Simulation of forest growth, applied to douglas fir stands in the Netherlands

anglais. Forest growth in relation to weather and soils is studied using a physiological simulation model. Growth potential depends on physiological characteristics of the plant species in combination with ambient weather conditions (mainly temperature and incoming radiation). For a given site, growth may be lower because of incomplete canopy closure, shortage of water and nutrients, and the occurrence of growth-disturbing factors such as pests, diseases, and damage to the plants, e.g. by windthrow or frost. Attention is focused on the main growth-limiting factors, i.e. canopy closure, and the availability of water, nitrogen and phosphorus, so that differences in growth between different sites can be explained as a function of the properties of plant and soil, and of the ambient weather at a particular site. The model is applied to even-aged Douglas fir stands in the Netherlands because of the availability of field data for testing and evaluating it.The life cycle of trees and forests encompasses many years, and in order to be able to study overall stand dynamics, the model aims at simulating growth over periods of several decades. This allows the results of the model to be evaluated against data from permanent field plots, that are also being used in conventional, descriptive research on growth and yield. Furthermore, simulating forest growth over such long periods makes the results from the model comparable with the results of practical forest management. Variations in growth during the year are caused by changes in incoming radiation, temperature and water availability in the soil. To simulate this, time intervals of one day are used for the main part of the simulation model.The particular value of simulation models in forestry lies in the possibility they offer of combining different aspects of growth in an overall approach, and of studying stand dynamics over a long period of time without having to rely entirely on expensive and time-consuming field trials. Moreover, in a situation where the environment for forest growth may change e.g. as the result of industrial pollution, or as a consequence of gradual climatic changes, modelling is one of the important means by which to assess these changes and potential damage.The subject of the study, an even-aged coniferous forest stand, is described in terms of the biomass components foliage, branches, stems and roots. These four components are the main state variables in the model. To enable comparisons to be made between the results from the model and the data from permanent field plots, only stem biomass and stem volume are considered, together with the number of trees. This reflects a top-down approach to growth, which is calculated as total stand growth per unit of soil surface area, before it is distributed over individual trees. In addition to state variables that denote biomass amounts, stand structure is also characterized by stand height, average dimensions of the tree crowns, and total depth of the rooted soil profile. All other state and intermediate variables of the trees (such as the Leaf Area Index of the stand), are derived from simulated biomass components and stand structure. In the model, ambient weather is characterized using meteorological data from a local weather station: total daily global radiation, daily minimum and maximum temperatures, daily vapour pressure of the air, average wind speed at 10 m above short vegetation, and precipitation. The latter is characterized by daily rainfall and the average number of rainfall events per day. Only the rooted soil profile is used to describe the soil compartment. Soil moisture retention properties are the main variables for the hydrological submodel. The simulation of nutrient dynamics is based on the total amount of nutrients retained in the rooted soil profile and incorporated in the stand biomass. Nutrient inputs to the system are described by forcing functions, and used as input to the model.Chapter 3 shows how primary production is calculated for the whole stand. Canopy assimilation is calculated from the distribution of photosynthetically active radiation over the foliage, together with the photosynthesis/light response curve at ambient temperature for the surface of an individual leaf. The assimilation submodel uses a three- point Gaussian integration, as described recently by Goudriaan (1986), and Spitters (1986). The distribution of photosynthetically active radiation over the foliage accounts for gaps in the canopy, and allows for clustering of the foliage, as in the case of grouping of needles around branches in old stands. Typical aspects of canopy assimilation in Douglas fir stands, are the evergreen habitus of the stand, and the generally low maximum photosynthesis rates, (around 15 kg CH 2 O ha -1h -1). These low rates of photosynthesis are coupled with high stomatal resistances for the diffusion of both carbon dioxide and water vapour.After canopy assimilation has been estimated, net growth is calculated by accounting for maintenance respiration, and by allocating the assimilates available for growth to the biomass components. Growth respiration is taken into account when converting assimilate products to structural dry matter. To calculate maintenance respiration, sapwood is distinguished from heartwood. It is