Post-Fire Carbon Dynamics in a UK Woodland: A Case Study from the Roaches Nature Reserve

Abstract

Forests play a crucial role in climate regulation through atmospheric CO₂ sequestration. However, disturbances like wildfires can severely compromise this function. This study assesses the ecological and economic consequences of a 2018 wildfire in The Roaches Nature Reserve, UK, focusing on post-fire carbon dynamics. A mixed woodland dominated by Pinus sylvestris L. and Larix decidua Mill. was evaluated via satellite imagery (remote sensing indices), dendrochronological analysis (wood cores sampling), and soil properties analyses. Remote sensing revealed areas of high fire severity and progressive vegetation decline. Tree-ring data indicated near-total mortality of L. decidua, while P. sylvestris showed greater post-fire resilience. Soil properties (e.g., soil organic carbon, biomass and microbial indices, etc.) assessed at a depth of 0–5 cm showed no significant changes. The analysis of CO₂ sequestration trends revealed a marked decline in burned areas, with post-fire sequestration reduced by approximately 70% in P. sylvestris and nearly 100% in L. decidua, in contrast to the stable patterns observed in the control stands during the same period. To estimate this important ecosystem service, we developed a novel CO₂ Sequestration Loss (CSL) index, which quantified the reduction in forest carbon uptake and underscored the impaired sequestration capacity of burned area. The decrease in CO₂ sequestration also resulted in a loss of regulating ecosystem service value, with burned areas showing a marked reduction compared to pre-fire conditions. Finally, a carbon loss of ~208 Mg ha⁻¹ was estimated in the burnt area compared to the control, mainly due to tree mortality rather than shallow soil carbon stock. Overall, our findings demonstrate that wildfire can substantially compromise the climate mitigation potential of temperate forests, highlighting the urgency of proactive management and restoration strategies.

1. Introduction

Forests are among the ecosystems most affected by climate change [1]. The increased frequency and intensity of extreme climatic events is reducing the resilience of these ecosystems, leading to widespread tree mortality [2,3,4,5,6]. Between 1990 and 2020, the global forest area decreased by around 5%, which has serious repercussions for mitigating global warming [7]. Forests act as carbon sinks by absorbing atmospheric CO₂ through the process of photosynthesis [8,9]. It is estimated that forests absorb a large amount of carbon dioxide each year, equivalent to nearly one-third of annual anthropogenic emissions [10]. This ecosystem service is not only ecologically relevant but also has significant economic value [11]. Several studies have estimated that annual carbon sequestration provides substantial global economic benefits, primarily through the avoidance of costs associated with the impacts of climate change caused by human activities [12]. For instance, within the European Union the annual value of forest carbon sequestration has been estimated at about €11.9 billion per year (≈0.08% of EU GDP) [13].

Forest carbon stock is primarily distributed between living biomass and soil, which are the most stable carbon reserves in forest ecosystems [14]. Soil represents the first terrestrial carbon reservoir [15] and the planet’s second largest carbon sink after the oceans [16,17]. However, ecological disturbances such as forest fires represent one of the most serious threats to the forest’s health [18,19]. As well as rapidly destroying plant biomass and reducing the ecosystem’s capacity to absorb CO₂, fires can also release large amounts of stored carbon from vegetation and soil in a very short time [20,21,22]. Notably, this contributes to increased atmospheric emissions and influence the global carbon balance [20,23,24].

In recent years, wildfires have increased in many regions of the world [25,26]. According to data provided by the European Forest Fire Information System (EFFIS), in 2024 alone, fires affected over 400,000 hectares of vegetated land in Europe, which is higher than the average recorded in the previous year [27]. Alarmingly, this phenomenon is spreading even to countries previously less prone to fires [28], including the United Kingdom [29,30,31]. Here, the wildfire risk is rising mainly due to anthropogenic activity and the climate change-induced effects, such as unprecedented heatwaves [32]. Although peatlands are most at risk in the UK’s humid temperate environment, woodlands are not immune, as demonstrated by recent forest fires, such as the Wareham and Swinley fires [31,33]. The UK Forestry Commission statistics indicated an annual average burnt area of around 6600 ha between 2009 and 2021 [34]. In the first few months of 2025 alone, the EFFIS detected over 150 large fires (>30 hectares), confirming an upward trend [27]. Currently, forests cover around 13% of the UK’s land area, significantly lower than the European average. This limited extent makes the UK’s forest heritage particularly vulnerable to environmental disturbances [34]. Even the loss of a few woodland areas can substantially reduce the capacity of these ecosystems to function as carbon sinks [35]. Furthermore, a reduction in tree cover can trigger serious cascading effects, such as the loss of local biodiversity or increased soil erosion, which contributes to the further reduction in the ecological resilience of UK forests [36,37,38]. In addition to these ecological consequences, reduced CO₂ sequestration capacity also translates into an economic loss. According to the UK Office for National Statistics (ONS), in 2021 woodlands removed 19.6 million tons of CO₂ equivalent, worth about €5.9 billion. Overall, UK woodland ecosystem services were valued at ~€11.6 billion annually, over twenty times the market value of timber and wood-fuel (€0.51 billion) [39].

