Simulation of forest carbon dynamics based on a dry-matter production model

A microcomputer model for forest carbon dynamics with five functional comparments (atmosphere, foliage, woody-parts, roots and dead biomass in the soil) is constructed which incorporates dry-matter production processes of trees such as photosynthesis, respiration and allocation of photosynthate. The effect of photosynthesis rate at saturated light and dark respiration rate of a single leaf upon surplus production (P _s) is three-dimensionally illustrated as a function of cumulative leaf area index (LAI) and extinction coefficient of light. Probable values of the physiological parameters in this model are determined by repeated simulation experiments. The successional pattern during a period of 100 years is simulated, demonstrating stable and perpetual occurrence of a tropical rainforest ecosystem composed of three strata. The model is also analyzed in terms of response of relative initial density of trees, thereby displaying the law of constant final yield in a forest ecosystem. The model outputs of carbon fluxes and phytomasses at the steady state agree quite well with field data already obtained from a tropical rainforest at Pasoh.     

Simulation of forest carbon dynamics based on a dry-matter production model

Journal of Plant Research - Effects of increasing atmospheric CO2 concentration upon a tropical rainforest ecosystem are analysed by employing the microcomputer model developed in a previous paper...     

Studies on the dynamic properties of terrestrial ecosystems based on a simulation model II. Tropical rainforest dynamics and stability as influenced by stem mortality

Further analysis of tropical rainforest dynamics and stability in relation to stem mortality has been conducted using a microcomputer model developed in a previous study (Oikawa, 1985). By simulation experiments covering a period of 100 years, the effects of changing stem mortality (δc) upon a tropical rainforest were investigated. Increasing stem mortality ranging from a standard value (3%yr⁻¹) to a 4-fold value (12%yr⁻¹) brings about decreases in stem biomass and thus total living biomass, and a contrasting increase of stem litterfall flux at the steady state of the forest ecosystem. At the same time, the decreased stem biomass at the steady state is predicted to result in increases of gross production (P _g) and net production (P _n), and an improvement in production efficiency of the model rainforest expressed as theP _n/P_g ratio. similar simulation experiments predict that the improved production efficiency in the forest with a 4-fold stem mortality is able to enhance tolerance to less productive environments such as a prolonged dry season or a reduced incident light flux density. On the other hand, the standard stem mortality (δ_c=3%yr⁻¹), which was estimated as a probable value for the Pasoh forest, West Malaysia, is considered to approximate the lower threshold necessary for attaining forest stability. Based on the results obtained, the significance of δ_c for the dynamics and stability of a tropical rainforest ecosystem is discussed in relation to the competition and tolerance of trees. In addition, the effectiveness of the simulation approach adopted here is emphasized. Titles are tentative translations by the author for original titles in Japanese.     

Studies on the dynamic properties of terrestrial ecosystems based on a simulation model I. Critical light conditions for stability of a tropical rainforest ecosystem

By employing a microcomputer model developed in a previous study (Oikawa, 1985), the stability of a tropical rainforest ecosystem composed of three strata was analyzed in relation to incident light flux density. Surplus production (P _s ), calculated as a function of the leaf area index and light attenuation coefficient, was remarkably affected by the maximum illuminance at noon (I _{0, max}). Simulation experiments for a period of 100 years demonstrated that the upper stratum was able to reach a steady state at about 50 years and thereafter, when a value ofI _{0, max} equal to or greater than 80 klux was assigned to the upper stratum, where the higher the value ofI _{0, max} , the greater the biomasses and the carbon fluxes at the steady state as a result of enhanced productivity. WhenI _{0, max} was assigned a value of 70 klux, on the other hand, this experiment predicted a failure of the upper stratum to maintain stability due to deficiency of surplus productivity. Moreover, it was also suggested that excessive luxuriance of the upper stratum due toI _{0, max} elevation may have a detrimental effect upon the survival of the middle stratum, since increasingI _{0, max} decreases the light energy available for the middle stratum even in absolute terms, resulting in disappearance of this stratum whenI _{0, max} is equal to or greater than 120 klux. These simulation experiments suggested that a tropical rainforest ecosystem composed of three strata is able to exist within a narrow range ofI _{0, max} between 80 and 110 klux, light conditions which are much higher than the light compensation point for canopy photosynthesis This article is dedicated to Professor Toshiro Saeki, University of Tokyo, in appreciation of the sincere encouragement he has given to the author.     

