Managing forests for wood yield and carbon storage: a theoretical study

Transcription

1 Tree Physiology 20, Heron Publishing Victoria, Canada Managing forests for wood yield and carbon storage: a theoretical study J. H. M. THORNLEY and M. G. R. CANNELL Institute of Terrestrial Ecology, Bush Estate, Penicuik, Midlothian EH26 0QB, U.K. Received August 24, 1999 Summary Which forest management regimes best achieve the dual objectives of high sustained timber yield and high carbon storage, including the carbon stored in soil and wood products? A mechanistic forest ecosystem simulator, which couples carbon, nitrogen and water (Edinburgh Forest Model), was calibrated to mimic the growth of a pine plantation in a Scottish climate. The model was then run to equilibrium (1) as an undisturbed forest, (2) removing 2.5, 10, 20 or 40% of the woody biomass each year (3) removing 50% of the woody biomass every 20 years, and (4) clear-felling and replanting every 60 years as in conventional plantations in this climate. More carbon was stored in the undisturbed forest (35.2 kg C m 2 ) than in any regime in which wood was harvested. Plantation management gave moderate carbon storage (14.3 kg C m 2 ) and timber yield (15.6 m 3 ha 1 year 1 ). Notably, annual removal of 10 or 20% of woody biomass per year gave both a high timber yield (25 m 3 ha 1 year 1 ) and high carbon storage (20 to 24 kg C m 2 ). The efficiency of the latter regimes could be attributed (in the model) to high light interception and net primary productivity, but less evapotranspiration and summer water stress than in the undisturbed forest, high litter input to the soil giving high soil carbon and N 2 fixation, low maintenance respiration and low N leaching owing to soil mineral pool depletion. We conclude that there is no simple inverse relationship between the amount of timber harvested from a forest and the amount of carbon stored. Management regimes that maintain a continuous canopy cover and mimic, to some extent, regular natural forest disturbance are likely to achieve the best combination of high wood yield and carbon storage. Keywords: carbon, forest, management, model, nitrogen, plantation, productivity, volume yield. Introduction Undisturbed forests yield no timber but have a high biomass and so store large amounts of carbon (Harmon et al. 1990). On the other hand, plantation forests which are periodically clear-felled can give high timber yields, but, averaged over the period from planting to harvesting, contain relatively little biomass and carbon compared with undisturbed forests (Cooper 1983, Cannell 1995). At first sight, it may be supposed that the more timber that is harvested from a forest the less carbon is stored. But, if timber were removed by regularly thinning, without clear-felling, would it be theoretically possible to obtain both a high sustained yield of timber and a large store of carbon? Is there a simple trade-off between these two objectives or is there an optimum management regime? The purpose of this study was to explore these questions. The answer is not self evident, because of the many interactions and feedbacks between plant and soil processes in a forest ecosystem, involving light, nutrients and water. Different management regimes perturb the system in different ways. Also, the answer would be difficult to derive by experimentation, because it would take centuries before valid estimates of sustained yield and carbon storage could be made. Transient responses would depend on the initial conditions and could differ in sign as well as magnitude from the equilibrium response. A model that represents all the essential interacting processes offers a way forward. In this paper, we use the Edinburgh Forest Model to estimate sustained timber volume yield and carbon storage in forests subjected to different harvesting regimes. The model is parameterized to simulate a pine forest in the climate of Scotland, but the principles elucidated may apply more widely. Materials and methods The Edinburgh Forest Model The Edinburgh Forest Model is a mechanistic evergreen forest ecosystem simulator that couples carbon, nitrogen and water and runs with a 20 minute timestep. It is programmed in ACSL (advanced continuous simulation language; Aegis Research Corporation, Huntsville, AL acsl-sales@aegisrc. com) and is available by anonymous ftp (username: anonymous; password: address) to budbase.nbu.ac.uk. The source program is FOREST.CSL in /pub/tree/forest/. The model is generic, assumes horizontal homogeneity and is composed of linked submodels, which are described elsewhere for the trees (Thornley 1991), soil and litter (Thornley 1998a, Chapter 5) and water (Thornley 1996, 1998a, Chapter 6). Flow diagrams of the submodel structures are shown as Figures A1 A4 in the Appendix, and a synopsis of the processes represented is given by Thornley and Cannell (1996)

