Site index estimation for clonal eucalypt plantations in Brazil: A modeling approach refined by environmental variables
Introduction
The productive potential of forest stands is commonly identified through assessing site quality (Burkhart and Tome, 2012) and site index has been historically applied for this purpose (Peng, 2000). The definition of productive potential of forest stands is a key point for developing silvicultural prescriptions and more recently for informing how climate variation affects forest productivity (Sabatia and Burkhart, 2014).
The main advantage of using site index, which is the dominant height at a defined base age, is its simplicity and robustness for assessing site quality. Dominant height is largely independent of stand density and it is believed that this variable integrates the history of the climate and soil effects on forest productivity (lles, 2003). At present, the use of this single variable for assessing site quality has been argued by some as incomplete (Martin et al., 2006). It seems doubtful, however, that the substitution of site index by soil and climate variables increases the prediction power of a growth modeling approach, especially when conceptually site index combines such variables (lles, 2003).
Burkhart and Tome (2012) discussed the relationship between model dimensionality and long projection lengths. They suggested that simpler models display higher inference power, which implies that the substitution of the single site index variable by several environmental variables (soil and climate) does not necessarily bring extra information as initially thought. The compounding error component when predicting or projecting forest growth through the use of the environmental variables can be substantial and multicollinearity among these variables can cause problems with respect to inference about forest growth (Weiskittel et al., 2011).
Empirical dominant height growth models provide estimates of productive potential of forest stands, although at the sacrifice of explanatory ability, which may lead to improper inferences if the climate history changes over time (Casnati, 2016). This fact does not imply the need for replacing site index when assessing site quality of forest stands, but it suggests the need for the inclusion of influential covariates to increase modeling performance. There are some approaches proposed in the literature to increase the explanatory ability of dominant height growth equations. One highlighted approach is using statistical growth equations with parameters refined by environmental variables (Weiskittel et al., 2011).
Sharma et al. (2015), following this approach, introduced climate variables in the dominant height modeling for Pinus banksiana Lamb. and Picea mariana [Mill.] B.S.P. stands in Canada. The authors highlighted that covariates such as growing season mean temperature and growing season total precipitation greatly increased the estimation accuracy of dominant height. Nunes et al. (2011) included soil and climate variables for modeling dominant height of Pinus pinaster Ait. stands in Portugal. They reported that the new growth equation outperformed the empirical dominant height model. Bravo-Oviedo et al. (2008) verified an increase in the dominant height estimation accuracy of Pinus pinaster Ait stands in Spain through the use of a modeling approach refined by environmental variables.
On the other hand, Wang et al. (2007) fitted a dominant height equation with a non-linear mixed model for Eucalyptus globulus Labill. stands in Australia and verified only a minor increase in dominant height estimation accuracy through the inclusion of environmental variables. They stated however that annual rainfall and average daily maximum temperature in July (winter) helped to greatly reduce the residual variability among the plots. Ung et al. (2001) also reported low impact after adding environmental covariates for modeling site index of Boreal tree species.
Scolforo et al. (2017) reported that the inclusion of a covariate, related to soil water availability in dominant height growth modeling, increased the explanatory ability when assessing site quality for clonal eucalypt stands in Brazil. In clonal eucalypt stands, the yearly rainfall variable provides a modest improvement in growth modeling by furnishing either more accurate site-specific curves or to mimic short-term climate changes (Scolforo et al., 2016). The productivity of clonal eucalypt plantations, as well as other tree species, is constrained by soil water availability (Spittlehouse, 2003), which implies that yearly rainfall plays a key role for forest growth, while potential evapotranspiration and soil water storage capacity are variables that should be considered as well.
There is a dearth of studies for clonal eucalypt plantations in Brazil that investigate the use of environmental variables in growth equations, where extra attention for the identification of the most proper environmental variable for increasing the explanatory ability of growth modeling should be taken (Sharma et al., 2015). In practical terms, there is a lack of generalized growth equations applied to Brazil with the capacity of precisely updating forest inventory when making long-term projections. Along the Brazilian environmental or climatic gradient there is a need to simulate how climate variation affects the site quality across this country. In addition, there is a lack of growth equations that can predict site quality for areas without forest plantation records.
