Nutrient Changes in Peaty Gley Soils after Clearfelling of Sitka Spruce Stands

The roots of second-rotation Sitka spruce (Picea sitchensis(Bong.) Carr.) planted on peaty gley sites are restricted tothe old litter (LFH) layer and are dependent on its decompositionfor availability of nutrients. A series of these sites of increasingage from felling were sampled to estimate changes in the nutrientcapital of the LFH horizon over time at Kielder Forest, Northumberland.Previous stand histories were reconstructed from stump data.Geographical, climatic, soil and mensurational data suggestedthat the use of a time series was justified. Nutrient capital in the LFH horizon generally declined overa 5 year period after clearfelling from approximately 997, 51and 83 kg ha^–1 to 676, 30 and 31 kg ha^–1 of N, Pand K, respectively. However, N concentration increased overa 5 year period from 11 mg g^–1 to 14 mg g^–1, P concentrationremained constant at about 0.6 mg g^–1, and K concentrationdecreased from 1.0 mg g^–1 to 0.7 mg g^–1. Nutrientconcentrations and contents of the LFH horizon were higher underthe brash (slash) swathes that resulted from the use of organizedfelling techniques than under clear strips devoid of brash. The patterned input of nutrient capital in brash as a resultof organized felling was also determined. Brash containing 219,20 and 71 kg ha^–1 of N, Pand K, respectively, was systematicallydistributed at a rate of 491 ha^–1 over 66 per cent ofthe site after harvesting. The needles and small branch fractionscontained 71 per cent of the N and 80 per cent of the P andK present in the brash.     

Effects of Refertilization on Growth and Nutrition of Lodgepole Pine (Pinus contorta Dougl.) Planted on a Minerotrophic Peatland in Central Newfoundland, Canada

A plantation of lodgepole pine (Pinus contorta Dougl.), whichwas established in 1973 on a ditched minerotrophic fen at plantingspacings of 1.2 m, 1.8 m, and 2.4 m, was refertilized in 1985with control, P, PK and NPK treatments consisting of 60 kg ha^–1phosphorus, 100 kg ha^–1 potassium and 200 kg ha^–1nitrogen. Although each seedling had been spot-fertilized in1973 with a mixture of 71 g urea, 99 g ground rock phosphateand 41 g potassium sulphate, height and needle nutrient assessmentsof the plantation in 1984 indicated poor growth, low concentrationsof needle nitrogen (1.26 per cent) and possible deficiency levelsof needle phosphorus (0.10 per cent) and potassium (0.36 percent). Between 1985 and 1991, height and diameter growth increasedsignificantly in all refertilization treatments. Growth responsewas greatest in the PK treatment and least in the P treatment.Needle weights, determined from 1987 to 1991, did not responduniformly to any of the refertilization treatments. Needle concentrationsof phosphorus increased to levels between 0.20 per cent and0.23 per cent following refertilization with P, PK, and NPK,but by 1988 had decreased to about 0.14 per cent. Needle concentrationsof potassium also increased sharply to 0.60 per cent in 1985following refertilization with PK and NPK, but by 1988 had decreasedto levels between 0.35 per cent and 0.45 per cent. Similarly,needle nitrogen concentrations increased to 1.84 per cent in1985 following refertilization with NPK, then decreased to levelsbetween 1.20 per cent and 1.33 per cent in 1986 and remainedat those levels each year thereafter. Although needle nutrientlevels fell sharply following an initial peak after refertilization,height and diameter growth did not decrease significantly, especiallyin the PK treatment, suggesting that nutrient levels remainedadequate for optimum growth.     

The Interaction of Green Spruce Aphid and Fertilizer Applications on the Growth of Sitka Spruce

In an 11-year-old stand of Sitka spruce (Picea sitchensis [Bong.]Carr.) application of nitrogen fertilizers, at a rate of 10kgN ha^–1 month^–1, increased mean diameter incrementby 12 per cent, while the further addition of phosphorus, at5 kg ha^–1 month^–1, resulted in a 23 per cent increase.An attack by the green spruce aphid (Elatobium albietinum Walker)occurred during the period of fertilizer addition. The mostseverely affected trees showed a reduction in diameter growthof 50 to 56 per cent but the severity of the attack betweentrees was unrelated to the treatments applied. However, fertilizerapplication did hasten the recovery of diameter growth afterdefoliation.     

