The need for reducffons in the emission of sulphur and nitrogen compounds tothe atmosphere in order to protect ecosystems ftom acidfficafion is now well established and has led to a series of international agreements under the auspices of the UN ECE andtheEU (Jenldns 1999). The most recent of ffiese agreements,theso called
“multi pollutant, muffi-effect” protocol signed in Goffienberg in 1999, utilisesthe crfflcal loads concept to target redudflons in sulphur, nitrogen and volaffle organic compounds, towards the most sensifive ecosystems within Europe. The crffical load is defined as the level ofacidicdeposifion below which damage to a particular part oftheecosystem does not occur. These are mapped throughout EuropeS0that effort can be focused on those areas which are atthegreatest envfronmental risk, ensuring thatthemost cost effecfive control strategy is implemented.
Crfficalloads for addfficafion have commonly been calculated using empiricafly based steady state modeis. This meffiod, however, provides only a measure of the riskfrom addfficaflon at present compared to some equilibrium time in the future, and ffierefore, gives no indication ofthetime over which chemical recovery will be acMeved in response to reduced addic oxide emissions. In addffion, such an approach is unable to aid assessment of the impact of future land use changes on water chemistry, such as de-forestation and re-plantation. It is important to be able to predict the behaviour of sensitive ecosystems to present and future levels of acidifying pollutants so as to allow an assessment of possible timescales over which any emissions reduction might be undertaken. Choice of timescale is important both from the point of view of limfflng damage to sensifive ecosystemsand also from the point of view of not pladng an unrealisfic burden onindustry.
The only way to assessthetime-dependence of crfflcal loadsand thepotential damage to areas where cdfical loads remain exceeded in the future is through dynamic models. This chapter describes application of the dynamic acidfflcafion model, MAGIC (Acidificafion of Groundwater in Catchments) to catchments in Wales, including the Alon Haften (GBO2)—one ofthe UKICP IM sites.
4.2 The MAGC Model
MAGIC is a process-oriented dynamic model forthelong term reconstruction and future prediction of soil and surface water addffication at the catchment scale (Cosby et aL1985). The model consists of; (i) soil-soil solufion equffibria equations in which
The Finnish Enronment 427
0
the chemical composffion of the soil solution is assumed to be governed by simultaneous reactions involving S04 adsorption, cation exchange, dissolution and speciation of inorganic and organic carbon; and (ii) mass balance equations in which fluxes of major ions to and from the soil and surface water are assumed to be governed by atmospheric inputs, mineral weathering, net uptake by biomass and loss to streamwater. MAGIC uses a Iumped parameter approach to; (i) aggregate the complex chemical and biological processes active at the catchment scale into a few readily described processes; and (ii) lumped catchment characterisfics to represent the spatia! heterogeneity of soil properties throughout the catchment.
Dynamic simulation of soil and stream water chemistry is achieved by coupling the equffibna equations with the dynamic mass balance equafions for each of the major cafions. Calibration of the model involves matching present day observed soil and streamwater base cation data through adjustment of base cation weathering rates and soi! selectivity coeffidents. In addition, soil C/N rafios are used to match observed NO3 leaching.
4.3 MuItipIe Site application of MAGIC to Wales
With the increasing recognifion of the importance of dynamic modeis in assessing the fime-dependence of critical loads and the potential damage to areas where crffical loads remain exceeded in the future, has come the desire to extrapolate model applications ftom individual catchments to whole regions. Such a development potentially provides policy makers with a spatia! resolution capable of better representing the impacts of emission reduction strategies and a statisfical basis on which to formu!ate economic assessments.
In the UK, the Welsh Add Waters Survey (WAWS: Stevenseta! 1997)undertaken in 1995 has provided a framework with which to undertake a regional, mu!fiple site application of MAGIC to 102 forested and moorland catchments in Wa!es. Rainfali, runoff, stream and soil chemistry, lanä use and atmospheric deposition data for each site were denved from a variety of sources at different spatial and tempora!
scaies.
