Aerosol Emission Patterns from the February 2019 Karnataka Fire

Abstract

Forests are vital for life on Earth but are threatened by forest fires, which have significant impacts on climate change both locally and globally. This study examines a forest fire that lasted from 15 to 26 February 2019 in Karnataka, India, using the Weather Research and Forecasting model with Chemistry (WRF-Chem) model to analyze the effects and atmospheric spread of fire-emitted aerosols. Model simulations are analyzed to understand the horizontal and vertical transport and radiative effects of the fire. The results show high aerosol levels and smoke particles reaching up to 3.5 km altitude and above. The fire raised near-surface air temperatures by ~1–1.5 °C. The net atmospheric forcing due to the fire over the affected area ranged from approximately 10 to 14 W/m², resulting in heating rates between about 0.002 and 0.005 K/day in the impacted region.

1. Introduction

Forest fires represent a major environmental challenge affecting ecosystems, climate, and human livelihoods. Global forests cover 4.06 billion hectares, accounting for 31% of the total land area [1]. Fire is considered to be one of the major causes of forest degradation in India [2,3]. Forest fires are responsible for significant damage to human property, as well as to wildlife, and they lead to loss of vegetation cover and soil degradation [4]. As these fires increase in frequency and intensity, there is an urgent need to deepen the understanding of their causes, symptoms, and significance at the regional level. Many studies focusing on the Indian region have discussed the consequences of fire hazards [5,6]. To better understand the unique scenario of forest fires in Karnataka, it is important to consider the nuances of the regional climate and environment. The development of accurate models for forest fire prediction and management is critical [7].

Karnataka’s forest ecosystems are highly susceptible to fire events, particularly during the dry season when high temperatures and low humidity create conducive conditions for wildfires. Previous research has shown that anthropogenic activities, including agriculture and shifting cultivation practices, significantly contribute to the incidence of forest fires in southern part of India [8,9]. The Western Ghats, which is a biodiversity hotspot, has been particularly impacted by frequent fire events, leading to a decline in species richness and changes in forest structure [10,11].

Aerosols also play a crucial role in the hydrological cycle through their well-documented ability to modify cloud properties [12]. As highlighted by previous studies [13,14], aerosol optical depth (AOD) is a key parameter for assessing the awareness and distribution of aerosols within the environment. Several studies focusing on the Karnataka region have emphasized the significant increase in AOD during fire events, which directly impact regional climate and weather patterns. During the peak fire season, Karnataka experiences elevated levels of AOD, contributing to significant changes in surface radiation and cloud formation [15]. It has been further explored how these fire-induced aerosol loads affect cloud microphysics, leading to changes in precipitation patterns and potential feedback mechanisms in the regional hydrological cycle [10].

A major component of fire is the heavy aerosol emissions in the atmosphere, including black carbon emissions, with resultant changes in columnar particulate matter 2.5-micron (PM2.5) and black carbon (BC) mass concentrations. This phenomenon is similar to that studied by [16,17,18], which indicates that more particles are produced during fires, which contributes to the increase in the aerosol optical depth. Understanding the implications of the presence of black carbon and the specific changes in columnar PM2.5 and BC mass concentrations requires an exploration of its radiative impacts on the atmosphere, as noted by [14,19], whose work emphasized the intricate relationships between aerosols, clouds, and radiative forcing. The emission of black carbon and other aerosols from these fires has been shown to significantly affect air quality and public health. Studies highlighted how increased levels of PM2.5 and BC during forest fire events are linked to respiratory and cardiovascular health issues in the affected populations [20,21]. Moreover, the deposition of BC on snow and ice surfaces in the Himalayan region, as noted by [22,23], can accelerate melting, leading to potential changes in water resources downstream, which is crucial for regions like Karnataka that rely on monsoon-fed rivers. The effect of aerosols like BC on monsoon onset and general meteorological dynamics during a given monsoon calendar period have also been studied [24,25,26].

