Not listed
THE CROP PROGRESS DATASET (SOWING DATE, EMERGENCE, CONDITION RATINGS, ETC.) IS A VERY IMPORTANT SOURCE FOR CROP MODELLING, YIELD FORECASTING, AND CROP GROWTH ASSESSMENT. CURRENTLY, CROP PROGRESS AND GROWTH CONDITION FOR DIFFERENT CROP TYPES IN THE USA ARE RELEASED WEEKLY AT STATE OR ADMINISTRATIVE DISTRICT LEVELS BY THE USDA (UNITED STATES DEPARTMENT OF AGRICULTURE) NASS (NATIONAL AGRICULTURAL STATISTICS SERVICE). THESE DATA ARE SUMMARIZED FROM MORE THAN 3,600 RESPONDENTS WHO MAKE VISUAL OBSERVATIONS AND CONTACT WITH FARMERS IN THEIR COUNTIES. HOWEVER, COLLECTING AND DISSEMINATING CROP DATA THROUGH FIELD SURVEYING AND REPORTING ARE TIME-CONSUMING, LABOR-INTENSIVE, HIGHLY SUBJECTIVE, AND ERROR PRONE. MOREOVER, THE REPORT AT A STATE OR ADMINISTRATIVE DISTRICT LEVEL IS TOO COARSE SPATIALLY FOR THE OPERATIONAL CROP MANAGEMENT AND YIELD ESTIMATION EFFORTS.SATELLITE-BASED MONITORING OF CROP PROGRESS AND CONDITION HAS SIGNIFICANT ADVANTAGES IN TERMS OF SPATIAL AND TEMPORAL RESOLUTIONS AND HAS RECEIVED TREMENDOUS ATTENTIONS OVER THE LAST THREE DECADES. IN CURRENTLY AVAILABLE SYSTEMS, HOWEVER, THE MAIN ISSUES ARE: (1) LACK OF SPATIAL DISTRIBUTION INFORMATION OF CROP TYPES WHEN IMPLEMENTING A NEAR REAL TIME (NRT) MONITORING, (2) COARSE SPATIAL RESOLUTION (>500 M) RESULTING IN GROWTH INFORMATION FROM THE MIXTURE OF MULTIPLE CROP TYPES OR CROP AND NATURAL VEGETATION, AND (3) INSUFFICIENT TIMELY AVAILABLE HIGH QUALITY SATELLITE OBSERVATIONS TO ACCURATELY TRACK CROP GROWTH DUE TO THE FREQUENT CLOUD CONTAMINATIONS. MOREOVER, ALTHOUGH CROP TYPE MAPPING IN A FIXED TIME DURING THE GROWING SEASONS HAS BEEN IMPROVED, THE ALGORITHMS ARE UNABLE TO BE APPLIED FOR OPERATIONALLY CLASSIFYING CROP TYPES DUE TO THE COMPLEXITY OF CLASSIFICATION MODELS AND THE LACK OF TIMELY AVAILABILITY OF CLOUD-FREE SATELLITE OBSERVATIONS.THIS PROPOSAL IS TO DEVELOP AN ENHANCED GEOSPATIAL TOOL FOR NRT MONITORING OF SPECIES-SPECIFIC CROP GROWTH USING THE FUSION OF MULTIPLE NEW GENERATION SATELLITE OBSERVATIONS. THE FUSION WILL BE PERFORMED BY EMPLOYING THE TIME SERIES OF LOW TEMPORAL BUT HIGH SPATIAL RESOLUTION DATA FROM THE NASA HARMONIZED LANDSAT-8/9 AND SENTINEL-2 (HLS) PRODUCT (3-DAY, 30-M) AND THE TEMPORAL SHAPE OF HIGH TEMPORAL (5-10 MIN) BUT LOW SPATIAL RESOLUTION (500 M) DATA FROM THE ADVANCED BASELINE IMAGER (ABI) ONBOARD GEOSTATIONARY OPERATIONAL ENVIRONMENTAL SATELLITE R SERIES (GOES-R). IMPLEMENTING THE TOOL OF THE NRT MONITORING WILL BE ABLE TO REPORT CROP PROGRESS AND CONDITION IN THE FIRST DAY OF THE GIVEN IMPLEMENTATION WEEK (A LATENCY LESS THAN ONE WEEK) AND PREDICT THE CROP GROWTH FOR THE FOLLOWING WEEK. TO DEVELOP THIS ENHANCED GEOSPATIAL