Temperature Dataset Research Articles

Estimating ocean heat content from the ocean thermal expansion parameters using satellite data

Abstract. Ocean heat content (OHC) is a depth-integrated physical oceanographic variable used to precisely measure ocean warming. Because of the limitations associated with in situ conductivity, temperature, and depth (CTD) data as well as ocean reanalysis system products, satellite-based approaches have gained importance in estimating the daily to decadal variability of OHC over the vast oceanic region. Efforts to minimize the biases in satellite-based OHC estimates are needed to realize the actual response of the ocean to the brunt of climate change. In the current study, an attempt has been made to better implement the satellite-based ocean thermal expansion method to estimate OHC at 17 depth extents ranging from the surface to 700 m. To achieve this objective, artificial neural network (ANN) models were developed to derive thermosteric sea level (TSL) from a given dataset of sea surface temperature, sea surface salinity, geographical coordinates, and climatological TSL. The model-derived TSL data were further used to estimate OHC changes based on the thermal expansion efficiency of heat. Statistical analysis showed high correlation coefficients and low errors in the validation of model-derived TSL and OHC for the 700 m modeling depth (N 388 469, R 0.9926 and 0.9922, RMSE 1.16 m and 1.56 GJ m−2, MBE −0.19 m and −0.24 GJ m−2, MBPE −0.46 % and −0.03 %, MAE 0.76 m and 1.03 GJ m−2, and MAPE 2.34 % and 0.13 %) and nearly similar results at the remaining modeling depths. These results suggest that the proposed ANN models are capable of generating satellite-based daily OHC maps by covering both shallower and deeper oceanic regions of varying bathymetry levels (≥20 m). In addition, the first-ever attempt to estimate the ocean thermal expansion component (i.e., TSL) from satellite data was successful, and the model-derived TSL can be used to obtain high-end sea level rise products in the global ocean.

Spatio-temporal Variations of Heat Extremes across the Yangtze River Delta during 2001-2023 Based on Remotely Sensed Seamless Air Temperature.

Heat extremes become increasingly frequent and severe, posing adverse risks to public health and environment. Previous research on extreme heat mostly used meteorological observations or reanalysis data, which cannot well capture detailed spatial patterns. This study developed a seamless air temperature (Ta) dataset from remote sensing data to characterize the spatio-temporal variations of heat extremes in the Yangtze River Delta (YRD) from 2001 to 2023. First, the daily maximum Ta of cloud-free pixels was estimated through machine learning algorithms from MODIS land surface temperature (LST) and other remote sensing data. Then, gaps in the estimated Ta caused by cloud cover were filled using the Temporal Fourier Analysis (TFA) method, generating a seamless daily maximum Ta dataset. The remotely sensed Ta achieved an overall MAE of 1.11°C. Based on the remotely sensed Ta, six heat indices were calculated to characterize heat extremes, including heat days (HTD), effective accumulated high temperature (EAHT), heatwave frequency (HWF), cumulative heatwave days (HWD), maximum heatwave duration (HWMD) and average heatwave duration (HWAD). Heat extremes occurred frequently in the YRD, with obvious spatial variability. Southern basins experienced intense heat with high frequency and duration, while southern mountains and northern areas experienced weaker heat extremes. Urban areas have substantially more intense heat events than suburbs, attributed to urban heat island effect. 2022 recorded the most severe heat, with notable events also in 2013 and 2003. This study provides valuable insights into heat events in the YRD and serves as a reference for remote sensing research on heat events.

Interannual deep-water renewal in the tropical fjord of Kao Bay of Halmahera Island, eastern Indonesia, linked to ocean dynamics in western equatorial Pacific