found that the hypothesis (Boysen Jensen, 1928; Kira and Shidei, 1967) that tree growth declines with age of the trees because maintenance requirements increase with accumulation of stem biomass does not hold when maintenance requirements are related to sapwood only. Sapwood (like foliage, branches and roots) has a limited life-span, and the maximum value it attains during stand development depends on site productivity. This maximum value is reached within 15 years of the time of maximum annual increment. Growth respiration is calculated by taking the chemical composition of the biomass formed into account. The allocation of assimilates to the biomass components is based on a distribution key derived from published data. The distribution of growth over the biomass components changes during stand development, and also depends on the productivity of the site. Stem dry weight increment is converted to volume increment by dividing the estimated dry weight increment by the basic density of the stem wood formed, i.e. the oven-dry weight per unit of fresh volume. Individual tree increment is calculated by dividing total stem volume increment by the number of trees in the stand, and only an average value for diameter at breast height is calculated from tree volume and height, using an empirical regression equation.Chapter 4 describes the hydrological part of the model. The three main aspects considered in the model are: a) interception of precipitation by the canopy and the resulting net infiltration to the soil compartment; b) the soil moisture balance; c) and uptake and transpiration of soil moisture by the trees. Coniferous forests in western Europe are often located on sandy soils with a limited soil moisture holding capacity and restricted capillary rise. This means that in periods of drought, availability of soil moisture becomes limiting for growth. In the model, therefore, it suffices to simulate water availability with an elementary summary model that keeps track of soil moisture. Soil moisture content and the rate of infiltration are simulated by assuming that the soil horizons are filled to field capacity by a sharp wetting front proceeding from the top of the soil profile downwards. Root uptake is assumed to proceed until soil moisture is depleted to the wilting point. Field capacity and wilting point are derived from soil suction curves, and depend on physical soil characteristics.Tall forest stands have considerable aerodynamic roughness, and this means that the aerodynamic resistance to the transport of water vapour from the surface of the foliage to the overlying atmosphere is small (around 10 s m -1). Besides, the large stomatal resistance of Douglas fir needles results in a minimum canopy resistance for the transpiration flux of 100 to 200 s m -1; therefore, precipitation intercepted by the vegetation will evaporate at rates several times the transpiration rate under the same atmospheric conditions. Therefore, interception represents a real loss that has to be accounted for. To estimate interception, the amount of intercepted precipitation is subtracted from daily precipitation.Daily transpiration is estimated with the Penman-Monteith combination equation, with total canopy resistance as one of the input variables. This resistance depends on: a) the vapour pressure deficit of the air (here assumed to pose a lower limit on stomatal resistance), b) the water status of the foliage, expressed in terms of needle water potential, and c) the stomatal opening resulting from photosynthesis. All three effects on stomatal resistance are calculated independently, and the largest resistance is used in the model to estimate total canopy resistance. The influence of vapour pressure and plant water status (through needle water potential), Is assumed to be the same for all foliage in the canopy. The stomatal resistance estimated from net photosynthesis rates varies with varying photosynthesis rates inside the canopy. As in the calculation of canopy assimilation, a Gaussian integration procedure is used to estimate the weighted average foliage resistance. The resulting transpiration rates are found to be unexpectedly low during the growing season. Total annual transpiration, however, is in accordance with published data, and the simulated change in soil moisture during 1983 compares well with measurements from the field plots. It is concluded that on dry soils like those frequently occupied by coniferous stands in the Netherlands, water shortage may have considerable influence on growth, even though transpiration rates are low. In its present state the model can be used to calculate the reduction In growth caused by water shortage, for different sites, and for stands of different structure.In chapter 5 the simulation of nutrient dynamics and the influence of nitrogen and phosphorus on growth are described. As it has been shown many times that nitrogen and phosphorus may limit growth of coniferous stands on sandy soils, only these two elements are incorporated in the model. No attempt is made to model the dynamics of nitrogen and phosphorus in detail; instead, an elementary model with time steps of one year is used in combination with the simulations of daily canopy assimilation and hydrology. Soil supply of nitrogen and phosphorus is estimated from total soil content, by taking into account an unstable and a stable pool of