USING FIELD AND SATELLITE NOVEL MEASURES TO IMPROVE CROP YIELD MODELING

• Sponsoring Institution: National Institute of Food and Agriculture
• Project Status: COMPLETE
• Funding Source: HATCH
• Reporting Frequency: Annual
• Accession No.: 1011541
• Project Start Date: Dec 1, 2016
• Project End Date: Sep 30, 2021
• Program Code: [(N/A)]- (N/A)

Recipient Organization
UNIVERSITY OF ILLINOIS
2001 S. Lincoln Ave.
URBANA,IL 61801

Performing Department
Natural Resources & Environmental Sciences

Non Technical Summary
Global food and biofuel demand keeps increasing with the growing population and economic developments, and meeting this rapidly rising demand under a warmer and changing climate poses one of the greatest challenges to humanity for the coming decades. The emerging space-borne sun-induced chlorophyll fluorescence (SIF) provides great promise to measure photosynthetic activity and crop productivity at a continental scale and in real time. SIF retrievals from the newly launched NASA Orbiting Carbon Observatory-2 (OCO-2), along with other SIF retrievals from existing satellites, provide an unprecedented opportunity to study the theoretical foundations between SIF and photosynthesis and to improve our capability of monitoring and forecasting global and regional food production. The overarching goal of the propose work is to develop a new satellite-based algorithm for measuring crop productivity, including Gross Primary Production (GPP), plant autotrophic respiration (Ra), Net Primary Production (NPP), and crop yield, using sun-induced fluorescence from the NASA OCO-2 satellite, and apply this to the U.S. Corn Belt. We will integrate field measurements, satellite data/algorithms, and process-based modeling to achieve this overarching goal.

Animal Health Component

50%

Research Effort Categories

Basic

50%

Applied

50%

Developmental

(N/A)

Classification

Knowledge Area (KA)	Subject of Investigation (SOI)	Field of Science (FOS)	Percent
205	2410	2020	100%

Knowledge Area
205 - Plant Management Systems;

Subject Of Investigation
2410 - Cross-commodity research--multiple crops;

Field Of Science
2020 - Engineering;

Keywords

crop productivity

gross primary production

plant autotrophic respiration

net primary production

crop yield

sun-induced fluorescence

Goals / Objectives
The overarching goal of the propose work is to develop a new satellite-based algorithm for measuring crop productivity, including Gross Primary Production (GPP), plant autotrophic respiration (Ra), Net Primary Production (NPP), and crop yield, using sun-induced fluorescence from the NASA OCO-2 satellite, and apply this to the U.S. Corn Belt. We will integrate field measurements, satellite data/algorithms, and process-based modeling to achieve this overarching goal. Specifically, our science questions and tasks include:(1) To set up long-term and continuous SIF measurements in the four eddy-covariance flux towers in the soybean and corn fields that are currently managed by the PI's collaborators (3 in the U.S. Corn Belt). We will use these SIF measurements coupled with eddy-covariance carbon flux data to study: What are observed relationships between long-term field-measured SIF, ETR and GPP across various environmental conditions, C3 and C4 crops, and different temporal scales (diurnal and seasonally)?(2) To study the SIF scaling processes from leaf, canopy, and landscape to satellite footprint by combining field-level measurements and a radiative transfer model that has been coupled with photosynthesis and fluorescence processes to answer the following questions: How much spatial and temporal variation of SIF is due to plant physiology, canopy structure, sun angles or view angles? How do we robustly use satellite SIF retrievals to derive crop productivity?(3) To utilize two independent approaches to estimate crop autotrophic respiration at the field and/or satellite scales, and develop a new model for crop autotrophic respiration using satellite-measurable co-variates to answer the following questions: What are the spatial and temporal variations of crop autotrophic respiration at the landscape and satellite scales? Can we use existing satellite and/or field-level data to model crop autotrophic respiration?(4) To use measured and derived datasets from the above tasks and data-model fusion techniques to improve the parameterizations in the crop modeling in an Earthy System Model (CESM) and to study: How can we use derived crop carbon budgets from SIF to improve the simulation of photosynthesis, plant phenology and heat stress responses in the crop module of the CESM model?