Effects of fragmentation on forest structure and litter dynamics in Atlantic rainforest in Pernambuco, Brazil

This study evaluates the effects of fragmentation on the spatial and temporal dynamics of small litterfall production in Atlantic rainforest in Pernambuco State, Brazil. Litterfall was collected for 24 months at two 0.2 ha sites, located in the forest edge zone and the forest interior, within a rainforest patch of about 300 ha. Structural parameters of both forest sites were recorded. Litter was sorted into six fractions (foliage, twigs, buds/flowers, fruits/seeds, peduncles, rest), dried and weighed. The interior forest plot contained 314 live trees with a dbh 5 cm and a stand basal area of 41.8±8.7 m², whereas the forest edge contained 211 live trees and a stand basal area of 23.4±3.6 m². Total small litterfall was extraordinarily high and totalled 12.62±4.73 t ha⁻¹ yr⁻¹ in interior forest and 14.74±2.78 t ha⁻¹ yr⁻¹ in forest edge. High litterfall rates are probably due to a pronounced periodicity, edge effects alter litterfall strongly.     

A dissipative particle dynamics model of carbon nanotubes

A mesoscale dissipative particle dynamics model of single wall carbon nanotubes (CNTs) is designed and demonstrated. The coarse-grained model is produced by grouping together carbon atoms and by bonding the new lumped particles through pair and triplet forces. The mechanical properties of the simulated tube are determined by the bonding forces, which are derived by virtual experiments. Through the introduction of van der Waals interactions, tube–tube interactions were studied. Owing to the reduced number of particles, this model allows the simulation of relatively large systems. The applicability of the presented scheme to model CNT based mechanical devices is discussed.     

Simulation of the effects of deer browsing on forest dynamics

I introduced forest dynamics to a deterministic herbivore-vegetation model to examine the effects of vertically stratified forest structure on the dynamics of the herbivore–vegetation system, the resilience and stable states of vegetation, and the interactions between deer populations and vegetation. I constructed a model based on data from field studies performed in Hokkaido, northern Japan. Three phases of state were identified for a given deer density: (1) understory vegetation is maintained with a equilibrium proportion of canopy gaps in the absence of deer; (2) if the equilibrium proportion of canopy gaps is greater than that in the equilibrium state in the absence of deer, the understory vegetation can be maintained; and (3) the understory vegetation cannot be maintained. At the boundary between phases 2 and 3, the herbivore population level had discontinuous effects on vegetation. When the deer density was held below the threshold, forest vegetation had resilience to recover to the equilibrium stable state at the given deer density, although the equilibrium canopy gap ratio and vegetation biomass differed with deer density. However, the forest vegetation–herbivore system could not be maintained in a stable state without artificial deer population management if food limitation was the only mechanism to keep the deer population at low levels. The deer population must be kept below the boundary between phases 1 and 2 to maintain the forest regeneration processes. The level cannot be determined by observing the deer population; careful observation of forest regeneration processes is required.     

用动态模型研究森林群落中物种间的竞争

研究了以过程为基础的模型在物种竞争试验模型及实践中的应用。应用森林动态林窗模型模拟了群落内物种竞争规律。结果表明：同一群落中具有不同生物、生态学特征的物种竞争过程与环境条件有密切关系。由此确定,竞争是与外部条件密切相关的,不是一个单独稀量物种特征的指标;具有相似生物和生态学特性的物种,在森林植被发育过程中,存在绝对竞争,具有较弱更新与生长能力的物种,始终处于竞争劣热,不宜于在这样的群落中生长,而且有较强更新和生长的树种在群落中处于优势地位。     

