Studying the Pre-Industrial to Present-Day Effective Radiative Forcing from Wildfire Emissions Using EC-Earth ^†

Abstract

The current study focuses on the interconnection between wildfires and the atmosphere and more precisely on the radiative effect of wildfire emissions on a global scale. Specifically, the effective radiative forcing (ERF) of present-day wildfire emissions relative to pre-industrial conditions is determined. Atmosphere-only simulations were performed using EC-Earth3, and three wildfire-emission datasets were introduced: BB4CMIP6 and two reconstructed alternatives, one derived from the BB4CMIP6 dataset and one derived from the fire model LPJ-LMfire. Our simulations indicate that the main drivers of ERF are the changes in cloud cover and surface albedo caused by the present-day wildfire emissions.

1. Introduction

Wildfires are an important part of the Earth’s climate system, occurring both naturally and anthropogenically, and are often in the epicentre of climate studies due to the radiative perturbations they cause via their emissions [1,2]. Wildfires emit a variety of gases important for climate dynamics, such as greenhouse gases (carbon dioxide (CO₂) and methane (CH₄)), ozone and aerosol precursors (carbon monoxide (CO), nitrogen oxides (NO_x) and non-methane volatile organic compounds (NMVOCs)), and aerosols (black carbon (BC) and organic carbon (OC)) [1,2]. Therefore, wildfire emissions intervene with incoming shortwave radiation and outgoing longwave radiations in both direct (via absorbing, scattering and reflecting the radiation) and indirect ways (via inhibiting or promoting cloud formation and affecting cloud characteristics), eventually causing so-called radiative forcing [1,2,3]. Briefly, radiative forcing is the quantification of the imbalance between the incoming shortwave radiation and the outgoing longwave radiation caused by climate perturbations, such as the aforementioned wildfire emissions, between two time periods [1,3]. Radiative forcing is measured in W/m² and can be positive (warming effect) or negative (cooling effect).

In this study, we use the Earth System Model (ESM) EC-Earth3 [4] in atmosphere-only simulations with prescribed climatological sea surface temperatures (SSTs) and sea ice concentrations (SICs) to determine the effect of present-day wildfire emissions relative to the pre-industrial era. Specifically, we calculate the Effective Radiative Forcing (ERF) [5,6] which includes the radiative forcing and the consequent rapid adjustments (feedbacks). Furthermore, this study investigates the combined effects of all wildfire emissions, i.e., aerosols, aerosol precursors, and ozone precursors, aiming to provide a more comprehensive assessment of their impact on the atmosphere. It is noted that three (3) sets of simulations were performed, one with the original BB4CMIP6 dataset [7], one with the reconstructed BB4CMIP6 dataset, and one with the reconstructed LPJ-LMfire dataset [8]. Apart from the global analysis, we also identify which regions experience stronger forcing from the aforementioned wildfire emissions and separate the roles of emissions and secondary wildfire effects in the forcing. Overall, this analysis contributes to a better understanding of the historical evolution of radiative forcing and of the role of wildfires in the climate system.

2. Materials and Methods

For this work, atmosphere-only simulations were performed using the EC-Earth3-AerChem configuration. In particular, the general circulation model (GCM) IFS cycle 36r4 (Integrated Forecasting System) from the European Centre for Medium-Range Weather Forecasts (ECMWF) and the chemistry and transport model (CTM) TM5 (Tracer Model version 5) were coupled, both in standard configurations [4,9,10].

To determine the ERF caused by the present-day wildfire emissions compared to the corresponding emissions of the pre-industrial era, 10-year atmosphere-only simulations with SSTs and SICs fixed to their 1850 values were performed. In addition, the mixing ratios of CH₄ and CO₂ were fixed to their 1850 values as well, along with the anthropogenic emissions, while the natural emissions were applied according to van Noije et al. [9,10]. Regarding wildfire emissions, they were fixed to a 10-year mean, as presented in

Table 1. Wildfire emissions applied in each simulation.

Dataset	Simulated Period	10-Year Mean
BB4CMIP6 [7]	Pre-industrial	1850–1859
	Present-day	2010–2015
reconstructed BB4CMIP6 [8]	Pre-industrial	1850–1859
	Present-day	2001–2010
reconstructed LPJ-LMfire [8]	Pre-industrial	1850–1859
	Present-day	2001–2010

, for each simulated period. Moreover, the corresponding global total emission values are presented in Figure 1. Finally, Welch’s t-test [11] was used to determine the statistically significant difference between the pre-industrial and present-day simulations.

