferous forests; deciduous forests; natural grasslands; and moors and heathlands. The CORINE land-cover database for Ireland Žscale 1:100 000, OSI, 1993. was used to describe their distribution. The model input parameters and critical limits are given in Table 2. Spatial mapping was carried out using a raster Žgridbased. geographical information system ŽGIS..

J. Aherne, E.P. Farrell r The Science of the Total En¨ ironment 288 (2002) 183᎐197

Fig. 1. Stages in the multi-receptor, multi-criterion modelling approach.

190

J. Aherne, E.P. Farrell r The Science of the Total En¨ ironment 288 (2002) 183᎐197

191

Raster systems tend to be very rapid in the evaluation of problems that involve various mathematical combinations of data in multiple layers, since geographical space is uniformly defined in a simple and predictable fashion ŽRonald-Eastman, 1999.. A geographic model registered to the Irish national grid co-ordinate system ŽOSI, 1996., with a grid resolution of 1 = 1 km, was used to store all spatial input data. In stage 2, the maximum critical loads of acidity for sulphur and nitrogen were determined for each receptor. Estimates were based on a multicriterion approach, with the most sensitive criterion being selected according to Eq. Ž10.. Finally in stage 3, the critical loads for each receptor were combined and their co-ordinates converted into the 150 = 150-km EMEP grid projection. In a single EMEP grid many critical load values for each receptor were calculated. These critical loads were sorted by magnitude, taking into account the area of the receptor ecosystem they represent, and the so-called ‘cumulative distribution function’ ŽCDF. was constructed ŽPosch et al., 1995.. From the CDF the 5th, 50th and 95th percentiles were calculated. The 150 = 150km grids that cover Ireland are given in Fig. 2.

3. Results and discussion 3.1. Spatial databases Critical loads of acidity were estimated using four receptor ecosystems. Based on the CORINE land-cover database the receptor areas were estimated to be: coniferous 2454 km2 ; deciduous 1813 km2 ; grasslands 2055 km2 ; and moors and heathlands 2636 km2 . The percentage cover of these receptors is: coniferous 3.49%; deciduous 2.58%; grasslands 2.92%; and moors and heathlands 3.75%. The total area for which critical loads were estimated is 8957 km2 , which is equivalent to 12.74% of the area of the Republic of Ireland. Løkke et al. Ž2000. state that the weathering rate is probably the most important term in the critical load calculation for acidity. Using the Skokloster classification, a map of base cation

Fig. 2. EMEP grid IDs.

weathering rate was produced from the general soil map of Ireland. Soils with the lowest weathering rates are associated with granite, schist, gneiss and quartzite parent material and organic soils ŽTable 1.. Approximately 16% of the soils were allocated to the range 0᎐200 mol c hay1 yeary1 , 13% to 200᎐500 mol c hay1 yeary1 , 24% to 500᎐1000 mol c hay1 yeary1 , 7% to 1000᎐2000 mol c hay1 yeary1 and 40% to greater than 2000 mol c hay1 yeary1 . The mid-point of each range was used to define base cation weathering, except for the first Žmost sensitive. class, which was set at 100 mol c hay1 yeary1 , and the final Žnon-sensitive. class, which was set at 4000 mol c hay1 yeary1 .

J. Aherne, E.P. Farrell r The Science of the Total En¨ ironment 288 (2002) 183᎐197

192

A basic validation of the weathering rate, for selected soils with measured mineralogy Ž20 samples., was carried out using the PROFILE model ŽTable 3.. In general, there was good agreement between PROFILE and Skokloster estimates. The differences that exist can be attributed to the simple modelling approach used, which relied heavily on default input data and the uncertainty in matching soil samples to their correct parent material. Unfortunately, there was considerable confusion concerning the exact co-ordinates and soil types for samples obtained from the national soil archive ŽAherne and Farrell, 2001.. However, overall there was good agreement between methods, indicating that the Skokloster classification approach was adequate for the national weathering rate map. The general soil map of Ireland was also used to describe the spatial distribution of K gibb and Ni . For K gibb , approximately 16% of the soils 6 were allocated to 9.5 m6 moly2 c , 7% to 100 m y2 6 y2 mol c and 76% to 300 m mol c . For Ni , 53% of the soils were set to 145 mol c hay1 yeary1 , the remaining 47% were set to 215 mol c hay1 yeary1 .

