Using plant functional attributes to quantify site productivity and growth patterns in mixed forests

doi:10.1016/S0378-1127(96)03972-2

Forest Ecology and Management

Volume 94, Issues 1–3, 30 June 1997, Pages 149-163

https://doi.org/10.1016/S0378-1127(96)03972-2 Get rights and content

Abstract

Forest growth models form one of several important prerequisites for sustainable management. The complexity of tropical moist forest means that there is often little objective information to classify sites and species for growth modelling and yield prediction. Classification based on observable morphological characteristics may be a useful surrogate for, or supplement to, other alternatives. This study investigated the utility of plant functional attributes (PFAs) for site and species classification. PFAs describe a plant in terms of its photosynthetic and vascular support system, and the sum of individual PFAs for all species on a plot provides an efficient summary of vegetation features at the site. Preliminary observations suggested that the PFA summary may also indicate site productivity, and that specific PFAs may be used to group species for modelling growth and yield. Data from 17 permanent plots in the tropical rainforests of North Queensland were used to test these preliminary observations. Standard PFA proformas were completed for each plot in January 1995, and the relationships between the PFAs, site productivity and specific growth patterns were examined using discriminant analysis, linear regression and standard statistical tests. Results indicate that mean leaf size, and the incidence of species with vertical leaf inclination (more than 30° above horizontal) are significantly correlated with site productivity. Of the PFAs assessed, five elements appear to offer a useful basis for grouping species for modelling: leaf size and inclination, a furcation index (i.e. relative height to first fork or break in the main stem), and the presence of lenticels and chlorophyllous tissue on the main stem. The restricted nature of our database limits comment on the general utility of the method, but results suggest that further work on PFAs is warranted.

References (21)

I.G.W. Corns et al.
Vegetational indicators as independent variables in forest growth prediction in west-central Alberta, Canada
For. Ecol. Manage.
(1984)
J.K. Vanclay
A growth model for north Queensland rainforests
For. Ecol. Manage.
(1989)
J.K. Vanclay
Aggregating tree species to develop diameter increment equations for tropical rainforests
For. Ecol. Manage.
(1991)
J.K. Vanclay
Assessing site productivity in tropical moist forests
For. Ecol. Manage.
(1992)
C.G.G.J. van Steenis
Rejuvenation as a factor for judging the status of vegetation types: the biological nomad theory
A.N. Gillison
Towards a functional vegetation classification
A.N. Gillison
A.N. Gillison et al.
A generic plant functional attribute set and grammar for vegetation description and analysis
CIFOR Working Paper No 3
(1994)
D. Mueller-Dombois et al.
C. Raunkiaer

There are more references available in the full text version of this article.

Cited by (23)

