CN114254802A - Prediction method of vegetation coverage space-time change under climate change drive - Google Patents

Prediction method of vegetation coverage space-time change under climate change drive Download PDF

Info

Publication number
CN114254802A
CN114254802A CN202111322031.5A CN202111322031A CN114254802A CN 114254802 A CN114254802 A CN 114254802A CN 202111322031 A CN202111322031 A CN 202111322031A CN 114254802 A CN114254802 A CN 114254802A
Authority
CN
China
Prior art keywords
year
data
pixel
vegetation
vegetation coverage
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Granted
Application number
CN202111322031.5A
Other languages
Chinese (zh)
Other versions
CN114254802B (en
Inventor
章文波
陈楷琳
郝玮
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Beijing Normal University
Original Assignee
Beijing Normal University
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Beijing Normal University filed Critical Beijing Normal University
Priority to CN202111322031.5A priority Critical patent/CN114254802B/en
Publication of CN114254802A publication Critical patent/CN114254802A/en
Application granted granted Critical
Publication of CN114254802B publication Critical patent/CN114254802B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00—Administration; Management
    • G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00—Administration; Management
    • G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
    • G06Q10/067—Enterprise or organisation modelling
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q50/00—Systems or methods specially adapted for specific business sectors, e.g. utilities or tourism
    • G06Q50/02—Agriculture; Fishing; Mining
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02A—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE
    • Y02A90/00—Technologies having an indirect contribution to adaptation to climate change
    • Y02A90/10—Information and communication technologies [ICT] supporting adaptation to climate change, e.g. for weather forecasting or climate simulation

Landscapes

  • Business, Economics & Management (AREA)
  • Engineering & Computer Science (AREA)
  • Human Resources & Organizations (AREA)
  • Strategic Management (AREA)
  • Economics (AREA)
  • Physics & Mathematics (AREA)
  • Entrepreneurship & Innovation (AREA)
  • Marketing (AREA)
  • Tourism & Hospitality (AREA)
  • General Business, Economics & Management (AREA)
  • General Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Game Theory and Decision Science (AREA)
  • Operations Research (AREA)
  • Quality & Reliability (AREA)
  • Development Economics (AREA)
  • Educational Administration (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Agronomy & Crop Science (AREA)
  • Animal Husbandry (AREA)
  • Marine Sciences & Fisheries (AREA)
  • Mining & Mineral Resources (AREA)
  • Health & Medical Sciences (AREA)
  • General Health & Medical Sciences (AREA)
  • Primary Health Care (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)

Abstract

The invention provides a method for predicting vegetation coverage space-time change under the drive of climate change, which comprises the following steps of 1: determining a research area and a data source for establishing a vegetation prediction model; step 2: taking the historical actually measured vegetation cover grid data as dependent variables, introducing different lags as independent variables, constructing pixel-by-pixel multiple linear regression models, synthesizing the decision coefficients of the models, the five-year average analysis method and the result of the trend rate analysis, and screening and constructing the optimal multiple linear regression model; and step 3: calculating vegetation coverage grid data in a future time period by taking the meteorological data in the future time period as input; and 4, step 4: carrying out downscaling processing according to the vegetation coverage grid data and the historical observation vegetation coverage data in the future period; and 5: and drawing a future annual change curve. The method adopts a pixel-by-pixel mode to construct a regression model, obtains the regression model with higher fitting degree in meteorological data with lower space-time resolution, and predicts the response of vegetation coverage to climate change in a future period.