Ocean dynamics in western equatorial Pacific (WEP) controls interannual deep-water renewal in the tropical fjord of Kao Bay, Halmahera Island of Indonesia. Processes such as boreal winter coastal upwelling at north New Guinea, New Guinea Coastal Undercurrent (NGCUC) and Halmahera Eddy (HE) were hypothesized to supply high density water at the vicinity of Kao Bay fjord to subsequently replenish the bottom of the fjord. To test the hypothesis, a large temporal dataset of temperature and salinity from HYCOM reanalysis model (1/12o) complemented by observations from TRITON T13 and T16 moorings and World Ocean Database (WOD) was employed within the NGCUC pathway from north New Guinea (T13) to Halmahera Island at T16 and offshore Kao Bay. Oceanic Niño Index (ONI) was used to unveil the connection between the coastal upwelling and ENSO. Additionally, tidal suction model was used to investigate when high density water from WEP can overflow the sill of Kao Bay. HYCOM simulation outputs were reasonably validated by the observations (RMSE: 0.37-1.67oC; 0.02-0.28 psu). The signal of the interannual variation of boreal winter coastal upwelling linked to ENSO variability (r=0.92) at the north New Guinea (at T13) was evident at offshore Kao Bay and T16, indicated by the concurrent of the shoaling of the 25oC isotherm between these locations. Pacific water mass reaching offshore Kao Bay and T16 resulting from the ocean teleconnection of the coastal upwelling was characterized as a mixture of North Pacific Tropical Water (NPTW) and South Pacific Tropical Water (SPTW). The upper layers of offshore Kao Bay and T16 showed similarity regarding a comparable mixture between NPTW and SPTW. The deep layers of these locations indicate the SPTW predominance (70-100% for offshore Kao Bay; 60% for T16).The comparable NPTW-SPTW mixture water at the upper layers of offshore Kao Bay fjord was identified to eligibly replenish the bottom of the fjord due to its temperature and salinity characteristics and its location within the range of local vertical tidal excursion based on tidal suction model – these two aspects of the intruding water were only achieved during boreal winter of normal and El Niño years. Due to this interannual nature, the deep layers of Kao Bay fjord can only be completely flushed via the cumulative effects of series of deep-water renewal intrusions spanning some El Niño years. This study has added an essential insight into the existing knowledge of deep-water renewal in Kao Bay fjord by demonstrating the more likely occurrence of interannual renewal than annual renewal, proposed by the early study of Van Aken and Verbeek (1988).

Weakening of subsurface ocean temperature seasonality over the past four decades

The seasonal cycle, responsible for much of the temperature variability in the upper ocean, exerts profound climatic and ecological influence. While surface intensification of temperature seasonality has been widely examined, changes beneath the ocean surface remain unknown. Here we analyze multiple ocean temperature datasets, revealing a robust, substantial weakening of subsurface seasonality by 5.7 ± 1.8% below the mixed layer in extratropical oceans since the 1980s. Using a hierarchy of climate models and an idealized diffusive model, we attribute this weakening to increased ocean heat uptake driven by rising greenhouse gases. This process strengthens upper ocean stratification, suppresses vertical mixing, and limits heat penetration into deeper ocean layers, resulting in a more quiescent subsurface ocean with reduced seasonal variability. Our findings highlight a new fingerprint of anthropogenic influence on subsurface ocean seasonality, with important implications for ocean biogeochemical processes and marine ecosystems.

Mobile Observation Field Experiment of Atmospheric Vertical Structure and Its Application in Precipitation Forecasts Over the Tibetan Plateau

AbstractWe carried out the first Mobile Field Observation Campaign of Atmospheric Profiles (MFOCAP) in the southeast Tibet and the Three‐River Source Region (TRSR) of the Tibetan Plateau (TP) by adopting two vehicle‐mounted integrated mobile observations (MO) system from July 18 to 30, 2021. Reliable MO data sets of air temperature (Ta), water vapor density (WVD) and relative humidity (RH) with high spatio‐temporal resolution over the TP were obtained and assimilated to improve precipitation forecast using the four‐dimensional variational (4DVAR) data assimilation (DA) method. The results show that Ta, WVD and RH profile data retrieved with the mobile microwave radiometer (MR) are credible over the TP. The atmospheric vertical structure measured by the mobile MR can reproduce the spatio‐temporal evolution characteristics of water vapor transport, temperature stratification and cloud structure. The distribution pattern of 24‐hr accumulated rainfall prediction with Ta profile DA was closer to measurements, and 6–12 hr forecasts for low to moderate rainfall in the central and western regions of Qinghai province were improved significantly. Data assimilation with air temperature retrievals from mobile MR observations were found beneficial for accurate simulation of water vapor transport, convergence and divergence of wind field, and upward motion associated with precipitation events. The finding of this study highlights the value of MR remote sensing observations in improving the rainfall monitoring and forecasts over the TP and downstream regions.

