CN108152212A - Meadow ground biomass inversion method based on high time resolution and spatial resolution Multi-spectral Remote Sensing Data - Google Patents

Abstract

The above-ground biomass inversion method of grassland based on multi-spectral remote sensing data with high temporal resolution and spatial resolution of the present invention, the specific steps are: 1) select multi-spectral remote sensing data covering the growing season of grassland, calculate NDVI, and synthesize and cover the research area 2) According to the monthly average temperature, monthly total precipitation, and monthly total solar radiation data of meteorological stations, interpolation generates monthly average temperature, monthly total precipitation, and monthly total solar radiation data covering the study area; 3) Based on the Light Use Efficiency (LUE) theory, the CASA (Carnegie‑Ames‑Stanford Approach) model was used to calculate the net primary productivity (Net Primary Productivity, NPP) of grassland; 4) According to NPP and aboveground biomass and underground Ratio of biomass to calculate aboveground biomass of grassland.

Description

Grassland based on high temporal and spatial resolution multispectral remote sensing data Biomass inversion method

technical field

The invention relates to the field of ecological index calculation, in particular to a method for calculating the aboveground biomass index of grassland based on remote sensing inversion technology.

Background technique

Biomass of plant communities is a comprehensive quantitative feature that reflects the structure and function of ecosystems. Changes in community biomass will affect other processes in ecosystems, such as terrestrial carbon cycle and global climate change. China's natural grassland accounts for more than 40% of the country's land area and is an important part of the terrestrial ecosystem. Therefore, a comprehensive understanding of grassland biomass has important scientific and practical value for long-term monitoring of carbon cycle in grassland ecosystems, assessment of grassland degradation, and rational use and management of natural grasslands.

The calculation methods of grassland biomass can be roughly divided into two categories: one is the field measurement method, and the other is the model simulation method. In the field measurement method, it is generally obtained by harvesting the quadrats when the existing amount of grassland reaches the maximum. This method is time-consuming, labor-intensive, and affected by traffic accessibility. The model simulation method can be subdivided into empirical-semi-empirical statistical regression method, mechanism model method and so on. Empirical-semi-empirical statistical regression methods usually use sample point data obtained from field measurements and some vegetation indicators, such as NDVI (Normalized Difference Vegetation Index, normalized difference vegetation index), to establish linear and nonlinear regression fitting formulas, and then obtain the entire area simulation results. This method lacks theoretical support, and the quality of the results is highly related to the quality of the data source used and the location of the region, so it is difficult to be extended to other regions. The mechanism model method simulates the physiological and ecological process of plants and its influencing factors, feedback mechanism, etc., mostly considering the soil-plant-atmosphere continuum as a whole, including photosynthesis, respiration, evapotranspiration, stomata, etc. Conductance and other sub-modules. The mechanism model has a complete theoretical system, but it involves many inputs, such as soil temperature, humidity, etc., so it can only be calculated near the observatory with relevant equipment, and it is difficult to extend to the entire area in the actual application process.

Contents of the invention

Aiming at the problems existing in the prior art, the present invention provides a remote sensing data utilizing high temporal resolution and spatial resolution (such as the multispectral data of Gaofen 1 satellite, its spatial resolution is 16 meters, and its temporal resolution is 2 -4 days), combined with meteorological observation data, and based on the mechanism model of grassland light energy utilization rate, the inversion calculation of large-scale grassland aboveground biomass is realized.

In order to achieve the above object, the above-ground biomass inversion method of grassland based on high time resolution and spatial resolution multi-spectral remote sensing data of the present invention, the specific steps are: 1) select multiple grassland growth season (such as May-September) Spectral remote sensing data, calculate NDVI, and synthesize monthly maximum NDVI covering the study area; 2) Interpolate monthly average temperature, monthly Total precipitation and total monthly solar radiation data; 3) Based on the Light Use Efficiency (LUE) theory, use the CASA (Carnegie-Ames-Stanford Approach) model to calculate the net primary productivity (Net Primary Productivity, NPP) of the grassland ; 4) According to NPP and the ratio of aboveground biomass to belowground biomass, the aboveground biomass of the grassland is calculated.

