AU2021105817A4 - Method for Reconstructing global Surface Temperature - Google Patents

Abstract

The invention belongs to the technical field of meteorological monitoring, and discloses a method for repairing earth's surface temperature, which comprises the following steps. Reconstructing scale surface temperature data by constructing a data repair model based on MODIS surface temperature data. The spatio-temporal variation characteristics of surface temperature in different time dimensions are analyzed from the whole, local and single pixel multiple spatial scales, and the abnormal variation areas of surface temperature are determined. Combined with the data of surface, atmosphere, ocean and social and economic activities, the driving factors of the spatio-temporal variation of surface temperature, especially the interannual abnormal variation, are analyzed. The method effectively improves the data quality of MODIS remote sensing surface temperature. The reconstructed monthly remote sensing surface temperature data covers a more complete surface area, and the practical application accuracy of the data reaches 1.5 K. It has important reference significance for the repair and reconstruction of remote sensing surface temperature data, climate variation monitoring, meteorological disaster early warning and agricultural production. 2/9 Using the comput er language program to obtain corresponding data in batches. and 101 perforning data extraction, coefficient conversion, projection conversion, cropping, site data sorting and other preprocessing on the obtained data 102 Combining LST remote sensing data, ground station data, altitude data and neighbor pixels, a data repair model is constructedfor data reconstruction, and the accuracy of thereconstructed data is verified by using independent monthly surface t temperature data of meteorological stations Generating high-quality surface temperature data, analyzing the spatio-temporal 103 variation of surface temperature by using GIS, IDL, SPSS and other tools. and det ermining the abnormal variation trend of surface t temperature S104 Combining various parameters of surface, atmosphere, ocean and socio-economic activities, using various geostatistical analysis methods to analyze the driving factors affecting the spatio-temporal variation of surface temperature Figure 2

Description

2/9

Using the comput er language program to obtain corresponding data in batches. and 101 perforning data extraction, coefficient conversion, projection conversion, cropping, site data sorting and other preprocessing on the obtained data

102 Combining LST remote sensing data, ground station data, altitude data and neighbor pixels, a data repair model is constructedfor data reconstruction, and the accuracy of thereconstructed data is verified by using independent monthly surface t temperature data of meteorological stations

Generating high-quality surface temperature data, analyzing the spatio-temporal 103 variation of surface temperature by using GIS, IDL, SPSS and other tools. and det ermining the abnormal variation trend of surface t temperature

S104 Combining various parameters of surface, atmosphere, ocean and socio-economic activities, using various geostatistical analysis methods to analyze the driving factors affecting the spatio-temporal variation of surface temperature

Figure 2

Method for Reconstructing Global Surface Temperature

TECHNICAL FIELD

The invention belongs to the technical field of meteorological monitoring, and

particularly relates to a method, a system and computer equipment for repairing the

earth's surface temperature.

BACKGROUND

Surface temperature is an important reference index to measure the earth's environment,

which has an important impact on the material and energy cycle, ecosystem balance and

human production and life. Traditionally, people mainly measure the surface temperature

by setting up ground stations and using thermometers. This measuring method is close to

the ground, and the measured data has high accuracy, which plays an important role in

scientific research in related fields. However, traditional methods are not only time

consuming and laborious, but also difficult to get involved in many areas due to the

limitation of terrain conditions and harsh climate environment, resulting in serious data

loss. The average daily cloud coverage in the world accounts for 60%-70%, and the data

missing is large. Ordinary interpolation methods are difficult to make up for the large

scale data missing, and it is difficult to guarantee the accuracy. In order to overcome this

difficulty, we make full use of the data of surface meteorological observation stations and

other assimilation products, and produce a set of global surface temperature data under

the condition of considering the influence of topography, which is of great significance to

global climate variation research and agricultural production monitoring.

SUMMARY

Aiming at the problems existing in the prior art, the invention provides a method, a

system and computer equipment for repairing land and ocean temperatures.

The invention is realized as follows: a method for repairing land and ocean temperatures

comprises the following steps.

Based on MODIS surface temperature data, large-scale surface temperature data is

reconstructed by constructing a data repair model. The spatio-temporal variation

characteristics of surface temperature in different time dimensions are analyzed from the

whole, local and single pixel multiple spatial scales, and the abnormal variation areas of

surface temperature are determined. Combined with the data of surface, atmosphere,

ocean and social and economic activities, the driving factors of the spatio-temporal

variation of surface temperature, especially the interannual abnormal variation, are

analyzed.

Further, the method for repairing land and ocean temperatures comprises the following

steps.

Step 1, in order to unify the standard and facilitate data processing, the computer

language program is used to obtain corresponding data in batches, and performing data

extraction, coefficient conversion, projection conversion, cropping, site data sorting and

other preprocessing on the obtained data.

Step 2, combining LST remote sensing data, ground station data, altitude data and

neighbor pixels, a data repair model is constructed for data reconstruction, and the

accuracy of the reconstructed data is verified by using independent monthly surface

temperature data of meteorological stations.

Step 3, generating high-quality surface temperature data, analyzing the spatio-temporal

variation of surface temperature by using GIS, IDL, SPSS and other tools, and

determining the abnormal variation trend of surface temperature.

Step 4, combining various parameters of surface, atmosphere, ocean and socio-economic

activities, using various geostatistical analysis methods to analyze the driving factors

affecting the spatio-temporal variation of surface temperature, thus providing information

for climate variation research and disaster prediction.

Further, in step 1, the data includes MODIS surface temperature data, ground station data

and other related data.

Other related data include altitude, NDVI, soil moisture, atmospheric water vapor, ENSO

index, NAO index and IOBW index.

Further, in step 2, the data reconstruction includes the following steps.

The quality control data set in MODIS original surface temperature data is used to

evaluate the quality of monthly surface temperature data, and pixels with accuracy lower

than 2 K and missing information are selected as the area to be repaired of monthly data.

Setting pixels with accuracy lower than 2 K and missing information in the surface

temperature daily data under the control of the area to be repaired as invalid pixels.

The ground station data in the corresponding time scale is assigned to some invalid pixels

according to longitude and latitude coordinates, and the remaining invalid pixels are

interpolated and repaired by using the neighbor value substitution method based on DEM.

New monthly data of surface temperature are generated by average synthesis of the

repaired daily data.

Further, the interpolation repair of the remaining invalid pixels by using the neighbor

value substitution method based on DEM includes the following steps.

Firstly, the pixel value of the original surface temperature is restored to the pixel value at

an altitude of 0 m, and the calculation formula is as follows:

LSTon = LST +0.55* (elel 100)

Wherein, ele represents the elevation of the pixel; LST represents MODIS surface

temperature daily data; LSTom indicates the surface temperature when MODIS surface

temperature daily data is restored to 0 m.

Secondly, the invalid pixels are interpolated and repaired by the neighbor pixel value

substitution method.

Finally, the interpolated pixel value is restored to the surface temperature value of the

corresponding elevation, and the calculation formula is as follows:

LSTeic = LSTibbl -0.55*(ele/100)

Wherein, LSTibbie represents the result of interpolation using the neighbor pixel

substitution, and LSTee represents the surface temperature restored to the corresponding

elevation after neighbor interpolation.