The estimated bulk deposition of base cations for the receptor ecosystems ranged between approximately 260 and 5400 mol c hay1 yeary1 with an average of 900 mol c hay1 yeary1 . In comparison, non-marine base cation bulk deposition ranged between approximately 50 and 400 mol c hay1 yeary1 with an average of 110 mol c hay1 yeary1 . Bulk depositions were scaled to total depositions using filter factors for each receptor ecosystem ŽTable 2.. Precipitation for the chosen ecosystems ranged between approximately 700 and 3500 mm yeary1 with an average of 1200 mm yeary1 . Precipitation surplus Žestimated as precipitation minus evapotranspiration and surface runoff. ranged between approximately 100 and 2300 mm yeary1 with an average of 600 mm yeary1 . 3.2. Multi-criterion SSMB Several chemical criteria were used in the determination of critical loads, five for mineral soils, two for organic soils. For each receptor ecosystem the range of chemical criteria were applied, for every 1 = 1-km grid-square, the minimum

Table 3 Comparison between PROFILE and Skokloster base cation weathering Žmol c hay1 yeary1 . Sample

Parent material

Profile 50 cm

Skokloster

1 2 3 4† 5 6† 7 8 9† 10 11 12 13† 14 15 16 17 18† 19† 20

Limestone glacial till Limestone gravely till Till of Irish sea origin with limestone and shale Sandstone, lower Avonian shale glacial till Mostly sandstone Morainic sands and gravels and blown sands Sandstone glacial till Mostly upper carboniferous limestone and shale Mostly granite or rhyolite glacial till Mica schist glacial till Mostly granite, sandstone Limestone Mixed sandstone, limestone glacial till Upper carboniferous shale glacial till Limestone glacial till Stony limestone glacial till Mostly granite, sandstone Sandstone, granite, mica schist glacial till Mostly sandstone Mostly granite, sandstone

38 158 10 727 48 149 257 751 62 749 23 812 1020 806 222 27 319 405 1331 74 208 91 974 377 1642 303 215

) 2000 ) 2000 ) 2000 500᎐1000 500᎐1000 200᎐500 500᎐1000 ) 2000 200᎐500 500᎐1000 200᎐500 ) 2000 ) 2000 1000᎐2000 ) 2000 ) 2000 200᎐500 500᎐1000 500᎐1000 200᎐500

†

Samples with differences between modelled and assigned weathering rates.

J. Aherne, E.P. Farrell r The Science of the Total En¨ ironment 288 (2002) 183᎐197

193

Table 4 Area Žkm2 ., and percentage of total receptor area, for each criterion in the determination of critical loads of acidity Criterion

pH, Al pH, Kgibb Al, Kgibb Bc:Al Bc:H Soil stability pH shift Total

Coniferous

Deciduous

Grassland

Heathland

All ecosystems

Area

%

Area

%

Area

%

Area

%

Area

%

1608

65

1359

75

1279

62

1830

69

6076

68

107 37 703

4 2 29

24 429

1 24

148 628

7 31

249 557

10 21

107 458 2316

1 5 26

2454

1813

ANC leŽcrit. estimated from the range of chemical criteria according to Eq. Ž10. was used to estimate the critical load. The contribution of each chemical criterion to the final distribution of critical load for each receptor is given in Table 4. The pH, K gibb criterion was dominant, determining 62᎐75% of the critical loads for each receptor, followed by the pH shift criterion, which determined 21᎐31% of the critical loads for each receptor ŽTable 4.. The combined distribution of the estimated critical loads for all receptors was determined by four criterion: pH, K gibb Ž68%.; Bc:H Ž1%.; soil stability Ž5%.; and pH shift Ž26%.. This indicates that 27% of the mapped receptors occur on organic soils ŽBc:H plus pH shift.. The Bc:Al chemical criterion, which is the most widely used in European critical load assessments ŽPosch et al., 1999., did not account for any critical loads in the final distribution ŽTable 4.. This is due to the inclusion of marine plus nonmarine base cation deposition in the estimation of ANC leŽcrit. according to Eq. Ž6. ŽHornung et al., 1995; Werner and Spranger, 1996., which has been previously discussed by Aherne et al. Ž2001.. The choice of critical limits and model parameters will effect the final critical load. Therefore, it is essential that appropriate critical limits are set to protect the biological indicator, and consequently, the receptor ecosystem, against long-term harmful effects. Moreover, current chemical criteria and critical limits should be critically evaluated and revised to include organic soils and receptor ecosystems other than forests.