Bulk deposihon samp!es were coilected during1995from 19 sites across the region and the data interpolated to provide wet deposition on a 20km grid (Stevens et al 1997). These va!ues were then enhanced to account for the variafion ofrainfali and jon concentrations in rainfail with aifitude. Within MAGIC, chloride (Cl) is treated as conservative and is assumed to be in steady state with respect to input output flux at each time step. Consequent!y, at moorland catchments the wet deposited C! input flux, derived from the 20km deposition grid, is compared to the stream output flux at each site and inputs are further enhanced as necessary assuming the difference to representdiydeposition. Suiphate ($04) is aiso assumed to be in steady-state at present, implying no current $ adsorption within the catchment systems, and the deposifion flux is enhanced to representdrydeposition of anthropogenic S to the catchment as SO,. The increased deposition flux attributable to the ability of the tree canopy to filter pollutants ftom the atmosphere isalsoknown to promote bigher stream output fluxes ($tevens eta! 1997). At forested sites, therefore, the extra deposifion required to ba!ance the output fluxes of Cl and
$04 represents both the dry deposifion enhancement assumed for moorland sites as well as the extra flux from the forest ffltering effect.
The cumulative effect of atmospheric deposifion since pre-industrial times is modeiled on anannuaitime step and is driven by changes in anthropogenic S and N emission over time. The histonca! trend in wet deposited non-marine $04 is assumed to foliow the sequence described by the UK Warren Spring Laboratory (DOE 1983, 1990). This sequence reflects the pattern of increased industrial growth from the onset of industrialisation in the mid to late nineteenth century. Nitrate (NO3) and
0
The Finnish Enwonment 427ammonium (NH4) in deposifion are assumed to have increased since industrialisafion in accordance with the estimated NO emissions in the UK (DOE 1983, 1990).
Three processes in MAGIC simulate the impact of afforestation on addffica%on of soils and surface water: (i) enhanced dry and occult deposition (Mayer and Ulrich 1977); (ii) ion uptake by growing forests (Mifier 1981); and (lii) decreased water yield concentrating pollutants in surface waters (Neal et al 1986). Uptake, enhanced deposition and sfream discharge are ali spedfied at each time step and are calculated with respect to forest age and the percentage of catchment covered. Of the WAWS sites, 57 are forested and the planting year and spafial coverage of each stand at each site was provided by the Forestry Enterprise (pers. comm).
Stream sampies were collected montffly during the WAWS through 1995 and analysed for ail major ions, annual geometric means were used for the model calibration procedure. fifteen sites have a mean acid neutralising capacity (ANC) of less than zero, indicating that many streams are fflcely to suffer adverse biological impacts in response to present day chemistry. Catchment runoff is not recorded at each site but is derived from estimates of ET and rainfail. It is estimated that ET ranges from 10% for a moorland catchment to 20% for a fully forested catchment.
Soil cation exchange capadty, bulk density and exchangeable base cation fractions were determined from each site ftom a database for England and Wales compiled by the Soil Survey and Land Research Centre.
4.4 Predicted future response
Once a successful calibration was achieved at each site, the model was used to predict future chemical recovery under the Gothenburg Protocol. Emissions under this agreement have been converted to a regional deposifion scenario at a 10km grid scale using the Huil Add Rain Model (HARM; Metcalfe and Whyatt 1995). In
The Finnish Enveonment 427
0
1999
ueqll
<0 0 to25 25 to 50 5Oto100
>100
2045
0
Figure 4. 1 ANC recover assessment of the Gothenburg scenario
addffion, afutureforest management scenario was incorporated for each forested site whereby each stand is felled at 50 years age, and immediately replanted such that the forest area remains constant.
The predicted response within the next 45 years, relative to present day, is generally a small recovery in surface waterANC across the whole region (Figure 4.1).Onlyone site is predicted to have anANCof less than zero by 2045. The more addified, lowANCsites are generally located in the central and northem parts of the region but no marked spatial pattem can be detected. This probably reflects the similarityin predicted decrease irL deposifion aaoss the region and spatial variation in the extent and age of forestry. A clear difference is predicted between recovery at moorland and forested sites: MeanANCrecovery by 2045 at moorland sites is 26 teq/1 in comparison to 16 teq/1 at forested sites, and, despite the emissions reductions, 3 forested sites are predicted to undergo a further drop inANC.Such multiple site assessments of recovery may necessitate the use of coarse datasets but provide a useful visual and stafisfical means of determining the impact of emissions reduction at a regional scale.
