Source: DAGAN, INC. submitted to NRP
COVER CROP INFORMATION SYSTEM: USING REMOTE SENSING AND MODELING TO MAP DETAILED INFORMATION ABOUT COVER CROPS ACROSS WIDE REGIONS
Sponsoring Institution
National Institute of Food and Agriculture
Project Status
COMPLETE
Funding Source
Reporting Frequency
Annual
Accession No.
1024310
Grant No.
2020-33530-33099
Cumulative Award Amt.
$600,000.00
Proposal No.
2020-06700
Multistate No.
(N/A)
Project Start Date
Sep 1, 2020
Project End Date
Aug 31, 2022
Grant Year
2020
Program Code
[8.4]- Air, Water and Soils
Recipient Organization
DAGAN, INC.
15 NEWMARKET RD
DURHAM,NH 038242815
Performing Department
(N/A)
Non Technical Summary
Decisions regarding the implementation of conservation and soil health practices in agricultural areas can have a significant effect on productivity and environmental outcomes, including soil erosion, water quality, and carbon sequestration. In addition, the effects of management can vary due to soil type and topographic conditions. While Dagan's OpTIS system produces wide area maps of the presence of cover crops, there is currently no systematic and cost-effective method for documenting winter cover crop quality, or the resulting effects of these cover crops, over a large region. During Phase I activities, we demonstrated the feasibility of using multi-sensor satellite observations and biogeochemical modeling for operational cover crop monitoring that will systematically provide detailed information about the spatial and temporal dynamics of cover crop adoption and vigor. In Phase II, we propose to operationalize these algorithms to facilitate deployment of the system over wide areas, back through time. This monitoring system will be tied to our existing Operational Tillage Information System (OpTIS) to provide the enhanced, detailed information that our data service clients require for sustainability reporting, ecosystem service market participation, and scenario planning.Under Phase II activities our goal is to bring the research demonstrated under Phase I activities into a fully functional and largely automated prototype system and evaluate the products in four demonstration regions. This goal will be accomplished through five technical objectives.This project targets five primary decision-makers as users of a cover crop information system: ag non-profit organizations (e.g. The Nature Conservancy), the developing ecosystem service markets (e.g. ESMC), watershed- and state-level water quality programs, the US Dept. of Agriculture NRCS and National Agriculture Statistics Service [NASS], and private corporations (e.g. General Mills and Bayer). Each of these organizations need accurate, timely, and spatially comprehensive information about the dynamic state of cover crops across large regions.
Animal Health Component
45%
Research Effort Categories
Basic
10%
Applied
45%
Developmental
45%
Classification

Knowledge Area (KA)Subject of Investigation (SOI)Field of Science (FOS)Percent
1020199208030%
1020199209030%
1020199106040%
Goals / Objectives
Primary Goal: Bring the research demonstrated under Phase I activities into a fully functional and largely automated prototype system, and we will evaluate the products for four demonstration regions.This goal will be accomplished through five technical objectives.Algorithm refinement objectivesTechnical Objective1: Refine and implement the automated data acquisition, pre-processing, and sensor integration framework outlined during Phase I activities.Technical Objective 2: Refine and implement the automated optimization of the fAPAR CSDM model that provides the foundation for estimating biomass, fractional cover, and crop calendar for the four Phase II implementation sites (a. Chesapeake Bay watershed, b. Minnesota, c. Northeastern US, and d. 90 sites across the South).Technical Objective 3: Refine and implement the automated integration of remote sensing-based estimates of biomass to constrain and improve the modeling of soil health and greenhouse gas emissions in the DNDC biogeochemical model.Product delivery objectiveTechnical Objective 4: Refine and implement the API delivery system outlined during Phase I activities.End-user integration and expanded marketing objectiveTechnical Objective 5: Work hand-in-hand with end-users at the Ecosystem Service Market Consortium, The Nature Conservancy, and General Mills to tailor our products and assist in integrating these products into their existing systems.
Project Methods
We will use our Geospatial Image Processing System (GIPS; open source software available at https://github.com/applied-geosolutions/gips) to acquire and pre-process multi-temporal Landsat, Sentinel-2A and 2B, and MODIS data for our testing locations. As noted in Phase I activities, modeling of canopy structure dynamics can be improved through the fusion of Landsat, Sentinel 2, and MODIS to create a seamless, gap-filled timeseries of NDVI. While our existing fusion module in GIPS is functional and was improved in Phase I, additional improvements to algorithms will be implemented in Phase II. The fusion process continues to produce sub-optimal results when the temporal gap between Landsat or Sentinel 2 extends beyond 40 days, particularly in the dynamic spring and fall seasons. These results can be improved by adjusting the way the weighting scheme is currently implemented in the algorithm.The Phase I implementation of the Canopy Structure Dynamic Model follows similar implementation of Liao et al., 2019 and Liu et al., 2010. This approach integrates patterns of daily temperature into the model. In Phase II, we will look to improve the CSDM approach by including the consideration of cumulative precipitation data as input into the modelling process (e.g. Sun and Schulz, 2017). In addition to temperature, precipitation is a primary controlling factor in the development of the crop canopy and integrating precipitation is expected to improve the fit between modeled and remote-sensing observed fAPAR.In Phase I, our team outlined two preferred approaches to using remote sensing observations for improved DNDC modeling of SOC, N2O, CH4, and other critical greenhouse gas emissions and soil health metrics: (1) to identify fields with similar growth patterns to be calibrated as a group, and (2) to directly calibrate seasonal growth at the field-scale. In Phase II, we will implement, demonstrate, and test these two approaches and report metrics on improvement and functionality. Our work on these two approaches will include both science research and implementation via software engineering and development.During Phase II, the team will build an API to receive data requests from customers and return detailed cover crop information back to the customers. These APIs will be tested with our partners and will be considered complete after our partners confirm successful use and implementation into their reporting systems.Communication between (a) the team's technical, scientific, and marketing staff and (b) our end-users and customers will result in improved products and more effective marketing. The world of soil health, ag conservation, and ecosystem service markets continues to grow and evolve rapidly. As end-user needs are refined, we will work to ensure the content and format of our data services, as well as our messaging, are well delivered to our clients. To ensure we understand our end-user needs, we will establish quarterly web meetings with representatives from our partner customers. These quarterly meetings will include updates on project progress provided by the project team, and updates from each organization regarding related projects and updates to their needs.