Evidence and Explanation for the 2023 Global Warming Anomaly

In 2023, the rapid increase in global temperature of around 0.25 °C caught the scientific community by surprise. Its cause has been investigated largely by exploring variations on a long-term trend, with little success. Building on previous work, this paper proposes an alternative explanation—on decadal timescales, observed temperature shows a complex, nonlinear response to forcing, stepping through a series of steady-state regimes. The 2023 event is nominated as the latest in the sequence. Step changes in historical and modeled global mean surface temperatures (GMSTs) were detected using the bivariate test. Each time series was then separated into gradual (trends) and rapid components (shifts) and tested using probative criteria. For sea surface, global and land surface temperatures from the NOAA Global Surface Temperature Dataset V6.0 1880–2022, the rapid component of total warming was 94% of 0.72 °C, 78% of 1.16 °C and 74% of 1.93 °C, respectively. These changes are too large to support the gradual warming hypothesis. The recent warming was initiated in March 2023 by sea surface temperatures (SSTs) in the southern hemisphere, followed by an El Niño signal further north. Global temperatures followed, then land. A preceding regime shift in 2014 and subsequent steady-state 2015–2022 was also initiated and sustained by SSTs. Analysis of the top 100 m annual average ocean temperature from 1955 shows that it forms distinct regimes, providing a substantial ‘heat bank’ that sustains the changes overhead. Regime shifts are also produced by climate models. Archived data show these shifts emerged with coupling of the ocean and atmosphere. Comparing shifts and trends with equilibrium climate sensitivity (ECS) in an ensemble of 94 CMIP5 RCP4.5 models 2006–2095 showed that shifts had 2.9 times the influence on ECS than trends. Factors affecting this relationship include ocean structure, initialization times, physical parameters and model skill. Single model runs with skill ≥75 showed that shifts were 6.0 times more influential than trends. These findings show that the dominant warming mechanism is the sudden release of heat from the ocean rather than gradual warming in the atmosphere. The model ensemble predicted all regime changes since the 1970s within ±1 year, including 2023. The next shift is projected for 2036, but current emissions are tracking higher than projected by RCP4.5. Understanding what these changes mean for the estimation of current and future climate risks is an urgent task.

Spatial and Temporal Variations of Surface Air Temperature (1962–2022) across Physiographic Regions in the Koshi Basin, Nepal

This study analyzed the seasonal and annual trends of maximum temperature (Tmax), minimum temperature (Tmin), and mean temperature (Tavg) across the physiographic regions (Himalaya, High Mountain, Middle Mountain, Siwalik, and Terai) of the Koshi Basin using observed data from 23 stations between 1962 and 2022. Missing temperature data were filled using the lapse rate method. The Mann-Kendall (M-K) test was employed to assess the consistency of the temperature dataset. The analysis revealed distinct regional and seasonal temperature trends in the Koshi Basin. The seasonal variations were prominent, especially in the Monsoon and Winter. In the Middle Mountain and High Mountain regions, Tmax increased significantly during the Pre-monsoon season, while Tmin decreased substantially during the Monsoon and Winter. Conversely, the Siwalik and Terai regions experienced more pronounced cooling Himalayas, especially during the Monsoon and Winter. Overall, the Himalayas and High Mountain regions exhibited a cooling trend until the late 1990s, followed by a warming trend. The Middle Mountain region demonstrated similar patterns, with a significant temperature increase after the 1990s. The Siwalik and Terai regions experienced a general cooling trend, although the Siwalik region exhibited some fluctuations. The significant regional variations in temperature trends were the key findings. Similarly, the Himalayas and High Mountains showed a cooling-warming shift in the late 1990s. All these findings underscore the importance of considering both regional and seasonal factors when studying temperature trends in the Koshi Basin.

Hierarchical Graph Attention Network with Positive and Negative Attentions for Improved Interpretability: ISA-PN.

With the advancement of deep learning (DL) methods in chemistry and materials science, the interpretability of DL models has become a critical issue in elucidating quantitative (molecular) structure-property relationships. Although attention mechanisms have been generally employed to explain the importance of molecular substructures that contribute to molecular properties, their interpretability remains limited. In this work, we introduce a versatile segmentation method and develop an interpretable subgraph attention (ISA) network with positive and negative streams (ISA-PN) to enhance the understanding of molecular structure-property relationships. The predictive performance of the ISA models was validated using data sets for aqueous solubility, lipophilicity, and melting temperature, with a particular focus on evaluating interpretability for the aqueous solubility data set. The ISA-PN model enables the quantification of the contributions of molecular substructures through positive and negative attention scores. Comparative analyses of the ISA, ISA-PN, and GC-Net (group contribution network) models demonstrate that the ISA-PN model significantly improves interpretability while maintaining similar accuracy levels. This study highlights the efficacy of the ISA-PN model in providing meaningful insights into the contributions of molecular substructures to molecular properties, thereby enhancing the interpretability of DL models in chemical applications.

