Student Number : 9402386H - PhD thesis - School of Animal, Plant and Environmental Sciences - Faculty of Science / The upper limit of forest production is constrained by the efficiency of radiation interception and the conversion of carbon dioxide into harvestable products, These processes are limited by water availability and nutrient supply. This study forms part of an international network of experiments aimed at demonstrating the potential yield of a forest species by eliminating water and nutrients as growth-limiting factors. The main aim of this study was to determine the climatic and nutritional controls on the growth of Eucalyptus under a given set of environmental conditions. A field experiement was initiated in the KwaZulu-Natal Midlands, South Africa, to establish the effects of varying levels of water and nutrient availability on the physiological processes controlling tree growth in E. grandis. The experimental design consisted of a control (C), irrigation (I), fertilisation (F), and irrigation and fertilisation (IF) treatments. These treatments aimed to eliminate both water (I and IF) and nutrients (F and IF) as growth-limiting factors. The study focused on three main areas, namely tree-water relations, nutrient dynamics and acquisition, and carbon allocation. Information on these processes was acquired through a combination of field data collection, the application of foliar diagnostic tools, and a modelling approach. The growth of the E.grandis trees in this trial was constrained by poor establishment and management practices. Inadequate vegetation control and vigorous weed growth were a source of competition during the establishment of the seedlings. In addition, the variable seed stock, late fertiliser application and damage from animals resulted in late canopy closure. Consequently, the potential growth of E.grandis at this site could not be determined, and comparisons were made between the different treatments rather than demonstrating maximum growth under optimal conditions. The boimass and nutrient pools were quantified to provide a mid-rotation value of the nitrogen (N) and phosphorus (P) contents contained in the soil, forest floor and above- and below-ground biomass. At age four years, irrigation had a more pronounced effect on growth and biomass accumulation than fertilisation. Increases in diameter, height,, basal area, standing volume and above- and below-ground biomass were observed in response to increased water availability. After four years, the respective totals of the above- and below-ground biomass pools accumulated by the C, I, F and IF treatments were 73, 82, 78 and 85 t ha-1. In irrigated (I and IF) treatments, significantly more above-ground biomass was allocated to the stem, compared to non-irrigated (C and F) trees. The lower root: shoot ratio in fertilised (F and IF) trees indicated that proportionally less mass was allocated to below-ground components compared to non-fertilised (C and I) trees. Irrigated trees accumulated more N and P than non-irrigated trees in above- and below-ground biomass fractions. The 3-PG model has shown considerable potential as a tool for predicting forest productivity. As part of a model validation exercise, 3-PG was applied to this stand. Simulations were run over a two-year period using appropriate initialisation data and a parameter set developed for E.grandis in summer rainfall region of South Africa. The model was validated using observed leaf area index (L) and foliage and above-ground woody biomass, and then used to predict stand volume. The 3-PG model produced accurate predictions of L, foliage and above-ground woody biomass. At four years, the model-predicted L over all treatments ranged from 3.8 to 5.0, in contrast to the observed range of 4.3 to 5.2. Stand volume at four years of age was under-estimated in the non-irrigated treatments. Actual stand volume at this age ranged from 100 to 118 m3 ha-1 across all treatments, compared to model estimates of 82 to 123 m3 ha-1. The 3-PG model can be used as a guide for future research direction, which should focus on a better understanding of nutritional dynamics operatins at a site. Use of a dynamic fertilty rating may improve model predictions and there is a possible scope for improvement in the water-balance component of the model. Results from this study have shown 3-PG to be a robust model by producing accurate predictions of growth under varying levels of resource availability. There is much current interest in predicting the maximum amount of water that can be transpired by Eucalyptus trees. It is possible that industrial waste water may be applied as irrigation water to eucalyptus and it is important ot predict the maximum transpiration rates of these plantations in an attempt to dispose of this contaminate water. A simple model was developed to predict maximum rates of daily transpiration by short-rotation E.grandis plantations experiencing no significant soil water deficits or fertility limitation. Daily sap flow data recorded in a single average tree in the irrigated treatment were used to estimate mean daily canopy conductance. Analysis of daily and seasonal variation in conductance confirmed that solar radiation and vapour pressure deficit are the dominant factors reducing canopy conductance below potential values, when soil water availability is high. A simple canopy conductance model based on these data was then used with the Penman-Monteith equation to predict daily transpiration rates by E.grandis trees at a site situated in Mpumalanga. The modelled transpiration rates agreed well with th oberved daily sap low (R2 = 0.79). The total observed annual sap flow at the Mpumalanga site was equivalent to 1320 mm compared to the modelled value of 1226 mm. Fertilisation is one of the most cost-effective methods of increasing and maintaining the productivity of E.grandis plantations in South Africa. This silvicultural pratice can be optimised by using the foliar nutrient ratios measured in plants at maximum growth as a guideline for fertiliser application. The foiar ratios present at optimum growth are similar across a wide range of plant apecies and can be defined in relation to N. A study was conducted to determine whether the proportions of elements relative to N could be refined specifically for E.grandis through the use of diagnosis and recommendation integrated system (DRIS) norms. The DRIS norms for N, P, potassium (K), calcium (Ca) and magnesium (Mg) wer derived from published studies on E.grandis and compared to targer values recommended by Linder (1995) for optimal growth. The DRIS norms for P and K relative to N corresponded to the target values, but the Ca:N and Mg:N ratios were much higher, indicating luxury consumption of theses elements. A similar trend was observed for foliar nutrient ratios calculated from published data on a wide range of plant species. The DRIS appears useful for diagnosing the nutrient status, provided accurate DRIS norms are used and indices are interpreted correctly. In most cases, the diagnosis made using the DRIS indices was confirmed with the optimal ratio approach. Diagnoses made using these techniques indicated that foliar P and K concentrations were sub-optimal at age four years. The results from this trial suggest that there was an apparent temporal separation of resource limitation experienced at this site. Initially, the trees appeared to be constrained by nutrient availability, whereas water was potentially more limiting to growth than nutrient supply at age four years. However, it is highly unlikely that irrigation of forest species would ever be implemented on a operational scale in this country. Rather than pursuing the potential reponses that could be obtained with unlimited water availability, the forest industry in South Africa would benefit greatly from improved fertiliser recommendations. On the basis of the optimal foliar technique and the DRIS approach, foliar ratios of 100 N: 8 P: 35 K: 2.5 Ca: 4 Mg are suggested for optimal growth of E.grandis trees.