Project Methods
We will use the following methods to address the scientific goals raised before:Task 1: Conduct field measurements of SIF and explore the SIF:GPP relationship.Dr. Guan and his team will conduct long-term SIF field-level measurements at three sites across the Corn Belt. Two of these sites are located in the Energy Farm of the University ofIllinois to the south of Champaign, Illinois (managed by Dr. Carl Bernacchi from the University of Illinois), and one is located in Mead, Nebraska (US-NE3 site, managed by Dr. Andrew Suyker from the University of Nebraska at Lincoln). For field-level SIF measurements, we choose to use the FluoSpec2 system that was developed by the collaborator Dr. Xi Yang. To complement the canopy-level measures of SIF from the FluoSpec2, we will also conduct biweekly leaf-level measures at the two Champaign sites, using LICOR-6400 with the leaf chamber PAM fluorometry (LICOR, Inc., Lincoln, NE).Task 2: Link field and satellite data to develop new algorithms for estimating crop productivity.The scaling of SIF signals from leaf to satellite has to go through three major steps: (1) from leaf to canopy (vertical heterogeneity and sun-sensor geometry); (2) from canopy to landscape (spatial heterogeneity within the flux-tower footprint); and (3) from landscape to satellite (more spatial heterogeneity from the mixture of C3 and C4 crops and other land covers). To separate and quantify these effects, we will combine our field-level measurements and a radiative transfer model - SCOPE. To corroborate and calibrate the SCOPE model performance, we will use both leaf-level measures fromTask 1 and the tower-fixed FluoSpec2 measures. Once a robust relationship of SIF, ETR and GPP is built at the landscape scale, our final step is to scale this relationship up to the satellite level. We will use the NASA OCO-2 SIF retrieval combined with the USDA NASS Crop Data Layer to downscale the SIF to the field level.Task 3: Constrain plant autotrophic respiration using multiple approaches at the field, landscape and satellite scales.We will measure the plant respiration through two approaches: (1) inversely calculated from crop yield-based NPP and estimated GPP; and (2) inverse from eddy-covariance measures and soil respiration measures. We will use both data to formulate a parsimonious model of Ra. We will adopt a classic frameworkthat separates Ra according to growth respiration and maintenance respiration and use environmental conditions, observed biomass, and GPP as inputs to parameterize this model.Task 4: Improving crop modeling in the CLM-Crop using multiple SIF measures.Once the three tasks above are achieved, we plan to use our derived SIF-based cropland GPP, Ra, and NPP to improve the crop modeling in the CLM-Crop, the land model of the CESM. We will focus on the following aspects to improve the model performance: (1) photosynthesis rate at diurnal and seasonal scales (optimize Vcmax and leaf phenology); and (2) simulation of Ra and Carbon-Use-Efficiency. We will use both field-collected and satellite-derived data to constrain and improve the model performance, followed by a Bayesian framework to optimize the model performance and calibrate the critical model parameters.

Progress 12/01/16 to 09/30/21

Outputs
Target Audience:The target audience for our research is scientists investigating crop yield estimation at large scales and the application of novel satellite products in agriculture domains. Other audiences include farmers and members of agricultural industries. Changes/Problems: Nothing Reported What opportunities for training and professional development has the project provided?The graduates and postdocs working on this project presented their outcomes from this project during the AGU fall meeting. We provide training to undergraduate interns working in the SIF field measurements and also to graduate students who enrolled in the PI's class "Terrestrial Remote Sensing Applications" at the University of Illinois. How have the results been disseminated to communities of interest?The results have been disseminated through our conference presentations, journal articles, and training opportunities provided to the community. What do you plan to do during the next reporting period to accomplish the goals? Nothing Reported

Impacts
What was accomplished under these goals? Accomplishments Under Goal One: We maintained the continuous sun-induced chlorophyll fluorescence (SIF) measurement sites at the Energy Farm at the University of Illinois at Urbana-Champaignsince growing season 2016 and also at two other Mead sites in Nebraska since 2017. All of these sites are planted with maize, soybean, or Miscanthus. We are continuing collecting SIF data at three ARPA-E SMARTFARM sites in Illinois. Accomplishments Under Goal Two: We have evaluated the SCOPE model at our SIF measure sites and analyzed the key factor affecting the scaling from leaf to canopy scale. We analyzed the spatiotemporal and inter-/intra-landcover variabilities of SIF yield in the U.S. Midwest using the OCO-2 and TROPOMI footprint data. A downscaling algorithm to disaggregate the TROPOMI SIF into crop-specific SIF has been developed and tested over the U.S. Midwest. We compiled near-surface sensing data from three agricultural sites in the U.S. Corn Belt. We found the near-infrared radiance of vegetation (NIRv,Rad), the product of NIR radiance and normalized difference vegetation index (NDVI), can accurately estimate corn and soybean GPP at daily and half-hourly time scales, benchmarked with multi-year tower GPP at three sites with different environmental and irrigation conditions. We investigated the timing of seasonal peaks of SIF and GPP in soybean fields by integrating tower data, satellite data, and process-based SCOPE model simulations. We found inconsistency between the seasonal peak timing of SIF and GPP in three of four soybean fields based on tower SIF and eddy-covariance measurements, which can be explained by a divergence in the seasonality between absorbed photosynthetic active radiation and canopy chlorophyll content. We found considerable differences in the SIF-GPP relationships between corn and miscanthus, with a stronger SIF-GPP relationship and higher slope of SIF-GPP observed in corn compared to miscanthus. These differences were mainly caused by leaf physiology. For miscanthus, high non-photochemical quenching (NPQ) under high light, temperature and water vapor deficit (VPD) conditions caused a large decline of fluorescence yield (ΦF), which further led to a strong SIF midday depression and weakened the SIF-GPP relationship. The larger slope in corn than miscanthus was mainly due to the higher GPP in middle summer, largely attributed to the higher leaf photosynthesis caused by less biochemical limitations Accomplishments Under Goal Three: We finished a data-driven model for simulating cropland carbon budget (Ra, Rh, NEE, and yield) across the U.S. Corn Belt. The new carbon balance model integrates advanced process-based model ecosys and satellite observations with deep learning techniques. Evaluation results show the data-driven model has good performance. Accomplishments Under Goal Four: We have implemented a new maize growth simulation scheme in CESM by integrating the strengths of both the CLM and APSIM models. We finished evaluation of this new model (CLM-ASPSIM) at fluxnet tower level, parameter sensitivity analysis and regional simulation over the U.S. Corn Belt. We have been developing the CLM-AgSys framework by expanding the CLM-APSIM maize model to other crop types through collaborating with NCAR team.