Progress 09/01/20 to 08/31/22

Outputs
Target Audience: Nothing Reported Changes/Problems: Nothing Reported What opportunities for training and professional development has the project provided? Nothing Reported How have the results been disseminated to communities of interest? Successful completion. Regrow has developed integrated data solutions for end-users and is actively serving data, through web applications and APIs. The following are a few examples of this work. External facing APIs have been built and deployed to support access for ESMC to OpTIS monitoring data and DNDC modeling data. ESMC is now using these APIs to seamlessly acquire the data needed to support their market program. Regrow has developed these APIs with third-party verification in mind as well and has drafted supporting documentation that has been reviewed and approved by voluntary carbon standards (Climate Action Reserve) and third-party verifiers (SustainCert). Regrow has delivered seven years (2015-2021) worth of OpTIS data for the entire continental United States to The Nature Conservancy. This data can also be accessed through Sustainability Insights, a tool to explore the regenerative agriculture landscape regionally in the US and Canada. General Mills has successfully completed an MRV assessment on various supply shed in the United States for 2021-2022 using OpTIS data to determine project adoption and model downstream outcomes. What do you plan to do during the next reporting period to accomplish the goals? Nothing Reported

Impacts
What was accomplished under these goals? Summary Through USDA-SBIR support, Regrow (formerly Dagan) has met the primary project goal of providing operational cover crop monitoring in the continental United States, Canada and parts of Europe for the last five years (2017-2021). This process was designed and engineered on cloud platforms to be deployed and scaled over large regions, allowing Regrow to quickly provide spatial and temporal insight for regenerative agriculture in real time. Regrow products have been optimized for internal integration with industry recognized models estimating outcomes for soil organic carbon (DNDC), water quality (APEX) and yield impacts (APSIM) at the field or regional scales. Through contributions from USDA-ARS, Regrow continues ongoing research towards algorithm refinement, large scale validation and product enhancement through incorporation of new sensor data (Landsat 9). To date, OpTIS data are served through three Regrow platforms, 1) the Monitor, Report and Validate (MRV) application which enables users can submit their own area of interest and generate a la carte intelligence, 2) the Sustainability Insights (SI) map allowing users to dynamically explore and query field level results over the entire OpTIS catalog and 3) The Monitor Application Programming Interface (API) giving users to power to plug into OpTIS data and organically build their own solutions. These market solutions provide academics, non-profits, commercial entities with the resources they need to make informed decisions on sustainability reporting, ecosystem service market participation, tracking conservation persistence on agricultural landscapes and scenario planning. USDA- ARS Engagement To support the mapping of cover crops presence and absence in the US, Regrow, USDA-ARS and USGS held 5 meetings in 2022. Over the last reporting period, with permission from the Maryland Department of Agriculture, USDA-ARS and USGS supplied five years of geospatial agronomic winter cover crop management data that represent 15% of fields enrolled in the Maryland Agricultural Water Quality Cost-Share (MACS) program from 2015-2021. MACS data were used by Regrow in this reporting period to calibrate the expansion of the OpTIS product to the Chesapeake Bay region. These data were essential for validating the accuracy of Regrow cover crop data in the region. Recent OpTIS cover crop data (fall of 2021- spring of 2022) for Mississippi are being evaluated for accuracy with USDA-USGS program enrollment data. This is being done by a postdoctoral researcher. Through this evaluation, USDA and USGS will also assess the role of cover crop performance (quality and quantity) and the presence/absence estimates to better recommend OpTIS product revisions. Furthermore, USDA-ARS will provide privacy protected geolocated data of cover crop presence from our on-farm trials to Regrow in Mississippi, Georgia, South and North Carolina (October, 2022). Additionally, in 2022 USDA-ARS USGS published a peer-reviewed article evaluating the efficacy of satellite radar and multispectral imagery to quantify cover crop biomass for which Regrow staff had previously provided input and advisement. Partnership with USDA-ARS has led to further data sharing opportunities with state, federal and non profit entities. The following organizations have agreed to help validate OpTIS data using internal ground control data on cover crop adoption - Practical Farmers of Iowa, the Illinois Department of Agriculture, and the USDA-NRCS in Minnesota and Indiana. Technical Objective1: Refine and implement the automated data acquisition, pre-processing, and sensor integration framework outlined during Phase I activities. Successful completion. The refined OpTIS algorithms for OpTIS have been