| Global climate problem has attracted more and more attention of mankind.The Urban heat island(UHI)caused by the rise of land surface temperature(LST)is one of the major impacts of human industrialization and urbanization on the earth's surface climate.Compared with traditional LST inversion using polar-orbiting satellite remote sensing data,LST products provided by synchronous satellites have higher timeliness and accuracy.Based on Global 5km hourly LST data from constellation of geostationary satellites,this study uses quantitative remote sensing technology and geographic information system(GIS)spatial analysis method to quantify LST by impervious surface fraction,explain the daily cycle characteristics and spatial heterogeneity of global UHI.At the same time,MODIS Enhanced Vegetation Index(EVI)products,NPP–VIIRS Nighttime Light(NL)products and SRTM3 global DEM data are quantified by impervious surface fraction and the driving factors of global UHI are analyzed comprehensively.The main conclusions are as follows:(1)Global LST cycles with local solar altitude.From the east-west direction,the overall LST from 6:00 to 17:00 is higher than that from 18:00 to 5:00.LST reaches the highest at9:00 with a value of 343.66 K.LST from 18:00 to 5:00 is between 250 K and 340 K,and reaches the highest at 21:00 with a value of 339.97 K.From the north-south direction,the changes of LST in different climatic zones are different,and generally show the curve fluctuation of rising and falling.From 6:00 to 15:00,LST in arid region is higher than that in other climate regions,and reached the maximum at 9:00,which is 310.53 K.During the period from 19:00 to 3:00,LST in the arid area keep a downward trend and reached the lowest at 3:00,which is 298.90 K.The LST in tropical region begin to decrease from 18:00 to23:00,which is the lowest in the whole day,292.36 K.As temperate region is more suitable for human social or economic development and contains more large cities,compared with other climate regions,the fluctuation is stronger and the threshold range is larger,especially the most significant between 15:00 and 17:00,and the LST is between 291.02 K and 306.31 K.(2)In this study,UHI is quantified by LST along the artificial impervious surface gradient,denoted by ? LST.The global diurnal ? LST is between-1 and 1,and there are more cities with values close to 0 during "day" than "night".During the "day",the average value of? LST is 0.002 K/%,and cities with positive ? LST account for the majority(175/294).These areas are mostly industrial bases or densely populated areas,forming metropolitan areas with the highest value of 0.77 K/%,and cities with negative ? LST account for 19.05 %.The minimum value is-0.95 K /%.Cities with positive ? LST at night were less than those in day(159/294),but the ? LST changes are more obvious.The average ? LST at night is 0.001 K/%,and 62.92% of the cities with values between-0.1 and 0.1 are mainly distributed in the eastern United States,Western Europe and southeast Asia.From the climate zone perspective,the mean values of ? LST between day and night in tropical region are-0.167 K /% and-0.165K/% respectively,showing a stable temporal variation.? LST values are the highest in "day" cold region(0.025 K /%),and the lowest in arid region(0.014 K /%).At night,? LST is the highest in arid(0.061 K/%),and the lowest in temperate region(0.005 K/%).In cold and arid region,? LST of "night" changes more sharply than that of "day".In addition,this study also discusses the distribution of UHI under elevation differentiation.The results show that the nighttime ? LST threshold range of each city is more stable under elevation differentiation than at night.When elevation gradient is low,the average ? LST value of daytime and night is closer to 0.When elevation gradient is high,the ? LST value of most cities is negative,and the more complicated the terrain,the greater the difference of ? LST value.(3)The spatial distribution of ? NL,? EVI and ? DEM shows variability globally.In most cities,the ? NL value is positive,accounting for 29% of the total value of over 0.5.The average value of ? LST at night in these regions is 0.039 K/%,which is much higher than the overall value of ? LST at night,indicating that the strong UHI is generated while the economic difference is large.The median value of ? NL in global cities is 0.33 and the average value is 0.29.The mean value of ? NL for tropical(0.38)and cold regions(0.44)is higher than global level,and the mean value of ? NL for temperate regions(0.27)and arid regions(0.16)is lower than global level.The global average value of ? EVI is 0.02,and more than half of the cities have negative ? EVI values(172/294).The climatic types from high to low are tropical(0.09),cold(0.03),temperate(0.02),arid(-0.04).The ? DEM values of33.33% of cities are positive but lower than 0.5,and 32.31% of cities are negative but higher than-0.5.Most of them are economically developed cities or urban agglomerations.The average value of ? DEM for cities around the world is 0.04 and the median value is 0.02,and the highest value is in temperate(0.07).It is followed by cold(0.05),arid(-0.03)and tropical(-0.13).(4)Using ? NL,? EVI and ? DEM as independent variables and ? LST as dependent variables,the results of multiple linear regression analysis show that independent variables can explain 31.1% of the variation of dependent variable.From a global perspective,the degree of influence on ? LST is ? EVI > ? NL > ? DEM.From the angle of correlation coefficient,? NL and ? LST show positive correlation(0.452),? EVI,? DEM and ? LST show negative correlation(-0.229,-0.311).The correlation coefficient between ? NL and?LST is always positive in the four climatic regions,with the highest value(0.654)in tropical and the lowest value(0.309)in the arid.The ? EVI and ? LST show negative correlation in most cities.Vegetation weakened UHI,and the negative correlation is-0.242 and-0.274 in tropical and temperate regions,respectively.The correlation between ? DEM and ? LST in the four climatic regions is negative,and the correlation coefficient is-0.683 in tropical.The influence of ? DEM on ? LST in temperate and cold regions is lower than that in tropical,and the correlation coefficient is-0.316. |
PDF Full Text Request