Trajectories of wildfire behavior under climate change. Can forest management mitigate the increasing hazard?

Mediterranean forests and fire regimes are closely intertwined. Global change is likely to alter both forest dynamics and wildfire activity, ultimately threatening the provision of ecosystem services and posing greater risks to society. In this paper we evaluate future wildfire behavior by coupling climate projections with simulation models of forest dynamics and wildfire hazard. To do so, we explore different forest management scenarios reflecting different narratives related to EU forestry (promotion of carbon stocks, reduction of water vulnerability, biomass production and business-as-usual) under the RCP 4.5 and RCP 8.5 climate pathways in the period 2020-2100. We used as a study model pure submediterranean Pinus nigra forests of central Catalonia (NE Spain). Forest dynamics were simulated from the 3rd National Forest Inventory (143 stands) using SORTIE-nd software based on climate projections under RCPs 4.5 and 8.5. The climate products were also used to estimate fuel moisture conditions (both live and dead) and wind speed. Fuel parameters and fire behavior were then simulated, selecting crown fire initiation potential and rate of spread as key indicators. The results revealed consistent trade-offs between forest dynamics, climate and wildfire. Despite the clear influence exerted by climate, forest management modulates fire behavior, resulting in different trends depending on the climatic pathway. In general, the maintenance of current practices would result in the highest rates of crown fire activity, while management for water vulnerability reduction is postulated as the best alternative to surmount the increasingly hazardous conditions envisaged in RCP 8.5.

Geospatial Modeling of Containment Probability for Escaped Wildfires in a Mediterranean Region.

Despite escalating expenditures in firefighting, extreme fire events continue to pose a major threat to ecosystem services and human communities in Mediterranean areas. Developing a safe and effective fire response is paramount to efficiently restrict fire spread, reduce negative effects to natural values, prevent residential housing losses, and avoid causalties. Though current fire policies in most countries demand full suppression, few studies have attempted to identify the strategic locations where firefighting efforts would likely contain catastrophic fire events. The success in containing those fires that escape initial attack is determined by diverse structural factors such as ground accessibility, airborne support, barriers to surface fire spread, and vegetation impedance. In this study, we predicted the success in fire containment across Catalonia (northeastern Spain) using a model generated with random forest from detailed geospatial data and a set of 73 fire perimeters for the period 2008-2016. The model attained a high predictive performance (AUC = 0.88), and the results were provided at fine resolution (25 m) for the entire study area (32,108 km2 ). The highest success rates were found in agricultural plains along the nonburnable barriers such as major road corridors and largest rivers. Low levels of containment likelihood were predicted for dense forest lands and steep-relief mountainous areas. The results can assist in suppression resource pre-positioning and extended attack decision making, but also in strategic fuels management oriented at creating defensive locations and fragmenting the landscape in operational firefighting areas. Our modeling workflow and methods may serve as a baseline to generate locally adapted models in fire-prone areas elsewhere.

Towards a comprehensive wildfire management strategy for Mediterranean areas: Framework development and implementation in Catalonia, Spain.

Southern European countries rely largely on fire suppression and ignition prevention to manage a growing wildfire problem. We explored a more wholistic, long-term approach based on priority maps for the implementation of diverse management options aimed at creating fire resilient landscapes, restoring cultural fire regimes, facilitating safe and efficient fire response, and creating fire-adapted communities. To illustrate this new comprehensive strategy for fire-prone Mediterranean areas, we developed and implemented the framework in Catalonia (northeastern Spain). We first used advanced simulation modeling methods to assess various wildfire exposure metrics across spatially changing fire-regime conditions, and these outputs were then combined with land use maps and historical fire occurrence data to prioritize different fuel and fire management options at the municipality level. Priority sites for fuel management programs concentrated in the central and northeastern high-hazard forestlands. The suitable areas for reintroducing fires in natural ecosystems located in scattered municipalities with ample lightning ignitions and minimal human presence. Priority areas for ignition prevention programs were mapped to populated coastal municipalities and main transportation corridors. Landscapes where fire suppression is the principal long-term strategy concentrated in agricultural plains with a high density of ignitions. Localized programs to build defensible space and improve self-protection on communities could be emphasized in the coastal wildland-urban interface and inner intermix areas from Barcelona and Gerona. We discuss how the results of this study can facilitate collaborative landscape planning and identify the constraints that prevent a longer term and more effective solution to better coexist with fire in southern European regions.

