scholarly journals Feasibility of Using MODIS Products to Simulate Sun-Induced Chlorophyll Fluorescence (SIF) in Boreal Forests

2020 ◽  
Vol 12 (4) ◽  
pp. 680 ◽  
Author(s):  
Meng Guo ◽  
Jing Li ◽  
Shubo Huang ◽  
Lixiang Wen
Keyword(s):  
Chlorophyll Fluorescence ◽  
Boreal Forest ◽  
Spatial Resolution ◽  
Vegetation Index ◽  
Mean Squared Error ◽  
Gross Primary Production ◽  
Coniferous Forest ◽  
Terrestrial Ecosystems ◽  
Linear Regression Models ◽  
Modis Evi

Solar-induced chlorophyll fluorescence (SIF) is a novel approach to gain information about plant activity from remote sensing observations. However, there are currently no continuous SIF data produced at high spatial resolutions. Many previous studies have discussed the relationship between SIF and gross primary production (GPP) and showed a significant correlation between them, but few researchers have focused on forests, which are one the most important terrestrial ecosystems. This study takes Greater Khingan Mountains, a typical boreal forest in China, as an example to explore the feasibility of using MODerate resolution Imaging Spectroradiometer (MODIS) products and Orbiting Carbon Observatory-2 (OCO-2) SIF data to simulate continuous SIF at higher spatial resolutions. The results show that there is no significant correlation between SIF and MODIS GPP at a spatial resolution of 1 km; however, significant correlations between SIF and the enhanced vegetation index (EVI) were found during growing seasons. Furthermore, the broadleaf forest has a higher SIF than coniferous forest because of the difference in leaf and canopy bio-chemical and structural characteristic. When using MODIS EVI to model SIF, linear regression models show average performance (R2 = 0.58, Root Mean Squared Error (RMSE) = 0.14 from Julian day 145 to 257) at a 16-day time scale. However, when using MODIS EVI and temperature, multiple regressions perform better (R2 = 0.71, RMSE = 0.13 from Julian day 145 to 241). An important contribution of this paper is the analysis of the relationships between SIF and vegetation indices at different spatial resolutions and the finding that the relationships became closer with a decrease in spatial resolution. From this research, we conclude that the SIF of the boreal forest investigated can mainly be explained by EVI and air temperature.

scholarly journals Modelling sun-induced fluorescence and photosynthesis with a land surface model at local and regional scales in northern Europe

2017 ◽  
Vol 14 (7) ◽  
pp. 1969-1987 ◽  
Author(s):  
Tea Thum ◽  
Sönke Zaehle ◽  
Philipp Köhler ◽  
Tuula Aalto ◽  
Mika Aurela ◽  
...  
Keyword(s):  
Chlorophyll Fluorescence ◽  
Land Surface ◽  
Large Scale ◽  
Model Simulation ◽  
Gross Primary Production ◽  
Regional Scale ◽  
Coniferous Forest ◽  
Flux Measurement ◽  
Surface Model ◽  
Fluorescence Measurements

Abstract. Recent satellite observations of sun-induced chlorophyll fluorescence (SIF) are thought to provide a large-scale proxy for gross primary production (GPP), thus providing a new way to assess the performance of land surface models (LSMs). In this study, we assessed how well SIF is able to predict GPP in the Fenno-Scandinavian region and what potential limitations for its application exist. We implemented a SIF model into the JSBACH LSM and used active leaf-level chlorophyll fluorescence measurements (Chl F) to evaluate the performance of the SIF module at a coniferous forest at Hyytiälä, Finland. We also compared simulated GPP and SIF at four Finnish micrometeorological flux measurement sites to observed GPP as well as to satellite-observed SIF. Finally, we conducted a regional model simulation for the Fenno-Scandinavian region with JSBACH and compared the results to SIF retrievals from the GOME-2 (Global Ozone Monitoring Experiment-2) space-borne spectrometer and to observation-based regional GPP estimates. Both observations and simulations revealed that SIF can be used to estimate GPP at both site and regional scales. At regional scale the model was able to simulate observed SIF averaged over 5 years with r2 of 0.86. The GOME-2-based SIF was a better proxy for GPP than the remotely sensed fAPAR (fraction of absorbed photosynthetic active radiation by vegetation). The observed SIF captured the seasonality of the photosynthesis at site scale and showed feasibility for use in improving of model seasonality at site and regional scale.