The Karnataka fire event has been a focal point of concern in relation to radiative perturbation, and it aligns well with what is known about aerosol-induced radiative forcing [27]. It will be helpful to assess how the fire incidence disturbed the radiational balance in the atmosphere in terms of BC abundance, as studied by [16], so that we can better understand regional climate response and potential feedback. It is important to assess the atmospheric heating rate, which is a key measure in understanding how energy is redistributed within the atmosphere.

This research seeks to provide a comprehensive analysis of the Karnataka fire event, examining a range of factors such as topography, meteorological data, aerosol interactions, and their links with regional climate change during both pre-fire and active fire periods. Through the integration of fire observations with the Weather Research and Forecasting with Chemistry (WRF-Chem) model, we offer a detailed analysis of aerosol transport toward the Western Ghats and the accumulation of aerosols, which in turn influence temperature and rainfall patterns. We also examine the vertical and horizontal distribution of fire-induced emissions to understand their impact on regional atmospheric forcing and heating. A study by [28] shows a strong spring snow cover decrease over the Tibetan Plateau due to absorbing aerosols. Another study by [25] over Uttarakhand region indicated that high atmospheric heating over the Himalayan region contributes significantly to the decrease in the snow cover due to an increase in smoke aerosols during a fire event. Hence, it is important to study the aerosols emitted by fires in different regions to understand their impacts. Thus, this study will enhance our understanding of regional environmental responses to intense biomass burning, thereby supporting more effective fire management and climate adaptation strategies.

2. Model and Observational Data Details

The WRF-Chem version 4.1.5 [29,30,31,32,33] is used to investigate aerosol radiative impacts from a fire event over Karnataka. Below are the key configurations included in the model:

• Physical and chemistry parameterization modules.
• Static surface-level geographical fields from USGS and MODIS.
• Initial and boundary meteorological conditions from NCEP’s GDAS data (0.25° resolution).
• Gas-phase chemistry using MOZART-4 and aerosol processes using GOCART.
• Chemical boundary conditions from EDGAR’s monthly varying emissions inventory (0.1° resolution).
• Fire inventory data from NCAR for plume-rise module.
• Biogenic emissions from MEGAN inventory.
• Radiation processes using RRTM scheme and Fast-J photolysis.
• Cloud microphysics with the Thompson graupel scheme.
• Land surface and surface-layer physics with the Unified Noah model and Revised MM5 Monin–Obukhov scheme.
• Boundary layer processes using the Yonsei University PBL scheme.

For gas-phase chemistry, the WRF-Chem model utilized the Model for Ozone and Related Chemical Tracers version 4 (MOZART-4) [34], and for aerosol processes, it incorporated the Goddard Chemistry Aerosol Radiation and Transport (GOCART) model. Anthropogenic emissions for different gaseous species and trace gases were derived from the EDGAR (Emission Database for Global Atmospheric Research) inventory, which provides monthly emission data at a 0.1° spatial resolution. Fire emissions were sourced from the National Center for Atmospheric Research’s Fire INventory (FINN) and incorporated into the model through an online plume-rise module [35]. The initial and lateral boundary meteorological conditions for the model simulation were provided by the National Centers for Environmental Prediction (NCEP) operational global analysis and forecast (GDAS) data, available every six hours at a horizontal resolution of 0.25° × 0.25°. The simulation period was from 15 to 26 February 2019 and covered the southern Indian subcontinent with a 5 km horizontal resolution and 35 vertical levels up to 50 hPa.

2.1. Details of Observational Datasets

A discussion of temperature and wind datasets obtained from ground-based and reanalysis data is provided to give an overall view of the region.

2.1.1. ERA5

ERA5 is a fifth-generation reanalysis from ECMWF, providing hourly estimates of atmospheric, land, and oceanic variables worldwide [36]. This dataset boasts a spatial resolution of 30 km and includes 137 vertical levels, reaching from the surface to 80 km altitude. By combining model outputs with historical observations, ERA5 creates a complete and consistent global dataset. In our analysis, we employed gridded ERA5 reanalysis data for temperature and winds at the 850 hPa level, utilizing a resolution of 0.25° × 0.25° to validate the results of our model.

