Altered growth conditions more than reforestation counteracted forest biomass carbon emissions 1990–2020

Understanding the carbon (C) balance in global forest is key for climate-change mitigation. However, land use and environmental drivers affecting global forest C fluxes remain poorly quantified. Here we show, following a counterfactual modelling approach based on global Forest Resource Assessments, that in 1990–2020 deforestation is the main driver of forest C emissions, partly counteracted by increased forest growth rates under altered conditions: In the hypothetical absence of changes in forest (i) area, (ii) harvest or (iii) burnt area, global forest biomass would reverse from an actual cumulative net C source of c. 0.74 GtC to a net C sink of 26.9, 4.9 and 0.63 GtC, respectively. In contrast, (iv) without growth rate changes, cumulative emissions would be 7.4 GtC, i.e., 10 times higher. Because this sink function may be discontinued in the future due to climate-change, ending deforestation and lowering wood harvest emerge here as key climate-change mitigation strategies.

Forest management required for consistent carbon sink in China’s forest plantations

Background Forest is the largest biomass carbon (C) pool in China, taking up a substantial amount of atmospheric carbon dioxide. Although it is well understood that planted forests (PFs) act as a large C sink, the contribution of human management to C storage enhancement remains obscure. Moreover, existing projections of forest C dynamics suffer from spatially inconsistent age and type information or neglected human management impacts. In this study, using developed PF age and type maps and data collected from 1371 forest plantation sites in China, we simulated biomass C stock change and quantified management impacts for the time period 2010–2050. Results Results show that future forest biomass C increment might have been overestimated by 32.5%–107.5% in former studies. We also found that age-related growth will be by far the largest contributor to PF biomass C increment from 2010 to 2050 (1.23 ± 0.002 Pg C, 1 Pg = 1015 g = 1 billion metric tons), followed by the impact of human management (0.57 ± 0.02 Pg C), while the contribution of climate is slight (0.087 ± 0.04 Pg C). Besides, an additional 0.24 ± 0.07 Pg C can be stored if current PFs are all managed by 2050, resulting in a total increase of 2.13 ± 0.05 Pg C. Conclusions Forest management and age-related growth dominate the biomass C change in PFs, while the effect of climatic factors on the accumulation is minor. To achieve the ambitious goal of forest C stock enhancement by 3.5 Pg from 2020 to 2050, we advocate to improve the management of existing forests and reduce the requests for more lands for forest expansion, which helps mitigate potential conflicts with agricultural sectors. Our results highlight that appropriate planning and management are required for sustaining and enhancing biomass C sequestration in China’s PF.

Historical and future carbon stocks in forests of northern Ontario, Canada

Background Forests in the Far North of Ontario (FNO), Canada, are likely the least studied in North America, and quantifying their current and future carbon (C) stocks is the first step in assessing their potential role in climate change mitigation. Although the FNO forests are unmanaged, the latter task is made more important by growing interest in developing the region’s natural resources, primarily for timber harvesting. In this study, we used a combination of field and remotely sensed observations with a land surface model to estimate forest C stocks in the FNO forests and to project their future dynamics. The specific objective was to simulate historical C stocks for 1901–2014 and future C stocks for 2015–2100 for five shared socioeconomic pathway (SSP) scenarios selected as high priority scenarios for the 6th Assessment Report on Climate Change. Results Carbon stocks in live vegetation in the FNO forests remained relatively stable between 1901 and 2014 while soil organic carbon (SOC) stocks steadily declined, losing about 16% of their initial value. At the end of the historical simulation (in 2014), the stocks were estimated at 19.8, 46.4, and 66.2 tCha−1 in live vegetation, SOC, and total ecosystem pools, respectively. Projections for 2015–2100 indicated effectively no substantial change in SOC stocks, while live vegetation C stocks increased, accelerating their growth in the second half of the twenty-first century. These results were consistent among all simulated SSP scenarios. Consequently, increase in total forest ecosystem C stocks by 2100 ranged from 16.7 to 20.7% of their value in 2015. Simulations with and without wildfires showed the strong effect of fire on forest C stock dynamics during 2015–2100: inclusion of wildfires reduced the live vegetation increase by half while increasing the SOC pool due to higher turnover of vegetation C to SOC. Conclusions Forest ecosystem C stock estimates at the end of historical simulation period were at the lower end but within the range of values reported in the literature for northern boreal forests. These estimates may be treated as conservatively low since the area included in the estimates is poorly studied and some of the forests may be on peat deposits rather than mineral soils. Future C stocks were projected to increase in all simulated SSP scenarios, especially in the second half of the twenty-first century. Thus, during the projected period forest ecosystems of the FNO are likely to act as a C sink. In light of growing interest in developing natural resources in the FNO, collecting more data on the status and dynamics of its forests is needed to verify the above-presented estimates and design management activities that would maintain their projected C sink status.