Description

Prediction method of vegetation coverage space-time change under climate change drive
Technical Field
The application relates to the technical field of vegetation cover coverage simulation, in particular to a prediction method of vegetation cover space-time change under the drive of climate change.
Background
In recent yearsGlobal warming is the main feature of climate change, and climate change evaluation reports indicate climate change and the impact on socioeconomic performance, and are adapted to recent research progress on possible countermeasures for mitigating climate change. The Normalized Difference Vegetation Index (NDVI) caused by climate change fluctuates back and forth within a certain range. In the terrestrial ecosystem, vegetation coverage changes are affected by a variety of factors, which are generally divided into two categories, terrain, climate, CO2The concentration, nitrogen and phosphorus sedimentation and other natural factors are human activity factors such as afforestation, agricultural modernization, population increase, economic development, urbanization, excessive reclamation, grazing and the like. The climate warming can affect the vegetation climate and the primary productivity of an ecological system, and the vegetation growth period in spring and autumn is increased, so that the vegetation coverage is increased. Research shows that under the future global warming background, the activity of the vegetation on the global, especially in the high latitude area in the northern hemisphere, is enhanced, and the vegetation develops towards greening. Meanwhile, the vegetation growth has certain hysteresis to the climate change, but the generated results are different due to different research scales.
At present, some scholars model and predict the change between meteorological data and vegetation coverage, but the range on the spatial scale is small, the meteorological data which can be acquired by a small scale is more specific and detailed, but in the national, continent and even global large-scale space, the detailed meteorological data or other data is difficult to acquire, and few scholars conduct vegetation coverage simulation research.
Therefore, it is desirable to provide a method for predicting the spatiotemporal variation of vegetation coverage under the drive of climate change.
Disclosure of Invention
In order to overcome the defects of the prior art, the invention aims to provide a method for predicting vegetation coverage space-time change under the drive of climate change. And carrying out downscaling processing on the vegetation cover data based on historical observation, and obtaining a final result of 24 and a half months of vegetation cover grid data year by year in the future at a kilometer resolution level. Obtaining a regression model with higher fitting degree in meteorological data with lower space-time resolution, predicting the response of vegetation coverage to climate change in a future period, and constructing a pixel-by-pixel multivariate linear regression model by taking historical actual measurement vegetation coverage grid data of 24 and a half months year after year as dependent variables and simultaneously taking meteorological grid data of the period as independent variables. In the process of constructing the regression model, the vegetation coverage data of the grid is used as a dependent variable, the temperature and precipitation data of half a month (month), one month later, three months later, one year and one half later and two years later are used as independent variables, the independent variables are obviously selected through simulation, the multivariate linear regression model of each grid point is obtained, and the accuracy comparison is carried out by taking the historical observed vegetation data as reference. By the method, a vegetation coverage space distribution map at any time period in the hundred years in the future under the background of global change or an annual change curve of vegetation coverage at a specified position point in different seasons can be drawn, and the response of the future vegetation coverage change to climate change is revealed.
In order to achieve the purpose, the solution adopted by the invention is as follows:
a method for predicting vegetation coverage space-time change under the drive of climate change comprises the following steps:
step 1: determining a research area and a data source for establishing a vegetation prediction model, specifically comprising the following steps:
step 11: determining a region to be predicted, wherein the region to be predicted comprises a spatial large-scale region with spatial diversity;
step 12: collecting a data source of the area to be predicted, wherein the data source comprises historical measured vegetation coverage grid data of 24 and a half months in a historical period year by year and meteorological grid data with the same time interval and the same space-time resolution, and the meteorological grid data comprise precipitation data and gas temperature data;
step 2: taking the historical actual measurement vegetation cover grid data of 24 and a half months year after year in the step 1 as a dependent variable, introducing hysteresis as an independent variable, constructing a pixel-by-pixel multiple linear regression model, and setting an optimal judgment condition to obtain an optimal multiple linear regression model;
and step 3: selecting meteorological data simulated in a future time period under a global climate mode as input according to the optimal multiple linear regression model obtained in the step 2, and calculating a preliminary result N of vegetation cover grid data of 24 and a half months year by year in the future time periodG；
And 4, step 4: according to the preliminary result of the vegetation coverage grid data of 24 and a half months year by year in the future period obtained in the step 3 and the historical observation vegetation coverage data N of the area to be predictedACarrying out downscaling processing to obtain a final result N of vegetation cover grid data of 24 and a half months in the future year by year, wherein the resolution of the area to be predicted can reach 250m multiplied by 250m at mostGnew：
NGnew＝α×(NG+γ)-γ
Figure BDA0003345859440000021
In the formula: gamma is a downscaling parameter; alpha is a downscaling proportionality coefficient;
and 5: according to the final result N of the vegetation coverage grid data of 24 and a half months in the kilometer resolution level of the area to be predicted obtained in the step 4 year by yearGnewDrawing a vegetation coverage space distribution map of a future time period; according to the final result N of the vegetation coverage grid data of 24 and a half months in the kilometer resolution level of the area to be predicted obtained in the step 4 year by yearGnewAnd drawing an annual variation curve covered by planting at a specified time, wherein the specified time comprises the year average, different seasons and different months in a specified range.
Preferably, the step 2 specifically includes the following steps:
step 21: the lag comprises meteorological grid data with the same space-time resolution ratio of the current half month, one month, three months, half year, one year and two years, and vegetation coverage data of each pixel is used as a dependent variable to obtain a multiple linear regression model of each pixel;
step 22: performing precision evaluation on the multiple linear regression model established in the step 21, specifically including calculating a decision coefficient of the model, calculating a correlation coefficient by adopting a five-year average analysis method aiming at the absolute change of the simulation result and the five-year absolute change of the measured data, calculating a tendency rate of the simulation result and the measured result by adopting a tendency rate analysis method, and performing correlation coefficient of the simulation result and the measured result;
