Search

Atmospheric extreme events cause severe damage to human societies and ecosystems. The frequency and intensity of extremes and other associated events are continuously increasing due to climate change and global warming. The accurate prediction, characterization, and attribution of atmospheric extreme events is, therefore, a key research field in which many groups are currently working by applying different methodologies and computational tools. Machine learning and deep learning methods have arisen in the last years as powerful techniques to tackle many of the problems related to atmospheric extreme events. This paper reviews machine learning and deep learning approaches applied to the analysis, characterization, prediction, and attribution of the most important atmospheric extremes. A summary of the most used machine learning and deep learning techniques in this area, and a comprehensive critical review of literature related to ML in EEs, are provided. The critical literature review has been extended to extreme events related to rainfall and floods, heatwaves and extreme temperatures, droughts, severe weather events and fog, and low-visibility episodes. A case study focused on the analysis of extreme atmospheric temperature prediction with ML and DL techniques is also presented in the paper. Conclusions, perspectives, and outlooks on the field are finally drawn. [ABSTRACT FROM AUTHOR]

Frost on agricultural products in spring imposes heavy financial losses to agriculture particularly in northwest of Iran's orchards. Frost early warning is an effective way in preventing frost risk in apple orchards. This paper aims to validate frost early warning system in apple orchards of northwestern Iran by predicting flowering date of apple tree and its combination with WRF (Weather Research and Forecasting) model simulations of the 2-m temperature. To this end, stepwise multiple linear regression model was used for predicting flowering date; therefore thermal variables data including mean temperature, mean maximum temperature, mean minimum temperature, last frost date, heat wave duration, heat wave intensity, GDD and maximum standard temperature data larger than 1 (Z > 1) for the period of 2007–2016 (from March 1st until flowering date) in Kahriz agrometeorological station in northwest of Iran were calculated and added to the model. Also, to evaluate the accuracy of WRF model simulations, the 72-h simulations of the 2-m air temperature for internal computational grid in northwestern Iran (West Azerbaijan Province) in 17 synoptic stations were compared to minimum air temperature observed in the stations. Results revealed that the proposed stepwise regression method is a relatively precise model with a mean absolute error of 2.7 days between the predicted and observed flowering dates. Findings were also indicative of an acceptable accuracy of 72-h minimum air temperature simulations of WRF model in the study area. [ABSTRACT FROM AUTHOR]

Drought is a natural phenomenon that occurred in Central Europe in the past, but it is becoming a more serious problem due to the changes in its occurrence and intensity related to climate change. The main aim of this paper was to determine changes in air temperature and precipitation amounts in Slovakia from 1931 to 2020 and to identify changes in drought occurrence and intensity using the Standardised Precipitation and Evapotranspiration Index (SPEI). Cluster analysis was used to determine regions with similar drought occurrences. For each of the identified five clusters, drought characteristics were determined and compared between two reference periods 1961–1990 and 1991–2020. While a decrease in the number of months with SPEI-1 below − 1 was observed in autumn, spring and summer months showed a drying tendency. Overall, we can say that we observed a shift in drought occurrence from autumn and winter months to spring and summer months. This is an important finding for the agriculture or forestry because it affects work management and planning. The clearest tendency in drought events was observed in the western part of Slovakia covering areas with agriculturally intensive land use. Besides prolonging drought events, there was also a higher accumulated deficit for each event. The increase in drought intensity was not spatially consistent over the cluster area, mostly reaching a slight decline of about − 0.1 to − 0.3, which means slight intensification of drought periods in 1991–2020. [ABSTRACT FROM AUTHOR]

We evaluated the performance of the Thornthwaite (ThW) method using two gridded climate datasets to estimate monthly average daily potential evapotranspiration (PET). The PET estimated from two gridded series were compared to PET and to reference evapotranspiration (ETo) determined, respectively, through the ThW and Penman-Monteith model parameterized on Food and Agriculture Organization–Irrigation and Drainage paper No 56 (PM-FAO56) using data from weather stations. The PET by ThM was based on monthly air temperature series (1961–2010) from two gridded datasets (Global Historical Climatology Network-GHCN and University of Delaware-UDel) and 21 weather stations of the National Institute of Meteorology (INMET) located in Southeastern Brazil. The ETo PM-FAO56 used monthly climate series (1961–2010) on sunshine duration, air temperature, relative humidity, and wind speed from weather stations of the INMET. The PET estimated using UDel gridded series was better overall performance than the GHCN series. Differences in altitude, latitude, and longitude were the main geographic factors determining the performance of the PET estimates using gridded climate series. Depending on the factors, some locations require bias correction, especially locations more than 10 km away from the grid point. The gridded datasets are an alternative for locations without climatic series data or with low-quality non-continuous data series. [ABSTRACT FROM AUTHOR]