In this context, the present study was conducted in a mixed woodland of Larix decidua Mill. (European larch) and Pinus sylvestris L. (Scots pine) located in central England affected by a wildfire in 2018. The research objective was to assess the impact of fire on the woodland’s carbon stock capacity and atmospheric CO₂ sequestration, by comparing a burned area with a nearby undisturbed one. The study also included an assessment of the fire’s economic impact, estimating the loss of the ecosystem service of CO₂ sequestration by trees. Consequently, it has been hypothesized that the wildfire would have caused a marked reduction in tree CO₂ fixation capacity and a parallel loss of carbon stored in the soil, thereby compromising the overall carbon sink function of the woodland. Finally, we translate this into projected economic losses, due to the decline in the forest’s capacity to absorb atmospheric CO₂. Overall, the outcomes of this study aim to contribute to a better understanding of the ecological and economic impacts of forest fires on temperate forests, particularly in the context of the UK, where such events are rare and available information remains limited [29,40].

2. Materials and Methods

2.1. Study Area

The study was carried out in a forested area of central England, located within The Roaches Nature Reserve (UTM 30U 566,962.40 m E, 5,890,975.00 m N) (Figure 1A). This area is part of the wider Peak District National Park and has also been designated a Special Area of Conservation (SAC—UK0030280) (Figure 1B). The study area includes two ecologically distinct habitats, separated by a rocky ridge. On one side, there is a mixed woodland dominated by Pinus sylvestris L. (Scots pine) and Larix decidua Mill. (European larch). On the other side, there is an open peatland, mainly populated by species from the Ericaceae family [41,42]. The climate in this area is typical of a temperate, humid region, with abundant rainfall throughout the year and moderate temperatures, without a clearly defined dry season. According to climate data from the ERA5-Land database for the period 1950–2023, the area’s average annual rainfall is approximately 990 mm, while the average annual temperature is around 8.4 °C. In 2018, a wildfire ignited by a barbecue within the woodland burned approximately 60 hectares of vegetation, affecting also the surrounding peatland.

A recent study [43] reported that the fire directly damaged 2.7 hectares of woodland (burned area), leaving 16.5 hectares unaffected (control area). The same study also highlighted a widespread mortality in L. decidua trees most likely linked to the wildfire effect coupled with an outbreak of Phytophthora ramorum Werres [44]. At the end of 2023, the Forestry England Commission began selectively cutting down infected trees in cooperation with the Staffordshire Wildlife Trust (SWT), which manages the Reserve, in an attempt to contain the spread of the pathogen [45].

2.2. Satellite Analysis

To assess the spatial extent and severity of the fire disturbance over time, this study employed the Normalized Burn Ratio (NBR) and its derivative index, the differenced Normalized Burn Ratio (dNBR). These indices allow for the identification of burned areas and the quantification of vegetation loss by comparing spectral reflectance before and after the fire event [46,47]. In particular, high NBR values indicate healthy vegetation, while low values indicate bare soil and areas affected by disturbance. Consequently, it can be assumed that the dNBR has higher values under conditions of reduced vegetation biomass/vitality. The burn severity classification used refers to the USGS ranges [48]. Sentinel-2A imagery were sourced from the Copernicus Data Space Ecosystem [49] for the same period to reduce phenological variation. NBR and dNBR were computed according to Equations (1) and (2), respectively.

$$\mathrm{N}\mathrm{B}\mathrm{R}={\displaystyle \frac{\mathrm{N}\mathrm{I}\mathrm{R}-\mathrm{S}\mathrm{W}\mathrm{I}\mathrm{R}2}{\mathrm{N}\mathrm{I}\mathrm{R}+\mathrm{S}\mathrm{W}\mathrm{I}\mathrm{R}2}}$$

(1)

where NIR = Band 8 (0.84–0.88 µm) and SWIR2 = Band 12 (2.11–2.29 µm).