step 23: integrating the decision coefficient of the model calculated in the step 22, the five-year average analysis method and the result of the trend rate analysis, judging whether the multiple linear regression model established in the step 21 meets the optimal judgment condition, and obtaining the optimal multiple linear regression model when the multiple linear regression model established in the step 21 meets the optimal judgment condition; otherwise, return to step 21.
Preferably, the decision coefficient of the calculation model itself in step 22 is specifically:
Figure BDA0003345859440000031
in the formula:
Figure BDA0003345859440000032
determining coefficients corresponding to pixel grids in the ith row and the jth column;
Figure BDA0003345859440000033
the predicted value of a multivariate linear regression model of the vegetation cover pixel value on a time sequence is obtained;
Figure BDA0003345859440000034
the mean value of the vegetation covering pixel value on the time sequence is obtained; y is(i,j)And covering the pixel value for the vegetation of the ith row and the jth column in the grid data.
Preferably, the step 22 of calculating a correlation coefficient for the absolute change of the simulation result and the five-year absolute change of the measured data by using a five-year average analysis method specifically includes: selecting a time interval corresponding to the data source in the step 1, simulating by using the multiple linear regression model of each pixel to obtain a dependent variable, actually measuring the dependent variable of each five-year early, middle and late time intervals in the time interval corresponding to the data source to be A, B, C respectively, calculating a five-year average value of the dependent variable and the actually measured dependent variable obtained by simulation, calculating absolute changes of the early, middle and late time intervals to be B-A, C-B and C-A to obtain an absolute change of a simulation result and an absolute change of actually measured data, and calculating a correlation coefficient of the absolute change of the simulation result and the absolute change of the actually measured data.
Preferably, in the step 22, the trend rate analysis method is adopted to calculate the trend rates of the simulation result and the actual measurement result, and the correlation coefficient between the simulation result and the actual measurement result is specifically: and (3) taking the time sequence as an independent variable and the vegetation coverage data of each pixel as a dependent variable to obtain a unitary linear equation of each pixel with respect to the time sequence:
Y(i,j)＝b(i,j)+a(i,j)t(i,j)
in the formula: y is(i,j)Covering pixel values for the vegetation of the ith row and the jth column of pixels in the grid data; t is t(i,j)The time sequence is the ith row and the jth column of the pixel; a is(i,j)Is one tenth of the tendency rate of the ith row and the jth column of the image elements;
calculating the tendency rate of the simulation result and the actual measurement result, and calculating the correlation coefficient of the simulation result and the actual measurement result.
Preferably, the multiple linear regression model for each pixel obtained in step 21 is specifically:
Y(i,j)＝β0(i,j)+β1(i,j)X1(i,j)+β2(i,j)X2(i,j)+……+βk(i,j)Xk(i,j)
in the formula: beta is a0(i,j)Constant items corresponding to the ith row and the jth column of pixel grids in grid data; beta is ak(i,j)Model coefficients corresponding to the k-th meteorological data independent variables of the ith row and jth column of pixels in the grid data; xk(i,j)And the independent variable of the kth meteorological data of the ith row and jth column pixel in the grid data.
Preferably, the optimal determination conditions in step 2 are:
PDI＞0.5＞70％
CIyear-avg＞0.9
CIavg＞0.5
CIrate＞0.6
in the formula: pDI＞0.5The pixel number proportion of a pixel number of a pixel-by-pixel multiple linear regression model of the region to be predicted, wherein the coefficient of determination DI is greater than 0.5; CIyear-avgThe correlation coefficient of the vegetation coverage results of two five-year average years in the same period in the five-year average analysis method; CIavgThe average value of the correlation coefficient of the five-year average space-time change is obtained; CIrateThe correlation coefficient is calculated by adopting a tendency rate analysis method.
Preferably, the area of the region to be predicted in step 11 is 50 km2The above.
Compared with the prior art, the invention has the beneficial effects that:
in the existing research, scholars establish model relations between meteorological data and vegetation coverage data, but the general scale is small and is mostly the scale of a certain small watershed or county level, the spatial diversity is not strong under the scale, a single model is mostly used for establishing relations, and the vegetation coverage data and the detailed meteorological data are easy to obtain. The method is used for establishing a model of meteorological data and vegetation coverage data one by one aiming at a large-scale area, fully considering the spatial diversity in the natural ecological environment, and obtaining a regression model which is more consistent with the natural law so as to simulate and predict the response of vegetation coverage to climate change in the future period. In addition, on the national scale, the continent or the global scale, the spatial-temporal resolution of the acquired meteorological data and the future climatic mode data is low, so that the resolution of the result of simulation prediction is low, and the requirements of some fine researches cannot be met.
Drawings
FIG. 1 is a block flow diagram of a method for predicting spatiotemporal changes in vegetation coverage driven by climate change in accordance with an embodiment of the present invention;
FIG. 2 is a graph of a hierarchical statistics of model decision coefficients for pixel-by-pixel linear hysteresis for one year in 1982-2012 in accordance with the present invention;
FIG. 3 is a graph of statistical significance grading of a one-year model of NDVI pixel-by-pixel linear hysteresis in 1982-2012 in accordance with an embodiment of the present invention;
FIG. 4 is a graph of the five-year average annual NDVI distribution of the VIP data of the present example of the invention in 1982-1986;
FIG. 5 is a graph of the five-year average annual average NDVI distribution of CRU data simulation in 1982-1986 according to an embodiment of the present invention;
fig. 6 is a graph of VIP data 2008-;
FIG. 7 is a graph of the CRU data simulation 2008-;
FIG. 8 is a graph of graded statistics of the five-year average absolute change in NDVI for the historical period VIP data 1982-1986 and 1995-1999 in accordance with an embodiment of the present invention;
FIG. 9 is a hierarchical statistical plot of the five-year average absolute change in NDVI for the historical period CRU data 1982-1986 and 1995-1999 according to an embodiment of the present invention;
FIG. 10 is a graph of graded statistics of the average absolute change in NDVI for five years in the historical period VIP data 2008-;
FIG. 11 is a graph of historical CRU data 2008-;
FIG. 12 is a graph of five-year average absolute change grading statistics of NDVI in the historical period VIP data 2008-;
FIG. 13 is a hierarchical graph of the five-year average absolute change NDVI of the CRU data 2008-;
