Reconstruction of U.S. Regional-Scale Soybean SIF Based on MODIS Data and BP Neural Network

doi:10.12133/j.smartag.SA202309006

Abstract

Abstract:

Objective SIF (Solar Induced Fluorescence) data obtained from satellites suffer from issues such as low spatial and temporal resolution, and discrete footprint because of the limitations imposed by satellite orbits. To address these problems, obtaining higher resolution SIF data, most studies are based on low-resolution satellite SIF for reconstruction. Moreover, the spatial resolution of most SIF reconstruction products is still not enough to be directly used for the study of crop photosynthetic rate at the regional scale. Although some SIF products boast elevated resolutions, it is crucial to underscore that these derive not from the original satellite SIF data reconstruct but instead evolve from secondary reconstructions based on preexisting SIF reconstruction products. A satellite named OCO-2 (The Orbiting Carbon Obsevatory-2) equipped with a high-resolution spectrometer, OCO-2 SIF has higher spatial resolution (1.29×2.25 km) compared to other original SIF products. It holds profound significance in advancing the realm of high-resolution SIF data reconstruction, particularly within the context of regional-scale crop studies. Methods This research primarily exploration SIF reconstruct at the regional scale, mainly focused on the partial soybean planting regions nestled within the United States. The selection of MODIS raw data hinged on a meticulous consideration of environmental conditions, the distinctive physiological attributes of soybeans, and an exhaustive evaluation of factors intricately linked to OCO-2 SIF within these soybean planting regions. This foundational research laid the groundwork for this subsequent investigation into reconstructing SIF using the original OCO-2 SIF data and MODIS products. The primary tasks encompassed reconstructing high resolution soybean SIF while concurrently executing a rigorous assessment of the reconstructed SIF's quality. During the dataset construction process, amalgamated SIF data from multiple soybean planting regions traversed by the OCO-2 satellite's footprint, the purpose was to retain as many of the available original SIF samples as possible. This approach has provided the subsequent SIF reconstruction model with a rich source of SIF data. SIF data obtained beneath the satellite's trajectory were matched with various MODIS datasets, including Enhanced Vegetation Index (EVI), Fraction of Photosynthetically Active Radiation (FPAR), and Land Surface Temperature (LST), resulting in the creation of a multisource remote sensing dataset ultimately used for model training. Because of the multisource remote sensing dataset encompassed the most relevant explanatory variables within each SIF footprint coverage area concerning soybean physiological structure and environmental conditions. Through the activation functions in the BP neural network, it enhanced the understanding of the complex nonlinear relationships between the original SIF data and these MODIS products. Leveraging these inherent nonlinear relationships, compared and analyzed the effects of different combinations of explanatory variables on SIF reconstruction, mainly analyzing the three indicators of goodness of fit R², root mean square error RMSE, and mean absolute error MAE, and then selecting the best SIF reconstruction model, generate a regional scale, spatially continuous, and high temporal resolution (500 m, 8 d) soybean SIF reconstruction dataset (BPSIF). Results and Discussions The research findings confirmed the strong performance of the SIF reconstruction model in predicting soybean SIF. After simultaneously incorporating Enhanced Vegetation Index (EVI), Fraction of Photosynthetically Active Radiation (FPAR), and Land Surface Temperature (LST) as explanatory variables to model, achieved a goodness of fit with an R² value of 0.84, this statistical metric validated the model's capability in predicting SIF data, it also reflected that the reconstructed 8 d time resolution of SIF data's reliability of applying to small-scale agricultural crop photosynthesis research with 500 m×500 m spatial scale. Based on this optimal model, generated the reconstructed SIF product (BPSIF). The Pearson correlation coefficient between the original OCO-2 SIF data and MODIS GPP stood were at a modest 0.53. In stark contrast, the correlation coefficient between BPSIF and MODIS Gross Primary Productivity (GPP) rosed significantly to 0.80. The increased correlation suggests that BPSIF could more accurately reflect the dynamic changes in GPP during the soybean growing season, making it more reliable compared to the original SIF data. This research selected soybean planting areas in the United States with relatively single crop cultivation as the research area, based on high spatial resolution (1.29 km×2.25 km) OCO-2 SIF data, greatly reduces vegetation heterogeneity under a single SIF footprint. Conclusions The resolution of most SIF reconstruction data is not yet sufficient for application in the agricultural field. Although some SIF reconstruction data have higher resolution, these data are not reconstructed from the original satellite SIF data, but evolved from the secondary reconstruction of pre-existing SIF reconstruction data, and the errors introduced cannot be ignored. BPSIF has significantly enhancing the regional and temporal continuity of OCO-2 SIF while preserving the time and spatial attributes contained in the original SIF dataset. Within the study area, BPSIF exhibits a significantly improved correlation with MODIS GPP compared to the original OCO-2 SIF. The proposed OCO-2 SIF data reconstruction method in this study holds the potential to provide a more reliable SIF dataset. This dataset has the potential to drive further understanding of soybean SIF at finer spatial and temporal scales, as well as find its relationship with soybean GPP.