$$\mathrm{d}\mathrm{N}\mathrm{B}\mathrm{R}={\mathrm{N}\mathrm{B}\mathrm{R}}_{\mathrm{p}\mathrm{r}\mathrm{e}\mathrm{f}\mathrm{i}\mathrm{r}\mathrm{e}}-{\mathrm{N}\mathrm{B}\mathrm{R}}_{\mathrm{p}\mathrm{o}\mathrm{s}\mathrm{t}\mathrm{f}\mathrm{i}\mathrm{r}\mathrm{e}}$$

(2)

where the difference from pre- and post-fire NBR (pre-fire NBR – post-fire NBR) is the satellite remote sensing bitemporal spectral index delta normalized burn ratio (dNBR).

Indices computation and map drawing were output via the open-source software QGIS Desktop 3.40.8. To provide geographical context, the “Bing VirtualEarth” basemap available through the “Basemap Manager” plugin was used as a background layer for maps export.

2.3. Sampling Protocol

Soil and wood samples were collected as described by Niccoli et al. [43]. Briefly, in September 2023, three circular experimental plots, each with a radius of 13 m and a total area of 539 m², were selected within burned and control areas, remaining within the same physiographic features and avoiding edge effects.

A forest survey was conducted in each plot, including a census of living and dead trees, to determine trees density before and after the fire. The diameter at breast height (DBH) and height of each tree were also measured. For dendrochronological analysis, 10 dominant trees were selected in each area (burned and control), and for each species. Wood cores were collected using an increment borer (Haglöfs, Långsele, Sweden), following standardized procedures for dendrochronology [50].

Within each experimental plot, soil samples were randomly collected, in triple field replicates, at a depth of 5 cm in the four corners and in the center of a 0.5 m² square. Then samples were mixed to create a composite soil sample. An undisturbed soil core for each replicate was also collected for bulk density determination.

2.4. Backdating of Tree Growth and Trees Carbon Dynamics

The dendrochronological approach used in this study allowed for reconstructing the atmospheric carbon uptake capacity of the two species over time.

Wood cores were mounted on dedicated supports and then sanded to clearly identify the tree rings. The samples were digitized using a high-resolution Epson Perfection 12000XL Pro scanner (Epson Corporation, Suwa, Japan) and the ring widths were measured using Windendro software (Version 2025a, Regent Instruments, Montreal, Canada).

These measures allowed the computation of annual growth chronologies for each tree. After a rigorous statistical analysis based on cross-dating [51], the chronologies were used to estimate the historical diameter at breast height (DBH) of the sampled trees. The DBH for each year was calculated by progressively subtracting the sum of ring widths from the DBH measured at the end of 2023, according to Equation (3).

$${\mathrm{D}\mathrm{B}\mathrm{H}}_t={\mathrm{D}\mathrm{B}\mathrm{H}}_{2023}-2\sum _{i=t+1}^{2023}wi$$

(3)

where DBH_t is the diameter at breast height in year t, DBH₂₀₂₃ is the diameter measured in field, and wi represents the tree ring width in year i. The factor 2 accounts for growth on both sides of the trunk.

Tree height in year t was estimated from the corresponding DBH_t using an asymptotic allometric equation commonly applied to European conifers [52,53,54].

$${\mathrm{T}\mathrm{r}\mathrm{e}\mathrm{e} \mathrm{h}\mathrm{e}\mathrm{i}\mathrm{g}\mathrm{h}\mathrm{t}}_t=1.3+a\cdot \left(1-{\mathrm{e}}^{\left\{-b\cdot {\mathrm{D}\mathrm{B}\mathrm{H}}_t\right\}}\right)$$

(4)

The parameters of the equation were defined as follows: (i) parameter a, representing the maximum asymptotic height, was calibrated individually for each sampled tree so that the fitted curve reproduced the actual tree height measured in the field at the end of 2023; and (ii) parameter b, which controls the steepness of the curve and the rate of height increase with DBH, was calibrated assigning the mean values reported in the literature for temperate European conifers growing in similar environmental contexts, providing consistent estimates for the studied species [55,56,57]. This procedure ensured that the model was anchored to local empirical data (through parameter a) while remaining consistent with generalized allometric relationships for the studied species (through parameter b).