scholarly journals Global Vegetation Photosynthetic Phenology Products Based on MODIS Vegetation Greenness and Temperature: Modeling and Evaluation

2021 ◽  
Vol 13 (24) ◽  
pp. 5080
Author(s):  
Xiaojun Xu ◽  
Yan Tang ◽  
Yiling Qu ◽  
Zhongsheng Zhou ◽  
Junguo Hu
Keyword(s):  
Land Surface ◽  
Vegetation Index ◽  
Deciduous Forest ◽  
Forest Type ◽  
Gross Primary Production ◽  
Plant Functional Types ◽  
Linear Regression Models ◽  
Climatic Regions ◽  
Functional Types ◽  
Global Vegetation

Land surface phenology (LSP) products that are derived from different data sources have different definitions and biophysical meanings. Discrepancies among these products and their linkages with carbon fluxes across plant functional types and climatic regions remain somewhat unclear. In this study, to differentiate LSP related to gross primary production (GPP) from LSP related to remote sensing data, we defined the former as vegetation photosynthetic phenology (VPP), including the starting and ending days of GPP (SOG and EOG, respectively). Specifically, we estimated VPP based on a combination of observed VPP from 145 flux-measured GPP sites together with the vegetation index and temperature data from MODIS products using multiple linear regression models. We then compared VPP estimates with MODIS LSP on a global scale. Our results show that the VPP provided better estimates of SOG and EOG than MODIS LSP, with a root mean square error (RMSE) for SOG of 12.7 days and a RMSE for EOG of 10.5 days. The RMSE was approximately three weeks for both SOG and EOG estimates of the non-forest type. Discrepancies between VPP and LSP estimates varied across plant functional types (PFTs) and climatic regions. A high correlation was observed between VPP and LSP estimates for deciduous forest. For most PFTs, using VPP estimates rather than LSP improved the estimation of GPP. This study presents a useful method for modeling global VPP, investigates in detail the discrepancies between VPP and LSP, and provides a more effective global vegetation phenology product for carbon cycle modeling than the existing ones.

scholarly journals Comparative Analysis on the Estimation of Diurnal Solar-Induced Chlorophyll Fluorescence Dynamics for a Subtropical Evergreen Coniferous Forest

2021 ◽  
Vol 13 (16) ◽  
pp. 3143
Author(s):  
Jinghua Chen ◽  
Shaoqiang Wang ◽  
Bin Chen ◽  
Yue Li ◽  
Muhammad Amir ◽  
...  
Keyword(s):  
Chlorophyll Fluorescence ◽  
Temporal Dynamics ◽  
Gross Primary Production ◽  
Stomatal Closure ◽  
Chlorophyll Concentration ◽  
Coniferous Forest ◽  
Diurnal Variations ◽  
Future Climate Change ◽  
Ecosystem Carbon ◽  
Seasonal And Diurnal Variations

Solar-induced chlorophyll fluorescence (SIF) is considered as a prospective indicator of vegetation photosynthetic activity and the ecosystem carbon cycle. The current coarse spatial-temporal resolutions of SIF data from satellite missions and ground measurements still cannot satisfy the corroboration of its correlation with photosynthesis and carbon flux. Practical approaches are needed to be explored for the supplementation of the SIF measurements. In our study, we clarified the diurnal variations of leaf and canopy chlorophyll fluorescence for a subtropical evergreen coniferous forest and evaluated the performance of the canopy chlorophyll concentration (CCC) approach and the backward approach from gross primary production (GPP) for estimating the diurnal variations of canopy SIF by comparing with the Soil Canopy Observation Photosynthesis Energy (SCOPE) model. The results showed that the canopy SIF had similar seasonal and diurnal variations with the incident photosynthetically active radiation (PAR) above the canopy, while the leaf steady-state fluorescence remained stable during the daytime. Neither the CCC nor the raw backward approach from GPP could capture the short temporal dynamics of canopy SIF. However, after improving the backward approach with a correction factor of normalized PAR incident on leaves, the variation of the estimated canopy SIF accounted for more than half of the diurnal variations in the canopy SIF (SIF687: R2 = 0.53, p < 0.001; SIF760: R2 = 0.72, p < 0.001) for the subtropical evergreen coniferous forest without water stress. Drought interfered with the utilization of the improved backward approach because of the decoupling of SIF and GPP due to stomatal closure. This new approach offers new insight into the estimation of diurnal canopy SIF and can help understand the photosynthesis of vegetation for future climate change studies.

