Metsäntutkimuslaitoksen tiedonantoja 586 The Finnish Forest Research Institute, Research Papers 586 Jyrki Hytönen Biomass production and nutrition of short-rotation plantations BIOMASS PRODUCTION AND NUTRITION OF SHORT-ROTATION PLANTATIONS Jyrki Hytönen ACADEMIC DISSERTATION To be presented with the permission of the Faculty of Agriculture and Forestry of the University of Helsinki, for public criticism in Auditorium II of Metsätalo, Unioninkatu 40 B, Helsinki, on February 23th 1996, at 12 o'clock noon. Metsäntutkimuslaitoksen tiedonantoja 586 The Finnish Forest Research Institute, Research Papers 586 Kannus 1996 2 Hytönen J. 1996. Biomass production and nutrition of short-rotation plantations. Metsäntutkimuslaitoksen tiedonantoja 586. The Finnish Forest Research Institute, Research Papers 586. 61 p. ISBN 951-40-1494-4, ISSN 0358-4283. The concept of short-rotation management comprises the establishment of stands of closely spaced, fast-growing trees and the application of intensive cultivation practices and repeated harvesting, using short cutting cycles and regeneration of subsequent crops via sprouts. Short-rotation plantations established on cut-away peatland areas and abandoned mineral soil agricultural land were studied. The tree species included in the study were mainly exotic willows. The amount of biomass produced and different methods of biomass estimation were compared. Factors affecting the regeneration of the plantations were studied. The development of stand structure in the course of intense competition was followed during a period of ten years. Special attention was focused on nutritional matters, the effects of fertilizer application on biomass production and the amount of nutrients bound in the biomass. Regression estimation using easily measured dimensional variables proved to accurately describe willow wood, bark and leaf mass as well as leaf area. Due to the non-destructive nature of this method, it is especially appropriate when the plantation is to be grown further or when estimates of annual biomass production are needed. Cutting of exotic willows during the summer led to an increase in mortality and decrease in biomass production during the following year. Short stumps gave higher yields than longer ones. Damage to the stumps in the harvesting phase decreased biomass production, especially in a young stand. In a dense willow plantation internal within-stand competition started already during the first growing season. It manifested itself in the form of skewed and bimodal stem frequency distributions. The smallest sprouts died and only one tenth of the original first year shoots survived to the end of the seven-year rotation. Fertilization in cut-away peatland areas was essential for achieving and maintaining high biomass production. In limed cut-away peatland areas only readily soluble phosphorus fertilizers increased the growth of willows and the amount of acid ammonium acetate extractable phosphorus in soil. Nitrogen fertilizer should be applied annually in order to maintain high productivity. Phosphorus and potassium could be applied at longer intervals. Analysis of soil extractable nutrients could be used to establish guidelines for determining the fertilization requirements of short rotation willow plantations. Willows bound high amounts of nutrients in their biomass. Almost half of the nutrients were in the foliage. With increase in age, the amount of wood and bark out of the total production increased and consequently the amount of nutrients bound per unit biomass decreased. Key words: Short-rotation plantations, biomass estimation, Salix, Populus, fertilization, nutrient uptake, sprouting, cutting time, stump height. Author's address: Finnish Forest Research Institute, Kannus Research Station, P.O. Box 44, FIN-69101 Kannus, Finland. 3 ACKNOWLEDGEMENTS This study has been carried out in the Finnish Forest Research Institute and has been part of research on wood as source of energy. For the guidance and expertice at the early years of the study I am grateful to Prof. Eero Paavilainen and Prof. Veli Pohjonen who have given me plenty of advice and support. Professor Seppo Kaunisto has guided me during all these years through many troubled waters and is especially thanked for his support and criticism in preparing this report. Prof. Matti Leikola encouraged me to join the pieces of this work together and at the final stages, Prof. Carl Johan Westman gave valuable criticism and advice. I am also grateful to Dr. Ilari Lumme, Mr. Juha Nurmi, Mr. Pekka Rossi, Mrs. Anna Saarsalmi, Dr. Timo Törmälä and Mr. Risto Lauhanen for their inspiring research collaboration. I especially remember the numerous, both scientific and personal discussions at different geographic locations with the late Dr. Ari Ferm. Besides colleagues in Finland, words of praise are also due to Dr. Stig Ledin, Prof. Paul Mitchell, Dr. Kurth Perttu, Dr. Lisa Sennerby-Forsse, Prof. Louis Zsuffa and many others who have inspired me at International Energy Agency Bioenergy Agreement meetings during the past seven years I have been a member in different activies and working groups. During the course of this research I have received professional help in the field, laboratory and office work from many persons at the Kannus Research Station. Especially Mr. Esa Heino, Mrs. Kaisa Jaakola, Mr. Keijo Polet, Mrs. Riitta Miettinen, Mrs. Arja Sarpola, Mrs. Sirpa Puranen, Mr. Kaarlo Sirviö and Mr. Seppo Vihanta have provided substancial help in collecting and analysing the data and contributed to the completion of the research. Mr. Arto Ketola patiently helped me with statistical analysis. My sincere thanks to all you. The assistance of numerous other people who have helped during the course of this study is also gratefully acknowledged. I wish to extend my warmest thanks to all my colleagues and friends, who with their stimulating discussions have created an inspiring atmosphere in which to work. I also want to thank Mr. Erkki Pekkinen for revising my English and Mrs. Minna Vihurila for the cover drawing. And last but not least I want to express my hearthful thanks to Maarit and Hanna for their patience and understanding during all these years. Kannus, January 1996 Jyrki Hytönen 4 CONTENTS LIST OF ORIGINAL PAPERS 5 1 INTRODUCTION 7 2 AIMS OF THE STUDY 12 3 MATERIAL AND METHODS 13 4 RESULTS AND DISCUSSION 18 4.1 Estimation of biomass 18 4.2 Establishment, regeneration and stand structure 21 4.3 Nutrition 25 4.3.1 Site characteristics 25 4.3.2 Effect of fertilization on soil nutrients 26 4.3.3 Nutrient concentrations in willow 27 4.3.4 Nutrients bound into the biomass 30 4.3.5 Nutrient cycling 33 4.3.6 The response of willow to fertilization 33 4.4 Biomass production 37 5 CONCLUSIONS 43 REFERENCES 46 ORIGINAL PAPERS (I-IX) 5 This thesis is based on the following publications, which are referred to in the text by their roman numerals. I Hytönen, J., Lumme, I. & Törmälä, T. 1987. Comparison of methods for estimating willow biomass. Biomass 14:39-49. II Hytönen, J. 1994. Effect of cutting season, stump height and harvest damage on coppicing and biomass production of willow and birch. Biomass and Bioenergy 6(5):349-357. 111 Hytönen, J. 1995. Ten-year biomass production and stand structure of Salix 'Aquati ca' energy forest plantation in southern Finland. Biomass and Bioenergy 8(2):63-71. IV Ferm, A., Hytönen, J. & Vuori, J. 1989. Effect of spacing and nitrogen fertilization on the establishment and biomass production of short rotation poplar in Finland. Biomass 18:95-108. V Hytönen, J. 1985. Teollisuuslietteellä lannoitetun vesipajun lehdetön maanpäällinen biomassatuotos. Abstract: Leafless above-ground biomass production of Salix Aquatica' fertilized with industrial sludge. Folia Forestalia 614. 16 p. VI Hytönen, J. 1986. Fosforilannoitelajin vaikutus vesipajun biomassatuotokseen ja ravinteiden käyttöön turpeennostosta vapautuneella suolla. Summary: Effect of some phosphorus fertilizers on the biomass production and nutrient uptake of Salix Aquati ca' in a peat cut-away area. Folia Forestalia 653. 21 p. VII Hytönen, J. 1995. Effect of fertilizer treatment on the biomass production and nutrient uptake of short-rotation willow on cut-away peatlands. Silva Fennica 29(l):21-40. VIII Hytönen, J. 1994. Effect of fertilizer application rate on nutrient status and biomass production in short-rotation plantations of willows on cut-away peatland areas. Suo 45(3):65-77. IX Hytönen, J. 1995. Effect of repeated fertilizer application on the nutrient status and biomass production of Salix Aquatica' plantations on cut-away peatland areas. Silva Fennica 29(2): 107-116. 6 7 1 INTRODUCTION The concept of short-rotation management includes the establishment of closely-spaced stands of fast-growing trees and the application of intensive cultivation practices and repeated harvesting, using short cutting cycles, regeneration of subsequent crops via sprouts or suckers, and a high degree of mechanization (e.g. Siren 1974, 1979, Pohjonen 1980, Pelkonen & Rossi 1984, Siren et al. 1984). Since the introduction of the concept by McAlpine et al. in 1966, short-rotation production practices have been under study in many countries (Coombs et al. 1990, Mitchell et al. 1992) with the maximization of the biomass growth of selected woody species as the goal. The characteristics of the ideotypes of the different tree species suitable for short-rotation cultivation have been outlined; e.g. by lEA's (International Energy Agency) forest energy working groups (Koski & Dickmann 1992). The criteria often used when selecting poplar and willow clones include the following: good rooting ability, fast and vigorous growth (especially at an early age), the capacity to grow as closely-spaced stands, good coppicing ability, resistance to pests, and, nowadays, also good quality of the wood produced for biomass utilization (e.g. Leikola 1976, Pohjonen 1977, Siren et al. 1984, Zsuffa & Gambles 1992, Mitchell 1992). Willows have been considered to possess several traits desirable in short-rotation tree species. Besides willows, especially the cultivation of alders and birches, applying longer rotations, has also been investigated in Finland (Saarsalmi et al. 1985, 1991, 1992, Ferm 1990, 1993). Willows have a long history of cultivation in Finland and Europe (Pohjonen 1984, 1987, 1991). For centuries, they have been the target of selection in Europe and they have been grown in closely-spaced plantations to produce material for the basket industry (Wa sielewski 1982, Pohjonen 1984). The first mention of willow plantations as a fuel source in Finland is from the year 1753 (Lithander, ref. Pohjonen 1984) and the first Finnish report on basket willow cultivation was published as early as in 1882 (Flinta 1882). At the beginning of this century Nordberg (1928) studied and propagated for cultivation of willows for basketry in Finland. In the 19505, willow cultivation and studies were initiated in Finland by Tapio (Tapio 1965). However, despite considerable inputs into research and development, basket willow cultivation in Finland did not reach the level of practical farming. At the beginning of the 19705, nearly all merchantable stemwood in Finland found industrial use and the country's forest industry was occasionally confronted by shortages