Method for estimating optimal temperature and optimizing vegetation photosynthesis model based on fluorescence efficiency

Abstract

The invention discloses a method for estimating an optimal temperature and optimizing a vegetation photosynthesis model based on fluorescence efficiency, and the method comprises the steps: obtaining a global TROPOMI satellite SIF product, and carrying out the simulation calculation through combining with the MODIS reflectivity, thereby obtaining the relative fluorescence efficiency phi F; constructing a relation curve between the phi F and the temperature Tair of the ERA5 in a 2-degree * 2-degree sliding window, and extracting a Tair value corresponding to the highest fluorescence efficiency to estimate Topt; replacing a biological group system Topt parameter in an original VPM model with the Topt inverted by the phi F, constructing an improved VPM model, and adjusting an LUE estimated value through an LUE temperature response function; and comparing original and improved model GPP estimation results at 203 flux tower stations in the world, and evaluating an improvement effect. The GPP remote sensing estimation precision is improved.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of remote sensing quantitative inversion and vegetation productivity simulation, and particularly relates to a method for estimating optimal temperature based on fluorescence efficiency and optimizing a vegetation photosynthesis model. BACKGROUND

[0002] One of the core processes of global carbon cycle is that terrestrial ecosystems absorb carbon dioxide in the atmosphere through photosynthesis, in which gross primary productivity (GPP) represents the total amount of CO2 fixed by vegetation through photosynthesis per unit time. Most of the current mainstream GPP remote sensing estimation methods are based on light use efficiency model (LUE model), in which the temperature response function is a key constraint term. The model represented by Vegetation Photosynthesis Model (VPM) widely adopts the optimal temperature (Topt) parameter fixed by biomes. However, actual observations show that the temperature response of vegetation in different regions has significant spatial heterogeneity after long-term adaptation to climate, and it is difficult for the fixed Topt parameter to adapt to the actual changes, resulting in systematic errors in GPP estimation. In recent years, solar-induced chlorophyll fluorescence (SIF) provides a new observation means for characterizing plant photosynthesis. Among them, the ratio of SIF to absorbed photosynthetically active radiation, i.e. fluorescence efficiency (ΦF), is considered to be a more physiological LUE index, which can reflect the energy distribution state of photosystem II. Therefore, ΦF is proposed to be used to invert the optimal point Topt of the temperature response curve, but there is still a lack of systematic method to integrate it into the ecological modeling process. SUMMARY

[0003] To solve the above technical problems, the application provides a method for estimating optimal temperature based on fluorescence efficiency and optimizing a vegetation photosynthesis model, which improves the accuracy of GPP remote sensing estimation.

[0004] To achieve the above purpose, the application provides a method for estimating optimal temperature based on fluorescence efficiency and optimizing a vegetation photosynthesis model, which comprises:

[0005] Global TROPOMI satellite SIF products are obtained, and relative fluorescence efficiency ΦF is simulated and calculated in combination with MODIS reflectance;

[0006] A relationship curve between ΦF and ERA5 air temperature Tair is constructed in a 2°x2° sliding window, and the Tair value corresponding to the highest fluorescence efficiency is extracted to estimate Topt;

[0007] The Topt inverted by ΦF replaces the biomes Topt parameter in the original VPM model, an improved VPM model is constructed, and the LUE estimated value is adjusted through the LUE temperature response function;

[0008] The improved effect is evaluated by comparing the GPP estimation results of the original and improved models at 203 flux tower sites around the world.

[0009] Technical effects of the present application:

[0010] (1) Improve the accuracy of remote sensing estimation of GPP: especially in the area where Topt deviates significantly (such as high latitude), the improved model R 2 increases by 0.05 and the RMSE decreases by 0.2 g / m 2 / d.