3. Results

This section presents the simulation results obtained using the three datasets described in Section 2, as depicted in Figure 2a–c. Specifically, Figure 2a shows the ERF caused by the present-day wildfire emissions compared to the pre-industrial times, and Figure 2b,c present the two components of ERF, i.e., downward shortwave radiation and downward longwave radiation, respectively. Moreover, Figure S1a–d depict the drivers of ERF: Aerosol Optical Depth (AOD), tropospheric ozone (O₃), cloud cover, and surface albedo.

Overall, while the ERF patterns coincide across all three datasets in certain regions, in most areas only two datasets show similar patterns. It is worth mentioning that both reconstructed datasets produce more pronounced ERF estimates than the original BB4CMIP6 dataset in almost all regions examined.

Regarding the drivers of ERF, all examined datasets indicate that in the mid- and high-latitude regions of the Northern Hemisphere, the change in surface albedo is the main ERF driver, while cloud cover change is the secondary driver. These changes in both surface albedo and cloud cover are driven by the effects of wildfire emissions on meteorological conditions [1,2]. In contrast, in the Southern Hemisphere and the tropical regions of the Northern Hemisphere, changes in cloud cover are the main driver of ERF. It is worth noting that cloud cover affects both shortwave and longwave radiation; depending on the region under investigation, ERF is driven by either component or a combination of both.

These findings underscore the stronger influence of aerosols compared to tropospheric O₃ on ERF. Although the secondarily produced tropospheric O₃ increases in tropical regions, especially in the Southern Hemisphere, the ERF in these areas is still mainly driven by the changes in cloud cover due to the wildfire-emitted aerosols.

In terms of AOD (Figure S1a), the BB4CMIP6 and the reconstructed BB4CMIP6 datasets agree in certain regions. However, in some areas—such as in regions of equatorial Asia—the reconstructed BB4CMIP6 dataset shows less pronounced changes, and in specific parts of Africa, even opposite changes. The reconstructed LMfire dataset aligns with the two aforementioned datasets in areas where AOD profoundly decreases in Asia. However, it presents a completely different change in AOD in other regions. For instance, in central South America, the LMfire dataset results in a less pronounced increase in AOD compared to the BB4CMIP6 datasets. In Africa, the differences are even more striking: in the tropical regions, the LMfire dataset shows a more pronounced AOD increase than the other two datasets. Meanwhile, in the Northern Hemisphere part of Africa, the LMfire dataset shows increases in AOD, in contrast to the decrease suggested by the BB4CMIP6 dataset.

Regarding tropospheric O₃ (Figure S1b), both BB4CMIP6 datasets capture the pronounced increase in central South America. However, the original BB4CMIP6 results in pronounced increases across the regions in equatorial Asia and northern Oceania, which does not coincide with the results derived from the reconstructed BB4CMIP6 dataset. Furthermore, the latter indicates tropospheric O₃ decreases across the whole Northern Hemisphere as opposed to the original BB4CMIP6 dataset. The LMfire dataset coincides with the reconstructed BB4CMIP6 dataset in the Northern Hemisphere, while it results in pronounced tropospheric O₃ increases in the tropical regions of southern hemispheric Africa, as opposed to the BB4CMIP6 datasets. In addition, similarly to the AOD results, the LMfire dataset is not in agreement in terms of the pronounced increase seen in the two other datasets, i.e., in the regions of central South America, equatorial Asia and northern Oceania.

4. Conclusions

This work determines the Effective Radiative Forcing (ERF) of the present-day wildfire emissions compared to the corresponding pre-industrial conditions, taking into consideration the synergistic effect of wildfire-emitted aerosols and O₃ precursors. The dependence of the findings on the emission input dataset used is also explored by employing three different such datasets. The findings of this study indicate that the wildfire-emitted aerosols are stronger drivers of ERF than the secondarily produced tropospheric O₃. Also, it is worth mentioning that the drivers of ERF depend on the region under investigation; our results indicate that the main driver of ERF in the mid- and high latitudes of the Northern Hemisphere is the changes in surface albedo induced by wildfire effects on meteorological conditions. Regarding the rest of the Northern Hemisphere, as well as the Southern Hemisphere, our results indicate that the main driver of ERF is cloud cover changes caused by present-day wildfire-emitted aerosols.