of raw material (Hakkila et al. 1979). First field trials established in 1973 by Pohjonen (1974) suggested that willows have considerable yield potential. Studies on the mass production of small-sized broadleaved trees were subsequently initiated. Especially at the end of the 19705, studies on domestic fuels were stepped up and the PERA project was started at the Finnish Forest Research Institute (Hakkila 1985). The emphasis in research was placed on existing small-tree and logging residue reserves, harvesting techniques, productivity, measur ing, improvement of chip quality, and reduction of procurement costs. Biomass production systems using the short-rotation management principles, including a wide variety of topics from basic biology to the practical aspects of growing and harvesting crops, were also 8 included in the sphere of research. Mainly Salix species have been used in the short-rotation experiments conducted in Finland and Sweden. Both the genera Polulus L. and Salix L. belong to the family Salicaceae and they are widely distributed in the northern hemisphere. The genus Salix is divided into three subgenera: Salix, Vetrix and Chamaetia. The majority of Salix species are shrubby. Most willow species and clones investigated in short-rotation experiments belong systematically to the subgenus Vetrix (shrub willows) of the genus Salix. The provenances, nomenclature and position in plant systematics of several willows used in short-rotation cultivation is not fully known (Stott 1984, Pohjonen 1987, Viherä-Aarnio 1991, Gullberg 1993). The taxonomic difficulties are biggest in the subgenus Vetrix (Pohjonen 1991). Of the indigenous willow species, S. myrsinifolia Salisb. or S. phylicifolia L. could be promising ones for short-rotation forestry (Pohjonen 1991, Honkanen 1994, Hytönen et ai. 1995). Salix viminalis L., widely used in Swedish practical short-rotation plantations (Sennerby-Forsse 1994), has also been tested in Finland (Rossi 1982, Hytönen 1987, Tahvanainen 1995). Salix 'Aquatica' (S. Aquatica Gigantea') originating from Denmark has also been widely used in short-rotation energy forest research in Europe. Its cultivation history in Finland has been reviewed in detail by Pohjonen (1987). Salix x dasyclados Wimm., a hybrid most likely between S. cinerea and S. viminalis (Pohjonen 1991, Hämet- Ahti et ai. 1989) and Salix Aquatica' resemble morphologically each other, even though some differences have been recognised (Heino 1982, Pohjonen 1987, Hämet-Ahti et ai. 1989). Some doubt even exists whether the clones cultivated in Finland as S. x dasyclados are true S. x dasyclados hybrids (Pohjonen 1987, 1991). Pohjonen (1987) suggests that nearly all clones of S. Aquatica' and S. x dasyclados Wimm. in Finland belong to the one and the same Siberian species, Salix burjatica Nasarov. A properly marked and registered clone has been the basic unit in short-rotation research. Best willow clones have commercial cultivar names in Sweden. Pohjonen (1991) even considers that in practical short-rotation forestry all taxons between species and clones are unnecessary. In many respect, willows are very rewarding breeding material. Wide genetic variation both within and between species, polyploid cellular structure, promising hybridization potential, flowering at an early age, and plentiful seed production offer a good basis for the selection and breeding of willows (Malmivaara et ai. 1971, Sennerby-Forsse et ai. 1983, Ager et ai. 1986, Zsuffa & Mosseler 1986, Hall 1986, Lumme & Törmälä 1988, Viherä- Aarnio 1988, 1991, Gullberg 1989). Early selection is possible when short rotation periods are utilized. In most willow species, clonal selection of individuals is possible due to easy vegetative propagation and thus increases the likelihood of finding superior combinations. Clonal selection from F,-progenies of planned crossings is likely to increase selection effi ciency (Viherä-Aarnio 1988, Pohjonen 1991, Viherä-Aarnio & Saarsalmi 1994). In the 1980 s, both exotic and indigenous willow species and clones were selected in Finland; this is described in detail by Pohjonen (1991) and Viherä-Aarnio (1991). Careful site selection, including site-clone matching, is necessary for satisfactory establishment and yield from intensively cultivated willow plantations (e.g. Leikola 1976, Hakkila et ai. 1979). Sites should be fertile and well-drained and free of stones. Mechan 9 ization of plantation establishment, plantation management and harvesting also influence site selection. The Finnish Energy Forest Committee (Energiametsätoimikunnan ... 1979, 1981) calculated that 550 000 ha of potentially suitable sites could be available for intensive short rotation forestry. Mainly marginal lands, e.g. former fields, open mires, cut-away peatland areas, powerlines, roadsides, and land formerly under water, have been suggested to be used for short-rotation forestry (Siren 1974, Leikola 1976, Hakkila et ai. 1979, Energiametsätoi mikunnan ... 1979, 1981). Short-rotation plantations have been studied in cut-away peatland areas besides in Finland also in Ireland and Estonia (Neenan 1983, Valk 1986, Kirt 1994). Nowadays, high quality agricultural lands are becoming available not only in Finland, but also in many other European countries (Ferm & Polet 1991, Christersson et al. 1993, Järvenpää et al. 1994, Hytönen 1995). Due to surplus production in agriculture, it is estimated that as much as 500 000 - 1 000 000 ha of arable land should be taken out of agricultural production in Finland (Järvenpää et al. 1994, Toivonen et al. 1994). Intensive site preparation and weed control are essential prior to planting. Especially during the planting year, willows and poplars planted as cuttings are intolerant of competition from weeds and competing vegetation can damage the plantation (Andersson et al. 1983, McElroy & Dawson 1986, Bowersox et al. 1988, Dawson 1988). Efficient weed control is a prerequisite for the establishment of productive plantations. Short-rotation willow plantations are usually established using unrooted stem cuttings which develop adventitious roots when planted. Willows generally root easier from cuttings than poplars do (Siren & Sivertsson 1976, Leikola & Rossi 1977, Pohjonen 1977, Ferm 1985 a). Cuttings offer a good possibility for using cloned material and their production, storage and planting are easy and can be mechanized (Rossi 1983, Harstela & Tervo 1983). The length of cuttings has generally been 20 cm with a minimum thickness of 5 mm. Special treatment of willow cuttings, e.g. soaking or hormone treatments has generally enhanced their good rooting ability only little (Rossi 1979 a, 1979b). Planting should be done as early in the spring as possible and cuttings should be planted in an upright position even though they root quite well also when planted horizontally if the soil is moist enough (Hytönen 1983). Short-rotation willow plantations are often cut-back after the first growing season. This increases the number of sprouts per stump and alters the shoot-root ratio. The ability to coppice or re-sprout after harvesting is one important characteristic of tree species suitable for short-rotation forestry. A critical aspect of biomass plantations is the sustainability of the system, especially its ability to withstand repeated harvesting (Sennerby-Forsse et al. 1992). The coppicing ability of tree species varies considerably; e.g. willows are considered to sprout well compared with birches (Ferm & Issakainen 1981, Ali- Alha 1987). The initial development of sprouts is often much faster than that of seedlings of the same tree species (Blake 1980, Kauppi et al. 1988). Several factors, both internal and external, influence the regeneration of growth from stumps. It has been shown with many tree species that factors such as season of cutting, cutting tool, stump height, growing site, tree diameter, tree age, spacing and rotation length influence re-sprouting (Heikinheimo 1930, Mikola 1942, Leikola & Mustanoja 1961, Etholen 1974, Moilanen & Oikarinen 1980, Blake & Raitanen 1981, Ferm & Issakainen 1981, Sennerby-Forsse et al. 1992). Though 10 extensively used in short-rotation cultivations, willows have not been studied in this respect. Knowledge of factors affecting coppice regeneration are necessary for the determination of cutting schedules and development of harvesting techniques. Besides the cultivation factors, knowledge of the physiology and morphology of sprout-producing buds and their development into sprouts is essential (Ferm & Kauppi 1990). High planting densities in willow cultivation (often 40 000 cuttings ha" 1 ) have been used in Finland (Hytönen 1993). Following cutting-back after the first growing season, the number of sprouts in S. 'Aquatica' plantations can exceed 300 000 ha" 1 (Hytönen 1988). The post-harvesting stand densities increase due to the increased amount of shoot-producing buds (Paukkonen et ai. 1992). Within-stand competition in such dense plantations is high. Willow growth and within-stand competition is further enhanced by fertilizer application. Competition-induced shoot mortality before the end of the rotation period is also high (Verwijst 1991b, Willebrand & Verwijst 1993). Special attention should be directed at the nutrient status of soil when practising short rotation forestry. Repeated harvesting of the above-ground biomass at short cutting cycles, even though excluding the foliage, promotes the loss of nutrients. Short-rotation willows bind considerable amounts of nitrogen, phosphorus, potassium and other nutrients into their biomass (Saarsalmi 1984, Ferm 1985 a). Frequent repetition of such nutrient drains could result in nutrient depletion of the site. An important objective of research programmes developing woody biomass plantations is to establish fertilization regimes optimizing growth with minimal adverse environmental consequences (Miegroet et al. 1994). Correct fertilization regime, with respect to timing and rates, is one of the most important ways to improve crop production. Our knowledge of the factors influencing the fertilizer reaction of short-rotation plantations (inc. nutrient status of the substrate, species and its stage of development, type and amount of fertilizer) is inadequate. Determining the need for ferti lization before the plantation is established should be based on site classification according to the natural nutrient status (soil analyses). Later on, foliar analyses and nutrient deficiency symptoms could also be used (Paavilainen 1979, Miller 1983). With nutrient physiology studies as the basis, trials have been conducted in Sweden using small amounts of liquid fertilizer, administered even daily (Ericsson 1981 a, Ingestad & Agren 1984, Christersson 1986, 1987). Yields have been estimated to have increased considerably by applying nutrients, especially nitrogen, during the growing season at the same rate as they are bound by the plants (Ingestad & Ägren 1984). The optimum pH for good development of willow root systems varies from species to species. The roots of S. pentandra L. and S. cinerea L. have developed almost as well in a hydroponic culture under pH 3.5 as under pH 5.0 (Lattke 1969). However, many willows used in short-rotation cultivation require rather high pH (5.0 - 6.0) levels of the substrate (e.g. S. viminalis, S. 