[0011] (2) Universal and adaptive: Topt inverted by ΦF has good consistency with the LUE inversion result (R 2 = 0.66), and is suitable for various vegetation types;

[0012] (3) Simple and efficient: it does not depend on the classification of the ecosystem, and only requires remote sensing ΦF and air temperature data to realize the spatialization of optimal temperature estimation;

[0013] (4) Strong scalability: after the high-resolution SIF data is provided by the future FLEX mission, the accuracy and adaptability of the method will be further improved. BRIEF DESCRIPTION OF DRAWINGS

[0014] The drawings constituting a part of this application are used to provide further understanding of the application, the illustrative embodiments of the application and the description thereof are used to explain the application, and do not constitute improper limitation on the application. In the drawings:

[0015] Figure 1 It is a process diagram for estimating the optimum temperature (Topt) by using the fluorescence efficiency (ΦF) and air temperature (Tair) for the embodiments of the present application;

[0016] Figure 2 It is a comparative analysis diagram of Topt under the LUE estimation of the embodiments of the present application and (A) Topt estimated based on SIF and (B) Topt estimated based on ΦF;

[0017] Figure 3 It is a global optimum temperature spatial distribution diagram based on relative fluorescence efficiency inversion for the embodiments of the present application;

[0018] Figure 4 It is a spatial distribution diagram of different types of vegetation for the embodiments of the present application, wherein (A) is the latitudinal distribution pattern of Topt of each vegetation type, and (B) is a box plot of each vegetation type;

[0019] Figure 5 It is a comparative diagram of flux tower GPP and GPP estimated by (A) the original VPM model and (B) the adjusted VPM model for the embodiments of the present application;

[0020] Figure 6 A flowchart of a method for estimating an optimal temperature based on fluorescence efficiency and optimizing a vegetation photosynthesis model according to an embodiment of the present application is shown in FIG. 1. DETAILED DESCRIPTION

[0021] It should be noted that the embodiments in the present application and the features in the embodiments can be combined with each other without conflict. The present application will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0022] It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions, and although a logical order is shown in the flowchart, in some cases, the steps shown or described herein can be executed in an order different from that shown herein.

[0023] As shown in FIG. 1, the method for estimating an optimal temperature based on fluorescence efficiency and optimizing a vegetation photosynthesis model according to an embodiment of the present application includes the following steps. Figure 6 Global TROPOMI satellite SIF products are obtained, and relative fluorescence efficiency ΦF is simulated and calculated in combination with MODIS reflectivity;

[0024] A relationship curve between ΦF and ERA5 air temperature Tair is constructed in a 2°x2° sliding window, and Tair value corresponding to the highest fluorescence efficiency is extracted to estimate Topt;

[0025] Topt inverted by ΦF is used to replace the biome Topt parameter in the original VPM model, an improved VPM model is constructed, and LUE estimated value is adjusted through a LUE temperature response function;

[0026] GPP estimated results of the original and improved models are compared at 203 global flux tower sites to evaluate the improvement effect.

[0027] The results show that Topt calculated by ΦF has a higher correlation with LUE estimated value (R 2 = 0.66), which is significantly better than the estimation method using SIF directly. After applying ΦF-Topt to the VPM model, the GPP simulation accuracy can be improved by more than 5% in the area with significant Topt deviation. The method has clear principles, strong applicability, and does not require biome classification information, and is suitable for Topt spatial modeling and carbon flux remote sensing inversion under various vegetation types and climate backgrounds, and improves the spatial authenticity and regional adaptability of carbon cycle simulation.

[0028]

[0029] ​Fluorescence efficiency extraction: The fluorescence efficiency ΦF was estimated from the SIF data acquired by TROPOMI satellite, after excluding the angle effect and low-quality data, and by normalizing the absorbed photosynthetically active radiation (APAR) as a proxy variable of temperature response. By aggregating ΦF and air temperature Tair in the same geographical area through a sliding window, the quantile method was used to filter out noise and extract the ΦF-Tair response curve.

[0030] Optimal temperature estimation based on ΦF: The two Tair points with the highest ΦF value in the ΦF-Tair curve were averaged to define the physiological optimal temperature Topt; Topt was calculated pixel by pixel in the global grid of 0.2°x0.2° to realize spatial distribution estimation; for different seasons and light periods, the daily average and instantaneous Topt were distinguished and converted at the VPM time resolution.

[0031] VPM model optimization: The original biome fixed Topt parameter in the VPM model was replaced by the ΦF inversion result; other temperature parameters (Tmin, Tmax) remained unchanged to ensure consistency of control variables; the temperature response function Tscalar was recalculated based on the optimized Topt, and the 8-day scale GPP was estimated.

