Final Report Summary - SPATFOREST (Spatial dynamics of tropical forest biomass change)
International climate change mitigation policies envisage carbon storage in tropical forests.
However, we have limited capacity to forecast future changes in tree biomass in response to climate change due to a lack of knowledge of how trees will respond to changing abiotic and biotic environments. Our project, known as SpatForest and funded by the EU, aimed to derive a better understanding of tropical carbon dynamics by developing statistical models of above-ground biomass, tree demography (tree growth and mortality) and below-ground carbon storage in tropical forests, that explicitly account for variation in local biotic and abiotic environments, such us topography and soils. In order to make progress on this topic we had to first develop improved equations to estimate biomass in tropical trees because of drawbacks in previous methods.
The first task was to examine fine-scale spatial distribution of tree biomass in a relatively undisturbed tropical forest on Barro Colorado Island, Panama. The key findings from this work were that above-ground tree biomass is not randomly at a small scale, as previous results have suggested, but depends on both biotic and abiotic conditions. Lianas (woody vines) reduce tree biomass accumulation but soils that are less acid and richer in nutrients (especially potassium) enhance above-ground biomass accumulation. The main practical implication of these findings is that it is possible to predict forest areas that sequester more carbon, which represents an opportunity for targeted conservation measures for forest carbon management. . Moreover, an increase in liana abundance has been observed in many tropical forests in recent years, and our results suggest that these changes may reduce their potential for carbon storage in above-ground biomass. In principle, lianas can be controlled by forest managers to enhance carbon storage without loss of biodiversity.
The second task was to uncover the factors that influence tree growth and mortality for individuals growing on a long-term forest dynamics plot, and here again the optimal study site was the plot on Barro Colorado Island, Panama, where data for 250 different species are available. Tree growth is a potentially important sink for atmospheric carbon dioxide and may contribute significantly to changes in above-ground biomass. Our results suggest that the main factor explaining three growth was competition among individual trees, suggesting that trees with a greater number of larger neighbours grow more slowly. However, tree growth was enhanced in soils with greater nutrient availability, which mitigated the impacts of competition among neighbours. Spatial variability was more important than temporal variation over 20 years in explaining the growth rates of trees, even though tree size also explained a component of variation in the growth of individuals. We explored the distribution of tree growth and mortality rates among different evolutionary lineages of trees and discovered that they have evolved in parallel. The main implications of these findings is that tree growth rates can be managed by reducing competition among individuals and is enhanced by soil nutrient availability. Most importantly, we observed that the largest trees have the highest growth rates, as well as store more carbon. These result suggest that conservation actions need to target the largest trees in a forest.
The third task was to compile data on harvested whole trees, including excavated root biomass, from a global network of collaborators. The final data-set from 409 sites distributed across all forested biomes around the globe. This data-set is the most extensive compiled to date on root biomass in forests, which is important for forest carbon balance studies because the ratio of above to below biomass is currently unknown in tropical forests and a uniform value of 20% in roots has been generally assumed. A modification to account for biome has recently being incorporated in the models developed by the Inter-governmental Panel on Climate Change. Nevertheless, plasticity in the allocation of carbon above and below-ground within species, and the large-scale drivers of differential allocation were unquantified until now. We found that tree size was the main predictor of root biomass, followed by climatic factors linked to water availability. Smaller (most likely younger) trees and trees in water-deficient areas invest relatively more of their biomass in roots, probably to reach deeper water sources. The main implication derived from this study is that the below ground carbon compartment in forests should not be considered as a unchanging black box, as it has been until now. Differences in carbon allocation among species, size classes and regions should be incorporated into global carbon cycle models. Our study quantifies that variation, and provides predicted values of below-ground allocation that should be considered in future studies.
SpatForest has identified important limitations over the ways that carbon is estimated, which is generally achieved using equations that define biomass based on easily measured variables such as diameter or height. We found that some widely-used equations that define tree shape and then estimate trees biomass are inaccurate and sometimes highly biased. We focused our study on allometric equations that describe tree height as a function of tree diameter. We compared all previously published equations and showed that the equation that was most successful at predicting above-ground biomass was rarely used in practice. We encourage future researchers to use this equation (the Weibull function) from now on. This recommendation is important because estimating the uncertainty of carbon pools is necessary to obtain a realistic view of current carbon storage and carbon dynamics in tropical forests. Moreover, many previous studies are likely to provide biased estimates of biomass due to the method of tree height estimation. These inaccurate and biased estimates of carbon storage may lead an inappropriate distribution of the benefits of carbon storage through the UN-REDD mechanism, because these are allocated in proportion to total estimated above-ground biomass. A fair and accurate mechanism for estimating carbon in trees and forests should therefore be implemented in the UN-REDD+ programme, to ensure the absence of discrimination among beneficiaries. The equations we proposed can be used in the future not only by the scientific community but by forest resource managers.
General implications of SpatForest:
The models developed during SpatForest will enhance our ability to forecast changes in biomass of tropical forests and contribute to policy development by the EU and the broader international community. Reviews of the global carbon cycle generated by the Intergovernmental Panel on Climate Change (IPCC) and policy initiatives to mitigate atmospheric carbon dioxide emissions are reliant on accurate quantification and assessment of biomass in forest vegetation. Under the UN-REDD+ programme, countries are rewarded for preserving carbon stocks in vegetation and for actions taken to reduce CO2 emissions through avoided deforestation and forest degradation. Our research suggests that carbon sequestration in natural tropical forests would be promoted by defining areas for conservation based on soil conditions and stand structure. It is also important to maintain old growth forest and large trees; because big trees capture more carbon than small trees, display faster growth rates and suffer less from competition. A further enhancement of carbon storage would occur in response to reduction in the abundance through management, which would counter recent increases in liana abundance observed in many tropical forests.