TOOL, THE SPECIFIC OBJECTIVES WE PROPOSE ARE TO: 1) DEVELOP A NOVEL ALGORITHM FOR GENERATING SYNTHETIC TIME SERIES OF CROP GREENNESS BY FUSING HLS AND ABI TIME SERIES; 2) INVESTIGATE MACHINE LEARNING MODELS FOR CLASSIFYING CROP TYPES WEEKLY WITHIN SEASON FROM THE SYNTHETIC TIME SERIES; 3) EXPLORE THE DETECTION OF SPECIES-SPECIFIC CROP,PROGRESS AND GROWTH CONDITION WEEKLY FROM TIMELY AVAILABLE HLS AND ABI OBSERVATIONS AND CLIMATOLOGICAL TIME SERIES; AND 4) PREPARE COMPUTER CODES FOR THE GEOSPATIAL TOOL AND DELIVER TO USDA NASS STAKEHOLDER.THE EXPECTED RESULT FROM THIS PROJECT IS A GEOSPATIAL TOOL FOR NRT MONITORING OF SPECIES-SPECIFIC CROP PROGRESS AND CONDITION AT A CROP FIELD. IT WILL PRODUCE THE FOLLOWING OUTPUTS: 1) ALGORITHMS AND COMPUTER CODES (WRITTEN IN C) TO PROCESS TIME SERIES SATELLITE DATA FROM HLS AND ABI OBSERVATIONS; 2) ALGORITHMS AND COMPUTER CODES TO FUSE HLS AND ABI TIME SERIES; 3) ALGORITHMS AND COMPUTER CODES FOR CROP PHENOLOGY DETECTIONS; 4) MACHINE LEARNING MODELS FOR CLASSIFYING CROP TYPES WITHIN CROP GROWING SEASON; 5) ALGORITHMS AND COMPUTER CODES FOR DETERMINING CROP GROWTH CONDITION; 6) RESULTS AND VALIDATION DATASETS FROM PHENOCAM OBSERVATIONS, UNMANNED AERIAL VEHICLES (UAV) OBSERVATIONS, AND USDA NASS CROP PROGRESS; 7) WEEKLY CROP PROGRESS AND GROWTH CONDITION AT A 30M FIELD FROM 2022-2026 ACROSS NINE STATES OF MIDWESTERN US; 8) SIX PEER-REVIEWED PAPERS TO BE PUBLISHED AT HIGH RANKED JOURNALS AND ~6 CONFERENCE PAPERS TO BE PRESENTED AT INTERNATIONAL CONFERENCES; 9) ONE PHD STUDENT AND ONE-POSTDOC TO BE TRAINED, AND 10) A GEOSPATIAL TOOL TO BE DELIVERED TO USDA NASS STAKEHOLDERS.THE OUTPUTS FROM THIS PROJECT WILL HAVE A BROAD IMPACT. THE PROPOSED NEW GEOSPATIAL TOOL IS ABLE TO PROVIDE NRT MONITORING OF CROP GROWTH CONDITION AT 30-M FIELD SCALES IN A WEEKLY BASIS. THE TIMELY MONITORING OF CROP GROWTH IN AN OPERATIONAL WAY WILL SERVE PRECISION AGRICULTURAL MANAGEMENT AND SUPPORT AGRICULTURAL FINANCE, SUCH AS FOR FARMERS TO MAKE DECISIONS OF MANAGEMENT PRACTICES AND INSURANCE COMPANIES AND POLICYMAKERS TO AVOID HUMAN AND LIVESTOCK FAMINE. WITH THE DELIVERY TO THE USDA NASS STAKEHOLDER FOR IMPLEMENTATION OF THIS TOOL, THE RESULT IS EXPECTED TO SIGNIFICANTLY IMPROVE NASS'S CROP PROGRESS REPORTS FOR SUPPORTING PRECISION CROP MANAGEMENT AND FINANCE AND INSURANCE DECISION MAKING. MOREOVER, THIS RESEARCH OUTPUT WILL IMPROVE OUR UNDERSTANDING OF SATELLITE CAPABILITY TO MONITOR CROP PROGRESS AND CONDITION AT THE FIELD SCALE. WITH THE PUBLICATIONS IN PEER-REVIEWED JOURNALS AND PRESENTATIONS AT INTERNAL CONFERENCES, THE RESULT WILL DELIVER THE SCIENCE-BASED KNOWLEDGE TO BROAD AUDIENCES.
$759.3k 7/24/23