This paper provides (1) a test of competing growth equations to simultaneously predict and project dominant height/site index in clonal eucalypt plantations across Brazil; (2) the identification and inclusion of the most influential environmental variable that increases the explanatory ability of the dominant height/site index equations; (3) a validation of the performance of a new compatible set of prediction and projection growth equations refined by environmental variables.
Section snippets
Study area and database
The data are composed of remeasurement information from 16 research sites distributed along tropical Brazil. These research sites labeled as: 2, 4, 5, 7, 9, 11, 13, 14, 20, 22, 24, 26, 29, 30, 31 and 33 (Fig. 1) are part of a research project denominated as TECHS (Tolerance of Eucalyptus Clones to Hydric, Thermal and Biotic Stresses, www.ipef.br/techs/en). The TECHS project was launched in 2011 to investigate clonal eucalyptus growth from northern Uruguay to northern Brazil. The project aimed
Clonal eucalypt dominant height growth under different climatic scenarios in Brazil
Statistically significant effects (at alpha = 0.05) of stand age, genotype and water scenario were observed on clonal eucalypt dominant height growth in Brazil. The statistical analysis highlighted that dominant height growth pattern changes according to the water scenario (Table 4). Table 1 indeed highlights that dominant height growth over time displays an overall reduction under drier climatic conditions (verified in the plots under the rainfall exclusion regime). This conclusion is
Discussion
This paper provides a set of prediction and projection dominant height/site index equations governed by water availability (environmental variable annual soil water deficit) for a variety of clonal eucalypt groups. Among all the tested growth equations, the compatible set of the Chapman-Richards prediction model, combined to the dynamic polymorphic model with single asymptote, resulted in the most accurate estimates of dominant height/site index for clonal eucalypt plantations in Brazil. The
Conclusions
Competing growth equations were simultaneously fitted for the prediction and projection of dominant height and site index in clonal eucalypt stands in Brazil. The compatible prediction/projection Chapman-Richards model, featuring polymorphism with single asymptote, displayed the most accurate estimates.
Among all the tested environmental variables, the annual soil water deficit was the single variable that best described dominant height growth. The refining of the common asymptote parameter with
CRediT authorship contribution statement
Henrique Ferraco Scolforo: Conceptualization, Methodology, Formal analysis, Validation, Writing - original draft. John Paul McTague: Conceptualization, Methodology, Writing - review & editing. Harold Burkhart: Conceptualization, Methodology, Writing - review & editing. Joseph Roise: Methodology, Writing - review & editing. Clayton Alcarde Alvares: Visualization, Writing - review & editing. Jose Luiz Stape: Funding acquisition, Methodology, Writing - review & editing.
Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Acknowledgments
The authors are especially grateful to the Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq - Brazil) for the scholarship provided to develop this research (249979/2013-6), to the Instituto de Pesquisas e Estudos Florestais (IPEF – Brazil) and to the TECHS project. We thank all the companies, universities and research institutions involved in the TECHS Project. The project was funded by the following 26 companies, with a main researcher: Anglo American (Andre Machado), Arauco
References (40)
- et al.
Mapping the effect of spatial and temporal variation in climate and soils on Eucalyptus plantation production with 3-PG, a process-based growth model
For. Ecol. Manage.
(2010) - et al.
The interactions of climate, spacing and genetics on clonal Eucalyptus plantations across Brazil and Uruguay
For. Ecol. Manage.
(2017) - et al.
Modeling dominant height growth of radiata pine (Pinus radiata D. Don) plantations in north-western Spain
For. Ecol. Manage.
(2005) Growth and yield models for uneven-aged stands: past, present and future
For. Ecol. Manage.
(2000)- et al.
Predicting site index of plantation loblolly pine from biophysical variables For
Ecol. Manage.
(2014) - et al.
Modeling dominant height growth of eucalyptus plantations with parameters conditioned to climatic variations
For. Ecol. Manage.
(2016) - et al.
Incorporating rainfall data to better plan Eucalyptus clones deployment in eastern Brazil
For. Ecol. Manage.
(2017) - et al.
Yield pattern of eucalypt clones across tropical Brazil: an approach to clonal grouping
For. Ecol. Manage.
(2019) - et al.
Modeling whole-stand survival in clonal eucalypt stands in Brazil as a function of water availability
For. Ecol. Manage.
(2019) - et al.