Changes in Amount and Distribution of Stem Growth in Pole-Stage Corsican Pine Following Application of Nitrogen Fertilizer

Ring-width measurements on trees felled in a nitrogen-deficientpole-stage crop of Corsican pine on the sand dunes of Culbinforest (Morayshire) showed that, in addition to a growth checkascribed to the low nitrogen content of the soil, there wasevidence for decreasing site quality, tree growth decliningprogressively with time. This feature is ascribed to the continuingimmobilization of the limited nitrogen capital in the newlydeveloping mor humus layer. Ammonium sulphate, at rates equivalentto 84, 168, 336, and 504 kg elemental nitrogen ha^–1, wasapplied to this crop when it was 36 years old, and the treatmentsrepeated for a further 2 years. Nitrogen levels in the foliagerose from 0.9 per cent to as high as 20 per cent in the firstyear, increasing to 2.6 per cent after the third year of application.Response in terms of basal area appeared in the year after firstapplication but was not accompanied by any changes in the patternof growth throughout the season, nor was there a significanteffect of tree size on the proportional change in relative basalarea growth induced by the fertilizer, except for a slightlyreduced response by the smallest trees to the lower rates ofapplication. A similar low response by the smallest trees wasfound for height, but, in addition, height growth after treatmentwas greatest for the middle-sized trees, despite the fact thatthese had been growing significantly more slowly than the largesttrees prior to application of the fertilizer. Form factor remainedunaltered by the treatments but taper changed slightly. Heightand basal area showed very different response patterns to thetreatments. Thus, whereas maximum height growth occurred atthe 168 kg N ha^–1 treatment rate (2.41 m over 7 years,1.4 times the control growth) maximum basal area growth occurredat the 504 kg N ha^–1 rate, the highest rate used (13.0m² ha^–1 over seven years, 2.4 times the control growth);volume growth was maximum at the 336 kg N ha^–1 rate (126m³ ha^–1 over 7 years, 2.6 times the control growth). Whenrelated to levels of foliar nitrogen in the previous year, annualheight growth was maximum at 1.6 per cent with maximum basalarea growth occurring nearer 2.2 per cent. It is estimated thatvolume growth would be maximised at just over 20 per cent nitrogen.     

Nutrient Leaching from the Litter Layer after Clearfelling of Sitka Spruce Stands on Peaty Gley Soils

The nutrient dynamics in the litter layer of former Sitka sprucestands were examined from 0 to 7 years after clearfelling usingzero-tension lysimeters on a time series of sites 0, 2 and 5years after felling. The loss of most nutrients monitored inleachate was independent of the time from clearfelling. However,54 per cent (43 kg ha^–1) of the 7-year net loss of potassiumwas leached out of the litter layer in the first year. Also,nitrate losses, although small, increased after 5 years. Leachinglosses of NH₄-N, PO₄-P, K, Ca and H⁺ exceeded precipitationinputs. However, leaching losses were less than precipitationinputs for NO₃-N, Na and Mg. Nutrient losses from under brashswathes were generally greater than from clear strips.     

Litter Production in Western Washington Douglas-Fir Stands

Litter production by Douglas-fir stands ranging in age from22 years old to 160 years old, is discussed. Typical leaf litterproduction was 2100 kg ha^–1 yr^–1 while total litterwas 2500 kg ha^–1 yr^–1. Annual fall of leaf litterincreases up to about 40 years of age and then becomes fairlyconstant while total litter continues to increase because ofwood production, although this increase may be quite irregular.Average nutrient returns to the forest floor are 21, 3, 7, 32,4, and 7 kg ha^–1 yr^–1 for N, P, K, Ca, Mg, and Mnrespectively.     