Publications

• Type: Journal Articles Status: Under Review Year Published: 2022 Citation: Genghong Wu, Kaiyu Guan, Chongya Jiang, Hyungsuk Kimm, Guofang Miao, Carl J. Bernacchi, Caitlin E. Moore, Elizabeth A. Ainsworth, Xi Yang, Joseph A. Berry, Christian Frankenberg and Min Chen. 2022. Attributing differences of SIF-GPP relationships between two C4 crops: corn and miscanthus. New Phytologist.
• Type: Journal Articles Status: Under Review Year Published: 2022 Citation: Genghong Wu, Chongya Jiang, Hyungsuk Kimm, Sheng Wang, Carl Bernacchi, Caitlin E. Moore, Andy Suyker, Xi Yang, Troy Magney, Christian Frankenberg, Youngryel Ryu, Benjamin Dechant and Kaiyu Guan. 2022. Difference in seasonal peak timing of soybean SIF and GPP explained by canopy structure and chlorophyll content. Remote Sensing of Environment.
• Type: Journal Articles Status: Published Year Published: 2020 Citation: Wu, G., Guan, K., Jiang, C., Peng, B., Kimm, H., Chen, M., Yang, X., Wang, S., Sukyer, A.E., Bernacchi, C., Moore, C.E., Zeng, Y., Berry, J. and Cendrero-Mateo, P. 2020. Radiance-based NIRv as a proxy for GPP of corn and soybean. Environmental Research Letters, 15, 034009.
• Type: Journal Articles Status: Published Year Published: 2020 Citation: Wang, C., Guan, K., Peng, B., Chen, M., Jiang, C., Zeng, Y., Wu, G., Wang, S., Wu, J., Yang, X., Frankenberg, C., K�hler, P., Berry, J., Bernacchi, C., Zhu, K., Alden, C. and Miao, G. 2020. Satellite footprint data from OCO-2 and TROPOMI reveal significant spatio-temporal and inter-vegetation type variabilities of solar-induced fluorescence yield in the U.S. Midwest. Remote Sensing of Environment, 241, 111728.
• Type: Journal Articles Status: Published Year Published: 2020 Citation: Peng, B., Guan, K., Zhou, W., Jiang, C., Frankenberg, C., Sun, Y., He, L. and K�hler, P. 2020. Assessing the benefit of satellite-based solar-induced chlorophyll fluorescence in crop yield prediction. International Journal of Applied Earth Observation and Geoinformation, 90, 102126.

Progress 10/01/19 to 09/30/20

Outputs
Target Audience:The target audience for our research is scientists investigating crop yield estimation at large scales and the applicationof novel satellite products in agriculture domains. Other audiences include farmers and members of agricultural industries. Changes/Problems: Nothing Reported What opportunities for training and professional development has the project provided?The postdocs and graduate students working on this project presented their outcome from this project during the AGU fall meeting. We provide training to undergraduate interns working on the SIF field measurements and also to graduate students who enrolled in the PI's class "Terrestrial Remote Sensing Applications" at the University of Illinois. How have the results been disseminated to communities of interest?The results have been disseminated through our conference presentations, journal articles, and training opportunities provided to the community. What do you plan to do during the next reporting period to accomplish the goals?We will continue to put efforts into maintaining the SIF measuring towers in Illinoisand use both site and satellite-based SIF data for crop monitoring at varied scales. We will also use satellite-based observations to constrain process-based models.