Further, in step 3, analyzing the spatio-temporal variation of the surface temperature

includes the following steps.

The least square method is used to calculate the interannual variation rate of surface

temperature in a long time series, and the calculation formula is as follows:

n0(kxTk)-0 k Tk b - k=1 k=1 k=1

n k 2 -( k) 2 k=1 k=1

Wherein, b is the variation rate of surface temperature with year; k is the time series

value; Tk is the average surface temperature in the k-th year; n is the total number of

years.

Significance test is performed on the calculated variation rate, and the area with

significant linear temperature increase or decrease is extracted as the abnormal variation

area of surface temperature.

Further, in step 4, analyzing the driving factors influencing the spatio-temporal variation

of the surface temperature by using a plurality of geostatistical analysis methods

comprises the following steps.

(1) Analyzing the correlation between surface temperature and vegetation index, soil

moisture, atmospheric water vapor and precipitation, expressing the correlation by

Pearson coefficient, and testing the significance of the calculated correlation coefficient.

The formula of correlation coefficient is as follows:

_1 Z(ai -5)(bi - b)

In the formula, r is Pearson index, which indicates Pearson relationship between surface

temperature and other parameters; ai is the surface temperature of different years; bi is

other surface atmospheric parameters in different years; a and b are the average values

of surface temperature and other parameters respectively; n is the total number of years.

(2) Using linear multiple stepwise regression model to analyze the influence degree of

latitude, longitude, altitude, vegetation index, soil moisture, atmospheric water vapor and

precipitation on surface temperature.

Further, the step of using linear multiple stepwise regression model to analyze the

influence degree of latitude, longitude, altitude, vegetation index, soil moisture,

atmospheric water vapor and precipitation on surface temperature comprises the

following steps.

Linear multiple stepwise regression models are established for different regions to

analyze the influence degree of latitude, longitude, altitude, vegetation index, soil

moisture, atmospheric water vapor and precipitation on surface temperature.

The linear model of the linear multiple stepwise regression is as follows:

P yz = Yzkxzk-+ k=1

Among them, yz and Xzk (k = 1, 2......p) represent standardized dependent and

independent variables respectively, p represents the number of factors, z represents

standardized coefficient, and c represents residual.

Further, the method for repairing land and ocean temperatures further comprises the

following steps: determining the lag time of the response variation of the surface

temperature in different regions to the climate mode by comparing the correlation

coefficient between the surface temperature anomaly value and the climate modal index

under different lag times.

Further, the significance test comprises the following steps: adopting a T test method as

follows:

R * n-2 T=T=1-R2

In the formula, R is the correlation coefficient between surface temperature and time; n is

the number of samples. Statistic Tc follows T distribution with degree of freedom n-2.

Given the significance level a, if ITcI > Ta, the original hypothesis is rejected, indicating

that the linear variation trend of surface temperature is significant.

The correlation coefficient between surface temperature and time is calculated as follows:

n (kxTk) JkjTk R =kI k I kI

nA k2 (k)2 x n Tk ( YT)2 k-1 k-1 k-1 k-1

Wherein, k is a time series value; T is the average surface temperature in the k-th year; n

is the total number of years.

Another object of the present invention is to provide a computer equipment, which

comprises a memory and a processor, wherein the memory stores a computer program,

and when the computer program is executed by the processor, the processor executes the

following steps. Reconstructing scale surface temperature data by constructing a data

repair model based on MODIS surface temperature data. The spatio-temporal variation

characteristics of surface temperature in different time dimensions are analyzed from the

whole, local and single pixel multiple spatial scales, and the abnormal variation areas of

surface temperature are determined. Combined with the data of surface, atmosphere,

ocean and social and economic activities, the driving factors of the spatio-temporal

variation of surface temperature, especially the interannual abnormal variation, are

analyzed.

Another object of the present invention is to provide a land and ocean temperature repair

system for implementing the method for repairing land and ocean temperature, and the

land and ocean temperature repair system comprises the following.

A data preprocessing module, which is used for obtaining corresponding data in batches

by using computer language programs, and performing data extraction, coefficient conversion, projection conversion, cropping, site data sorting and other preprocessing on the obtained data.

A data accuracy verification module, which is used for constructing a data repair model

for data reconstruction by combining LST remote sensing data, ground station data,

altitude data and neighbor pixels, and verifying the accuracy of the reconstructed data by

using independent monthly surface temperature data of meteorological stations.

An abnormal variation trend determination module, which is used for generating high

quality surface temperature data, analyzing the spatio-temporal variation of surface

temperature by using GIS, IDL, SPSS and other tools, and determining the abnormal

variation trend of surface temperature.

A driving factor analysis module, which is used for analyzing the driving factors

influencing the spatio-temporal variation of the surface temperature by combining

various parameters of the surface, the atmosphere, the ocean and social and economic

activities and utilizing various geostatistical analysis methods.

Combined with all the above technical schemes, the invention has the advantages and

positive effects that the data repair model constructed based on ground stations, neighbor

pixels and altitude effectively improves the data quality of MODIS remote sensing

surface temperature. The reconstructed monthly remote sensing surface temperature data

covers a more complete surface area, and the average accuracy of the data reaches 1.5 K.

It has important reference significance for the repair and reconstruction of remote sensing

surface temperature data, climate variation monitoring, meteorological disaster early

warning and agricultural production.

BRIEF DESCRIPTION OF THE FIGURES

Figure 1 A schematic diagram of a method for repairing land and ocean temperatures

provided by an embodiment of the present invention

Figure 2 A flow chart of the method for repairing land and ocean temperatures provided

by the embodiment of the present invention

Figure 3 A schematic diagram of research objects and partitions provided by an

embodiment of the present invention

Figure 4 A schematic diagram of accuracy verification results of surface temperature data

in different months provided by an embodiment of the present invention

Figure 5 A schematic diagram of the accuracy verification result of reconstructed LST

data provided by an embodiment of the present invention

Figure 6a A schematic diagram of the spatial distribution of global annual average

surface temperature from 2002 to 2018 (unit:°)

Figure 6b A schematic diagram of the spatial distribution of the annual average surface

temperature in China from 2002 to 2018 (unit:°)

Figure 7 A schematic diagram of the spatial distribution of daytime and night average

surface temperatures in China from 2002 to 2018 provided by an embodiment of the

present invention

Figure 8 A schematic diagram of spatial distribution of average surface temperature in

different seasons in China from 2002 to 2018 provided by an embodiment of the present

invention

Figure 9 A schematic diagram of spatial distribution of temperature difference between

day and night and temperature difference between winter and summer provided by an

embodiment of the present invention

Figure 10 A schematic diagram of the variation trend of the annual average LST anomaly

value in China from 2002 to 2018 provided by an embodiment of the present invention

Figure 11 A schematic diagram of the variation trend of the day-night and seasonal

average LST anomaly values in China from 2002 to 2018 provided by an embodiment of

the present invention

Figure 12 A schematic diagram of the annual variation trend of the interannual average

LST in China from 2002 to 2018 provided by an embodiment of the present invention

Figure 13 A schematic diagram of the interannual variation trend of the average LST

during day and night in China from 2002 to 2018 provided by an embodiment of the

present invention

Figure 14 A schematic diagram of the interannual variation trend of the average LST in

four seasons in China from 2002 to 2018 provided by an embodiment of the present

invention.