2 478 THORNLEY AND CANNELL and Cannell et al. (1998). The model was calibrated to simulate a pine forest in a U.K. upland climate. Two developments have been made in the model since it was described (Thornley 1991, Thornley and Cannell 1996). First, the physiology of tree growth has been improved by including the acclimation of photosynthesis to light, nitrogen, carbon dioxide and temperature (Thornley 1998b), and by separating out some of the components of maintenance respiration, including those associated with phloem loading, uptake of nutrients and nitrate reduction (Cannell and Thornley 2000, Thornley and Cannell 2000). Second, the model was modified for this application to run in different management modes. As originally configured, it simulated the growth of a conifer plantation of identical trees of the same age, with an initial density of 2500 trees ha 1, which was regularly thinned, clear-felled and replanted every 60 years. The model has now been modified to simulate an undisturbed, natural forest too. This was achieved by assuming a proportional stem mortality rate that depended on tree nutrient status (the product of the carbon substrate and nitrogen substrate concentrations), water stress, and leaf area index (LAI) as a surrogate for between-tree competition. The proportional stem regeneration rate was then assumed to depend on stem mortality (as a result of the formation of gaps), tree nutrient status, irradiance at ground level and leaf area per stem. For the purposes of this study, which was to make comparisons between harvesting regimes rather than quantitative predictions, the carbon contained in wood products was estimated on the assumption that thinnings had a half-life of 5 years and clear-felled timber a half-life of 20 years. Environment The models were run in a constant annual climate simulating the 30-year mean conditions at Eskdalemuir in northern Britain, latitude 55 o 19 N, 242 m a.s.l. (Meteorological Office 1982). Wind speed (at 50-m height) was constant at 4 m s 1. Other quantities varied sinusoidally throughout the year: photosynthetically active radiation varied from a maximum of 7.1 MJ m 2 day 1 on June 21 to 0.5 MJ m 2 day 1 6 months later; daily maximum and minimum air temperatures varied from annual maxima of 18 and 9 C on July 26 to minima of 4 and 1 C 6 months later; soil temperatures were diurnally constant and varied from 14.5 C on July 27 to 1.5 C 6 months later; rainfall varied from 5.6 mm day 1 on November 20 to 2.8 mm day 1 6 months later (totalling 1530 mm year 1 ); daily maximum (dawn) and minimum (1500 h) relative humidities varied from annual maxima of 0.91 and 0.88 on December 30 to minima of 0.77 and months later. Daily variation in radiation was assumed to be a full sine wave between dawn and dusk; daily variation in air temperature was sinusoidal, with the minimum at dawn and the maximum at 1500 h. The forest was assumed to receive 10 kg N ha 1 year 1 from the atmosphere, evenly spread throughout the year. This is a reasonable average for combined dry and wet deposition of oxidized and reduced nitrogen in many areas of Britain and Europe. Model evaluation and calibration During the development of the model (Thornley and Cannell 1992), the assumptions and ways in which processes are represented were progressively modified until the dynamics of the system were stable and the model was capable of simulating known trends during a 60-year rotation, in the relative mass of tree parts, LAI, the main elements of the C budget (gross photosynthesis, respiration and carbon loss to litter) and the N budget (N uptake and N loss to litter). For this application, the parameters listed by Thornley and Cannell (see Table 2A in Thornley and Cannell 1996) were adjusted so that the values of the principal output variables were within measured ranges for pine forests in the U.K. uplands, averaged over a 60-year rotation (Table 1). The Yield Class was 15.6 m 3 ha 1 year 1 (Christie and Lines 1979). Net primary production varied from zero to 15 Mg C ha 1 year 1 over a rotation (averaging 4.5 Mg C ha 1 year 1 ) and at the end of the rotation the mean tree height was 25 m, with LAI = 