Impacts
What was accomplished under these goals? This year we have been making good progress of using solar-induced chlorophyll fluorescence (SIF) to monitor crop productivity at both ground and satellite scales. Main achievements are summarized as follows: 1. Radiance-based NIRV as a proxy for GPP of corn and soybean We compiled near-surface sensing data from three agricultural sites in the U.S. Corn Belt. We found the near-infrared radiance of vegetation (NIRv,Rad), the product of NIR radiance and normalized difference vegetation index (NDVI), can accurately estimate corn and soybean GPP at daily and half-hourly time scales, benchmarked with multi-year tower GPP at three sites with different environmental and irrigation conditions. Overall, NIRv,Rad explains 84% and 78% variations of half-hourly GPP for corn and soybean, respectively, outperforming near-infrared reflectance of vegetation (NIRv,Ref), enhanced vegetation index (EVI), and solar-induced fluorescence at 760 nm (SIF760). The strong linear relationship between NIRv,Rad and absorbed photosynthetically active radiation by green leaves (APARgreen), and that between APARgreen and GPP, explain the good NIRv,Rad:GPP relationship. The NIRv,Rad:GPP relationship is robust and consistent across sites. The scalability and simplicity of NIRv,Rad indicate great potential to estimate daily or sub-daily crop GPP from high-resolution or long-term remote sensing data. Publication: Wu, G., Guan, K., Jiang, C., Peng, B., Kimm, H., Chen, M., Yang, X., Wang, S., Sukyer, A.E., Bernacchi, C., Moore, C.E., Zeng, Y., Berry, J. andCendrero-Mateo, P. 2020. Radiance-based NIRv as a proxy for GPP of corn and soybean. Environmental Research Letters, 15, 034009. 2. Satellite footprint data from OCO-2 and TROPOMI reveal significant spatio-temporal and inter-vegetation type variabilities of solar-induced fluorescence yield in the U.S. Midwest We utilized recent SIF data with high spatial resolution from two satellite instruments, OCO-2 and TROPOMI, together with multiple other datasets. We estimate apparent SIF yield across space, time, and different vegetation types in the U.S. Midwest during crop growing season (May to September) from 2015-2018. We find that apparent SIF yield of croplands is larger than noncroplands during peak season (July-August). However, apparent SIF yield between corn (C4 crop) and soybean (C3 crop) does not show a significant difference. Apparent SIF yield of corn, soybean, forest, and grass/pasture show clear seasonal and spatial patterns. The spatial variability of precipitation during the growing season can explain the overall spatial pattern of apparent SIF yield. Further analysis by decomposing apparent SIF yield into ΦF and fesc using near-infrared reflectance of vegetation (NIRV) suggests that fesc may be the major driver of the observed variabilities of apparent SIF yield. Publication: Wang, C., Guan, K., et al. 2019. Satellite footprint data from OCO-2 and TROPOMI reveal significant spatiotemporal and inter-vegetation type variabilities of solar-induced fluorescence yield in the U.S. Midwest. 3.Assessing the benefit of satellite-based solar-Induced chlorophyll fluorescence in crop yield prediction We assessed the benefits of using two satellite-based SIF products with higher resolution in yield prediction for maize and soybean in the U.S. Midwest: One is the gap-filled SIF from Orbiting Carbon Observatory 2 (OCO-2) and the other one is new SIF retrievals from the TROPOspheric Monitoring Instrument (TROPOMI). The yield prediction performances of using SIF data were benchmarked with those using satellite-based vegetation indices (VIs), including normalized difference vegetation index (NDVI), enhanced vegetation index (EVI), near-infrared reflectance of vegetation (NIRv), and land surface temperature (LST). Five machine-learning algorithms were used to build yield prediction models with both remote-sensing-only and climate-remote-sensing-combined variables. We found that using SIF products gave the best forward predictions for both maize and soybean yields in 2018, indicating great potential of using satellite-based SIF products for crop yield prediction. However, using currently available high-resolution SIF products did not guarantee consistently better yield prediction performances than using other satellite-based remote sensing variables in all the evaluated cases. The relative performances of using different remote sensing variables in yield prediction depended on crop types (maize or soybean), out-of-sample testing methods (five-fold-cross-validation or forward), and record length of training data. We also found that using NIRv could generally lead to better yield prediction performance than using NDVI, EVI, or LST, and using NIRv could achieve similar or even better yield prediction performance than using the SIF products. We concluded that satellite-based SIF products could be beneficial in crop yield prediction with more high-resolution and good-quality SIF products accumulated in the future. Publication:Peng, B., Guan, K., Zhou, W., Jiang, C., Frankenberg, C., Sun, Y., He, L. andKöhler, P. 2020. Assessing the benefit of satellite-based solar-induced chlorophyll fluorescence in crop yield prediction. International Journal of Applied Earth Observation and Geoinformation, 90, 102126. 4. Develop a high-resolution, land cover-specific product of solar-induced fluorescence downscaled from TROPOMI footprint data for the U.S. Midwest We proposeda methodology to spatially downscale the TROPOMI footprint SIF data by using a semi-empirical model expressing the SIF of a mixed pixel as the sum of SIF of each vegetation type weighted by the land cover fraction. SIF of a specific vegetation type is expressed as the product of photosynthetically active radiation (PAR), the fraction of PAR absorbed by the canopy (fPAR), and canopy SIF yield (SIFyield). The approach estimates each component in the SIF model using TROPOMI footprint SIF and some other datasets. It then assembles these variables to reconstruct the SIF value. We generated a new SIF dataset by applying this methodology. The dataset covers the U.S. Midwest during the crop growing season (May-September) with a spatial resolution of fourkm and a temporal frequency of eightdays in 2018. The new SIF product provides not only SIF of a mixed pixel but also SIF, the land cover fraction, PAR, fPAR, and SIFyield of four dominated vegetation types in the area (i.e., corn, soybean, grass/pasture, forest). Validation with an independent TROPOMI SIF dataset (R2=0.79), OCO-2 SIF dataset (R2=0.75), and ground measurements from three field sites (R2=0.85-0.93) showed a good agreement between the reconstructed SIF and the independent observations. This new SIF dataset provides unique values in vegetation productivity estimating and vegetation stress detecting. Publication: Wang, C. et al.2021. A high-resolution, land cover-specific product of solar-induced fluorescence downscaled from TROPOMI footprint data for the U.S. Midwest. Under review at Earth System Science Data.