The aboveground biomass (AGB) for each year t was estimated following the methodological framework of the IPCC Guidelines for National Greenhouse Gas Inventories [58] and the Winrock International manual for aboveground biomass estimation [59] according to Equation (5):

$${\mathrm{A}\mathrm{G}\mathrm{B}}_t\left(\mathrm{M}\mathrm{g}\right)=\mathrm{V} \mathrm{s}\mathrm{t}\mathrm{e}\mathrm{m}\times \mathrm{W}\mathrm{D}\times \mathrm{B}\mathrm{C}\mathrm{E}\mathrm{F}$$

(5)

where V stem is the stem volume derived from tree height and DBH, WD is the tree species wood density (Mg m⁻³), and BCEF is the factor converting stem biomass into total aboveground biomass. The adopted BCEFs for Pinus sylvestris and Larix decidua were derived from [60], while species-specific wood density values were obtained from [61].

Then, AGB_t was converted to yearly tree carbon stock using a carbon fraction value of 0.51, as suggested by the IPCC guidelines (Equation (6)) [58]. Finally, carbon stock was converted to the yearly CO₂ equivalent by multiplying by 3.67, the ratio of the molecular weight of CO₂ (44) to that of carbon (12) (Equation (7)) [62]. Both tree carbon stock and CO₂ stored were expressed per hectare, considering the pre- and post-fire tree density measured by the previous study conducted in the area [43].

$$\mathrm{T}\mathrm{r}\mathrm{e}\mathrm{e} {\mathrm{C}}_{\mathrm{s}\mathrm{t}\mathrm{o}\mathrm{c}\mathrm{k}t}\left({\displaystyle \frac{\mathrm{M}\mathrm{g}}{\mathrm{h}\mathrm{a}}}\right)=\mathrm{A}\mathrm{G}\mathrm{B}\times 0.51$$

(6)

$${\mathrm{C}\mathrm{O}}_2 t\left({\displaystyle \frac{\mathrm{M}\mathrm{g}}{\mathrm{h}\mathrm{a}}}\right)=\mathrm{T}\mathrm{r}\mathrm{e}\mathrm{e} {\mathrm{C}}_{\mathrm{s}\mathrm{t}\mathrm{o}\mathrm{c}\mathrm{k}}\left({\displaystyle \frac{\mathrm{M}\mathrm{g}}{\mathrm{h}\mathrm{a}}}\right)\times 3.67$$

(7)

The annual CO₂ fixed per hectare by each tree species in each area was calculated as the difference between the CO₂ stored at the end of the current growing season (year t) and that of the previous season (year t − 1), as shown in Equation (8) [21].

$${\mathrm{C}\mathrm{O}}_2{\mathrm{f}\mathrm{i}\mathrm{x}\mathrm{e}\mathrm{d}}_t={\mathrm{C}\mathrm{O}}_2{\mathrm{s}\mathrm{t}\mathrm{o}\mathrm{r}\mathrm{e}\mathrm{d}}_t-{\mathrm{C}\mathrm{O}}_2{\mathrm{s}\mathrm{t}\mathrm{o}\mathrm{r}\mathrm{e}\mathrm{d}}_{t-1}$$

(8)

Box plots were used to compare the annual CO₂ fixed during the pre-fire (2012–2017) and post-fire (2018–2023) periods by the two tree species across the control and burned area. Statistical significance between groups was assessed using an independent t-test (α = 0.05).

Finally, to evaluate the economic repercussions of the wildfire on CO₂ sequestration, the monetary value of the CO₂ fixed in each area during the same periods was estimated [21]. This calculation was based on the amount of CO₂ fixed, multiplied by the average 2023 CO₂ price (75 €/tCO₂), as reported by the European Commission in its assessment of the carbon market [63].

2.5. Ecosystem CO₂ Sequestration Loss

To assess the functional impact of the wildfire, a novel CO₂ Sequestration Loss (CSL) index has been employed, designed to quantify the relative reduction in this ecosystem service. The index is based on ratio between the annual sum (year t) of CO₂ adsorbed by the two species in each area (Equation (9)). The calculation included the year preceding the fire (2017) and the post-fire period (2018–2023). The control stand was chosen as the reference because it represents the best local performance of CO₂ sequestration in the absence of disturbance, thus allowing us to express the functional loss in the burned area as a relative measure against this baseline.

$${\mathrm{C}\mathrm{S}\mathrm{L}}_t=1-{\displaystyle \frac{{({\mathrm{C}\mathrm{O}}_2 \mathrm{f}\mathrm{i}\mathrm{x}\mathrm{e}\mathrm{d} \mathrm{i}\mathrm{n} \mathrm{b}\mathrm{u}\mathrm{r}\mathrm{n}\mathrm{e}\mathrm{d} \mathrm{a}\mathrm{r}\mathrm{e}\mathrm{a})}_t}{{({\mathrm{C}\mathrm{O}}_2 \mathrm{f}\mathrm{i}\mathrm{x}\mathrm{e}\mathrm{d} \mathrm{i}\mathrm{n} \mathrm{c}\mathrm{o}\mathrm{n}\mathrm{t}\mathrm{r}\mathrm{o}\mathrm{l} \mathrm{a}\mathrm{r}\mathrm{e}\mathrm{a})}_t}}$$

(9)

Higher values of CLS index indicate a greater loss of CO₂ sequestration capacity, while lower values indicate a recovery of sequestration capacity.