Eucalyptus growth and yield system: linking individual-tree and stand-level growth models in clonal Eucalypt plantations in Brazil
For. Ecol. Manage.
(2019)
Köppen’s climate classification map for Brazil
Meteorol. Z.
Modeling dominant height growth: eucalyptus plantations in Portugal
For. Sci.
Base-age invariant polymorphic site curves
For. Sci.
Dominant height growth equations including site attributes in the generalized algebraic difference approach
Can. J. For. Res.
Modeling Forest Trees and Stands
Modeling dominant height growth based on nonlinear mixed-effects model: a clonal Eucalyptus plantation case study
For. Ecol. Manage.
Hybrid mensurational-physiological models for Pinus taeda and Eucalyptus grandis in Uruguay
Generalized algebraic difference approach: theory based derivation of dynamic site equations with polymorphism and variable asymptotes
For. Sci.
Site classification for eucalypt stands using artificial neural network based on environmental and management features
Cerne
Cited by (13)
Climate change altered the dynamics of stand dominant height in forests during the past century – Analysis of 20 European tree species
2024, Forest Ecology and ManagementSoil morphological, physical and chemical properties affecting Eucalyptus spp. productivity on Entisols and Ultisols
2023, Soil and Tillage ResearchFertilization management with sewage sludge sustains Eucalyptus productivity in Cerrado infertile soil
2022, Forest Ecology and ManagementCitation Excerpt :To check the regressions, the adjusted coefficient of determination (R2 adj.), root mean square error (RMSE), and mean absolute error (MAE) were calculated. MAE and RMSE statistics closer to zero indicate a better fit (Scolforo et al., 2020). In order to verify the best SS fertilization management (fixed variables) for the Eucalyptus plantation, the accumulated TCSA and the wood biomass production (random variables) data of each inventory (22, 35, and 54 months after planting) were subjected to analysis of variance by the F test (p < 0.05) and, when significant, multiple comparison tests were performed using the Duncan test (p < 0.05).
Thinning-response modifier term in growth models: An application on clonal Tectona grandis Linn F. stands at the amazonian region
2022, Forest Ecology and ManagementCitation Excerpt :Thus, meristematic activity depends on site conditions, and more specifically on the duration of dry seasons for tropical species such as teak (Tondjo et al., 2018). Including the dominant height prior to thinning in f thin allows for the capture of the remaining forest growth for different productive capacity classes, as dominant height is an effective integrator of key biological growth determinants (Kuehne et al., 2016; Scolforo et al., 2020; Weiskittel et al., 2011) that is less affected by thinning, and is an indicator of site productivity (Burkhart and Tomé 2012; Kuehne et al. 2016; Scolforo et al. 2020). Although essential to explain growth and production, the effect of thinning on plant physiology was not evaluated in this study.
Variable selection for estimating individual tree height using genetic algorithm and random forest
2022, Forest Ecology and ManagementCitation Excerpt :The use of soil variables indicates a direct effect between tree height and climatic characteristics. As the Rainforest has taller trees and the Cerrado (Brazilian Savannah) has shorter trees, it becomes reasonable to associate the soil metrics as explanatory variables (Bugmann and Solomon, 2000; Nunes et al., 2011; Scolforo et al., 2020, 2019). Generally, dominant height is responsible for biogeographic patterns that affect the tree height and the vegetation pattern (Ferraz Filho et al., 2018; Nishizono et al., 2014; Scolforo et al., 2016; Sharma et al., 2011).
Developing a site index model for P. Pinaster stands in NW Spain by combining bi-temporal ALS data and environmental data
2021, Forest Ecology and ManagementCitation Excerpt :In agreement with other studies in eucalyptus and pine stands in Brazil (Scolforo et al., 2013; Campoe et al., 2016; Scolforo et al., 2020), our findings suggest that the inclusion of Pe as predictor could integrate the effect of physiographical, soil and climate variables in dominant height growth modelling, thereby increasing the ability of the model to estimate site quality. The SI model proposed by Scolforo et al. (2020) included mean annual soil water deficit (mm year−1), which used the monthly differences between potential and actual evapotranspiration, calculated by the water balance model of Thornthwaite-Mather. Campoe et al. (2016) found that Pe, along with other variables such as maximum temperature and vapour deficit, were related to the annual increase in tree basal area in pine and eucalyptus plantations in Brazil.