Pastoral wildfires in the Mediterranean: understanding their linkages to land cover patterns in managed landscapes.

The pastoral use of fire to regenerate rangelands is a major cause of wildfires in many Mediterranean countries. Despite producing important environmental impacts, this phenomenon has hardly ever been studied separately from other wildfire ignition causes. As extensive livestock breeding relies on the available pasture resources, we hypothesised that a higher rate of pastoral wildfire ignitions could be associated with land cover patterns, as these reflect the spatial arrangement of human activities in managed landscapes. To investigate these patterns, we studied landscape structure and the pastoral wildfires recorded between 1988 and 2000 in 24 Nature Park landscapes in Andalusia (Spain). The CORINE Land Cover map was reclassified according to five levels of grazing use and landscape metrics were calculated. Neural networks were developed to model the relationship between landscape metrics and pastoral wildfires, obtaining a set of significant variables which are discussed in the frame of land and livestock management in the region. We conclude that pastoral wildfire ignitions are more likely in landscapes where the pattern of being dominated by a matrix composed of several large patches of low to moderate grazing use, and having abundant small and elongated patches of higher grazing use, is more extreme. This pattern could be reflecting the persistence of numerous small livestock farms within an increasingly abandoned agrarian landscape. To prevent pastoral wildfires, land management could attempt to enlarge and merge those small patches of higher grazing use, reducing the amount of interface and their intermixture with the surrounding poorer pasture resources.

Convergence in critical fuel moisture and fire weather thresholds associated with fire activity in the pyroregions of Mediterranean Europe.

Wildfires are becoming an increasing threat to many communities worldwide. There has been substantial progress towards understanding the proximal causes of increased fire activity in recent years at regional and national scales. However, subcontinental scale examinations of the commonalities and differences in the drivers of fire activity across different regions are rare in the Mediterranean zone of the European Union (EUMed). Here, we first develop a new classification of EUMed pyroregions, based on grouping different ecoregions with similar seasonal patterns of burned area. We then examine the thresholds associated with fire activity in response to different drivers related to fuel moisture, surface meteorology and atmospheric stability. We document an overarching role for variation in dead fuel moisture content (FMd), or its atmospheric proxy of vapor pressure deficit (VPD), as the major driver of fire activity. A proxy for live fuel moisture content (EVI), wind speed (WS) and the Continuous Haines Index (CH) played secondary, albeit important, roles. There were minor differences in the actual threshold values of FMd (10-12%), EVI (0.29-0.36) and CH (4.9-5.5) associated with the onset of fire activity across pyroregions with peak fire seasons in summer and fall, despite very marked differences in mean annual burned area and fire size range. The average size of fire events increased with the number of drivers exceeding critical thresholds and reaching increasingly extreme values of a driver led to disproportionate increases in the likelihood of a fire becoming a large fire. For instance, the percentage of fires >500 ha increased from 2% to 25% as FMd changed from the wettest to the driest quantile. Our study is among the first to jointly address the roles of fuel moisture, surface meteorology and atmospheric stability on fire activity in EUMed and provides novel insights on the interactions across fire activity triggers.

Has COVID-19 halted winter-spring wildfires in the Mediterranean? Insights for wildfire science under a pandemic context.