Further, the multispectral remote sensing data in step 1) needs to be corrected for atmospheric radiation, and the 6S or FLASSH module of ENVI is used for correction.

Further, after calculating the NDVI of each scene of remote sensing data, the maximum value of the NDVI of the same month is synthesized, and the calculation is performed with the Mosaic module of ArcGIS.

Further, the interpolation method in step 2) adopts the thin disk spline interpolation method, which is realized with the ANUSPLIN tool.

Furthermore, the interpolation result is in a raster format, and its spatial resolution is consistent with that of the multispectral data.

Further, the specific steps of grassland NPP calculation in step 3) are: according to NDVI and grassland type, calculate the absorption ratio (FPAR) of grassland to photosynthetically active radiation; calculate the optimum temperature for grassland growth according to the monthly average temperature when the NDVI value reaches the highest value in a year. Temperature index: Calculate the coefficient of reduction of light energy utilization rate of grassland under the condition of deviating from the optimum temperature according to the average temperature of the month when the NDVI value reaches the highest value in a year and the average temperature of each month; calculate the growth of grassland according to the precipitation of each month Influencing factors of water stress; the monthly solar radiation data, FPAR, the most suitable temperature index for grassland growth, the coefficient of reduction of light energy use efficiency of grassland under the condition of deviating from the optimum temperature, and the water stress influencing factors of grassland growth The monthly NPP result of the grassland is obtained through the effective radiation conversion rate coefficient of the grassland; finally, the NPP results of each month in the growing season are added to obtain the annual NPP result of the grassland.

Further, the grassland type data involved in the calculation in step 3) is in raster format, which can be obtained from vector data such as grassland type maps through vector-raster conversion, and its spatial resolution is consistent with that of multispectral data.

Further, the grassland effective radiation conversion rate coefficient in step 3) corresponds to the grassland type, and each grassland type has a corresponding grassland effective radiation conversion rate coefficient value, which can be obtained through literature review or actual measurement.

Further, the specific steps for calculating the aboveground biomass of grassland in step 4) are as follows: calculate the annual NPP results of the grassland and the root-to-shoot ratio, root turnover rate, carbon content of the underground part, and carbon content of the aboveground part of the grassland type according to the formula to obtain the result.

Furthermore, in step 4), the coefficients of root-to-shoot ratio, root turnover rate, underground carbon content, and aboveground carbon content of different grassland types can be obtained through literature review or actual measurement.

The invention proposes an inversion model algorithm based on multi-spectral remote sensing images for the biomass on the grassland. The algorithm of the present invention combines the characteristics of large-scale monitoring data acquired by remote sensing images and the theoretical advantages of mechanism models, can more accurately simulate the aboveground biomass of grasslands, and can be used for ecological status monitoring and grazing management of natural grasslands.

Description of drawings

Fig. 1 is the flow chart of the aboveground biomass inversion method of grassland based on high temporal resolution and spatial resolution multispectral remote sensing data of the present invention

Detailed ways

The calculation process of the aboveground biomass inversion method of grassland based on multispectral remote sensing data with high temporal resolution and spatial resolution in the present invention is shown in FIG. 1 . The specific introduction of the algorithm principle of this method is as follows:

1. Calculation of NPP principle based on light energy utilization rate

NPP represents the total amount of organic matter that green plants can fix per unit time and per unit area. The light use efficiency model calculates the NPP of the grassland through the absorbed photosynthetically active radiation (APAR) and the effective radiation conversion rate (ε) absorbed by the grassland:

NPP=APAR(x,t)×ε(x,t)

In the formula: x is the spatial position, t is the time

APAR(x, t) represents the photosynthetically active radiation absorbed by the pixel in month t, and the unit is MJ.m ^-2 .month ^-1 . ε(x,, t) represents the actual effective radiation conversion rate of the pixel in month t, and the unit is gC.MJ ^-1 . The photosynthetically active radiation (APAR) absorbed by the grassland is determined by the photosynthetically active radiation (PAR) in the total solar radiation and the proportion of photosynthetically active radiation absorbed by the grassland (FPAR). FPAR uses the normalized difference vegetation index (NDVI) and Grassland type representation, ε is the efficiency of grassland to convert absorbed photosynthetically active radiation (FPAR) into organic carbon, which is mainly affected by temperature and moisture.