However, we have limited capacity to forecast future changes in tree biomass in response to climate change due to a lack of knowledge of how trees will respond to changing abiotic and biotic environments. Our project, known as SpatForest and funded by the EU, aimed to derive a better understanding of tropical carbon dynamics by developing statistical models of above-ground biomass, tree demography (tree growth and mortality) and below-ground carbon storage in tropical forests, that explicitly account for variation in local biotic and abiotic environments, such us topography and soils. In order to make progress on this topic we had to first develop improved equations to estimate biomass in tropical trees because of drawbacks in previous methods.
The first task was to examine fine-scale spatial distribution of tree biomass in a relatively undisturbed tropical forest on Barro Colorado Island, Panama. The key findings from this work were that above-ground tree biomass is not randomly at a small scale, as previous results have suggested, but depends on both biotic and abiotic conditions. Lianas (woody vines) reduce tree biomass accumulation but soils that are less acid and richer in nutrients (especially potassium) enhance above-ground biomass accumulation. The main practical implication of these findings is that it is possible to predict forest areas that sequester more carbon, which represents an opportunity for targeted conservation measures for forest carbon management. . Moreover, an increase in liana abundance has been observed in many tropical forests in recent years, and our results suggest that these changes may reduce their potential for carbon storage in above-ground biomass. In principle, lianas can be controlled by forest managers to enhance carbon storage without loss of biodiversity.
The second task was to uncover the factors that influence tree growth and mortality for individuals growing on a long-term forest dynamics plot, and here again the optimal study site was the plot on Barro Colorado Island, Panama, where data for 250 different species are available. Tree growth is a potentially important sink for atmospheric carbon dioxide and may contribute significantly to changes in above-ground biomass. Our results suggest that the main factor explaining three growth was competition among individual trees, suggesting that trees with a greater number of larger neighbours grow more slowly. However, tree growth was enhanced in soils with greater nutrient availability, which mitigated the impacts of competition among neighbours. Spatial variability was more important than temporal variation over 20 years in explaining the growth rates of trees, even though tree size also explained a component of variation in the growth of individuals. We explored the distribution of tree growth and mortality rates among different evolutionary lineages of trees and discovered that they have evolved in parallel. The main implications of these findings is that tree growth rates can be managed by reducing competition among individuals and is enhanced by soil nutrient availability. Most importantly, we observed that the largest trees have the highest growth rates, as well as store more carbon. These result suggest that conservation actions need to target the largest trees in a forest.
The third task was to compile data on harvested whole trees, including excavated root biomass, from a global network of collaborators. The final data-set from 409 sites distributed across all forested biomes around the globe. This data-set is the most extensive compiled to date on root biomass in forests, which is important for forest carbon balance studies because the ratio of above to below biomass is currently unknown in tropical forests and a uniform value of 20% in roots has been generally assumed. A modification to account for biome has recently being incorporated in the models developed by the Inter-governmental Panel on Climate Change. Nevertheless, plasticity in the allocation of carbon above and below-ground within species, and the large-scale drivers of differential allocation were unquantified until now. We found that tree size was the main predictor of root biomass, followed by climatic factors linked to water availability. Smaller (most likely younger) trees and trees in water-deficient areas invest relatively more of their biomass in roots, probably to reach deeper water sources. The main implication derived from this study is that the below ground carbon compartment in forests should not be considered as a unchanging black box, as it has been until now. Differences in carbon allocation among species, size classes and regions should be incorporated into global carbon cycle models. Our study quantifies that variation, and provides predicted values of below-ground allocation that should be considered in future studies.
SpatForest has identified important limitations over the ways that carbon is estimated, which is generally achieved using equations that define biomass based on easily measured variables such as diameter or height. We found that some widely-used equations that define tree shape and then estimate trees biomass are inaccurate and sometimes highly biased. We focused our study on allometric equations that describe tree height as a function of tree diameter. We compared all previously published equations and showed that the equation that was most successful at predicting above-ground biomass was rarely used in practice. We encourage future researchers to use this equation (the Weibull function) from now on. This recommendation is important because estimating the uncertainty of carbon pools is necessary to obtain a realistic view of current carbon storage and carbon dynamics in tropical forests. Moreover, many previous studies are likely to provide biased estimates of biomass due to the method of tree height estimation. These inaccurate and biased estimates of carbon storage may lead an inappropriate distribution of the benefits of carbon storage through the UN-REDD mechanism, because these are allocated in proportion to total estimated above-ground biomass. A fair and accurate mechanism for estimating carbon in trees and forests should therefore be implemented in the UN-REDD+ programme, to ensure the absence of discrimination among beneficiaries. The equations we proposed can be used in the future not only by the scientific community but by forest resource managers.
General implications of SpatForest:
The models developed during SpatForest will enhance our ability to forecast changes in biomass of tropical forests and contribute to policy development by the EU and the broader international community. Reviews of the global carbon cycle generated by the Intergovernmental Panel on Climate Change (IPCC) and policy initiatives to mitigate atmospheric carbon dioxide emissions are reliant on accurate quantification and assessment of biomass in forest vegetation. Under the UN-REDD+ programme, countries are rewarded for preserving carbon stocks in vegetation and for actions taken to reduce CO2 emissions through avoided deforestation and forest degradation. Our research suggests that carbon sequestration in natural tropical forests would be promoted by defining areas for conservation based on soil conditions and stand structure. It is also important to maintain old growth forest and large trees; because big trees capture more carbon than small trees, display faster growth rates and suffer less from competition. A further enhancement of carbon storage would occur in response to reduction in the abundance through management, which would counter recent increases in liana abundance observed in many tropical forests.