The Winter Foehn Footprint Across McMurdo Dry Valleys of Antarctica Using a Satellite‐Derived Data Set‐AntAir v1.0

AbstractContinental‐scale mosaics of satellite‐based surface brightness temperature from thermal infrared band measurements and derived near‐surface air temperatures from geostatistical modeling provide new opportunities for understanding wintertime Foehn wind warming and its potential impacts on the valley floor warming. We have detected and assessed Foehn signatures using a combined data analysis approach from previously developed and validated Antarctic‐wide near‐surface Air temperature data set (AntAir), automatic weather stations from the McMurdo Dry Valleys, and regional climate model simulations at 10 km spatial grid resolution. Self‐organizing maps and data compositing methods on regional climate model outputs provided meteorological context for the AntAir‐derived surface climate information. We conclude that AntAir is suitable for surface climatological analyses and improvements are underway to enhance the spatial resolution to sub‐kilometer grid scales. Finally, by applying a Foehn detection algorithm over 13 years, we present the spatial climatological footprints of Foehn‐induced warming across the Dry Valleys of Antarctica for the first time over the austral winters.

Fast generation of high-dimensional spatial extremes

Widespread extreme climate events cause many fatalities, economic losses and have a huge impact on critical infrastructure. It is therefore of utmost importance to estimate the frequency and associated consequences of spatially concurrent extremes. Impact studies of climate extremes are severely hampered by the lack of extreme observations, and even large ensembles of climate simulations often do not include enough extreme or record-breaking climate events for robust analysis. On the other hand, weather generators specifically fitted to extreme observations can quickly generate many physically or statistically plausible extreme events, even with intensities that have never been observed before. We propose a Fourier-based algorithm for generating high-resolution synthetic datasets of rare events, using essential concepts of classical modelling of (spatial) extremes. Here, the key feature is that the stochastically generated datasets have the same spatial dependence as the observed extreme events. Using high-resolution gridded precipitation and temperature datasets, we show that the new algorithm produces realistic spatial patterns, and is particularly attractive compared to other existing methods for spatial extremes. It is exceptionally fast, easy to implement, scalable to high dimensions and, in principle, applicable for any spatial resolution. We generated datasets with 10000 gridpoints, a number that can be increased without difficulty. Since current impact models often require high-resolution climate inputs, the new algorithm is particularly useful for improved impact and vulnerability assessment.

The Loss of Beneficial Thermal Priming on Global Coral Reefs.

Warm-season marine heatwaves (MHWs) have greatly increased in frequency, severity, and extent over the last few decades, driving more frequent and severe coral bleaching episodes. Given the grave near-term threat to coral reefs imposed by MHWs, it is important to assess the mechanisms by which corals may acquire higher thermal tolerance. Recent field and laboratory studies have demonstrated that exposure to sublethal heat stress, known as "priming," can reduce bleaching susceptibility during a subsequent MHW. Little is known, however, about how often priming conditions occur, and how effective those conditions may be at protecting coral reefs. We employed a global historical coral bleaching database and a high-resolution sea surface temperature dataset to assess the frequency of priming and examine its effect on coral bleaching sensitivity on a global scale. The analysis showed that coral reefs in parts of the western to central tropical Pacific experienced priming on average over twice a decade and had a higher likelihood of priming protection. Mixed-effects regression models indicated that priming conditions could mitigate coral bleaching response by up to 12% in advance of a moderate MHW. However, the protective effect of priming decreased, and even became harmful, with more severe MHWs. We detected spatial variations in priming frequency that could provide insight for conservation planning and explain some variations in bleaching sensitivity to MHWs. Even so, our findings suggest that thermal priming will not be sufficient to protect most coral reefs from MHWs in the future, without substantial efforts to mitigate climate change.