Publications

• Type: Journal Articles Status: Published Year Published: 2020 Citation: Wu, G., Guan, K., Jiang, C., Peng, B., Kimm, H., Chen, M., Yang, X., Wang, S., Sukyer, A.E., Bernacchi, C., Moore, C.E., Zeng, Y., Berry, J. and Cendrero-Mateo, P. 2020. Radiance-based NIRv as a proxy for GPP of corn and soybean. Environmental Research Letters, 15, 034009.
• Type: Journal Articles Status: Published Year Published: 2020 Citation: Wang, C., Guan, K., Peng, B., Chen, M., Jiang, C., Zeng, Y., Wu, G., Wang, S., Wu, J., Yang, X., Frankenberg, C., K�hler, P., Berry, J., Bernacchi, C., Zhu, K., Alden, C. and Miao, G. 2020. Satellite footprint data from OCO-2 and TROPOMI reveal significant spatio-temporal and inter-vegetation type variabilities of solar-induced fluorescence yield in the U.S. Midwest. Remote Sensing of Environment, 241, 111728.
• Type: Journal Articles Status: Published Year Published: 2020 Citation: Peng, B., Guan, K., Zhou, W., Jiang, C., Frankenberg, C., Sun, Y., He, L. and K�hler, P. 2020. Assessing the benefit of satellite-based solar-induced chlorophyll fluorescence in crop yield prediction. International Journal of Applied Earth Observation and Geoinformation, 90, 102126.
• Type: Journal Articles Status: Under Review Year Published: 2021 Citation: Wang, C., Guan, K., Peng, B., Jiang, C., Peng, J., Wu, G., Frankenberg, C., K�hler, P., Yang, X., Cai, Y. and Huang, Y. 2021. A high-resolution, land cover-specific product of solar-induced fluorescence downscaled from TROPOMI footprint data for the U.S. Midwest. Under review at Earth System Science Data.

Progress 10/01/18 to 09/30/19

Outputs
Target Audience:The target audience for our research is scientists investigating crop yield estimation at large scales and the applications of novel satellite products in agriculture domains. Other audiences include farmers and members of agricultural industries. Changes/Problems: Nothing Reported What opportunities for training and professional development has the project provided?The postdocs working on this project presented their outcome from this project during the AGU fall meeting. We provide trainning to undergraduate interns working on the SIF field measurements and also to graduate students who enrolled in the PI's class "Terrestrial remote sensing applications" at the University of Illinois. How have the results been disseminated to communities of interest?The results have been disseminated through our conference presentations, journal articles, and training opportunities provided to the community. What do you plan to do during the next reporting period to accomplish the goals?We will continue to put efforts into maintaining the SIF measuring towers in Illinois and Nebraska and use both site and satellite-based SIF data for crop monitoring at varied scales.