APAR(x,t)=SOL(x,t)×FPAR(x,t)×0.5

SOL(x, t) represents the total amount of solar radiation at the pixel (position) x in month t, and the unit is MJ.m ^-2 ; FPAR is the absorption ratio of the grass layer to the incident photosynthetically active radiation (no unit); the constant is 0.5 Indicates the proportion of solar effective radiation (wavelength 0.38-0.71 μm) that can be used by grassland to total solar radiation.

Grassland has the maximum light utilization efficiency under ideal conditions, while the maximum light utilization efficiency under realistic conditions is mainly affected by temperature and moisture:

ε(x, t)=T _ε1 (x, t)×T _ε2 (x, t)×W _ε (x, t)×ε _max

The first two factors represent the stress effects of low temperature and high temperature on light energy utilization efficiency. The third factor represents the influence coefficient of water stress, which reflects the influence of water conditions. The fourth item is the maximum utilization rate of light energy under ideal conditions, that is, the effective radiation conversion rate (gC.MJ ^-1 ).

Based on the above formula, the calculation of NPP becomes:

NPP＝SOL(x,t)×FPAR(x,t)×0.5×T _ε1 (x,t)×T _ε2 (x,t)×W _ε (x,t)×ε _max

Among them: (1) SOL(x, t) is the total solar radiation, which is obtained by interpolating the observed data of meteorological stations:

(2) Calculation method of FPAR(x, t)

There is a linear relationship between FPAR(x, t) and NDVI, which can be determined according to the maximum and minimum values of NDVI of a grassland type and the corresponding maximum and minimum values of FPAR:

The values of FPARmax and FPARmix in the formula have nothing to do with the type of grassland, they are 0.95 and 0.001, respectively. NDVIi, max, NDVIi, min are the maximum and minimum values of NDVI for the i-th grassland type, respectively.

Further research shows that there is also a good linear relationship between FPAR(x, t) and the ratio grassland index SR:

SRi, max, SRi, min are NDVIi, max, NDVIi, min of the corresponding grassland type i, respectively:

The FPAR estimated by NDVI is higher than the measured value, while the FPAR estimated by SR is lower than the measured value, but the error is smaller than the result directly estimated by NDVI, so the mean of the two is taken as the estimated FPAR value:

FPAR(x, t) = α×FPAR _NDVI + (1-α)×FPAR _sR

α is the adjustment coefficient.

(3) Calculation method of T _ε1 (x, t)

T _ε1 (x, t) reflects the reduction of net primary productivity due to the limitation of photosynthesis by the internal biochemical action of plants at low temperature and high temperature, which can be expressed by the following formula:

T _ε1 (x, t)＝0.8+0.02T _opt (x)-0.0005[T _opt (x)] ²

T _opt (x) is the monthly average temperature when the NDVI value reaches the highest value in a certain area in a year. The size and change of NDVI can reflect the growth status of plants. When the NDVI reaches the highest value, the plants grow the fastest. The temperature at this time can be To a certain extent, it represents the optimum temperature for plant growth.

(4) Calculation method of T _ε2 (x, t)

T _ε2 (x, t) represents the trend that the light utilization rate of plants gradually decreases when the ambient temperature changes from the optimum temperature (T _opt ) to high and low temperatures, because at low and high temperatures, high respiration consumption will inevitably reduce light utilization. Utilization efficiency, when grown under conditions deviating from the optimum temperature, its light utilization efficiency will definitely decrease:

T(x, t) represents the average temperature at pixel x in month t. When the average temperature of a month is 10° higher or 13° lower than the optimum temperature (T _opt ), T _ε2 (x, t) is equal to the monthly average temperature (T(x, t)) as the optimum temperature (T _opt ) is half of T _ε2 (x, t).