The Internal Transfer of Nutrients in a Scots Pine Stand 2. The Patterns of Transfer and the Effects of Nitrogen Availability

The fall of fine litter was 4.5 tonnes ha^–1 y^–1,(78% needles) returning to the ground 36.6 kg N, 2.2 kg P, 7.6kg K, 12.1 kg Ca and 2.0 kg Mg. Nutrient input in rainfall (880mm in 1978) was 2.6 kg N, 0.4 kg P, 4.3 kg K, 7.6 kg Ca and1.4 kg Mg. Throughfall was 534 mm. The method of Miller, Cooperand Miller (1976) was used to estimate aerosol contributions,leaching and absorption. Retranslocation estimates were 59.6kg N ha^–1 y^–1 (55% of Requirement), 8.5 kg P (64%R), 42.0 kg K (56% R) and 1.2 kg Mg (14% R). Retranslocationwas unchanged, but uptake increased with added N % P and decreasedwith added sawdust and glucose. Other typical effects of nutrientamendment were observed: with added N % P growth increased,current tissues had higher concentrations of N and P and litterfalldecreased in the first year but was heavier in the second year:with added sawdust and glucose, growth declined but concentrationsof N and P remained substantially the same.     

The Application of Fertilizers to Drained Peat 2. Uptake by Vegetation and Residual Distribution in Peat

Analyses of peat cores taken from lysimeters, which had beenused to monitor leaching losses from a drained peat on a raisedbog for three years, showed that much of the applied fertiliser,rock phosphate and potassium chloride, was retained in the uppermost0.3 m of the peat. Undisturbed bog vegetation had accumulated10–13 per cent of the applied P and K and losses to drainagewater amounted to about 10 and 35 per cent respectively. Whenboth nutrients were applied together leaching losses were reduced,vegetation uptake was increased and more was retained in thepeat. Within the fairly wide 95 per cent confidence limits mostof the P and Ca could be accounted for but K showed a deficit.The Ca in the rock phosphate was almost entirely retained inthe peat.     

Nitrogen critical loads for spruce plantations in Wales: is there too much nitrogen?

We have used the mass balance approach for calculating nitrogencritical loads (CL(N)) to avoid eutrophication for Sitka spruceplantation forestry in Wales. The various approaches for assigningvalues to the parameters in the mass balance equation are discussedwith particular reference to the soil nitrogen immobilizationvalue. A CL(N) value of 11 kgN ha^–1 a^–1 was calculatedfor an intensively studied site in Wales of Yield Class 14 ona freely draining acid soil. If this site is assumed to representa typical spruce stand, application of the CL(N) value meansthat 97 per cent of the area of coniferous forest in Wales,which is predominantly Sitka spruce, is currently receivingnitrogen deposition in excess of the CL(N). The area of coniferousforest at risk is reduced to 72 per cent if the proposed empiricalCL(N) for managed acidic coniferous forests to prevent ecologicalchanges (10–20 kgN ha^–1 a^–1) is applied andto 45 per cent if the empirical CL(N) to prevent nitrogen saturation,nitrate leaching and depletion of soil base cations is applied(10^–25kgN ha^–1 a^–1). Irrespective of the choiceof CL(N) values, the implications of critical load exceedanceneed urgent investigation. Available information at presentindicates that the main known consequence of chronic atmosphericnitrogen deposition to coniferous forest ecosystems is enhancednitrate and associated aluminium leaching to freshwaters. Thereis insufficient information regarding the potential adverseeffects of eutrophication of soils and waters and of impactson tree health and production.     

Productivity of Sitka Spruce in Northern Britain 2. Prediction from Site Factors

Productivity and site data from 187 temporary sample plots wereanalysed by multiple regression analysis to derive models inwhich site variables accounted for 78–86 per cent of thevariation in Sitka spruce productivity (General Yield Class,GYC). Climatic variables (accumulated temperature and windiness)extrapolated from meteorological data and tatter flag resultsaccounted for up to 78 per cent of the variation, the contributionof edaphic factors being small. The best regression models wereassociated with confidence limits of about ± 2.5 m³ ha^–1y^–1 and the mean error for predicting GYC for a forestblock (acquisition) was calculated to be ±1 m³ ha^–1y^–1 These figures were confirmed by the results of a validationsurvey and the application to field prediction of productivityis described.     