Abnormal changes in air temperature cause natural disasters such as droughts, hailstorms, and storms, thereby affecting the normal lives of human beings. Consequently, timely and accurate air temperature prediction is essential for human production and livelihood. Traditional air temperature prediction methods are less accurate and less consider the spatial relationship between air temperature in different regions. In this paper, we propose a new deep learning model, convolutional long short-term memory based on channel attention and spatial autocorrelation (ConvLSTM-CASA), which focuses on the spatial correlation between ambient air temperatures and can effectively capture the interaction of air temperatures in different regions. The results show that the ConvLSTM-CASA model has an average R2 of 0.954 and MSE of 5.245 for pentad-mean temperature prediction over the Yangtze River basin. Compared with baseline forecasting models, the MSE accuracy by the ConvLSTM-CASA model improved by 72.45%, 48.95%, 48.97%, 47.79%, and 22.63% over the decision tree regression (DTR), multiple linear regression (MLR), random forests (RF), long short-term memory (LSTM), and ConvLSTM models, respectively. The ConvLSTM-CASA model is expected to outperform the ConvLSTM model over 90% of the area, suggesting robust prediction skill improvement over space. The ConvLSTM-CASA model provides new insights for data-driven pentad-mean air temperature prediction by including elaborate channel and spatial feature modeling, which aid individuals in comprehending the intricate patterns of air temperature fluctuations. [ABSTRACT FROM AUTHOR]

Currently, changes in rainfall regimes have intensified biomass burning in central Brazil. These fires, resulting mainly from anthropic action, emit large amounts of gases and aerosols into the atmosphere. The emission of these constituents in the dry season can drastically alter the functioning of ecosystems. In this work, it was verified that the high concentrations of Black Carbon (BC) have a direct relationship with the air temperature and influence the carbon capture process. This study was the result of a long-term campaign carried out in the northern region of the Brazilian Pantanal, between 2017 and 2019. Measurements of CO2 fluxes were evaluated together with measurements of temperature, radiation and BC concentrations. The results showed that in some episodes, the increase in BC concentration inhibits photosynthetically active radiation and increases air temperature. As a consequence of the increase in BC concentration, a reduction in carbon capture was observed. In a specific episode in October 2017, this work estimated a deficit of 1.65 kg.m−2 in CO2 capture due to the emission of BC in the Pantanal. [ABSTRACT FROM AUTHOR]

The annual, seasonal, and monthly trends of air temperatures were analyzed for thirteen urban and five rural meteorological stations in Kazakhstan for the 1963–2020 period. The non-parametric Mann–Kendall (M–K) rank correlation and Sen's slope estimator methods and the parametric least-squares linear regression (LSLR) were used to determine whether there were positive or negative statistically significant trends in mean, average maximum, and minimum air temperature time series along with diurnal temperature ranges (DTRs), and temperature differences between five large and small cities. In addition, Kazakhstan's annual and seasonal air temperature series were analyzed in terms of autocorrelation (serial correlation) coefficients. Coefficients of variations indicated that mean annual temperature variability is the highest in northern cities. Results of the M–K trend test indicated that the highest and lowest increases in the mean air temperatures were observed in spring and autumn, respectively. The magnitudes of the significant increasing trends in annual air temperature ranged between 0.23 °C/decade at Karagandy and 0.54 °C/decade at Kyzylorda. Annual and seasonal diurnal temperature ranges (DTRs) reveal insignificantly decreasing and increasing trends at most of the stations characterized by urbanization. According to the results of both M–K and LSLR tests, annual and winter air temperature differences of some station pairs tend to significantly increase, which may mean that the differences in the calculated temperature range between large and small cities might have widened significantly. However, significance test for the calculated autocorrelation coefficients of the annual and seasonal air temperature data showed that most of the series clearly appear as a low-frequency variability on the significantly increased long-term averages. [ABSTRACT FROM AUTHOR]