(5) Calculation method of W _ε (x, t)

Water stress influence coefficient W _ε (x, t), reflects the effect of available water conditions that plants can use on light use efficiency. As the available moisture in the environment increases, W _ε (x, t) gradually increases, and its value ranges from 0.5 (under extreme drought conditions) to 1 (under very humid conditions), and the formula is:

EET(x, t) is the regional actual evapotranspiration (mm/month), PET(x, t) is the regional potential evapotranspiration (mm/month). EET(x, t) is calculated according to the actual evapotranspiration model of the area:

In the formula, r is the precipitation amount (mm), and R _n is the net radiation amount (mm). Since the general meteorological observation stations do not observe surface net radiation, the calculation method of R _n is as follows:

R _n = (E _p ×r) ^1/2 [0.369+0.598 (E _p /r) ^1/2 ]

E _p is the local potential evapotranspiration, that is, the evapotranspiration of a small patch of fully wetted ground under local climate conditions. The relationship between E _p and the monthly average temperature T (°C) is as follows, the unit is mm/month:

E _p ＝16×(10T/I) ^α

in:

α=(0.675×I ³ -77.1×I ² +17920×I+492390)×10 ^-6

I is the heat index of the 12-month sum, and a is a constant that varies from place to place and is a function of I. This relationship is only valid for air temperatures between 0° and 26.5°. Potential evapotranspiration is set as 0 when the air temperature is lower than 0°, and when it is higher than 26.5°, the potential evapotranspiration only increases with the increase of temperature and has nothing to do with I.

Complementary relationship among regional actual evapotranspiration, local potential evapotranspiration, and regional potential evapotranspiration:

EET(x,t)+E _p =2×PET(x,t)

So far, W _ε (x, t) can be obtained according to monthly rainfall and temperature.

2. The principle of calculating aboveground biomass based on the ratio of NPP to aboveground and belowground biomass

According to the ratio of the NPP result to the aboveground biomass, the aboveground biomass of the grassland, that is, the grass production, can be calculated in g/m ² ,

NPP＝ANPP+BNPP

ANPP is the aboveground biomass of grassland, and BNPP is the belowground biomass of grassland. Therefore, the amount of grassland grass can be estimated by the productivity ratio of the aboveground part and the underground part of various grasslands:

GY is the amount of grass produced by grass. The calculation method of BNPP is as follows:

BNPP＝BGB×(live BGB/BGB)×turover

Turover＝0.0009×ANPP+0.25

Among them, BGB is the biomass of the underground part of the grassland (root system), live BGB/BGB is the ratio of living root biomass to the total root biomass, and BGB×(liveBGB/BGB) is equal to the ratio of underground living biomass to aboveground biomass (R /S), turnover is the root turnover value of grassland plants.

It should be pointed out that any modification made according to the specific embodiments of the present invention shall not deviate from the spirit of the present invention and the scope described in the claims.

Claims (8)