A rapid tree diversity assessment method for cocoa agroforestry systems
2021, Ecological Indicators
Citation Excerpt :
A literature review was carried out, specifically on existing alternatives to apply a rapid field assessment of biodiversity (Greco et al., 2012). After a preliminary selection, eight different methods were explored in detail, among which those based on the survey of plant functional types (Gillison, 2002; Gillison, 1996; Gillison et al., 2004; Vanclay et al., 1997), on significant or unique species (Mccullough et al., 2008; McCullough et al., 2007), on a plant assemblage analysis (Kuncoro et al., 2006) or that were specifically applied in agroforests (Oke and Odebiyi, 2007; Sonwa et al., 2009). The strengths and weaknesses of each method and their compatibility with the traceability and mapping system were analysed and discussed with the stakeholders (Greco et al., 2012).
Biodiversity is recognized as an essential part of sustainable development efforts, however reducing biodiversity loss is a key global challenge that requires updated data on biodiversity status at different scales. Cocoa agroforests include tree species besides cocoa, a practice beneficial to biodiversity, ecosystem conservation and farming households. We present a stepwise procedure to test and select a method that rapidly assesses biodiversity in cocoa agroforests based primarily on species richness and counts of non-cocoa trees. Three rapid assessment methodologies (RapidBAM) with different sampling procedures were tested in three phases: calibration, testing and evaluation. Results showed the method using the lowest number of sample plots with a minimum area coverage and a consistent sampling time (regardless of farm context) provided the most accurate and straightforward assessment. Farmers accurately reported qualitatively on species, complimenting quantitative data produced by RapidBAM. Collecting biodiversity data with RapidBAM proved valuable to collect data at large-scales and is applicable to different landscapes. Monitoring biodiversity with fewer required resources than conventional methods is a relevant outcome, which can help defining efficient biodiversity-friendly farming practices.
A new nonlinear model applied to the height-DBH relationship in Gmelina arborea
2017, Forest Ecology and Management
Citation Excerpt :
Although measurement of total tree height is relatively complex, time consuming, and expensive, DBH is conventionally measured for all trees sampled, and height is then estimated from the height-diameter model equations. As cited in the forest growth and yield articles and textbooks Vanclay et al. (1997) and Weiskittel et al. (2011), height-DBH models are important for use in models of forest growth and yield. Such models often rely on DBH measurements and a good model for height rather than direct height measurements.
Modeling of the relationship between diameter and height of trees generally employs one of several possible nonlinear growth models, with Richards and Schnute among the most commonly used in the literature. The recently defined hyperbolastic models are nonlinear models of sigmoidal growth noted for their flexibility in representation of biological phenomena. We apply these models to a set of data representing the relationship between diameter and height for a stand of Gmelina arborea in Ibadan, Nigeria and compare the performance with several other nonlinear models. We also compare these nonlinear models with a standard allometric relation, as well as an expanded version. In addition we display the use of the multivariable form of these hyperbolastic models for use in models representing height as a function of diameter as well as other additional explanatory variables.
Quantification of Quercus petraea Liebl. forking based on a 23-year-long longitudinal survey
2012, Forest Ecology and Management
Citation Excerpt :
In even-aged and two-storey stands, it defines the boundary between the environment of the photosynthetic active canopy and that of the understorey, which have different microclimatic conditions in terms of light, wind, temperature and air moisture (Aussenac and Ducrey, 1977). The location of the lowest fork has been identified as a useful attribute of plant function to distinguish species groups when quantifying the site productivity of mixed forests (Vanclay et al., 1997). The leaf area index, which is related to forest productivity, is also affected by the forking pattern (Kull and Tulva, 2000).
For Quercus petraea (sessile oak), the origin of forks is either traumatic or related to crown base formation (“metamorphosis”). The objectives of this study were to quantify forking, distinguish the two origins and test the effect of initial stocking density and tree attributes on the number of forks and height to the lowest one.
The experiment took place in a 23-year-old sessile oak plantation in Normandy, France, where three initial planting densities were compared (1333, 2667 and 5333 seedlings per ha). The forks were observed every 3 years after the plantation. Fork number was modelled with a system of simultaneous linear equations that respectively predicted tree girth, height and the log-transformed fork frequency, with age and initial stocking density as explanatory variables. This model 1 was fitted as a seemingly unrelated regression with random effects to account for the tree and plot clustering. For trees having at least one fork, lowest fork height was quantified using a linear mixed model including predictors compatible with the response variables of model 1, age and stocking.
The results showed that the number of forks can be predicted using the expected explanatory variables and it is significantly correlated with height and girth as well. Traumatic and metamorphosis-related origins were not clearly distinguished, but significant differences between the initial seedling densities were found, with the lowest density showing the highest number of forks. Lowest fork height depended on response variables of model 1, age and stocking. The stocking effect was significant but small.
This study presents the first statistical quantification of tree forking ever made. From this quantification, it seemed obvious that the initial density of 2667 seedlings per ha could be a relevant initial density to avoid the emergence of too many forks and the insertion of the lowest forks too low down.
Modelling the impacts of various thinning intensities on tree growth and survival in a mixed species eucalypt forest in central Gippsland, Victoria, Australia
2008, Forest Ecology and Management