'Aquatica', S x dasyclados) (Ericsson & Lindsjö 1981, Ferm & Hytönen 1988). The pH of cut-away peatland areas is generally quite low (e.g. Kaunisto 1980, 1982 a, Hytönen 1984, Ferm & Kaunisto 1983, Lumme et ai. 1984). The soils of cultivated fields generally have pH values exceeding 5 (Kurki 1982, Viljavuustutkimuksen ... 1992). Cut-away peatland areas are characterized by their variable peat thickness, low pH and 11 high nitrogen concentrations and low phosphorus and potassium concentrations (Kaunisto 1979, Ferm & Kaunisto 1983, Lumme et ai. 1984, Lehtonen & Tikkanen 1986, Ferm & Hytönen 1988, Kaunisto & Viinamäki 1991). The nutrient status of agricultural soils shows wide variation (Kurki 1982, Viljavuustutkimuksen... 1992). Afforestation of cut-away peatland areas, and especially of peatland fields, may be confronted by nutritional problems (Raitio 1979, Veijalainen 1983, Hytönen & Ekola 1993, Aro 1995). Research results show that success in the afforestation of cut-away peatland areas depends a lot on fertilization (Kaunisto 1979, 1986, 1987 a, Valk 1986, Ferm & Hytönen 1988, Aro 1995). Fertilization and soil amelioration are probably also the most important factors affecting the biomass production of short-rotation plantations in cut-away peatland areas (Hytönen 1982, 1987, Kaunisto 1983, Ferm & Hytönen 1988, Lumme 1989). It is especially important to optimize nitrogen application because of the high cost of fertilizer nitrogen. Besides commercial fertilizers, also the recycling of wood ash and utilization of nutrients in sludges could offer ecologically and environmentally interesting alternatives for soil amelioration and fertilization in short-rotation plantations (Kaunisto 1983, Ferm 1985 a, Lumme & Laiho 1988, Lumme 1989). However, the need for repeating fertilization and the duration of the fertilizer effect are poorly known. Dry mass is more useful than volume as a parameter for depicting the value of the raw material for industrial purposes focusing on the production of pulp, chemicals and energy (Hakkila 1989). Dry-mass production of a short-rotation plantation can be measured with different methods varying greatly in the amount of work and destructiveness they involve. Determination of the current annual yield requires annual measurements by a non destructive method so as not to influence the future development of the plantation. Especially for research purposes, the composition of the produced biomass should be described; e.g. the chemical and physical composition of wood differs considerably from bark. Chemical and physical properties can play an important role in the efficiency of most energy conversion processes. Many biotic and abiotic factors influence the biomass production of short-rotation plantations. According to simulation models, the annual variation in radiation and temperature can cause considerable year-to-year variation in biomass production (Eckersten et al. 1983, Nilsson & Eckersten 1983, Sievänen 1983, Perttu et ai. 1984, Nilsson 1985, Ec kersten 1985, Eckersten et al. 1987). In the case of some soil types and weather conditions, the availability of water can also be a limiting factor (Grip 1980 a, 1980b, 1981, Grip & Perttu 1982, Kaunisto 1983, Kowalik & Eckersten 1984, Saarsalmi 1984). Early summer frosts can considerably decrease the biomass production of willow plantations (Christersson et al. 1982, 1984, Ericsson et al. 1983, Ahola 1987, Fircks 1992). Production may also be severely reduced by the susceptibility of the species or clones to pests and diseases (Rossi 1982, Larsson & Wiren 1981, Hubbes 1983, Morris 1983, Ronnberg-Wastljung & Gun nerbeck 1985, Tahvanainen et al. 1985, Larsson et al. 1986, Royle & Hubbes 1992). The productive period of a short-rotation willow plantation is expected to be 20-25 years (Anderson et al. 1983, Ledin et al. 1992). Studies on the development of short-rotation plantations have usually been confined to the first years after establishment. Very few 12 reports include records of long-term biomass production and survival, matters of crucial importance for the success of such ventures. Varying views have been expressed in regard to the biomass yields that can be achieved in short-rotation plantations; these have been based on both simulation models and field experiments. Rather high yields (even 30-40 t ha" 1 a"' dry-mass) have been considered possible (Pohjonen 1974, 1980, Siren & Sivertsson 1976, Siren 1979, Hathaway 1979, 1980, Linder & Lohammer 1982, Christersson 1986, 1987, White et al. 1989). The reliability and generalizability of yield results obtained in connection with field experiments is hindered by matters such as smallness of sample plots and exceptionally intensive management (Cannell 1989). 2 AIMS OF THE STUDY The general objectives of the studies reviewed in this paper were to obtain more thorough understanding of the silvicultural and nutritional prerequisites of growing short-rotation forests on cut-away peatland areas and abandoned farmlands. In order to achieve this the following problem areas were studied: 1. Biomass estimation methods. Accurate biomass measurements provide the basis for biomass estimation and nutrient uptake studies. The aim of the first investigation in this study was to compare different biomass estimation methods and evaluate their suitability for the determination of short-rotation biomass production (I). 2. Stand establishment and coppice regeneration. The objective of several studies was to evaluate the success of plantation establishment (11, IV, VI-IX). Lack of knowledge of critical factors affecting the sustainability of coppice regeneration systems was an impetus for research which focused on cultural factors influencing coppicing (II). 3. Stand development and biomass production. The aims were to study the effects of within-stand competition and stand structure (III), effects of planting density on biomass production (IV), and biomass production and its allocation in different compartments (leaves, bark, wood, stump, roots) (IV-IX). 4. Fertilization and nutritional requirements of short-rotation plantations. The specific objectives were to study the effects of fertilizer nitrogen, phosphorus and potassium and their combinations, application rates, repetition of the fertilizer application, the effect of solubility of fertilizer phosphorus and sludge application on the biomass production and nutrition of short-rotation willows in cut-away peatland areas (IV-IX). One of the main aims in investigations concerned with nutrition of short-rotation plantations was to study the nutrient concentrations in the above-ground biomass, amounts of nutrients bound in the above-ground biomass, biomass production and the effects of fertilizer application (IV-IX). In order to gain better understanding of nutrition, biomass, nutrient concentrations, and nutrient uptake of different above ground compartments was studied (VI-VII). The possibility of utilizing soil analysis for determination of fertilization regime was evaluated. 13 3 MATERIAL AND METHODS The field experiments were established on two types of sites potentially suitable for short rotation cultivation: cut-away peatland areas and abandoned farmland (Table 1, Fig. 1). The species and clones used in the different studies are presented in Table 1. All the experimental short-rotation plantations were established using unrooted, 20-25 cm long cuttings, except 111 and V where rooted cuttings were used. Study II also included naturally established stands of willow and birch. The willows were planted in rows: the space between the rows was 70 cm (in 111 and V: 80 cm) and the distance between the cuttings in the rows was 35 cm (planting density 41 000 cuttings ha" 1 , in III,V: 36 000 cuttings ha" 1). Mechanical and manual weeding was done during the planting year and in some cases later on, too. Figure 1. Location of the field experiments. 14 Table 1. Location of the field experiments, soil types and species used in the different studies. Different methods of measuring willow biomass production of Salix 'Aquatica' and S. triandra were compared (I). The harvesting method consisted of cutting the willow to a stump height of 10 cm after which the sprouts were gathered into bundles and weighed immediately. Six sprouts of different sizes were sampled from each plot and their leafless fresh-mass and dry-mass were determined. In the mean stool method, all the sprouts from stools selected at random were cut to a stump height of 10 cm with the sampling covering 7% of the total number of the stools. The total number of the stools on sample plots was counted. The dry-mass was obtained by multiplying the dry-mass of the average stool with the number of stools. When regression estimation was used, height and diameter distributions of the willows were determined by systematic sampling, which, on average, covered 10 - 13% of the area of the sample plots. Thirty sprouts from each willow clone were sampled. The sample sprouts were then cut back to a stump height of 10 cm and their length, diameter and dry-mass were determined. The independent variable in the dry-mass equations (Y = aX b ) was diameter measured at 10 cm height (d 01) and total height (h) combined into a single variable (d 2 h). Prior to computing the dry-mass, the stump height of 10 cm was subtracted from the willow height measurements. The effects of cutting season on the survival, coppicing and growth of S. viminalis, S. 'Aquatica', S. x dasyclados, S. phylicifolia and S. pentandra mixture, and B. pubescens) were studied in a cut-away peatland area, mineral-soil and peat-soil fields at Haapavesi (11, Fig. 1, Table 1). At each of the 32-36 or 53 cutting times 20-30 stools were cut down. At Haapavesi, the effect of stump height on the survival and biomass production of Salix 'Aquatica' was also studied by cutting back one-year-old sprouts in 1983 to stump heights of 0, 10, 20 and 40 cm on plots sized 60 m 2 (II). Randomized block design with four replications was used. Survival and biomass were measured in 1985, 1987 and 1990, when the willows were cut to their initial stump heights. Study Soil type Species and clones I, VII, VIII, IX Cut-away peatland area S. 'Aquatica' (V769) VI, VII S. 'Aquatica' (E4856) II, VII, VIII S. x dasyclados (P601 1) I S. triandra (P6010) I S. triandra (P6291) II S. viminalis (E7901) II Peat field S. phylicifolia, S. pentandra II B. pubescens II Mineral soil field S. 'Aquatica' (E4856) II " S. x dascylados (V761) III, V " S. 'Aquatica' (V769) IV " P. x rasumowskyana 15 The effects of harvesting damage on the sprouting and biomass production of S. 