FIG. 14 is a graph of the annual NDVI pixel-by-pixel trend rate grading statistics of VIP data in 1982-2012 in accordance with the present invention;
FIG. 15 is a graph of the image element-by-image element tendency rate grading statistics of the CRU data in 1982-2012 lagging by an image element for one year model simulation of the annual average NDVI;
FIG. 16 is a statistical chart of significance grading of the annual average NDVI per pixel tendency rate of VIP data in 1982-2012 in accordance with the present invention;
FIG. 17 is a statistical chart showing the significance of the pixel-by-pixel tendency rate of the CRU data in 1982-2012 lagging by one year model simulation annual average NDVI;
FIG. 18 is a graph of statistics of the annual NDVI per-pel trend rate determination coefficients of VIP data in 1982-2012 in accordance with the present invention;
FIG. 19 is a graph showing the statistics of the decision coefficient grading of the pixel-by-pixel tendency rate of the CRU data in 1982-2012 lagging by one year model simulation annual average NDVI;
FIG. 20 is a graph of year-averaged NDVI simulation 2020 for GFDL-ESM2M model data in the pre-downscaling RCP4.5 emissions scenario in accordance with an embodiment of the present invention;
FIG. 21 is a graph of year-averaged NDVI simulation 2020 for GFDL-ESM2M mode data in a post-downscaling RCP4.5 emissions scenario in accordance with an embodiment of the present invention;
FIG. 22 is a graph of simulated 2050 year-averaged NDVI distribution of GFDL-ESM2M model data in a pre-downscaling RCP4.5 emission scenario in accordance with an embodiment of the present invention;
FIG. 23 is a graph of simulated 2050 year-averaged NDVI distribution of GFDL-ESM2M model data in a post-downscaling RCP4.5 emissions scenario in accordance with an embodiment of the present invention;
FIG. 24 is a graph of year-averaged NDVI simulation 2020 for GFDL-ESM2M model data in the RCP4.5 emissions situation after downscaling according to an embodiment of the present invention;
FIG. 25 is a graph of simulated 2050 year-averaged NDVI distribution of GFDL-ESM2M model data in the RCP4.5 emissions scenario after downscaling in accordance with an embodiment of the present invention;
fig. 26 is a graph illustrating the change of annual curves of the average annual NDVI within a specified range in a future period, for example, in the beijing area, according to an embodiment of the present invention.
Detailed Description
Hereinafter, embodiments of the present invention will be described with reference to the drawings.
The invention provides a method for predicting vegetation coverage space-time change under the drive of climate change, which comprises the following steps:
step 1: determining a research area and a data source for establishing a vegetation prediction model, specifically comprising the following steps:
step 11: and determining a region to be predicted, wherein the region to be predicted comprises a spatial large-scale region with spatial diversity.
Step 12: collecting a data source of an area to be predicted, wherein the data source comprises historical measured vegetation coverage grid data of 24 and a half months in a historical period year by year and meteorological grid data with the same time interval and the same space-time resolution, and the meteorological grid data comprises precipitation data and air temperature data;
step 2: taking the historical actual measurement vegetation cover grid data of 24 and a half months year after year in the step 1 as dependent variables, introducing lag as independent variables, constructing a pixel-by-pixel multiple linear regression model, setting an optimal judgment condition, and obtaining an optimal multiple linear regression model;
and step 3: according to the optimal multiple linear regression model obtained in the step 2, selecting meteorological data simulated in a future time period in a global climate mode as input, and calculating a preliminary result N of vegetation coverage grid data of 24 and a half months year by year in the future time periodG；
And 4, step 4: according to the preliminary result of the vegetation coverage grid data of 24 and a half months year by year in the future period obtained in the step 3 and the historical observation vegetation coverage data N of the area to be predictedACarrying out downscaling processing to obtain a final result N of vegetation coverage grid data of 24 and a half months in the future year with the highest resolution of the area to be predicted up to 250m multiplied by 250mGnew：
NGnew＝α×(NG+γ)-γ
Figure BDA0003345859440000071
In the formula: gamma is a downscaling parameter; alpha is a downscaling proportionality coefficient;
and 5: according to the kilometer resolution level of the area to be predicted obtained in the step 4, the final result N of 24 and a half months of vegetation coverage grid data year by yearGnewDrawing a vegetation coverage space distribution map of a future time period; according to the kilometer resolution level of the area to be predicted obtained in the step 4, the final result N of 24 and a half months of vegetation coverage grid data year by yearGnewAnd drawing an interplanetary change curve covered in planting at a specified time, wherein the specified time comprises the annual average, different seasons and different months in a specified range.
The main and further preferred embodiments of the present invention will be described below by way of a specific example.
The embodiment of the invention provides a method for predicting Vegetation coverage space-time change under the drive of climate change, as shown in fig. 1, historical measured Vegetation coverage grid Data of 24 and a half months year by year are used as dependent variables, namely Data of a Multi-Sensor Vegetation Index and a normalized Vegetation Index product (VIP-NDVI) corresponding to a geochemical Earth Science Data record in 1982 and 2012 and every 15 days and periods, and Data of a meteorological grid are used as independent variables, namely weather Data of a clinical Research institute (CRU) of England university in British university in 2012 in 1982 and 2012, so as to construct a pixel-by-pixel multiple linear regression model. In the process of constructing the regression model, the vegetation coverage data of the grid is used as a dependent variable, and the temperature and precipitation data of the current half month/month, one month delay, three months delay, one year delay, one and one half delay and nonlinear terms are used as independent variables, so that the historical periods are integrated, for example: the five-year average analysis, the trend rate and the contemporaneous correlation coefficient comparison of the actually measured data VIPNDVI obtained in 1982-2012 and the simulated data obtained by taking CRU meteorological data as independent variables are carried out, and the specific standards are as follows:
1. the pixel number proportion P of the pixel number of the pixel-by-pixel multiple linear regression model of the region to be predicted, wherein the determination coefficient DI of the pixel-by-pixel multiple linear regression model of the region to be predicted is more than 0.5DI＞0.5Over 70%.
2. In the five-year average analysis method, two five-year average annual vegetation in the same periodCorrelation coefficient CI covering the resultyear-avgMean value CI of correlation coefficient of more than 0.9 and five-year mean space-time variationavgGreater than 0.5.
3. Correlation coefficient CI calculated by adopting trend rate analysis methodrateGreater than 0.6.
The three conditions are simultaneously satisfied, and an optimal multiple linear regression model is obtained.