4.1 (Miller et al. 1980). Total plant respiration averaged 0.40 of gross photosynthesis (within the range for pines: Ryan et al. 1997). The mean C:N ratios of soil organic matter, harvested wood and of the whole ecosystem were 11, 289 and 24, respectively. Non-symbiotic N 2 fixation averaged 7 kg N ha 1 year 1 (Sprent and Sprent 1990). Gaseous losses of N averaged 2 kg N ha 1 year 1 as NH 3 from foliage, 4 kg N ha 1 year 1 NH 3 volatilized from the soil (Sutton et al. 1993) and 1kgNha 1 year 1 lost by both nitrification and denitrification (Williams et al. 1992). Leaching losses were small, averaging 1kgNha 1 year 1, or 10% of atmospheric N deposition (Binkley and Hogberg 1997, Wright et al. 1995, Emmett et al. 1993). Over a rotation, 30% of the annual precipitation was lost by evaporation of intercepted rain, 21% by tree evapotranspiration and 49% by drainage to groundwater, which is within the measured range for U.K. upland forests (Johnson 1990). Model runs The model was run to equilibrium with the following management regimes: (1) undisturbed natural forest; (2) thinned natural forest, with 2.5, 10, 20 or 40% of the woody biomass removed each year simulating theoretical regular harvesting of whole trees or prunings; (3) thinned natural forest, with 50% of the woody biomass removed every 20 years; and (4) thinned plantation forest, clear-felled and replanted every 60 years, which is close to the rotation period for maximum mean annual volume increment (cf. Thornley and Cannell 1996, Cannell et al. 1998). Results Carbon sequestration and volume yield The simulated effects of the seven forest management techniques are summarized in Table 2. At equilibrium, the undisturbed natural forest stored a total of 35.2 kg C m 2 in biomass and soil organic matter (Figure 1). This is more than was stored in any of the management scenarios in which wood was TREE PHYSIOLOGY VOLUME 20, 2000

3 MANAGING FORESTS FOR YIELD AND CARBON STORAGE 479 Table 1. Parameters in the Edinburgh Forest Model were calibrated to give the output values shown below, simulating a pine plantation in the climate of Eskdalemuir, Scotland. All quantities are means for a 60-year rotation, unless otherwise stated. Net photosynthesis is gross canopy photosynthesis minus whole-plant respiration. The simulated Yield Class is 15.6 m 3 ha 1 year 1 (stemwood yield at harvest divided by 60). Parameter Output value Carbon budget (kg C m 2 year 1 ) Gross canopy photosynthesis 0.75 Net photosynthesis 0.45 Ratio (net/gross) 0.60 (dimensionless) Local growth respiration 0.11 Residual maintenance respiration 0.14 Soil respiration 0.23 Leaching Products 0.22 Nitrogen budget (kg N ha 1 year 1 ) Deposition 10 Fixation 7 NH 3 emission from foliage 2 Soil NH 3 volatilization 4 Nitrification 1 Denitrification 1 Soil gaseous N emission 5 System gaseous N emission 7 Leaching 1 Products 8 Water budget (m year 1 ) Annual rainfall 1.53 Intercepted and evaporated rain 0.46 Plant evapotranspiration 0.32 Drainage 0.75 End-of-rotation variables Leaf area index 4.1 Stem height Stem mass 25 m 1202 kg structural dry mass stem 1 Some other variables Stem height:diameter ratio (constant) 60 Stomatal conductance on July 1 at 1500 h m s 1 (cf fully open) Soil mineral N concentration kg N m 2 Carbon sequestered (kg C m 2 ) System 14.3 Soil (SOM, litter, biomass, Csol) 6.4 Tree 3.8 Products 4.0 C:N ratio of system: 23.5 kg C (kg N) 1 C:N ratio of soil organic matter: 10.7 kg C (kg N) 1 C:N ratio of total product pool: 289 kg C (kg N) 1 harvested, including the carbon stored in wood products (the equilibrium product pool allowing for decay). In the natural forest, biomass carbon reached an equilibrium of 13.2 kg C m 2, which occurred when tree respiration plus senescence equalled gross production. Continuous high light interception ensured moderately high canopy photosynthesis (despite low stomatal conductance on summer days; Table 1) and all of the net primary production was input to the soil, giving rise to a relatively high equilibrium soil carbon store of 22.1 kg C m 2 (Figure 1). The thinned and clear-felled plantation forest stored a total of 14.3 kg C m 2 (the sum of biomass, soil and product carbon; bottom of Figure 1), less than half that stored in the undisturbed forest. Clear-felling every 60 years and regular thinning limited the mean standing biomass to 3.8 kg C m 2, less than one-third of that in the undisturbed natural forest. Clear-felling and thinning also limited light interception, so that rotation-averaged canopy photosynthesis