THIS PROJECT WILL DEVELOP A DEEP LEARNING COMPUTATIONAL PLATFORM THAT INTEGRATES MULTI-SCALE MULTI-SENSOR SOIL MOISTURE OBSERVATIONS, SEASONAL CLIMATE FORECAST, AND USER INPUT IN REAL-TIME TO IMPROVE SKILL OF SOIL MOISTURE FORECAST ON SUB-SEASONAL TO SEASONAL TIMESCALES. A MULTI-DISCIPLINARY APPROACH IS PROPOSED BY COMBINING EXPERTISE FROM EARTH SYSTEM MODELING, BIG-DATA TECHNOLOGY, AND DROUGHT MONITORING AND FORECAST. THE THREE SCIENCE OBJECTIVES OF THE PROJECT ARE: (1) DEVELOP NEW ALGORITHMS FOR INTEGRATING SOIL MOISTURE DATA FROM DIFFERENT SOURCES INCLUDING USER INPUT, REMOTELY SENSED SOIL MOISTURE OBSERVATIONS, AND EARTH SYSTEM MODELING, (2) BUILD A SCALABLE BIG-DATA INFRASTRUCTURE, AND DEEP LEARNING ANALYTICS PLATFORM FOR REAL-TIME AND INTERACTIVE SOIL MOISTURE FORECAST APPLICATIONS, AND (3) DEVELOP NEW OR IMPROVED FORECAST ATTRIBUTES AT INTERFACE OF HUMAN-TECHNOLOGY-DATA INTERACTIONS. WE WILL DEVELOP THE PROPOSED PROJECT IN TWO STAGES: (1) A 25-KM RESOLUTION NATIONAL SOIL MOISTURE FORECAST USING A THREE-TIER MODELING APPROACH (2) DOWNSCALING THE SOIL MOISTURE FORECAST TO 1-KM RESOLUTION AND USER SPECIFIED LOCATION USING DEEP LEARNING. WE WILL USE CONTAINER-BASED SOFTWARE DEFINED STORAGE TECHNOLOGY THAT ENABLES ON-THE-FLY DOWNSCALING AND DEEP LEARNING, DYNAMICALLY AND FLEXIBLY AT USER LOCATION FOR UNKNOWN NUMBER OF USER REQUESTS. WE WILL TEST THE USEFULNESS AND USABILITY OF THE SYSTEM WITH STAKEHOLDERS IN THE APALACHICOLA-CHATTAHOOCHEE-FLINT RIVER BASIN AND CENTRAL GREAT PLAINS BY ORGANIZING TWO WORKSHOPS. OVERALL, WE PUT FORWARD A VISION OF THE NEXT GENERATION-INTERACTIVE SOIL MOISTURE FORECASTING SYSTEM THAT WILL REDUCE INPUT COST TO US AGRICULTURE BY PROVIDING A COST EFFECTIVE BIG-DATA TECHNOLOGY AND SKILLFUL SOIL MOISTURE FORECAST AT SUB-SEASONAL TO SEASONAL TIME SCALES.