The response of tree survival and diameter growth to thinning treatments was examined over 29 years, in various thinning treatments established in a 21-year-old even-aged mixed species regenerating forest in Victoria, Australia. The treatments were control, crown release, strip thinning and three different intensities of thinning from below (light, moderate, and heavy). Each treatment was replicated three times in a complete randomised design. Logistic and multilevel regression analyses showed that tree survival, growth and thinning response (change of tree growth due to a thinning treatment) were functions of tree species, size, age, removed and remaining competition, as well as time since the treatment. Mean annual tree diameter growth in unthinned stands was highest for Eucalyptus sieberi L. Johnson (1.9 mm) followed by Eucalyptus baxteri (Benth.) Maiden & Blakely ex J. Black (1.6 mm), and lowest for both Eucalyptus consideniana (Maiden) and Eucalyptus radiata (Sieber ex DC) combined (0.7 mm). Diameter growth increased with tree size for both E. sieberi and E. baxteri, but not for E. consideniana and E. radiata. Smaller trees were more likely to die due to shading and suppression than their larger counterparts. A mortality model suggested, however, that both shading and suppression had very little effect on trees in both E. consideniana and E. radiata species, which were less likely to die compared to trees in the other species. This result indicates that both E. consideniana and E. radiata species may be relatively shade tolerant compared with the other species. Total thinning response was a sum of positive (increased growing space) and negative (thinning stress) effects. Following thinning, smaller trees showed signs of thinning stress for the first one or two years, after which the highest percentage thinning response was observed. While larger trees were initially less responsive to thinning, the rate of decrease in the response for subsequent years was greater in smaller trees than larger ones. The average amount of thinning response showed similar trends to diameter growth increasing from E. sieberi (1.7 mm) through E. baxteri (0.6 mm) to both E. consideniana and E. radiata (0.5 mm). This translates into low average percentage thinning response in E. baxteri (34%), twice as much in both E. consideniana and E. radiata (69%) and highest overall percentage response in E. sieberi (87%). Thinning response and the duration of this response appeared to increase with thinning intensity and was still evident 29 years after thinning. Heavy thinning did, however, reduce the number of trees to a severely under-stocked condition, which prohibited optimum site occupancy, requiring 29 years of post-thinning development for the heavily thinned stands to regain their pre-thinning stand basal area.
Efficient floristic inventory for the assessment of tropical tree diversity: A comparative test of four alternative approaches
2006, Forest Ecology and Management
Biodiversity inventory as a tool for guiding conservation planning at a local scale is under used, especially in tropical countries where technical capacity is often limited. Here four inventory protocols (one ad hoc, one fixed count and two fixed area methods) are tested for their efficiency and statistical robustness when applied to the woody floras of a Mexican seasonally dry tropical forest. They are tested for their ability to distinguish between sites not only by number of species (richness), but also by number of threatened species, and finally in their application to complementary reserve selection exercises. Whilst the ad hoc method is shown to have the advantage of high efficiency and is technically undemanding, where possible the statistically more robust fixed count or fixed area methodologies are recommended. However, the much used fixed area method of Gentry [Gentry, A.H., 1982. Patterns of neotropical plant species diversity. Evol. Biol. 15, 1–84] that uses 2 m × 50 m plots is shown to be inefficient. Only when comparison with other previously published assessments is a priority should the 2 m × 50 m fixed area method be considered, and then efficiency may be increased by widening the plots and distinguishing between those stems that fall within a notional 2 m plot and those outside of it. Both prioritisation by number of species and complementarity analysis, using the ‘greedy’ algorithm, are shown to be sensitive to choice of inventory protocol, suggesting that consideration is given to how species lists are obtained in reserve selection exercises.
Assessing biodiversity at landscape level in northern Thailand and Sumatra (Indonesia): The importance of environmental context
2004, Agriculture, Ecosystems and Environment
Most biodiversity assessment methods tend to sample isolated areas of land cover such as closed forest or local land use mosaics. Contemporary methods of assessing biodiversity are briefly reviewed and focus on the relative roles of the Linnean species and plant functional types (PFTs). Recent case studies from central Sumatra and northern Thailand indicate how the range distributions of many plant and animal species and functional types frequently extend along regional gradients of light, water and nutrient availability and corresponding land use intensity. We show that extending the sampling context to include a broader array of environmental determinants of biodiversity results in a more interpretable pattern of biodiversity. Our results indicate sampling within a limited environmental context has the potential to generate highly truncated range distributions and thus misleading information for land managers and for conservation. In an intensive, multi-taxa survey in lowland Sumatra, vegetational data were collected along a land use intensity gradient using a proforma specifically designed for rapid survey. Each vegetation sample plot was a focal point for faunal survey. Whereas biodiversity pattern from samples within closed canopy rain forest was difficult to interpret, extending the sample base to include a wider variety of land cover and land use greatly improved interpretation of plant and animal distribution. Apart from providing an improved theoretical and practical basis for forecasting land use impact on biodiversity, results illustrate how specific combinations of plant-based variables might be used to predict impacts on specific animal taxa, functional types and above-ground carbon. Implications for regional assessment and monitoring of biodiversity and in improving understanding of the landscape dynamics are briefly discussed.

View all citing articles on Scopus

View full text

Using plant functional attributes to quantify site productivity and growth patterns in mixed forests

Abstract

For. Ecol. Manage.

For. Ecol. Manage.

For. Ecol. Manage.

For. Ecol. Manage.

Towards a functional vegetation classification

A generic plant functional attribute set and grammar for vegetation description and analysis

CIFOR Working Paper No 3