Wildfire Becoming More Severe and Expanding into New Areas in Alaska: Increasing Climate Awareness and Resilience

The boreal ecosystem in Alaska, predominantly situated between the Brooks Range and the Alaska Range, is naturally prone to seasonal wildfire. However, increasing heat and drought extremes are influencing the severity of wildfire in this region, and are contributing to extreme wildfire behavior in new areas. Not only are temperature extremes during the peak fire season helping to boost fire danger, but record spring warmth and low snow cover are promoting an earlier start to the fire season. Most recently, Southcentral Alaska in 2019, and Southwest Alaska in 2022, experienced their worst wildfire seasons of record with over two-million acres burned each year statewide. Health studies in the region have shown that exposure to smoke and extreme heat led to increased respiratory and cardiovascular effects in populations unaccustomed to prolonged exposure to these environmental hazards. Thus, there are needs among the fire management and environmental health communities for improved climate information about seasonal fire prediction and decadal-scale projections of fire- and health-relevant indicators. Certain large-scale meteorological patterns promote active fire seasons in Alaska and their prevalence is associated with Arctic and Pacific teleconnection indices. For example, strong El Niño phases have been linked to increased wildfire in Alaska, due to higher temperatures and decreased precipitation. The last strong El Niño prior to the ongoing one occurred in 2015, when over five million acres burned in Alaska, the state’s second-highest total of record. However, the teleconnection indices are based on statistical relationships that connect weather phenomena between two disparate regions of the world, and these relationships are subject to strengthen or weaken through time. Climate warming, which is occurring in the Arctic at twice the rate of the global average, is impacting regional sea ice coverage, snow season length, and atmospheric moisture availability. All of these factors have a major influence on a given fire season’s evolution, and changes to them will affect the skill of any seasonal fire forecast that uses historical teleconnection relationships. Current results from this research have identified significant relationships (p < 0.05; Kendall’s tau) between annual area burned in Alaska and the June value of four teleconnection indices, including the: Alaska Blocking Index, East Pacific/North Pacific pattern, Niño 4 region, and the Southern Oscillation Index. These correlations, which were calculated monthly from May-July, used a 30-year reference period from 1993-2022. The primary mechanisms for an active fire season appear to be the establishment of a high-pressure ridge across Alaska in June, and a southeasterly flow pattern in July and August that delays the seasonal transition to cool moist southwesterly flow off of the Bering Sea and North Pacific. The former promotes hot and dry conditions that dries out the fine fuels on the surface as well as the deeper duff layer, while the latter advects warm continental air that supports ample daytime heating for increased instability and convection. Given that these teleconnection patterns are, or can be, forecast in advance, there is an opportunity to include these results in seasonal fire outlooks. Development is underway of an interactive model that takes inputs of preferred teleconnection indices, weighting (e.g., skill or no skill), training period, and outputs categorical forecasts (e.g., above normal, normal, below normal) of temperature, precipitation, and burned area in Alaska. This tool will further categorize results regionally by Predictive Service Area, and temporally by fire subseason, which are all used operationally by fire management to help determine how to best allocate fire suppression resources. The fire subseasons in Alaska include the: wind-driven (1 April-10 June), duff-driven (11 June-20 July), cumulative drought (21 July-9 August), and diurnal-effect (10 August-30 September). This model will be validated using the Heidke skill score, which is used for categorical forecasts. The recent severe wildfire seasons in Alaska have also exposed vulnerabilities to human health that need to be better understood, especially given the rate of observed climate warming. For example, a heat index of 70˚˚ F has been associated with an increased number of emergency room visits, according to data from the Alaska Health Facilities Data Reporting Program. Greater impacts were identified when this heat threshold was exceeded for multiple consecutive days and impacts were further exacerbated at times by unhealthy levels of smoke. Recent years have also seen a continuation of the trend for an increasing number of smoky days in Alaska, one that is expected to worsen as annual area burned increases. Climate information is needed to support environmental health preparedness, including decisions about changing air filtration systems in public spaces, making cooling centers available, and developing emergency evacuation plans for wildfire. This research is designed to facilitate climate adaptation efforts by producing decadal-scale projections of fire- and health-relevant indices, often in the context of extreme scenarios. Heat domes, for example, have impacted large regions of the Earth during the past few summer seasons, including in the Pacific Northwest in 2021 where persistent high pressure promoted and preceded extreme wildfire conditions. Alaska has yet to experience such an event, which allows us to translate what heat domes from other regions of the world would mean for Alaska before it happens. Heat dome case studies are being identified, and standardized anomalies of meteorological variables, relative to the region of interest, are being extracted from the ERA5. These standardized anomalies will then be applied to the corresponding distributions of the meteorological variables in Alaska to identify their translated magnitudes. These values will be used to calculate the fire- and health-relevant indicators to construct meaningful extreme scenarios. Projected distributions of the relevant meteorological variables will be extracted, bias-corrected, and the indicators calculated from 1) an ensemble of model runs from CMIP6, and 2) a set of dynamically downscaled simulations, driven from a subset of the CMIP6 models. Given these distributions, an assessment of the future probability of similar extreme events can be made.