2055

2635

8957

The existence of the SSMB is based solely on the need for a simple model that can be generally applied on a national scale with limited data. However, there are a number of limitations in the current formulation of the SSMB. Leaching of organic anions should be included, especially for organic soils. An aluminium᎐organic complexation model should also be included, to account for complexation of organic ions by aluminium. Holmberg et al. Ž2001. proposed tentative modification to the SSMB towards this. In addition, there are weaknesses in the gibbsite equilibrium relationship suggesting it may be more appropriate to use laboratory and field observations of soil solution chemistry to develop an empirical relationship between hydrogen and aluminium concentrations. Similar relationships have also been observed for organic soil horizons, which are generally under-saturated with respect to gibbsite. 3.3. Critical loads of acidity at the EMEP resolution The 5th, 50th and 95th percentile for maximum critical loads of acidity for sulphur and nitrogen are presented in Table 5 Žsee Fig. 2 for grid-square IDs.. The 5th percentile Žpentile. protects 95% of the total receptor area mapped within each EMEP grid-square, and has been used to represent the sensitivity in a grid-square under the UNECE CLRTAP ŽPosch et al., 1999.. The estimated pentile critical loads are lower than those previously estimated for Ireland ŽAherne and Farrell, 2001; Posch et al., 1999.. For example, the two most

J. Aherne, E.P. Farrell r The Science of the Total En¨ ironment 288 (2002) 183᎐197

194

Table 5 Fifth, 50th and 95th percentile for maximum critical loads of sulphur and nitrogen Žmol c hay1 yeary1 . for each 150 = 150-km EMEP grid Grid ID

1311 1312 1313 1314 1315 1411 1412 1413 1414 1415 1512 1513

Area km2

Percentile CLma x ŽS. 5th

50th

95th

5th

50th

95th

64 1500 1728 1726 49 26 1502 1963 165 2 1 232

1002 571 424 585 708 712 470 320 493 589 1325 461

1219 1699 1663 1119 886 1113 1248 1122 1396 824 1342 1278

2390 2785 5505 5181 1835 2308 4535 4701 4960 936 4626 4515

1302 899 789 965 996 1167 890 694 953 875 1541 920

1503 1990 1931 1400 1172 1377 1667 1477 1682 1110 1558 1689

2635 3058 5726 5447 2121 2594 4872 4975 5246 1222 4842 4823

sensitive grids Ž1313. and Ž1413. have pentile maximum critical loads for sulphur of 424 mol c hay1 yeary1 and 320 mol c hay1 yeary1 , respectively. Previously these grids were estimated to be 805 mol c hay1 yeary1 and 649 mol c hay1 yeary1 , respectively. Critical loads will invariably be smaller using a multi-criterion approach because the most sensitive criterion is selected. However, in general the critical loads should not change to a large extent compared to a single criterion approach, assuming reasonable limits were applied. The large increase in sensitivity in this application is due to the more complete inclusion of organic soils, which represented 27% of the mapped receptors. The lower pentile critical loads are more in line with those of the UK and Scandinavian countries ŽPosch et al., 1999.. The 5th, 50th and 95th percentile critical loads for each receptor ecosystem are presented in Table 6. In general, coniferous ecosystems had the lowest critical load, which was due to the high percentage occurring on organic soils Ž33%, Table 4.. In addition, coniferous forests were the only receptor that had a removal of base cations from harvesting. The lowest pentile maximum critical loads for sulphur, for the two most sensitive grids Ž1313. and Ž1413., are for coniferous forests at 374 mol c hay1 yeary1 and 415 mol c hay1 yeary1 , respectively.

Percentile CLmax ŽN.