DESCRIPTION OF THE INVENTION

In order to make the purpose, technical scheme and advantages of the present invention

clearer, the present invention will be further described in detail with reference to the

following es. It should be understood that the specific embodiments described herein are

only used to explain the present invention, and are not used to limit the present invention.

In view of the problems existing in the prior art, the invention provides a method, a

system and computer equipment for repairing land and ocean temperatures, and the

invention is described in detail with reference to the drawings.

As shown in Figure 1, the method for repairing land and ocean temperatures provided by

the embodiment of the present invention includes the following steps.

Reconstructing scale surface temperature data by constructing a data repair model based

on MODIS surface temperature data. The spatio-temporal variation characteristics of

surface temperature in different time dimensions are analyzed from the whole, local and

single pixel multiple spatial scales, and the abnormal variation areas of surface

temperature are determined. Combined with the data of surface, atmosphere, ocean and

social and economic activities, the driving factors of the spatio-temporal variation of

surface temperature, especially the interannual abnormal variation, are analyzed.

As shown in Figure 2, the method for repairing land and ocean temperatures provided by

the embodiment of the present invention includes the following steps.

S101, obtaining corresponding data in batches by using computer language programs, and

performing data extraction, coefficient conversion, projection conversion, cropping, site

data sorting and other preprocessing on the obtained data.

S102, combining LST remote sensing data, ground station data, altitude data and

neighbor pixels, a data repair model is constructed for data reconstruction, and the

accuracy of the reconstructed data is verified by using independent monthly surface

temperature data of meteorological stations.

S103, generating high-quality surface temperature data, analyzing the spatio-temporal

variation of surface temperature by using GIS, IDL, SPSS and other tools, and

determining the abnormal variation trend of surface temperature.

S104, combining various parameters of surface, atmosphere, ocean and socio-economic

activities, using various geostatistical analysis methods to analyze the driving factors

affecting the spatio-temporal variation of surface temperature.

The data provided by the embodiment of the invention includes MODIS surface

temperature data, ground station data and other related data.

Other related data include altitude, NDVI, soil moisture, atmospheric water vapor, ENSO

index, NAO index and IOBW index.

The data reconstruction provided by the embodiment of the invention comprises the

following steps.

The quality control data set in MODIS original surface temperature data is used to

evaluate the quality of monthly surface temperature data, and pixels with accuracy lower

than 2 K and missing information are selected as the area to be repaired of monthly data.

Setting pixels with accuracy lower than 2 K and missing information in the surface

temperature daily data under the control of the area to be repaired as invalid pixels.

The ground station data in the corresponding time scale is assigned to some invalid pixels

according to longitude and latitude coordinates, and the remaining invalid pixels are

interpolated and repaired by using the neighbor value substitution method based on DEM.

New monthly data of surface temperature are generated by average synthesis of the

repaired daily data.

The interpolation repair of the remaining invalid pixels by using the neighbor value

substitution method based on DEM provided by the embodiment of the present invention

includes the following steps.

Firstly, the pixel value of the original surface temperature is restored to the pixel value at

an altitude of 0 m, and the calculation formula is as follows:

LSTon= LST+0.55*(ele/100)

Wherein, ele represents the elevation of the pixel; LST represents MODIS surface

temperature daily data; LSTom indicates the surface temperature when MODIS surface

temperature daily data is restored to 0 m.

Secondly, the invalid pixels are interpolated and repaired by the neighbor pixel value

substitution method.

Finally, the interpolated pixel value is restored to the surface temperature value of the

corresponding elevation, and the calculation formula is as follows:

LSTeic = LSTibbl -0.55*(ele/100)

Wherein, LSTibbie represents the result of interpolation using the neighbor pixel

substitution, and LSTee represents the surface temperature restored to the corresponding

elevation after neighbor interpolation.

The analysis of the spatio-temporal variation of the surface temperature provided by the

embodiment of the present invention includes the following steps.

The least square method is used to calculate the interannual variation rate of surface

temperature in a long time series, and the calculation formula is as follows:

n0(kxTk)-0 k Tk b - k=1 k=1 k=1

n k 2 -( k) 2 k=1 k=1

Wherein, b is the variation rate of surface temperature with year; k is the time series

value; Tk is the average surface temperature in the k-th year; n is the total number of

years.

Significance test is performed on the calculated variation rate, and the area with

significant linear temperature increase or decrease is extracted as the abnormal variation

area of surface temperature.

According to the embodiment of the invention, analyzing the driving factors influencing

the spatio-temporal variation of the surface temperature by using a plurality of

geostatistical analysis methods comprises the following steps.

(1) Analyzing the correlation between surface temperature and vegetation index, soil

moisture, atmospheric water vapor and precipitation, expressing the correlation by

Pearson coefficient, and testing the significance of the calculated correlation coefficient.

The formula of correlation coefficient is as follows:

_1 Z(ai -5)(bi - b)

In the formula, r is Pearson index, which indicates Pearson relationship between surface

temperature and other parameters; ai is the surface temperature of different years; bi is

other surface atmospheric parameters in different years; a and b are the average values

of surface temperature and other parameters respectively; n is the total number of years.

(2) Using linear multiple stepwise regression model to analyze the influence degree of

latitude, longitude, altitude, vegetation index, soil moisture, atmospheric water vapor and

precipitation on surface temperature.

The step of using linear multiple stepwise regression model to analyze the influence

degree of latitude, longitude, altitude, vegetation index, soil moisture, atmospheric water

vapor and precipitation on surface temperature provided by the embodiment of the

invention comprises the following steps.

Linear multiple stepwise regression models are established for different regions to

analyze the influence degree of latitude, longitude, altitude, vegetation index, soil

moisture, atmospheric water vapor and precipitation on surface temperature.

The linear model of the linear multiple stepwise regression is as follows:

P yz = zkxzk-+ k=1

Among them, yz and Xzk (k = 1, 2......p) represent standardized dependent and

independent variables respectively, p represents the number of factors, z represents

standardized coefficient, and c represents residual.

The step of repairing land and ocean temperatures provided by the embodiment of the

invention further comprises the following steps: determining the lag time of the response

variation of the surface temperature in different regions to the climate mode by

comparing the correlation coefficient between the surface temperature anomaly value and

the climate modal index under different lag times.

The significance test provided by the embodiment of the invention comprises the

following steps: adopting a T test method as follows:

R * n-2 T=T=1-R2

In the formula, R is the correlation coefficient between surface temperature and time; n is

the number of samples. Statistic Tc follows T distribution with degree of freedom n-2.

Given the significance level a, if ITcI > Ta, the original hypothesis is rejected, indicating

that the linear variation trend of surface temperature is significant.

