Marcel Van Oijen

On the relative magnitudes of photosynthesis, respiration, growth and carbon storage in vegetation

BACKGROUND AND AIMS The carbon balance of vegetation is dominated by the two large fluxes of photosynthesis (P) and respiration (R). Mechanistic models have attempted to simulate the two fluxes separately, each with their own set of internal and external controls. This has led to model predictions where environmental change causes R to exceed P, with consequent dieback of vegetation. However, empirical evidence suggests that the R : P ratio is constrained to a narrow range of about 0.4-0.5. Physiological explanations for the narrow range are not conclusive. The aim of this work is to introduce a novel perspective by theoretical study of the quantitative relationship between the four carbon fluxes of P, R, growth and storage (or its inverse, remobilization). METHODS Starting from the law of conservation of mass - in this case carbon - equations are derived for the relative magnitudes of all carbon fluxes, which depend on only two parameters: the R : P ratio and the relative rate of storage of carbon in remobilizable reserves. The equations are used to explain observed flux ratios and to analyse incomplete data sets of carbon fluxes. KEY RESULTS The storage rate is shown to be a freely varying parameter, whereas R : P is narrowly constrained. This explains the constancy of the ratio reported in the literature. With the information thus gained, a data set of R and P in grassland was analysed, and flux estimates could be derived for the periods after cuts in which plant growth is dominated by remobilization before photosynthesis takes over. CONCLUSIONS It is concluded that the relative magnitudes of photosynthesis, respiration, growth and substrate storage are indeed tightly constrained, but because of mass conservation rather than for physiological reasons. This facilitates analysis of incomplete data sets. Mechanistic models, as the embodiment of physiological mechanisms, need to show consistency with the constraints.

Radiation and nitrogen use at the leaf and canopy level by wheat and oilseed rape during the critical period for grain number definition

During the critical period for grain number definition, the amount of biomass produced per unit absorbed radiation is more sensitive to nitrogen (N) supply in oilseed rape than in wheat, and reaches a higher value at high N. This response was investigated by combining experimental and modelling work. Oilseed rape and wheat were grown at three levels of N supply, combined with two levels of plant density at high N supply. Canopy photosynthesis and daytime radiation use efficiency (RUEA) were calculated with a model based on observed N-dependent leaf photosynthesis and observed canopy vertical distribution of light and leaf N. In oilseed rape, RUEA was higher than in wheat and, in contrast to wheat, the sensitivity to canopy leaf N content increased from the start to the end of the critical period. These results were partly explained by the higher leaf photosynthesis in oilseed rape vs wheat. In addition, oilseed rape leaves were increasingly shaded by the inflorescence. Thus, RUEA increased because more leaves were operating at non-saturating light levels. In both species, the vertical distribution of leaf N was close to that optimising canopy photosynthesis. The results are discussed in relation to possibilities for improvement of N productivity in these crops.

Coffee agroforestry systems in Central America: II. Development of a simple process-based model and preliminary results

Research on coffee agroforestry systems in Central America has identified various environmental factors, management strategies and plant characteristics that affect growth, yield and the impact of the systems on the environment. Much of this literature is not quantitative, and it remains difficult to optimise growing area selection, shade tree use and management. To assist in this optimisation we developed a simple dynamic model of coffee agroforestry systems. The model includes the physiology of vegetative and reproductive growth of coffee plants, and its response to different growing conditions. This is integrated into a plot-scale model of coffee and shade tree growth which includes competition for light, water and nutrients and allows for management treatments such as spacing, thinning, pruning and fertilising. Because of the limited availability of quantitative information, model parameterisation remains fraught with uncertainty, but model behaviour seems consistent with observations. We show examples of how the model can be used to examine trade-offs between increasing coffee and tree productivity, and between maximising productivity and limiting the impact of the system on the environment.

Modelling the dynamics of snow cover, soil frost and surface ice in Norwegian grasslands

Studying the winter survival of forage grasses under a changing climate requires models that can simulate the dynamics of soil conditions at low temperatures. We developed a simple model that simulates depth of snow cover, the lower frost boundary of the soil and the freezing of surface puddles. We calibrated the model against independent data from four locations in Norway, capturing climatic variation from south to north (Arctic) and from coastal to inland areas. We parameterized the model by means of Bayesian calibration, and identified the least important model parameters using the sensitivity analysis method of Morris. Verification of the model suggests that the results are reasonable. Because of the simple model structure, some overestimation occurs in snow and frost depth. Both the calibration and the sensitivity analysis suggested that the snow cover module could be simplified with respect to snowmelt and liquid water content. The soil frost module should be kept unchanged, whereas the surface ice module should be changed when more detailed topographical data become available, such as better estimates of the fraction of the land area where puddles may form.