'Aquatica' was studied at Haapavesi on a cut-away peatland area and at Nurmijärvi on a mineral-soil field (11, Fig. 1). One-year-old willows at Haapavesi were cut using (A) seca teurs, resulting in a smooth cutting surface, and (B) a brushsaw, leaving a rougher cutting surface. In addition, half of the stumps were damaged manually with a sledge-hammer in both treatments. Randomized block design with four replications was used. The survival, number of sprouts per stump, and the height of the tallest sprout in each stool were measu red. The dry-mass of the willows was determined in 1985, 1987 and 1990 using the harves ting method. At Nurmijärvi, eight-year-old sprouts were cut with a chainsaw or with a brushsaw (II). The treatments including a control (A), light-weight forwarder driving on the row of stumps (B), and manual damaging of the stumps using a sledge-hammer (C) were replicated three times in a randomized block design. The height of the sprouts and the number of sprouts per stool were measured after one growing season. The dry-mass per stool was calculated using dry-mass equations. The long-term biomass production and stand structure of S. Aquatica' was studied on a limed mineral soil field in Nurmijärvi (111, Fig. 1, Table 1). The willows were planted in 1982 and fertilized with sludge. The willow harvest took place in 1985, three growing seasons after planting. Sludge fertilization was renewed after the harvest, but the nutrient amounts applied were small (18 -70 kg N ha" 1 ). The second harvest took place at the end of the seven-year rotation in 1992. The leafless above-ground biomass was determined annually using allometric dry-mass equations. The stand structure at different ages was studied by constructing equal-interval frequency distribution histograms of willow diameter. The third (g,) moment about the mean, a measure of the skewness of the distribution, and the fourth moment (g2), a measure of kurtosis of the distribution, were also calculated. The effects of spacing (35 000, 15 000, 5000 stems ha" 1 ) on the biomass production of Populus x rasumowskyana , planted using cuttings in 1981, was studied on a former cultivated mineral soil field of satisfactory fertility under pH of 6.1 - 6.9 at Paimio (IV, Fig. 1, Table 1). The poplars were fertilized in 1981 using a multi-nutrient fertilizer (48 kg N ha" 1 , 21 kg P ha" 1 , 39 kg K ha" 1). A corresponding trial at Kannus (Fig. 1) was destroyed by frost the next year and thus only the first year's data are available. The various spacing treatments were replicated twice on the plots (size 400 m 2) selected at random. In the autumn after the first growing season, the longest shoot on each stool was left to grow. The effect of fertilization on the biomass production, foliar, bark and wood nutrient concentrations and the nutrient uptake of willows (V-IX) and poplar (IV) was studied in field experiments established on mineral soil agricultural fields (IV, V) and cut-away peatland areas (VI-IX). The effect of fertilization treatments on soil properties were also studied. The fertilizer application treatments are presented in Table 2. The sizes of the experimental plots were 56 - 80 m 2 (VII-IX), 225 m 2 (IV, VI) or 300 m 2 (V). The willow cuttings were planted at a density of 36 000 (V) or 41 000 cuttings (VI-IX) per hectare. The planting density for poplars was 15 000 cuttings per hectare (IV). The sprouts were cut back after the first growing season (VI-IX) or cut back leaving the longest shoot on each stool (IV). The experimental design used consisted of randomized blocks (VII-IX) or fully 16 randomized treatments (IV, VI, Table 2). The diameter and height distribution of the willow sprouts on the experimental plots were measured each year. The number of living and dead stools was also recorded. Allometric dry-mass equations, based on sample sprouts selected according to the size distribution of the sprouts in the plots with base diameter (d 01) or the product of base diameter squared and height (d 2 h) as an independent variable, were constructed. Sample sprouts were dried to constant weight at 105° C (stem, branches, bark and wood) or at 80° C (leaves). Dry-mass equations were calculated for the leafless above-ground mass, but in VI-IX also for leaf, bark and wood mass and in IV for branch mass. Root and stump masses were calculated in VII-IX with equations based on randomly selected stools (including stems and roots) dug up annually. The independent variable in the stump and root dry-mass equations was the dry-mass of all the sprouts on a stool. The stump mass equations also included the number of sprouts per stool as a variable. Foliar samples were collected from the fertilization experiments V-IX from each plot for nutrient analysis mostly in late August or early September (Table 2) (Halonen et ai. 1983). The nutrient concentrations of the bark and the wood were also determined in VI and VII. Soil samples (composed of five subsamples) were taken from the 0 - 10 cm top soil layer on all the plots (V-IX, Table 2). The samples were stored in freezer (-20 °C). The pH of the dried samples was analyzed in distilled water (V/V 1:5). Soil total nitrogen (Kjeldahl, V-IX), ammonium and nitrate nitrogen (VI), acid ammonium acetate (pH 4.65) extractable phosphorus, potassium, calcium, and magnesium (IV-IX) were also determined (mg 1"', volume determined in laboratory) (Halonen et ai. 1983). 17 1) Prior to planting all experimental fields were limed with 6 000 kg ha' 1 of dolomite lime, except IV (no soil amelioration) and Valkeasuo (VII) where either 12 000 kg ha" 1 wood ash or 12 000 kg ha' 1 dolomite lime were used. In VII-IX Piipsanneva and Paloneva experimental areas also basic fertilization with NPK and in Valkeasuo limed area with PK. 2) In VI and VII also wood and bark samples were collected. 3) sf = superphosphate, rf = rock phosphate, ap = apatite. Table 2. Fertilization treatments in studies IV - IX. See figure 1 for the location of the study sites. Study Species and clones No. repli- cations Fertilization treatments 0 Fertilizer application year Soil samples year Foliar samples' 2 year IV P. x rasymowskyana 2 O, N, 2N, 3N (N = 100 kg N ha 1 ) 1981, 1982 - - V S. 'Aquatica' 3 NPK, Sludge 30, 60, 120 m" 3 ha 1 Sludge: 1982, NPK: 1982,1983,1984 Spring 1983, Spring 1984 1983, 1984 VI VII S. 'Aquatica' S. 'Aquatica' S. x dasyclados S. 'Aquatica' S. 'Aquatica' S. 'Aquatica' 4 3 3 3 4 4 0, NK, NP sf K (3 , NP rf K, NP ap K (N=200 kg N ha 1 , P = 87 kg P ha 1 , K = 166 kg K ha" 1 ) 0, P, K, N, PK, NK, NP, NPK (N = 100 kg N ha" 1 , P = 30 kg P ha 1 , K = 40 kg K ha 1 . 1981, 1983 1983, 1984, 1985 Autumn 1983 August 1985 1983 1983, 1984, 1985 VIII S. 'Aquatica' 5. 'Aquatica' S. x dasyclados 3 3 3 Fife N (0, 50, 100, 150, 200 kg ha" 1 ), P (0, 15, 30, 45, 60 kg ha" 1 ) and K (0, 20, 40, 60, 80 kg ha" 1 ) levels. When the amount of one of the nutrients in NPK-fertilization was changed others remained unchanged (N 100, P 30 and K 40 kg ha' 1 ). 1983, 1984, 1985 August 1985 1983, 1984, 1985 IX S. 'Aquatica' S. 'Aquatica' 3 3 PK, NP, NK, NPK, N2 PK, NP 2 K, NPK 2 , where N = 100, N 2 = 200 kg N ha" 1 , P = 30, P 2 = 60 kg ha ', K = 40, K 2 = 80 kg ha 1 . Reference treatment fertilized in 1984 compared with repeated fertilizations in 1984, 1985, 1986. Annual: 1983, 1984, 1985 Once: 1983 August 1985 1983, 1984, 1985 18 4 RESULTS AND DISCUSSION 4.1 Estimation of biomass Small experimental plots (even less than 1 m 2 in area) have often been used in research on short-rotation cultivation (e.g. Pohjonen 1974, 1977, Lepistö 1978, Hathaway 1979, 1980). Even though small plots can be effective in screening clones or fertilizer treatments, it must be taken into consideration that the biomass production figures obtained from such small plots can be biased due to errors caused by the so-called edge effect (Zavitkovski 1981). Trees growing at the borders of experimental plots may be in more advantageous or disadvantageous positions regarding nutrients, moisture, radiation and temperature than trees growing in the interior parts of the sample plots. Biomass production in the border areas is generally much higher than in the central parts of the plots (Zavitkovski 1981, Stott et al. 1983). The early literature often includes biased results from small plots expressed as production per hectare (Cannell & Smith 1980); e.g. early production figures of Populus in Wisconsin based on small plots have been revised from 25-30 t ha'a 1 to about 10 t h a"'a ' (Isebrands et al. 1979). Yields of S. 'Aquatica' on square plots (50-70 m 2) were 20-30% lower when one border row was excluded (Stott et al. 1983). According to Stott et al. (1983) one outside row could suffice for the elimination of the edge effect for up to 3-year old willows (10,000 - 40,000 plants ha' 1 , 40-70 m 2 plot size), but two would allow the 'safe' assessment of production per hectare. In all the studies (I-IX) the practice has been to always exclude one, usually two or more, border rows from being measured. The said studies also meet the criteria set by Cannell and Smith (1980) concerning the stipulatum that the ratio of the height of the measured trees (inside the plots) to their distance from the edge of the plot (i.e. the border width) should not exceed four. Most methods for the determining of the dry-mass of a tree involve the measurement of the moisture content of sample trees and of their different compartments (for woody compartments usually by weighing green, drying at 105 °C to constant weight and reweighing). Accurate determination of the moisture content is extremely important in biomass estimation (I, Ferm & Hytönen 1984). The moisture content in a tree varies; e.g. by tree species, tree age and size, longitudinally along the stem, by season and even the time of day, weather, growing site and fertilization (I, Hytönen & Ferm 1984, Ferm 1985b, Hytönen 1987, Hakkila 1989). There are several sources of error related to sampling for moisture content, storage of moisture samples and measurement of these samples. Determining the mean moisture content of a single willow sprout using one sample is more difficult than with birch (Hytönen & Ferm 1984, Hakkila 1979, Auclair & Metayer 1980, Björklund & Ferm 1982). The green mass of the moisture samples should be measured without delay, since the moisture content of the samples decreases during storage depending on the duration and method of storage (Ferm & Hytönen 1984). Incorrect storage methods can cause even losses of dry-matter; e.g. when dealing with foliage samples (Nilsson 1983). The drying time should be long enough to allow water to leave the sample. Many sources 19 of error related to sampling and subsampling for moisture content were avoided by taking entire sprouts as samples (I-111, V-IX). The harvesting method, where all material within a unit area is harvested and the biomass is subsequently weighed, is also subject to various errors (I). Accurate measurement of the harvested area is important. Even the actual cutting height has a significant effect on the amount of biomass harvested (I). The weighing technique may be another source of error (Björklund & Ferm 1982) and weighing large biomass amounts is labour intensive. The harvesting method is not well suited for determination of biomass compartments. This is also true for the mean tree method. A single tree of mean dimensions will not provide mean weight for all biomass compartments. The mean stool method, where all the shoots of a number of stools are cut and weighed, is better suited for coppice biomass estimation. The most common method for estimating tree biomass is through the use of regression analysis (Hitchcock & McDonnell 1979, Crow & Schlaegel 1988). Generally the dry weights of destructively harvested sample trees and their compartments are related by regression equations to a readily measurable dimension or combination of dimensions. The dimensions most frequenctly