4. Conclusions The modelling methodology presented in this paper is a simple approach to estimating critical loads of acidity using multiple criteria and receptors. Model inputs and critical limits for the range of recommended chemical criteria have been obtained using a combination of national data and default literature values. The most important chemical criteria, in the determination of critical loads of acidity for the Republic of Ireland, were based on a critical pH limit ŽwAlx crit estimated using K gibb . for mineral soils or a critical pH shift in relation to pristine conditions for organic soils. Approximately 68 and 26% of the final distribution of critical loads were estimated using these criteria, respectively. The pentile critical loads of acidity were more sensitive than those previously estimated for Ireland due to the more complete inclusion of organic soils. Furthermore, coniferous ecosystems had the lowest critical load, due to the high percentage occurring on organic soils and the removal of base cations through harvesting. The results demonstrate that it may be more appropriate to use multiple criteria and receptors to produce more realistic critical loads and ensure adequate protection of biological indicators. The approach is recommended for determining

J. Aherne, E.P. Farrell r The Science of the Total En¨ ironment 288 (2002) 183᎐197

national critical loads of acidity because it takes a broader range of semi-natural ecosystems into

195

account and sets the most sensitive limits for protection against damage. In addition, it at-

Table 6 Fifth, 50th and 95th percentile for maximum critical loads of sulphur and nitrogen Žmol c hay1 yeary1 . for each receptor ecosystem and 150 = 150-km EMEP grid Grid ID

Receptor

Area km2

Percentile CLma x ŽS.

Percentile CLmax ŽN.

5th

50th

95th

5th

50th

95th

1311

Coniferous Deciduous Grassland Heathland

2 5 9 48

871 1287 990 1016

910 2419 1282 1208

1274 2475 1852 2178

1396 1573 1276 1302

1435 2635 1568 1493

1729 2756 2138 2464

1312

Coniferous Deciduous Grassland Heathland

105 140 422 832

365 608 562 728

984 2141 1714 1697

2770 4923 2835 2520

801 894 848 1014

1509 2427 1969 1980

3225 5207 3055 2776

1313

Coniferous Deciduous Grassland Heathland

388 660 508 172

374 559 484 502

681 2825 1827 1551

4811 5576 5546 5419

767 845 770 788

1113 3041 2043 1767

5291 5792 5762 5635

1314

Coniferous Deciduous Grassland Heathland

330 113 556 729

405 1007 627 757

803 1802 1074 1139

5008 5457 5322 2367

893 1293 913 1043

1326 2088 1337 1395

5465 5673 5538 2588

1315

Coniferous Deciduous Grassland Heathland

-1 1 7 41

473 857 602 717

473 1826 778 897

1614 1956 1625 1825

998 1143 888 1003

998 2112 1064 1183

2139 2242 1911 2111

1411

Coniferous Deciduous Grassland Heathland

2 5 4 15

712 1075 618 1009

735 1927 1160 1062

1765 2223 2045 2354

1167 1361 904 1247

1190 2213 1376 1348

2220 2509 2331 2640

1412

Coniferous Deciduous Grassland Heathland

689 212 249 351

415 1009 566 674

1102 1693 1272 1366

2432 4903 2798 2515

871 1242 852 960

1627 1966 1551 1649

2887 5129 3014 2731

1413

Coniferous Deciduous Grassland Heathland

746 571 244 401

216 419 420 493

941 1350 1185 1183

4401 4822 4586 2229

589 705 706 779

1463 1625 1460 1457

4888 5070 4809 2447

1414

Coniferous Deciduous Grassland Heathland

53 51 39 22

435 531 649 965

1208 1616 1116 1195

4856 4890 5086 1598

960 817 935 1181

1663 1832 1332 1411

5311 5173 5302 1814

J. Aherne, E.P. Farrell r The Science of the Total En¨ ironment 288 (2002) 183᎐197

196 Table 6 Ž Continued. Grid ID

1415

1512

1513

Receptor

Coniferous Deciduous Grassland Heathland Coniferous Deciduous Grassland Heathland Coniferous Deciduous Grassland Heathland

Area km2

Percentile CLma x ŽS. 5th

Percentile CLmax ŽN.

50th

95th

5th

50th

95th

1 1

589 701

814 930

824 936

875 987

1100 1216

1110 1222

1

1325

1342

4626

1541

1558

4842

137 53 16 25

381 918 866 687

1166 1539 1430 1316

4312 4708 4730 4515

906 1134 1082 973

1669 1755 1715 1602

4774 4924 5016 4801

tempts to elucidate which chemical criteria are setting the most sensitive limits. The choice of critical limits and model parameters will ultimately effect the critical loads. Therefore, it is important that appropriate critical limits are chosen to protect the biological indicators and receptor ecosystems from long-term damage.

