Annual 30 m soybean yield mapping in Brazil using long-term satellite observations, climate data and machine learning
Year: 2022
Type: article
Abstract: Long-term spatially explicit information on crop yield is essential for understanding food security in a changing climate. Here we present a study that combines twenty-years of Landsat and MODIS data, climate and weather records, municipality-level crop yield statistics, random forests and linear regression models for mapping crop yield in a multi-temporal, multi-scale modeling framework. The study was conducted for soybean in Brazil. Using a recently developed 30 m resolution, annual (2001–2019) soybean classification map product, we aggregated multi-temporal phenological metrics derived from Landsat and MODIS data over soybean pixels to the municipality scale. We combined phenological metrics with topographic features, long-term climate data, in-season weather data and soil variables as inputs to machine learning models. We trained a multi-year random forests model using yield statistics as reference and subsequently applied linear regression to adjust the biases in the direct output of the random forests model. This model combination achieved the best performance with a root-mean-square-error (RMSE) of 344 kg/ha (12% relative to long-term mean yield) and an r2 of 0.69, on the basis of 20% withheld test data. The RMSE of the leave-one-year-out model assessment ranged from 259 kg/ha to 816 kg/ha. To eliminate the artifacts caused by the coarse-resolution climate and weather data, we developed multiple models with different categories of input variables. Employing the per-pixel uncertainty estimates of different models, the final soybean yield maps were produced through per-pixel model composition. We applied the models trained on 2001–2019 data to 2020 data and produced a soybean yield map for 2020, demonstrating the predictive capability of trained machine learning models for operational yield mapping in future years. Our research showed that combining satellite, climate and weather data and machine learning could effectively map crop yield at high resolution, providing critical information to understand yield growth, anomaly and food security.
Institutions: Texas Tech University, University of Maryland, College Park
Language: English
Subfield: Ecology
Field: Environmental Science
Domain: Physical Sciences
SDGs: Zero hunger, Climate action
Open Access status: bronze
Awards: THIS PROPOSED FOOD SECURITY AND AGRICULTURE CONSORTIUM (FSAC) WILL ADDRESS VARIOUS ASPECTS OF THE APPLICATION OF EARTH OBSERVATIONS (EO) TO DOMESTIC AND GLOBAL AGRICULTURE AND FOOD SECURITY THROUGH AN UNPRECEDENTED GROUP OF RESEARCH AND OPERATIONAL PARTNERS. WE HAVE ASSEMBLED A CONSORTIUM OF PRINCIPALS DRAWN FROM LEADING ORGANIZATIONS IN THE U.S. AND AROUND THE WORLD THAT USE EO AND DERIVED INFORMATION TO ADVANCE AGRICULTURE AND FOOD SECURITY. THE PROPOSED FSAC ANCHORS A BROAD NETWORK OF PARTNERS DRAWN FROM UNIVERSITIES GOVERNMENT AGENCIES AND RESEARCH ORGANIZATIONS NGOS INTERNATIONAL ORGANIZATIONS AND THE PRIVATE SECTOR REPRESENTING END USERS INFORMATION AND SERVICE PROVIDERS AND APPLICATIONS RESEARCHERS. FSAC PARTNERS LEAD THE WORLD IN ENGAGING