Carbon storage and sequestration in the forests of Northern Ireland

The rate of accumulation of carbon in forests and woodlandsin Northern Ireland was estimated using the record of forestplanting since 1900 and a model that calculated the flow ofcarbon from the atmosphere to trees, litter, soil, wood productsand back to the atmosphere. It was assumed that all coniferforests had the carbon accumulation characteristics of Piceasitchensis, and upper and lower estimates of carbon storagewere calculated assuming Yield Class 16 m³ha^–1 a^–1unthinned and Yield Class 14 m³ ha^–1 a^–1 thinned.Broadleaved woodlands were assume to have the carbon accumulationcharacteristics of Fagus sylvatica, Yield Class 6 m³ha^–1a^–1. Northern Ireland currently has about 78 300 ha offorest, 83 per cent of which is coniferous, 77 per cent state-owned,mostly planted since 1945, with peak planting in 1960–1975.In 1990, conifer forests contained 3–4 MtC (trees + litter)and broadleaved wdlands contained about 0.8 MtC (trees + litter+ new forest soil). In 1990, conifer forests were sequestering0.15–0.20 MtC a^–1 and broadleaved woodlands about0.025 MtC a^–1. To maintain these sink sizes, new coniferforests need to be planted at 1500–2000 ha a^–1,and new broadleaved woodland at100–150 ha a^–1 inaddition to full restocking. Current carbon sequestration byNorthern Ireland forests represents around 6.5–8.2 percent of the total for UK forests and is greater per hectar thanin Britain because the average forest age is younger in NorthernIreland     

Forest biomass estimated from MODIS and FIA data in the Lake States: MN, WI and MI, USA

This study linked the Moderate Resolution Imaging Spectrometerand USDA Forest Service, Forest Inventory and Analysis (FIA)data through empirical models established using high-resolutionLandsat Enhanced Thematic Mapper Plus observations to estimateaboveground biomass (AGB) in three Lake States in the north-centralUSA. While means obtained from larger sample sizes in FIA datasetscan be used as reference numbers over large scales, remote sensing(RS)-based observations have the ability to reflect spatialvariation of properties of interest within a given area. Thus,combining two national on-going datasets may improve our abilityto accurately estimate ecological properties across large scales.Using standard and consistent data sources can reduce uncertaintyand provide more comparable results at both temporal and spatialdimensions. We estimated total forest AGB in the region was1479 T_g (10¹² g, dry weight) in 2001 with mean AGB value of95 mg ha^–1 ranging from 4 to 411 mg ha^–1 (within95 per cent percentiles). Mixed forests featured 66 per centof the total AGB while deciduous and evergreen forests contained32 and 2 per cent of the total AGB, respectively, at 1-km pixelresolution. Spatially, AGB values increased from north-westto south-east in general. The RS-based estimates indicated agreater range in AGB variations than the FIA data. Deciduousforests were more variable (both in absolute and relative terms)than evergreen forests. The standard deviation of AGB for deciduousforests was 137 mg ha^–1, or a coefficient of variationof 92 per cent, that for evergreen forests was 24 mg ha^–1,or a coefficient of variation of 37 per cent.     

Nitrogen Transformations and Decomposition in Litter and Humus from beneath Closed-Canopy Sitka Spruce

Samples of litter and humus from beneath 10 m tall, closed-canopySitka spruce planted on a brown forest soil were incubated underboth field and laboratory conditions to measure mineral nitrogenproduction and carbon dioxide evolution. Mineral nitrogen productionin enclosed samples over 12 months was equivalent to 50 and17 kg N ha^–1 in litter and humus, respectively. Applicationsof fertilizer NPK (200 kg N ha^–1 as ammonium nitrate,100 kg P ha^–1 as unground rock phosphate and 150 kg Kha^–1 as potassium chloride), 18 months previously, decreasedthese values slightly, but stimulated the production of nitratein both litter and humus. Compared with samples kept under laboratoryconditions at 10°C, those incubated in the field at a similarmean temperature released less carbon dioxide and, in the caseof fertilized humus produced smaller amounts of mineral nitrogen.     

Response of Pole-stage Sitka Spruce to Applications of Fertilizer Nitrogen, Phosphorus and Potassium in Upland Britain

Although there are many reports of growth responses to fertilizerN, P or K in young stands, and to fertilizer N in old stands,there are relatively few reported responses in pole-stage ormiddle-aged stands. Among reasons for this may be lack of experimentalinterest in stands of this age or a relatively scarce occurrenceof a need for additional fertilizer inputs at a time when nutrientcycling, both within the tree and through the litter layer,is very efficient. To explore this, fertilizers were appliedto six stands of spruce, aged 25 or 30 years, in contrastingregions of Scotland and North England. In four experiments nogrowth response was recorded, either to a single dressing ofup to 400 kg N ha^–1, 200 kg P ha^–1 or 300 kg K ha^–1,or various combinations of levels of these elements, or to subsequentheavier applications. In a further experiment on peat therewas a weak response to P. In the remaining experiment a short-termresponse to PK was recorded but it is suggested that this wasdue to transient nutrient stress as the trees recovered fromthinning. The pattern of responses, other than the last mentioned,accorded with predictions based on foliar analysis. Taken togetherthese results seem to confirm the supposition that efficientnutrient cycling in middle-aged stands means that fertilizerresponses are unlikely (but not impossible) at this stage.     