nutrients in the soil, each with different turnover rates. The demand for nitrogen and phosphorus by the growing vegetation depends on the concentrations of these elements in the tissue, and on the amounts redistributed before dead biomass is shed, in combination with an estimated rate of biomass increment. By adjusting the concentration in the tissue for the next period of growth, demand and supply are balanced, and the influence of nutrient availability on growth during the following year is estimated using an empirical relationship between foliage nutrient concentration and growth. This approach assumes the existence of maximum and minimum concentrations of both nutrients in the tissue. Above the maximum concentration there is no further uptake; below the minimum, growth ceases.The final results from the model, together with the measurement series from permanent field plots are given in chapter 6. The field plots used to calibrate the model are discussed first; after this the model is tested against an independent set of data. Overall model behaviour seems to follow field measurements reasonably, both in the field plots used for calibration and in the independent (control) plot. Maximum increment rates as measured in the field are well reflected in the simulations, as is the decline in stem increment in older trees. Most of the discrepancies between predicted and real values are found to occur at higher ages of the stand. It is concluded that this is probably because the model overestimates light interception, because it takes no account of effects of uneven distributions of the trees in the field. This becomes more important when stands are thinned at high ages, when the crowns have only a limited ability to occupy the available growing space.Together with the evaluation of model behaviour, the value of the use of modelling in forestry in general, and of the use of a physiologically-based model like the one used here, is discussed. These models are needed for analysing growth and yield, and for contributing to the understanding of forest primary production. Moreover, they can be used to bridge the gap between widely different aspects of forest growth such as forest hydrology and forest nutrition. By integrating the main aspects of forest growth, these models also allow the main factors that determine total stand growth to be ascertained. As a result, possibilities for yield improvement, and the areas where research is mostly needed, can be identified. In the present case study, it appears that canopy growth often declines in the course of years because of decreased light interception. Current forestry practice in the Netherlands often includes an intensive thinning programme aimed at creating space for the individual crop trees. But this decreases stand growth. In general, this is not the intention, and therefore the efficacy thinning operations at higher stand ages that open up the stand to a degree that can no longer be utilized by the remaining trees, has to be reassessed.Not only does availability of soil moisture limit growth; nitrogen and phosphorus availability may also play an important role in determining the production level of a stand. The elementary model used indicates the extent to which both nitrogen and phosphorus may influence stand growth, and the results are evaluated against the results of fertilizer experiments carried out in Douglas fir on a range of sites during the 1950s and the 1960s (Blok et al., 1975). The increase In atmospheric input of nitrogen, resulting from, among others, intensive livestock farming and manure-spreading on agricultural lands, has greatly increased nitrogen supply. As a result, widespread phosphorus deficiency has become apparent. In the Netherlands, all but the best sites currently available suffer from severe phosphorus deficiency. This situation, where widespread nitrogen deficiency has changed into a deficiency of phosphorus, demands attention from researchers and forest managers. Increasing phosphorus availability through additional. fertilization can be expected to boost primary production and thus increase yield.One of the possible applications of the model is to calculate the growth potential of a wide range of available soil types and growing conditions, thereby allowing potential forest growth to be assessed. It can also be used to evaluate management interventions. If employed in a target-oriented mode the model could be used to evaluate the efficacy of applying fertilizer. Some of the growth- or stand-disturbing factors will have to be incorporated in the model before it can be used to calculate economic yield or optimal felling regimes.The simulation programme is available upon request.

néerlandais; flamand. M.b.v. een fysiologisch simulatiemodel van de groei van een bos is de relatie tussen klimaat, bodem en produktie bestudeerd. Het onderzoek is vooral gericht op de belangrijkste groeibeperkende factoren, zoals de mate van kroonsluiting en de beschikbaarheid van water, stikstof en fosfor, waarbij de laatste beide voedingselementen de enige zijn, waarvan de invloed op de primaire produktie in het model is opgenomen. De beschikbaarheid van water, stikstof en fosfor wordt gebruikt om verschillende groeisnelheden op verschillende groeiplaatsen te verklaren m.b.v. bodem- en planteigenschappen. Het model is ontwikkeld en toegepast in gelijkjarige Douglasbossen