Wildfires in the Mediterranean are strongly tied to human activities. Given their particular link with humans, which act as both initiators and suppressors, wildfire hazard is highly sensitive to socioeconomic changes and patterns. Many researchers have prompted the perils of sustaining the current management policy, the so-called 'total fire exclusion'. This policy, coupled to increasingly fire-prone weather conditions, may lead to more hazardous fires in the mid-long run. Under this framework, the irruption of the COVID-19 pandemic adds to the ongoing situation. Facing the lack of an effective treatment, the only alternative was the implementation of strict lockdown strategies. The virtual halt of the system undoubtedly affected economic and social behavior, triggering cascading effects such as the drop in winter-spring wildfire activity. In this work, we discuss the main impacts, challenges and consequences that wildfire science may experience due to the pandemic situation, and identify potential opportunities for wildfire management. We investigate the recent evolution of burned area (retrieved from the MCD64A1 v006 MODIS product) in the EU Mediterranean region (Portugal, Spain, France, Italy and Greece) to ascertain to what extent the 2020 winter-spring season was impacted by the public health response to COVID-19 (curfews and lockdowns). We accounted for weather conditions (characterized using the 6-month Standardized Precipitation Evapotranspiration Index; SPEI6) to disregard possible weather effects mediating fire activity. Our results suggest that, under similar drought-related circumstances (SPEI6 ≈ -0.7), the expected burned area in 2020 during the lockdown period in the EU (March-May) would lay somewhere within the range of 38,800 ha ± 18,379 ha. Instead, the affected area stands one order of magnitude below average (3325 ha). This stresses the need of considering the social dimension in the analysis of current and future wildfire impacts in the Mediterranean region.

Modeling initial attack success of wildfire suppression in Catalonia, Spain.

In southern European regions, the few fires that escape initial attack (IA) account for most of the burned area. Nonetheless, limited effort has been conducted to develop spatiotemporal models aiming at improving pre-positioning and deployment of fire-fighting brigades on the first dispatch. To this end, we calibrated a model to assess the probability of containment of fire by IA in Catalonia (northeastern Spain). The model was trained using machine learning algorithms from georeferenced historical fire ignition locations, fire response and weather conditions. Our results indicated that early detection, ground accessibility, and aerial support governed the broad spatial pattern of fire containment probability, with strong gradients that ranged from lowest chances of containment in northwestern mountains to highest in the coastal belt. In turn, weather conditions and fire simultaneity were crucial to explain the differences during wildfire season. We found that fires igniting above the 85th percentile of temperature and wind speed, during simultaneous fire episodes (n > 10), and 12.5 km away from the nearest fire station will probably escape IA, and grow into large events. These hazardous fire danger conditions were met 13 days per year on average during the period 1998-2015, with 5 fire simultaneous episodes escaping IA that burned 1546 ha in total. Results were provided as a set of high-resolution raster grids (100 m), which replicated the most typical weather and fire occurrence scenarios that first responders are likely to face during the wildfire season. This study reveals existing limitations in the dominant fire exclusion policy of Mediterranean areas and advocates for a comprehensive long-term wildfire management solution. Our model may help inform science-based decision-making on IA and general fire response planning in the study area.

Spatial stratification of wildfire drivers towards enhanced definition of large-fire regime zoning and fire seasons.

The area affected by wildfires is experiencing an overall decrease in the Mediterranean European region. However, there is no clear trend associated to the incidence of large fire events, which continue to pose an important threat to assets-at-risk, while debates on control by meteorological or fuel drivers are ongoing. Understanding the underlying spatial and temporal patterns of large-fire drivers is of critical importance for a more efficient and science-based management, and specifically for improving wildfire season definition and informing fuel management. Taking advantage of the reliable wildfire data available in Spain, we analyzed large fires (>100 ha) in the period 2010-2015 to outline homogenous spatial-temporal regions in terms of the influence of the main drivers of large-fire activity: temperature, wind speed, slope, distance to populated places and roads, and proximity to agricultural lands. We combined Geographically Weighted Logit Regression (GWLR) models to parameterize the marginal influence of the drivers, with optimized hierarchical clustering to define uniform regions in terms of the underlying driving factors. These regions were subsequently analyzed for monthly distribution of fire occurrence and associated fuel models. We identified four different zones in terms of drivers' features, capturing dissimilar intra-annual patterns of fire activity and affected fuels: one covering the Mediterranean and two along the northern coast, and a fourth aggregation in the hinterlands that seems to act as transition area. The Mediterranean and hinterland were linked to weather-related summer ignitions, late and early summer respectively. The northern cluster gathers most winter fires starting in remote locations under steep slopes and strong wind conditions. The northwestern cluster accounts for most of the fire activity in Spain, related to complex relief and shrub-type fuels.