and net primary production were less than in the undisturbed forest, although the plantation forest suffered less water stress in summer (Table 1). Low net primary production and the removal of biomass in thinnings and at clear-felling decreased litter input to the soil, so that equilibrium carbon storage in the soil was only 6.4 kg C m 2, less than one-third of that in the undisturbed forest. The sustained wood yield was quite large, at 15.6 m 3 ha 1 year 1. Because the clear-felled timber had a half-life of 20 years (compared with 5 years for thinnings in this and the other management scenarios), the time-averaged product carbon pool for the plantation was also quite large similar in size to the biomass carbon pool (4.0 and 3.8 kg C m 2, respectively; Figure 1). However, the carbon stored in wood products (clear-felled timber and thinnings) did not compensate for the much smaller amounts of carbon stored in biomass and soils in the plantation forest compared with the undisturbed forest. Thinning the natural forest to remove just 2.5% of woody biomass each year yielded 12.2 m 3 ha 1 year 1, 78% of that yielded by the plantation forest, while storing a total of 28.1 kg Cm 2 in biomass, soil and products, twice that stored in the plantation forest (Figure 1). Thinning the natural forest to remove 10 and 20% of the woody biomass each year yielded 25.4 and 25.7 m 3 ha 1 year 1, respectively, about 60% more than that yielded by the plantation forest (Figure 1) while storing a total of 23.7 and 20.5 kg C m 2, respectively, more than stored in the plantation forest. Regularly thinning natural forests clearly resulted in stands that were considerably more efficient at generating, storing and yielding carbon than conventionally clear-felled plantations. The high performance of the 10 and 20% thinned natural forests was associated with high canopy photosynthesis and net primary productivity, resulting from a combination of moderately high light interception (with sustained LAIs of about 4), but lower evapotranspiration and less water stress on summer days than in the undisturbed forest, plus a lower respiratory load (net/gross photosynthesis was about 0.65 compared with 0.60 in plantations and undisturbed forests; Table 1). The biomass in these forests was decreased by thinning compared with the undisturbed forest, but even with 20% of the biomass removed each year, there was 3.0 kg C m 2 in biomass, about 80% of that in the plantation forest averaged over a rotation. More importantly, high sustained net primary production maintained a high input of litter to the soil, generat- TREE PHYSIOLOGY ON-LINE at

4 480 THORNLEY AND CANNELL Table 2. Simulated effect of seven scenarios of forest management on long-term (equilibrium) mean gross and net photosynthesis, and on water relations indicated by stomatal conductance on July 1 at 1500 h. Scenario Gross photosynthesis Net photosynthesis Stomatal conductance on (kg C m 2 year 1 ) (kg C m 2 year 1 ) 1 July 1 at 1500 h (m s 1 ) Undisturbed natural forest % of woody biomass harvested each year % of woody biomass harvested each year % of woody biomass harvested each year (fully open) 40% of woody biomass harvested each year % of woody biomass harvested every 20 years % of woody biomass harvested every 60 years (plantation) Equivalent to net primary production, equal to total litter input to the soil (because the system is at equilibrium) plus harvested wood products. ing over twice as much soil carbon in forests subject to 10 and 20% annual thinning as in the plantation forest. However, there is clearly a limit to the severity of annual thinning, beyond which the forest has such a low LAI that it is no longer productive. This limit was surpassed with the removal of 40% woody biomass every year, which resulted in very low LAIs, yield and carbon storage (Figure 1). Figure 1. Simulated effects of seven forest management regimes on forest biomass (continuous line), leaf area index (LAI, dashed line), and long-term equilibrium values of carbon sequestration and woody volume yield. Removing 50% of the woody biomass every 20 years gave similar wood yields and carbon storage to removing 2.5% every year the numerical equivalent (Figure 1). Similarly, removing 20% of the woody biomass every 4 years (not shown) gave similar results to removing 5% every year. Thus, thinning does not need to be done every year to achieve a combination of high yield and carbon storage. Within limits, it is the average removal rate that is important, giving flexibility