scholarly journals The Links between Canopy Solar-Induced Chlorophyll Fluorescence and Gross Primary Production Responses to Meteorological Factors in the Growing Season in Deciduous Broadleaf Forest

2021 ◽  
Vol 13 (12) ◽  
pp. 2363
Author(s):  
Xiangfen Cheng ◽  
Yu Zhou ◽  
Meijun Hu ◽  
Feng Wang ◽  
Hui Huang ◽  
...  
Keyword(s):  
Remote Sensing ◽  
Chlorophyll Fluorescence ◽  
Primary Production ◽  
Forest Canopy ◽  
Meteorological Factors ◽  
Gross Primary Production ◽  
Terrestrial Ecosystems ◽  
Growing Season ◽  
Near Surface ◽  
The Relationship

Solar-induced chlorophyll fluorescence (SIF) is a hopeful indicator, which along with remote sensing, is used to measure the photosynthetic efficiency and gross primary production (GPP) of vegetation in regional terrestrial ecosystems. Studies have found a significant linear correlation between SIF and GPP in a variety of ecosystems. However, this relationship has mainly been established using SIF and GPP data derived from satellite remote sensing and continuous ground-based observations, respectively, which are difficult to accurately match. To overcome this, some studies have begun to use tower-based automatic observation instruments to study the changes of near-surface SIF and GPP. This study conducts continuous simultaneous observation of SIF, carbon flux, and meteorological factors on the forest canopy of a cork oak plantation during the growing season to explore how meteorological factors impact on canopy SIF and its relationship with GPP. This research found that the canopy SIF has obvious diurnal and day-to-day variations during the growing season but overall is relatively stable. Furthermore, SIF is greatly affected by incident radiation in different weather conditions and can change daily. Meteorological factors have a major role in the relationship between SIF and GPP; overall, the relationship shows a significant linear regression on the 30 min scale, but weakens when aggregating to the diurnal scale. Photosynthetically active radiation (PAR) drives SIF on a daily basis and changes the relationship between SIF and GPP on a seasonal timescale. As PAR increases, the daily slopes of the linear regressions between SIF and GPP decrease. On the 30 min timescale, both SIF and GPP increase with PAR until it reaches 1250 μmol·m−2·s−1; subsequently, SIF continues to increase while GPP decreases and they show opposite trends. Soil moisture and vapor pressure deficit influence SIF and GPP, respectively. Our findings demonstrate that meteorological factors affect the relationship between SIF and GPP, thereby enhancing the understanding of the mechanistic link between chlorophyll fluorescence and photosynthesis.

scholarly journals Modelling Daily Gross Primary Productivity with Sentinel-2 Data in the Nordic Region–Comparison with Data from MODIS

2021 ◽  
Vol 13 (3) ◽  
pp. 469
Author(s):  
Zhanzhang Cai ◽  
Sofia Junttila ◽  
Jutta Holst ◽  
Hongxiao Jin ◽  
Jonas Ardö ◽  
...  
Keyword(s):  
Spatial Resolution ◽  
Primary Productivity ◽  
Vegetation Index ◽  
Regional Scale ◽  
Gross Primary Productivity ◽  
Environmental Drivers ◽  
Linear Regression Models ◽  
Biophysical Parameters ◽  
Moderate Resolution Imaging Spectroradiometer ◽  
Sentinel 2

The high-resolution Sentinel-2 data potentially enable the estimation of gross primary productivity (GPP) at finer spatial resolution by better capturing the spatial variation in a heterogeneous landscapes. This study investigates the potential of 10 m resolution reflectance from the Sentinel-2 Multispectral Instrument to improve the accuracy of GPP estimation across Nordic vegetation types, compared with the 250 m and 500 m resolution reflectance from the Moderate Resolution Imaging Spectroradiometer (MODIS). We applied linear regression models with inputs of two-band enhanced vegetation index (EVI2) derived from Sentinel-2 and MODIS reflectance, respectively, together with various environmental drivers to estimate daily GPP at eight Nordic eddy covariance (EC) flux tower sites. Compared with the GPP from EC measurements, the accuracies of modelled GPP were generally high (R2 = 0.84 for Sentinel-2; R2 = 0.83 for MODIS), and the differences between Sentinel-2 and MODIS were minimal. This demonstrates the general consistency in GPP estimates based on the two satellite sensor systems at the Nordic regional scale. On the other hand, the model accuracy did not improve by using the higher spatial-resolution Sentinel-2 data. More analyses of different model formulations, more tests of remotely sensed indices and biophysical parameters, and analyses across a wider range of geographical locations and times will be required to achieve improved GPP estimations from Sentinel-2 satellite data.