Neural Bayes Estimators for Irregular Spatial Data using Graph Neural Networks

Neural Bayes estimators are neural networks that approximate Bayes estimators in a fast and likelihood-free manner. Although they are appealing to use with spatial models, where estimation is often a computational bottleneck, neural Bayes estimators in spatial applications have, to date, been restricted to data collected over a regular grid. These estimators are also currently dependent on a prescribed set of spatial locations, which means that the neural network needs to be re-trained for new data sets; this renders them impractical in many applications and impedes their widespread adoption. In this work, we employ graph neural networks (GNNs) to tackle the important problem of parameter point estimation from data collected over arbitrary spatial locations. In addition to extending neural Bayes estimation to irregular spatial data, the use of GNNs leads to substantial computational benefits, since the estimator can be used with any configuration or number of locations and independent replicates, thus amortising the cost of training for a given spatial model. We also facilitate fast uncertainty quantification by training an accompanying neural Bayes estimator that approximates a set of marginal posterior quantiles. We illustrate our methodology on Gaussian and max-stable processes. Finally, we showcase our methodology on a data set of global sea-surface temperature, where we estimate the parameters of a Gaussian process model in 2161 spatial regions, each containing thousands of irregularly-spaced data points, in just a few minutes with a single graphics processing unit.

Multiple timescale variations in fronts in the Seto Inland Sea, Japan

Abstract. The Seto Inland Sea (SIS) is a critical semi-enclosed coastal sea in Japan, characterized by intricate coastlines and narrow straits that give rise to various fronts. Despite extensive research on tidal fronts, knowledge gaps persist regarding their spatiotemporal dynamics, particularly in certain poorly documented regions. Additionally, the understanding of thermohaline fronts, which emerge during winter, requires further investigation. We aimed to enhance our understanding of tidal and thermohaline fronts in the SIS by analyzing their dynamic processes, including intra-tidal and spring–neap tidal cycles, seasonal variations, and anomalous frontal variability. Using a gradient-based algorithm with an advanced contextual feature-preserving median filter, we processed the high-resolution sea surface temperature dataset to detect and quantify tidal and thermohaline fronts. Our analysis revealed the presence of numerous tidal fronts, predominantly influenced by the M2 tide, across the SIS, with substantial spatial variations in amplitude due to complex coastlines and narrow straits. Intra-tidal movements of tidal fronts corresponded to ebb and flood currents, while spring–neap tidal cycles and seasonal shifts influenced frontal positions and intensities. Additionally, thermohaline fronts were identified in certain regions during winter, characterized by large horizontal temperature and salinity gradients. This study enhances the understanding of tidal and thermohaline fronts in the SIS, emphasizing the importance of intra-tidal and wind-driven influences on frontal dynamics. However, limited observational coverage and resolution emphasize the need for further research to explore long-term temporal changes and better grasp the influences of ambient currents and wind patterns. Such insights are vital for effective coastal management and environmental monitoring in the SIS region.

Temperature-Based Long-Term Stabilization of Photoacoustic Gas Sensors Using Machine Learning.

In this study, we address the challenge of estimating the resonance frequency of a photoacoustic detector (PAD) gas cell under varying temperature conditions, which is crucial for improving the accuracy of gas concentration measurements. We introduce a novel approach that uses a long short-term memory network and a self-attention mechanism to model resonance frequency shifts based on temperature data. To investigate the impact of the gas mixture temperature on the resonance frequency, we modified the PAD to include an internal temperature sensor. Our experiments involved multiple heating and cooling cycles with varying methane concentrations, resulting in a comprehensive dataset of temperature and resonance frequency measurements. The proposed models were trained and validated on this dataset, and the results demonstrate real-time prediction capabilities with a mean absolute error of less than 1 Hz for frequency shifts exceeding 30 Hz over four-hour periods. This approach allows continuous, real-time tracking of the resonance frequency without interrupting the laser operation, significantly enhancing gas concentration measurements and contributing to the long-term stabilization of the sensor. The results suggest that the proposed approach is effective in managing temperature-induced frequency shifts, making it a valuable tool for improving the accuracy and stability of gas sensors in practical applications.

Simulation and projection of photovoltaic energy potential over a tropical region using CMIP6 models

