Abstract

This article introduces the Tropical Biomass & Carbon Application – the ‘TB&C App’, a web application available on the permanent link www.tropicalbiomass.com. The TB&C App requires as input attributes ‘the smallest and largest diameters’, ‘number of trees ha􀀀 1’, basal area ha􀀀 1, and ‘parameters of the diameter (beta) distribution’ describing stand structure. The App delivers outputs at two levels: (1) Stand level, including mean aboveground biomass (AGB) and carbon (AGC), in Mg ha􀀀 1, along with confidence intervals (CIs) as measures of uncertainty, and; (2) Tree level estimates, with AGB and diameter for every simulated tree. Phase 1 of the project TB&C comprises four Brazilian forest (and non-forest) formations: Campinarana, Floresta estacional, Floresta ombrofila, and Savana. This article aims to (i) describe the algorithm written for the TB&C App, and (ii) present results of Phase 1. This first phase counts on a standardized database of 1,428 trees with field-measured dry AGB, from plots across the different formations, which is the largest tree-biomass database compiled so far in Brazil. Model uncertainties were incorporated into the modeling process, allowing computation of CIs through an uncertainty approach. The total variance of residuals of AGB was also modeled, aiming at predicting CIs as a function of the quantity of AGB. An analysis of reliability of the equations implemented in the TB&C App indicates that more than 95% (n = 64,000) of the true AGB’s fit into the CI outputted by the TB&C App. A comparison with other approaches in the literature shows significant agreement with previous estimates and more conservative estimates where previously-published estimates disagreed with the TB&C App. We cite as advantages of the TB&C App; (i) reliability of the outputs, (ii) a user-friendly layout, (iii) AGB and AGC estimates provided along with robust CIs, and (iv) estimates at the stand and tree levels with consistent totals. A biomass dataset containing information on 64,000 plots is also delivered as supplement of this paper.

Abstract

The sources of uncertainty in wall-to-wall AGB maps propagate from the tree to pixel, but uncertainty due to forest cover mapping is rarely incorporated into the error propagation process. This study aimed to (1) elaborate an analytical procedure to incorporate forest-mapping-related uncertainty into the error propagation from plot and pixel predictions; (2) develop a stratified estimator with a model-assisted estimator for small and large areas; and (3) estimate the effect of ignoring the mapping uncertainty on the confidence intervals (CIs) for totals. Data consist of a subset of the Brazilian national forest inventory (NFI) database, comprising 75 counties that, once aggregated, served as strata for the stratified estimator. On-ground data were gathered from 152 clusters (plots) and remotely sensed data from Landsat-8 scenes. Four major contributions are highlighted. First, we describe how to incorporate forest-mapping-related uncertainty into the CIs of any forest attribute and spatial resolution. Second, stratified estimators perform better than non-stratified estimators for forest area estimation when the response variable is forest/non-forest. Comparing our stratified estimators, this study indicated greater precision for the stratified estimator than for the regression estimator. Third, using the ratio estimator, we found evidence that the simple field plot information provided by the NFI clusters is sufficient to estimate the proportion forest for large regions as accurately as remote-sensing-based methods, albeit with less precision. Fourth, ignoring forest-mapping-related uncertainty erroneously narrows the CI width as the estimate of proportion forest area decreases. At the small-area level, forest-mapping-related uncertainty led to CIs for total AGB as much as 63% wider in extreme cases. At the large-area level, the CI was 5–7% wider.

Abstract

Background: This study reveals the surprising impact of large trees on biomass modeling and estimation in tropical forests. Findings emerged from viewing tropical forests as an Extremistan environment—a domain where a small number of extreme events disproportionately impact overall outcomes. The aims were to: (i) determine whether humid tropical forests can be characterized as an Extremistan environment, (ii) quantify the impact of large trees on the biomass quantification, and (iii) recommend better practices to mitigate the impact of large trees. The methods included forest simulation, biomass model calibrated with multi datasets and extensive examination of the impact of large trees on model performance and mean biomass estimation. Results: The select group of the 1% heaviest trees account for 25–35% of the total biomass, a concentration analogous to the wealth concentration in developed countries. Additionally, a “tyranny” of the 5% heaviest trees (diameter >18–31 cm) was observed, in which 50-75% of the total biomass is retained, significantly affecting biomass modeling and mean biomass estimation regardless of the model used. Conclusions: This study confirms that humid tropical forests behave as an Extremistan environment. For biomass and carbon inventories, installing 10,000-m² sample units is recommended to mitigate the “tyranny” effect of the 5% heaviest trees, with a minimum size threshold of 4000 m².