The correlation coefficient between surface temperature and time is calculated as follows:

n (kxTk) JkjTk

nA k2 (k)2 x n Tk ( Tk)2 k-1 k-1 k-1 k-1

Wherein, k is a time series value; Tk is the average surface temperature in the k-th year; n

is the total number of years.

The following will further explain the technical scheme of the present invention with

specific embodiments.

Embodiment 1:

1 Overview and methods

1.1 China is located in the east of Asia-Europe continent, on the west coast of Pacific

Ocean, from the center of Heilongjiang River near Mohe River in the north, to Zengmu

Shoal in Nansha Islands in the south, from Pamirs Plateau in the west, and to the

confluence of Heilongjiang and Wusuli River in the east. The north and south span the

tropics, subtropics and temperate zones, from west to east, and go through plateaus,

mountains, hills and plains. The climate types are complex and diverse. According to the

terrain and climate characteristics of our country, our country is divided into four regions,

which are: I arid and semi-arid region in northwest China, II Qinghai-Tibet Plateau

region, III northern region and IV southern region.

1.2 Method

1.2.1 Surface temperature data reconstruction

The low accuracy of data in some areas caused by cloud interference is a common

problem faced by optical remote sensing, which will affect the accuracy of data analysis.

The method comprises the following steps. Firstly, preprocessing surface temperature

products, including data set extraction, coefficient conversion, projection conversion,

cropping and the like, to obtain monthly surface temperature data of North America from

2002 to 2018. On this basis, the invention uses the quality control data set in MODIS

original surface temperature data to evaluate the quality of monthly surface temperature

data, and selects pixels with accuracy lower than 2 K and missing information as the area

to be repaired for monthly data. Then, the pixels with accuracy lower than 2 K and

missing information in the surface temperature data controlled by this range are set as

invalid pixels. After determining the pixels to be repaired, the invention performs repair

interpolation on them. The invention first assigns the ground station data under the

corresponding time scale to partial invalid pixels according to the longitude and latitude

coordinates, and then performs interpolation repair on the remaining invalid pixels by

using the neighbor value substitution method based on DEM. After the invalid pixels are

repaired, the invention generates a new monthly surface temperature data by average

synthesis of the repaired daily data. Finally, the invention verifies the accuracy of the

reconstructed data by using the monthly surface temperature data of independent

meteorological stations. The technical flow chart of data repair is shown in Figure 1.

The neighbor pixel value substitution method used in the invention is to substitute and

repair the pixel values of the neighbor areas under the same time scale, and assign the

nearest pixel values to invalid pixels. Before using this method for interpolation, the

invention has interpolated some pixels using ground station data, and other invalid pixels are basically distributed sporadically. In this case, the neighbor value substitution method can guarantee certain interpolation accuracy. In order to further improve the repair accuracy, the invention considers the influence of altitude on the surface temperature.

The present invention is based on the generally accepted vertical temperature decreasing

law in geography, in which the temperature drops by 0.55°C for every 100 m above sea

level without changing other conditions (PAYTAN, 2012). According to this law, firstly,

the pixel value of the original surface temperature is restored to the pixel value when the

altitude is 0 m, and the calculation formula is shown in Formula (1). On this basis, the

invalid pixels are interpolated and repaired by the neighbor pixel value substitution

method. Finally, according to this law, the interpolated pixel value is restored to the

surface temperature value of the corresponding elevation, as shown in Formula (2). The

vertical temperature decreasing law, when the altitude rises by 100 m and the temperature

drops by 0.55°C, is suitable for most areas. Therefore, for pixels with close distance and

large altitude difference, the method can improve the accuracy of surface temperature

data repair to a certain extent.

LSTom=LST+0.55*(ele100) (1)

LSTeic = LSTnibbl-0.55*(ele/100) (2)

In the formula, ele is the elevation of pixel, and the unit is m; LST is MODIS surface

temperature daily data; LSTom is the surface temperature when MODIS surface

temperature daily data is restored to m; LSTibble is the result of interpolation using the

neighbor pixel substitution, and LSTee is the surface temperature restored to the

corresponding elevation after neighbor interpolation.

1.2.2 Monitoring of abnormal variation in surface temperature

In order to reveal the interannual variation law of surface temperature from 2002 to 2018

from pixel scale, the least square method is used to calculate the interannual variation rate

of surface temperature in the long time series from 2002 to 2018, and the calculation

formula is shown in Formula (3). Secondly, significance test is performed on the

calculated variation rate, and the area with significant linear temperature increase or

decrease is extracted as the abnormal variation area of surface temperature. The T test

method is adopted in the invention, and the test method is shown in Formula (4).

n$(kx Tk)-$ k$ T b = k=1 k=1 k=1

2 n k -( k) 2

k=1 k=1 (3)

In the formula, b is the rate of variation of surface temperature with year; k is a time

series value, which is 1-17 in turn from 2002 to 2018; T is the average surface

temperature in the k-th year; n is the total number of years, which is taken as 17 in the

present invention.

R* 2 TT= =3 l-R 2 (3)

In the formula, R is the correlation coefficient between surface temperature and time, and

the calculation Formula (5) is as follows; n is the number of samples, which is taken as

17 in the present invention. The statistic Tc follows the T distribution with the degree of

freedom (n-2). Given the significance level a, if ITcI > Ta, the original hypothesis is

rejected, and the linear variation trend of surface temperature is considered to be

significant.

n (kxT) JkjT, R = k-I k-I k

nik2 ( k)2 x n Tk ( Tk)2 k-1 k- k-I k-1 (5)

In the formula, k is a time series value, which is 1-17 in turn from 2002 to 2018; T is the

average surface temperature in the k-th year; n is the total number of years, which is

taken as 17 in the present invention.

1.2.3 Study on driving factors of surface temperature variation

Correlation analysis can only show the correlation between surface temperature and other

parameters, but can not reveal the influence mechanism of various parameters on surface

temperature. Therefore, on the basis of correlation analysis, the regression relationship

between surface temperature and other parameters is studied. The linear multiple

stepwise regression model is used to analyze the influence degree of seven factors,

namely latitude, longitude, altitude, vegetation index, soil moisture, atmospheric water

vapor and precipitation, on the surface temperature. In order to improve the accuracy of

regression analysis, the present invention establishes regression models for different

regions respectively, and the partition basis is shown in Figure 3. Multiple stepwise

regression is an important method in regression analysis. This method is mainly used to

determine the number of independent variables from more factors to be selected and

select independent variables to establish the equation with the most regression

relationship. The invention selects a stepwise regression scheme. The basic idea is to

select the factor with the largest variance contribution from the p factors to be selected,

first introduce the regression equation for it, and select until the last factor. After

introducing a factor, the existing factors in the regression equation should be checked one

by one, and if the regression effect becomes insignificant, it should be eliminated immediately. The regression equation obtained at last is optimal until no factor can be introduced and no factor can be eliminated from the regression equation. The linear model of linear multiple stepwise regression is shown in Formula (6).

yz = Zzkxzk p +8 k=1 (6)

yz and xz (k = 1, 2......p) represent standardized dependent and independent variables

respectively, p is the number of factors, z is standardized coefficient, and c is residual.