to adopt various thinning regimes. Nitrogen inputs and outputs In addition to the 10 kg N ha 1 year 1 received from the atmosphere, the forests received N as a result of non-symbiotic N 2 fixation. Fixation was assumed to be positively related to the amount of soil organic matter (which determined the biomass of N 2 -fixing microbes) and the amount of carbon entering the soil (which influenced microbial N 2 fixation rates). The plantation forest fixed only about 7 kg N ha 1 year 1, primarily because it had a relatively small amount of soil organic matter (6.4 kg C m 2 ), whereas the undisturbed natural forest fixed over8kgnha 1 year 1. Natural forests that were thinned to remove 10 or 20% of the woody biomass each year, fixed about 9 kg N ha 1 year 1, more than that fixed by the undisturbed forest, owing to their high net primary production (Figure 2, Table 1). By definition, at equilibrium, ecosystem N outputs equaled N inputs (Figure 2). However, the fraction of N lost by leaching, in wood products and as gases (by nitrification, denitrification and volatilization), differed greatly among the forest management scenarios. In the undisturbed natural forest, leaching losses were small, because soil water drainage was small, and the N balance was maintained by gaseous loss. In the plantation forest, almost half of the N output was in harvested products, but leaching losses were also relatively high, owing to mineralization of litter and high drainage during the years after clear-felling, In the 10 and 20% annually thinned natural forests, about 47% of the N output was in harvested products, whereas leaching losses were no greater than in the undisturbed forest, owing to low drainage and continued depletion of the soil mineral pools by root uptake (driven by tree growth). Thus, these thinned natural forests avoided the N loss to groundwaters that occurred in plantations and yet captured TREE PHYSIOLOGY VOLUME 20, 2000

5 MANAGING FORESTS FOR YIELD AND CARBON STORAGE 481 Figure 2. Simulated effects of seven forest management regimes on long-term (equilibrium) system N inputs and outputs. Definitions: N deposition = atmospheric N deposition; N 2 fixation = non-symbiotic N 2 fixation; products = N removed in harvested wood; gas = losses of N 2, NO, N 2 O and NH 3 from denitrification, nitrification, volatilization from the soil, and emissions by foliage through the stomates; and leaching = losses to ground water from the soil nitrate pool. a substantial fraction of the N that was otherwise lost as gases in undisturbed forests. Essential features of undisturbed, thinned natural and plantation forests The essential features of the simulated undisturbed, thinned natural and plantation forests are illustrated in Figure 3. Differences between these forest types in mean LAI and biomass affected evapotranspiration and light interception, which in turn affected the C and N dynamics. The undisturbed forest maintained a high LAI and biomass, resulting in high light interception, but also high evapotranspiration and summer water stress, so that net primary production was only moderately high. However, leaching losses were small (because of low drainage) and, because no biomass was removed, soil organic matter levels and N 2 fixation rates were moderately high. The net result was high carbon sequestration in biomass and soils, but no yield and a high loss of N as gases. Averaged over a rotation, the LAI and biomass of the plantation forest were relatively small, resulting in relatively low Figure 3. Summary of the main features of simulated forests, managed in three ways, when at equilibrium. No marking indicates a relatively small value of the process or quantity, underlining a moderate value, and bold type a large value. Abbreviations: LAI = leaf area index; GPP = gross primary production or gross photosynthesis; NPP = net primary production or net photosynthesis; SOM = soil organic matter; C seqn. = carbon sequestration or storage in kg C m 2 ; and harvest = woody volume removed in m 3 ha 1 year 1. light interception, low evapotranspiration, high drainage and N leaching, leading to low net primary production, despite relatively little summer water stress. Low net primary production, combined with the removal of wood, resulted in low soil organic matter levels and N 2 fixation rates. The outcome was a combination of moderate carbon sequestration, moderate wood yield, and considerable N loss to groundwaters. Natural forests that were thinned to remove about 10% of the biomass each year (or 50% every 5 years) maintained a moderately high, continuous LAI and biomass, giving moderate light interception and evapotranspiration, resulting in relatively little summer water stress. Consequently, net primary production was high, sustaining both a high yield and litter in- TREE PHYSIOLOGY ON-LINE at