Forest structure and carbon dynamics in Amazonian tropical rain forests

Living trees constitute one of the major stocks of carbon in tropical forests. A better understanding of variations in the dynamics and structure of tropical forests is necessary for predicting the potential for these ecosystems to lose or store carbon, and for understanding how they recover from disturbance. Amazonian tropical forests occur over a vast area that encompasses differences in topography, climate, and geologic substrate. We observed large differences in forest structure, biomass, and tree growth rates in permanent plots situated in the eastern (near Santarém, Pará), central (near Manaus, Amazonas) and southwestern (near Rio Branco, Acre) Amazon, which differed in dry season length, as well as other factors. Forests at the two sites experiencing longer dry seasons, near Rio Branco and Santarém, had lower stem frequencies (460 and 466 ha^–1 respectively), less biodiversity (Shannon–Wiener diversity index), and smaller aboveground C stocks (140.6 and 122.1 Mg C ha^–1) than the Manaus site (626 trees ha^–1, 180.1 Mg C ha^–1), which had less seasonal variation in rainfall. The forests experiencing longer dry seasons also stored a greater proportion of the total biomass in trees with >50 cm diameter (41–45 vs 30% in Manaus). Rates of annual addition of C to living trees calculated from monthly dendrometer band measurements were 1.9 (Manaus), 2.8 (Santarém), and 2.6 (Rio Branco) Mg C ha^–1 year^–1. At all sites, trees in the 10–30 cm diameter class accounted for the highest proportion of annual growth (38, 55 and 56% in Manaus, Rio Branco and Santarém, respectively). Growth showed marked seasonality, with largest stem diameter increment in the wet season and smallest in the dry season, though this may be confounded by seasonal variation in wood water content. Year-to-year variations in C allocated to stem growth ranged from nearly zero in Rio Branco, to 0.8 Mg C ha^–1 year^–1 in Manaus (40% of annual mean) and 0.9 Mg C ha^–1 year^–1 (33% of annual mean) in Santarém, though this variability showed no significant relation with precipitation among years. Initial estimates of the C balance of live wood including recruitment and mortality as well as growth suggests that live wood biomass is at near steady-state in Manaus, but accumulating at about 1.5 Mg C ha^–1 at the other two sites. The causes of C imbalance in living wood pools in Santarém and Rio Branco sites are unknown, but may be related to previous disturbance at these sites. Based on size distribution and growth rate differences in the three sites, we predict that trees in the Manaus forest have greater mean age (~240 years) than those of the other two forests (~140 years).     

热带季节雨林林冠碳通量不同校正方法的比较分析

利用西双版纳热带季节雨林2003年3月1～9日的CO₂和水汽通量数据,比较了不同校正方法的差异.结果表明,对季节雨林林冠碳通量观测数据进行校正是必要的,但不同校正方法所起的作用有所差异;昼间WPL校正的作用最大,虚温校正的作用最小;t检验证明,经WPL修正后的碳通量值与基准值、无显著差异;聚类分析表明,进行自然风3次旋转校正有利于数值精度的提高.夜间是自然风旋转校正的贡献最大、虚温校正贡献最小;t检验表明,经过WPL修正和平面旋转校正及其组合得到的碳通量值与基准值均有显著差异;而经包含有自然风旋转校正处理后的碳通量值与基准值均无显著差异;聚类分析表明,经过自然风坐标旋转后,再进行WPL修正,有利于数值精度的提高.     

Advances in spatial,individual‐based modelling of forest dynamics

Many individual‐based models of forest dynamics lack spatial complexity. Although, in certain cases, spatially simple models may not be substantially inferior to spatially complex models, advances in vegetation science indicate potential weaknesses, particularly the lack of consideration of propagule availability in horizontal space, and varying patch (or canopy gap) dimensions. Models with vertical and horizontal spatial complexity can address these issues, but, thus far, evidence that they outperform patch (or gap) models is limited. Comparison of projections from models that differ only in their spatial complexity is needed to address the effects of propagule availability in space, spatial pattern of canopy tree mortality, and spatial resolution.     

A spatial model of forest dynamics

Effects of spatial processes on temperate deciduous forest structure and dynamics were investigated with a spatial simulator derived from a forest gap model. The multi-species neighborhood model accounted for competitive interactions and endogenous disturbance in the form of small canopy gaps. Simulated and actual spatial pattern of old-growth stands were compared. The 400 yr simulations produced a pattern scale (0.07–0.2 ha patches) similar to that of an actual stand; simulated pattern intensity was greater than actual intensity, however. Distances to nearest neighbor were somewhat similar for trees in the simulated and actual stands; yet the frequency distributions of distance to nearest neighbor values differed substantially. The simulated stand patterns were generally less random than the actual patterns. Spatial pattern changed markedly during the course of simulated succession. Pattern approached a random dispersion in early succession. Intensity peaked at mid-succession (ca. 150 yr) with a hyperdispersed overstory and a strongly clumped understory. Pattern intensity diminished in late succession as a mixed size structure developed. Old-growth patch size was greater than the neighborhood (or gap) size, suggesting the gap-sized areas do not behave independently.     