1.2.4 Analysis of lag effect

Because the occurrence centers of different climate modes are different and there is a

certain spatial distance from China, the influence of climate modes on surface

temperature variation in China has a certain lag. The invention determines the lag time of

the response variation of the surface temperature in different regions of China to the

climate mode by comparing the correlation coefficient between the surface temperature

anomaly value and the climate mode index under different lag times. The basis of

determination is that when a pixel has the strongest correlation between its surface

temperature anomaly value and climate modal index under a certain lag time (0 month, 1

month, 2 months ... 12 months), the lag time is determined as the lag time of the pixel's

response variation of the surface temperature to the climate mode.

1.3 Data

1.3.1 Remote sensing surface temperature data

MODIS sensors are carried on Terra and Aqua satellites, which were launched in 1999

and 2002 respectively, with transit time of 10: 30 am and 01: 30 pm respectively. MODIS

sensor has 36 discrete spectral bands, which can simultaneously provide information

reflecting land surface conditions, cloud characteristics, aerosol, surface temperature, ozone, ocean and other characteristics. At present, MODIS surface temperature products include seven categories: MOD1I _ L2, MOD1IA1, MOD11A2, MOD1IBI,

MODIIC1, MODI1C2, MODI1C3, and their detailed parameters are shown in the table.

According to the invention, a monthly surface temperature product MODI1C3 covering

the whole world with a resolution of 0.050 is used, and the time scale covers 2002-2018.

The MODI1C3 product is obtained by day and night inversion, projection, splicing,

resampling and average synthesis.

Table 1 Introduction of MODIS Series Surface Temperature Data

Numberof Spatial Time Type rows and resolution resolution Projection columns/range MOD1I L2 2030x1354 1 km Column No MODIA1 1200x1200 1 km day Sinusoidal projection MODI1A2 1200x1200 1 km 8 days Sinusoidal projection MODIIBI 240x240 6 km day Sinusoidal projection MOD1ICI 3600x1800 0.05°x0.050 day Isometric projection MOD11C2 3600x1800 0.05x0.050 8 days Isometric projection MOD11C3 3600x1800 0.05x0.050 Month Isometric projection 1.3.2 Data of surface meteorological stations

The meteorological station data used in the invention comes from the official website

website of the National Center for Environmental Information (NCEI) of the National

Oceanic and Atmospheric Administration of the United States. The NCEI land-based

observation data collected instruments from different continents and regions, including

temperature, dew point, relative humidity, precipitation, wind speed and direction,

visibility, atmospheric pressure, hail, fog and thunder, etc., and can obtain data on various

time scales such as hours, days, months, year and years. In the present invention, firstly,

four hour scales from 2002 to 2018 provided by this institution are selected, including 01:

am, 10: 00 am, 13: 00 pm and 22: 00 pm, which respectively approximate to 01: 30 am, 10: 30 am, 13: 30 pm and 22: 30 pm in MODIS surface temperature daily data. The

MODIS surface temperature daily data at different times are interpolated and repaired

with hourly ground station data. After the data is repaired, the surface temperature

monthly data of meteorological stations from 2002 to 2018 provided by the institution are

re-selected to verify the accuracy of the reconstructed surface temperature data.

1.3.3 Surface atmospheric parameters

The influencing factors of surface temperature are complex. According to the related

research results of predecessors on the formation mechanism of surface temperature, the

invention selects four surface atmospheric parameters, namely normalized difference

vegetation index (NDVI), soil moisture, atmospheric water vapor content and

precipitation, and analyzes the driving factors of the spatio-temporal difference of surface

temperature by integrating longitude, latitude and altitude. Detailed introduction of four

surface air parameter products is shown in Table 2. The vegetation index, atmospheric

water vapor content and precipitation data used in the invention are all from MODIS data

provided by NASA (National Aeronautics and Space Administration), and the soil

moisture data in the invention is synthesized by using AMSR-E, SOMS and AMSR-2

data.

Table 2 Introduction of Other Surface Atmospheric Data

Type Data name Spatial Time resolution Spatial range resolution NDVI MODI1C2 Global Monthly scale 0.050* 0.050 Soil moisture Passive Global Monthly scale 0.050* 0.050 microwave soil moisture Atmospheric MODO8_M3 Global Monthly scale 10* 10 water vapor Precipitation TRMM - TMPA 50°N - Monthly scale 0.250* 0.250 500 S

1.3.4 Marine climate modal index

The ocean on the earth's surface accounts for about 71% of the global total area, which is

an important factor affecting the thermal distribution, atmospheric circulation, weather

variation and climate differences on the earth's surface. El Nino-La Nina phenomenon

refers to the continuous warming and cooling of sea surface temperature in the equatorial

central and eastern Pacific Ocean. This phenomenon affects the water vapor energy cycle

between land and sea, and has an important impact on temperature and rainfall in many

parts of the world. They are the strongest signal of interannual variation of climate

system. Generally, it is defined as an El NINO event that the sea surface temperature

anomaly index in NINO region (50 N-5° S, 90° W-150° W) reaches more than 0.5°C for

6 consecutive months, and it is defined as a La Nina event that the SST anomaly index in

Nino region reaches or falls below-0.5°C for at least 6 consecutive months. The North

Atlantic Oscillation (NAO) refers to the relationship between the reverse pressure

variation between the azores high and Iceland depressions in the North Atlantic, and it is

the most significant mode of the North Atlantic atmosphere. The positive phase of NAO

reflects that the high latitude area of the North Atlantic is lower than the normal pressure,

while the central part of the North Atlantic is higher than the normal pressure. The

negative phase of NAO reflects that the high latitude area of the North Atlantic is higher

than the normal pressure, while the middle part of the North Atlantic is lower than the

normal pressure. NAO index can explain 31% of the variance of average temperature in

winter in northern hemisphere. The consistent sea surface temperature mode (IOBW) in

the whole tropical Indian Ocean is defined as the average sea surface temperature anomaly in the tropical Indian Ocean region. This mode is the most important mode of sea surface temperature variation in the tropical Indian Ocean, which usually begins to develop in winter and reaches its strongest in the spring of the following year. The eastern coast of China is low-lying and close to the ocean, and the surface temperature is greatly affected by the ocean. In the present invention, the sea surface temperature index

(NINO3.4) in NINO3.4 area, the North Atlantic Oscillation Index (NA0) and the tropical

Indian Ocean region-wide consistent sea surface temperature index (IOBW) are used to

analyze the impacts of three climate modes on the spatio-temporal variation of surface

temperatures in different regions of China. The specific introduction of the three indexes

is shown in Table 3.

Table 3 Introduction of Climate Modal Index

Data name Abbreviations source Time Time range scale Sea temperature NINO3.4 China Month 2002-2018 index in NINO3.4 meteorological area administration North Atlantic NAO NOAA Month 2002-2018 Oscillation Index Indian Ocean SST IOBW China Month 2002-2018 Index meteorological administration 1.3.5 Social and economic data

With the advancement and development of urbanization and industrialization, China's

urban population is growing, the scale of cities is expanding, and urban and rural

infrastructure construction is constantly improving. Under this background, many factors

affecting regional surface temperature variation have appeared in many areas of China,

such as land use type variation, population aggregation, urban heat island and so on.