Acknowledgements Part-funded by the Irish Environmental Protection Agency under the Environmental Monitoring R& D Sub-programme of the Operational Programme for Environmental Services. References Aherne J, Farrell EP. Assessment of the application of empirical critical loads of acidity for Irish soils. Irish J Agric Food Res 2000;39:349᎐358. Aherne J, Farrell EP, editors. Final report: determination and mapping of critical loads for sulphur and nitrogen and critical levels for ozone in Ireland. Dublin: Environmental Protection Agency, 2001:212. Aherne J, Farrell EP, Hall J, Reynolds B, Hornung M. Selecting chemical criteria for critical loads of acidity in maritime regions. Water Air Soil Pollut: Focus 2001;1:75᎐90. Anon, Irish National Forest Standard. Forest Service, Department of the Marine and Natural Resources, Dublin, 2000, p. 52. Bobbink R, Heil GW, Raessen MBAG. Atmospheric deposi-

tion and canopy exchange processes in heathland ecosystems. Environ Pollut 1992;75:29᎐37. COFORD, Pathway to progress, A programme for forest research and development, National Council for Forest Research and Development, University College Dublin, 1994, p. 132. Cresser MS. The critical loads concept: milestone or millstone for the new millennium? Sci Tot Environ 2000;249:51᎐62. Cresser MS, Smith C, Sanger L. Critical loads for peat soils. In: Hornung M, Skeffington RA, editors. Critical loads: concept and application, pp. 34᎐39 ITE Symposium No. 28. London: HMSO publications, 1993:134. de Vries W. Methodologies for the assessment and mapping of critical loads and of the impact of abatement strategies on forest soils. Report 46. The Netherlands: DLO The Winand Staring Centre Wageningen, 1991:109. Downing RJ, Hettelingh J-P, de Smet PAM, editors. Calculation and mapping of critical loads in Europe: status report 1993. The Netherlands: RIVM Bilthoven, 1991:163. RIVM Report No. 259101003. Emmett B, Reynolds B. Nitrogen critical loads for spruce plantations in Wales: is there too much nitrogen? Forestry 1996;69:205᎐214. Gardiner M, Radford J. Soil associations of Ireland and their land use potential. Soil survey bulletin No. 36. Dublin: An Foras Taluntais, 1980:143. Hall J, Reynolds B, Langan S, Hornung M, Kennedy F, Aherne J. Investigating the uncertainties in the simple mass balance equation for acidity critical loads for terrestrial systems in the United Kingdom. Water Air Soil Pollut: Focus 2001a;1:43᎐56. Hall J, Reynolds B, Aherne J, Hornung M. The importance of selecting appropriate criteria for calculating acidity critical loads for terrestrial ecosystems using the simple mass balance equation. Water Air Soil Pollut: Focus 2001b;1:29᎐41.