Key words: spaceborne SIF data, MODIS data, BP neural networks, soybean SIF refactoring

YAO Jianen, LIU Haiqiu, YANG Man, FENG Jinying, CHEN Xiu, ZHANG Peipei. Reconstruction of U.S. Regional-Scale Soybean SIF Based on MODIS Data and BP Neural Network[J]. Smart Agriculture, doi: 10.12133/j.smartag.SA202309006.

Figures/Tables 11

Fig. 1

Fig. 2

Table 1

Table 2

Fig. 3

Table 3

Fig. 4

Fig. 5

Fig. 6

Fig. 7

References 35

1	ZHANG Z Y, ZHANG Y G, ZHANG Y, et al. The potential of satellite FPAR product for GPP estimation: An indirect evaluation using solar-induced chlorophyll fluorescence[J]. Remote sensing of environment, 2020, 240: ID 111686.
2	BAKER N R. Chlorophyll fluorescence: A probe of photosynthesis in vivo [J]. Annual review of plant biology, 2008, 59: 89-113.
3	ZARCO-TEJADA P J, CATALINA A, GONZÁLEZ M R, et al. Relationships between net photosynthesis and steady-state chlorophyll fluorescence retrieved from airborne hyperspectral imagery[J]. Remote sensing of environment, 2013, 136: 247-258.
4	HAO P Y, DI L P, ZHANG C, et al. Transfer learning for crop classification with cropland data layer data (CDL) as training samples[J]. Science of the total environment, 2020, 733: ID 138869.
5	KÖHLER P, GUANTER L, JOINER J. A linear method for the retrieval of Sun-induced chlorophyll fluorescence from GOME-2 and SCIAMACHY data[J]. Atmospheric measurement techniques, 2015, 8(6): 2589-2608.
6	JOINER J, GUANTER L, LINDSTROT R, et al. Global monitoring of terrestrial chlorophyll fluorescence from moderate-spectral-resolution near-infrared satellite measurements: Methodology, simulations, and application to GOME-2[J]. Atmospheric measurement techniques, 2013, 6(10): 2803-2823.
7	JOINER J, YOSHIDA Y, VASILKOV A P, et al. First observations of global and seasonal terrestrial chlorophyll fluorescence from space[J]. Biogeosciences, 2011, 8(3): 637-651.
8	SUN Y, FRANKENBERG C, JUNG M, et al. Overview of Solar-Induced chlorophyll Fluorescence (SIF) from the Orbiting Carbon Observatory-2: Retrieval, cross-mission comparison, and global monitoring for GPP[J]. Remote sensing of environment, 2018, 209: 808-823.
9	FRANKENBERG C, O'DELL C, BERRY J, et al. Prospects for chlorophyll fluorescence remote sensing from the orbiting carbon observatory-2[J]. Remote sensing of environment, 2014, 147: 1-12.
10	章钊颖, 王松寒, 邱博, 等. 日光诱导叶绿素荧光遥感反演及碳循环应用进展[J]. 遥感学报, 2019, 23(1): 37-52.
	ZHANG Z Y, WANG S H, QIU B, et al. Retrieval of Sun-induced chlorophyll fluorescence and advancements in carbon cycle application[J]. Journal of remote sensing, 2019, 23(1): 37-52.
11	ZHANG L L, ZHANG Z, LUO Y C, et al. Combining optical, fluorescence, thermal satellite, and environmental data to predict county-level maize yield in China using machine learning approaches[J]. Remote sensing, 2019, 12(1): ID 21.
12	CAO J, ZHANG Z, TAO F L, et al. Integrating multi-source data for rice yield prediction across China using machine learning and deep learning approaches[J]. Agricultural and forest meteorology, 2021,297: ID 108275.
