Building capacity for estimating fire emissions from tropical peatlands; a worked example from Indonesia.

Tropical peatlands are globally significant in the terrestrial carbon cycle as they are comprised of a large forest carbon sink and a large peat carbon store-both of which can potentially be exchanged with the atmosphere on decadal time frames. Greenhouse gas emissions from fire-disturbance and development of tropical peatlands over the last few decades, and the potential for ongoing emissions, highlights the need for policy to slow or halt emissions and to activate mechanisms to sequester carbon through restoration of degraded peatlands. The UN REDD + scheme provides a means for developing countries to receive payments for avoided deforestation and forest degradation, but the steps to achieve REDD+ compliance are rigorous and the details required can be a barrier to activating benefits-especially for peatlands where repeated cycles of fire interrupt forest recovery and create a range of recovery classes. Therefore, to improve estimates of peat fire emissions and of carbon balance of tropical peatlands, the biomass and combustion factor parameters need to be developed and applied according to forest recovery stage. In this study we use published activity data from the extensive 1997 fires in the peatlands of Indonesian Borneo to detail a transparent and accountable way to estimate and report emissions from tropical peatland fires. This example for estimating and reporting emissions is provided to assist REDD+ countries to efficiently develop their capacity for improving emissions estimates from fire-impacted tropical peatlands.

Tree mortality and carbon emission as a function of wildfire severity in south-eastern Australian temperate forests.

Disturbance trends over recent decades indicate that climate change is resulting in increased fire severity and extent in Australia's temperate Eucalyptus forests. As disturbance cycles become shorter and more severe, empirical measurements are required to identify potential change in forest carbon (C) stock and emissions. However, such estimates are rare in the literature. The 2019-2020 wildfires burnt through 6 to 7 million ha of mainly temperate open Eucalyptus forest in south-east Australia, with top down emission estimates ranging from 97 to 130 tonnes CO2 ha-1. Study sites that had been assessed for all aboveground C pools prior to the wildfires, were burnt in January 2020 by wildfire that varied in severity. Here we quantify the impact of high and low/moderate fire severities on tree mortality, C loss and C redistribution and assess implications for future C storage in these temperate Eucalyptus forests. Higher fire severity resulted in greater overstorey tree mortality but not understorey or loss of dead standing trees than in low/moderate severity fires. High severity fires combusted almost twice as much C from live trees (42 Mg C ha-1) as low/moderate severity fires (25 Mg C ha-1), while C loss from dead standing trees was similar among fire severity classes (average 17 Mg C ha-1). Total aboveground C lost across study sites was 42 Mg C ha-1 for high and 47 Mg C ha-1 for low/moderate severity, with an average of 45 Mg C ha-1 equivalent to 15 % (high severity) and 14 % (low/moderate severity) of AGC. Extrapolating our findings to other tall to medium open Eucalyptus forests across Victoria revealed that 37.33 ± 12.25 Tg C (mean ± s.e.) or 152 ± 50 Mg CO2 ha-1 was lost to the atmosphere from the 0.9 million ha of these productive forests, equating to about 20 % of Australia's total net annual emissions.

Effects of prescribed fire frequency on wildfire emissions and carbon sequestration in a fire adapted ecosystem using a comprehensive carbon model.