According to the standards, the final model is determined to be a prediction model with the independent variables lagging for one year, namely the independent variables are current month rainfall, air temperature, previous month rainfall and air temperature, previous March accumulated rainfall and month temperature equilibrium, previous June accumulated rainfall and month temperature equilibrium, previous December accumulated rainfall and month temperature equilibrium, a multiple linear regression model of each lattice point is obtained, and accuracy comparison is carried out by taking historical observation vegetation data as reference.
The method comprises the steps of taking meteorological data of a global climate mode simulation in a certain period of time in the future as input, selecting the 2.0 Version (GFDL-ESM 2M) mode data of an Earth System Model of a United states Geophysical Fluid Dynamics Laboratory, and predicting to obtain a preliminary result of vegetation coverage grid data of 24 and a half months year by year in a pixel-by-pixel multiple regression Model. In the document, the historical period in the pattern data is calculated, namely 1982-2005, the pattern data is semilunar meteorological data, the CRU data is lunar data, and in order to keep the data magnitude and time scale consistent, the following changes are calculated when the vegetation index in the future period is predicted:
namely, the independent variables are: the accumulated precipitation and the average temperature of the current half month and the last half month, the accumulated precipitation and the average temperature of the first two and half months, the accumulated precipitation and the half month temperature of the first six and half months, the accumulated precipitation and the half month temperature of the first twelve and half months, and the accumulated precipitation and the half month temperature of the first twenty-four and half months. Here, the unit is a half month, for example, the first two half months means the first month.
And carrying out downscaling processing on the vegetation cover data based on historical observation, and obtaining a final result of 24 and a half months of vegetation cover grid data year by year in the future at a kilometer resolution level.
By the method provided by the invention, the vegetation coverage space distribution map of any time period in the future hundred years under the background of global change can be drawn, or the annual change curve of vegetation coverage in different seasons at a specified position point is shown, so that the response of the future vegetation coverage change to climate change is revealed. The method specifically comprises the following steps:
firstly, determining coefficient and significance of a one-year prediction model are delayed;
firstly, determining a model: the resolution of this model was 0.5 ° x 0.5 °.
Then, calculating a decision coefficient, specifically:
Figure BDA0003345859440000081
wherein:
Figure BDA0003345859440000082
the decision coefficient corresponding to the pixel grid of the ith row and the jth column,
Figure BDA0003345859440000083
is a predicted value of a multivariate linear regression model of vegetation cover pixel values on a time sequence,
Figure BDA0003345859440000084
is the average value of the vegetation covering pixel value over the time series, Y(i,j)And covering the pixel value for the vegetation of the ith row and the jth column in the grid data.
The independent variables are current month rainfall, air temperature, previous March accumulated rainfall, month temperature, previous June accumulated rainfall, month temperature, previous December accumulated rainfall, month temperature, and as shown in FIG. 2, the model decision coefficient is subjected to hierarchical statistics in a linear delay one year by one pixel in 2012 in 1982 plus materials, and Table 1 is data corresponding to a specific analysis range:
TABLE 1
Figure BDA0003345859440000091
Finally, significance statistics is performed, as shown in fig. 3, the significance grading statistics of the model for one year is performed by means of pixel-by-pixel linear hysteresis of NDVI in 1982-2012, and table 2 is the significance grading statistics of the linear hysteresis model:
TABLE 2
Linear lag model significance rank statistics
Figure BDA0003345859440000092
And secondly, analyzing the five-year average space-time variation of the VIP NDVI data, the NDVI obtained by CRU calculation and the NDVI obtained by GFDL calculation. The time scale of NDVI calculated by the VIP NDVI data and CRU is 2012 in 1982 and 2005 in GFDL, and the correlation coefficient comparison is only compared with that in the same time period;
the five-year average NDVI was calculated as shown in fig. 4 for VIP data 1982-:
TABLE 3
Figure BDA0003345859440000101
Fig. 6 shows the average annual average NDVI distribution in five years of 2008-:
TABLE 4
Figure BDA0003345859440000102
Figure BDA0003345859440000111
Table 5 is the five year average NDVI correlation coefficient for two time periods:
TABLE 5
1982-1986 VIP 2008 + 2012 VIP
1982-CRU in 1986 0.9990 -
2008 + 2012 CRU - 0.9979
Thirdly, the VIP NDVI data, the NDVI obtained by CRU calculation, the NDVI obtained by GFDL calculation, the annual average NDVI tendency rate of the three, the corresponding determination coefficient and the significance, and the correlation coefficient comparison is only compared with that of the three in the same time period;
the five-year average time-space change obtains the annual average NDVI from the annual monthly results predicted by the regression model, and the historical period, here the five-year average results of three periods of early, middle and late in 1982-2012, is selected to compare the time-space absolute change and the correlation coefficient, so that the accuracy of the model is determined to be higher. The years corresponding to the early, middle and late stages are 1982-1986, 1995-1999 and 2008-2012, respectively.
FIG. 8 is the five-year average absolute change ranking statistics of historical period VIP data 1982-1986 and 1995-1999, FIG. 9 is the five-year average absolute change ranking statistics of historical period CRU data 1982-1986 and 1995-1999, and Table 6 is the five-year average absolute change ranking statistics of historical period VIP data 1982-CRU data 1986 and 1995-1999:
TABLE 6
Figure BDA0003345859440000121
Fig. 10 is the historical period VIP data 2008-:
TABLE 7
Figure BDA0003345859440000122
Figure BDA0003345859440000131
Fig. 12 is the historical period VIP data 2008-:
TABLE 8
Figure BDA0003345859440000132
Figure BDA0003345859440000141
Table 9 correlation coefficients for VIP results:
TABLE 9
1999-1986 2012-1999 2012-1986
CRU 0.3815 0.6943 0.5204
The independent variable of the NDVI historical period tendency rate is 1-31, namely 31 years in 1982-2012, the dependent variable is the annual average value of each set of data pixel by pixel year, namely averaging every month from 1 to 12 months, and the 'determination coefficient' and 'significance' of the part are both the determination coefficient and the significance corresponding to the grid tendency rate equation of each pixel. And respectively calculating prediction data obtained by calculating historical actually measured data VIP-NDVI and CRU data according to the tendency rates, and comparing the results of the two tendency rates by using correlation coefficients.
Fig. 14 is a graded statistics of the annual average NDVI pixel-by-pixel tendency rate of VIP data in 1982-:
watch 10
Figure BDA0003345859440000142
Figure BDA0003345859440000151
Fig. 16 is statistics of significance grading of annual average NDVI pixel-by-pixel tendency rate of VIP data in 1982 + 2012, fig. 17 is statistics of significance grading of annual average NDVI pixel-by-pixel tendency rate of CRU data in 1982 + 2012 by pixel-by-pixel lag one-year model simulation, and table 11 is statistics of significance grading of annual average NDVI pixel-by-pixel tendency rate of VIP data in 1982 + 2012 by pixel-by-pixel lag one-year model simulation:
