Quantitative Inversion Of Temperature And Precipitation By Remote Sensing And Its Influence On Vegetation Index

Global climate change makes people pay more attention to ecological environment,especially climate and vegetation.Northeast China is an important agricultural and forestry production base in China and a sensitive area of temperature and precipitation.Therefore,it is necessary to monitor the dynamic changes in the long time scale.Based on this,the temperature and precipitation daily data of remote sensing and stations in northeast China regions were used as data sources and the optimal corresponding relationship model between the two kinds of data is estimated under different time scales.By using the optimal relationship,the monthly remote sensing data of temperature and precipitation from 2000 to2018 were used to obtain the annual average temperature and annual precipitation of the corresponding period.Correlation analysis was used to study the relationship between annual temperature,precipitation and vegetation index?NDVI?.The results are as follows:?1?The surface temperature in the remote sensing data of MODIS can better invert the mean air temperature,maximum temperature and minimum temperature on the scale of day,month,season and year,especially the average temperature.1)Daily scale: the inversion results were successively the average temperature > the lowest temperature > and the highest temperature,and the best goodness of fit and verification accuracy were under the average temperature?r² = 0.976,RMSE = 2.2986 ?,MAE = 1.7909 ??.2)Monthly scale: the model performance was the best in October?r²,RMSE,MAE were 0.9851,0.5145 ?,0.4162 ?,respectively?,and the worst in August?r²,RMSE,MAE were 0.8324,0.9214 ?,0.7327 ?,respectively?.3)Seasonal scale: The verification accuracy of inversion estimation in the four seasons was high,with r² greater than 0.89,RMSE and MAE less than 1 ?.4)Annual scale:The estimation result was the best?k = 0.978,r² = 0.974?,and both RMSE and MAE were less than 0.5 ?.?2?IMERG remote sensing data perform well in the estimation of precipitation at the seasonal and annual scales,while the estimation effect at the daily and monthly scales needs to be improved.1)Daily scale: The inversion range of daily precipitation was 0 75 mm,the fitting effect was poor?r² = 0.1351?,and the accuracy of non-precipitation and light rain in the magnitude frequency analysis was high.2)Monthly scale: There was an underestimation ofhigh precipitation?except in October?,with the best fitting in April?r² = 0.8955?and the worst fitting in September?r² = 0.066?.3)Quarterly scale: The fitting effect of the four seasons was good?r² between 0.8792 0.9351?,and the total precipitation was the best in summer?MAE= 4.3698 mm?and the worst in winter?MAE = 30.0149 mm?.4)Annual scale: The annual precipitation range was 100 1100 mm,and the fitting estimation effect was the best?r² =0.8476,RMSE = 89.9179 mm,MAE = 67.3797 mm?.?3?The changes of NDVI,temperature and precipitation in northeast China from 2000 to 2018 were shown as follows: 1)NDVI showed an upward trend of fluctuation(the overall growth rate was 0.0025?a^-1,p < 0.01),on the spatial scales,the area proportion of NDVI growth trend was 84.98%,the area with k between 0 0.005 was the largest,and the area with trend less than-0.05 was the least.2)During the 19 years,the overall temperature fluctuated and increased(the growth rate was 0.0296 ?·a^-1).In terms of space,the temperature change was shown as the area proportion of decreasing trend was more?52.95%?,the area of trend between-0.5 0 was the largest,and the area of trend greater than 1 was the least.3)The precipitation fluctuates and increases with time(the growth rate was 6.9587 mm·a^-1).In terms of space,the proportion of regions with increasing trend of precipitation in space was 94.43%,among which the regions with trend between 5 7.5 was the largest and the areas with trend less than-5 was the least.?4?The correlation coefficient range of annual average temperature,annual total precipitation and annual average NDVI in 2000²018 was-0.93⁰.98,-0.88⁰.95,respectively.The influence of temperature and precipitation on NDVI was mainly negatively correlated and positively correlated,respectively,and the regions with insignificant correlation are the largest.The regions where temperature has a great influence on vegetation index are mainly in the southwest of northeast China,the central of Jilin province and Heilongjiang province,and parts of the Lesser Khingan Mountains and Changbai Mountains.The regions where precipitation has a great influence on vegetation index are mainly in the central and southwest of northeast China.