Impacts
What was accomplished under these goals? This year we have been making good progress of using solar-induced chlorophyll fluorescence (SIF) to monitor crop productivity at both ground and satellite scales. Main achievements are summarized as follows: Radiance-based NIRV as a proxy for GPP of corn and soybean We compiled near-surface sensing data from three agricultural sites in the U.S. Corn Belt. One rainfed site was located at the Energy Farm of the University of Illinois at Urbana-Champaign (UIUC, 40.0628 N, 88.1959 W). Another two sites were located at the Eastern Nebraska Research and Extension Center of University of Nebraska-Lincoln, with one irrigated (UNL irrigated, 41.1649 N, 96.4701 W) and one rainfed (UNL rainfed, 41.1797 N, 96.4397 W) site. The mean annual temperature and precipitation of 1990-2018 were (11.5 °C, 1036 mm) and (10.1 °C, 770 mm) at UIUC(Willard Airport weather station) and two UNL sites (National Climate Data Center, Nebraska, Mead 6S weather station), respectively. The UIUC site had a corn-corn-soybean rotation, whereas the two UNL sites were corn-soybean rotation. The growing season (from planting to harvesting) was generally May-October for both crops across all the three sites. During 2016-2018, a total of four and three site-year observations were made for corn and soybean, respectively. Fluospec2 systems (Miao et al., 2018;Yang et al., 2018) were installed to acquire vegetation indices and solar-induced chlorophyll fluorescence (SIF), paired with EC systems for measuring GPP. Downwelling and upwelling photosynthetic active radiation (PAR) were measured above and below canopy by multiple point or line quantum sensors (LI-COR Inc., USA) at the two UNL sites, from which the fraction of absorbed PAR (FPAR) were derived at half-hourly intervalwe found the near-infrared radiance of vegetation (NIRv,Rad), the product of NIR radiance and normalized difference vegetation index (NDVI), can accurately estimate corn and soybean GPP at daily and half-hourly time scales, benchmarked with multi-year tower GPP at three sites with different environmental and irrigation conditions. Overall, NIRv,Rad explains 84% and 78% variations of half-hourly GPP for corn and soybean, respectively, outperforming near-infrared reflectance of vegetation (NIRv,Ref), enhanced vegetation index (EVI), and solar-induced fluorescence at 760 nm (SIF760). The strong linear relationship between NIRv,Rad and absorbed photosynthetically active radiation by green leaves (APARgreen), and that between APARgreen and GPP, explain the good NIRv,Rad:GPP relationship. The NIRv,Rad:GPP relationship is robust and consistent across sites. The scalability and simplicity of NIRv,Rad indicate great potential to estimate daily or sub-daily crop GPP from high-resolution or long-term remote sensing data. Publication: Wu, G., Guan, K., et al (2019). Radiance-based NIRV as a proxy for GPP of corn and soybean. Under review at Environmental Research Letters. Satellite footprint data from OCO-2 and TROPOMI reveal significant spatio-temporal and inter-vegetation type variabilities of solar-induced fluorescence yield in the U.S. Midwest Solar-induced chlorophyll fluorescence (SIF) measured from space has been increasingly used to quantify plant photosynthesis at regional and global scales. Apparent canopy SIF yield, determined by fluorescence yield (ΦF) and escaping ratio (fesc), together with absorbed photosynthetically active radiation (APAR), is crucial in driving spatio-temporal variability of SIF. While strong linkages between apparent SIF yield and plant physiological responses and canopy structure have been suggested, spatio-temporal variability of apparent SIF yield at regional scale remains largely unclear, which limits our understanding of the spatio temporal variability of SIF and its relationship with photosynthesis. In this study, we utilize recent SIF data with high spatial resolution from two satellite instruments, OCO-2 and TROPOMI, together with multiple other datasets. We estimate apparent SIF yield across space, time, and different vegetation types in the U.S. Midwest during crop growing season (May to September) from 2015-2018. We find that apparent SIF yield of croplands is larger than non-croplands during peak season (July-August). However, apparent SIF yield between corn (C4 crop) and soybean (C3 crop) does not show a significant difference. Apparent SIF yield of corn, soybean, forest, and grass/pasture show clear seasonal and spatial patterns. The spatial variability of precipitation during the growing season can explain the overall spatial pattern of apparent SIF yield. Further analysis by decomposing apparent SIF yield into ΦF and fesc using near-infrared reflectance of vegetation (NIRV) suggests that fesc may be the major driver of the observed variabilities of apparent SIF yield. Publication: Wang, C., Guan, K., et al. (2019). Satellite footprint data from OCO-2 and TROPOMI reveal significant spatio-temporal and inter-vegetation type variabilities of solar-induced fluorescence yield in the U.S. Midwest. Under review at Remote Sensing of Environment. Assessing the benefit of satellite-based Solar-Induced Chlorophyll Fluorescence in crop yield prediction Large-scale crop yield prediction is critical for early warning of food insecurity, agricultural supply chain management, and economic markets. Satellite-based Solar-Induced Chlorophyll Fluorescence (SIF) products have revealed hot spots of photosynthesis over global croplands, such as in the U.S. Midwest. However, to what extent these satellite-based SIF products can enhance the performance of crop yield prediction when benchmarking against other existing satellite data remains unclear. Here we assessed the benefits of using two satellite-based SIF products with higher resolution in yield prediction for maize and soybean in the U.S. Midwest: one is the gap-filled SIF from Orbiting Carbon Observatory 2 (OCO-2) and the other one is new SIF retrievals from the TROPOspheric Monitoring Instrument (TROPOMI). The yield prediction performances of using SIF data were benchmarked with those using satellite-based vegetation indices (VIs), including normalized difference vegetation index (NDVI), enhanced vegetation index (EVI), near-infrared reflectance of vegetation (NIRv), and land surface temperature (LST). Five machine-learning algorithms were used to build yield prediction models with both remote-sensing-only and climate-remote-sensing-combined variables. We found that using SIF products gave the best forward predictions for both maize and soybean yields in 2018, indicating great potential of using satellite-based SIF products for crop yield prediction. However, using currently available high-resolution SIF products did not guarantee consistently better yield prediction performances than using other satellite-based remote sensing variables in all the evaluated cases. The relative performances of using different remote sensing variables in yield prediction depended on crop types (maize or soybean), out-of-sample testing methods (five-fold-cross-validation or forward), and record length of training data. We also found that using NIRv could generally lead to better yield prediction performance than using NDVI, EVI, or LST, and using NIRv could achieve similar or even better yield prediction performance than using the SIF products. We concluded that satellite-based SIF products could be beneficial in crop yield prediction with more high-resolution and good-quality SIF products accumulated in the future. Publication: Peng, B., Guan, K., et al. (Under review). Assessing the benefit of satellite-based Solar-Induced Chlorophyll Fluorescence in crop yield prediction Agricultural and forest meteorology.

Publications

• Type: Journal Articles Status: Under Review Year Published: 2019 Citation: Wu, G., Guan, K., Jiang, C., Peng, B., Kimm, H., Chen, M., Yang, X., Wang, S., Suyker, S., Bernacchi, C., Moore, C., Zeng, Y., Berry, J. and Cendrero-Mateo, M. P. 2019. Radiance-based NIRV as a proxy for GPP of corn and soybean. Under review at Environmental Research Letters.
• Type: Journal Articles Status: Under Review Year Published: 2019 Citation: Wang, C., Guan, K., Peng, B., Chen, M., Jiang, C., Zeng, Y., Wang, S., Wu, J., Yang, X.,Frankenberg, C., K�hler, P., Berry, J., Bernacchi, B., Zhu, K., Alden, C. and Miao, G. 2019. Satellite footprint data from OCO-2 and TROPOMI reveal significant spatio-temporal and inter-vegetation type variabilities of solar-induced fluorescence yield in the U.S. Midwest. Under review at Remote Sensing of Environment.
• Type: Journal Articles Status: Under Review Year Published: 2019 Citation: Peng, B., Guan, K., Zhou, W., Jiang, C., Frankenberg, C., Sun, Y., He, L. and K�hler, P. 2019. Assessing the benefit of satellite-based solar-induced chlorophyll fluorescence in crop yield prediction agricultural and forest meteorology.

