| The Jiangsu,Zhejiang and Shanghai regions on the eastern coast of China are important economic development belts in China.Due to the impact of global warming,sea level rises and climate change is abnormal,which has a great impact on urban development and social environment in the region.Therefore,based on the 8-day surface temperature products provided by MODIS,this paper reconstructs the surface temperature data series from 2003 to 2017 by harmonic analysis and analyzes its spatial and temporal changes.At the same time,using the monthly mean data of meteorological stations in the study area in 2017,the multivariate linear regression method was used to fit the five factors of longitude,latitude,altitude,surface temperature and NDVI,and the spatial distribution of average temperature in each month in 2017 was obtained.The final conclusion is as follows:(1)Overall,the annual average data availability of MODIS LST data is 53.25%,and the data availability of Terra Star is higher than Aqua Star.The valid data for each month is above 40%,and the highest availability in March is 69.40%.The data quality is lower in June,July and August.The difference between the north and the south is large in space,and the data quality in the north is higher than that in the south(2)The data availability after reconstruction using harmonic analysis method has increased from 53.25%to over 95%,and the integrity is better.The accuracy of the reconstructed data and the original QC=0 data is verified.The results show that the average error is 2.34?,and the ratio of the absolute difference between the reconstructed LST and the original LST in 1? is 21.65%,0?2?.The proportion is 56.05%.(3)The average surface temperature of Jiangsu,Zhejiang and Shanghai for many years from 2003 to 2017 was 17.20?,showing a fluctuating upward trend,with an average annual increase of 0.0414?.In the past five years,the surface temperature has increased significantly,reaching a peak in 2017,with an average surface temperature of 17.86?.Comparing the coefficient of variation of surface temperature in the four seasons from 2003 to 2017,the coefficient of variation in spring,summer and autumn is small,with the smallest in summer,1.53%,and the coefficient of variation in winter is 11.17%,indicating the largest change in LST in winter;After linear fitting of surface temperature,it is found that the surface temperature rise trend is more obvious in summer and winter.The surface temperature of each month in 15 years showed an upward trend,and the upward trend was the most obvious in January.The average annual increase was 0.1107?,and the increase in November was the smallest,and the trend line was close to the level.(4)There is a difference in the spatial distribution of surface temperature.The spatial variation of the average surface temperature in Jiangsu,Zhejiang and Shanghai from 2003to 2017 is between 13.29 and 21.03?.The northern temperature is low,and the surface temperature in the south is staggered.The difference in surface temperature in Zhejiang is greater than that in Jiangsu.According to the trend analysis method,the interannual variation analysis of Jiangsu,Zhejiang and Shanghai in 2003?2017 shows that the trend slope is greater than 0,showing an upward trend.The rising area accounted for 43.93%,mainly concentrated in Jiangsu,Shanghai and Zhejiang,Jiaxing,Ningbo,Hangzhou,the main urban area,southern Wenzhou,Taizhou coastal areas;the decline area accounted for12.99%,mainly concentrated in Zhejiang's higher elevation forest areas;The regions with significant rises and significant declines are extremely rare,accounting for only 0.97% and 0.70%.(5)Different land uses have different effects on surface temperature.From 2003 to 2017,the land use with the highest average surface temperature in Jiangsu,Zhejiang and Shanghai regions was 17.45?,followed by unused land.The average surface temperature of water bodies was 16.63?,and the difference between cultivated land,grassland and forest land was small.Between the highest and lowest values,the woodland temperature is slightly lower than the cultivated land and grassland.The temperature of construction land and unused land is higher than the average surface temperature,which has an effect on the surface temperature,while the temperature increase effect of the construction land is obvious;the water body and the forest land are lower than the average surface temperature,and the cooling effect on the surface temperature is obvious.(6)In this paper,the five factors affecting the surface temperature acquired by the reconstructed Terra star at night time,the NDVI data obtained by Terra star,the longitude,latitude and altitude are the same as the month of the weather station of Jiangsu,Zhejiang and Shanghai 2/3 in 2017.A multivariate linear regression model was established for the average temperature.After the fitted model was tested with the monthly average temperature of the remaining 1/3 of the site,the average R~2 of 12 months was 0.824,and the average RMSE was 0.59.The overall error was The data of 0~0.5? accounts for 63.44%,and the precision is high.According to the established multiple linear regression model,the temperature distribution map of the average monthly gas in the Jiangsu,Zhejiang and Shanghai regions in 2017 is available.The temperature simulation results in January and November are relatively smooth,and the temperature distribution in April and July is spatially distributed.The details are better. |
PDF Full Text Request