Study On The Influence Of Urban Morphology On Annual And Diurnal Surface Temperature Variation In Local Climate Zones

The rapid development of urbanization has led to changes in land cover,affecting the surface energy balance and thus having a significant impact on local urban climate change,as well as exacerbating urban ecological problems such as environmental pollution and the urban heat island(UHI)effect,which poses a serious threat to the social environment and people’s health.The annual and diurnal cycle of urban land surface temperature(U-LST)and its influencing factors are crucial for improving the urban thermal environment,urban planning,and sustainable development.In recent years,scholars have used thermal infrared remote sensing data to study the multi-temporal changes in surface temperature,but a single satellite-borne thermal infrared sensor cannot provide data with both high spatial and temporal resolution simultaneously.In addition,urban land surface temperature is influenced by two-dimensional and three-dimensional urban morphological parameters(UMPs),but research on the relationship between UMPs and surface temperature in different local climate zones(LCZs)is still lacking.This study explores how urban morphology affects annual and diurnal surface temperature cycles under the background of local climatic zoning,focusing on the central areas within and around the fifth ring road of Beijing.Firstly,the shortcomings of the classic Spatial-Temporal Integrated Temperature Fusion Model(STITFM)algorithm for integrating multi-source remote sensing data of land surface temperature are analyzed.An improved Flexible Spatiotemporal Data Integrated Temperature Fusion Model(FSDITFM)is proposed,which integrates Landsat-8 surface temperature inversion,Moderate-resolution Imaging Spectroradiometer(MODIS)diurnal surface temperature data,and Fengyun-2F(FY-2F)hourly surface temperature data to generate 30-meter resolution surface temperature data on an hourly basis.The fusion results are evaluated for accuracy using actual meteorological station data and actual urban land surface temperature(U-LST)data.The Landsat LST is used for daytime U-LST verification,and the Advanced Spaceborne Thermal Emission and Reflection Radiometer(ASTER)is used for nighttime verification.By calculating surface temperature characteristic parameters,the spatial and temporal distribution characteristics of urban surface temperature in 15 different local climatic zoning types in the study area in 2018are analyzed from the annual,diurnal,and hourly perspectives.Finally,based on local climatic zoning theory and using high-resolution remote sensing and multi-source geospatial data,two-and three-dimensional urban morphology parameters are extracted,and statistical methods such as Pearson correlation analysis and e Xtreme Gradient Boosting(XGBoost)regression analysis are used to investigate the effects of two-and three-dimensional urban morphology parameters on the annual and diurnal cycles of surface temperature under the background of local climatic zoning.The main conclusions of this study are as follows:(1)Qualitative and quantitative analysis was conducted on the surface temperature prediction results of the STITFM and FSDITFM models using both actual meteorological station data and U-LST data.The accuracy evaluation results based on actual U-LST data indicate that the FSDITFM model(r=0.920,RMSE=0.4 K,MAE=0.3 K)outperforms the STITFM model(r=0.901,RMSE=0.4 K,MAE=0.3 K)in surface temperature prediction,especially in terms of integrating details such as building outlines.Furthermore,the fusion effect of the daytime model is better than that of the nighttime model.In addition,the accuracy evaluation based on actual meteorological stations indicates that the fusion effect of the FSDITFM model is slightly better than that of the STITFM model in terms of robustness and detail feature processing under different land cover types,but the differences in each evaluation index are within a small range(R~2<0.01,RMSE<1,MAE<0.1).(2)The urban heat island effect in the study area is more significant at night than during the day,and the urban heat island effect is more pronounced in summer during the day than in other seasons.Significant differences in surface temperature exist among different local climate zones.The built-up area(LCZ 1-9)is a key factor in the production of urban heat islands,with compact low-rise buildings(LCZ 2-3)being the main heat source in the study area,while water bodies and vegetation can help cool the city.There are some differences in intra-LCZ variability between different seasons,but the overall standard deviation is no higher than 5.0 K.In addition,the diurnal temperature range is largest in the transition season,followed by winter,and smallest in summer.Moreover,the diurnal temperature range is greatest in compact and large low-rise buildings(LCZ 3,LCZ 7),while water bodies and dense/sparse forests have a smaller diurnal temperature range.(3)Most urban forms are significantly correlated with seasonal surface temperature.The impact of urban forms on summer surface temperature is the most significant,with normalized vegetation index,average tree height,and sky view factor being the main cooling factors.Taking average surface temperature as an example,Pearson correlation coefficients are-0.444,-0.145,and-0.304,respectively,and the P values are less than 0.01.In winter,the relationship between urban form parameters and surface temperature is mostly opposite to that of summer.Vegetation plays a weaker role in regulating temperature and instead acts as insulation.The relationship between surface temperature and urban form parameters in the transitional season shows higher consistency with winter but weaker than that in winter.In addition,urban morphological parameters(except MTH)have a significant correlation with daily temperature range(DTR)and relative temperature range(ΔDTR)of seasonal surface temperature,especially in summer.The urban form parameters have a completely opposite effect on DTR andΔDTR.This indicates that using only the highest and lowest temperatures may not accurately reflect the daily changes in summer surface temperature,and further emphasizes the importance of introducingΔDTR.(4)The influence of urban morphology on land surface temperature varies among different local climate zones.In summer,NDVI is the dominant factor in reducing U-LSTs.In transitional and winter seasons,in some LCZs,the absolute dominance of NDVI is replaced by SVF.Additionally,the combined effect of two and three-dimensional urban morphology parameters is significant(R~2:0.2-0.75).Apart from open-building areas,the contribution of two-dimensional urban morphology(R~2:0.2-0.7)to the land surface temperature in the study area is much higher than that of three-dimensional urban morphology(R~2:0.1-0.4).The use of only three-dimensional urban morphology parameters cannot predict changes in land surface temperature well.Furthermore,all LCZ models perform well in simulating seasonal DTRs(R~2:0.3-0.8;MSE<3 K).The results show that NDVI is more significant than other morphology parameters in affecting seasonal DTRs,and in most LCZs,NDVI has a higher impact on transitional season DTRs than in summer and winter seasons.The findings of this study contribute to a deeper understanding of urban heat islands in both two-dimensional and three-dimensional spaces,providing ideas and references for mitigating urban heat islands,optimizing urban planning and management,and sustainable development strategies.