used in regression analysis and describing the allometric structure of trees include tree height and diameter. Often the measurement of diameter alone is adequate since adding height into the models together with diameter (d 2 h) increases the degree of determination only little (I, V, VI, Payandeh 1981, Björklund & Ferm 1982, Ferm & Kaunisto 1983, San Miguel & Cancio 1985, Hakkila 1989). The most widely used biomass model is the allometric model y, = where y—the weight of the ith sample tree, X, = the value of the independent variable of the ith sample tree, a and b = the model intercept (a) and slope (b), e, = the random error associated with estimating the weight of the ith sample tree, i = represents any one of the sample trees (Crow & Schlaegel 1988). Logarithmic transformations, likely to equalize the variance over the range of y-values, have been used to linearize this allometric equation (InY; = lna + blnXj + Ine,). When untransformed arithmetical units are desired, several procedures to correct the slight underestimation caused by logarithmic transformation (Satoo & Madgwick 1982) have been advocated, and that proposed by Meyer (1941) is most often used (correction with s 2 e /2, where sis the residual of the equation). This correction was usually slight (11, VI). Logarithmic transformations and the subsequent linear regression of biomass on tree dimensions may result in biased estimates when a non-zero intercept is present in the untransformed data (Verwijst 1991 a). This bias was avoided by taking the diameter measurements at harvest level (I, 111-IX). Besides the power function, several other regression models (e.g. weighed linear regression) have been used (e.g. Cunia & Briggs 1984, Crow & Schlaegel 1988). The stem, bark, wood, and leaf mass, and also the leaf area of willow were accurately described by allometric regression equations (I, 111-IX). Since the allometric exponent b in the biomass equations (Y = aX b ) changed with age (111- VII) generalized biomass equations should most probably be age-specific. With increase of age the growth form and density of willow bark and wood change (Hytönen & Ferm 1984). The number of sample trees sacrificed in fitting models in previous studies has varied from less than ten to over a hundred. Usually, however, as in the present studies (I, 111-IX), 20 this has been between twenty and thirty. Biomass equations based on smaller number of sample trees have also been used (e.g. 5-6: Gholz et al. 1979, 7-9 Finer 1989, 1991). The most cost-effective means of increasing the precision of biomass estimation would be to sacrifice more trees to fit the models or to sample larger areas in the stand; more sophisticated models would have little effect on the error (Woods et al. 1991). The sample trees often need to be subsampled. The representativeness of internal sampling should be carefully considered. On the part of leaf mass and leaf area, the time of the year when sampling is done can lead to errors: leaf mass and leaf area reach their maximum in central Finland in late August (Ahola 1987), and in Sweden between between early August and late September (Nilsson 1985, Nilsson & Ericsson 1986). The sampling in the present studies was mostly done in late August (VII-IX), but in the case of VI so late that the leaf mass was already declining. Sometimes the regression models used can lead to inaccuracies. A common problem is the poor additivity of the masses obtained using dry-mass equations for the different compartments of a sampled tree. When the dry-mass equations for the different compartments are calculated independently, and when the sampling errors of the different compartments differ, the result might be a group of regression equations behaving irra tionally with respect to each other (Cunia & Briggs 1984). Also, missing data for some compartments can hinder additivity (Kozak 1970). Several procedures for solving the problem have been presented (see Jacobs & Cunia 1980, Cunia & Briggs 1984, Chieynda & Kozak 1984). In the present experiments, the same independent variables were used in the dry-mass models for the different compartments except for root and stump mass (IV, VI-IX). It should especially be kept in mind that extrapolation over the sample tree range can cause considerable errors, but predictions from allometric equations can be extrapolated more readily than from weighted linear equations, for instance (Crow & Schlaegel 1988). Thus, major errors in estimating of dry-matter production per hectare may be caused by poor sampling and by the regression equations used (Cannell 1989). All biomass figures were calculated supposing a cutting height of 10 cm. In Sweden a fixed cutting height of 5 cm above ground level has been used (Telenius & Verwijst 1995). Since the actual stool height following harvesting with brush saw or commercial willow harvesters has been 10-20 cm (I, Telenius & Verwijst 1995) biomass estimates differ from the actual amount of harvested biomass. Height and diameter distribution have to be measured when calculating the biomass using dry-mass equations. Similar systematic sampling procedure than used in I, 111, V-IX gives according to Telenius and Verwijst (1995) goood estimate of the stem frequency distribution. It is often difficult to measure the height of willows accurately, especially of clones with contorted stems (I, V). Usually, there are many small sprouts in a willow stand. Even though the number of small sprouts is high, their biomass is very small compared to the mass of the larger sprouts (III). It would have been possible, when measuring the one year-old plantation, to omit over 50% of the sprouts without underestimating the stem mass by more than 5%. The tallest 10% of sprouts contained 45% out of the dry-mass. When the dominant height of willows was 200 cm, all sprouts shorter than 110 cm could have been 21 left unmeasured. The amount of work can, therefore, be considerably reduced and the measurements can be concentrated on sprouts containing the highest amount of biomass. The various methods for measuring the biomass of short-rotation plantations produced some differences between the biomass estimates (I). The basal area ratio method, sometimes used in biomass studies consistently overestimated the willow biomass (I). The most laborious method was the harvesting method, whereas the amount of work and expense incurred in the other methods were less. Regression estimation as a non-destructive method is especially suitable when the plantation is to be grown further or when estimates of annual biomass production during a sequence of years are needed. 4.2 Establishment, regeneration and stand structure The establishment of the short-rotation plantations with cuttings was successful (IV, VI-IX). While the percentage of rooted cuttings was high additional planting was done to secure the evenness of the plantations. The size of the poplar cuttings (IV) had a significant effect on sprouting and early sprout growth. Thicker willow and poplar cuttings root easier and develop longer and thicker sprouts during the first growing season (IV, Singh & Chaukiayal 1983, Koo et al. 1986, Burgess et al. 1990, Rossi 1991). Short-rotation forestry utilizes the exceptional growth rates of coppice shoots. The reasons for the rapid early development of coppice shoots are not fully understood, but alte red shoot-root ratios and large root systems are important factors. A factor critical to biomass plantations is the viability of the coppice system over several successive rotations (Sennerby-Forsse et al. 1992). Several factors, both internal and external, influence the regeneration of stumps following cutting (Kauppi 1989, Ferm & Kauppi 1990, Paukkonen et al. 1992). Many cultivation-related factors, e.g. stump diameter, age of the tree, site quali ty, spacing and the harvesting cycle, influence coppicing. The timing of cutting had a marked impact on the height growth, biomass production, and survival of the exotic willow species (II). The best regrowth occurred when the plants were cut during the dormant stage, i.e. between late autumn and early spring. Cutting at the end of July, or beginning of August, had a highly detrimental effect on survival. Besides willows, many other species also exhibit seasonal variation in coppicing, the dormant season being superior in minimizing mortality and usually also increasing the number and growth of the resulting sprouts (Deßell & Alford 1972, Anderson 1979, Belanger 1979, Blake & Raitanen 1981, Blake 1981, Harrington 1984, Webley et al. 1986). Short-rotation willow plantations should be harvested during the dormant period as recommended in old textbooks on the cultivation of basket willows (11, Flinta 1882, Nordberg 1914, 1928). Thus, the harvesting of protein-rich leaves for fodder (as suggested by, for example, Näsi & Pohjonen 1981, Näsi 1983, Pohjonen & Näsi 1983) during the growing season would markedly decrease the vitality of the plantations. There are considerable inter-species differences in the reaction to the timing of cutting 22 (Blake 1981, Blake & Raitanen 1981, Webley et ai. 1986). This was most clearly evident in regard to survival (III). Contrary to the behaviour of exotic willows, the survival of downy birch or indigenous willow species was not affected by the timing of cutting; their survival exceeded 80% throughout. In the case of birch, this has been observed to be the case also in earlier studies (Etholen 1974, Johansson 1992 a). As in other experiments, the height growth of birch was slightly affected by the timing of cutting, with a minimum in June - July (11, Heikinheimo 1930, Mikola 1942, Leikola & Mustanoja 1961, Andersson 1966, Etholen 1974, Moilanen & Oikarinen 1980, Ferm & Issakainen 1981, Ali-Alha 1987, Johansson 1992 b). This was also true of native willow (II). Differences in the height growth of birch caused by the timing of cutting levelled off within seven years. Initial differences caused by the timing of cutting may thus disappear later on (Ciancio & Menguzzato 1985). The reasons for differences in coppicing due to timing of the cutting are not fully un derstood. Some earlier studies reviewed by Blake (1981) have linked differences in sprout growth to differences in the levels of carbohydrate reserves in the roots of parent trees in the season of cutting. However, the carbohydrate levels have been shown to be adequate for coppicing under most conditions (Blake 1981, Blake & Raitanen 1981, Johansson 1992 a, Blake 1983) and carbohydrate concentrations and sprout growth do not correlate well. The highest number of sprouts for downy birch resulted from being cut back in the summer. Sycamore ( Platanus occidentalis L.) produced most sprouts when cut in July (Belanger 1979). Also the buds of exotic willow species burst even when cut in late summer or early autumn (II). In the beginning of winter such sprouts are small and their moisture content is high. One reason for poor coppicing vigour and increased stump mortality following late autumn cutting may, thus, be in the death of these small sprouts due to frost. Mikola (1942) and Johansson ( 1992a,b) considered frost risk to be considerable for birch, too. The axillary buds of downy birch do usually burst only after winter dormancy, but the tree is unable to maintain