Prescribed fire to reduce forest fuels has been routinely applied to reduce wildfire risk in many parts of the world. It has also been proposed that prescribed fire can be used to mitigate greenhouse gas (GHG) emissions. Although prescribed fire creates emissions, if the treatment also decreases the incidence of subsequent wildfires, it is possible for the net outcome to be an emissions decline. Previous studies have suggested prescribed fire, at the frequencies required to materially impact wildfire occurrence, generally leads to net emissions increases. A focus on emissions means any change in carbon storage within the ecosystem remains unaccounted for; because living, dead, and soil carbon pools are characterized by different residence times, a re-distribution of carbon amongst these pools may either reduce or increase long-term ecosystem carbon stores. A full ecosystem carbon model has been developed to investigate the implications of prescribed fire management on total Net Ecosystem Carbon Balance (NECB), inclusive of both emissions and carbon storage. Consistent with previous work, the results suggested limited potential for reducing net GHG emissions through applying prescribed fire, with higher emissions from prescribed fire approximately offset by lower emissions and avoided carbon losses from the subsequent reduction in wildfire frequency. For example, shortening the prescribed fire interval from 25 to 10 years resulted in a NECB sequestration that was typically less than ±0.4 Mg C ha-1 yr-1, or less than approximately 0.1% of the total ecosystem carbon storage. Hence, whilst there was limited opportunity for achieving emission abatement outcomes from changing prescribed fire management, there were no significant emission penalties for doing so. These results suggest land managers should be free to adopt prescribed fire regimes to target specific management outcomes, without significantly impacting net emissions or total ecosystem carbon storage over the long term.

Carbon balance of tropical peat forests at different fire history and implications for carbon emissions.

Accurate assessment of tropical peat forest carbon stocks and impact of fires on carbon pools is required to determine the magnitude of emissions to the atmosphere and to support emissions reduction policies. We assessed total aboveground carbon (AGC) in biomass pools including trees, shrubs, deadwood, litter and char, and peat carbon to develop empirical estimates of peat swamp forest carbon stocks in response to fire and disturbance. In contrast to the common assumption that peat fires combust all AGC, we observed that about half of undisturbed forest AGC, equivalent to about 70 Mg C ha-1, remains after one or two recent fires - mainly in dead trees, woody debris and pyrogenic carbon. Both recently burnt and repeatedly burnt peat forests store similar amounts of carbon in the top 10 cm of peat when compared with undisturbed forests (70 Mg C ha-1), mainly due to increased peat bulk density after fires that compensates for their lower peat C%. The proportion of fuel mass consumed in fire, or combustion factor (CF), is required to make accurate estimates of peat fire emissions for both AGC and peat carbon. This study estimated a CF for AGC (CFAGC) of 0.56, comparable to the default value of the Intergovernmental Panel on Climate Change (IPCC). This study estimated a varying CF for peat (CFPEAT) that ranged from 0.4 to 0.68 as depth of burn increased. This revised CFPEAT is one third to one half of the IPCC default value of 1.0. The current assumption of complete combustion of peat (CF = 1.0) is widely acknowledged in the literature as oversimplification and is not supported by our field observations or data. This study provides novel empirical data to improve estimates of peat forests carbon stocks and emissions from tropical peat fires.

Identifying and addressing knowledge gaps for improving greenhouse gas emissions estimates from tropical peat forest fires.

Tropical peatlands are areas of high carbon density that are important in biosphere-atmosphere interactions. Drainage and burning of tropical peatlands releases about 5% of global greenhouse gas (GHG) emissions, yet there is great uncertainty in these estimates. Our comprehensive literature review of parameters required to calculate GHG emissions from burnt peat forests, following the international guidelines, revealed many gaps in knowledge of carbon pools and few recent supporting studies. To improve future estimates of the total ecosystem carbon balance and peatfire emissions this study aimed to account for all carbon pools: aboveground, deadwood, pyrogenic carbon (PyC) and peat of single and repeatedly burnt peat forests. A further aim was to identify the minimum sampling intensity required to detect with 80% power significant differences in these carbon pools among long unburnt, recently burnt and repeatedly burnt peat swamp forests. About 90 Mg C ha-1 remains aboveground as deadwood after a single fire and half of this remains after a second fire. One fire produces 4.5 ± 0.6 Mg C ha-1 of PyC, with a second fire increasing this to 7.1 ± 0.8 Mg C ha-1. For peat swamp forests these aboveground carbon pools are rarely accounted in estimates of emissions following multiple fires, while PyC has not been included in the total peat carbon mass balance. Peat bulk density and peat carbon content change with fire frequency, yet these parameters often remain constant in the published emission estimates following a single and multiple fires. Our power analysis indicated that as few as 12 plots are required to detect meaningful differences between fire treatments for the major carbon pools. Further field studies directed at improving the parameters for calculating carbon balance of disturbed peat forest ecosystems are required to better constrain peatfire GHG emission estimates.