successfully integrated into a Google Cloud Platform (GCP). This gives Regrow the capacity to run geospatial processing at scale with cloud computing, store the outputs strategically and access the data quickly. This new framework allowed Regrow to process and serve OpTIS products on hundreds of millions of agricultural acres in less than 6 months leveraging products like Google Earth Engines (GEE) and Google Cloud Storage (GCS). Technical Objective 2: Refine and implement the automated optimization of the fAPAR CSDM model that provides the foundation for estimating biomass, fractional cover, and crop calendar as outlined during Phase I activities for the four implementation sites. Successful completion. In partnership with USDA-ARS, Regrow has developed a scalable model for biomass as a time series for cover crop estimation and for input into downstream modeling like DNDC. To date, Regrow has made substantial progress in integrating a working model using sensor data to track biomass accumulation (see Figure 1.) as fAPAR influenced by water and temperature stress. Figure 1. Correlation between USDA-field samples and modeled biomass Technical Objective 3: Refine and implement the automated integration of remote sensing-based estimates of biomass to constrain and improve the modeling of soil health and greenhouse gas emissions in the DNDC biogeochemical model. Successful completion. Regrow's DNDC calculator has been optimized to integrate new data and process large areas. This enhancement has allowed Regrow to begin testing the impacts that biomass input data has on DNDC outputs and to quickly return results for validation and interpretation. Technical Objective 4: Refine and implement the API delivery system outlined during Phase I activities. Successful completion. Regrow has completed the build out and deployment of a functional OpTIS API called Monitor equipped with user documentation request protocols. Engineering efforts to optimize efficiency are ongoing, and are informed by user testing and customer feedback. Technical Objective 5: Work hand-in-hand with end-users at the Ecosystem Service Market Consortium, The Nature Conservancy, and General Mills to tailor our products and assist in integrating these products into their existing systems. Conclusion Through support of the USDA-SBIR, Regrow (formally Dagan) has risen to an industry leader in providing regenerative agricultural solutions. Building off of strong private and public relationships, Regrow leverages its ability to turn global earth observations and scientific modeling into meaningful solutions. These data help to make the agricultural landscape more resilient in the face of climate change, and provide opportunities for intervention and research on a global scale. Regrow recognizes the importance that resources from SBIR had on building out this capacity.

Publications


    Progress 09/01/20 to 08/31/21

    Outputs
    Target Audience:We made progress reaching two types of target audiences in Year 1: Consumer packaged goods companies that need information for supply chain reporting or for participation in ecosystem service markets; Non-govermental Organizations who seek to improve an understanding of the effectiveness of their incentive programs; Our cover crop monitoring system can help meet needs from both types of groups. Our discussions with organizations in both of these groups has involved both marketing/business development staff and science/technical staff. Changes/Problems:In the project period, our company has merged with Flurosat and rebranded as Regrow Agriculture. The merger has expanded our team of scientists, developers, and business development experts. The merger has hada positive impact on the project. What opportunities for training and professional development has the project provided?This project has funded a post-doc at the USDA-ARS in Beltsville, MD. How have the results been disseminated to communities of interest?The results have been disseminated to our research partners at the USDA-ARS in Beltsville, MD. What do you plan to do during the next reporting period to accomplish the goals?In Year 2 we will focus on: integrating RS based estimates into the DNDC model; (TO 3) integratingour cover crop products into the workflows of our partners at ESMC, TNC, and General Mills(TO 5)

    Impacts
    What was accomplished under these goals? TO1 - we have deployed our cover crop mapping algorithms within our Geospatial Image Processing System (GIPS) which is now deployed as a service; this technical objective is complete; TO2 - we have completed the prototype implementation of a new model for mapping biomass, and the developers are working to implement this algorithm to function across large scales - TO 2 is approximately 70% complete; TO 3 - we have enhance our outline and plan for integrating RS into DNDC - TO 3 is approximately 10% complete TO 4 - we have created an API called the OpTIS Field Service to deliver cover crop estimates to customers; work remains to add customer tracking to the API; TO 4 is approximately 60% complete TO 5 - we have signed additional contracts with our primary customers including ESMC, TNC, and General Mills and continue to work closely to provide critical data on cover cropping; TO 5 is 20% complete;

    Publications