Characterising the holdover phase of lightning-ignited wildfires in Catalonia

HIghlights

• Lightning efficiency in the triggering of wildfires is of one in ∼ 840 lightning.

• Lightning-ignited fire distribution is not related to flash density but to land-use.

• Most lightning fires flare up in the short term, latent fires above 24 h are rare.

• The Ångström Index is an effective means to predict potential lightning fires.

• Low-intensity lightning fires should be considered as a tool to reduce forest fuels.

Abstract

In the present work, we analysed fuel moisture conditions during ignition, holdover, and detection of flaming combustion of lightning-ignited wildfires in Catalonia (north-eastern Spain) between 2003 and 2018. First, we identified the most probable lightning candidate for each wildfire, implementing a matching algorithm between historical lightning-caused fire data and cloud-to-ground lightning records. The bulk of lightning-fire ignitions (80%) occurred between June and September during the warm season. Conifer forests concentrate almost half of the lightning-ignited wildfires. Then, we spatially interpolated air temperature and relative humidity data from automatic weather stations to calculate a weather index describing the evolution of fuel moisture content at the specific wildfire location, from the time of the lightning-caused ignition to the time of the fire detection. Results showed that fuel moisture content drives lightning-ignited wildfires since most ignitions around midday turn into flaming combustion almost immediately, when fuel moisture content reaches the minimum of the day. The holdover duration increased in late afternoon lightning-ignited fires, which remain smouldering overnight and evolve to flaming combustion in the next solar cycle. We found that latent fires above 24 h were rare (15%), and only 3% of the fires had a holdover period above three days. Only 1 in ∼840 cloud-to-ground flashes started a wildland fire. Our outcomes provide valuable insight to improve the modeling and management of natural wildfires in the Mediterranean areas.

Introduction

Lightning wildfires play a significant role in the conservation of fire-adapted Mediterranean ecosystems. Lightning damage in trees facilitates insect and disease attacks (McMullen and Atkins, 1962; Schmitz and Taylor, 1969) and promotes forest regeneration (Komarek, 1968; Latham and Williams, 2001). When lightning strikes a tree, a continuous groove of bark is stripped out, following a crack along the trunk (Mäkelä et al., 2009). In extreme cases, the lightning discharge can blow off the bark, torch the crown, or even start a wildfire (Plummer, 1912; Taylor, 1969; Fuquay and Baughman, 1969). The fire severity and extent of the burned area depend on fuel loads, topographic position, cumulative drought, and fire-weather conditions (Román-Cuesta et al., 2009; Stephens et al., 2018).

The physical process involved in a lightning-ignited wildfire (LIW) is divided into three stages: (i) lightning-caused ignition, (ii) ignition survival or holdover phase, and (iii) flaming combustion plus wildfire growth (Anderson, 2002). In fact, not all cloud-to-ground (CG) lightning flashes can cause a fire ignition. The lightning efficiency (LE) ratio refers to the lightning ignitions that eventually cause a fire, in relation to the total number of CG lightning (Podur et al., 2003). Despite the energetic electrical discharge, the peak power of the CG return stroke is too short to raise the temperature to the ignition level (Latham and Williams, 2001). Instead, it is generally accepted in the scientific literature that only lightning with a long continuing current (LCC) phase can cause the ignition of forest fuels. That is, continuing currents lasting longer than ∼ 40 ms are often sufficient to heat the fuels and start a wildfire (Fuquay and Baughman, 1969; Fuquay et al., 1972). High-speed video observations have shown that LLC is present in both negative (-CG) and positive (+CG) flashes, but not all flashes show an LLC (Campos et al., 2007; Saba et al., 2010; Pineda et al., 2014). The lightning-caused incipient fire may rarely grow since the rainfall from the associated thunderstorm system is often sufficient to suppress most ignitions. The fraction of ignitions caused by “dry lightning” is more likely to end in a flaming wildfire because precipitation reaching the ground is weak or non-existent (Rorig and Ferguson, 1999; Pineda and Rigo, 2017). If weather conditions are conductive, lightning-caused ignitions can immediately spread as an active fire. Contrarily, if the fuel moisture content is high but less than the moisture content of extinction, the fire may smolder as a “holdover fire” (Flannigan and Wotton, 1991; Martell and Sun, 2008).

Indeed, fuel moisture during lightning occurrence plays an essential role in the probability of reaching the critical ignition temperature. A low fuel moisture content would allow the energy and heat of the lightning to the critical ignition temperature (Fuquay et al., 1979; Latham and Schlieter, 1989; Morin et al., 2015). Therefore, assessing fuel moisture content is an essential tool for fire management (Viney, 1991; Sharples et al., 2009). In addition, fuel type and loads have their part since fine fuels such as litter, dead pine needles and slash are more prone to ignite (Viegas et al., 1992; Meisner et al., 1993; Vecín-Arias et al., 2016). Besides, a greater lightning fire occurrence associated with conifer stands has been reported in different parts of the world, from boreal Picea spp stands to mixed coniferous forests in the Alps (Krawchuk et al., 2006; Reineking et al., 2010; Müller et al., 2013). Compared to other forest species, conifers in Mediterranean areas often present hazardous conditions due to poor management (Nuñez-Reguerira et al., 2000).

Low-intensity LIWs show a beneficial effect in many Mediterranean forest systems such as cork oak (Quercus suber L.) woodlands and black pine (Pinus nigra Arn.) old-growth habitats (Domènech et al., 2018). The thick bark, high crown, and deep root systems of mature trees in these species make them resistant to surface fires (Fulé et al., 2008; Camarero et al., 2019). Nonetheless, the bulk of the burned area concentrates in summer when high temperatures and strong winds after prolonged droughts provide the conditions for extreme wildfires (Ganteaume et al., 2013). Humans largely cause these fires, which burn extensive areas, and concentrate environmental and economic losses to forest values and property (Salis et al., 2019). The core causative factors are the rapid fuel build-up resulting from a fire exclusion policy and the rural exodus (Brotons et al., 2013; Cervera et al., 2016). Since lightning is a minor cause of forest fires in the southern EU countries (5–10%) (San-Miguel-Ayanz et al., 2013), LIW are commonly perceived as irrelevant. Nonetheless, LIW must be considered a major disruptive agent in Mediterranean-climate regions. Although most LIWs burn less than 1 ha, some of the largest fires recorded in Spain were caused by lightning (Fernandes et al., 2021).

Previous works on LIWs conducted in Catalonia focused on describing their general characteristics (Pineda et al., 2014), studied the influence of rainfall (Pineda and Rigo, 2017), and characterised the type of thunderstorm causing the LIW (Soler et al., 2021). These studies, among others, have reported that LIWs do not necessarily flare up and spread immediately upon the lightning-caused ignition. Holdover LIWs present a challenge to its real-time detection and prediction. In this study, we analyze the influence of the fuel moisture conditions during the different LIWs stages, and the holdover phase in particular, to better understand the natural fire regime in Catalonia. This knowledge is essential for integrating unplanned fires into fuels management and landscape restoration plans. Taking advantage of a high-resolution spatial interpolation scheme, we calculated air temperature and relative humidity for each LIW ignition location. Once combined into a fuel moisture content index, we used the index to accurately characterize evolving conditions along the three stages of the LIW: lightning-caused ignition, holdover period, and flaming combustion initiation (i.e., fire detection). Moreover, we analyze three historical LIW episodes showing a long holdover phase.