Recent Carbon Storage and Burial Exceed Historic Rates in the San Juan Bay Estuary Peri-Urban Mangrove Forests (Puerto Rico, United States)

Mangroves sequester significant quantities of organic carbon (C) because of high rates of burial in the soil and storage in biomass. We estimated mangrove forest C storage and accumulation rates in aboveground and belowground components among five sites along an urbanization gradient in the San Juan Bay Estuary, Puerto Rico. Sites included the highly urbanized and clogged Caño Martin Peña in the western half of the estuary, a series of lagoons in the center of the estuary, and a tropical forest reserve (Piñones) in the easternmost part. Radiometrically dated cores were used to determine sediment accretion and soil C storage and burial rates. Measurements of tree dendrometers coupled with allometric equations were used to estimate aboveground biomass. Estuary-wide mangrove forest C storage and accumulation rates were estimated using interpolation methods and coastal vegetation cover data. In recent decades (1970–2016), the highly urbanized Martin Peña East (MPE) site with low flushing had the highest C storage and burial rates among sites. The MPE soil carbon burial rate was over twice as great as global estimates. Mangrove forest C burial rates in recent decades were significantly greater than historic decades (1930–1970) at Caño Martin Peña and Piñones. Although MPE and Piñones had similarly low flushing, the landscape settings (clogged canal vs forest reserve) and urbanization (high vs low) were different. Apparently, not only urbanization, but site-specific flushing patterns, landscape setting, and soil fertility affected soil C storage and burial rates. There was no difference in C burial rates between historic and recent decades at the San José and La Torrecilla lagoons. Mangrove forests had soil C burial rates ranging from 88 g m–2 y–1 at the San José lagoon to 469 g m–2 y–1 at the MPE in recent decades. Watershed anthropogenic CO2 emissions (1.56 million Mg C y–1) far exceeded the annual mangrove forest C storage rates (aboveground biomass plus soils: 17,713 Mg C y–1). A combination of maintaining healthy mangrove forests and reducing anthropogenic emissions might be necessary to mitigate greenhouse gas emissions in urban, tropical areas.

Machine Learning for Prediction of Survival Outcomes with Immune-Checkpoint Inhibitors in Urothelial Cancer

Machine learning (ML) may enhance the efficiency of developing accurate prediction models for survival, which is critical in informing disease prognosis and care planning. This study aimed to develop an ML prediction model for survival outcomes in patients with urothelial cancer-initiating atezolizumab and to compare model performances when built using an expert-selected (curated) versus an all-in list (uncurated) of variables. Gradient-boosted machine (GBM), random forest, Cox-boosted, and penalised, generalised linear models (GLM) were evaluated for predicting overall survival (OS) and progression-free survival (PFS) outcomes. C-statistic (c) was utilised to evaluate model performance. The atezolizumab cohort in IMvigor210 was used for model training, and IMvigor211 was used for external model validation. The curated list consisted of 23 pretreatment factors, while the all-in list consisted of 75. Using the best-performing model, patients were stratified into risk tertiles. Kaplan–Meier analysis was used to estimate survival probabilities. On external validation, the curated list GBM model provided slightly higher OS discrimination (c = 0.71) than that of the random forest (c = 0.70), CoxBoost (c = 0.70), and GLM (c = 0.69) models. All models were equivalent in predicting PFS (c = 0.62). Expansion to the uncurated list was associated with worse OS discrimination (GBM c = 0.70; random forest c = 0.69; CoxBoost c = 0.69, and GLM c = 0.69). In the atezolizumab IMvigor211 cohort, the curated list GBM model discriminated 1-year OS probabilities for the low-, intermediate-, and high-risk groups at 66%, 40%, and 12%, respectively. The ML model discriminated urothelial-cancer patients with distinctly different survival risks, with the GBM applied to a curated list attaining the highest performance. Expansion to an all-in approach may harm model performance.

Enhanced growth – more than reforestation – counteracted biomass carbon emissions (1990-2020)

Understanding the carbon (C) balance in global forest biomass is key for climate change mitigation. However, land-use and environmental drivers affecting forest C fluxes remain poorly isolated and quantified. Following a counterfactual modelling approached based on global Forest Resource Assessments, we show that in the hypothetical absence of changes in forest (i) area, (ii) harvest or (iii) burnt area, global forest biomass would reverse from an actual cumulative net C source of c. 2.4 GtC to a net C sink of 23.6, 2.7 and 5.2 GtC, respectively in 1990–2020. In contrast, cumulative emissions would be 12.9 GtC, i.e. 6 times higher, in the absence of forest growth changes. A typology systematically assessing C flux drivers reveals that enhanced growth, more than reforestation, counteracted C emissions. This sink function may, however, be discontinued in the future, thus alerting to the need for safer strategies to effectively preserve or enhance C sequestration.