2.6. Soil Processing

For soil sampling methodologies see [43].

Here, the soil organic carbon stock was calculated according to Equation (10) [64,65,66]:

$${\mathrm{S}\mathrm{o}\mathrm{i}\mathrm{l} \mathrm{C}}_{\mathrm{s}\mathrm{t}\mathrm{o}\mathrm{c}\mathrm{k}}\left({\displaystyle \frac{\mathrm{M}\mathrm{g}}{\mathrm{h}\mathrm{a}}}\right)={\%\mathrm{C}}_{\mathrm{o}\mathrm{r}\mathrm{g}}\times \mathrm{B}\mathrm{D}\left({\displaystyle \frac{\mathrm{g}}{{\mathrm{c}\mathrm{m}}^3}}\right)\times \mathrm{D}(\mathrm{c}\mathrm{m})$$

(10)

where %C_org is the weight percentage of soil organic carbon, BD is the bulk density and D is the depth of the soil’s samples.

Furthermore, two microbial indices, microbial carbon and nitrogen ratio (C_mic/N_mic) [67,68] and microbial quotient (C_mic%C_org) [66], were calculated by microbial biomass (C_mic and N_mic) and organic carbon content (C_org) values (Table S1).

Descriptive statistics were computed, and non-normal variables were log-transformed [69]. Independent t-tests (α = 0.05) assessed significant differences between burned and control plots.

2.7. Tree–Soil Carbon Stock (2023)

The tree–soil carbon stock for 2023 was estimated for both the control and burned areas by summing the carbon stored in the living above-biomass of the two tree species in 2023 (computed as the cumulative annual Tree C stock) and the shallow soil organic carbon content (0–5 cm) [70], according to the following Equation (11):

$$\mathrm{T}\mathrm{r}\mathrm{e}\mathrm{e}-\mathrm{s}\mathrm{o}\mathrm{i}\mathrm{l} {\mathrm{C}}_{\mathrm{s}\mathrm{t}\mathrm{o}\mathrm{c}\mathrm{k}} \left({\displaystyle \frac{\mathrm{M}\mathrm{g}}{\mathrm{h}\mathrm{a}}}\right)=\mathrm{T}\mathrm{r}\mathrm{e}\mathrm{e} {\mathrm{C}}_{\mathrm{s}\mathrm{t}\mathrm{o}\mathrm{c}\mathrm{k}} \left({\displaystyle \frac{\mathrm{M}\mathrm{g}}{\mathrm{h}\mathrm{a}}}\right)+\mathrm{S}\mathrm{o}\mathrm{i}\mathrm{l} {\mathrm{C}}_{\mathrm{s}\mathrm{t}\mathrm{o}\mathrm{c}\mathrm{k}}\left({\displaystyle \frac{\mathrm{M}\mathrm{g}}{\mathrm{h}\mathrm{a}}}\right)$$

(11)

An independent t-test (α = 0.05) was then used to statistically compare the results of the two areas. Although the calculation excluded dead non-functional biomass and aboveground herbaceous vegetation, it provides a representative estimate of the ecosystem’s carbon pool per hectare five years after the fire.

3. Results

3.1. Spatial Analysis of the Fire Impact

The spatial impact of the fire was assessed using satellite analysis of the woodland for the years 2018 (fire event) and 2023 (year of the sampling). The NBR and dNBR maps provided complementary information on post-fire dynamics: NBR describing the state of woodland cover, and dNBR quantifying the degree of accumulated disturbance.

The NBR index (Figure 2A) highlighted a wider impacted area in 2023 compared to 2018. The dNBR index (Figure 2B) allowed classification of disturbance severity, highlighting differences between pre- and post-fire conditions. Five years after the event, the 2023 dNBR map shows an expansion of areas classified as high severity (from yellow to red pixels).