The study examines the potential impact of climate change on photovoltaic energy (PV) in Nigeria. Solar radiation and temperature datasets from 13 regional climate models (CMIP6) for 2015–2099 were used to evaluate the photovoltaic energy under moderate (SSP245) and high (SSP585) emission scenarios for near-future (2023–2053), mid-future (2054–2084), and far-future (2084–2099) periods. The precision of the models for the simulation of PV energy was validated with MERRA-2 reference data using the compromise programming index (CPI). Models with the lowest CPI were selected for regional PV energy projections. The findings showed varying numbers of increase and decrease projected changes across the four regions under SSP245 and SSP585 scenarios for the near, mid and far future timescales. Specifically, in the SSP245 scenario, the model with lowest CPI was the CMCC-CESM2 model, it projected a decrease in PV energy in the Sahel (−2.30), an increase in the Guinea Savannah (+2.80), the Rainforest (+1.20) and the coastal region (+4.80) for the far future period (2085–2099). In the SSP585 scenario, the AWI-CM-1.1-MR model projected a decrease in the Sahel region (−4.60), while the MPI-ESM1-2-LR model projected an increase in the Guinea Savannah region (+1.80), and the ACCESS-CM2 model projected an increase in the Rainforest (+10.20) and Coastal regions (+13.20) for the far-future. All values in the parentheses are measured in watts-hour per square-meters. The projected changes in PV energy revealed the need for a regional-specific approach to the planning and implementation of energy transition mix in Nigeria.

Estimating the vertical profile of water quality variables in reservoirs: Application of remotely sensed data and machine learning techniques

Water quality assessment and management of reservoirs depend on accurate, large-scale, and continuous monitoring of the vertical profile of Water Quality Variables (WQVs). Remote sensing data have been widely used to retrieve high spatiotemporal water quality data; however, their application has practically been limited to evaluating surface WQVs. In this paper, a novel and efficient approach is introduced for assessing the profile of WQVs in reservoirs that depend on stratification, by taking into account the shape of profile as prior knowledge. First, an appropriate function is fitted to the WQVs vertical profile, and second, the parameters of that function as representative of the WQVs vertical profile are estimated using machine learning techniques. The model's inputs are day, maximum depth of point, and remote sensing data. Finally, PAWN sensitivity analysis is applied to show the extent to which each input influences different parts of the vertical profile. This method is applied in the Wadi Dayqah Reservoir, the largest dam in Oman, to evaluate water temperature, dissolved oxygen, pH, and chlorophyll-a profile. The results show that the predicted profiles are properly representative of in situ measurements, with a mean absolute error of 0.28 °C, 0.25 mg/L, 0.052, and 0.33 μg/L on test data sets of water temperature, dissolved oxygen, pH, and chlorophyll-a, respectively. Finally, PAWN sensitivity analysis illustrates that satellite data not only influence the parameters representing surface WQVs but also contribute to the estimation of other curve parameters.

Long-term trends of heat waves and ecosystem responses in Jiaozhou Bay, the Yellow Sea

With global warming, extreme weather frequently occurs, yet the consequences remain unexplored. A total of 156 heat waves and their characteristics were detected on the basis of a temperature dataset from 1954 to 2022 in Jiaozhou Bay (JZB). The increment is 0.62 times decade−1 for the number of heat waves, 6.65 days decade−1 for the sum of participating heat wave days and 0.66 days for the duration of each heat wave. The intensity of heat waves showed regular fluctuations with progressively shorter periods. Based on the dataset of 12 stations in JZB from 2003 to 2022, the zooplankton abundance was significantly greater during heat waves, which was strongly attributed to the greater abundance of copepods and gelatinous zooplankton during heat waves. However, the responses of plankton to heat waves were seasonally heterogeneous. Our study provides new insight into and a scientific basis for understanding the effects of heat waves on offshore plankton ecosystems.

The role of antecedent southwest summer monsoon rainfall on the occurrence of premonsoon heat waves over India in the present global warming era

