Investigating important urban characteristics in the formation of urban heat islands: a machine learning approach

Abstract

IntroductionThe warming trend of US cities becomes significant since the late 1970s and its rate and magnitude of this trend severed during the late 1990s [1][2][3]. Land use changes due to urbanization can modify the energy balance in cities, and in turn, this affects the urban thermal environment, resulting in the urban heat islands (UHIs) phenomenon, meaning urban areas have higher air and surface temperature than their rural surroundings [4][5][6]. Long recognized in many disciplines of physical science, such as geography as an important climatological phenomenon, however, the UHIs has only recently gained serious attention in the field of urban planning [7]. As UHIs occur as a result of land cover transformation, mainly replacement of natural vegetation and natural land cover by impervious surfaced associated with urban land uses, key questions for the discipline of urban planning include what are the most contributing urban physical characteristics to AbstractDespite the urban heat islands phenomenon has long been recognized as a major urban environmental problem, it was not until recently that this urban phenomenon gained attention from the discipline of urban planning. To integrate the findings of the urban heat islands research into the planning practice, the relationship between land surface temperatures and urban physical and socioeconomic characteristics should be addressed at the planning relevant spatial scale, a land parcel. Using a parcel as a unit of analysis, this study proposed to use a machine learning approach to identify important variables in the formation of urban heat islands in Indianapolis, Indiana. Applying random forest method to planning zones, this study identified planning zone specific urban physical and socioeconomic characteristics that are important for the interpretation of urban heat islands phenomenon of Indianapolis, Indiana. The main contribution of this study is twofold: to integrate urban physical and socioeconomic characteristics into a land parcel for the better interpretation of the result of urban heat islands study into planning practice and to apply machine learning approach to identify highly determinant variables in the formation of urban heat islands. Yoo J Big Data (2018) Despite its commonly acknowledged importance as a major urban environmental issue, above stated questions were primarily investigated and addressed by the discipline of geography in the context of remote sensing and the assessment of social vulnerability. Moreover, the UHIs phenomenon has not been properly addressed by urban planners even as they have tried to become more environmentally conscious in both their thinking and practice [8]. The Indianapolis Metropolitan Statistical Area (MSA), which is the main subject of this study, is one of those US cities where UHIs have grown in largest numbers in recent years, but both state and local government have not yet to provide any concrete and effective solutions [2]. RESEARCHThis lack of attention and response to UHIs and its environmental and...

Multidisciplinary Understanding of the Urban Heating Problem and Mitigation: A Conceptual Framework for Urban Planning

“…The types of zoning in urban areas influence temperature in association with those of zoning in adjacent areas [ 26 , 27 , 28 , 29 , 30 ]. Commercial zones show higher UHI intensity than do residential zones [ 27 , 31 ]. The effect of UHI is especially high in areas where high-rise buildings are closely situated [ 9 , 19 ].…”

“…Others suggest that, in a cluster of buildings with significant height variation, buildings absorb more solar radiation, thus increasing the temperature of the area [ 33 ]. High population density in a high-density area also contributes to the rise in temperature [ 31 , 34 ]; in a small, high-density area, poor air circulation can increase nighttime temperatures [ 35 , 36 ].…”

Identifying Surface Urban Heat Island Drivers and Their Spatial Heterogeneity in China’s 281 Cities: An Empirical Study Based on Multiscale Geographically Weighted Regression

“…Previous empirical research models for SUHI driver analysis can generally be divided into two main categories. The first type includes global-scale analysis methods, such as correlation coefficients [28,35,36], ordinary least square (OLS) [37][38][39], generalized additive model (GAM) [40], and various models for machine learning [41][42][43]. The biggest problem with these methods is that they cannot adequately analyze the spatial variations of SUHI drivers, making them feasible for small regional studies but leading to obvious bias at large spatial scales, such as in China.…”

A generalizability analysis of a data-driven method for the Urban Heat Island phenomenon assessment

“…Nutkiewicz et al [8] developed a methodology to characterize and model the energy performance of buildings at multiple spatial and temporal scales. Regarding the domain of urban climates, a handful of machine-learning approaches have been developed in order to predict temperatures in specific urban settings [9][10][11][12][13]. For instance, Vulova et al [10] developed a deep learning approach to discover hot spots in Berlin, Germany.…”