Effect of thinning and burning fuel reduction treatments on forest carbon and bushfire fuel hazard in Eucalyptus sieberi forests of South-Eastern Australia.

More frequent hot and windy weather in fire prone forested landscapes requires that a full suite of fuel reduction measures be investigated for effectiveness in fuel hazard reduction, environmental impact and carbon (C) outcomes. Although prescribed fire and thinning are routinely applied in forests of North America to reduce fuel loads, there are few detailed studies from Australia. We report the impacts of fuel reduction treatments including burning, mechanical thinning and the combination of both on forest C and fuel hazard in open forests dominated by Eucalyptus sieberi in south-eastern Australia. Carbon losses to the atmosphere and redistribution within the forest were calculated from stocks within each fuel category before and after treatment. Mechanical thinning + burning was the most effective treatment for reducing aboveground C and fuel hazard, with major reductions in dead trees, stumps and understorey, as well as stems removed for sale as pulpwood. However forest floor fuel loads increased in thinned treatments relative to control forests. The overall fuel hazard rating in the burn only treatment was significantly reduced from extreme to low immediately following burning. In thinned only stands, the overall fuel hazard rating did not change from the pre-treatment rating of extreme, due to high surface and forest floor fuel loads and loose and flammable bark on the retained overstorey trees. This result suggests the current fuel hazard guide in use in Australia should be revised to enable it to better describe the benefits of thinning for fuel reduction - in this case the removal of about 50% of aboveground C mostly as overstorey trees, and a significant reduction in understorey, dead trees and stumps.

Aboveground carbon of community-managed Chirpine (Pinus roxburghii Sarg.) forests of Nepal based on stand types and geographic aspects.

On a global scale, about 15.5% of forests are administered through community-based forestry programs that offer the opportunity for enhanced carbon sequestration while maintaining the supply of more traditional goods and services such as cooking fuels, animal fodder and bedding. A challenge in community forest (CF) management is to realize their carbon value without compromising their role in the provision of these traditional goods and services. In this study of CF dominated by Pinus roxburghii in the Phalebas region of Nepal, the impacts of stand composition and geographic aspect on aboveground forest carbon is investigated as a means to optimize CF management for both traditional values and for emerging carbon market values. The aboveground carbon of mixed and monospecific stands of Pinus roxburghii was estimated using a combination of destructive sampling and species-specific allometric equations. On average, monospecific stands contained 106.2 Mg C ha-1 in aboveground tree biomass, significantly more than mixed stands at 73.1 Mg C ha-1 (p = 0.022). Similarly, stands growing on northern aspects (northeast 124.8 Mg C ha-1, northwest 100.9 Mg C ha-1) stored significantly more carbon (p = 0.002) than southern aspects (southeast 75.3 Mg C ha-1, southwest 57.6 Mg C ha-1), reflecting the more favorable growing conditions of northern aspects. These results suggest monospecific stands planted on northern aspects may be best suited for management to achieve carbon benefits, whilst mixed-species stands on southern aspects may be better suited for biodiversity conservation and supporting livelihoods. To maintain and increase carbon value, community forestry may need to implement nutrient return practices to limit the impact of sustained nutrient removals on stand productivity.

Fire intensity effects on post-fire fuel recovery in Eucalyptus open forests of south-eastern Australia.