scholarly journals A Convolutional Neural Network for Spatial Downscaling of Satellite-Based Solar-Induced Chlorophyll Fluorescence (SIFnet)

2021 ◽  
Author(s):  
Johannes Gensheimer ◽  
Alexander Jay Turner ◽  
Philipp Köhler ◽  
Christian Frankenberg ◽  
Jia Chen
Keyword(s):  
Neural Network ◽  
United States ◽  
Chlorophyll Fluorescence ◽  
Convolutional Neural Network ◽  
Spatial Resolution ◽  
Vegetation Index ◽  
Spatial Scales ◽  
The United States ◽  
Cycle Space ◽  
Auxiliary Data

Abstract. Gross primary productivity (GPP) is the sum of leaf photosynthesis and represents a crucial component of the global carbon cycle. Space-borne estimates of GPP typically rely on observable quantities that co-vary with GPP such as vegetation indices using reflectance measurements (e.g., NDVI, NIRv, and kNDVI). Recent work has also utilized measurements of solar-induced chlorophyll fluorescence (SIF) as a proxy for GPP. However, these SIF measurements are typically coarse resolution while many processes influencing GPP occur at fine spatial scales. Here, we develop a Convolutional Neural Network (CNN), named SIFnet, that increases the resolution of SIF from the TROPOspheric Monitoring Instrument (TROPOMI) on board of the satellite Sentinel-5P by a factor of 10 to a spatial resolution of 500 m. SIFnet utilizes coarse SIF observations together with high resolution auxiliary data. The auxiliary data used here may carry information related to GPP and SIF. We use training data from non-US regions between April 2018 until March 2021 and evaluate our CNN over the conterminous United States (CONUS). We show that SIFnet is able to increase the resolution of TROPOMI SIF by a factor of 10 with a r2 and RMSE metrics of 0.92 and 0.17 mW m−2 sr−1 nm−1, respectively. We further compare SIFnet against a recently developed downscaling approach and evaluate both methods against independent SIF measurements from Orbiting Carbon Observatory 2 and 3 (OCO-2/3). SIFnet performs systematically better than the downscaling approach (r = 0.78 for SIFnet, r = 0.72 for downscaling), indicating that it is picking up on key features related to SIF and GPP. Examination of the feature importance in the neural network indicates a few key parameters and the spatial regions these parameters matter. Namely, the CNN finds low resolution SIF data to be the most significant parameter with the NIRv vegetation index as the second most important parameter. NIRv consistently outperforms the recently proposed kNDVI vegetation index. Advantages and limitations of SIFnet are investigated and presented through a series of case studies across the United States. SIFnet represents a robust method to infer continuous, high spatial resolution SIF data.

scholarly journals A global spatially contiguous solar-induced fluorescence (CSIF) dataset using neural networks

2018 ◽  
Vol 15 (19) ◽  
pp. 5779-5800 ◽  
Author(s):  
Yao Zhang ◽  
Joanna Joiner ◽  
Seyed Hamed Alemohammad ◽  
Sha Zhou ◽  
Pierre Gentine
Keyword(s):  
Vegetation Index ◽  
Normalized Difference Vegetation Index ◽  
Gross Primary Production ◽  
Terrestrial Ecosystems ◽  
Temporal Variations ◽  
Clear Sky ◽  
Moderate Resolution Imaging Spectroradiometer ◽  
Regression Slopes ◽  
The Individual ◽  
Sky Conditions