Progress 10/01/17 to 09/30/18

Outputs
Target Audience:The target audience for our research is scientists investigating crop yield estimation at large scales and the applications of novel satellite products in agriculture domains. Other audiences include farmers and members of agricultural industries. Changes/Problems: Nothing Reported What opportunities for training and professional development has the project provided?The postdocs working on this projectpresented their outcome from this project during the AGU fall meeting. We provide trainning to undergraduate interns working on the SIF field measurements and also to graduate students who enrolled in the PI's class "Terrestrial remote sensing applications" at the University of Illinois. How have the results been disseminated to communities of interest?The results have been disseminated through our conference presentations, journal articles, and training opportunities provided to the community. What do you plan to do during the next reporting period to accomplish the goals?We will continue to put efforts inmaintaining the SIF measuring towers in Illinois and Nebraska in order to collect more data for both soybean and maize fields. We will use the SCOPE radiative transfer model to simulate canopy scale SIF which encounters the effect from canopy structures and viewing geometry. We will finish developing the CLM-AgSys-Soybean model and run regional simulations for both maize and soybean across the CONUS.

Impacts
What was accomplished under these goals? Accomplishment under Goal One: We maintained the continuous sun-induced chlorophyll fluorescence (SIF) measurement sites at the Energy Farm at the University of Illinois at Urbana-Champaign, Illinois since growing season 2016 and also at two other Mead sites in Nebraskasince 2017. In the growing season of 2018, one site (rainfed) is in Illinois and two sites (one rainfed and one irrigated) in Nebraska. All of these three sites are planted with maize. Accomplishment under Goal Two: We analyzed the spatiotemporal and inter-/intra-landcover variabilities of SIF yield in the U.S. Midwest using the OCO-2 footprint data. A downscaling algorithm to disaggregate the GOME-2 SIF into crop-specific SIF has been developed and tested over the U.S. Midwest. A more advanced downscaling approach leveraging GOME-2, OCO-2 and TROPOMI SIF data is under development. Accomplishment under Goal Three: We are developing the data-driven model for simulating the plant autotrophic respiration across the U.S. Corn Belt. Accomplishment under Goal Four: Previously, we have implemented a new Maize growth simulation scheme in CESM by integrating the strengths of both the CLM and APSIM models, i.e. CLM-AgSys-Maize. We finished global parameter sensitivity analysis of the CLM-AgSys-Maize model in this year. Regional simulation over the U.S. Corn Belt has also been finished. We are currently implementing the soybean model under the framework of CLM-AgSys.

Publications

• Type: Journal Articles Status: Published Year Published: 2018 Citation: Cai, Y., Guan, K., Peng, J., Wang, S., Seifert, C., Wardlow, B. and Li, Z. 2018. A high-performance and in-season classification system of field-level crop types using time-series Landsat data and a machine learning approach. Remote Sensing of Environment, 210, 35-47.
• Type: Journal Articles Status: Published Year Published: 2018 Citation: Miao, G., Guan, K., Yang, X., Bernacchi, C.J., Berry, J.A., DeLucia, E.H., Wu, J., Moore, C.E., Meacham, K., Cai, Y., Peng, B., Kimm, H. and Masters, M.D. 2018. Sun-induced chlorophyll fluorescence, photosynthesis, and light use efficiency of a soybean field. Journal of Geophysical Research: Biogeosciences, 123, 610-623.
• Type: Conference Papers and Presentations Status: Published Year Published: 2018 Citation: Cai, Y., Guan, K., Lobell, D., Potgieter, A., Wang, S., Peng, J., Xu, T., Asseng, S., Zhang,Y., You, L. and Peng, B. 2018. Integrating satellite and climate data to predict wheat yield in Australia using machine learning approaches. American Geophysical Union, Fall General Assembly 2018, Washington, D.C.
• Type: Conference Papers and Presentations Status: Published Year Published: 2018 Citation: Kimm, H., Guan, K., Wu, G.,Peng, B., Burroughs, C., Kumagai, E., Bernacchi, C. and Ainsworth, E. 2018. Can sun-induced chlorophyll fluorescence (SIF) detect heat-stress of soybean? Evidence from field measurement utilizing a portable SIF system at the Temperature Free-Air Controlled Enhancement (T-FACE) experiment. American Geophysical Union, Fall General Assembly 2018, Washington, D.C.
• Type: Conference Papers and Presentations Status: Published Year Published: 2018 Citation: Peng, B., Guan, K., Chen, M., Lawrence, D., Pokhrel, Y. and Lombardozzi, D. 2018. Representing agricultural systems in Earth system model: Implementation, calibration, and multi-scale validation of CLM-AgSys. American Geophysical Union, Fall General Assembly, Washington, D.C.
• Type: Conference Papers and Presentations Status: Published Year Published: 2018 Citation: Peng, B., Guan, K. et al. 2018. CLM-AgSys: Simulating crop growth and managementin the Earth system models. 3rd Annual Crops in Silico Symposium, Urbana, Illinois.
• Type: Journal Articles Status: Under Review Year Published: 2019 Citation: Guofang Miao, Kaiyu Guan, Andrew E. Suyker, Xi Yang, Timothy J. Arkebauer, Elizabeth A. Walter-Shea, Hyungsuk Kimm, Gabriel Y. Hmimina, John A. Gamon, Trenton E. Franz, Christian Frankenberg, Joseph A. Berry and Genghong Wu. 2019. Varying contributions of drivers on the relationship between canopy photosynthesis and far-red sun-induced fluorescence for two maize sites at different temporal scales (Submitted).
• Type: Conference Papers and Presentations Status: Published Year Published: 2018 Citation: Guofang Miao, Kaiyu Guan, Andrew Suyker, Zeng Yelu, Xi Yang, Genghong Wu, Youngryel Ryu, Benjamin Dechant, Timothy J Arkebauer, Elizabeth Anne Walter-Shea, John Arthur Gamon, Gabriel Hmimina, Tom Avenson, Ryan Moore and Hyungsuk Kimm. 2018. Structural and physiological effects on the relationship between solar-induced fluorescence and gross primary production: A comparison study between nadir and hemispherical fluorescence observations. American Geophysical Union, Fall General Assembly 2018, Washington, D.C.