This is a study of the re-accumulation of bushfire fuels following both prescribed fire of low fireline intensity (<700 kW m-1) and wildfire of high intensity (>10,000 kW m-1) in Australian Eucalyptus open forests of differing annual rainfall. Repeated measurements over 5 to 7 years of litter, elevated fuels, coarse woody debris, and bark revealed more rapid fuel recovery in higher rainfall forests compared with lower rainfall forests, following prescribed fire. In prescribed-burnt forests with mean annual rainfall 900-950 mm all fuel categories recovered to very high within seven years, with elevated fuels exceeding pre-fire loads by up to 200%. No fuels in prescribed-burnt forests with mean annual rainfall 600-650 mm recovered to pre-fire loads after six years suggesting that rainfall is an important driver of the rate of fuels recovery. High intensity wildfire in lower rainfall forests (600-650 mm) stimulated the rapid recovery of elevated fuels to over 600% of pre-fire loads - effectively transforming open forest formations into shrublands over the 6 years after fire. The recovery of elevated fuels following both prescribed fire in high rainfall forests and wildfire in low rainfall forests did not follow a gradual negative exponential increase often approximated by an Olson curve, but peaked early after fires. This suggests that the Olson recovery function, the default for predicting loads for these fuels in the operational fire behaviour models in use in south-eastern Australia, may not be appropriate in all cases. Fire simulations were run for forests burnt in wildfires using default (forest) and observed (shrubland) vegetation types. Under weather conditions similar to the previous wildfire, predictions for fireline intensities and the rate of spread would be at least 50% greater in transitional shrubland than forest, emphasizing the importance of accounting for vegetation dynamics for safe response management.

A data - Model fusion methodology for mapping bushfire fuels for smoke emissions forecasting in forested landscapes of south-eastern Australia.

The increasing regional and global impact of wildfires on the environment, and particularly on the human population, is becoming a focus of the research community. Both fire behaviour and smoke dispersion models are now underpinning strategic and tactical fire management by many government agencies and therefore model accuracy at regional and local scales is increasingly important. This demands accuracy of all the components of the model systems, biomass fuel loads being among the more significant. Validation of spatial fuels maps at a regional scale is uncommon; in part due to the limited availability of independent observations of fuel loads, and in part due to a focus on the impact of model outputs. In this study we evaluate two approaches for estimating fuel loads at a regional scale and test their accuracy against an extensive set of field observations for the State of Victoria, Australia. The first approach, which assumes that fuel accumulation is an attribute of the vegetation class, was developed for the fire behaviour model Phoenix Rapid-Fire, with apparent success; the second approach applies the Community Atmosphere Biosphere Land Exchange (CABLE) process-based terrestrial biosphere model, implemented at high resolution across the Australian continent. We show that while neither model is accurate over the full range of fine and coarse fuel loads, CABLE biases can be corrected for the full regional domain with a single linear correction, however the classification based Phoenix requires a matrix of factors to correct its bias. We conclude that these examples illustrate that the benefits of simplicity and resolution inherent in classification-based models do not compensate for their lack of accuracy, and that lower resolution but inherently more accurate carbon-cycle models may be preferable for estimating fuel loads for input into smoke dispersion models.

Importance of disturbance history on net primary productivity in the world's most productive forests and implications for the global carbon cycle.

Analysis of growth and biomass turnover in natural forests of Eucalyptus regnans, the world's tallest angiosperm, reveals it is also the world's most productive forest type, with fire disturbance an important mediator of net primary productivity (NPP). A comprehensive empirical database was used to calculate the averaged temporal pattern of NPP from regeneration to 250 years age. NPP peaks at 23.1 ± 3.8 (95% interquantile range) Mg C ha-1  year-1 at age 14 years, and declines gradually to about 9.2 ± 0.8 Mg C ha-1  year-1 at 130 years, with an average NPP over 250 years of 11.4 ± 1.1 Mg C ha-1  year-1 , a value similar to the most productive temperate and tropical forests around the world. We then applied the age-class distribution of E. regnans resulting from relatively recent historical fires to estimate current NPP for the forest estate. Values of NPP were 40% higher (13 Mg C ha-1  year-1 ) than if forests were assumed to be at maturity (9.2 Mg C ha-1  year-1 ). The empirically derived NPP time series for the E. regnans estate was then compared against predictions from 21 global circulation models, showing that none of them had the capacity to simulate a post-disturbance peak in NPP, as found in E. regnans. The potential importance of disturbance impacts on NPP was further tested by applying a similar approach to the temperate forests of conterminous United States and of China. Allowing for the effects of disturbance, NPP summed across both regions was on average 11% (or 194 Tg C/year) greater than if all forests were assumed to be in a mature state. The results illustrate the importance of accounting for past disturbance history and growth stage when estimating forest primary productivity, with implications for carbon balance modelling at local to global scales.