According to the data of urban population, urban GDP scale, national road traffic network and the like, the influence of social and economic development on China's surface temperature is analyzed. These socio-economic data come from National Bureau of

Statistics, Urban Statistical Yearbook and ENVI system database.

2.1 surface temperature data reconstruction

After repairing and reconstructing MODIS monthly scale surface temperature data by

using ground stations and neighbor pixels, the accuracy of the reconstructed data is

verified. In this study, some meteorological stations in China were randomly selected

again, and the reconstructed remote sensing surface temperature data were verified by

monthly scale station surface temperature data. The accuracy verification results of

surface temperature data in different months are shown in Figure 4. It can be seen that the

average accuracy error between the reconstructed surface temperature data and the

surface temperature of the ground station is about 1.5K, R 2 is above 0.95, and the data

covers a more complete pixel area, which can better meet the accuracy requirements of

spatio-temporal analysis. Meanwhile, the invention finds that the accuracy errors of

surface temperature data in different months are different. Generally, the data accuracy in

summer and autumn is relatively low and that in winter and spring is relatively high.

2.2 Analysis of spatio-temporal variation of surface temperature

2.2.1 Spatial distribution of surface temperature

Based on the reconstructed surface temperature data set, firstly, the spatial distribution of

surface temperature in China is studied and analyzed. The spatial distribution of China's

average surface temperature from 2002 to 2018 is shown in Figure 6. It can be found that

the spatial distribution of China's surface temperature varies greatly due to the

comprehensive influence of latitude, altitude, atmospheric circulation, land cover and other factors. The highest annual average surface temperature is located in tropical

Hainan Island, which can reach 27.85°C, mainly due to low latitude and sufficient heat.

The lowest temperature is located along the Altai Mountains in northern Xinjiang, and

the annual average surface temperature can be as low as -21.87°C, which is mainly

caused by higher latitude and higher altitude. The maximum difference between surface

high temperature and low temperature in China is 49.72°C. In addition, based on the

transit time of Terra and Aqua satellites, the average value of 01: 30 am and 10: 30 pm is

calculated as the night average surface temperature, and the average value of 10: 30 am

and 01: 30 pm is calculated as the daytime average surface temperature. By analyzing the

surface temperature conditions at day and night in China (Figure 7, Table 4), it is found

that there are obvious differences in the spatial distribution of surface temperature

between day and night in China. During the day, the highest temperature in China is

located in Tarim Basin, Xinjiang, and the highest temperature can reach 38.26°C, which

is mainly due to the sunny weather and wide desert in this area, which leads to the

absorption of solar radiation by the surface during the day and a large temperature rise.

The highest temperature area at night is located in the tropical Hainan island area, mainly

due to low latitude and sufficient heat. The spatial difference of surface temperature in

China at night (51.1°C) is smaller than that in daytime (54.7°C). In addition, in order to

analyze the distribution of surface temperature in different seasons in China, the average

values of March-May, June-August, September-November and December-February of

the following year are taken as the average surface temperatures in spring, summer,

autumn and winter (Figure 8, Table 4), and it is found that the average values of surface

temperatures in the four seasons in the whole region are arranged from high to low: summer > spring > autumn > winter. Except for summer, the highest temperatures in spring, autumn and winter are located in tropical Hainan Island, which are 29.47°C,

28.56°C and 23.31°C respectively. The highest temperature in summer is 43.43°C in

Turpan Basin, Xinjiang. The lowest temperature regions in the four seasons are mainly

located along the mountains in the west and northeast of China (such as Altai Mountains,

Tianshan Mountains, Kunlun Mountains and Daxing 'anling Mountains), and the lowest

average surface temperature in winter can reach -32.49°C, which is located in the west of

Daxing 'anling Mountains. The standard deviation of regional surface temperature

reflects the spatial difference of regional surface temperature. Comparing the standard

deviation of surface temperature in four seasons, it is found that the spatial difference of

surface temperature in winter is the largest (standard deviation =10.36°C), and the

difference between high temperature and low temperature can reach 55.8°C.

Table 4 Statistical Indicators of Surface Temperature in Different Time dimensions

Day Night Spring Summer Autumn Winter Annual Average 18.17 0.58 10.97 20.93 9.85 -4.34 9.31 value ('C) Maximum 38.26 22.13 29.47 43.43 28.56 23.31 27.85 value (°C) Minimum -16.44 -28.97 -21.30 -12.61 -22.23 -32.49 -21.87 value (°C) Range (°C) 54.7 51.1 50.77 56.04 50.79 55.8 49.72 Standard deviation 7.65 9.36 8.11 8.51 7.78 10.36 7.82 ( 0C)

On this basis, the spatial distribution of the temperature difference between day and night

and the temperature difference between winter and summer in China is analyzed based on

the average conditions of day and night and four seasons of surface temperature (Figure

9). In terms of temperature difference between day and night, due to the difference of thermal properties between land and sea, the temperature difference between day and night in eastern coastal areas is smaller than that in western inland areas as a whole.

Secondly, due to the differences in land cover types and land and water thermal

properties, the temperature difference between day and night in rivers and lakes is

significantly lower than that in other surrounding land areas, such as the Yellow River

Basin, the Yangtze River Basin, the Qinghai-Tibet Plateau Lake Area and the Yangtze

River estuary lakes. However, the arid and semi-arid areas in northwest China, especially

the desert areas in Xinjiang, have the largest temperature difference between day and

night due to the special sand cover, which provides convenient conditions for the

production of fruits and other agricultural products in this area. In terms of seasonal

temperature difference, affected by tropical climate, the area south of Tropic of Cancer

has sufficient heat and stable climate all year round, and the overall seasonal temperature

difference is small, especially in Yunnan and Hainan. Affected by the monsoon

circulation in winter and summer, the temperature difference between winter and summer

is larger in the area north of 40 degrees north latitude, especially in inland areas,

especially in northern Xinjiang and the west of Daxing 'anling Mountains. In addition, the

temperature difference between winter and summer is relatively small in mountains and

plateaus such as Tianshan Mountains, Kunlun Mountains and the southern part of

Qinghai-Tibet Plateau due to low perennial temperature and snow cover. Seasonal

temperature difference in different regions has an important impact on the geographical

features, vegetation growth, agricultural production and human living habits in the

region. On the whole, topography, land type and land-sea distribution are important

factors affecting the spatial distribution of temperature difference between day and night in China. The spatial difference of temperature difference between winter and summer is mainly related to latitude, terrain conditions and land cover.

2.2.2 Time series variation trend of surface temperature

Firstly, the anomaly value of annual average surface temperature in China from 2002 to

2018 is calculated (Figure 10), and the interannual variation trend of surface temperature

in China is analyzed as a whole. The analysis results show that China's surface

temperature is in a fluctuating growth state from 2002 to 2018, with a slow growth rate of

0.008°C/year on average. According to the interannual variation range of LST, the

interannual fluctuation of surface temperature in China is severe, especially from 2012 to

2013, with an interannual variation range of 1.22°C. During 2002-2018, LST was higher

than average in 10 years, especially in 2007 (0.57°C). LST was lower than average in 7

years, especially in 2012 (-0.92°C). To some extent, the drastic interannual variation of

LST reflect the unstable climatic conditions and frequent meteorological disasters in

China, which pose a serious threat to agricultural production in different regions of

China.