6 482 THORNLEY AND CANNELL put to the soil, resulting in moderately high soil content of organic matter and high N 2 fixation rates. The outcome was moderately high carbon sequestration and high wood yield. Discussion The salient conclusion from this study is that there is no simple inverse relationship between the amount of timber harvested from a forest and the amount of carbon stored in the ecosystem and wood products. The method of harvesting is important. In particular, regular removal of timber from a forest (annually or every few years) in a way that maintains a continuous canopy is likely to give substantially higher sustained yields and amounts of carbon storage than periodical clear-felling, as in conventional plantations. Runs of the model in other temperate climates and with different calibrations and N deposition rates suggested that, qualitatively, this conclusion might apply in a wide range of forest types. It should be stressed, however, that we claim only to have elucidated a principle, not to have made quantitative predictions, much less to have considered the practicalities, costs and other implications of different harvesting regimes. The study suggested that, if the objective were simply to maximize timber volume yield (regardless of cost or value) the order of optimal management system would be regularly thinned forest > plantation > undisturbed forest, whereas if the objective were to maximize carbon storage, the order would be undisturbed forest > regularly thinned forest > plantation. If the objective were to minimize uncontrolled N emissions to the environment, the order would be the same as that to maximize timber volume yield. In the simulations, more carbon was stored in the undisturbed forest than in any management regime in which wood was harvested, including wood products, supporting previous analyses that compared plantations with undisturbed forests (cf. Cannell 1995) and suggesting that any biomass removal from a forest will lower carbon storage, without unrealistic assumptions on the rates of decay of harvested wood. Plantations offered the worst combination of yield and carbon storage and regularly thinned forests the best, provided thinning removed not less than about 5% and not more than about 25% of the woody biomass each year. There seemed to be some scope for flexibility in thinning frequency, in that, for instance, harvesting 5% of the woody biomass per year gave similar results to harvesting 25% every 5 years. Why were regularly thinned forests so efficient in the model? The main features were as follows: (1) the continuous canopy and moderately high LAI (about 4) gave high light interception and NPP; (2) there was no period of slow recovery, which occurs after clear-felling; however, the LAI was less than in an undisturbed forest and so evapotranspiration was less, with less risk of water stress on dry summer days, also enhancing NPP; (3) high NPP ensured high litter input to the soil and a large equilibrium soil carbon store, also favoring non-symbiotic N 2 fixation; (4) regular thinning meant that the forest had a lower biomass than an undisturbed forest and was continually growing, resulting in less maintenance respiration; and (5) continuous growth also meant that the soil mineral nitrate pool remained depleted and N losses by leaching were reduced. The conclusion that regular thinning is better than clear-felling is in keeping with much of the current discussion regarding forest management to sustain multiple functions, including the maintenance of biodiversity, preserving soil fertility, and preventing erosion (Gale and Cordray 1991, Wiersum 1995). It is increasingly recognized that the many demands made on forests may best be met by maintaining an intact forest ecosystem or ecological integrity (Armstrong 1999). This analysis offers some scientific basis for those terms, with regard to carbon dynamics in the soil plant system, as affected by nitrogen dynamics and the water balance. Finally, the regular thinning regimes that were optimal in this study may be viewed as emulating the natural regular disturbance that many forests experience from insects, other animals, storms and minor fires. Simulating nature may, by chance or design, be an optimal strategy to maximize yield and carbon storage, as it may be in the natural disturbance model of boreal forest management proposed by Hunter (1993). Acknowledgments This work was supported in part by the U.K. Department of Environment, Transport and Regions in Contract EPG1/1/39. References Armstrong, G.W A stochastic characterization of the natural disturbance regime of the boreal mixed wood forest with implications for sustainable forest management. Can. J. For. Res. 29: Binkley, D. and P. Hogberg Does atmospheric deposition of nitrogen threaten Swedish forests? For. Ecol. Manage. 