The effect of nitrogen enrichment on the carbon sink in coniferous forests: Uncertainty and sensitivity analyses of three ecosystem models

Estimates of the global carbon sink induced by nitrogen enrichment range vary widely, from nearly zero to 2.3 Gt C year−1. It is necessary to reduce this uncertainty if we are to make accurate predictions of the future magnitude of the terrestrial carbon sink. Here, we present a Monte Carlo approach to uncertainty and sensitivity analysis of three ecosystem models, Century, BGC and Hybrid. These models were applied to a coniferous forest ecosystem in Sweden. The best estimate of the change in total carbon content of the ecosystem with the cumulative change in nitrogen deposition over 100 years, ΔCtotal/ΔNdeposition was 20.1 kg C (kg N)−1 using the pooled mean, with a pooled standard deviation of 13.8 kg C (kg N)−1. Variability in parameters accounted for 92% of the total uncertainty in ΔCtotal/ΔNdeposition, and only 8% was attributable to differences between models. The most sensitive parameters were those which controlled the allocation of assimilate between leaves, roots and stem. In particular, an increase in allocation to fine roots led to a large reduction in ΔCtotal/ΔNdeposition in all models, because the fine roots have a very high turnover rate, and extra carbon allocated there is soon lost through mortality and decomposition.

On the use of specific publication criteria for papers on process-based modelling in plant science

Abstract The use of process-based modelling is increasing in most disciplines of plant science, but criteria for what is acceptable modelling work have rarely been made explicit. Unlike empirical work, modelling does not provide direct information on nature, and the goal is generally to explain natural phenomena, or predict their consequences, rather than reveal new phenomena or test specific hypotheses. Goals, methods and results of modelling and empirical work thus differ and they should be judged by different criteria. This note shows how the four criteria of explanation, testability, novelty and good modelling practice can be applied to modelling papers in the plant sciences. The practical implementation of these criteria is discussed, and a checklist of eight questions is derived with which papers can be evaluated. To illustrate the use of the criteria, they are applied to a classic publication from the crop modelling literature [Simulation of Assimilation, Respiration and Transpiration of Crops. Pudoc, Wageningen].

A novel probabilistic risk analysis to determine the vulnerability of ecosystems to extreme climatic events

We present a simple method of probabilistic risk analysis for ecosystems. The only requirements are time series—modelled or measured—of environment and ecosystem variables. Risk is defined as the product of hazard probability and ecosystem vulnerability. Vulnerability is the expected difference in ecosystem performance between years with and without hazardous conditions. We show an application to drought risk for net primary productivity of coniferous forests across Europe, for both recent and future climatic conditions.

Integrating parameter uncertainty of a process-based model in assessments of climate change effects on forest productivity

The parameter uncertainty of process-based models has received little attention in climate change impact studies. This paper aims to integrate parameter uncertainty into simulations of climate change impacts on forest net primary productivity (NPP). We used either prior (uncalibrated) or posterior (calibrated using Bayesian calibration) parameter variations to express parameter uncertainty, and we assessed the effect of parameter uncertainty on projections of the process-based model 4C in Scots pine (Pinus sylvestris) stands under climate change. We compared the uncertainty induced by differences between climate models with the uncertainty induced by parameter variability and climate models together. The results show that the uncertainty of simulated changes in NPP induced by climate model and parameter uncertainty is substantially higher than the uncertainty of NPP changes induced by climate model uncertainty alone. That said, the direction of NPP change is mostly consistent between the simulations using the standard parameter setting of 4C and the majority of the simulations including parameter uncertainty. Climate change impact studies that do not consider parameter uncertainty may therefore be appropriate for projecting the direction of change, but not for quantifying the exact degree of change, especially if parameter combinations are selected that are particularly climate sensitive. We conclude that if a key objective in climate change impact research is to quantify uncertainty, parameter uncertainty as a major factor driving the degree of uncertainty of projections should be included.

Bayesian Methods for Quantifying and Reducing Uncertainty and Error in Forest Models

Forest models are tools for analysis and prediction of productivity and other services. Model outputs can only be useful if possible errors in inputs and model structure are recognized. However, errors cannot be quantified directly, making uncertainty inevitable. In this paper, we aim to clarify terminological confusion around the concepts of error and uncertainty and review current methods for addressing uncertainty in forest modelling. Modellers increasingly recognize that all uncertainties—in data, model inputs and model structure—can be represented using probability distributions. This has stimulated the use of Bayesian methods for quantifying and reducing uncertainty and error in models of forests and other vegetation. The Achilles’ heel of Bayesian methods has always been their computational demand, but solutions are being found. We conclude that future work will likely include (1) more use of Bayesian methods, (2) more use of hierarchical modelling, (3) replacement of model spin-up by Bayesian calibration, (4) more use of ensemble modelling and Bayesian model averaging, (5) new ways to account for model structural error in calibration, (6) better software for Bayesian calibration of complex models, (7) faster Markov chain Monte Carlo algorithms, (8) more use of model emulators, (9) novel uncertainty visualization techniques, (10) more use of graphical modelling and (11) more use of risk analysis.