[0032] Vorticity flux observation verification: The vorticity flux data of 203 sites such as AmeriFlux, OzFlux, Europe-Flux, etc. were collected, and high-quality half-hour level GPP records were selected; the simulated GPP was compared with the measured value to evaluate the precision of the VPM model before and after improvement; special attention was paid to the sites with ΔTopt exceeding 5℃, and the applicability of the method under extreme deviation conditions was verified.

[0033] One specific application example of the present application:

[0034] Fluorescence efficiency calculation: The ΦF value is obtained by inversion from TROPOMI SIF data, as follows:

[0035]

[0036] FCVI = NIR-VIS (2);

[0037] where SIFobsrepresents the observed TROPOMI SIF data. PAR (photosynthetically active radiation) is converted from shortwave radiation data from the fifth generation ECMWF reanalysis product (ERA5) with hourly temporal resolution and 0.25° spatial resolution. PAR at the TROPOMI satellite overpass time is obtained by linear interpolation. Unlike SIF, which is susceptible to PAR, the fluorescence efficiency ΦFis extremely small in sensitivity to PAR. FCVI (fluorescence-structure vegetation index) is defined as the difference between near-infrared reflectance (NIR) and visible reflectance (VIS) in the 400-700 nm band, which is calculated using the same sun-observation geometry as SIF. Since TROPOMI SIF product does not provide VIS reflectance, the RTLSR (Ross Thick-Li Sparse R) BRDF model is used to simulate reflectance at the sun-observation geometry of SIF. The BRDF parameters required to drive the RTLSR model are from the MCD19A3 BRDF / albedo product of MODIS with a spatial resolution of about 1 x 1 km 2 .

[0038] In addition, the MODIS reflectance simulated by RTLSR has been widely used in the SIF community in recent years to correct the effects of angle and canopy structure on SIF. VIS reflectance is calculated as a weighted sum of red (R), green (G), and blue (B) band reflectance according to the following equation (3):

[0039] VIS = 0.331R + 0.424B + 0.246G (3).

[0040] ΦF-Taircurves are constructed to calculate Topt:

[0041] ΦFand its corresponding Tairare used to estimate Toptat each 0.2° grid. Tairis derived from hourly data of ERA5 and the actual Tairat the TROPOMI satellite overpass time is obtained by linear interpolation. Figure 1 An example of the Toptestimation process is shown. To increase the number of samples, all ΦFand Tairpair data (grey dots in Figure 1 ) are combined within a 2° x 2° sliding window. Subsequently, these data are divided into twenty intervals according to the 0%, 5%, 10%, …, 100% quantiles of Tair, and the 90% quantile value of relative ΦFwithin each interval is calculated (circle dots in Figure 1 ). To reduce the uncertainty of Toptdetermination, the corresponding highest two ΦF( Figure 1Topt is estimated by averaging the Tair corresponding to the red dots in the scatter plot) as Topt. The estimated Topt represents the true Topt for the instantaneous photosynthesis during the 12:00-14:00 period, and can be converted to the daily Topt through the relationship between the LUE-derived instantaneous and daily Topt. Figure 1 An example demonstrates the process of estimating the optimum temperature (Topt) using the fluorescence efficiency (ΦF) and air temperature (Tair). The grey dots represent all paired values of ΦF and Tair within a 2° x 2° moving window during 2018-2021. The circles represent the 90th percentile of ΦF within each 5% quantile of Tair, and the red circles correspond to the two highest ΦF values among them.

[0042] The VPM model is widely used to estimate GPP, which takes photosynthetically active radiation (PAR), chlorophyll-related effective vegetation fraction (FPARchl), and light use efficiency (LUE) as inputs, as shown in equation (4). PAR is converted from shortwave radiation through an empirical relationship. FPARchl is estimated from the enhanced vegetation index (EVI), as shown in equation (5). LUE is scaled from the maximum light use efficiency (LUEmax) by a water adjustment factor (Wscalar) and a temperature adjustment factor (Tscalar), as shown in equation (6). The value of LUEmax is set to 0.42 gC / mol for C3 plants and 0.63 gC / mol for C4 plants in the original model. Wscalar is calculated as a function of the land surface water index (LSWI) and its annual maximum (LSWImax), as shown in equation (7). To reduce the effect of observation angle, EVI and LSWI are calculated using the MCD19A3-based nadir BRDF-adjusted reflectance with an 8-day time resolution. LSWImax takes the annual maximum in snow-free seasons. In the original VPM model, the temperature adjustment factor Tscalar is calculated based on the minimum temperature (Tmin), maximum temperature (Tmax), and optimum temperature (Topt) set for specific biomes.