TABLE 11
Figure BDA0003345859440000152
Fig. 18 is a statistics of ranking coefficients of the annual average NDVI per-pixel tendency rate determination of VIP data in 1982 + 2012, fig. 19 is a statistics of ranking coefficients of the annual average NDVI per-pixel tendency rate determination of CRU data in 1982 + 2012 by pixel model simulation, and table 12 is a statistics of ranking coefficients of the annual average NDVI per-pixel tendency rate determination of VIP data in 1982 + 2012 by pixel model simulation:
TABLE 12
Figure BDA0003345859440000153
Figure BDA0003345859440000161
Table 13 is the correlation coefficient with VIP propensity:
watch 13
Correlation coefficient with VIP tendency rate
CRU 0.7208
Fourthly, comparing the NDVI before and after the dimension reduction of the future time is predicted;
the objective was to convert the simulated NDVI product with a resolution of 0.5 ° × 0.5 ° to a resolution of 10km × 10km (0.083 ° × 0.083 °). The method specifically comprises the following steps:
1. data preparation
Firstly, calculating the VIP NDVI with the resolution of 0.05 degrees multiplied by 0.05 degrees for a half month and a plurality of years to obtain the multi-year average NDVI of each half month in 1982 + 2005, and filling up the high latitude areas, wherein the calculation of the simulated high latitude areas uses the climate data with the resolution of 0.5 degrees multiplied by 0.5 degrees and the DEM with the resolution of 0.0083 degrees multiplied by 0.0083 degrees to resample to 0.05 degrees multiplied by 0.05 degrees to obtain the full global multi-year average NDVI of each half month, and the calculation of the simulated high latitude areas adopts bilinear interpolation to resample to 0.083 degrees multiplied by 0.083 degrees.
Next, the 0.5 ° × 0.5 ° NDVI results from the history period 1982-. Meanwhile, the annual semilunar NDVI of 0.5 degrees multiplied by 0.5 degrees of the historical period, the future period RCP4.5 and the RCP8.5 obtained by simulating the GFDL climate mode data is resampled to be 0.083 degrees multiplied by 0.083 degrees by a nearest neighbor method.
2. Spatial fusion scaling factor calculation
The two sets of NDVI data have a relation between each pixel:
Figure BDA0003345859440000171
wherein: n is a radical ofARepresents VIP-NDVI, NGRepresenting the NDVI obtained by simulating GFDL-ESM2M mode data, gamma is a downscaling constant, 2 is taken,alpha is a scaling coefficient fused with each half-moon space, namely a scaling coefficient.
The calculation of the coefficients is carried out in MATLAB: and (3) simulating the semimonthly perennial average NDVI obtained by using the semimonthly perennial average of the VIP-NDVI and the GFDL-ESM2M mode data, and substituting the two sets of data into a formula (2) to obtain the spatial fusion proportionality coefficient alpha of each half month.
3. Downscaling computation
Transforming the formula (1) to obtain
NG0.083＝α×(NG0.5+γ)-γ (3)
Wherein: n is a radical ofG0.083Is a downscaled NDVI, N with a resolution of 0.083 DEG x 0.083 DEGG0.5The resolution of the GFDL-ESM2M mode data simulation is 0.5 degrees multiplied by 0.5 degrees, NDVI which is resampled to be 0.083 degrees multiplied by 0.083 degrees by a nearest neighbor method is adopted as gamma, and alpha is a scaling coefficient of space fusion of each half month, namely a scaling coefficient.
The NDVI of each half month of the emission scenes RCP4.5 and RCP8.5 in the historical period, namely 1982-2005 and the future period, namely 2006-2099 are processed by the formula (3) to obtain the result with the resolution of 0.083 degrees multiplied by 0.083 degrees. Pre-and post-downscaling comparison of RCP4.52020, such that the GFDL-ESM2M mode data in the pre-downscaling RCP4.5 emission scenario simulates 2020 year-averaged NDVI distribution with a resolution of 0.5 ° x 0.5 °, as shown in fig. 20, such that the GFDL-ESM2M mode data in the post-downscaling RCP4.5 emission scenario simulates 2020 year-averaged NDVI distribution with a resolution of 0.083 ° x 0.083 °, as shown in table 14 for pre-and post-downscaling data comparison:
TABLE 14
Figure BDA0003345859440000172
Figure BDA0003345859440000181
Pre-and post-downscaling comparison of RCP4.52050, with the GFDL-ESM2M mode data simulating a mean-per-year NDVI distribution of 2050 with a resolution of 0.5 ° x 0.5 ° for the pre-downscaling RCP4.5 emission scenario in fig. 22, and with the GFDL-ESM2M mode data simulating a mean-per-year NDVI distribution of 2050 with a resolution of 0.083 ° x 0.083 ° for the post-downscaling RCP4.5 emission scenario in fig. 23, and with the pre-and post-downscaling data comparison in table 15:
watch 15
Figure BDA0003345859440000182
Fifth, relevant analysis of NDVI tendency rate in future 2020-2099 year under the RCP4.5 situation;
the spatial distribution map of the global vegetation coverage data at a certain period in the future is obtained through calculation, for example, fig. 24 shows that the model data of the GFDL-ESM2M under the RCP4.5 emission situation after the downscaling simulates the annual average NDVI distribution of 2020, the resolution is 0.083 degrees × 0.083 degrees, for example, fig. 25 shows that the model data of the GFDL-ESM2M under the RCP4.5 emission situation after the downscaling simulates the annual average NDVI distribution of 2050, the resolution is 0.083 degrees × 0.083 degrees, and fig. 26 shows the annual curve change of the annual average NDVI within a certain specified range at a certain period in the future, taking the beijing area as an example.
Compared with the prior art, the method for predicting the vegetation coverage space-time change under the drive of the climate change is suitable for simulating the vegetation coverage in a small-scale area, the method is mainly used for establishing the relation between meteorological data and vegetation coverage data by a single model in the large-scale area, and is mainly used for simulating the vegetation coverage data in a historical period or a current period, and the method is used for establishing the model relation between the meteorological data and the vegetation coverage data one by one aiming at a single pixel, fully considering the space diversity in the natural ecological environment and suitable for a large-scale space range with a large area so as to obtain a regression model which is more in line with the natural law, thereby simulating and predicting the response of the vegetation coverage in the future period to the climate change. However, because the spatial resolution of the input data is low, the requirement of some fine researches cannot be met, and the invention finds out the size reduction method for effectively improving the spatial resolution on the basis of low spatial resolution, for example, in the embodiment, the original spatial resolution is 0.5 degrees multiplied by 0.5 degrees, the spatial resolution is improved to 0.083 degrees multiplied by 0.083 degrees after size reduction, so that geographic objects such as rivers, lakes and the like are clearer in the distribution edge on the space, the prediction result is finer, and the prediction result with large size and high resolution is achieved.
The above-mentioned embodiments are merely illustrative of the preferred embodiments of the present invention, and do not limit the scope of the present invention, and various modifications and improvements made to the technical solution of the present invention by those skilled in the art without departing from the spirit of the present invention shall fall within the protection scope defined by the claims of the present invention.