3.2. CO₂ Uptake Dynamics and Ecosystem Service Costs

3.2.1. Pinus sylvestris L.

P. sylvestris trees in The Roaches Nature Reserve exhibited a consistent capacity to absorb CO₂ over time across all analyzed areas (Figure 3A). An overall increase in assimilation was observed from the 1980s onwards, which may be related to climate trends, stand development, or past management activities. Although trees in the control area showed slightly higher values, both groups followed similar trends until the year of the fire. From 2018 onwards, however, trees experienced a marked decline in CO₂ assimilation, dropping from 4.1 Mg/ha of CO₂ in 2017 to just 1.2 Mg/ha of CO₂ by 2023. In contrast, control trees maintained a stable assimilation rate, with values of 4.5 Mg/ha of CO₂ in both 2017 and 2023.

Statistical analyses (Figure 3B) confirm that CO₂ sequestration values were comparable between burned and control areas during the pre-fire period (2012–2017) (p > 0.05). However, from 2018, a significant difference in sequestration capacity was observed (p < 0.001).

The economic valuation of the CO₂ absorption ecosystem service, estimated as the avoided cost of CO₂ emissions in euros per hectare, is presented in Figure 4. During the pre-fire period (2012–2017), the ecosystem service value was similar across both areas, ranging between €250 and €380 per hectare. In the post-fire period (2018–2023), however, a clear contrast was observed: in the control, the estimated value ranged from €260 to €320 per ha, while in the burned area it declined substantially, reflecting the impaired CO₂ sequestration capacity between 100 and 200 euro/ha.

3.2.2. Larix decidua Mill.

L. decidua trees showed closely aligned trends for CO₂ fixation capacity over time in both control and burned area until the fire event (Figure 5A). Also in this case, from the 1980s onwards, an overall upward trend in CO₂ assimilation was evident until the fire event. Statistical analysis (Figure 5B) confirmed no significant differences during the pre-fire period (2012–2017; p > 0.05).

After the fire (2018–2023), the difference became highly significant (p < 0.001), with contrasting patterns: carbon dioxide uptake in control trees slightly declined from 10.9 Mg/ha of CO₂ in 2017 to 7.8 Mg/ha of CO₂ in 2023, whereas burned trees experienced a dramatic drop from 11.5 Mg/ha of CO₂ in 2017 to nearly zero by 2023.

This had a direct impact on the economic value of the CO₂ sequestration ecosystem service provided by this species (Figure 6B). In the pre-fire period (2012–2017), the associated economic values were comparable between the two areas, ranging between €700 and €980 per ha, with the burned area occasionally showing slightly higher values. In the post-fire period (2018–2023), while trees in the control area maintained relatively stable values (in average €500/ha annually), in the burned area, the values dropped to nearly zero, reflecting the extensive tree mortality caused by the fire.

3.3. Post-Fire CO₂ and Tree–Soil Carbon Stock Balance

The CO₂ Sequestration Loss (CSL) index, based on the annual ratio between the cumulative CO₂ stocked of burned and control trees for both species, allowed for the quantification of the loss of this ecosystem service across The Roaches Nature Reserve woodland. As shown in Figure 7, in the year prior to the fire (2017), the index recorded values close to zero (CSL = 0.002), indicating an essentially equivalent CO₂ uptake capacity between the two areas. However, starting from 2018 there was a marked increase in the index (CSL = 0.80), reflecting a significant loss of sequestration capacity in the burned area. Following a slight decline in 2020 (CSL = 0.70), the index rose again, reaching a peak in 2023 (CSL = 0.89).

By summing the carbon stored in the surviving trees of both species with that measured in the soil, the overall carbon balance for 2023 was estimated at each area (Figure 8A). In the control area, the tree–soil carbon stock was 317 Mg/ha, whereas in the burned area this value dropped to just 109 Mg/ha, corresponding to a loss of over 66%.

The distribution of carbon among compartments (Figure 8B) indicated that this reduction was primarily attributable to the mortality of L. decidua trees. Meanwhile, soil carbon stock remained essentially unchanged, and no statistically significant differences were detected between the burned and control areas.

4. Discussion

The multidisciplinary approach adopted in this study allowed for an integrated assessment of the 2018 wildfire’s impact on the woodland of The Roaches Nature Reserve, analyzing both the ecological implications in terms of CO₂ sequestration and the economic consequences associated with the loss of this critical ecosystem service.

Remote sensing analysis provided essential information for quantifying the spatial extension and severity of the fire in the medium term [21,71]. In line with methodologies applied in previous studies, the comparison of NBR and dNBR indices between 2018 and 2023 revealed a progressive decline in vegetation cover indicative of persistent indirect effects [42,72]. These results suggested not only a limited capacity for post-fire recovery, but also a progressive intensification of fire-related impacts over time, driven by delayed defoliation, limited natural regeneration, and increasing tree mortality rates [6,73,74,75]. Such processes have significant implications for the structural integrity and functional performance of the forest ecosystem, particularly by undermining its capacity to assimilate atmospheric CO₂ and to store carbon in living biomass [76].