[0043] GPP = PAR x FPARchl x LUE (4) chl

[0044] FPARchl = (EVI - 0.1) x 1.25 (5) chl

[0045] LUE = LUEmax x Wscalar x Tscalar (6) max scalar scalar

[0046] ​​​​​

[0047] In the improved VPM model, the Topt, based on the biome setting in the original VPM model, is replaced with the global optimum temperature (Topt) obtained from fluorescence efficiency (ΦF) inversion. Since the spatial distributions of the minimum temperature (Tmin) and maximum temperature (Tmax) are still unclear, the improved model retains the biome setting from the original VPM model. Therefore, the difference in GPP estimation results between the original and improved models is solely due to the Topt replacement. The original and improved VPM models were run separately to estimate the 8-day GPP, and the results were compared with the GPP observed by flux towers.

[0048] Topt verification of SIF and ΦF inversion:

[0049] The Topt values ​​derived from SIF and ΦF inversion were validated against those derived from Light Energy Use Efficiency (LUE). Prior to this, the instantaneous-scale Topt values ​​were converted to diurnal-scale values ​​according to the relationship shown in Figure S6 of the supplementary material. The results show that the correlation between Topt derived from SIF and LUE is weak, and its coefficient of determination R0 is low. 2 The value was 0.22, the root mean square error (RMSE) was 6.43℃, and the relative root mean square error (RRMSE) was 27.1%. Figure 2 A) indicates that the SIF method performs poorly in Topt estimation.

[0050] In contrast, ΦF shows a stronger correlation with Topt derived from LUE, R 2 The RMSE decreased to 3.74℃, reaching 0.66, and the RRMSE also decreased to 15.8%. Figure 2 B). These results demonstrate that the ΦF-based method has higher accuracy and reliability in estimating Topt compared to the SIF-based method.

[0051] Spatial distribution map of the world's optimal temperature (Topt):

[0052] The annual average Topt values ​​of global vegetation regions (excluding sparse vegetation) obtained from fluorescence efficiency (ΦF) inversion range from 10 to 33°C, exhibiting significant spatial heterogeneity (see...). Figure 3 ). Figure 3The global spatial distribution of the optimal temperature (Topt, unit: °C) based on the inversion of relative fluorescence efficiency (ΦF) with a spatial resolution of 0.2°. White areas on land represent low vegetation cover in arid and semi-arid regions, where the relative ΦF cannot be reliably estimated by the FCVI-based method. The maximum Topt (close to 33 °C) is mainly distributed in western India, southeastern China, and northern Australia, while the minimum Topt is concentrated in high-latitude regions. This geographical dependence of Topt can be clearly seen from its latitudinal distribution pattern (see Figure 4 A). In addition, among the 12 vegetation types, the average Topt of evergreen broadleaf forest (EBF) is the highest, at 26.3 ± 2.3 °C, while the average Topt of the remaining types is in the range of 20-25 °C (see Figure 4 B). Although the Topt of biomes set in the original VPM model is generally consistent with the trend of differences between different vegetation types, it ignores the spatial variation within each vegetation type (see Figure 4 B). This result indicates that the Topt estimation with spatial resolution can more accurately reflect the response of vegetation to temperature in different regions of the world than the biome average.

[0053] GPP estimation based on the adjusted VPM model:

[0054] A comprehensive comparison of the performance of the original and adjusted VPM models in GPP estimation was conducted. The results showed that the adjusted model performed slightly better (R 2 = 0.70 vs. 0.72, see Figure 5 A-B). In particular, at sites where the difference in Topt exceeded 5 °C, the R 2 of the adjusted model was significantly higher than that of the original model (0.70 vs. 0.65, see Figure 5 C-D). These results indicate that the adjusted VPM model can improve the accuracy of GPP estimation by more realistically simulating the response of vegetation to temperature.