2.1.2. IMD

For the period, 1990–2023, the India Meteorological Department (IMD) high-resolution 1 × 1-degree gridded daily temperature data are used [37]. This temperature dataset is based on 395 stations and gridded data were developed by interpolating station data into regular grids using a modified version of Shepard’s angular distance weighting algorithm [37].

2.1.3. Details of the Event

Figure 1a–c shows the domain for study region over the Karnataka state (India). Karnataka is a state located in the southern region of India that encompasses a vast geographical area of 191,791 km², accounting for 5.83% of the country’s total landmass. Situated between latitudes 11°30′ and 18°25′ N and longitudes 74°10′ and 78°35′ E, it boasts 5 national parks and 21 wildlife sanctuaries. The state is blessed with an extensive recorded forest area spanning approximately 38,284 km², which constitutes around two-fifths of its total geographic expanse. The fire occurred in the region of Mysore (12.2958 °N, 76.6394 °E) and Chamrajnagar (11.9261 °N, 76.9437 °E) district of Karnataka state over Bandipur National Park (centered 11.78 °N, 76.45 °E) that is situated between 75°12′17″ E and 76°51′32″ E and 11°35′34″ N and 11°57′02″ N, where the Deccan Plateau meets the Western Ghats, and the altitude of the park ranges from 680 m, covering 868.63 km² [38].

The forest fires over the region burned down 10,920 acres of forest land. Bandipur National Park, located in Chamarajanagar district in the Indian state of Karnataka, was designated as a tiger reserve under Project Tiger in 1973 [39]. The park features diverse vegetation types, including dry deciduous forests, moist deciduous forests, and shrublands [40]. The forest fire took place from 21 to 26 February 2019. The fire also spread to the Mudumalai forest range in Tamil Nadu, causing damage that spanned a few acres.

Traditional land use practices over time, along with changing weather patterns, have affected the frequency of forest fires in this region, causing fires to ignite earlier than in previous years due to dry grass. In tropical deciduous forests, fire is a natural occurrence, especially due to high water stress during the summer. Frequent fires contribute significantly to forest degradation in India. Annual fires can inhibit the growth of grasslands, shrubs, and forests, potentially leading to increased soil erosion [41].

3. Result

3.1. Climatology of Mean Temperature

Figure 2a shows the variation in the mean maximum and minimum temperature for February, showing slight yearly fluctuations and a slight upward trend. The highest recorded temperatures appear around 1999, but after 2005, the Tmax values stabilize, remaining relatively consistent between 23 and 25 °C. This stability aligns with findings from [42], which noted minor fluctuations in pre-monsoon maximum temperatures across India. The lack of significant change post-2005 could be attributed to moderating climatic factors, such as increased cloud cover or changes in land use patterns. The mean minimum temperature shows a more noticeable upward trend compared to Tmax. Tmin starts at around 10–12 °C in the early 1990s and gradually increases, reaching approximately 12–13 °C by 2023. This increase in Tmin is consistent with warming trends documented in various studies, such as [43], which linked rising minimum temperatures to anthropogenic influences and broader climate change patterns. Figure 2b shows the variation in mean maximum and minimum temperature during 21–26 February. The mean maximum temperature during 21–26 February mirrors the overall February trend. During fire, many events, e.g., aerosols blocking sunlight, trapping heat, the loss of vegetation, and altered atmospheric conditions occur in tandem to create a specific scenario. Thus, daytime temperatures decrease due to reduced solar radiation, while nighttime temperatures increase due to trapped heat and lack of air cooling from damaged vegetation. The smoke from fires contains particulate matter (PM) that can trap outgoing longwave radiation (referred to as the greenhouse effect) from the fire event, leading to warmer nighttime temperatures. Figure 2c showcases the effect of increased nighttime temperature (minimum temperature) during a fire event. Additionally, aerosol-induced radiative effects during the fire period could contribute to this trend by trapping outgoing longwave radiation, thus limiting nocturnal cooling [44]. Research by [44] suggests that high aerosol optical depth (AOD) during the fire period may reduce Tmax by scattering and absorbing sunlight. While this scattering limits daytime heating, aerosols can also trap heat at night, which may have contributed to the observed rise in Tmin. Climatologically, the values of Tmax and Tmin appear to be within the normal range (Figure 2a,b). Further, we have also examined anomalies in the spatial distribution of surface temperature.