Carbon dynamics of terrestrial ecosystems on the Tibetan Plateau during the 20th century: an analysis with a process‐based biogeochemical model

Aim The Tibetan Plateau accounts for about a quarter of the total land area of China and has a variety of ecosystems ranging from alpine tundra to evergreen tropics. Its soils are dominated by permafrost and are rich in organic carbon. Its climate is unique due to the influence of the Asian monsoon and its complex topography. To date, the carbon dynamics of the Tibetan Plateau have not been well quantified under changes of climate and permafrost conditions. Here we use a process‐based biogeochemistry model, the Terrestrial Ecosystem Model (TEM), which was incorporated with a soil thermal model, to examine the permafrost dynamics and their effects on carbon dynamics on the plateau during the past century. Location The Tibetan Plateau. Methods We parameterize and verify the TEM using the existing data for soil temperature, permafrost distribution and carbon and nitrogen from the region. We then extrapolate the model and parameters to the whole plateau. Results During the 20th century, the Tibetan Plateau changed from a small carbon source or neutral in the early part of the century to a sink later, with a large inter‐annual and spatial variability due to changes of climate and permafrost conditions. Net primary production and soil respiration increased by 0.52 and 0.22 Tg C year^?1, respectively, resulting in a regional carbon sink increase of 0.3 Tg C year^?1. By the end of the century, the regional carbon sink reached 36 Tg C year^?1 and carbon storage in vegetation and soils is 32 and 16 Pg C, respectively. On the plateau, from west to east, the net primary production, soil respiration and net ecosystem production increased, due primarily to the increase of air temperature and precipitation and lowering elevation. In contrast, the decrease of carbon fluxes from south to north was primarily controlled by precipitation gradient. Dynamics of air temperature and associated soil temperature and active layer depth resulted in a higher plant carbon uptake than soil carbon release, strengthening the regional carbon sink during the century. Main conclusions We found that increasing soil temperature and deepening active layer depth enhanced soil respiration, increasing the net nitrogen mineralization rate. Together with the effects of warming air temperature and rising CO₂ concentrations on photosynthesis, the stronger plant nitrogen uptake due to the enhanced available nitrogen stimulates plant carbon uptake, thereby strengthening the regional carbon sink as the rate of increase net primary production was faster than that of soil respiration. Further, the warming and associated soil thermal dynamics shifted the regional carbon sink from the middle of July in the early 20th century to early July by the end of the century. Our study suggests that soil thermal dynamics should be considered for future quantification of carbon dynamics in this climate‐sensitive region.     

Modelling carbon dynamics in coniferous forest soils in a temperature gradient

Climate change and changes in land use will alter the stores of carbon and turnover of soil organic matter. We have used a theory for carbon cycles in terrestrial ecosystems to analyse changes in soil organic matter turnover in coniferous forests. The central concepts of the theory are a continuously changing substrate quality, a constant decomposer efficiency and a climatically controlled decomposer growth rate. Measurements on litter production and soil carbon stores from field experiments have been used to successfully validate the model predictions. Measured litter production increased with increasing temperature but the response was not identical for forests of different vegetation types which reflect variations in productivity. The temperature response of needle-litter production and decomposition rate were strongest in the most productive forests and weakest for the low productive forests. Initial decay rates of soil C store from steady state showed the same trend in temperature response as decay of a single litter cohort did, but the absolute values are 16% of the decay rates of a single litter cohort. Predicted soil C ranged from 5 to 9 kg C m^–2. There exists a remarkable variation in forest soil C store response to temperature; the magnitude and even the sign depends on productivity as defined by vegetation type. The assumption that, in general, decomposition rates increase more than NPP with temperature, and consequently, soil C stores should decrease in response to a climate warming, seems therefore too simplistic.     

Individual-based model of spatio-temporal dynamics of mixed forest stands

This research presents the results of constructing and parameterizing an individual-based model of spatiotemporal dynamics of mixed forest stands. The model facilitates computerized experiments with forest stands having different combinations of species and age structures. These forest stands grow on temperate areas where light is the main system-forming factor that shapes and develops forest ecosystems. The model TEMFORM (TEMperate FORests Model) is developed with few equations and parameters, most of which can be estimated using standard forest inventory data. Parameterization of the model used the growth tables of a set of basic forest-forming species in Far East Russia. Simulation results of the development of the natural single- and mixed-species stands and the effects of different types of disturbances on the stand dynamics and compositions are presented.     