13	GUO M, LI J, HUANG S B, et al. Feasibility of using MODIS products to simulate Sun-induced chlorophyll fluorescence (SIF) in boreal forests[J]. Remote sensing, 2020, 12(4): ID 680.
14	YANG P Q, VANDERTOL C, VERHOEF W,et al.Using reflectance to explain vegetation biochemical and structural effects on sun-induced chlorophyll fluorescence[J]. Remote sensing of environment, 2019, 231: ID 110996.
15	DUVEILLER G, FILIPPONI F, WALTHER S, et al. A spatially downscaled Sun-induced fluorescence global product for enhanced monitoring of vegetation productivity[J]. Earth system science data, 2020, 12(2): 1101-1116.
16	MA Y, LIU L Y, LIU X J, et al. An improved downscaled Sun-induced chlorophyll fluorescence (DSIF) product of GOME-2 dataset[J]. European journal of remote sensing, 2022, 55(1): 168-180.
17	LI X, XIAO J F. A global, 0.05-degree product of solar-induced chlorophyll fluorescence derived from OCO-2, MODIS, and reanalysis data[J]. Remote sensing, 2019, 11(5): ID 517.
18	YU L, WEN J, CHANG C Y, et al. High-resolution global contiguous SIF of OCO-2[J]. Geophysical research letters, 2019, 46(3): 1449-1458.
19	DUVEILLER G, CESCATTI A. Spatially downscaling Sun-induced chlorophyll fluorescence leads to an improved temporal correlation with gross primary productivity[J]. Remote sensing of environment, 2016, 182: 72-89.
20	WEN J, KÖHLER P, DUVEILLER G, et al. A framework for harmonizing multiple satellite instruments to generate a long-term global high spatial-resolution solar-induced chlorophyll fluorescence (SIF)[J]. Remote sensing of environment, 2020, 239: ID 111644.
21	LIU X J, LIU L Y, BACOUR C, et al. A simple approach to enhance the TROPOMI solar-induced chlorophyll fluorescence product by combining with canopy reflected radiation at near-infrared band[J]. Remote sensing of environment, 2023, 284: ID 113341.
22	KANG X Y, HUANG C P, ZHANG L F, et al. Downscaling solar-induced chlorophyll fluorescence for field-scale cotton yield estimation by a two-step convolutional neural network[J]. Computers and electronics in agriculture, 2022, 201: ID 107260.
23	YOSHIDA Y, JOINER J, TUCKER C, et al. The 2010 Russian drought impact on satellite measurements of solar-induced chlorophyll fluorescence: Insights from modeling and comparisons with parameters derived from satellite reflectances[J]. Remote sensing of environment, 2015, 166: 163-177.
24	YANG X, TANG J W, MUSTARD J F, et al. Solar-induced chlorophyll fluorescence that correlates with canopy photosynthesis on diurnal and seasonal scales in a temperate deciduous forest[J]. Geophysical research letters, 2015, 42(8): 2977-2987.
25	KIRA O, Y-Y CHANG C, GU L, et al. Partitioning net ecosystem exchange (NEE) of CO₂ using solar-induced chlorophyll fluorescence (SIF)[J]. Geophysical research letters, 2021, 48(4): ID e2020GL091247.