Carbon Losses from Decomposing Windrowed Sitka Spruce Woody Debris Over a 16-Year Chronosequence

Meeting the reporting requirements of the Kyoto Protocol has focused attention on the potential of forests in sustainably sequestering carbon (C) to mitigate the effects of rising levels of atmospheric CO2. Much uncertainty remains concerning the ultimate effect of management on such sequestration effects. The management of woody debris (WD) and other deadwood stocks is an example of a management intervention with the scope of affecting the source-sink dynamics of forest C. Windrowing is the most commonly employed approach to the management of post-harvest WD. This study investigated the quantities of windrowed deadwood C across a chronosequence of reforested commercial Sitka spruce stands in Ireland and how its decomposition rate affected its contribution to forest C sequestration. The C stocks in windrowed WD ranged from 25 to 8 t C ha−1 at the 4- and 16-year-old stands, respectively. Losses due to the decomposition of these stocks ranged from 5.15 t C ha−1 yr−1 at the youngest site (4 years old) to 0.68 t C ha−1 yr−1 at the oldest site (16 years old). Using a visual decay-class categorization of WD components and an assessment of wood density, decay rate constants were estimated for logs, branches, and stumps (the main WD constituents of windrows) as 0.037, 0.038, and 0.044, respectively. These results, derived from stand stock evaluations, were placed into context with data previously published from the same chronosequence that characterized the day-to-day fluxes to or from this pool. This comparison indicated that though only a very small quantity of C was lost in dissolved leachate form, the most significant pathway for loss was respiratory and ranged from 16 to 8 t C ha−1 yr−1 at the 9- and 16-year-old sites. These estimates were many times greater in extent than estimates made using a density-loss approach, the difference indicating that fragmentation and weathering play a large role in woody decomposition in intensively managed forests.

Carbon fractions in the world’s dead wood

A key uncertainty in quantifying dead wood carbon (C) stocks—which comprise ~8% of total forest C pools globally—is a lack of accurate dead wood C fractions (CFs) that are employed to convert dead woody biomass into C. Most C estimation protocols utilize a default dead wood CF of 50%, but live tree studies suggest this value is an over-estimate. Here, we compile and analyze a global database of dead wood CFs in trees, showing that dead wood CFs average 48.5% across forests, deviating significantly from 50%, and varying systematically among biomes, taxonomic divisions, tissue types, and decay classes. Utilizing data-driven dead wood CFs in tropical forests alone may correct systematic overestimates in dead wood C stocks of ~3.0 Pg C: an estimate approaching nearly the entire dead wood C pool in the temperate forest biome. We provide for the first time, robust empirical dead wood CFs to inform global forest C estimation.

Cumulative effects of natural and anthropogenic disturbances on the forest carbon balance in the oil sands region of Alberta, Canada; a pilot study (1985–2012)

Background Assessing cumulative effects of anthropogenic and natural disturbances on forest carbon (C) stocks and fluxes, because of their relevance to climate change, is a requirement of environmental impact assessments (EIAs) in Canada. However, tools have not been developed specifically for these purposes, and in particular for the boreal forest of Canada, so current forest C assessments in EIAs take relatively simple approaches. Here, we demonstrate how an existing tool, the Generic Carbon Budget Model (GCBM), developed for national and international forest C reporting, was used for an assessment of the cumulative effects of anthropogenic and natural disturbances to support EIA requirements. We applied the GCBM to approximately 1.3 million ha of upland forest in a pilot study area of the oil sands region of Alberta that has experienced a large number of anthropogenic (forestry, energy sector) and natural (wildfire, insect) disturbances. Results Over the 28 years, 25% of the pilot study area was disturbed. Increasing disturbance emissions, combined with declining net primary productivity and reductions in forest area, changed the study area from a net C sink to a net C source. Forest C stocks changed from 332.2 Mt to 327.5 Mt, declining by 4.7 Mt at an average rate of 0.128 tC ha−1 yr−1. The largest cumulative areas of disturbance were caused by wildfire (139,000 ha), followed by the energy sector (110,000 ha), insects (33,000 ha) and harvesting (31,000 ha) but the largest cumulative disturbance emissions were caused by the energy sector (9.5 Mt C), followed by wildfire (5.5 Mt C), and then harvesting (1.3 Mt C). Conclusion An existing forest C model was used successfully to provide a rigorous regional cumulative assessment of anthropogenic and natural disturbances on forest C, which meets requirements of EIAs in Canada. The assessment showed the relative importance of disturbances on C emissions in the pilot study area, but their relative importance is expected to change in other parts of the oil sands region because of its diversity in disturbance types, patterns and intensity. Future assessments should include peatland C stocks and fluxes, which could be addressed by using the Canadian Model for Peatlands.