Self-shading, carbon gain and leaf dynamics: a test of alternative optimality models

A simple model of shoot-level carbon gain is presented addressing the optimal number and life span of leaves in relation to alternative optimality criteria: (1) maximizing carbon export from the shoot, or (2) maximizing the rate of leaf production at the shoot tip. Additionally, the processes that cause declining assimilation with leaf age are considered in relation to (1) leaf position on the shoot (e.g., self-shading) versus (2) leaf age per se. Using these alternative scenarios, only a model based on position-dependent assimilation and maximization of leaf production rates resulted in quantitative predictions for all aspects of leaf dynamics on the shoot (i.e., leaf number, life span, and birth rate), while other approaches predicted that one or more parameters would be infinite. This formulation of the model also predicted that leaves should be maintained on the shoot until the diurnal carbon balance declines to zero, in contrast with other scenarios which predict that leaves should be shed while maintaining a positive carbon balance. Predictions of the model were supported by the results of a field study of carbon gain and leaf dynamics in saplings of three species of tropical pioneer trees (Carica papaya, Cecropia obtusifolia, and Hampea nutricia) which differ in the number of leaves per shoot. The results illustrate that in these fast-growing plants, leaf production and height growth may be more appropriate measures of performance than net carbon export from the shoot, and suggest that leaf senescence is primarily a function of the position of a leaf within the canopy, rather than its chronological age. Received: 13 October 1998 / Accepted: 27 January 1999     

Implications of floristic and environmental variation for carbon cycle dynamics in boreal forest ecosystems of central Canada

Abstract. Species composition, detritus, and soil data from 97 boreal forest stands along a transect in central Canada were analysed using Correspondence Analysis to determine the dominant environmental/site variables that differentiate these forest stands. Picea mariana stands were densely clustered together on the understorey DCA plot, suggesting a consistent understorey species composition (feather mosses and Ericaceae), whereas Populus tremuloides stands had the most diverse understorey species composition (ca. 30 species, mostly shrubs and herbs). Pinus banksiana stands had several characteristic species of reindeer lichens (Cladina spp.), but saplings and Pinus seedlings were rare. Although climatic variables showed large variation along the transect, the CCA results indicated that site conditions are more important in determining species composition and differentiating the stand types. Forest floor characteristics (litter and humus layer, woody debris, and drainage) appear to be among the most important site variables. Stands of Picea had significantly higher average carbon (C) densities in the combined litter and humus layer (43530 kg‐C.ha^‐1) than either Populus (25 500 kg‐C.ha^‐1) or Pinus (19 400 kg‐C.ha^‐1). The thick surface organic layer in lowland Picea stands plays an important role in regulating soil temperature and moisture, and organic‐matter decomposition, which in turn affect the ecosystem C‐dynamics. During forest succession after a stand‐replacing disturbance (e.g. fires), tree biomass and surface organic layer thickness increase in all stand types as forests recover; however, woody biomass detritus first decreases and then increases after ca. 80 yr. Soil C densities show slight decrease with ages in Populus stands, but increase in other stand types. These results indicate the complex C‐transfer processes among different components (tree biomass, detritus, forest floor, and soil) of boreal ecosystems at various stages of succession.     

