Identifying most extreme heatwaves in Latvia based on a skewed temperature distribution.

Heatwaves are periods of extremely high temperature. They are one of the most hazardous extreme meteorological events for the developed countries. Droughts caused by heatwaves can lead to crop failures. Human mortality due to cardiovascular diseases and drowning significantly increases during heatwaves. High temperatures can lead to infrastructure damage. Electrical grid overload might occur due to high air conditioning use. Railroad tracks expand due to heat, which can lead to irreversible damage. Overall, the longer and more intense is a heatwave, the more impact it has on society. Identifying historical heat extremes and evaluating their return period is important to better prepare for similar events in the future.

In this work historical air temperature observation data in Latvia since 1966 was used to develop a statistical model. The model was used to evaluate yearly cycle of probability distribution of temperature related meteorological variables. Temperature related variables chosen in this work were mean temperature, highest and lowest maximum temperature, highest and lowest minimum temperature. Daily temperature is autocorrelated in time, which makes calculation of return periods of heatwaves different from calculations of return periods of just daily mean temperature. Therefore, distributions of these temperature variables were calculated for periods with length ranging between 1 and 30 days.

Usually, probability distribution functions of meteorological variables are calculated based on the reference period assuming normal distribution. The model used in this study considered downsides of such approach. First, during the summer distribution of daily mean temperature in Latvia is skewed towards high temperatures, therefore normal distribution is not suitable as the probability function. Therefore, in this work skewed Student-t distribution was used for temperature. Second, due to climate change temperature has increased. Therefore, it is more likely that heatwaves are identified in the recent years. To solve this problem, a trend was added to the mean of the probability distribution.

Based on the statistical model most extreme heat events between May and September in Latvia were identified. For the specific event return period is highly dependent on the analysed temperature variable. For example, during July 2021 heatwave record for the highest nighttime temperature was broken (record was broken once again on 6^th of August 2023). Such high lowest daily temperature on 21^st June has a return period of 4300 years. However, daily mean temperature reached on 21^st June 2021 has return period of 186 years and daily maximum temperature has return period of only 41 years.

The most intensive heatwave events were chosen for further analysis to investigate mechanisms that have caused them. Trajectories of air parcels that bring warm air into the region were identified using Lagrangian tracing backwards in time. Air temperature, potential temperature and humidity of air parcels was traced along the trajectories to identify the mechanisms behind the extreme heat. Modelling of the events was performed using WRF to establish issues that arise when forecasting such events.