the same control over its basal buds (Kauppi et ai. 1987). Also, the proportion of frost-hardy internodes of sugar maple (Acer saccharum Marsh.) in sprouts arising following post-growing-period cutting decreased with each successive cut (Mac Donald & Powell 1985). In this study (II), harvest damage had a negative effect on the survival, height growth and biomass production (26 - 54% in different rotations) of a young plantation of S. Aquatica'. In the older, well-established plantation (II), the number of sprouts per living stool and the biomass per stool was slightly lower following harvest damage caused by a mini-forwarder driven on the stumps. However, these differences were not statistically significant. Damage to birch stumps has not been observed to affect the survival or height growth of sprouts (Mikola 1942, Leikola & Mustanoja 1961, Ferm & Issakainen 1981). The differences between birch and willow in relation to the effects of harvest damage may be due to the location of the sprout-producing buds. About 90% of the basal buds of birch are located below ground level (Kauppi et ai. 1987, 1988) while most buds of Salix 'Aquatica' are above ground level (Paukkonen et ai. 1992). In willow coppice most sprouts originate from the axillary bud groups located on the remaining basal parts of the previously harvested stems (Sennerby-Forsse et al. 1992). Thus, harvesting damage may have more 23 serious effects on willow than on birch regeneration. Contrary to instructions on the harvesting of willow with a sharp blade without damaging the bark (Nordberg 1914, 1928, Tapio 1965), smooth cutting surfaces did not give any better coppicing results than rougher cutting surfaces made with a brushsaw (II). The sustainability of the coppice system should be taken into account when designing willow harvesters. Stump height seems to have only a minor effect on the first rotation's yield (II) of willow and on the sprouting of poplar (Populus trichocarpa : Deßell & Alford 1972). Howe ver, in successive rotations, stump height was of crucial importance, with short stumps producing more biomass (as much as 70%) than high stumps (II). This is in agreement with the recommendation to cut willows at ground level (Nordberg 1928). However, according to the results obtained, stumps as high as 10 cm could be used without decreasing producti on. The shoots on low-cut stumps are more likely to be connected to individual roots, and this is believed to give some advantage in terms of availability of water and metabolites (Sennerby-Forsse et al. 1992). High stumps may increase the risk of fungal infection and decay and increased breaking away of sprouts from the stumps. Salix 'Aquatica' is able to withstand several repeated harvests (Paukkonen et al. 1992). However, the coppicing ability of some tree species decreases if they are coppiced using short rotations (Platanus occidentalis: Steinbeck & Brown 1976, Populus : Strong 1989). The productive period of the plantations has been estimated to exceed 20 years (Anderson et al. 1983, Ledin et al. 1992). In this study (III), stool mortality increased year by year. After 10 years from the establishment of the stands, over one-third of the stools had died. In closely spaced poplar and willow stands, stool mortality is higher than in wider spacing (IV, Heilman et al. 1972, Bowersox & Ward 1976, Hytönen 1982). Due to mortality, the initial differences in spacing are reduced during the rotation (IV, Hytönen 1982). Increasing the supply of nutrients intensifies the competition process (Ford 1984). Nitrogen fertilization was shown to increase stool mortality (VIII) and decrease the rooting of cuttings (Hytönen 1984). Despite a stool mortality of 34%, the plantation (III) was still productive, containing approx. 24 000 live stools per hectare, which is more than the planting density currently used in Sweden (Sennerby-Forsse & Johansson 1989). The planting density could, most probably, be much lower if longer rotations were applied. Increased post-harvesting coppicing compensates for the loss of stools and increases the number of sprouts by three to four times (III). After harvesting, the number of sprouts was 339 000 ha" 1 (III). Thus, during the second and following rotations, canopy closure proceeds faster than during the first rotation. This implies that competition should start earlier in subsequent rotations (111, Willebrand & Verwijst 1993). Stool mortality (III) was most probably due to competition, especially during the second seven year long rotation. The reduction in the number of sprouts with increase in stand age was an indication of high within-stand competition (III). The number of living sprouts in the stand was much smaller than the total number of standing sprouts. Altogether 40% of the first year's initial shoots survived to the end of first three-year-long rotation and only 13% to the end of the second seven-year-long rotation (III). Similar cumulative shoot mortality has been observed also in a Salix viminalis plantation in Sweden after three years (Verwijst 1991b). Mortality was at 24 its greatest between the ages of one and two years (111, Verwijst 1991b). Already after the first growing season, 13% (first rotation) and 20% (second rotation) of the total number of standing shoots were dead. Standing dead stems in the plantation are, however, harvestable. Because standing dead stems were lost by breaking, their number remained constant during the second rotation even though shoot mortality continued. Thus, besides the decrease in the wood density of the dead shoots (Verwijst 1991b), more biomass is probably lost as the result of breaking off of dead shoots. Competition changes the size and weight distributions in a population. In this study (III), the diameter distribution of one- and four-year-old sprouts was clearly bimodal as described also in connection with dense, even-aged, single-species stands by Ford (1975) and Mohler et al. (1978). The majority of the sprouts were short and thin. Even though their number may be high, the biomass of declining sprouts, which die at a later stage, is very small compared to the mass of larger sprouts. The death of the smallest sprouts, i.e. the first peak in the distribution, began during the first growing season. The second, smaller peak, was made up of the group of dominant sprouts. The significance of the second peak in height and diameter distribution is more important, however, from the viewpoint of the total sprout dry-mass. The positive skewness of the weight and size distributions reaches a maximum immediately before the suppressed plants begin to die (Mohler et al. 1978, Wille brand & Verwijst 1993). This phenomenon manifested itself already before the end of the first growing season (III). The basal axillary buds in willow consist of a single bud scale covering three shoot primordia, the larger one in the middle giving rise to taller shoots than the two laterally placed buds (Brunkener 1984, Paukkonen et al. 1992). This may have contributed to the first year's bimodality in the distributions. In the stand of willow, (III), the second self-thinning phase began at the age of four years and manifested itself as renewed bimodality of the stem diameter frequency distribution. It seems that self-thinning proceeds slower during the second self-thinning phase. The death of bigger stems is probably a slower process than the death of one-year old sprouts. Reduction in the skewness caused by the death of the most suppressed plants, as self-thinning proceeded, was observed in the frequency distributions. Kalela (1962) has also reported a death peak in stands of Salix caprea at the age of seven years with another mass death peak expected ten years later. According to Ford (1975) and Ford & Newbould (1971), bimodality of the stem diameter distribution indicates a disjunct distribution in growth rates. The increase in the relative growth rate from small plants to large is not uniform throughout the range of plant sizes and depends on the competition process (Ford & Newbould 1971). In coppice stands, there is competition between large shoots in the upper parts of the canopy where leaves receive direct radiation. However, once a shoot is overtopped, it exists in a markedly less favourable, but fairly constant, environment of diffuse radiation (Ford 1975). 25 4.3 Nutrition 4.3.1 Site characteristics The short-rotation plantations monitored in this study were established in cut-away peatland areas and on abandoned mineral-soil agricultural fields. Mineral and peat soils differ from each other not only by their physical properties but also in regard to their biological and chemical characteristics. The greatest differences are in their pore volume, the proportions of organic and inorganic matter, and in the amounts of nutrients bound in them (Kaunisto & Päivänen 1985, Westman 1991). There were some marked differences between the cut away peatland areas regarding their nutrient concentrations (VI-IX). Generally, it has been reported that nitrogen concentrations of cut-away peatland areas are high and those of potassium and phosphorus low (Table 3, Kaunisto 1979, 1982 a, 1983, Ferm & Kaunisto 1983, Lehtonen & Tikkanen 1986, Hytönen 1987, Ferm & Hytönen 1988, Lumme 1989, Kaunisto & Viinamäki 1991). Especially at Paloneva, peat nitrogen concentrations were high, but phosphorus, potassium and calcium concentrations low compared with the other experimental areas (VI-IX, Table 3). According to the classification system of cultivated soils applied in Finland, the extractable phosphorus concentration in all the cut-away peatland areas (Table 3) was poor (< 2.0 mg l" 1), and that of potassium and calcium also poor (K < 30 mg l" 1 , Ca < 600 mg l" 1 ) and only exceptionally rather poor (Viljavuustutkimuksen.... 1992). In peatlands, especially in the deeper layers, only a small proportion of the total phosphorus is available to plants and and as much as 80-95% of the total phosphorus may be in the organic form (Kaila 1956). According to Aro (1995), cut-away peatland areas may contain phosphorus in amounts similar to natural nitrogen-rich peatlands. Potassium is concentrated in the surface layers of peatlands and the potassium concentration is very low in the deeper layers (Kaila & Kivekäs 1956, Pakarinen & Tolonen 1977). Most of the potassium in peat is exchangeable. Potassium as a monovalent cation is bound only loosely in organic soils. Thus, part of the potassium can be lost through leaching, especially if high doses of potassium are used (Ahti 1983). Mixing soil from ditches, soil cultivation and basic fertilization before the start of some of the experiments (VII-IX) increased the site-to-site variation; e.g. mixing clay mineral soil into peat fields, a common practice in Finland, has increased the amounts of total potassium in soil. Peat layer in the experimental fields was so thick, over 30 cm, that willow roots most probably did not penetrate into the subsoil (Ericsson et al. 1983, Ericsson 1984, Elowson & Rytter 1984). The pH of cut-away peatland areas is almost without exception too low for the cultivation of many willow species (Lattke 1969, Ericsson & Lindsjö 1981, Ferm & Hytö nen 1988). In terms of their pH, all the areas would be classified as poor or fairly poor compared with cultivated soils (Viljavuustutkimuksen... 