1. A grassland aboveground biomass inversion method based on high temporal and spatial resolution multispectral remote sensing data, the specific steps are: 1) Select the multispectral remote sensing data covering the growing season of the grassland, and calculate the normalized normalized grassland index NDVI , and synthesize the monthly maximum NDVI covering the study area; 2) According to the monthly average temperature, monthly total precipitation, and monthly total solar radiation data of meteorological stations, interpolate to generate the monthly average temperature, monthly total precipitation, and monthly total solar radiation covering the study area The spatial resolution of the solar radiation data is consistent with that of the multispectral remote sensing data; 3) Based on the Light Use Efficiency (LUE) theory, the CASA (Carnegie-Ames-Stanford Approach) model is used to calculate the net The calculation process of the first productivity (Net Primary Productivity, NPP) is as follows: firstly, according to the normalized difference vegetation index (NDVI) and the ratio vegetation index (SR) of each grassland type, the absorption ratio of photosynthetically active radiation (FPAR) of the grassland is obtained, Secondly, calculate the most suitable temperature index T _ε1 (unitless) for grassland growth based on the temperature of the most vigorous month of grassland growth, and then calculate the temperature index of grassland under the condition of deviating from the optimum temperature according to the most suitable temperature index of grassland and the monthly average temperature The coefficient T _ε2 (unitless) for the reduction of light energy use efficiency, and then calculate the water stress influence coefficient W _ε (unitless) of grassland, and finally compare the above indicators with the monthly total solar radiation SOL and grassland effective radiation conversion rate ε _max Multiply the monthly NPP in the growing season, and add up the NPP of each month in the growing season to get the annual NPP of the grassland, as shown in formula (1): NPP＝SOL(x, t)×FPAR(x, t)×0.5×T _ε1 (x, t)×T _ε2 (x, t)×W _ε (x, t)×ε _max (1), SOL( x, t) represents the total solar radiation at the pixel (position) x in month t, and the unit is MJ.m ^-2 , FPAR is the absorption ratio of the grass layer to the incident photosynthetically active radiation (unitless), ε(x, t) represents the actual effective radiation conversion rate of the pixel in month t, and the unit is gC.MJ ^-1 . 4) According to NPP and the ratio of aboveground biomass to belowground biomass, the aboveground biomass of grassland is calculated. 2. the biomass inversion method on grassland based on high temporal resolution and spatial resolution multispectral remote sensing data as claimed in claim 1, is characterized in that, when calculating FPAR, as shown in formula (2): The values of FPARmax and FPARmix in the formula have nothing to do with the type of grassland, they are 0.95 and 0.001, respectively. NDVIi, max, NDVIi, min are the maximum and minimum values of NDVI of the i-th grassland type respectively, SRi, max, SRi, min are the NDVIi, max, NDVIi, min, SR of the i-th grassland type respectively (x,t)=(1+NDVI(x,t))/(1-NDVI(x,t)). 3. the aboveground biomass inversion method based on high time resolution and spatial resolution multispectral remote sensing data as claimed in claim 1, is characterized in that, calculates the most suitable temperature index for grassland growth, as shown in formula (3) : T _ε1 (x, t)=0.8+0.02T _opt (x)-0.0005[T _opt (x)] ² (3) T _opt (x) is the average temperature of the month when the NDVI value reaches the highest in a certain area within a year. 4. the biomass inversion method on the grassland based on high temporal resolution and spatial resolution multispectral remote sensing data as claimed in claim 1, is characterized in that, the temperature T _opt (x) when calculating the most suitable temperature index for grassland growth , is calculated based on the average monthly mean temperature when the grassland NDVI reaches the highest value for many years. 5. The method for retrieving biomass on grassland based on multi-spectral remote sensing data with high temporal resolution and spatial resolution as claimed in claim 2, characterized in that, the light energy utilization of the grassland under the condition of deviating from the optimum temperature is calculated The coefficient of rate reduction, as shown in formula (4): T(x, t) represents the average temperature at pixel x in month t. 6. the aboveground biomass inversion method of grassland based on high time resolution and spatial resolution multispectral remote sensing data as claimed in claim 2, is characterized in that, calculates the water stress influence coefficient of grassland, as shown in formula (5) : EET(x, t) is the regional actual evapotranspiration (mm/month), PET(x, t) is the regional potential evapotranspiration (mm/month). 7. the aboveground biomass inversion method based on high time resolution and spatial resolution multispectral remote sensing data as claimed in claim 2, is characterized in that, in step 4) according to NPP and aboveground biomass and underground biomass The aboveground biomass of the grassland is calculated. The calculation process is as follows: firstly, according to the carbon content of the underground part and the aboveground part of the grassland, the root turnover rate and the root-to-shoot ratio parameters, the ratio of the underground NPP to the aboveground NPP of the grassland (BNPP/ANPP) is obtained. , and then calculate the aboveground biomass of grassland by annual grassland NPP and grassland BNPP/ANPP, as shown in formula (6): GY is grass yield, ANPP is aboveground biomass of grassland, and BNPP is belowground biomass of grassland. 8. the aboveground biomass inversion method based on high temporal resolution and spatial resolution multispectral remote sensing data as claimed in claim 1, is characterized in that, BNPP calculation is as shown in formula (7): BNPP＝BGB×(live BGB/BGB)×turover (7) Among them, BGB is the biomass of the underground part of the grassland (root system), live BGB/BGB is the ratio of live root biomass to the total root biomass, and BGB×(live BGB/BGB) is equal to the ratio of underground live biomass to aboveground biomass ( R/S), turnover is root turnover value of grassland plants.