Progress 12/01/16 to 09/30/17

Outputs
Target Audience:The target audience for our research is scientists investigating crop yield estimation at large scales and the applications of novel satellite products in agriculture domains. Other audiences include farmersand members of agriculturalindustries. Changes/Problems: Nothing Reported What opportunities for training and professional development has the project provided?The postdoc working onthis project gave a presentation during the SIF workship in Nebraska and presentedthe outcome from this project during the AGU fall meeting. We provide trainning to undergraduate interns working on the SIF field measurements and also to graduate students who enrolled in the PI's class "Terrestrial remote sensing applications" at the University of Illinois. How have the results been disseminated to communities of interest?The results have been disseminated through our conference presentations, journal articles, and training opportunities provided to the community. What do you plan to do during the next reporting period to accomplish the goals?We will continue to maintain the SIF measuring towers in Illinois and Nebraska to accumulate more data for both soybean and maize. We will use a new radiative transfer model to simulate canopy scale SIF which encounters the effect from canopy structures and viewing geometry. We will finish developing an algorithm to estimate plant autotrophic respiration using satellite data and Bayesian parameter calibration of the new crop model in CESM.

Impacts
What was accomplished under these goals? Accomplishment under Goal One:Continuous sun-induced chlorophyll fluorescence (SIF) measurements were conducted at the soybean and maizeplots at the Energy Farm of the University of Illinois at Urbana and Champaign, USA (40.065791 N, 88.208387 W, and about 220 meters above sea level) since growing season 2016 and also at two another Mead sites in Nebraska since 2017. Accomplishment under Goal Two: We have evaluated the SCOPE model at our SIF measure sites and analyzed the key factor affecting the scaling from leaf to canopy scale. We also developed adownscaling algorithm to disaggregate the coarse resolution SIF products from GOME-2 by leveraging the benefits from the high-resolution OCO-2. Accomplishment under Goal Three: We are still exploring the most important variables in modeling the plant autotrophic respiration. Accomplishment under Goal Four: We have implemented a new Maize growth simulation scheme in CESM by integrating the strengths of both the CLM and APSIM models. We finished evaluation of this new model (CLM-ASPSIM) at fluxnet tower level and we are currently doing work related to parameter sensitivity analysis and regional simulation over the U.S. Corn Belt.

Publications

• Type: Conference Papers and Presentations Status: Published Year Published: 2016 Citation: Miao, G., Guan, K., Yang, X., Bernacchi, C., DeLucia, E. H., Cai, Y., ... and Peng, B. 2016. Continuous Measurements of Canopy-level Solar-Induced Chlorophyll Fluorescence for Inferring Diurnal and Seasonal Dynamics of Photosynthesis in Crop Fields in the Midwestern USA. In AGU Fall Meeting Abstract.
• Type: Journal Articles Status: Under Review Year Published: 2018 Citation: Miao, G., Guan, K. and Yang, Xi, et al. 2018. Sun-Induced Chlorophyll Fluorescence, Photosynthesis, and Light Use Efficiency of a Soybean Field. Journal of Geophysical Research-Biogeoscience.
• Type: Journal Articles Status: Published Year Published: 2017 Citation: Guan, K., Wu, J., Anderson, M.C., Kimball, J., Frolking, S., Li, B. and Lobell, D. 2017. The Shared and Unique Value of Optical, Flourescence, Thermal and Microwave Satellite Data for Estimating Large Scale Crop Yield. Remote Sensing of Environment.
• Type: Conference Papers and Presentations Status: Published Year Published: 2017 Citation: Guan, K., Peng, B., Chen, M., Lawrence, D., Pokhrel, Y. and Lu, Y. 2017. Improving the Maize Growth Processes in the Community Land Model: Implementation and Evaluation. In NCAR CESM Workshop on Joint Societal Dimensions and Land Model Working Groups.
• Type: Conference Papers and Presentations Status: Published Year Published: 2016 Citation: Peng, B., Guan, K. and Chen, M. 2016. Parsing Multiple Processes of High Temperature Impact on Corn Yield Using a Newly Developed CLM-APSIM Modeling Framework. In American Geophysical Union, Fall Meeting 2016.
• Type: Journal Articles Status: Published Year Published: 2018 Citation: Peng, B., Guan, K., Chen, M., Lawrence, D.M., Pokhrel, Y., Suyker, A., Arkebauer, T. and Lu, Y. 2018. Improving Maize Growth Processes in the Community Land Model: Implementation and Evaluation. Agricultural and Forest Meteorology, 250251, 64-89.