An improved analysis of forest carbon dynamics using data assimilation

There are two broad approaches to quantifying landscape C dynamics – by measuring changes in C stocks over time, or by measuring fluxes of C directly. However, these data may be patchy, and have gaps or biases. An alternative approach to generating C budgets has been to use process‐based models, constructed to simulate the key processes involved in C exchange. However, the process of model building is arguably subjective, and parameters may be poorly defined. This paper demonstrates why data assimilation (DA) techniques – which combine stock and flux observations with a dynamic model – improve estimates of, and provide insights into, ecosystem carbon (C) exchanges. We use an ensemble Kalman filter (EnKF) to link a series of measurements with a simple box model of C transformations. Measurements were collected at a young ponderosa pine stand in central Oregon over a 3‐year period, and include eddy flux and soil CO₂ efflux data, litterfall collections, stem surveys, root and soil cores, and leaf area index data. The simple C model is a mass balance model with nine unknown parameters, tracking changes in C storage among five pools; foliar, wood and fine root pools in vegetation, and also fresh litter and soil organic matter (SOM) plus coarse woody debris pools. We nested the EnKF within an optimization routine to generate estimates from the data of the unknown parameters and the five initial conditions for the pools. The efficacy of the DA process can be judged by comparing the probability distributions of estimates produced with the EnKF analysis vs. those produced with reduced data or model alone. Using the model alone, estimated net ecosystem exchange of C (NEE)=?251±197 g C m^?2 over the 3 years, compared with an estimate of ?419±29 g C m^?2 when all observations were assimilated into the model. The uncertainty on daily measurements of NEE via eddy fluxes was estimated at 0.5 g C m^?2 day^?1, but the uncertainty on assimilated estimates averaged 0.47 g C m^?2 day^?1, and only exceeded 0.5 g C m^?2 day^?1 on days where neither eddy flux nor soil efflux data were available. In generating C budgets, the assimilation process reduced the uncertainties associated with using data or model alone and the forecasts of NEE were statistically unbiased estimates. The results of the analysis emphasize the importance of time series as constraints. Occasional, rare measurements of stocks have limited use in constraining the estimates of other components of the C cycle. Long time series are particularly crucial for improving the analysis of pools with long time constants, such as SOM, woody biomass, and woody debris. Long‐running forest stem surveys, and tree ring data, offer a rich resource that could be assimilated to provide an important constraint on C cycling of slow pools. For extending estimates of NEE across regions, DA can play a further important role, by assimilating remote‐sensing data into the analysis of C cycles. We show, via sensitivity analysis, how assimilating an estimate of photosynthesis – which might be provided indirectly by remotely sensed data – improves the analysis of NEE.     

Simulation of terrestrial carbon equilibrium state by using a detachable carbon cycle scheme

Determining the equilibrium state of terrestrial carbon is a prerequisite for scientific analysis on the carbon cycle. However, the mechanism through which the carbon cycle reaches the equilibrium state remains unclear. Moreover, the carbon cycle in most of the short–term field experiments rarely reaches the equilibrium state. In this study, a detachable carbon cycle (DCC) model was proposed to simulate the equilibrium state of each carbon pool. The model was established based on a pool–and–flux scheme and contained 14 carbon pools, or carbon flow processes, each process could be detached from the main model and evaluated as an independent component. The environmental scalar algorithms of the Integrated Terrestrial Ecosystem Carbon budget model (InTEC) and Community Atmosphere Biosphere Land Exchange (CABLE) were incorporated in the DCC model. Four situations were compared using the two environmental scalar algorithms and model structure (9 vs. 14 carbon pools). Furthermore, the size and turnover time of each carbon pool were analyzed at the equilibrium state. A sensitivity analysis was then conducted to investigate the responses of carbon density and equilibrium time to 12 key parameters of the model. Results indicated that the combination of the CABLE environmental scalar algorithm and 14 pools exhibited improved performance on carbon storage simulation than that of the other combinations, and the effect of the environmental scalar algorithm was considerably larger than that of the carbon pool number. Sensitivity analysis indicated that the carbon density of grassland and cropland was more vulnerable and sensitive to key parameters of the model than that of the other biomes. This study elucidates influencing factors and underlying control mechanisms in the carbon accumulation, and provides a framework for quantitative analysis of each component of the carbon cycle.     

Episodic stemflow inputs of magnesium and potassium to a tropical forest floor during heavy rainfall events

Summary Stemflow inputs of magnesium and potassium were measured from 57 canopy trees representing eight species under heavy rainfall conditions in two tropical forest sites in northeast Queensland, Australia. In the premontane tropical moist forest site on the Atherton Tableland, the stemflow input per unit trunk basal area of 51 canopy trees was found to be 0.46 g m^-2 of Mg²⁺ and 4.22 g m^-2 of K⁺ for an average wet season rainday of 99 mm. In the wetter montane tropical rainforest site on Mount Bellenden Ker, the stemflow input per unit trunk basal area of six canopy trees was 5.55 g m^-2 of Mg²⁺ and 9.12 g m^-2 of K⁺ for a wet season rainday of 38 mm. These stemflow inputs from single raindays are greater than the mean annual rainfall input and are almost of the same order of magnitude as the mean annual throughfall input of these cations to areas equal to the trunk basal area from which the stemflow was collected. Stemflow cation fluxes of this magnitude are mainly attributable to the funnelling of large quantities of rainwater down the trunks of these canopy trees by their thoroughly wetted, upwardly inclined branches.