Figure 3 displays the anomaly for February 2019 and 21 to 26 February 2019, relative to more than 30 years of climatology from 1990 to 2023. A clear gradient in temperature anomalies from north to south and east to west over the selected region is observed, with warmer anomalies in the north transitioning to cooler anomalies in the south (Figure 3a,b). According to a study by [45], temperature gradients along the Western Ghats are common due to orographic lifting and coastal proximity, leading to warmer conditions in the northern regions compared to the southern regions. Similarly, [46] highlights that the Western Ghats significantly influence regional climate by altering temperature gradients, with elevated northern areas experiencing warmer temperatures relative to the southern regions. A report [47] highlighted that southern India has been experiencing significant warming over the last few decades, with maximum temperatures rising by about 0.6 °C per decade in some parts of the region. The warming trend depicted in Figure 3a is corroborated by [48], who note that climate change has led to increased temperature extremes in South India, with significant warming being particularly evident in the northern regions. Figure 2c and Figure 3b illustrate the significant impact of a fire with an approximate area of 44 km² on local temperature increases. However, linking this localized temperature rise to broader climate change phenomena remains challenging. Figure 3b shows an elevated temperature anomaly from 21 to 26 February 2019, which is associated with the fire period over Karnataka region, as compared to Figure 3a. Forest fires can lead to localized warming due to the release of heat and aerosols, which can influence temperature distribution on a wider scale.

3.2. Impact of a Fire Event on Surface-Level Pollutants in Karnataka

During the pre-fire and fire periods, there are significant changes in the levels of smoke aerosols released in the atmosphere and at the surface. Figure 4 shows the model-simulated average surface level PM2.5 from 15 to 26 February 2019 averaged over the entire Karnataka state. Surface-level PM2.5 represents the number of pollutants that are emitted locally. These pollutants ascend to higher altitudes and travel long distances from their sources due to transport and favorable wind conditions. During the fire period, surface-level pollutants in Karnataka increased by more than two-fold. The model-simulated PM2.5 concentrations were mostly ~5 µg m⁻³ (lowest value) during the pre-fire period and intensified significantly to ~15 µg m⁻³ (highest value) on the 23 February 2019 (Figure 4). The release of these additional pollutants into the atmosphere often causes a non-negligible impact on local as well as downwind regions. As they are transported, these pollutants likely alter the atmospheric radiation balance along their pathways. Furthermore, the possibility of these pollutants accumulating over the Western Ghats also has not been ruled out. For simplicity and clarity, we use the terminology “pre-fire period” (18–20 February 2019) and “fire period” (21–26 February 2019) throughout the case study to discuss and estimate the fire’s impact.

3.3. Meteorological Conditions

Many fundamental atmospheric and surface parameters change during the pre-fire and fire periods. Figure 5 illustrates the differences in average temperature and winds over the domain during the fire periods, as simulated by the WRF-Chem model compared to ERA5 at the surface level. The temperature patterns from the model highlight significant discrepancies from ERA5, largely due to ERA5’s coarser resolution, which fails to capture the fire’s impact. While wind directions align with ERA5 data, the WRF-Chem model shows higher wind magnitudes, which could influence aerosol formation and distribution. The model simulates surface winds as predominantly northwesterly over the Arabian Sea and easterly over the source region.

The average difference in surface-level temperature varied between 0.6 °C and 2 °C over the source region, as observed by the model, while ERA5 data show a difference of 0 °C to 0.6 °C. The model simulation shows an increase in the average surface temperature across the entire region due to fire events, which is not clearly reflected in the ERA data; therefore, it is necessary to study such events to simulate the local impact and improve the forecast.

3.4. Aerosol Optical Depth

Figure 6 shows the spatial variability of aerosol optical depth averaged both prior to and during the fire event within the study region. The model shows an increase in AOD values during the fire event compared to the pre-fire period over the entire Western region (Figure 6) due to aerosol emissions from the fire source. Model-simulated AODs range from 0.15 to 0.25 above the fire source region. Lower AODs over the southern Indian region are observed, coinciding with the wind patterns. Previous studies indicated that the WRF-Chem model can effectively capture the spatial characteristics of AOD but tends to slightly underestimate its values [25,49,50,51].

The difference (fire–prefire) in aerosol optical depth due to fire is a matter of concern, as the pollutants emitted are transported toward Western Ghats. The high AOD values were observed to have accumulated over the Western Ghat region due to the vertical and horizontal transport of the aerosols during the fire event, as shown in Figure 7. These aerosols play an important role in altering the weather over the region and need to be considered, as the impact may be non-negligible.

3.5. Black Carbon Concentration

Figure 8a shows the daily variability of surface observations of black carbon (BC) mass concentration from 15 February to 26 February 2019 simulated by the WRF-Chem model. During the pre-fire period, the model captured the values of BC concentrations up to 0.2 µg m⁻³. Consistent with the model-simulated AOD, BC concentrations rose significantly during the fire period from 21 to 26 February. An up to 0.6 µg m⁻³ variation in BC concentration was observed during the fire period.

Our analysis thus far clearly indicates that the fire event led to a widespread increase in aerosol concentrations. To understand the vertical extent of aerosol transport, we have plotted the vertical profile of BC. Figure 8b shows the elevated BC derived from the WRF-Chem model during the fire period. High values of BC are observed at the surface, while a peak at a height of ~650 hPa (~3.5 km) is observed during the fire period. This enhanced peak can have significant localized and regional impacts on the surface temperature and radiative forcing, and these impacts are analyzed further. Additionally, vertically transported BC can significantly influence cloud properties. It has been shown by [52] that an increase in aerosols over central India can cause reduction in cloud fraction due to the enhanced radiative impacts. Model simulations indicate that smoke aerosols are vertically transported to mid-tropospheric altitudes, potentially having significant impacts on local and regional climate change.

3.6. Change in Columnar PM2.5 and BC Mass Concentrations

Figure 9 presents spatial maps of changes in columnar average PM2.5 and BC mass concentrations during the fire period at the near-surface, 900 hPa, and 850 hPa levels. (1 to 1.5 km). A significant increase in PM2.5 is observed over Karnataka regions, specifically the Western Ghat region, along the valley close to the Western Ghats.

We also see that the PM2.5 mass has also increased substantially at the higher altitudes, which is a matter of concern. The BC mass concentration contributes around 3% to PM2.5 and has also been found to be accumulated over the Western Ghats (Figure 9). A non-negligible increase in both PM2.5 and BC mass concentrations has also been observed towards the Arabian sea. From Figure 9, it can be inferred that the smoke from the Karnataka fires was primarily confined to the Western Ghat region and did not reach the eastern Indian region at the lower altitude due to the prevailing wind directions. Nevertheless, there is evidence of BC being transported to higher altitudes, as depicted in Figure 9b, which in turn can cause these aerosols to be transported towards oceanic regions and have radiative impacts. Subsequently, these pollutants have the potential to undergo further dispersion and deposition, which can also be hazardous to human health.

3.7. Variability of Clouds

Figure 10 illustrates the vertical distribution of cloud fraction, showing the variation in cloud cover for the pre-fire and fire period. This vertical distribution during the pre-fire and fire period provides valuable insights into the impact of forest fires on local atmospheric conditions.