Figure 11 shows the variation trend of LST anomaly in day, night and four seasons. From

the perspective of day and night, the variation trend of surface temperature during day

and night is generally consistent, and the interannual variation range at night is smaller

than that during day, and both of them reach the lowest temperature in 2012. From the

seasonal dimension, the interannual variation trend of surface temperature in the four

seasons is significantly different. On the whole, the interannual variation of LST is the

largest in winter, followed by spring and autumn, and the smallest in summer. From 2002

to 2018, there are great differences in the years when the highest temperature and the lowest temperature appear in each season. The highest temperatures in spring, summer, autumn and winter appears in 2018, 2013, 2006 and 2016 respectively, and the lowest temperatures appears in 2010, 2003, 2012 and 2010 respectively. The variation trend of surface temperature in different seasons in the same year is significantly different, even showing a reverse trend, such as summer, autumn, winter and spring in 2010 and winter and autumn in 2016. The surface temperature of the four seasons only keeps the same trend in 2012-2013.

2.3 Abnormal variation of surface temperature in long time series

On the basis of the research on the variation trend of regional integral LST, the

interannual variation of surface temperature in different regions in China is analyzed by

using the least square method from the pixel scale. Firstly, the LST interannual variation

rates of all regions in China from 2002 to 2018 are calculated by using a linear model.

Secondly, the significance test is carried out on the calculated LST interannual variation

rates, and the regions with significant linear variation in LST are emphatically analyzed.

Figure 12 shows the interannual variation trend of annual average LST in China from

2002 to 2018. It is found that in the past 17 years, most areas in China have shown a

warming trend, which is mainly distributed in arid and semi-arid areas in northwest

China, southern Qinghai-Tibet Plateau, eastern North China Plain and Yangtze River

Delta. Among them, the central Inner Mongolia, the area east of Taihang Mountain in

North China Plain, the area south of Qilian Mountain in Qinghai-Tibet Plateau, Chengdu

Plain and Yangtze River Delta showed a significant warming trend, with an average

interannual warming rate of 0.08°C/year. During 2002-2018, the main cooling regions in

China were distributed in the northeast and most of the south, especially in the south of

Guangxi and Guangdong provinces. The average interannual variation rate of LST in all

significant linear cooling regions reached -0.07°C/year.

From the day-night dimension (Figure 13), there are significant differences in spatial

distribution of LST trends between day and night. Relatively speaking, the warming area

during the day is larger than that at night, especially the area with significant linear

warming is much larger than that at night. The results show that the interannual variation

rate of LST in daytime is 0.02°C/year, and the obvious linear warming areas are the

central Inner Mongolia, the southern Qinghai-Tibet Plateau, the eastern North China

Plain, the Chengdu Plain, the middle and lower reaches of the Yangtze River Plain and

the Yangtze River Delta, with an average interannual variation rate of 0.12°C/year. The

areas with a decreasing trend of LST during the day are mainly distributed in Northeast

China and most areas in the south, especially Guangxi and Guangdong provinces and

parts of Northeast Plain and Sanjiang Plain. The interannual variation rate of LST in all

significant linear cooling areas reaches -0.11°C/year. On the whole, the surface

temperature at night is still on the rise, and the interannual variation rate of LST at night

in the whole region reaches 0.01°C/year. Significant linear warming areas are mainly

distributed in Tarim River Basin, Yellow River Basin along Helan Mountain-Yinshan

Mountain, northeast of Sanjiang Plain and some rivers and lakes in Qinghai-Tibet

Plateau, with an interannual variation rate of 0.08°C/year.

From the seasonal dimension (Figure 14), the regional distribution of warming and

cooling in spring, summer, autumn and winter is also significantly different, and the

overall warming trend is in spring (0.04°C/year), summer (0.03°C/year) and winter

(0.04°C/year), and the overall cooling trend is in autumn (-0.01°C/year). The warming areas in spring are mainly distributed in the arid and semi-arid areas of Northwest China and the eastern part of North China Plain, especially in the central Inner Mongolia, along the eastern Taihang Mountains and in the Yangtze River Delta, showing a significant linear warming trend, with an interannual variation rate of LST of 0.16°C/year. The warming areas in summer are widely distributed, and the significant linear warming areas are mainly distributed in Tarim Basin, along the yinshan mountains, southern Liaodong

Peninsula, eastern Taihang Mountains, Shandong Peninsula and Yangtze River Delta, etc.

The interannual variation rate of LST in these areas is 0.1°C/year. There are few warming

areas in autumn, and the obvious linear warming area is the southwest of Qinghai-Tibet

Plateau, with a warming range of 0.12°C/year. Secondly, the northern part of Xinjiang

shows a great trend of cooling, with the interannual variation rate of LST reaching

0.1°C/year, and the linear trend of cooling is obvious. In winter, the warming area is

relatively small, but the warming range is large. Significant linear warming areas are

mainly distributed in the Yellow River Basin along Tianshan Mountains and Yinshan

Luliang Mountains, and the average interannual variation rate of LST is 0.19°C/year.

The above is only a specific embodiment of the present invention, but the protection

scope of the present invention is not limited to this. Any modification, equivalent

substitution and improvement made within the spirit and principle of the present

invention by any person familiar with the technical field within the technical scope

disclosed by the present invention should be covered within the protection scope of the

present invention.

Claims (10)

THE CLAIMS DEFINING THE INVENTION ARE AS FOLLOWS:

1. A method for repairing earth's surface temperature, characterized in that: the method

for repairing land and ocean temperatures comprising the following steps: reconstructing

scale surface temperature data by constructing a data repair model based on MODIS

surface temperature data; the spatio-temporal variation characteristics of surface

temperature in different time dimensions are analyzed from the whole, local and single

pixel multiple spatial scales, and the abnormal variation areas of surface temperature are

determined; combined with the data of surface, atmosphere, ocean and social and

economic activities, the driving factors of the spatio-temporal variation of surface

temperature, especially the interannual abnormal variation, are analyzed.

2. The method for repairing land and ocean temperatures according to claim 1,

characterized in that the method for repairing land and ocean temperatures comprises the

following steps:

step 1, using the computer language program to obtain corresponding data in batches, and

performing data extraction, coefficient conversion, projection conversion, cropping, site

data sorting and other preprocessing on the obtained data;

step 2, combining LST remote sensing data, ground station data, altitude data and

neighbor pixels, a data repair model is constructed for data reconstruction, and the

accuracy of the reconstructed data is verified by using independent monthly surface

temperature data of meteorological stations;

step 3, generating high-quality surface temperature data, analyzing the spatio-temporal

variation of surface temperature by using GIS, IDL, SPSS and other tools, and

determining the abnormal variation trend of surface temperature;

Step 4, combining various parameters of surface, atmosphere, ocean and socio-economic

activities, using various geostatistical analysis methods to analyze the driving factors

affecting the spatio-temporal variation of surface temperature.