1992). Therefore, all experimental fields in the cut-away peatland areas were limed or ash fertilized (VI-IX). The pH of agricultural fields can be high enough for the cultivation of many willow species (IV, Kurki 1982, Viljavuustutkimuksen... 1992). Even though liming and application of ash increased 26 the pH of the cut-away peatland areas, it still remained below the optimum for S. viminalis root growth (Ericsson & Lindsjö 1981) but probably did not limit willow growth even at the Valkeasuo site, where the pH was the lowest (VII, Ferm & Hytönen 1988). The amounts of liming agents should be fairly high in order to increase the low soil pH of cut-away peatland areas to 5.0 - 5.5 for the lifetime of the plantation throughout the soil-tilling and root zone. Table 3. Ash content, pH, total nitrogen, acid ammonium acetate extractable phosphorus, potassium, calcium and magnesium concentrations of unfertilized cut-away peatland areas according to some investigations conducted in Finland. 1) Location of cut-away peatland areas: Aitoneva at Kihniö, Hirvineva at Liminka, Katinhännnsuo at Vihti, Paloneva at Rantsila/Ruukki. Piipsanneva at Haapavesi, Valkeasuo at Tohmajärvi, Osmanginsuo at Kiuruvesi. 2) Liming. 3) From organic matter. 4) Basic fertilization with PK or NPK four years before analysis 4.3.2 Effect of fertilization on soil nutrients In cut-away peatland areas fertilization with phosphorus (superphosphate) and potassium (potassium chloride) increased the amounts of the corresponding acid ammonium acetate extractable nutrients in soil manyfold (often tenfold) compared to the control plots (VI-IX). However, even after two or three annual PK fertilizer applications, the extractable phosphor us and potassium concentrations in peat were rather poor (VI-IX) and only on the experi mental field ameliorated with wood ash the P and K concentrations were good (VII) Area 0 pH Ash Peat N P K Ca Mg Author no amelio- content, depth, tot., % mg l"1 mg r 1 mg r 1 mg r 1 ration % cm Aitoneva 3.8 _ . 1.8° . _ - Kaunisto 1982a Aitoneva 3.7 14 - 1.7 - - - Kaunisto 1983 Aitoneva 3.9 - 55-87 2.4° - - - Kaunisto 1987a Aitoneva 3.6 16 38 1.7 2.8 55 - Ferm & Kaunisto 1983 Aitoneva 3.9 - - 1.4 1.7 15 350 77 Ferm & Hytönen 1988 Hirvineva 4.5-5.0 39 - 1.5 0.8 10 679 110 Lumme et ai. 1984 Hirvineva 4.5 - 10-100 1.5 0.8 11 216 39 Lehtonen & Tikkanen 1986 Hirvineva 4.8 - 20-40 2.0 0.0 9 520 87 Lumme & Törmälä 1988 Katinhännänsuo - - >100 1.6 1.6 50 2670(2 132 (2 Hytönen 1987 Paloneva 4.9 - - 2.4 1.2 7 618 Hytönen 1984 Paloneva 4.9 - - 2.7 1.3 13 488 98 Ferm & Hytönen 1988 Paloneva 4.1 6-10 152 3.0 0.9 8-15 533(2 157 (2 VII-IX, Hytönen unpubl. Paloneva - 91 2.7 1.9 5 675 244 VI Paloneva - 10 - 2.5 1.1 10 650(2 171 (2 Hytönen 1987 Piipsanneva 3.9 39 - 1.4 (2.4°) - - - Kaunisto 1982a, 1983 Piipsanneva 3.9 - 59 3.4 - - - Kaunisto 1987a Piipsanneva 4.8 - - 2.1 1.0 5 660 100 Ferm & Hytönen 1988 Piipsanneva 4.2 16-29 40-58 1.8 0.2-1.8(4 12-22 14 830-1430 (2 210-520' 2 VII-IX Piipsanneva 4.8 81 1.7 1.1 12 742 Hytönen et ai. 1995 Osmanginsuo 4.5 - 40-74 2.5,2.4° - - - Kaunisto 1982a, 1987a Valkeasuo 4.3 - - 1.4 1.2 5 450 35 Ferm & Hytönen 1988 Valkeauso 4.2 - - 2.3 2.2 13 - Hytönen 1984 Valkeasuo 3.9 24-28 42-46 1.5 1.2 17 322 45 VII, Hytönen unpubl. Valkeasuo 4-5 - - 0.4 20 420 32 Heikkilä 1986 Valkeasuo 4.2 - 94 1.8 1.1 37 465 - Hytönen et ai. 1995 27 according to the classification system for cultivated soils in Finland (VI-IX, Kurki 1982, Viljavuustutkimuksen... 1992). Also the phosphorus and potassium concentrations of mineral soil fields were low (V) or satisfactory (IV) according to the classification system applied. Changes in peat phosphorus and potassium concentrations depended on the amount of nutrients given in the fertilizer applications (VIII, IX, Kaila 1959, Saarela 1982, Hytönen 1987) and in the case of phosphorus on the type of the phosphorus fertilizer (VI). The higher the phosphorus fertilizer amount the higher the soil phosphorus concentration after three year study period. Single-application of 60 kg ha" 1 of fertilizer phosphorus could be detected in soil analysis after three years from fertilization. Phosphorus from easily soluble compounds increased the acid ammonium acetate extractable peat phosphorus concentration (VI). On limed peatland sites, slowly soluble phosphorus fertilizers (rock phosphate, apatite) failed to increase the amount of extractable phosphorus in the substrate and did not ensure the availability of phosphorus for willows (VI, Kaunisto 1983, Yli-Halla & Lumme 1987). This is probably due to the fact that a rise in pH caused by liming decreases considerably the solubility of apatite (Salonen 1968, Karsisto 1973, 1976). Since the solubility of phosphorus in wood ash is greater than that in rock phosphate ash fertilization increases considerably the extractable phosphorus concentrations in peat (VII, Kaunisto 1983, Ferm & Hytönen 1988, Lumme & Laiho 1988). In Carex peat readily soluble phosphorus can be bound tightly by A 1 and Fe (Yli-Halla & Lumme 1987). Besides commercial fertilizers also wood ash can be used to increase the soil potassium concentration (VII, Kaunisto 1983, Ferm & Hytönen 1988). Fertilization with nitrogen only decreased the peat phosphorus and potassium concentrations (VII, cf. Ferm & Hytönen 1988). This was probably due to the higher biomass production and increased phosphorus and potassium utilisation of willows fertilized with nitrogen. 4.3.3 Nutrient concentrations in willow Analysis of foliar mineral nutrient concentrations has been used for a long time to diagnose nutrient status and the fertilization need of trees (e.g. Paarlahti et al. 1971, Paavilainen 1979). The time of leaf sampling is important since foliar nutrient concentrations change during the growing season. Towards autumn the foliar nitrogen and phosphorus concentrations of willow decrease while that of calcium increases and that of potassium may increase or decrease (Lehtonen & Tikkanen 1986, Saarsalmi 1984, Elowson & Rytter 1988, Rytter & Ericsson 1993). Because the foliar nutrient concentrations of willows (e.g. S. Aquatica', Lehtonen & Tikkanen 1986) vary also from the base to the top of the shoot, subsampling of foliage from different parts of the shoot can also cause variation in the results. In these studies, foliar samples were taken along the whole length of the shoots and mainly at the end of August (VII-IX), but in studies V and VI from the upper parts of the shoots. According to Rytter and Ericsson (1993), the most appropriate time for leaf sampling of S. viminalis is during the phase of the most intensive growth. Samples for bark and wood nutrient analysis included whole shoots, with wood and bark separated, thus 28 avoiding possible errors related to subsampling (VI, VII). Foliar nitrogen, phosphorus and potassium concentrations of fertilized S. 'Aquatica' and S. x dasyclados have mainly been in the range of 23 -42 g kg' 1 , 2- 5 g kg" 1 and 10 - 24 g kg" 1 , respectively, in experiments conducted in Finland (V-IX, Table 4, Lumme et ai. 1984, Saarsalmi 1984, Ferm 1985 a). The willows in this study, VI-IX, responded readily to fertilizer applications: as in other studies, nitrogen, phosphorus and potassium fertilization increased the corresponding foliar nutrient concentrations (Kaunisto 1983, Hytönen 1987, Ferm & Hytönen 1988). According to Ericsson (1981b), the decisive factor governing mineral uptake in Salix is the rate of nutrient supply. Response was related to the amount of nutrients applied (VIII), the number of applications (IX), the concentrations of soil nutrients (VII, VIII) and, in the case of phosphorus, also to the type of fertilizer (VI). Only superphosphate increased the phosphorus concentration in the foliage, bark and wood, while concurrently slightly decreasing their potassium concentrations (VI). Annual fertilizer treatment is needed in order to keep foliar nitrogen concentrations at a high level, whereas higher application rates of phosphorus and potassium at the establishment phase may compensate for the effect of annually repeated fertilization (IX). Fertilizer application can be used to adjust foliar nitrogen, phosphorus and potassium concentrations and also foliar nutrient ratios; e.g. the foliar nitrogen concentrations of willow clones demonstrating good production have been 30 - 40 g kg" 1 (Rytter & Ericsson 1993) and foliar nitrogen concentrations of short-rotation poplar should be maintained at a level of over 30 g kg" 1 to achieve good growth (Hansen et al. 1988). Sludge fertilization increased foliar nitrogen concentrations the more the higher the amount of sludge used, but decreased foliar phosphorus concentration at the same time (V, Lumme & Laiho 1988, Lumme 1989, Simon 1989 a, 1989b). Nitrogen fertilization increased the nitrogen concentration of bark by 1 - 3 mg g" 1 and phoshorus fertilization increased the phosphorus concentration in both the bark (0.2 - 0.4 mg g" 1 ) and wood (0.2 - 0.5 mg g" 1 ) (VI, VII). Potassium fertilization increased foliar potassium concentrations but not those of wood and bark (VII). Especially the nutrient concentrations in the bark and wood, but in some cases also that in leaves, changed with increase in willow age, so that older willows tended to have lower nutrient concentrations (VII). Especially nitrogen, phosphorus and potassium concentrations in one-year-old willow bark were high (VII, Table 4). However, the phosphorus and potassium concentrations in the wood changed only little with increase in age. Bark calcium concentrations increased with age (VII). Differences in nutrient concentrations in the bark and wood are high (VI, VII, Ferm 1985 a). Some of the clonal and between-species differences that have been found in some studies in stem (wood and bark) nutrient concentrations of willow sprouts (Viherä- Aarnio & Saarsalmi 1994) can be attributed to differences in sprout size and consequently the varying amounts of wood and bark in the stem biomass. 29 Table 4. Nutrient concentrations (mg g1) in the above-ground parts of some tree species at different ages (years) according to the results of some investigations conducted in Finland. All willows were fertilized with NPK, except those in V (sludge), Viherä-Aarnio & Saarsalmi (1994) (no fertilization), Lehtonen & Tikkanen (1986) (N+peat ash). Also the birch, pine and spruce stands of Fin6r (1989) were fertilized with NPK. Alders (Saarsalmi et al. 1985, 1992) were fertilized with wood ash and birches (Saarsalmi et al. 1992) were fertilized with N + wood ash. The growing sites were mineral soil fields in V, Saarsalmi et. al. 1985, 1992, Viherä-Aarnio & Saarsalmi 1994, cut-away peatland areas in VI, VII, VIII, IX, Lehtonen & Tikkanen 1986, Vaccinium type and Myrtillus type forest soils in Mälkönen 1974, Oxalis-Myrtillus type forest soils in Saarsalmi et al. 1991 and in Mälkönen 1977 ( Betula pubescens 84%, B. pendula 16%). The birches in Finer 1989 were growing on a herbrich sedge pine mire and pines on a herbrich pine mire and on an ordinary sedge pine mire and spruces on a Vaccinium myrtillus spruce mire. The pines in Paavilainen 1980 were growing on a dwarf shrub pine swamp and in Finer 1992 on a low-shrub pine bog. Species Age Nitrogen Phosphorus Potassium Calcium Magnesium Author years Foliage Bark Wood Foliage Bark Wood Foliage Bark Wood Foliage Bark Wood Foliage Bark Wood S. 