Abstract. Satellite-retrieved solar-induced chlorophyll fluorescence (SIF) has shown great potential to monitor the photosynthetic activity of terrestrial ecosystems. However, several issues, including low spatial and temporal resolution of the gridded datasets and high uncertainty of the individual retrievals, limit the applications of SIF. In addition, inconsistency in measurement footprints also hinders the direct comparison between gross primary production (GPP) from eddy covariance (EC) flux towers and satellite-retrieved SIF. In this study, by training a neural network (NN) with surface reflectance from the MODerate-resolution Imaging Spectroradiometer (MODIS) and SIF from Orbiting Carbon Observatory-2 (OCO-2), we generated two global spatially contiguous SIF (CSIF) datasets at moderate spatiotemporal (0.05∘ 4-day) resolutions during the MODIS era, one for clear-sky conditions (2000–2017) and the other one in all-sky conditions (2000–2016). The clear-sky instantaneous CSIF (CSIFclear-inst) shows high accuracy against the clear-sky OCO-2 SIF and little bias across biome types. The all-sky daily average CSIF (CSIFall-daily) dataset exhibits strong spatial, seasonal and interannual dynamics that are consistent with daily SIF from OCO-2 and the Global Ozone Monitoring Experiment-2 (GOME-2). An increasing trend (0.39 %) of annual average CSIFall-daily is also found, confirming the greening of Earth in most regions. Since the difference between satellite-observed SIF and CSIF is mostly caused by the environmental down-regulation on SIFyield, the ratio between OCO-2 SIF and CSIFclear-inst can be an effective indicator of drought stress that is more sensitive than the normalized difference vegetation index and enhanced vegetation index. By comparing CSIFall-daily with GPP estimates from 40 EC flux towers across the globe, we find a large cross-site variation (c.v. = 0.36) of the GPP–SIF relationship with the highest regression slopes for evergreen needleleaf forest. However, the cross-biome variation is relatively limited (c.v. = 0.15). These two contiguous SIF datasets and the derived GPP–SIF relationship enable a better understanding of the spatial and temporal variations of the GPP across biomes and climate.

scholarly journals Sun-induced Chlorophyll fluorescence and PRI improve remote sensing GPP estimates under varying nutrient availability in a typical Mediterranean savanna ecosystem

2015 ◽  
Vol 12 (14) ◽  
pp. 11891-11934 ◽  
Author(s):  
O. Perez-Priego ◽  
J. Guan ◽  
M. Rossini ◽  
F. Fava ◽  
T. Wutzler ◽  
...  
Keyword(s):  
Remote Sensing ◽  
Chlorophyll Fluorescence ◽  
Vegetation Index ◽  
Near Infrared ◽  
Gross Primary Production ◽  
High Spectral Resolution ◽  
N Availability ◽  
P Availability ◽  
N Content ◽  
Near Surface

Abstract. This study investigates the performances of different optical indices to estimate gross primary production (GPP) of herbaceous stratum in a Mediterranean savanna with different Nitrogen (N) and Phosphorous (P) availability. Sun-induced chlorophyll Fluorescence yield computed at 760 nm (Fy760), scaled-photochemical reflectance index (sPRI), MERIS terrestrial-chlorophyll index (MTCI) and Normalized difference vegetation index (NDVI) were computed from near-surface field spectroscopy measurements collected using high spectral resolution spectrometers covering the visible near-infrared regions. GPP was measured using canopy-chambers on the same locations sampled by the spectrometers. We hypothesized that light-use efficiency (LUE) models driven by remote sensing quantities (RSM) can better track changes in GPP caused by nutrient supplies compared to those driven exclusively by meteorological data (MM). Particularly, we compared the performances of different RSM formulations – relying on the use of Fy760 or sPRI as proxy for LUE and NDVI or MTCI as fraction of absorbed photosynthetically active radiation (fAPAR) – with those of classical MM. Results showed significantly higher GPP in the N fertilized experimental plots during the growing period. These differences in GPP disappeared in the drying period when senescence effects masked out potential differences due to plant N content. Consequently, although MTCI was tightly related to plant N content (r2 = 0.86, p < 0.01), it was poorly related to GPP (r2 = 0.45, p < 0.05). On the contrary sPRI and Fy760 correlated well with GPP during the whole measurement period. Results revealed that the relationship between GPP and Fy760 is not unique across treatments but it is affected by N availability. Results from a cross validation analysis showed that MM (AICcv = 127, MEcv = 0.879) outperformed RSM (AICcv = 140, MEcv = 0.8737) when soil moisture was used to constrain the seasonal dynamic of LUE. However, residual analyses demonstrated that MM is predictively inaccurate whenever no climatic variable explicitly reveals nutrient-related changes in the LUE parameter. These results put forward that RSM is a valuable means to diagnose nutrient-induced effects on the photosynthetic activity.

scholarly journals Asymmetric Effects of Daytime and Nighttime Warming on Boreal Forest Spring Phenology

