Plain Language Summary
Wildfire is a dominant disturbance agent that drives ecosystem change, climate forcing, and carbon cycle in Arctic tundra. Tundra fires can exert a considerable influence on the local ecosystem functioning and contribute to climate change. However, the drivers and mechanisms of tundra fires are still poorly understood. Research on modeling contemporary fire occurrence in the tundra is also lacking. Here we examined the key environmental factors that drive tundra fire occurrence with numeric weather prediction and statistical models. We found that tundra fire occurrence is primarily controlled by cloud-to-ground lightning. Warmer and drier fire weather conditions also support burnings in the tundra. We recommend the integration of lightning modeling with numeric weather prediction model for fire monitoring and forecasting in the data-scarce regions like the Arctic.
1 Introduction
Wildfire plays an essential role in altering ecosystem functioning, driving land cover change, and affecting carbon balance in boreal forest and tundra ecosystems (Bret-Harte et al., 2013; Mack et al., 2011; Randerson et al., 2006; Rocha and Shaver, 2011; van Wees et al., 2021; Wang et al., 2021). Though typically less severe than the boreal forest fires, tundra fires are widespread across the pan-Arctic region. Particularly, Alaskan tundra burns more than any other tundra region across the globe, according to satellite-based observations (He et al., 2019; Loboda et al., 2017). In recent years, several large fire seasons have occurred in Alaskan tundra, including the 2010 fire season in the Noatak River Valley, the 2015 fire season in Southwest Alaska, and the now infamous extreme 2007 Anaktuvuk River fire on the North Slope.
Tundra fires can lead to shrub expansion, alter organic soil properties and affect the surface energy budget in the local ecosystems (Bret-Harte et al., 2013; Frost et al., 2020; He et al., 2021; Rocha and Shaver, 2011). They also have the potential to release the ancient carbon stored in the frozen organic soil and cause widespread permafrost degradation and thermokarst development (Jones et al., 2015; Mack et al., 2011). Moreover, habitat suitability and forage availability for numerous wildlife species, e.g. caribou, are threatened by such fires, affecting the living resources of local human societies (Gustine et al., 2014; Joly et al., 2012). Under the rapid climate warming in the Arctic, the tundra could become more vulnerable to burnings due to the increased danger of lightning activity and extreme fire weather (French et al., 2015; McCarty et al., 2021; Young et al., 2017), which will threaten permafrost carbon and result in substantial feedbacks into regional to global climate systems, and circumpolar indigenous and nonnative communitites (Bogdanova et al., 2021; Chen et al., 2021; Forbes, 2013; Hu et al., 2015). However, tundra fires attract less scientific attention compared to fires in other ecosystems. Current research primarily focuses on evaluating post-fire impacts with comparatively little attention to understanding driving mechanisms and modeling tundra fire occurrence.
Fire occurrence results from a combination of ignition and propagation. Cloud-to-ground (CG) lightning and, to a lesser extent, human activity (due to minimal human presence) are the primary ignition sources in tundra ecosystems. Three types of forces generally control fire propagation: fuel, weather, and topography, as summarized by the “Fire Environment Triangle” (Pyne et al., 1996). Fuel type, representing properties of the fuel itself, and fuel moisture state, related to vegetation moisture content, are critically important factors controlling fire-environment interactions by affecting fuel flammability and fire characteristics. Topography also influences fire propagation directly by altering wind patterns or upslope preheating, and indirectly by controlling fuel moisture state through exposure to sunlight and moisture pooling. Finally, fire weather is frequently the dominant contributor to wildfire occurrence across different temporal scales through impacts on fuel moisture state and ignition source. Various fire danger rating systems, that implicitly or explicitly bundle weather impacts on fuel moisture, have been developed to capture the broader impact of weather on expected fire growth and quantify the potential fire risk. Specifically, the National Fire Danger Rating System (NFDRS) implemented in the US and the Canadian Forest Fire Weather Index System (CFFWIS) are the best known and most broadly used in the high northern latitudes (HNL).
Previous studies in the HNL have not reached a consensus regarding the relative impacts of various environmental factors on wildfire occurrence. The majority of the existing studies focused on the boreal forests when examining the environmental drivers of wildfire behaviors. Liu et al. (2012) found out that lightning-ignited fires were controlled by fuel moisture and vegetation type in the boreal forests of Northeast China. While studies in North America emphasized the impacts of atmospheric stability, count of lightning strikes, and dry weather on boreal forest fires (Peterson et al., 2010). Veraverbeke et al. (2017) suggested that lightning activity explained the burned area trends in the boreal forests of North America during recent large fire years. Though lightning characteristics like polarity and peak current were found significant in modeling fire occurrences (Müller and Vacik, 2017; Vecín-Arias et al., 2016), they did not function as major contributors in other studies (Adámek et al., 2018; Pineda et al., 2014).
Nevertheless, these findings in the boreal forests are not readily transferrable to the treeless tundra, as the land-atmosphere interactions differ substantially between the two ecosystems (Chambers et al., 2005; Dissing and Verbyla, 2003; Jiang et al., 2015; Van Heerwaarden and Teuling, 2014). Previous studies have modeled historical or future tundra fire regimes with ecosystem or statistical models (Higuera et al., 2011; Joly et al., 2012; Sae-Lim et al., 2019; Young et al., 2017). Specifically, Young et al. (2017) modeled future fire occurrence probability in Alaska accounting for climate and landscape features. Masrur et al. (2018) found that warm and dry conditions affect the spatiotemporal patterns across the circumpolar Arctic tundra. Yet, efforts on examining the driving mechanisms and contemporary modeling of fire occurrence have been lacking in the tundra ecosystems in existing research. Critical factors such as lightning, were not considered in these studies.
This study investigates the key environmental factors controlling fire occurrences in Arctic tundra via contemporary modeling during 2001 – 2019. We defined the wildfire occurrence as the start of an individual fire event detected by satellite sensors. We developed an empirical-dynamical framework to predict the fire occurrence probability by combining numerical weather prediction (NWP) and machine learning models. We considered factors that control wildfire behaviors, including fuel, fire weather, topography, and ignition source.
2 Materials and Methods
2.1 Data and variable preparation