92: Cannell, M.G.R Forests and the global carbon cycle in the past, present and future. European Forest Institute. Research Report 2. Joensuu, Finland, 66 p. Cannell, M.G.R. and J.H.M.Thornley Modelling the components of plant respiration. I: some guiding principles. Ann. Bot. In press. Cannell, M.G.R., J.H.M. Thornley, D.C. Mobbs, and A.D. Friend UK conifer forests may be growing faster in response to increased N deposition, atmospheric CO 2 and temperature. Forestry 71: Christie, J.M. and R. Lines A comparison of forest productivity in Britain and Europe in relation to climatic factors. For. Ecol. Manage. 2: Cooper, C.F Carbon storage in managed forests. Can. J. For. Res. 13: Emmett, B.A., B. Reynolds, P.A. Stevens, D.A. Norris, S. Hughes, J. Gorres and I. Lubrecht Nitrate leaching from afforested Welsh catchments interactions between stand age and nitrogen deposition. Ambio 22: Gale, R.P. and S.M. Cordray What should forests sustain? Eight answers. J. For. 89: Harmon, M.E., W.K. Ferrell and J.F. Franklin Effects on carbon storage of conversion of old growth forests to young forests. Science 247: Hunter, M.L Natural fire regimes as spatial models for managing boreal forests. Biol. Conserv. 65: TREE PHYSIOLOGY VOLUME 20, 2000

7 MANAGING FORESTS FOR YIELD AND CARBON STORAGE 483 Johnson, R.C The interception, throughfall and streamflow in a forest in Highland Scotland and the comparison with other upland forests in the UK. J. Hydrol. 118: Meteorological Office Tables of temperature, relative humidity, precipitation and sunshine for the world. Part III. Europe and the Azores. HMSO, London, 228 p. Miller, H.G., J.D. Miller and J.M. Cooper Biomass and nutrient accumulation at different growth rates in thinned plantations of Corsican pine. Forestry 53: Ryan, M.G., M.B. Lavigne and S.T. Gower Annual carbon cost of autotrophic respiration in boreal forest ecosystems in relation to species and climate. J. Geophys. Res. 102: Sprent, J. and P. Sprent Nitrogen fixing organisms: pure and applied aspects. Chapman and Hall, London, 256 p. Sutton, M.A., C.E.R. Pitcairn and D. Fowler The exchange of ammonia between the atmosphere and plant communities. Adv. Ecol. Res. 24: Thornley, J.H.M A transport-resistance model of forest growth and partitioning. Ann. Bot. 68: Thornley, J.H.M Modelling water in crops and plant ecosystems. Ann. Bot. 77: Thornley, J.H.M. 1998a. Grassland dynamics: an ecosystem simulation model. CAB International, Wallingford, Oxon, pp xii and 241. Thornley, J.H.M. 1998b. Dynamic model of leaf photosynthesis with acclimation to light and nitrogen. Ann. Bot. 81: Thornley, J.H.M. and M.G.R. Cannell Nitrogen relations in a forest plantation-soil organic matter ecosystem model. Ann. Bot. 70: Thornley, J.H.M. and M.G.R. Cannell Temperate forest responses to carbon dioxide, temperature and nitrogen: a model analysis. Plant Cell Environ. 19: Thornley, J.H.M. and M.G.R. Cannell Modelling the components of plant respiration. II: representation and realism. Ann. Bot. In press. Wiersum, K.F years of sustainability in forestry: lessons from history. Environ. Manage. 19: Williams, E.J., G.C. Hutchinson and F.C. Fehsenfeld Nox and N 2 O emissions from soils. Global Biogeochem. Cycles 6: Wright, R.F., J.G.M. Roelofs, M. Bredemeier, K. Blanck, A.W. Boxman, B.A. Emmett, P. Gundersen, H. Hultberg, O.J. Kjonaas, F. Moldan, A. Tietema, N. van Breeman and H.F.G. van Dijk NITREX: responses of coniferous forest ecosystems to experimentally changed deposition of nitrogen. For. Ecol. Manage. 71: Figure A1. Overall structure of the Edinburgh Forest Model Appendix Figure A2. Flow diagram of the tree submodel of the Edinburgh Forest Model. TREE PHYSIOLOGY ON-LINE at

8 484 THORNLEY AND CANNELL Figure A3. Flow diagram of the soil submodel of the Edinburgh Forest Model. Figure A4. Flow diagram of the water submodel of the Edinburgh Forest Model. TREE PHYSIOLOGY VOLUME 20, 2000