Edge type affects leaf-level water relations and estimated transpiration of Eucalyptus arenacea.

While edge effects on tree water relations are well described for closed forests, they remain under-examined in more open forest types. Similarly, there has been minimal evaluation of the effects of contrasting land uses on the water relations of open forest types in highly fragmented landscapes. We examined edge effects on the water relations and gas exchange of a dominant tree (Eucalyptus arenacea Marginson & Ladiges) in an open forest type (temperate woodland) of south-eastern Australia. Edge effects in replicate woodlands adjoined by cleared agricultural land (pasture edges) were compared with those adjoined by 7- to 9-year-old eucalypt plantation with a 25m fire break (plantation edges). Consistent with studies in closed forest types, edge effects were pronounced at pasture edges where photosynthesis, transpiration and stomatal conductance were greater for edge trees than interior trees (75m into woodlands), and were related to greater light availability and significantly higher branch water potentials at woodland edges than interiors. Nonetheless, gas exchange values were only ∼50% greater for edge than interior trees, compared with ∼200% previously found in closed forest types. In contrast to woodlands adjoined by pasture, gas exchange in winter was significantly lower for edge than interior trees in woodlands adjoined by plantations, consistent with shading and buffering effects of plantations on edge microclimate. Plantation edge effects were less pronounced in summer, although higher water use efficiency of edge than interior woodland trees indicated possible competition for water between plantation trees and woodland edge trees in the drier months (an effect that might have been more pronounced were there no firebreak between the two land uses). Scaling up of leaf-level water relations to stand transpiration using a Jarvis-type phenomenological model indicated similar differences between edge types. That is, transpiration was greater at pasture than plantation edges in summer months (most likely due to greater water availability at pasture edges), resulting in significantly greater estimates of annual transpiration at pasture than plantation edges (430 vs. 343lm(-2)year(-1), respectively). Our study highlights the need for landscape-level water flux models to account for edge effects on stand transpiration, particularly in highly fragmented landscapes.

Interactive effects of high irradiance and moderate heat on photosynthesis, pigments, and tocopherol in the tree-fern Dicksonia antarctica.

Effects of high irradiance and moderate heat on photosynthesis of the tree-fern Dicksonia antarctica (Labill., Dicksoniaceae) were examined in a climate chamber under two contrasting irradiance regimes (900 and 170 µmol photons m-2 s-1) and three sequential temperature treatments (15°C; 35°C; back to 15°C). High irradiance led to decline in predawn quantum yield of photochemistry, Fv/Fm (0.73), maximal Rubisco activity (Vcmax; from 37 to 29 µmol m-2s-1), and electron transport capacity (Jmax; from 115 to 67 µmol m-2 s-1). Temperature increase to 35°C resulted in further decreases in Fv/Fm (0.45) and in chlorophyll bleaching of high irradiance plants, while Vcmax and Jmax were not affected. Critical temperature for thylakoid stability (Tc) of D. antarctica was comparable with other higher plants (c. 47°C), and increases of Tc with air temperature were greater in high irradiance plants. Increased Tc was not associated with accumulation of osmotica or zeaxanthin formation. High irradiance increased the xanthophyll cycle pigment pool (V+A+Z, 91 v. 48 mmol mol-1 chlorophyll-1), de-epoxidation state (56% v. 4%), and α-tocopherol. Temperature increase to 35°C had no effect on V+A+Z and de-epoxidation state in both light regimes, while lutein, ß-carotene and α-tocopherols increased, potentially contributing to increased membrane stability under high irradiance.