Carbon stocks and sequestration in plantation forests in the Republic of Ireland

The United Nations Framework Convention on Climate Change andits Kyoto Protocol (KP) have created a clear need for methodsthat enable accurate accounting of carbon (C) stocks and stockchanges in forest ecosystems. The rate of accumulation of Cin plantation forests in the Republic of Ireland was estimatedfor the period 1906–2002 using the record of afforestationand a dynamic C accounting model (C-flow). Projections for theperiod 2003–2012 were made using different afforestationrates. It was assumed that Sitka spruce planted in the period1906–1989 was Yield Class (YC) 16 m³ ha^–1 year^–1and that after 1990 this increased to 20 m³ ha^–1 year^–1.All other conifers were assumed to have the growth characteristicsof YC 8 m³ ha^–1 year^–1 lodgepole pine. Broadleaveswere assumed to have the growth characteristics of YC 6 m³ ha^–1year^–1 beech. In 2002, the total forest C stock was 37.7Mt C representing an increase of 14.8 Mt C since 1990. In 2002,the rate of increase in trees, products, litter and soil was0.7, 0.1, 0.1 and 0.5 Mt C, respectively. Under a business-as-usualscenario, afforestation since 1990 is estimated to create anannual average C sink of 0.9 Mt C year^–1 during the period2008–2012. This accounts for 22 per cent of Ireland'sreduction commitment under the KP. Afforestation on peat soilswas found to reduce the net C sink, although the extent to whichit does so is highly dependent on assumptions regarding therate of peat C loss.     

Carbon pools and sequestration in forest ecosystems in Britain

British vegetation is estimated to contain 113.8 million tC,80 per cent of which is in forests and woodlands (91.9 milliontC). Sitka spruce plantations, although covering 21.4 per centof the forest/woodland area, contain only 8.2 per cent of theforest/woodland carbon, because the plantations are young andhave an average of only 14.1 tC ha^–1. Broadleaved woodlandsin Britain have an average of 61.9 tC ha^–1 and contain46.8 per cent of the total carbon in all vegetation. A breakdownis given of the carbon density (tC ha^–1) and content ofdifferent tree species. A carbon density map of Britain highlightsthe concentration of carbon in the broadleaved woodlands insouthern England and in the large conifer plantations in southernScotland and northern England. Carbon storage in the trees, products, litter and soil can beevaluated in terms of long-term equilibrium storage or short-termrate of storage. These two components vary among forest typesin Britain and globally. Plantations harvested at the time ofmaximum mean annual increment (MAI) will not store as much carbonas mature, old-growth forests on the same site unless they havelong-lasting products and/or are very fast growing. Maximumequilibrium carbon storage is generally achieved by harvestingat the time of maximum MAI when the lifetime of products exceedsthe time to maximum MAI. Undisturbed peatlands sequester CO₂and emit CH₄, and may be greenhouse neutral. When peatlandsare drained and planted with trees, they stop emitting CH₄ andstore carbon in the trees, forest litter, forest soil and woodproducts. However, these greenhouse gas ‘gains’are offset by the oxidation to CO₂ of the peat, and the gainsare exceeded by CO₂ losses when 20–40 cm depth of peathas been oxidized. Forests in Britain are currently sequestering1.5–1.7 million tC a^–1 in trees, 0.3–0.5 tCa^–1 in litter and 0.5 million tC a^–1 in wood products,totalling about 2.5 million tC, equivalent to about 1.5 percent of the carbon currently emitted by burning fossil fuelsin the UK. In order to maintain the current forest carbon sink,the forest area needs to continue to expand at about 25 000ha a^–1 of upland conifers or 10 000 ha a^–1 of poplarson good land.     