J. Aherne, E.P. Farrell r The Science of the Total En¨ ironment 288 (2002) 183᎐197 Hammond RF. The Peatlands of Ireland. Soil survey bulletin No.35. Dublin: An Foras Taluntais, 1981:60. Hettelingh J-P, Downing RJ, de Smet PAM, editors. Mapping critical loads for Europe: CCE technical report no. 1. The Netherlands: RIVM Bilthoven, 1991:86. RIVM Report No. 259101001. Holmberg M, Mulder J, Posch M, Starr M, Forsius M, Johansson M, Bak J, Ilvesniemi H, Sverdrup H. Critical loads of acidity for forest soils: tentative modifications. Water Air Soil Pollut: Focus 2001;1:91᎐101. Hornung M, Sutton MA, Wilson RB, editors. Mapping and modelling of critical loads for nitrogen: a workshop report. UK: Institute of Terrestrial Ecology Merlewood, 1995:207. Jonsson C, Warfvinge P, Sverdrup H. Uncertainties in predict¨ ing weathering rate and environmental stress factors with the PROFILE model. Water Air Soil Pollut 1995;81:1᎐23. Langan SJ, Reynolds B, Bain DC. The calculation of base cation release from mineral weathering in soils derived from Palaeozoic greywackes and shales in upland UK. Geoderma 1996;69:275᎐285. Langan SJ, Sverdrup HU, Coull M. The calculation of base cation release from the chemical weathering of Scottish soils using the PROFILE model. Water Air Soil Pollut 1995;85:2497᎐2502. Løkke H, Bak J, Falkengren-Grerup U, Finlay RD, Ilvesniemi H, Nygaard PH, Starr M. Critical loads of acidic deposition for forest soils: is the current approach adequate? Ambio 1996;25:510᎐516. Løkke H, Bak J, Bobbink R, Bull K, Curtis C, FalkengrenGrerup U, Forsius M, Gundersen P, Hornung M, Skjelkvale ˚ BL, Starr M, Tybirk K. Critical Loads Copenhagen 1999, 21᎐25 November 1999. Conference report prepared by members of the conference’s secretariat, the scientific committee and chairmen and rapporteurs of its workshops in consultation with the UNECE secretariat, Critical loads, National Environmental Research Institute, Denmark, 2000, p. 48. Lovblad G, Andersen B, Joffre S, Pedersen U, Hovmand M, ¨ Reissell A. Mapping deposition of sulphur, nitrogen and base cations in the Nordic countries. Report B 1055. Stockholm: Swedish Environmental Research Institute ŽIVL., 1992, p. 33. Nilsson J, Grennfelt P, editors. Critical loads for sulphur and nitrogen. Miljorapport 1988, p. 15. Copenhagen: Nordic ¨ council of Ministers, 1988:418. OSI. 1:100 000 CORINE land cover project ŽIreland.. Phoenix Park, Dublin: Ordnance Survey of Ireland, 1993. OSI. The Irish grid: a description of the co-ordinate reference

197

system used in Ireland. Phoenix Park, Dublin: Ordnance Survey of Ireland, 1996, p. 41. Posch M, de Smet PAM, Hettelingh J-P, Downing RJ, editors. Calculation and mapping of critical thresholds in Europe: status report 1995. The Netherlands: RIVM Bilthoven, 1995:198. RIVM Report No. 259101004 Posch M, de Smet PAM, Hettelingh J-P, Downing RJ, editors. Calculation and mapping of critical thresholds in Europe: status report. The Netherlands: RIVM Bilthoven, 1999:165. RIVM Report No. 259101009 Reynolds B. An evaluation of critical loads of soil acidity in areas of high sea salt deposition. Sci Total Environ 2000;253:169᎐176. Ronald-Eastman J. Clark University MA, USA: Clark Labs, 1999:193. Smith C, Cresser MS, Mitchell DJ. Sensitivity to acid deposition of dystrophic peat in Great Britain. Ambio 1993;22:22᎐26. Sverdrup H, de Vries W. Calculating critical loads for acidity with the simple mass balance method. Water Air Soil Pollut 1994;72:143᎐162. Sverdrup H, Warfvinge P. The effect of soil acidification on the growth of trees, grass and herbs as expressed by the ŽCa q Mgq K.rAl ratio. Reports in ecology and environmental engineering 1993, p. 2. Lund University, Sweden: Department of Chemical Engineering II, 1993:177. Sverdrup H, de Vries W, Henriksen A. Mapping critical loads. Miljorapport 1990, p. 14. Copenhagen: Nordic Council of ¨ Ministers, 1990:124. Ulrich B. Interaction of forest canopies with atmospheric constituents: SO 2 , alkali and earth alkali cations and chloride. In: Ulrich B, Pankrath J, editors. Effects of accumulation of air pollutants in forest ecosystems, pp. 33᎐45. The Netherlands: Reidel Dordrecht, 1983:389. Ulrich B. Stability, elasticity and resilience of terrestrial ecosystems with respect to matter balance. In: Schulze ED, Zwolfer H, editors. Potentials and limitations of ecosystem analysis, Ecological studies 61. Berlin, Heidelberg: Springer-Verlag, 1987:11᎐49. Warfvinge P, Sverdrup H. Calculating critical loads of acid deposition with PROFILE: a steady state soil chemistry model. Water Air Soil Pollut 1992;63:119᎐143. Werner B, Spranger T, editors. Manual on methodologies and criteria for mapping critical levelsrloads and geographical areas where they are exceeded. Berlin: Federal Environmental Agency ŽUmweltbundesamt., 1996:144.