Claims (8)

1. A method for predicting vegetation coverage space-time change under the drive of climate change is characterized by comprising the following steps:
step 1: determining a research area and a data source for establishing a vegetation prediction model, specifically comprising the following steps:
step 11: determining a region to be predicted, wherein the region to be predicted comprises a spatial large-scale region with spatial diversity;
step 12: collecting a data source of the area to be predicted, wherein the data source comprises historical measured vegetation coverage grid data of 24 and a half months in a historical period year by year and meteorological grid data with the same time interval and the same space-time resolution, and the meteorological grid data comprise precipitation data and gas temperature data;
step 2: taking the historical actual measurement vegetation cover grid data of 24 and a half months year after year in the step 1 as a dependent variable, introducing hysteresis as an independent variable, constructing a pixel-by-pixel multiple linear regression model, and setting an optimal judgment condition to obtain an optimal multiple linear regression model;
and step 3: selecting meteorological data simulated in a future time period under a global climate mode as input according to the optimal multiple linear regression model obtained in the step 2, and calculating a preliminary result N of vegetation cover grid data of 24 and a half months year by year in the future time periodG；
And 4, step 4: according to the preliminary result of the vegetation coverage grid data of 24 and a half months year by year in the future period obtained in the step 3 and the historical observation vegetation coverage data N of the area to be predictedAPerforming downscaling treatment to obtain theThe highest resolution of the area to be predicted can reach 250m multiplied by 250m, and the final result N of the vegetation coverage grid data of 24 and a half months in the future year by yearGnew：
NGnew＝α×(NG+γ)-γ
Figure FDA0003345859430000011
In the formula: gamma is a downscaling parameter; alpha is a downscaling proportionality coefficient;
and 5: according to the final result N of the vegetation coverage grid data of 24 and a half months in the kilometer resolution level of the area to be predicted obtained in the step 4 year by yearGnewDrawing a vegetation coverage space distribution map of a future time period; according to the final result N of the vegetation coverage grid data of 24 and a half months in the kilometer resolution level of the area to be predicted obtained in the step 4 year by yearGnewAnd drawing an annual variation curve covered by planting at a specified time, wherein the specified time comprises the annual average, different seasons and different months in a specified range.
2. The method of claim 1, wherein step 2 comprises the steps of:
step 21: the lag comprises meteorological grid data with the same space-time resolution ratio of the current half month, one month, three months, half year, one year and two years, and vegetation coverage data of each pixel is used as a dependent variable to obtain a multiple linear regression model of each pixel;
step 22: performing precision evaluation on the multiple linear regression model established in the step 21, specifically including calculating a decision coefficient of the model, calculating a correlation coefficient by adopting a five-year average analysis method aiming at the absolute change of the simulation result and the five-year absolute change of the measured data, calculating a tendency rate of the simulation result and the measured result by adopting a tendency rate analysis method, and performing correlation coefficient of the simulation result and the measured result;
step 23: integrating the decision coefficient of the model calculated in the step 22, the five-year average analysis method and the result of the trend rate analysis, judging whether the multiple linear regression model established in the step 21 meets the optimal judgment condition, and obtaining the optimal multiple linear regression model when the multiple linear regression model established in the step 21 meets the optimal judgment condition; otherwise, return to step 21.
3. The method for predicting spatiotemporal changes in vegetation coverage under the drive of climate change according to claim 2, wherein the decision coefficient of the calculation model in the step 22 is specifically:
Figure FDA0003345859430000021
in the formula:
Figure FDA0003345859430000022
determining coefficients corresponding to pixel grids in the ith row and the jth column;
Figure FDA0003345859430000023
the predicted value of a multivariate linear regression model of the vegetation cover pixel value on a time sequence is obtained;
Figure FDA0003345859430000024
the mean value of the vegetation covering pixel value on the time sequence is obtained; y is(i，j)And covering the pixel value for the vegetation of the ith row and the jth column in the grid data.
4. The method for predicting spatiotemporal changes in vegetation coverage under the drive of climate change according to claim 2, wherein the calculating of the correlation coefficient for the absolute change of the simulation result and the absolute change of the measured data by the five-year average analysis method in step 22 is specifically as follows: selecting a time interval corresponding to the data source in the step 1, simulating by using the multiple linear regression model of each pixel to obtain a dependent variable, actually measuring the dependent variables of the early, middle and late time intervals of every five years in the time interval corresponding to the data source to be A, B and C respectively, calculating the five-year average value of the dependent variable and the actually measured dependent variable obtained by simulation, calculating the absolute change of the early, middle and late time intervals to be B-A, C-B and C-A to obtain the absolute change of a simulation result and the five-year absolute change of actually measured data, and calculating the correlation coefficient of the absolute change of the simulation result and the five-year absolute change of the actually measured data.
5. The method for predicting vegetation coverage space-time variation under the drive of climate variation according to claim 2, wherein the step 22 of calculating the trend rates of the simulation result and the actual measurement result by using a trend rate analysis method and calculating the correlation coefficient between the simulation result and the actual measurement result specifically comprises: and (3) taking the time sequence as an independent variable and the vegetation coverage data of each pixel as a dependent variable to obtain a unitary linear equation of each pixel with respect to the time sequence:
Y(i，j)＝b(i，j)+a(i，j)t(i，j)
in the formula: y is(i，j)Covering pixel values for the vegetation of the ith row and the jth column of pixels in the grid data; t is t(i，j)The time sequence is the ith row and the jth column of the pixel; a is(i，j)Is one tenth of the tendency rate of the ith row and the jth column of the image elements;
calculating the tendency rate of the simulation result and the actual measurement result, and calculating the correlation coefficient of the simulation result and the actual measurement result.
6. The method for predicting vegetation coverage spatiotemporal variation under the drive of climate variation according to claim 5, wherein the multiple linear regression model for each pixel obtained in the step 21 is specifically:
Y(i，j)＝β0(i，j)+β1(i，j)X1(i，j)+β2(i，j)X2(i，j)+……+βk(i，j)Xk(i，j)
in the formula: beta is a0(i，j)Constant items corresponding to the ith row and the jth column of pixel grids in grid data; beta is ak(i，j)Model coefficients corresponding to the k-th meteorological data independent variables of the ith row and jth column of pixels in the grid data; xk(i，j)And the independent variable of the kth meteorological data of the ith row and jth column pixel in the grid data.
7. The method for predicting vegetation coverage space-time variation under the drive of climate variation according to claim 1, wherein the optimal judgment conditions in the step 2 are as follows:
PDI＞0.5＞70％
CIyear-avg＞0.9
CIavg＞0.5
CIrate＞0.6
in the formula: pDI＞0.5The pixel number proportion of a pixel number of a pixel-by-pixel multiple linear regression model of the region to be predicted, wherein the coefficient of determination DI is greater than 0.5; CIyear-avgThe correlation coefficient of the vegetation coverage results of two five-year average years in the same period in the five-year average analysis method; CIavgThe average value of the correlation coefficient of the five-year average space-time change is obtained; CIrateThe correlation coefficient is calculated by adopting a tendency rate analysis method.
8. The method of claim 1, wherein the area of the area to be predicted in step 11 is 50 km/km2The above.
CN202111322031.5A 2021-11-09 2021-11-09 Prediction method for vegetation coverage space-time change under climate change drive Active CN114254802B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202111322031.5A CN114254802B (en) 2021-11-09 2021-11-09 Prediction method for vegetation coverage space-time change under climate change drive