2019 ◽  
Vol 11 (14) ◽  
pp. 1651 ◽  
Author(s):  
Guorong Deng ◽  
Hongyan Zhang ◽  
Xiaoyi Guo ◽  
Yu Shan ◽  
Hong Ying ◽  
...  
Keyword(s):  
Boreal Forest ◽  
Vegetation Index ◽  
Terrestrial Ecosystem ◽  
Terrestrial Ecosystems ◽  
Maximum Temperature ◽  
Spring Phenology ◽  
Asymmetric Effects ◽  
Terrestrial Ecosystem Model ◽  
Diurnal Temperatures ◽  
Nighttime Warming

Vegetation phenology is the most intuitive and sensitive biological indicator of environmental conditions, and the start of the season (SOS) can reflect the rapid response of terrestrial ecosystems to climate change. At present, the model based on mean temperature neglects the role of the daytime maximum temperature (TMAX) and the nighttime minimum temperature (TMIN) in providing temperature accumulation and cold conditions at leaf onset. This study analyzed the spatiotemporal variations of spring phenology for the boreal forest from 2001 to 2017 based on the moderate-resolution imaging spectro-radiometer (MODIS) enhanced vegetation index (EVI) data (MOD13A2) and investigated the asymmetric effects of daytime and nighttime warming on the boreal forest spring phenology during TMAX and TMIN preseason by partial correlation analysis. The results showed that the spring phenology was delayed with increasing latitude of the boreal forest. Approximately 91.37% of the region showed an advancing trend during the study period, with an average advancement rate of 3.38 ± 0.08 days/decade, and the change rates of different land cover types differed, especially in open shrubland. The length of the TMIN preseason was longer than that of the TMAX preseason and diurnal temperatures showed an asymmetrical increase during different preseasons. The daytime and nighttime warming effects on the boreal forest are asymmetrical. The TMAX has a greater impact on the vegetation spring phenology than TMIN as a whole and the effect also has seasonal differences; the TMAX mainly affects the SOS in spring, while TMIN has a greater impact in winter. The asymmetric effects of daytime and nighttime warming on the SOS in the boreal forest were highlighted in this study, and the results suggest that diurnal temperatures should be added to the forest terrestrial ecosystem model.

scholarly journals 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

2020 ◽  
Vol 12 (7) ◽  
pp. 1111
Author(s):  
Yun Gao ◽  
Songhan Wang ◽  
Kaiyu Guan ◽  
Aleksandra Wolanin ◽  
Liangzhi You ◽  
...  
Keyword(s):  
Chlorophyll Fluorescence ◽  
Crop Yield ◽  
Vegetation Index ◽  
Crop Yields ◽  
Growing Season ◽  
Supplementary Information ◽  
Climate Data ◽  
Modis Evi ◽  
Combining Data ◽  
The U.S

Satellite sun-induced chlorophyll fluorescence (SIF) has emerged as a promising tool for monitoring growing conditions and productivity of vegetation. However, it still remains unclear the ability of satellite SIF data to predict crop yields at the regional scale, comparing to widely used satellite vegetation index (VI), such as the Enhanced Vegetation Index (EVI) from the Moderate Resolution Imaging Spectroradiometer (MODIS). Additionally, few attempts have been made to verify if SIF products from the new Orbiting Carbon Observatory-2 (OCO-2) satellite could be applied for regional corn and soybean yield estimates. With the deep neural networks (DNN) approach, this study investigated the ability of OCO-2 SIF, MODIS EVI, and climate data to estimate county-level corn and soybean yields in the U.S. Corn Belt. Monthly mean and maximum SIF and MODIS EVI during the peak growing season showed similar correlations with corn and soybean yields. The DNNs with SIF as predictors were able to estimate corn and soybean yields well but performed poorer than MODIS EVI and climate variables-based DNNs. The performance of SIF and MODIS EVI-based DNNs varied with the areal dominance of crops while that of climate-based DNNs exhibited less spatial variability. SIF data could provide useful supplementary information to MODIS EVI and climatic variables for improving estimates of crop yields. MODIS EVI and climate predictors (e.g., VPD and temperature) during the peak growing season (from June to August) played important roles in predicting yields of corn and soybean in the Midwestern 12 states in the U.S. The results highlighted the benefit of combining data from both satellite and climate sources in crop yield estimation. Additionally, this study showed the potential of adding SIF in crop yield prediction despite the small improvement of model performances, which might result from the limitation of current available SIF products. The framework of this study could be applied to different regions and other types of crops to employ deep learning for crop yield forecasting by combining different types of remote sensing data (such as OCO-2 SIF and MODIS EVI) and climate data.

Sign in / Sign up

Export Citation Format

Share Document