Global warming has significantly increased the risk of heat waves (HWs) globally, with India being particularly vulnerable during the summer months (March-June; MAMJ). This study investigated the critical relationship between Indian summer monsoon rainfall (ISMR) and the occurrence of premonsoon HWs in subsequent years across the Indian subcontinent. It has been hypothesized that droughts during the ISMR could lead to more frequent HWs in the following MAMJ period. Using the Indian Meteorological Department's (IMD) gridded observed surface air daily maximum temperature (Tmax) dataset for the period 1951–2023, we analyzed the climatic patterns, interannual variability (IAV), and coefficient of variation (CV) of Tmax across India. The analysis compares two distinct periods: 1951–1999 (P1) and 2000–2023 (P2), with focus on Tmax trends and HW duration, distinguishing between short-duration HWs (SHWs, 2 days) and long-duration HWs (LHWs, 5 days or more). A key purpose of this study is to examine the relationship between the preceding all India summer monsoon rainfall (AISMR) and the occurance of various types of HW in the subsequent premonsoon season. In particular extreme AISMR events, such as droughts or excess rainfall, influence HW occurrence. The findings reveal a significant rise in Tmax across many regions of India during the MAMJ period, with the highest temperatures (> 37 °C) observed in northwestern, central, and eastern coastal areas. Northern India, particularly the Himalayan region, exhibits a greater interannual variability in Tmax, with June showing the most pronounced fluctuations. The study also highlights an increase in the frequency and intensity of HWs, especially in central and southern India, with the Chandigarh-Haryana-Delhi region recording the highest occurrences. A critical finding is the strong inverse relationship between the AISMR and conditions in the subsequent premonsoon season. Specifically, drought in the antecedent AISMR results in reduced soil moisture, which is strongly associated with higher premonsoon Tmax and an increased frequency of extreme heat events across India, particularly in regions prone to severe heat during this season. Drought conditions during AISMR are closely linked to higher HW frequencies in the following summer, especially in the central, northeast-central, and east-coastal regions. The frequencies of HW days, SHWs, and LHWs are significantly greater in years following AISMR droughts than in those following excess rainfall, indicating that drought years are more likely to lead to widespread HW activity. Despite the overall warming trends, some regions, such as the Indo-Gangetic Plain and parts of the Himalayan region, show cooling trends, although these trends are less widespread. The onset of the monsoon in June tends to reduce the intensity and spatial extent of warming, particularly in the central and eastern coastal regions, although significant HW trends persist in northwestern India and along the east coast. This study underscores the crucial role of AISMR in influencing HW events across India and highlights the need for adaptive strategies that account for the interactions between monsoon rainfall and HW risk, providing valuable insights for mitigating the impacts of HWs in the context of global warming.

Quantifying thermo-physiological stress exposure during extreme events: Developing a morphological case study in Istanbul

This study aimed to determine thermal comfort conditions during Extreme Heat Events (EHE), particularly focus on Very Hot Day and Heat Wave Event in Istanbul. The analysis was based on a comprehensive dataset of hourly temperature records collected from six meteorological stations over the period 1991–2021. Subsequently, Land Surface Temperature (LST) was mapped on the basis of EHE days. Finally, specific areas with different street morphologies values from the referenced LST were obtained, enabling the evaluation of heat stress for the EHE day and for the period between 2006 and 2021. The study identified August 5th, 2021 as the day of extreme heat stress in Istanbul. LST obtained for August 5th was divided into three groups: low (23.3 °C), moderate (26 °C), and high (29.3 - 30 °C), and findings regarding heat stress were obtained through thermal indices. Physiologically Equivalent Temperature (PET) values on August 5th, 2021, were significantly higher than the 15-year average, showing a shift from moderate and slight heat stress to strong and extreme heat stress levels. PET values indicated strong heat stress, especially around noon, reaching up to 40.5 °C in lower LST, and extreme heat stress level of 45.7 °C in the afternoon in higher LST. Extreme thermal conditions were generally observed on the right side of streets oriented between 105° and 165° in both regions.

A method for filling missing values in multivariate sequence bidirectional recurrent neural networks based on feature correlations

Multivariate real-life time series data often contain missing values. These missing values often affect subsequent prediction tasks. Traditional imputation methods generally consider only some of the characteristics of multivariate time series data. This can easily lead to inaccurate filling results. In this paper, a feature correlation-based bidirectional recurrent network (BRNN-FR) is proposed to solve the problem of missing values in multivariate sequence data. First, this method involves the design of a bidirectional prediction network based on time intervals and the use of forward and reverse time series information between data points to obtain the characteristics of data changes with time to the greatest extent. Second, considering the correlation between features, a combined feature selection strategy based on the Pearson correlation coefficient and mutual information was proposed. A multiple regression model was established to predict between features. Finally, a bidirectional network ensemble filling algorithm based on the relationships between features is established to predict missing values. Comprehensive experiments on four public datasets show that the mean absolute error (MAE), root mean square error (RMSE) and maximum R2 value (R2_score) of the BRNN-FR algorithm in the direct imputation test are better than those of the other comparison methods in most cases. BRNN-FR also achieved a better area under the curve (AUC) in the indirect comparison experiment of two classifications of in-hospital death after filling the medical dataset. Using the AIR air quality dataset and the power transformer temperature dataset from the ETTH1 interpolation regression to predict the next 3 hours and 6 hours of average numerical results, most of the optimal regression results are obtained.