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202111322031.5A CN114254802B (en) 2021-11-09 2021-11-09 Prediction method for vegetation coverage space-time change under climate change drive

Publications (2)

Publication Number Publication Date
CN114254802A true CN114254802A (en) 2022-03-29
CN114254802B CN114254802B (en) 2023-09-08

Family

ID=80790625

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202111322031.5A Active CN114254802B (en) 2021-11-09 2021-11-09 Prediction method for vegetation coverage space-time change under climate change drive

Country Status (1)

Country Link
CN (1) CN114254802B (en)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN116628979A (en) * 2023-05-23 2023-08-22 中国科学院重庆绿色智能技术研究院 Multiple regression total primary productivity pixel-by-pixel prediction method, device and medium
CN116975784A (en) * 2023-09-19 2023-10-31 四川省水利科学研究院 High-space-time resolution MPDI data set construction method, system and storage medium

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105719236A (en) * 2016-01-14 2016-06-29 北京师范大学 Method for generating complete high-resolution vegetation coverage image reason change sequence
KR101809629B1 (en) * 2016-11-28 2017-12-15 한국외국어대학교 연구산학협력단 Method and system for forecasting detailed information of urban heat using the um ldaps
CN107632967A (en) * 2017-09-14 2018-01-26 青海省基础地理信息中心 A kind of meadow grass yield evaluation method
CN111241912A (en) * 2019-12-18 2020-06-05 安徽易刚信息技术有限公司 Multi-vegetation index rice yield estimation method based on machine learning algorithm
CN111898809A (en) * 2020-07-15 2020-11-06 闫慧强 Flood season climate trend prediction system based on TCN

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105719236A (en) * 2016-01-14 2016-06-29 北京师范大学 Method for generating complete high-resolution vegetation coverage image reason change sequence
KR101809629B1 (en) * 2016-11-28 2017-12-15 한국외국어대학교 연구산학협력단 Method and system for forecasting detailed information of urban heat using the um ldaps
CN107632967A (en) * 2017-09-14 2018-01-26 青海省基础地理信息中心 A kind of meadow grass yield evaluation method
CN111241912A (en) * 2019-12-18 2020-06-05 安徽易刚信息技术有限公司 Multi-vegetation index rice yield estimation method based on machine learning algorithm
CN111898809A (en) * 2020-07-15 2020-11-06 闫慧强 Flood season climate trend prediction system based on TCN

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
钱娅: "深圳市植被覆盖度动态变化及驱动力分析", 《中国优秀硕士学位论文全文数据库 中南林业科技大学大学》 *
马天啸;宋现锋;赵昕;李润奎;: "2000―2010年黄河源区植被覆盖率时空变化及其影响因素", 干旱区研究, no. 06 *

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN116628979A (en) * 2023-05-23 2023-08-22 中国科学院重庆绿色智能技术研究院 Multiple regression total primary productivity pixel-by-pixel prediction method, device and medium
CN116628979B (en) * 2023-05-23 2024-04-30 中国科学院重庆绿色智能技术研究院 Multiple regression total primary productivity pixel-by-pixel prediction method, device and medium
CN116975784A (en) * 2023-09-19 2023-10-31 四川省水利科学研究院 High-space-time resolution MPDI data set construction method, system and storage medium
CN116975784B (en) * 2023-09-19 2023-12-29 四川省水利科学研究院 High-space-time resolution MPDI data set construction method, system and storage medium

Also Published As

Publication number Publication date
CN114254802B (en) 2023-09-08

Similar Documents

Publication Publication Date Title
van Bussel et al. From field to atlas: Upscaling of location-specific yield gap estimates
Rauthe et al. A Central European precipitation climatology–Part I: Generation and validation of a high-resolution gridded daily data set (HYRAS)
Crochemore et al. Bias correcting precipitation forecasts to improve the skill of seasonal streamflow forecasts
de Hipt et al. Modeling the effect of land use and climate change on water resources and soil erosion in a tropical West African catch-ment (Dano, Burkina Faso) using SHETRAN
Boer et al. Kriging and thin plate splines for mapping climate variables
Tveito et al. A GIS-based agro-ecological decision system based on gridded climatology
Miyazaki et al. Global lightning NO x production estimated by an assimilation of multiple satellite data sets
Singh et al. Inter-comparisons and applicability of CMIP5 GCMs, RCMs and statistically downscaled NEX-GDDP based precipitation in India
Huang et al. Assessing model performance of daily solar irradiance forecasts over Australia
Solaiman et al. Development of probability based intensity-duration-frequency curves under climate change
Montzka et al. Modelling the water balance of a mesoscale catchment basin using remotely sensed land cover data
Arowolo et al. Comparison of spatial interpolation techniques to generate high‐resolution climate surfaces for Nigeria
CN114254802A (en) Prediction method of vegetation coverage space-time change under climate change drive
CN111260111A (en) Runoff forecast improvement method based on meteorological big data
Kiesel et al. Improving hydrological model optimization for riverine species
ZHANG et al. Developing a process-based and remote sensing driven crop yield model for maize (PRYM–Maize) and its validation over the Northeast China Plain
Liu et al. Unrevealing past and future vegetation restoration on the Loess Plateau and its impact on terrestrial water storage
Li et al. Terrestrial transect study on pattern and driving mechanism of ecosystem services in the China–Mongolia–Russia Economic Corridor
Cheng et al. Improving soil available nutrient estimation by integrating modified WOFOST model and time-series earth observations
Faiz et al. Effects of land use and climate variability on the main stream of the Songhua River Basin, Northeast China
Zhou et al. Daily rainfall model to merge TRMM and ground based observations for rainfall estimations
Giles et al. Optimization of linear stream temperature model parameters in the soil and water assessment tool for the continental United States
CN114677020A (en) Quantitative evaluation method for regional scale soil wind erosion change driving mechanism
Londhe et al. Comparative assessment of evapotranspiration in Bhima sub-basin using spatial analysis for normal and ENSO years
Clayer et al. Sources of skill in lake temperature, discharge and ice-off seasonal forecasting tools

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant