Amazonian Fire Events Disturbed the Global Carbon Cycle: A Study from 2019 Amazon Wildfire Using Google Earth Engine ^†

Abstract

An unprecedented number of wildfire events during 2019 throughout the Brazilian Amazon caught global attention, due to their massive extent and the associated loss in the Amazonian forest—an ecosystem on which the whole world depends. Such devastating wildfires in the Amazon has strongly hampered the global carbon cycle and significantly reduced forest productivity. In this study, we have quantified such loss of forest productivity in terms of gross primary productivity (GPP), applying a comparative approach using Google Earth Engine. A total of 12 wildfire spots have been identified based on the fire’s extension over the Brazilian Amazon, and we quantified the loss in productivity between 2018 and 2019. The Moderate Resolution Imaging Spectroradiometer (MODIS) GPP and MODIS burned area satellite imageries, with a revisit time of 8 days and 30 days, respectively, have been used for this study. We have observed that compared to 2018, the number of wildfire events increased during 2019. But such wildfire events did not hamper the natural annual trend of GPP of the Amazonian ecosystem. However, a significant drop in forest productivity in terms of GPP has been observed. Among all 11 observational sites were recorded with GPP loss, ranging from −18.88 gC m⁻² yr⁻¹ to −120.11 gC m⁻² yr⁻¹, except site number 3. Such drastic loss in GPP indicates that during 2019 fire events, all of these sites acted as carbon sources rather than carbon sink sites, which may hamper the global carbon cycle and terrestrial CO₂ fluxes. Therefore, it is assumed that these findings will also fit for the other Amazonian wildfire sites, as well as for the tropical forest ecosystem as a whole. We hope this study will provide a significant contribution to global carbon cycle research, terrestrial ecosystem studies, sustainable forest management, and climate change in contemporary environmental sciences.

1. Introduction

Ecosystem fragmentation due to the countless number of wildfires and the extreme rate of deforestations in the Amazon every year seriously threatens conservation practices, associated biodiversity, and species richness [1,2]. The Brazilian Amazon landscape is universally recognized for its rich biodiversity and species richness, and is also considered as a global repository of ecosystem services [3]. Despite several measures that have been taken to promote conservation of the Amazon and its rich biodiversity, several studies have indicated that the loss in forest cover due to wildfires and deforestation are still occurring and running continuously [1,4,5]. Future projections have suggested that such wildfires and deforestation activities will keep on occurring in the Amazon region time to time [4]. Though wildfires are a very regular phenomenon in the Amazonian landscape, an unprecedented number of wildfire events happened during 2019 throughout the Brazilian Amazon, which caught global attention due to their massive extent and the associated loss, causing serious environmental impacts [5]. Figure 1 shows the number of wildfire events that occurred during 2018 and 2019, and indicates that the number of wildfire events in 2019 has been increased compared to 2018.

The wildfire events that occurred from June to September 2019 caught the global media attention due to their massive expansion and devastation. According to INEP, the wildfire events had an increase of 17% in 2019 compared to 2018 (https://www.bbc.com/news/world-latin-america-49971563). The number of active fires in August 2019 were nearly three times higher than in August 2018, and the highest since 2010 [6]. The belch smoke and soot emitted from the wildfire zones polluted the air massively and disturbed the wildlife along their path, destroying a significant part of one of the most important carbon storehouses left on the planet. After September 2019, the intensity of these wildfires decreased over time, but such a huge wildfire in the Brazilian Amazon in 2019 considerably disturbed the global carbon cycle.

In this study, we have compared the carbon sequestration in terms of gross primary productivity (GPP) between 2018 and 2019 from 12 wildlife spots, in order to understand whether such a massive fire event such as occurred in 2019 had any impact on carbon sequestration and GPP. The 12 wildfire spots at the Brazilian Amazon were selected based on the spatial extension of the 2019 wildfire incident. Several studies, such as Gerwing [7], Kauffman et al. [8], Hughes et al. [9], Cochrane and Schulze [10], and Nepstad et al. [11] have significantly covered the Brazilian Amazon wildfire events for a long time. However, no such study was found on the 2019 Amazon wildfire that particularly addresses the impact on the carbon cycle in terms of GPP. This study was conducted in the open-source Google Earth Engine (GEE) platform.

2. Study Area

This study was conducted at 12 pilot spots spread over the different parts of the Brazilian Amazon. The detailed locations of these spots are provided in

Table 1. Location of the 12 pilot spots located at different parts of the Brazilian Amazon.

Spots	Coordinates (W, S)	Reference Places
P1	−51.45, −11.14	Luciará (Mato Grosso, Brazil)
P2	−51.21, −10.74	Porto Alegre o Norte (Mato Groso, Brazil)
P3	−50.33, −10.90	Inawebohona (Tocantis, Brazil)
P4	−51.22, −12.74	Ribeirão Cascalheira (Mato Grosso, Brazil)
P5	−52.68, −15.02	Novo Sao Joaquim (Mato Grosso, Brazil)
P6	−48.04, −12.90	Paraná (Tocantis, Brazil)
P7	−59.14, −17.02	San Matías (Brazil–Bolivia border)
P8	−59.12, −16.53	San Matías (Brazil–Bolivia border)
P9	−59.29, −15.95	Porto Esperidião (Mato Grosso, Brazil)
P10	−57.84, −20.41	Corumbá (Mato Grosso del Sur,Brazil)
P11	−56.31, −19.88	Miranda (Mato Grosso del Sur, Brazil)
P12	−57.26, −20.60	Kadiwéu (Porto Murtinho; Mato Grosso del Sur, Brazil)

and Figure 2.

3. Materials and Methods

3.1. Datasets

This study incorporated 2018 and 2019 MODIS GPP and MODIS burned area data, obtained using the cloud-based geospatial processing platform GEE. MODIS cumulative eight-day composite GPP products (MOD17A2H) with a 500 m resolution have been used to estimate the plant productivity for both years. The MOD17A2 product derived from the MODIS sensor provides the accumulated value of GPP based on the concept of the efficiency of solar radiation used by vegetation [12]. Similarly, MODIS Terra and Aqua, combining burned area monthly data products (MCD64A1 Version 6) with a 500 m resolution, were used to map the burned areas for both years.

3.2. Methods

To understand the trend of GPP for 2018 and 2019, each overpass record of MODIS GPP with the interval of eight days have been considered for all 12 pilot points. Furthermore, the annual mean of each pilot spot has been calculated. To understand the gap between 2018 and 2019 GPP, the annual mean for each spot has been compared. The whole operation has been conducted in the open-source GEE platform, accessed from the Earth Engine home page (https://earthengine.google.com/).

4. Result and Discussion

4.1. Result

Figure 2 shows the distribution of wildfire zones during 2018 and 2019 over the Brazilian Amazonian landscape. As clearly visualized in Figure 2, the wildfire intensities and spread have significantly increased during 2019, compared to the previous year.

Even though several wildfire events took place in the Brazilian Amazon landscape, we have observed a common general trend of GPP for the year 2018 and 2019 (see Figure 3). Such condition indicates that the wildfire events did not hamper the general trend of GPP of the Amazon forest, but did affect productivity negatively in small-scale scenarios. The highest GPP in 2018 was observed at site 2, with 507.05 gC m⁻² yr⁻¹, and the lowest GPP was observed at site 1, with 181.88 gC m⁻² yr⁻¹, whereas the highest GPP and lowest GPP for the year 2019 were recorded at site 2 and site 4, with 457.67 gC m⁻² yr⁻¹ and 147.38 gC m⁻² yr⁻¹, respectively (see Figure 4). However, a significant drop was observed in GPP in terms of productivity for the year 2019 compared to 2018 in all other 11 pilot spots except site 3 (See Figure 5). Site 11, site 12, and site 8 recorded the highest decreases in GPP, with rates of −120.11 gC m⁻² yr⁻¹, −99.32 gC m⁻² yr⁻¹, and −79.70 gC m⁻² yr⁻¹, respectively. The details of a site-wise decrease in GPP in 2019 compared to 2018 have been provided in Figure 5.

4.2. Discussion

Our study demonstrates that the wildfire events that occurred in 2019 at the Brazilian Amazon significantly impacted the primary productivity of the forest, as well as the Amazonian ecosystem, compared to 2018. We estimated an 8.95% reduction in GPP during 2019 at the 12 pilot study plots in the Brazilian Amazon compared to 2018. Amazon forest is globally recognized as a prime producer of oxygen and receiver of carbon. However, such destructive wildfire events often turn this carbon sink zone into a carbon source, due to wildfires. It has also been observed that though wildfire events do not hamper the general trend of GPP for the forest ecosystem, they significantly reduce forest productivity, resulting in low consumption of carbon for photosynthetic activity and a decrease in autotrophic and heterotrophic respiration [13]. Thus, it can be assumed that the ecosystem respiration over the different wildfire zones of the Amazon forest during 2019 decreased due to forest fires and induced extreme heat. Such disturbances, particularly extreme wildfires events like in 2019, significantly hampers the structure and diversity of the landscape over time by exerting selection pressure, controlling succession, and affecting the ecosystem functioning, including the carbon and nutrient cycles [14]. Thus, it is evident that the wildfire events not only disturbed the structure, composition, and functionality of the terrestrial ecosystem, but also influenced the global CO₂ fluxes and their feedback to the global climate system as a whole [15].

5. Conclusions

The intensity and number of wildfires have significantly increased globally over time due to climate change and global warming. Such situations have accelerated the possibilities of tree mortality, destruction of forest ecosystems, and loss of biodiversity. Quantifying such trends are highly necessary to detect the early signs of ecosystem degradation. Advanced satellite remote sensing technology provides us such a great opportunity to monitor, measure, and take required mitigation measures on a timely basis. However, more advanced vegetation signals like sun-induced fluorescence (SIF) can be incorporated with modern machine-learning models to monitor, predict, and quantify productivity losses in real-time and implement sustainable management techniques to restore the forest ecosystem and biodiversity [16]. We assume that an increase in future wildfire events could turn forest ecosystems into carbon sources, contributing towards positive carbon–climate feedback, which is already anticipated in the tropics [17].