Abstract

Researchers on forest biomass seek to develop or find methods of simple application, whose estimators reveals the best performance and provide the additivity of the biomass parts. Tree biomass normally is obtained by scaling tree parts (stem, branches, and leaves) separately. When the biomass is modeled by parts of the tree, their sum does not match the total, which means lack of additivity in the tree biomass estimation. We aimed to test and propose a modeling technique for estimating total aboveground biomass, so that consistency between tree parts will be achieved. The techniques ‘ratio estimators’ and ‘weighted apparently unrelated nonlinear regressions (WNSUR)’ were tested using a biomass dataset with 387 trees. Data were collected in eight sites located in Paraná and Rio Grande do Sul, Brazil, where information was collected on diameter at 1.30 m aboveground (dbh), heights (total, stem and crown), total biomass aboveground and its parts: stem, branches and leaves. All trees were identified at species level. To adjust the ratio equations, the cylinder volume (vcl), based on the mean square diameter (d²qi), was used as the independent variable and the total aboveground biomass and the respective parts: stem, branches, and leaves, as dependent variables. The occurrence of heteroscedasticity in the ratio estimators indicated that the data should be stratified into two stages, the first by dbh and the second by the slope of the ratio between the total aboveground biomass and the vcl of each tree. To group the trees in the second stage, a posteriori, a discriminant analysis was applied. The standard errors of the estimate resulted in less than 4% for the ratio estimators, while for the WNSUR they reached values larger than 40% for the total biomass aboveground and its parts. The statistics: bias, MAE, MSE and RMSE showed better performance by ratio estimators when compared to the WNSUR, for the total biomass and its parts except for the bias towards the stem. Ratio estimators naturally provide additivity to biomass parts making it possible to improve the precisions of the estimates. In future work with forest biomass and its parts, more canopy variables should be collected to improve the stratification of trees within the diameter stratum. The total volume of the trees should also be estimated, if possible, via xylometry, to approximate the ratio coefficients of the total aboveground biomass, as much as possible, to the actual values of the specific gravity of the wood.

Abstract

In Brazil, studies addressing soil carbon (C) stocks under native vegetation are scarce and often limited to land use change comparisons at the plot level, while studies across larger areas are based on legacy data that address multiple land uses, use wet-oxidations methods to determine C concentrations (Cc) and lack proper soil bulk density (ρb) determinations. Together, these factors may result in either underestimations or biased soil C stocks, thereby hampering the development of adequate land use policy. In this study, we aimed to (i) determine the soil C stocks at different depths (0–10, 10–20, 20–40, 40–60 and 60–100 cm) in Cerrado woodland remnants in Minas Gerais (MG), southeastern Brazil, (ii) study the spatial distribution of soil C stocks using regression-kriging (RK), and (iii) provide a description of the soil C stock distribution and its relationship with variables that drive the soil C stock in the MG Cerrado. We designed a large soil inventory, in which soil samples were collected for soil Cc and ρb determination, to compute soil C stocks in the Cerrado biome of MG. Afterwards, we used RK to model the soil C stocks across the MG Cerrado to a depth of 1 m. Our results show that using RK to map C stocks across different soil layers in the MG Cerrado resulted in accurate estimates. The soil C stocks in the MG Cerrado tend to increase in the west-east direction, are positively correlated with altitude and silt+clay content, and are negatively correlated with the mean annual temperature and total annual precipitation. On average, the C stored at the 0–40 cm layer accounted for 56% of the total C stocked in the top 1m of soil. Approximately 1.03 Pg of C is stored in the MG Cerrado soils at a depth of 1 m. The average soil C stock in the MG Cerrado is approximately 53% larger than the average indicated for the Brazilian Cerrado. Our results highlight the need for more detailed regional scale soil C inventories in the country to refine national C accounting. We believe our study can aid the development of adequate land use policy and could be used as a baseline for land use change studies in the MG Cerrado.

Abstract

Forest ecosystems play an important role in the global carbon cycle and with this there is an increasing need for quantifying carbon at large scales. The aim of this research was to develop a system for classifying tropical forests in Brazil into carbon stock classes, applicable to large areas, emphasizing different sets of stand and climate variables. We used data from forests inventoried in two Brazilian biomes: Atlantic Forest and Savanna. We applied discriminant analysis to generate a classification rule by biome. Three types of variables were used: climatic (mean annual temperature and precipitation, or MAT and MAP), geographical (latitude and longitude), and stand variables (density of trees, mean height or h, mean square diameter or dg, and basal area or G). We combined these into three scenarios for analysis: (1) all variables; (2) all variables, except h; (3) all variables, except h, dg, and G, to determine their contribution to classifying carbon stocks. We also assessed each set of variables in the presence/absence of MAP and MAT, used simultaneously or not. The best classification rules resulted in 83.9% and 98.5% of correct classifications for Atlantic Forest and Savanna biomes, respectively. Stand variables contributed significantly to successful classification; for the Atlantic Forest biome, dg and G contributed from 36% to 42% and h from 2% to 5%, yet for the Savanna biome the gains ranged from 31% to 42% and 6%–9%, respectively. For the climate variables, the simultaneous use of MAT and MAP played an important role in the classification in all cases in the Atlantic Forest biome, contributing up to 9.2% for the classification. In the Savanna biome, we found significant positive gains by the simultaneous use in the absence of h, dg, and G, on the other hand, the simultaneous use exerted negative effects when h was used. We concluded that climate variables are most helpful when stand variables are not included in the analysis. In terms of carbon stock variation, the Atlantic Forest biome tended to be more sensitive to both MAT and MAP, whereas the Savanna biome had no significant climatic dependence in the classification. The variable h exerted a greater effect in the Savanna biome than in the Atlantic Forest, however, basal area and mean square diameter were the most important in both biomes.