[0055] The above merely describes the preferred embodiments of the present application, but the scope of protection of the present application is not limited thereto. Any changes or replacements within the technical scope disclosed by the present application can be easily conceived by those skilled in the art, and should be encompassed within the scope of protection of the present application. Therefore, the scope of protection of the present application should be subject to the scope of protection of the claims.

Claims

1. A method for estimating optimal temperature and optimizing a vegetation photosynthesis model based on fluorescence efficiency, characterized in that, The global TROPOMI satellite SIF product is obtained, and the relative fluorescence efficiency ΦF is simulated by combining the MODIS reflectance; The relationship curve between ΦF and ERA5 air temperature Tair is constructed within a 2° × 2° sliding window, and the Tair value corresponding to the highest fluorescence efficiency is extracted to estimate Topt; The Topt parameter of the original VPM model is replaced by the Topt estimated by ΦF to construct an improved VPM model, and the LUE estimated value is adjusted through the LUE temperature response function; The GPP estimated results of the original and improved models are compared at 203 global flux tower sites to evaluate the improvement effect. FCVI = NIR-VIS (2); 2. The method of claim 1, wherein the optimal temperature is estimated based on the fluorescence efficiency and the model of photosynthesis of vegetation is optimized. Fluorescence efficiency Φ F From the inversion of TROPOMISIF data: The visible light reflectance VIS is the weighted sum of the reflectances of red R, green G, and blue B bands, and is calculated as follows: where SIF obs denotes observed TROPOSIF data, PAR denotes photosynthetically active radiation, FCI denotes fluorescence canopy index, NIR denotes near-infrared reflectance, and VIS denotes visible reflectance.

3. The method of claim 2, wherein the optimal temperature is estimated based on the fluorescence efficiency and the model of photosynthesis of vegetation is optimized. VIS = 0.331R + 0.424B + 0.246G (3). The Topt estimated by ΦF replaces the biome Topt parameter in the original VPM model, including:

4. The method of claim 1, wherein the optimal temperature is estimated based on the fluorescence efficiency and the model of photosynthesis of vegetation is optimized. EVI and LSWI use the near-nadir bidirectional reflectance distribution function correction reflectance calculation based on MCD19A3; The VPM model is widely used to estimate GPP, which takes photosynthetically active radiation (PAR), chlorophyll-related fraction of absorbed photosynthetically active radiation (FPAR), and light use efficiency (LUE) as inputs chl and light use efficiency (LUE) as inputs PAR is converted from shortwave radiation by an empirical relationship, FPAR chl is then estimated by the enhanced vegetation index, EVI LUE is the maximum light use efficiency LUE max by the water regulation factor W scalar and the temperature regulation factor T scalar scaled; LUE max The value is set to 0.42 g C / mol for C3 plants and 0.63 g C / mol for C4 plants in the original model; W scalar is calculated as a function of the surface water index LSWI and its annual maximum LSWI max max Estimating GPP includes: LSWI max Taking the annual maximum in the snow-free season, the temperature adjustment factor T scalar in the original VPM model depends on the minimum temperature T min , the maximum temperature T max and the optimum temperature T opt set for a particular biome.

5. The method of claim 4, wherein the optimal temperature is estimated based on the fluorescence efficiency and the model of photosynthesis of vegetation is optimized. Calculating the light use efficiency LUE includes: GPP = PAR x FPAR chl x LUE (4).

6. The method of claim 4, wherein the optimal temperature is estimated based on the fluorescence efficiency and the model of photosynthesis of vegetation is optimized. Computing the fraction of absorbed photosynthetically active radiation FPAR chl comprising: FPAR chl = (EVI - 0.1) x 1.25 (5).

7. The method of claim 4, wherein the optimal temperature is estimated based on the fluorescence efficiency and the model of photosynthesis of vegetation is optimized. ​ LUE = LUE max x W scalar x T scalar (6).

8. The method of claim 4, wherein the optimal temperature is estimated based on the fluorescence efficiency and the model of photosynthesis of vegetation is optimized. Computing a moisture adjustment factor W scalar comprises:

9. The method of claim 4, wherein the optimal temperature is estimated based on the fluorescence efficiency and the model of photosynthesis of vegetation is optimized. Calculating a temperature adjustment factor T scalar comprising:

Images