A deeper understanding of these ecosystem-level dynamics can be achieved by examining species-specific responses to post-fire stress [77,78]. According to the previous study conducted in the area [43], L. decidua exhibited a high post-fire mortality rate, reaching 97% by 2023, whereas P. sylvestris showed a more moderate mortality (39%). This contrasting response reflects the differing eco-physiological sensitivities of the two species to both the direct effects of fire and secondary stressors [4,6,79]. Among these, the spread of Phytophthora ramorum within The Roaches Nature Reserve [44], an oomycete responsible for widespread L. decidua mortality across Europe, may have played a significant role in the death of larches already severely weakened by fire. Thus, the delayed tree mortality appeared to result from a complex interaction between biotic and abiotic factors persisting long after the initial disturbance, as also documented elsewhere [1,80,81]. These factors may have triggered eco-physiological processes leading to chronic tree decline, such as xylem cavitation and carbon starvation. These mechanisms can severely impair tree resilience, diminishing their capacity to sustain key ecosystem services over time [82,83]. Furthermore, soil investigations did not show significant changes in the medium term after fire in properties supporting aboveground vegetation, such as water content, soil organic matter, and microbial biomass [43]. This suggested that the limited woodland recovery was most likely driven by trees eco-physiological dynamics rather than belowground conditions. Indeed, it can be assumed that the soil properties were not significantly affected by the fire because it quickly developed into a crown fire and spread to adjacent peatland [84], without maintaining the intensity or residence time needed to produce detectable ground-level effects up to 5 years later. This has been further corroborated by microbial indices, which showed comparable values between burned and control areas (Table S1), highlighting no stress in the microbial community [85,86,87].

The results highlighted a stable uptake of CO₂ by both species over time, with a marked increase starting in the 1980s. This trend can be attributed not only to global factors, such as rising temperatures and atmospheric CO₂ concentrations, which stimulate photosynthesis and tree growth [7,88], but also to stand-level dynamics, as many trees reached maturity during this period, thereby enhancing their CO₂ assimilation capacity [89]. A contribution from past forest management activities, such as thinning, cannot be excluded, although no historical records are available for the area. However, the 2018 fire caused a drastic decline in CO₂ sequestration capacity in the affected areas, with a particularly significant reduction for L. decidua, due to its high post-fire mortality. The loss of living aboveground biomass directly reduces the amount of carbon dioxide fixed through photosynthesis, further compromising the woodland carbon sink function [90,91]. In the absence of effective natural recovery, the capacity for CO₂ sequestration remains limited in the medium to long term, with consequent negative impacts on local and global climate stability [92]. Therefore, traits that confer fire resistance are crucial for species resilience. Indeed, P. sylvestris showed a greater capacity for survival, probably thanks to its thick bark, deep roots, self-pruning and broad canopy [77], while L. decidua was shown to be more vulnerable, with a dense crown, thin bark and low resprouting capacity [93].

The loss of CO₂ sequestration capacity entails not only ecological damage but also significant economic consequences [11]. The role of forests in mitigating climate change through atmospheric CO₂ uptake is recognized as one of the most important regulating ecosystem services [14,94,95]. The economic value of this service can be estimated using the carbon credit market, which was around €75 per Mg of CO₂ for 2023 [63]. In this case study, the post-fire sharp decline in CO₂ sequestration value highlighted how even relatively small-scale fires can significantly undermine the climate mitigation potential of temperate forests. However, it should be acknowledged that carbon prices are highly variable and influenced by policy and market dynamics, which means that the monetary estimates reported here represent only one possible scenario.

Wildfires cause a double economic loss: first, by releasing previously stored carbon into the atmosphere, especially when mature trees are affected, and second, by reducing the forest’s future capacity to sequester additional CO₂ due to tree mortality as well as slow and incomplete recovery [12,96]. Therefore, the loss of woodland cover observed in burned areas of The Roaches Nature Reserve translated into a potentially long-term economic loss.

