Using deep-learning to forecast the magnitude and characteristics of urban heat island in Seoul Korea

Abstract

Urban heat island (UHi), a phenomenon involving increased air temperature of a city compared to the surrounding rural area, results in increased energy use and escalated health problems. to understand the magnitude and characteristics of UHi in Seoul and to accommodate for the high temporal variability and spatial heterogeneity of the UHi which make it inherently challenging to analyze using conventional statistical methods, we developed two deep learning models, a temporal UHi-model and a spatial UHi model, using a feed-forward deep neural network (Dnn) architecture. Data related to meteorological elements (e.g. air temperature) and urban texture (e.g. surface albedo) were used to train and test the temporal UHi-model and the Spatial UHi-model respectively. Also, we develop and propose a new metric, UHI-hours, that quantifies the total number of hours that UHI exists in a given area. our results show that UHi-hours is a better indicator of seasonal UHi than the commonly used index, UHi-intensity. consequently, UHi-hours is likely to provide a better measure of the cumulative effects of UHI over time than UHI-intensity. UHI-hours will help us to better quantify the effect of UHI on, for example, the overall daily productivity of outdoor workers or heat-related mortality rates.The world's population is increasing dramatically; new cities are being built, while existing cities are getting overpopulated. Currently, more than 50% of the world population resides in cities 1 . This increased development in urban areas has resulted in changes in urban morphology and surfaces, as well as an increase in the amount of anthropogenic heat released to the atmosphere 2 . The combined effect of such changes leads to higher air temperature in urban areas than in the surrounding rural or suburban areas, hence the formation of heat islands 3 . These heat islands, especially in summers, considerably decrease the outdoor air quality and trigger heat-related diseases and deaths in urban areas 4,5 . Moreover, extreme temperatures have been reported to be the leading cause of weather-related deaths 6 . This is particularly prevalent among the elderly (people of 65 years of age and above) and those with cardiovascular and respiratory health issues. For example, Paravantis et al. 7 analyzed the impact of temperature and heat waves on the number of deaths caused by cardiovascular and respiratory health issues in people above 65 years of age in Athens, Greece. They reported a U-shaped exposure-response curve, indicating reduced mortality rates at moderate temperatures and 20% and 35% increase in mortality rates at very low and very high temperatures, respectively. Furthermore, due to the increase in urban temperatures, there has been an increase in air-conditioning system usage, which in turn has led to high electricity demands and overall increased building energy consumption 8 . At the same time, passive cooling methods, such as natural and night ventilation, have become ineffective for the thermal comfort of building occupants 9 .Due to...

Synergies between urban heat island and heat waves in Seoul: The role of wind speed and land use characteristics

“…Following such guidelines, we chose the Neuggok station located in the northern area of Seoul at a longitude of 126.47˚E and latitude of 37.37˚N. The Neuggok weather station has also been used in several UHIrelated studies in Seoul [45][46][47]. UHII was therefore calculated as T urban −T rural with T urban representing air temperature values observed at the earlier mentioned ten stations and T rural representing air temperature values recorded at the Neuggok station.…”

A novel deep learning‐based integrated photovoltaic, energy storage system and electric heat pump system: Optimising energy usage and costs

“…Subsequently, it provides a solution to the shortcomings, as mentioned In regards to prediction techniques used to forecast building energy demand and RES energy generation, previous studies have mostly employed traditional predictive methods which are particularly limited in mapping complex behaviour. [45][46][47][48][49][50] For example, Perez et al 65 developed a model predictive control for real-time power generation for an electric system having a PV system and ESS, and used a mathematic process model to forecast PVgenerated energy. Another study by Lee et al 28 used short-term and very short-term day-ahead forecasting of PV-generated energy and developed a scheduling algorithm for an ESS-PV system.…”

“…[42][43][44][45] (b) ANN-based and empirical models also tend to overfit and offer less generalisation ability when used with large datasets. [45][46][47][48] Besides such problems associated with modelling approaches, there is also a limited number of studies that discuss efficient methods to combine EHP, PV, and ESS for energy optimisation.…”

“…29,30,[33][34][35][36]41 To overcome such limitations and improve prediction accuracy, we employ the DNN algorithm to forecast PV energy generation and EHP energy demand and subsequently determine the ESS charging/discharging schedules. [45][46][47][48][49][50] Another critical aspect of our study is related to the parameters used in forecasting EHP energy consumption and PV energy generation. Previous studies mentioned above have mainly used a large number of parameters to predict/forecast EHP energy demand and PV energy generation.…”

“…ANNs and empirical mathematical models employed in previous studies discussed above have several shortcomings: (a) Empirical mathematical models require user knowledge on the dataset and the working mechanisms of the model to ensure accuracy during model application 42‐45 . (b) ANN‐based and empirical models also tend to overfit and offer less generalisation ability when used with large datasets 45‐48 . Besides such problems associated with modelling approaches, there is also a limited number of studies that discuss efficient methods to combine EHP, PV, and ESS for energy optimisation.…”