The wildland-urban interface raster dataset of Catalonia.

We provide the wildland urban interface (WUI) map of the autonomous community of Catalonia (Northeastern Spain). The map encompasses an area of some 3.21 million ha and is presented as a 150-m resolution raster dataset. Individual housing location, structure density and vegetation cover data were used to spatially assess in detail the interface, intermix and dispersed rural WUI communities with a geographical information system. Most WUI areas concentrate in the coastal belt where suburban sprawl has occurred nearby or within unmanaged forests. This geospatial information data provides an approximation of residential housing potential for loss given a wildfire, and represents a valuable contribution to assist landscape and urban planning in the region.

Wildfire spread, hazard and exposure metric raster grids for central Catalonia.

We provide 40 m resolution wildfire spread, hazard and exposure metric raster grids for the 0.13 million ha fire-prone Bages County in central Catalonia (northeastern Spain) corresponding to node influence grid (NIG), crown fraction burned (CFB) and fire transmission to residential houses (TR). Fire spread and behavior data (NIG, CFB and fire perimeters) were generated with fire simulation modeling considering wildfire season extreme fire weather conditions (97th percentile). Moreover, CFB was also generated for prescribed fire (Rx) mild weather conditions. The TR smoothed grid was obtained with a geospatial analysis considering large fire perimeters and individual residential structures located within the study area. We made these raster grids available to assist in the optimization of wildfire risk management plans within the study area and to help mitigate potential losses from catastrophic events.

Optimizing prescribed fire allocation for managing fire risk in central Catalonia.

We used spatial optimization to allocate and prioritize prescribed fire treatments in the fire-prone Bages County, central Catalonia (northeastern Spain). The goal of this study was to identify suitable strategic locations on forest lands for fuel treatments in order to: 1) disrupt major fire movements, 2) reduce ember emissions, and 3) reduce the likelihood of large fires burning into residential communities. We first modeled fire spread, hazard and exposure metrics under historical extreme fire weather conditions, including node influence grid for surface fire pathways, crown fraction burned and fire transmission to residential structures. Then, we performed an optimization analysis on individual planning areas to identify production possibility frontiers for addressing fire exposure and explore alternative prescribed fire treatment configurations. The results revealed strong trade-offs among different fire exposure metrics, showed treatment mosaics that optimize the allocation of prescribed fire, and identified specific opportunities to achieve multiple objectives. Our methods can contribute to improving the efficiency of prescribed fire treatment investments and wildfire management programs aimed at creating fire resilient ecosystems, facilitating safe and efficient fire suppression, and safeguarding rural communities from catastrophic wildfires. The analysis framework can be used to optimally allocate prescribed fire in other fire-prone areas within the Mediterranean region and elsewhere.

Human-caused wildfire risk rating for prevention planning in Spain.

This paper identifies human factors associated with high forest fire risk in Spain and analyses the spatial distribution of fire occurrence in the country. The spatial units were 6,066 municipalities of the Spanish peninsular territory and Balearic Islands. The study covered a 13-year series of fire occurrence data. One hundred and eight variables were generated and input to a dedicated Geographic Information System (GIS) to model different factors related to fire ignition. After exploratory analysis, 29 were selected to build a predictive model of human fire ignition using logistic regression analysis. The binary model estimated the probability of high or low occurrence of forest fires, as defined by an ignition danger index that is currently used by the Spanish forest service (number of fires divided by forest area in each municipality). Thirteen explanatory variables were identified by the model. They were related to agricultural landscape fragmentation, agricultural abandonment and development processes. The prediction agreement found between the model binary outputs and the historical fire data was 85.3% for the model building dataset (60% of municipalities). A slightly lower predictive power (76.2%) was found for the validation data (the remaining 40%). The probabilistic output of the logistic was significantly related to the raw ignition index (Spearman correlation of 0.710) used by the Spanish Forest Service. Therefore, the model can be considered a good predictor of human-caused fire risk, aiding spatial decisions related to prevention planning in Spanish municipalities.