26	GAO Y, WANG S H, GUAN K Y, et al. The ability of Sun-induced chlorophyll fluorescence from OCO-2 and MODIS-EVI to monitor spatial variations of soybean and maize yields in the Midwestern USA[J]. Remote sensing, 2020, 12(7): ID 1111.
27	MOHAMMED G H, COLOMBO R, MIDDLETON E M, et al. Remote sensing of solar-induced chlorophyll fluorescence (SIF) in vegetation: 50 years of progress[J]. Remote sensing of environment, 2019, 231: ID 111177.
28	梁守真, 马万栋, 王猛, 等. 冠层绿色FPAR与植被指数关系及其对气溶胶的敏感性分析[J]. 测绘与空间地理信息, 2018, 41(12): 11-14.
	LIANG S Z, MA W D, WANG M, et al. Analysis on the relationship between green FPAR and vegetation indices and their sensitivity to aerosol optical depth[J]. Geomatics & spatial information technology, 2018, 41(12): 11-14.
29	KÖEHLER P, FRANKENBERG C, MAGNEY T S, et al. Global retrievals of solar induced chlorophyll fluorescence with TROPOMI: First results and inter-sensor comparison to OCO-2[J]. Geophysical research letters, 2018, 45(19): 10456-10463.
30	FRANKENBERG C, FISHER J B, WORDEN J, et al. New global observations of the terrestrial carbon cycle from GOSAT: Patterns of plant fluorescence with gross primary productivity[J]. Geophysical research letters, 2011, 38(17): 2011GL048738.
31	GUANTER L, FRANKENBERG C, DUDHIA A, et al. Retrieval and global assessment of terrestrial chlorophyll fluorescence from GOSAT space measurements[J]. Remote sensing of environment, 2012, 121: 236-251.
32	PARAZOO N C, BOWMAN K, FRANKENBERG C, et al. Interpreting seasonal changes in the carbon balance of southern Amazonia using measurements of XCO₂ and chlorophyll fluorescence from GOSAT[J]. Geophysical research letters, 2013, 40(11): 2829-2833.
33	GENTINE P, ALEMOHAMMAD S H. Reconstructed solar-induced fluorescence: A machine learning vegetation product based on MODIS surface reflectance to reproduce GOME-2 solar-induced fluorescence[J]. Geophysical research letters, 2018, 45(7): 3136-3146.
34	LI X, XIAO J F, HE B B. Chlorophyll fluorescence observed by OCO-2 is strongly related to gross primary productivity estimated from flux towers in temperate forests[J]. Remote sensing of environment, 2018, 204: 659-671.
35	MA Y, LIU L Y, CHEN R N, et al. Generation of a global spatially continuous TanSat solar-induced chlorophyll fluorescence product by considering the impact of the solar radiation intensity[J]. Remote sensing, 2020, 12(13): ID 2167.

数据	时间分辨率/d	空间分辨率/m	波段
EVI	8	500	Gap-Filled and Timesat-Smoothed EVI
FPAR	8	500	Composed FPAR from MODIS and Converted FPAR
LST	8	1 000	8-day daytime 1 km grid LST
GPP	8	500	MODIS/Aqua GPP 8-Day L4 Global 500m SIN Grid

指标	机器学习算法			深度学习算法
指标	决策树回归	随机森林回归	支持向量机回归归	CNN	BP神经网络
R ²	0.64	0.68	0.70	0.742	0.841
RMSE	0.123	0.118	0.114	0.241	0.152
MAE	0.096	0.093	0.089	0.193	0.039

月份	EVI	FPAR	LST
5月	0.030 1	2.90E-96	2.96E-322
6月	0.013 3	1.44E-44	4.32E-164
7月	3.59E-37	3.66E-8	4.14E-8
8月	7.02E-14	0.000 2	4.94E-182
9月	6.38E-14	0.042 2	3.27E-12