During the pre-fire period, a small peak is observed between the altitude of 850 to 700 hPa, which is denoted by the dashed line in Figure 10. There is a decline in cloud fraction values during the fire period of up to 500 hpa. This suggests that there is a reduction in cloud cover during the forest fire, possibly due to the significant quantities of smoke and aerosols emitted. This might have led to the dissipation and suppression of cloud formation [16].

3.8. Radiative Perturbation Due to Karnataka Fire Event

Aerosol radiative forcing from fire events is calculated as the difference in clear-sky net flux (downward minus upward) averaged over the fire and pre-fire periods at a specified atmospheric layer [25]. The difference in radiative forcing between the top of the atmosphere (TOA) and the surface (SFC) yields the atmospheric forcing (ATM) caused by the fire aerosols. The model accounts for all aerosols present in the atmosphere during the fire event to calculate aerosol radiation, meaning the radiative forcing here represents the radiative disturbance experienced by the atmosphere and surface due to additional aerosols from the forest fire period. Figure 11 displays the net aerosol radiative forcing from fire events across the entire domain area at the TOA, ATM, and surface.

The radiative impact of a forest fire is most pronounced in the vicinity of the source region but is also significant over the Western Ghats region and over the Arabian sea, where the pollutants were transported. The NET atmospheric forcing is mostly positive and varied between ~10 and 14 Wm⁻² over the affected areas. Conversely, the net surface radiative forcing takes on negative values, spanning from ~0 to −4 Wm⁻². A positive radiative force within the atmosphere can exert a significant influence on the vertical temperature profile, quantified in terms of the heating rate.

3.9. Atmospheric Heating Rate

The net atmospheric heating rate is positive and exhibits a range that varies between 0.002 and 0.005 K/day. The notable increase in the atmospheric heating rate aligns closely with the presence of an elevated black carbon layer observed in the region (Figure 12). The atmospheric heating rate attributed to fire event has also expanded its reach over the Western Ghat and coastal areas (Figure 12). The heating rates of approximately 0.005 K/day, based on the estimations made in this study, in the lower atmosphere could have implications for atmospheric stability.

4. Conclusions

The WRF-Chem model-simulated spatial patterns, meteorological conditions (winds and temperature), as well as aerosol concentrations and their distributions were studied in detail for the Karnataka fire event. While the fire event was less intense and lasted for a shorter duration, it still resulted in a notable near-surface temperature increase of ~1.5 °C. Forest fires result in local warming resulting from the emission of heat and aerosols, potentially altering temperature distribution over a wider region. The consistent rise in Tmin is linked to anthropogenic influences and broader climate change patterns. The model also predicted PM2.5 concentrations of approximately 5 µg m⁻³ during the pre-fire period, which increased significantly to around 15 µg m⁻³ along with a wide distribution of black carbon (BC) at altitudes of 850 to 650 hPa (1.5 to 3.5 km), with concentrations reaching peaks of 0.1 to 0.2 µg m⁻³ and a reduction in cloud fraction values up to 500 hpa during the fire event. The net atmospheric radiative forcing, influenced by both shortwave and longwave radiation, ranged from 10 to 14 W/m², contributing to an atmospheric heating rate of up to 0.005 K/day. During the fire event, there was a local increase in Tmin. These findings align with those of previous studies indicating the significant impact of forest fires on local climate conditions [53,54]. The unique topography of the Western Ghats contributed to the accumulation of elevated aerosol values, as it trapped aerosols through both vertical and horizontal transport during the fire event, potentially altering precipitation patterns in downwind regions. This underscores the necessity for detailed investigations into forest fire events and their broader atmospheric impacts. Comprehensive analyses employing in situ, satellite, and modeling approaches are essential to evaluate the effects of such fires and to inform effective mitigation policies and strategies. Addressing these challenges will be crucial in safeguarding both local ecosystems and broader climatic stability in the face of increasing forest fire incidents.