'Aquatica' 1 22 . . 4.8 - - 17 . . 10 . . 3 . . Viherä- Aarnio & Saarsalmi 1994 1 24-37 18-21 5-6 1.7-2.7 1.8-2.5 0.6-0.8 8-20 8.0-8.6 1.5-2.5 - - - - - - VII, IX " 1 27-38 - - 5.4-15.5 - - 5-17 - - 2-12 - - 2-11 - - • Lehtonen & Tikkanen 1986 S. x dasyclados 1 30-35 - - 2.1-2.4 - - 16-21 - - - - - - - - VII, IX S. 'Aquatica' 2 23-39 11-22 3-6 1.9-4.7 1.4-2.2 0.6-0.9 11-19 6.9-7.3 2.1-2.6 7-10 7-8 0.9-1.3 4-6 2 0.4-0.7 V, VI, VII, VIII, IX S. x dasyclados 2 24-34 - - 2.3-3.0 - - 14-24 - - 6-8 - - 5-6 - - VII, IX S. 'Aquatica' 3 29-42 11-13 2-3 2.3-3.8 1.5-1.7 0.7-0.8 10-18 5.9-7.9 1.7-3.0 8-12 8-9 1.1-1.3 4-7 2 0.5-0.6 V, VII, VIII, IX S. x dasyclados 3 27-35 9 3 2.7-3.3 1.4 0.8 14-18 5.4 1.8 7-8 7 1.2 5-6 1.7 1.2 VII, IX A. inc ana 4 36 17 5 2.7 1.6 0.7 16 4.7 1.8 11 7 1.0 2 0.9 0.3 Saarsalmi et ai. 1985 6 36-40 16 4 2.1-2.2 1.4 0.5 15 4.6 1.2 11-14 6 1.2 3 0.9 0.3 Saarsalmi et ai. 1985, 1992 10 36 - - 2.0 - - 15 - - 12 - - 2 - - Saarsalmi et ai. 1992 " 25 28 11 2 1.1 0.8 0.2 11 3.4 0.9 9 8 0.8 2 0.6 0.2 Saarsalmi et ai. 1991 B. pendula 6 29 . . 4.0 - . 10 . . 8 . . 3 . Saarsalmi et ai. 1992 " 10 27 - - 4.3 - - 11 - - 10 - 3 - - Saarsalmi et ai. 1992 B. pubescens 40 24 5 0.8 1.9 0.4 0.1 10 3.9 0.4 11 6 0.6 - - - Mälkönen 1977 40-60 24 5 1 1.5 0.3 0.06 7 1.1 0.3 10 5 0.5 3 0.4 0.1 Fin£r 1989 P. sylvestris 28-47 12-13 3-4 0.6-0.7 1.5-1.6 0.5-0.6 0.04-0.06 6-8 1.9-2.3 0.3-0.4 2 3-6 0.5-0.6 . . Mälkönen 1974 - 14-17 4-6 0.4 1.6-2.3 0.5-0.7 0.07 4-5 1.4-2.6 0.3-0.4 - - - - - - Paavilainen 1980 " 40-60 14-16 4-5 1 1.4-1.9 0.3-0.6 0.03-0.05 5 1.3-1.5 0.2-0.3 2 3 0.5 1 0.4-0.6 0.2 Finer 1989 85 11 4 0.6 1.4 0.4 0.03 5 1.4 0.3 2 3 0.5 1 0.5 0.1 Finer 1992 P. abies 100 14.3 5 0.6 1.6 0.5 0.03 5.7 1.7 0.3 4 10 0.7 1.0 0.7 0.1 Finer 1989 30 Even though the foliar nutrient concentrations in the fertilized willow stands were generally high as compared to Scots pine or silver birch or downy birch, the greatest differences between these tree species were noted in the nutrient concentrations in the wood and bark (V-IX, Table 4). The nitrogen, phosphorus and potassium concentrations in willow bark were two to three times higher than in birch or pine (Table 4). The nitrogen concentrations in willow wood at the age of three years were 2- 5 times higher, those of phosphorus 7-20 times higher, and those of potassium 5 - 8 higher than in pine, spruce or birch. Higher nitrogen concentrations in all the compartments in grey alder and similar or lower foliar phosphorus and potassium concentrations than those of willows in these studies have been reported (Table 4). The magnesium concentrations in willow bark were 2-3 times higher than in grey alder and the calcium concentrations in spruce bark were even higher than in willow. 4.3.4 Nutrients bound into the biomass The amount of nutrients bound in a short-rotation plantation is affected by the allocation of biomass among the tree compartments and the nutrient concentrations of the compartments. By far the highest proportion of nutrients was bound in the foliage (VI, VII). The foliage accounted for 21-23% of the above-ground biomass in a two- and three-year-old willow stand, but 40-64% of the nutrients (VI, VII). The proportion of bark out of the total biomass was 23% and the proportions of most nutrients in the bark varied within the range 20-23% (VII) or 26-30% (VI). Calcium was an exception; the bark contained 40% of the total calcium (VII). The proportion of wood in the biomass was 54-56% (VI, VII), but the percentages of most nutrients in the wood varied between 15-22%. The percentage of phosphorus (30-31%) was, however, clearly higher. The high proportion of phosphorus in the willow wood was also clear in the study by Ferm (1985 a). As mentioned earlier, also age affects considerably the nutrient concentrations in the biomass compartments. With increase in age and size, the percentage of compartments containing most nutrients (foliage, bark) decreases and the percentage of wood increases (VI-IX). Compared with the amounts given in fertilization, the willow stands in this study (VI, VII) bound considerably smaller amounts of nitrogen, and especially of phosphorus and potassium, into their biomass. Only in the fastest growing S. x dcisyclados stands was the amount of nitrogen in the biomass of the same order as given in fertilization if the nitrogen bound into the leaves during three years is also considered (VII). At the age of three years, the difference in the amounts of nitrogen bound into the above-ground leafless biomass between the control and NPK-fertilized plots was 12 - 16% of the nitrogen given in fertilization (VII). The recovery was similar to that reported in short-rotation (Miegroet et al. 1994) and conventional tree plantations studies (Paavilainen 1979, Ballard 1984). The amount of phosphorus, potassium and calcium bound into one metric tonne of willow biomass (VI, VII) was at the same level as in earlier investigations (Table 5) except that by Saarsalmi (1984) which reports higher amounts of bound potassium. With an 31 increase in willow age from one to two or three years the amount of nitrogen bound into one metric tonne of biomass decreased by 13 - 30%, and as much as by 60%, in a nitrogen rich (VII) area and by 42% in a S. viminalis stand in Sweden (Nilsson & Ericsson 1986). This has been considered to be an expression of the increased nutrient use efficiency of older shoots (Nilsson & Ericsson 1986). Thus, short rotations entail the removal of higher quantities of nitrogen and also potassium per harvested unit biomass than longer rotations (Table 5). Contrary to the behaviour of nitrogen and potassium, the amount of phosphorus bound into one metric tonne of willow biomass did not decrease, or decreased only slightly, with increasing age or yield because of the fairly high phosphorus concentration in willow wood (VI, VII). Unit biomass of young grey alder and birch contains equal amounts (but unit biomass of older alder or birch considerably less) of nitrogen compared to the three year-old willows in this study (Table 5). The unit biomass of Scots pine contains considerable smaller amounts of N, P, K, Ca and Mg even when the stands are equal in terms of their amounts of biomass (Table 5). Birch, as regards its nutrient requirements, is a demanding species compared to Scots pine (Mälkönen 1977) and willows are even more demanding (Table 5). Since the fertilization regime (VI, VII), tree age and size (VII) affect the amount of nutrients bound in one metric tonne of biomass, one should be careful when making conclusions on the nutrient requirements of different tree species based on these figures (VII, Table 5). Willow stands bind high amounts of nutrients into their biomass (VI, VII, Saarsalmi 1984, Ferm 1985 a, Nilsson 1985, Nilsson & Ericsson 1986, Hytönen et ai. 1995). Two to three-year-old willow stands may contain nitrogen, phosphorus and potassium (VI: N 228, P 21, KB4 kg ha, VII: N 196, P 26, K 101, Ca 47, Mg 37 kg ha" 1 ) in amounts equal to or even exceeding those of an advanced 40-year-old birch stand, a pole-sized Scots pine stand, an 85-year-old Scots pine stand (above-ground biomass 62 t ha" 1 ) or a 100-year-old Norway spruce stand (Mälkönen 1977, Paavilainen 1980, Finer 1989, 1991). Stands of grey alder (with above-ground biomasses of 24 -32 t ha" 1 ) have been found to bind considerably more nitrogen into their biomass, but equal amounts or less of phosphorus, potassium, calcium and magnesium, than the willow stand of 18 t ha" 1 examined in this study (VII, Saarsalmi et al. 1985, 1991, 1992). The ability of willows to utilize high amounts of nutrients can be exploited in the treatment of waste waters in sanitary landfills (Ferm 1985 a, Ettala 1987, 1988) or in the removal of nutrients from other wastewaters and sludges. Such vegetation filters could act as both economically and environmentally sound biological purification systems (Aronsson & Perttu 1994). There were marked site-to-site differences, especially in the foliar zinc, but also in copper, manganese and iron concentrations (V, VII, VIII). Some willow clones can accumulate high amounts of heavy metals (especially cadmium) and thus they could be used to remove heavy metals from polluted soil (Landberg & Greger 1994). 32 Table 5. Dry-mass (t ha- 1) and the amounts of nutrients bound into above-ground biomass (kg t-1) of some tree species at different ages (years) according to investigations conducted in Finland. Figures calculated from data presented in the publications. All willows were fertilized with NPK, except those in Ferm (1985) and Ettala (1987) with wastewater leachate. The birch, pine and spruce stands of Finer (1989) were also fertilized with NPK. Alders in Saarsalmi et al. (1985, 1992) were fertilized with wood ash and birches in Saarsalmi et al. (1992) were fertilized with N + wood ash. Alders in Hytönen et al. (1995) were fertilized with PK and birches with NPK. In the lysimeter experiment by Saarsalmi (1984), the amounts of N and K leached from soil (limed Sphagnum peat) was only 0.5 - 0.6% of the amounts added in fertilization and the plants received more nutrients with the rain than was lost through leaching. In a short rotation sycamore ( Platanus occidentalis L.) plantation the lowest nitrate leaching losses Species and clones Age Dry-mass N P K Ca Mg Author Salix 'Aquatica' 5 0.4 18.2 23 9.1 5.7 23 Hytönen et ai. 1995 1 05 22.9 1.8 73 - - VII " 1 1.0 15.8 1.3 9.8 - - VII " 1 2 16.3 2.0 10.9 4.6 1.2 Ferm 1985a, Ettala 1987 " 2 3.7 18.3 1.8 8.4 5.6 2.6 VII " 2 6.8 13.8 1.6 8.9 5.1 2.1 VII " 1 11.0 12.4 1.9 13.6 4.8 1.5 Saarsalmi 1984 " 3 123 10.6 15 5.2 4.5 1.9 VII " 3 12.6 12.9 1.5 7.0 5.1 2.0 VII " 2 13.1 173 1.6 63 - - VI " 1 17 11.5 1.4 9.5 3.7 0.8 Ferm 1985a S. x dasyclados 3 18.4 10.5 1.4 5.4 4.0 2.0 VII " 5 19.4 9.2 1.4 5.8 4.2 1.7 Hytönen et ai. 1995 S. 'Aquatica' 3 25 9.1 1.0 6.2 5.7 0.6 Ferm 1985a 2 34 10.4 1.3 6.8 5.8 0.6 Ferm 1985a Salix phylicifolia 5 43.0 6.5 13 4.5 4.5 1.0 Hytönen et ai. 1995 Alrnis incana 6 9.0 17.5 1.6 4.8 6.6 0.9 Saarsalmi et ai. 1992 " 4 15.9 11.4 1.2 4.3 3.6 0.7 Saarsalmi et ai. 1985 " 10 25.1 9.6 0.9 3.4 3.7 0.5 Saarsalmi et ai. 1992 " 7 29.5 12.3 1.0 3.4 4.1 1.1 Hytönen et ai. 1995 " 6 31.1 8.8 0.8 2.9 3.0 0.6 Saarsalmi et ai. 1985 " 35 33.6 6.8 0.5 1.9 3.1 - Saarsalmi & Mälkönen 1989 " 35 69.2 5.8 0.4 1.5 2.7 - Saarsalmi & Mälkönen 1989 P. x rasymowskyana 3 15 8.1 1.0 7.7 6.0 0.8 Ferm 1985a Betula pendula 6 3.8 103 1.4 3.8 43 1.0 Saarsalmi et ai. 1992 " 7 4.9 9.9 13 3.1 33 0.9 Hytönen et ai. 1995 " 10 12.9 6.5 1.0 25 3.0 0.7 Saarsalmi et ai. 1992 " 8 15.2 73 0.8 2.4 23 0.7 Hytönen et ai. 1995 " 7 24.4 7.7 1.1 3.3 3.0 1.1 Hytönen et ai. 1995 Betula pubescens 20 38.3 3.9 0.5 1.3 2.5 - Mälkönen & Saarsalmi 1982 40-60 38.6 2.7 0.2 0.7 2.0 0.3 Finer 1989 " 40 90.2 2.4 0.2 1.2 2.0 - Mälkönen 1977 " 40 120.3 2.6 0.2 0.9 1.6 - Mälkönen & Saarsalmi 1982 Pinus sylvestris 28 17.9 3.0 0.4 1.5 2.2 . Mälkönen 1974 " 47 41.9 2.6 0.3 1.1 1.2 - Mälkönen 1974 " 40-50 50.0 25 03 0.7 1.1 03 Finer 1989 " 40-60 52.4 1.4 0.1 0.1 0.8 0.2 Finer 1989 " 85 53.4 2.1 0.2 0.8 1.3 03 Finer 1991 " - 75.2 2.3 0.3 0.8 1.3 0.4 Paavilainen 1980 n 45 75.9 2.0 0.2 1.0 1.3 - Mälkönen 1974 Picea abies 100 121.9 2.6 0.2 0.9 2