3. The method for repairing land and ocean temperatures according to claim 2,

characterized in that in step 1, the data includes MODIS surface temperature data, ground

station data and other related data;

other related data include altitude, NDVI, soil moisture, atmospheric water vapor, ENSO

index, NAO index and IOBW index.

4. The method for repairing land and ocean temperatures according to claim 2,

characterized in that in step 2, the data reconstruction comprises the following steps:

the quality control data set in MODIS original surface temperature data is used to

evaluate the quality of monthly surface temperature data, and pixels with accuracy lower

than 2 K and missing information are selected as the area to be repaired of monthly data;

setting pixels with accuracy lower than 2 K and missing information in the surface

temperature daily data under the control of the area to be repaired as invalid pixels;

the ground station data in the corresponding time scale is assigned to some invalid pixels

according to longitude and latitude coordinates, and the remaining invalid pixels are

interpolated and repaired by using the neighbor value substitution method based on

DEM;

new monthly data of surface temperature are generated by average synthesis of the

repaired daily data.

5. The method for repairing land and ocean temperatures according to claim 4,

characterized in that the interpolation repair of the remaining invalid pixels by using the

neighbor value substitution method based on DEM comprises the following steps:

firstly, the pixel value of the original surface temperature is restored to the pixel value at

an altitude of 0 m, and the calculation formula is as follows:

LSTon= LST +0.55*(ele/ 100).

wherein, ele represents the elevation of the pixel; LST represents MODIS surface

temperature daily data; LSTom indicates the surface temperature when MODIS surface

temperature daily data is restored to 0 m;

secondly, the invalid pixels are interpolated and repaired by the neighbor pixel value

substitution method;

finally, the interpolated pixel value is restored to the surface temperature value of the

corresponding elevation, and the calculation formula is as follows:

LSTeic = LSTnibblc-0.55*(ele/100).

Wherein, LSTibbie represents the result of interpolation using the neighbor pixel

substitution method and LSTee represents the surface temperature restored to the

corresponding elevation after neighbor interpolation;

6. The method for repairing land and ocean temperatures according to claim 2,

characterized in that in step 3, analyzing the spatio-temporal variation of the surface

temperature comprises the following steps:

the least square method is used to calculate the interannual variation rate of surface

temperature in a long time series, and the calculation formula is as follows: n (kx Tk)- k Tk b= k=1 k= 1 n k 2 -( k) 2 k=1 k=1 wherein, b is the variation rate of surface temperature with year; k is the time series value;

Tk is the average surface temperature in the k-th year; n is the total number of years;

significance test is performed on the calculated variation rate, and the area with

significant linear temperature increase or decrease is extracted as the abnormal variation

area of surface temperature.

7. The method for repairing land and ocean temperatures according to claim 2,

characterized in that in step 4, analyzing the driving factors influencing the spatio

temporal variation of the surface temperature by using a plurality of geostatistical

analysis methods comprises the following steps:

(1) analyzing the correlation between surface temperature and vegetation index, soil

moisture, atmospheric water vapor and precipitation, expressing the correlation by

Pearson coefficient, and testing the significance of the calculated correlation coefficient;

the formula of correlation coefficient is as follows:

Y(at - a)(bi - b) _1

Z(ai -a) 2 * i=1 $(b i=I - b) 2

in the formula, r is Pearson index, which indicates Pearson relationship between surface

temperature and other parameters; ai is the surface temperature of different years; bi is

other surface atmospheric parameters in different years; a and b are the average values

of surface temperature and other parameters respectively; n is the total number of years;

(2) using linear multiple stepwise regression model to analyze the influence degree of

latitude, longitude, altitude, vegetation index, soil moisture, atmospheric water vapor and

precipitation on surface temperature.

8. The method for repairing land and ocean temperatures according to claim 7,

characterized in that the step of using linear multiple stepwise regression model to

analyze the influence degree of latitude, longitude, altitude, vegetation index, soil

moisture, atmospheric water vapor and precipitation on surface temperature comprises

the following steps:

linear multiple stepwise regression models are established for different regions to analyze

the influence degree of latitude, longitude, altitude, vegetation index, soil moisture,

atmospheric water vapor and precipitation on surface temperature;

the linear model of the linear multiple stepwise regression is as follows:

P y = $zkxzk-+ e k=1

among them, yz and Xzk (k = 1, 2......p) represent standardized dependent and

independent variables respectively, p represents the number of factors, z represents

standardized coefficient, and c represents residual;

the method for repairing land and ocean temperatures further comprises the following

steps: determining the lag time of the response variation of the surface temperature in

different regions to the climate mode by comparing the correlation coefficient between

the surface temperature anomaly value and the climate modal index under different lag

times;

the significance test comprises the following steps: adopting a T test method as follows:

R* n- 2 T=T1=1- R 2

. in the formula, R is the correlation coefficient between surface temperature and time; n is

the number of samples; statistic Tc follows T distribution with degree of freedom n-2;

given the significance level a, if ITcI > Ta, the original hypothesis is rejected, indicating

that the linear variation trend of surface temperature is significant;

the correlation coefficient between surface temperature and time is calculated as follows:

n (kxT) JkYT, R -k -I k-I k -I

nIk2 (k)2x n k T)2 k=I k=i k=i k=i

wherein, k is a time series value; T is the average surface temperature in the k-th year; n

is the total number of years.

9. A computer equipment, characterized in that the computer equipment comprises a

memory and a processor, wherein the memory stores a computer program, and when the

computer program is executed by the processor, the processor executes the following

steps: reconstructing scale surface temperature data by constructing a data repair model

based on MODIS surface temperature data; the spatio-temporal variation characteristics

of surface temperature in different time dimensions are analyzed from the whole, local

and single pixel multiple spatial scales, and the abnormal variation areas of surface

temperature are determined; combined with the data of surface, atmosphere, ocean and

social and economic activities, the driving factors of the spatio-temporal variation of

surface temperature, especially the interannual abnormal variation, are analyzed.

10. A land and ocean temperature repair system implementing the method for repairing

land and ocean temperatures according to any one of claims 1 to 8, characterized in that

the land and ocean temperature repair system comprises:

a data preprocessing module, which is used for obtaining corresponding data in batches

by using computer language programs, and performing data extraction, coefficient

conversion, projection conversion, cropping, site data sorting and other preprocessing on

the obtained data;

a data accuracy verification module, which is used for constructing a data repair model

for data reconstruction by combining LST remote sensing data, ground station data,

altitude data and neighbor pixels, and verifying the accuracy of the reconstructed data by

using independent monthly surface temperature data of meteorological stations;

an abnormal variation trend determination module, which is used for generating high

quality surface temperature data, analyzing the spatio-temporal variation of surface

temperature by using GIS, IDL, SPSS and other tools, and determining the abnormal

variation trend of surface temperature;

A driving factor analysis module, which is used for analyzing the driving factors

influencing the spatio-temporal variation of the surface temperature by combining

various parameters of the surface, the atmosphere, the ocean and social and economic

activities and utilizing various geostatistical analysis methods.