Assessing an ecosystem’s post-disturbance resilience requires the use of functional indicators, not just structural ones [97]. For this purpose, the CO₂ Sequestration Loss (CSL) index, implemented in this study, has proven to be a particularly effective tool for assessing the loss of the climate-regulating ecosystem service. Although this specific index is novel, the application of quantitative indices to evaluate carbon uptake dynamics and the effects of forest disturbances is well established in the literature [98]. Similar concepts have been applied to evaluate forest recovery and carbon dynamics through indices of resilience and functional potential [76,97,98,99]. In this study, the CLS index was intentionally designed as a relative metric, expressing the performance of burned stands in comparison with the best local reference, represented by the control area. This approach captures the functional impairment caused by fire directly against the undisturbed sequestration potential, thereby highlighting the ecological relevance of the observed differences.

The obtained results further confirmed that the impact of the fire is not limited to the direct effects, but may trigger prolonged functional decline, implying a decrease in the ecosystem’s capacity to absorb CO₂, undermining climate stability and altering its ecological and economic value [99].

The carbon balance stored by the woodland in 2023, estimated by considering the carbon content in living biomass and topsoil soil, allowed for a direct comparison of carbon stock between burned and control areas. Five years after the fire, carbon loss estimated was around 208 Mg ha⁻¹ in the burned area compared to the control, a difference attributable mainly to the drastic reduction in living tree biomass. Indeed, soil carbon stock did not show statistically significant variations between the two areas. However, a slightly higher absolute carbon stock value was found in the burned area, which may be due to the incorporation into the soil of more recalcitrant forms of organic residues resulting from combustion [87,100,101]. The estimated carbon balance in the burned area (~109 Mg C ha⁻¹) was consistent with other post-fire contexts, falling within the lower range of documented values for mixed conifer forests (101–282 Mg C ha⁻¹ [76]). By comparison, the same fire event determined a reduction of up to 70% of the organic carbon content in the nearby peat bog area of the Nature Reserve [42], with an estimated average carbon loss of around 51.1 Mg ha⁻¹ [102].

It is important to reiterate that the carbon stock estimated in this study refers exclusively to the carbon stored in the living above-ground biomass of P. sylvestris and L. decidua and the soil organic carbon pool at 0–5 cm depth. Notably, other potentially relevant carbon pools, such as below-ground root biomass, understory vegetation, litter, and coarse woody debris, were not included. Therefore, the reported tree–soil carbon stock should be interpreted with caution, as it captures only a portion of the ecosystem’s total carbon pools. Accordingly, the near-total mortality of L. decidua in the burned area raises uncertainties regarding the fate of carbon contained in the dead biomass. Depending on whether this material remains as coarse woody debris, decomposes over time, or is consumed by subsequent fire events, it may act either as a temporary carbon stock or as a significant source of emissions. For this reason, future monitoring of stand recovery and decomposition processes will therefore be crucial to assess this woodland role as a carbon sink or carbon source. Finally, the contrasting carbon loss patterns observed between the woodland and the adjacent peatland underline the complexity and variability of fire impacts within the same geographical area. In the UK context, where forests are relatively scarce, highly fragmented, and often embedded within diverse landscapes, accounting for such heterogeneity is essential for developing site-specific management and restoration strategies aimed at enhancing carbon conservation.

5. Conclusions

The results of this study partially confirmed the initial hypotheses. The 2018 wildfire had a significant impact on the carbon sink of The Roaches Nature Reserve, causing a marked reduction in CO₂ sequestration capacity (−70% in P. sylvestris and nearly −100% in L. decidua), primarily due to the near-total mortality of larch tress. The loss of tree biomass compromised the carbon sink function of the woodland, with persistent effects over time, leading to an overall reduction of ~208 Mg C ha⁻¹ in burned area compared to the control. However, contrary to the initial hypotheses, soil analyses did not show significant changes in carbon stock, suggesting that the edaphic component remained quite stable in the medium term. Nevertheless, the estimated carbon stock (here evaluated as living tree biomass and soil carbon stock) highlighted the need to monitor stand recovery and decomposition processes to assess whether the site will regain its role as a carbon sink or evolve into a persistent carbon source.

From an economic perspective, the strong decline in trees’ CO₂ absorption capacity translated into a substantial loss of ecosystem service. In the post-fire period, the estimated value of this service was reduced by approximately 50% for P. sylvestris, while for L. decidua it was almost completely lost. In the UK, where temperate forests cover relatively limited areas and wildfires are increasing in frequency, these impacts carry greater significance. Even a single fire event can potentially reduce the climate-related benefits provided by forests for several decades, with serious environmental and financial implications.

Therefore, there is a clear need to shift forest management toward a more proactive strategies aimed at enhancing ecosystem functionality and resilience, in order to mitigate forest losses in the UK under increasingly uncertain climate conditions.