4.5 Site Specific Application 0fMAGIC to the Afon Hafren (GBO2)
The streamwaters of the Afon Hafren are addic, and characterised by a mean pH andANCof 5.4and 6 ieq/1respectively. Ihis is a reflecfion of the addic bedrock and soils, coupled with a significant flux of acidic oxides in deposition and accentuated by the growth of forests. future recovery at the site, predicted by MAGIC under the Gothenburg Protocol (Figure 4.2), indicates that recovery m streamwaterANC will be modest, rising only slightly from present day values. During the forecast period, the site undergoes cycles of transient recovery as trees are felled, followed by addification as stands grow and uptake and forest ifiter deposifion increase.
This pattern is superimposed upon a number of stands of different ageswithinthe catchment and the predicted response is therefore complex. It is clear however, that ecosystem recovery from aadification wffl be limited at GBO2 even under the emissions reducfions of the latest protocol. Futhermore, it is important to note that the prediction of mean annualANCdoes not account for acidic pulses observed at this, and other sites, during high winter flows. These pulses may cause significant effects on freshwater biota whilst the mean annual ANC indicates less cause for concem. The fflusfrated time series also shows that significant addification has occurred from pre-industrial times, reaching a minimum sfreamwaterANC of -31ieqI1 in 1980, pnor to the impiementafion of emissions protocols.
The site appllcafion of MAGIC to GBO2 (Figure 4.2) fflustrates the key ufflity of dynamic modeis in acidffication studies. In contrast to empirical methods for calculafing cri%cal loads, the fimescale for hydrochemical recovery is clearly observable underfuturedeposition and Iand use scenarios. As a consequence, the time taken, and cost to industry to achieve a certain target chemistry, for example, zeroANCto protect brown trout, can be readily assessed.
0
The Finnish Environment 427Acknowledgements
This research has beenfunded by the Department of the Environment, Transport and the Regions (Project No. 1/3/133)
References
Cosby, B.J., Hornbergei, G.M., Wright, R.f and Galloway, J.N. 1985. Modeffing the effects of add deposifion: Assessment of a lumped parameter model of soil water and
sfreamwater chemistry Water Resources Research, 21, 51-63
Department of the Environment (DOE) 1983 Acid deposffion m the United Kmgdom Warren Spnng Laboratory Stevenage, UK
Department of the Environment (DOE) 1990. Aud deposifion m the United Kingdom 1986-1988. HMSO, London, UK
Jenkins, A. 1999. End of the acid reign ? Nature, 401.
Mayer, R. and Ulrich, B. 1977. Acidity of predpitation as influenced by the ffltering of atmospheric 5 and N compounds-its role in the element balance and effect on soil.
Water, Afr and Soil Pollufion, 7,409-416.
Metcalle, S.E. and Whyatt, JD. 1995. Modelling future add deposifion with HARM. In: Acid rain and its impact: the crifical loads debate (Battarbee, R.W (Ed)), 27-37. ENSIS Publishing, London.
Miller, H G 1981 Forest fertilisation-some guidmg concepts forestry 54 157-167 Neal, C., Whitehead, PG., Neale, R, and Cosby, B.J. 1986. Modelling the effects of acidic
depos;fion and conifer afforestation on stream acid;ty m the Bribsh uplands Journal of Hydro1og 86, 15-26.
Stevens, PA., Ormerod, 5,1. and Reynolds, B. 1997. Final Report on the Add Waters Survey for Wales, Volume 1, Mam Text (ITE Project T07072R5), 224pp.
The Rnnish Envwonment 427 120
100 80
0z 60 40 20 0
1895 1925 1955
-40
2015 2045
Figure 4.2 MAGIC predicted streamwater ANC under the Gothenburg Protocol.