END USERS IN USING EO-DERIVED INFORMATION FOR AGRICULTURE. THE FSAC EXPANDS ON EXISTING AND PAST NASA INVESTMENTS AND IS SPECIFICALLY DESIGNED TO AMPLIFY THE BENEFITS BROADLY THROUGH PUBLIC AND PRIVATE SECTORS IN THE US AND INTERNATIONALLY. LEVERAGING OUR CURRENT AND PAST WORK TO BUILD THE GEO G20 GLOBAL AGRICULTURAL MONITORING (GEOGLAM) INITIATIVE THE CURRENT PROPOSAL IS ESSENTIAL TO THE CONTINUED IMPLEMENTATION AND DEVELOPMENT OF GEOGLAM WHICH HAS ACHIEVED UNPRECEDENTED SUCCESS IN ENGAGING END-USERS INVOLVED WITH CROP PRODUCTION MONITORING AND FORECASTING. THROUGH THIS PROPOSAL THE CROP MONITORS LED BY UMD WILL CONTINUE TO DELIVER EO-DERIVED INFORMATION TO THE AGRICULTURAL MARKET INFORMATION SYSTEM (AMIS) AN INTERNATIONAL BODY OF CHIEF ECONOMISTS FROM MAJOR CROP EXPORT COUNTRIES AND PROVIDE A MONTHLY BULLETIN OF CROP CONDITIONS FOR FOOD INSECURE COUNTRIES WITH AN EXPANDING GROUP OF INTERNATIONAL AND NATIONAL FOOD SECURITY ORGANIZATIONS. THE FSAC END USER ENGAGEMENT WILL BE BROAD - ADDRESSING NOT ONLY INTERNATIONAL CROP PRODUCTION FORECASTS MARKETS AND TRADE ALONGSIDE REGIONAL/NATIONAL FOOD SECURITY EARLY WARNING DISASTER RESPONSE AND POLICY - BUT ALSO DOMESTIC COMMERCIAL AGRICULTURE AND FARM MANAGEMENT AND SMALLHOLDER FARM MANAGEMENT RESILIENCE AND MICRO-INSURANCE. IT WILL INCLUDE STRATEGIC SYSTEMATIZED ENGAGEMENT OF NON-TRADITIONAL USERS OF NASA DATA (INSURANCE INDUSTRY HUMANITARIAN ORGANIZATIONS AND INTELLIGENCE). THE FSAC MEMBERS WERE SELECTED STRATEGICALLY FOR THEIR EXPERTISE AND RELATIONSHIPS WITH A BROAD RANGE OF USER-GROUPS. OUR APPLICATIONS R&D WILL FOCUS ON CROP CONDITION MONITORING AND RISK ASSESSMENT CROPLAND CROP TYPE AND AREA MAPPING CROP YIELD AND PRODUCTION MODELING AND FORECASTING AND LINKING EO PRODUCTS TO SOCIO-ECONOMIC DATA TO ADDRESS FOOD SECURITY QUESTIONS. THE RESULTING SUITE OF INNOVATIVE PRODUCTS AND SERVICES WILL COMBINE EO PRODUCTS FROM DIFFERENT SENSORS WITH METEOROLOGICAL GROUND AND SURVEY DATA. OTHER MAJOR WORK ELEMENTS WILL INCLUDE: THE ECONOMIC AND SOCIAL VALUATION OF THE BENEFITS OF EO THROUGH ECONOMIC ANALYSIS OF EO FOR PRODUCTION FORECASTING AND REPRESENTATIVE END USER CASE STUDIES; LINKAGE TO NEW NASA SENSORS ENHANCING THE EARLY ADOPTERS PROGRAM; DEVELOPMENT OF AN FSAC PORTAL FOR INFORMATION COLLECTION DISSEMINATION AND INNOVATIVE COMBINATION; AND A PROFESSIONAL COMMUNICATIONS INITIATIVE GIVING BROAD OUTREACH FOR THE PROGRAM. THE DOMESTIC AGRICULTURE COMPONENT OF THE PROPOSAL WILL TEST NEW EO-BASED APPLICATIONS TO IMPROVE FARM MANAGEMENT AND HELP USDA NASS IMPROVE THE RELIABILITY OF ITS US STATISTICS (AREA/YIELD). FSAC WILL HOLD A SERIES OF THEMATIC COMMUNITY WORKSHOPS IN THE US ESTABLISHING A US GEOGLAM COMMUNITY OF PRACTICE (COP) TO COMPLEMENT THE INTERNATIONAL COP. THESE ACTIVITIES TOGETHER WILL DELIVER NASA AN OUTSTANDING COORDINATED AND WELL-BALANCED PROGRAM APPLYING EO TO FOOD SECURITY AND AGRICULTURE WITH STRONG CONNECTIONS TO RELEVANT NASA PROGRAMS SCIENCE TEAMS AND NEW MISSIONS. A GOAL OF THE FSAC IN ADDITION TO RESPONDING TO THE RFP IS TO EXPAND ITS ACTIVITIES PARTNERSHIPS AND COLLABORATIONS WITH NON-NASA FUNDING SUSTAINING THE CONSORTIUM AND ITS FUNCTIONS BEYOND THE 5-YEAR FUNDING PERIOD., G6075595415
OpenAlex ID: W4297978097