The tolerance of young trees to applications of clopyralid alone and in mixture with foliar-acting herbicides

The selective herbicide clopyralid is often used to controlcompeting Cirsium arvense in newly planted woodlands. When appliedas an overall spray at different dates in the spring (at 0.2kg acid equivalent (a.e.) ha^–1) to 10 tree species (Fraxinusexcelsior, Prunus avium, Quercus robur, Acer pseudoplatanus,Populus x canadensis cv. ‘Ghoy’, Pseudotsuga menziesii,Pinus nigra ssp. laricio, Larix kaempferi, Picea abies and Piceasitchensis) it did not reduce survival, and had little effecton growth. However, some species showed distortion of the youngestsprayed leaves or needles for several weeks after treatment,particularly F. excelsior, L. kaempferi and P. x canadensis.Sequential applications of clopyralid (first at 0.1 kg a.e.ha^–1 followed by 0.2 kg a.e. ha^–1 after 3 weeks),which are often required to control C. arvense, did not leadto increased leaf damage or growth reduction. Mixtures of clopyralidwith selective graminicides (cycloxydim at 0.45 kg active ingredient(a.i.) ha^–1; fluazifop-p-butyl at 0.38 kg a.i. ha^–1and propaquizafop at 0.15 kg a.i. ha^–1) did not causesignificant adverse effects on survival or growth of any species.If herbicides are required to control mixed stands of susceptibleproblem weeds such as C. arvense and grasses which are overtoppingyoung trees, these herbicide mixtures, applied as overall sprays,are less likely to cause damage to trees than attempts to usedirected applications of broad-spectrum foliar-acting herbicides.     

Characteristics of lowland broadleaved woodland being restocked by natural regeneration

A survey of 78 sites in southern England with approved managementplans for restocking by natural regeneration was made duringthe summers of 1993 and 1994. The following features were recorded:species, stocking, canopy cover and seed-bearing potential oftrees present in the overstorey; species and canopy cover ofthe understorey; ground cover; species, browsing damage, numberand heights of tree seedlings. In general, sites were poorlystocked with overstorey trees having an average of 135 stemsha^–1 and a mean canopy cover of 36 per cent. Similarly,the understoreys were poorly developed with an average coverof only 23 per cent. Twenty-nine species of tree were foundin the overstorey, Quercus spp. and Fraxinus excelsior werethe most common. Many of the stems present had poor seed-bearingpotential. Cover of the ground flora often exceeded 75 per cent.Seedlings were present on most sites, with F. excelsior andBetula pendula being most abundant with mean seedling densitiesexceeding 10000 ha^–1. Most seedlings were >20 cm tall,few exceeded 120 cm. About 30 per cent of seedlings were browsed.The results are related to current guidance and the future prospectsfor use of natural regeneration.     

The effect of geographical variation of planting rate on the uptake of carbon by new forests of Great Britain

The planting rates from 1921 to 1996 of new coniferous and broadleavedforests for 11 regions of Great Britain were assembled for thestate and private sectors. Over that period new planting totalled231 kha of conifers and 132 kha of broadleaves in England, 141kha of conifers and 16 kha of broadleaves in Wales and 881 khaof conifers and 52 kha of broadleaves in Scotland. These time series and regional values of Yield Class were usedas input data for an accounting model of carbon in the trees,litter, soils and products to produce estimates of their netuptake of carbon by the forests from the atmosphere (i.e. increasein the carbon pools). On the assumption that conifer and broadleafplanting could be represented by Sitka spruce and beech treesrespectively, litter and forest soil in Great Britain were accumulatingcarbon at 2.42 Mt a^–1 in 1995–96. Coniferous forestaccounted for 89 per cent of this uptake. Scottish conifer andbroadleaf forests took up 68 per cent and mapping the uptakeshowed that the greatest rate occurred in western Scotland.The pool of carbon in wood products increased in 1995–96by 0.31 Mt a_–1. The estimated uptake rates were sensitive to the relative amountsof conifer and broadleaf forest planted (particularly in relationto increases in the pool of carbon in wood products) but notto regional differences in Yield Class. Use of any single YieldClass in the range 10–16 m³ ha^–1 a^–1 for allSitka spruce planting produced estimates of uptake rate in GreatBritain to trees, litter and soil within ±10 per centof that, assuming yield varied across the country. Lack of preciseknowledge on the parameters of the model was estimated to introducean uncertainty of ±30–70 per cent into estimatesof carbon uptake.