GROWERS OF PERENNIAL FRUITS (E.G. GRAPES, APPLES, OR CHERRIES) OFTEN USE VARIOUS TECHNIQUES TO ESTIMATE PLANT WATER DEMAND WITH THE INTENT OF TRANSLATING THE DATA INTO PROPER IRRIGATION SCHEDULING (I.E. TIME AND VOLUME OF WATER TO BE APPLIED). THESE TECHNIQUES INCLUDE MEASUREMENT OF PLANT WATER POTENTIAL, SOIL MOISTURE, AND CANOPY SPECTRAL INDICES. HOWEVER, THESE METHODS FOR ESTIMATING WATER STATUS ARE LIMITED BY LARGER UNCERTAINTY AND VARIABILITY OVER SPACE AND TIME. GROWERS, THEREFORE, OFTEN RELY ON THEIR JUDGMENT FOR IRRIGATION SCHEDULING DECISIONS, WHICH MAY LEAD TO INEFFICIENT WATER USE. THE LONG TERM GOAL OF THIS RESEARCH IS TO IMPROVE CROP QUALITY AND YIELD WITH MINIMAL WATER FOOTPRINT BY USING SMART IRRIGATION. TO ACHIEVE THIS GOAL, MULTI-SOURCE DATA (E.G. HYPERSPECTRAL IMAGES, SOIL MOISTURE AND CLIMATE DATA) WILL BE COLLECTED AND NOVEL BIG DATA ANALYTICS TECHNIQUES WILL BE USED TO INTEGRATE AND ANALYZE THOSE DATA, WHICH IS EXPECTED TO LEAD TO MORE RELIABLE ESTIMATION OF PLANTWATER REQUIREMENTS. THIS ACTIONABLE INFORMATION WILL THEN BE USED FOR REAL-TIME CONTROL OF A SMART, PRECISION IRRIGATION SYSTEM IN ACTUAL CROP FIELDS. THESE RESEARCH ACTIVITIES WILL CONTRIBUTE TO NIFA'S RESEARCH INTEREST IN NEW APPROACHES TO"EXTRACT ACTIONABLE INFORMATION" FROM LARGE AGRICULTURAL DATA.MAJOR FUNDAMENTAL CONTRIBUTION OF THE PROJECT WILL BE TO SCALE THE INTEGRATION OF MULTIPLE TYPES OF SPATIO-TEMPORAL DATA (OR D4, E.G. SENSOR IMAGES, WEATHER DATA, AND MOISTURE PROBE DATA STREAMS) WITH DATA FUSION AND MINING TOOLS, AND HUMAN-IN-THE-LOOP MACHINE LEARNING SYSTEMS THAT WILL THEN ALLOW FOR LARGE-SCALE, DATA-DRIVEN DECISIONS IN PRECISION IRRIGATION. THIS IS A KEY DEPARTURE FROM EXISTING IRRIGATION MANAGEMENT APPROACHES. AS CURRENT ESTIMATION TECHNIQUES FOR PLANT WATER STATUS SUFFER FROM WIDE VARIABILITY AND UNCERTAINTY, POINT MEASUREMENTS CANNOT BE RELATED TO PLANT WATER DEMAND WITH A DESIRED LEVEL OF CONFIDENCE. THIS STUDY WILL GENERATE EFFICIENT TECHNIQUES FOR KNOWLEDGE FUSION, CORRELATION, AND PATTERN RECOGNITION FROM DATASETS THAT ARE SPECIFICALLY OPTIMIZED FOR THE ANALYSIS OF PLANT WATER STRESS LEVELS. THIS TECHNIQUE ENABLES TRANSLATION OF COLLECTABLE DATA TO ACTIONABLE INFORMATION FOR GROWERS TO MAKE RELIABLE DECISIONS FOR IRRIGATION SCHEDULING.THE SUCCESSFUL COMPLETION OF THE PROJECT WILL LEAD TO OPTIMAL USE OF IRRIGATION WATER WHILE IMPROVING CROP YIELD AND QUALITY THUS MAKING A POSITIVE IMPACT IN REDUCING THE WASTEFUL CONSUMPTION OF NATURAL RESOURCES AND INCREASING ECONOMIC, SOCIAL, AND ENVIRONMENTAL SUSTAINABILITY OF RURAL AGRICULTURAL COMMUNITIES (WHERE MAJORITY OF FARMS ARE LOCATED). NOVEL BIG DATA ANALYTICS TOOLS FOR HUGE AGRICULTURAL DATA AND INTEGRATED CYBER PHYSICAL SYSTEM DEVELOPED IN THIS PROJECT WILL ALSO HAVE APPLICABILITY IN OTHER AREAS SUCH AS HEALTHCARE, TRANSPORTATION, DISASTER MANAGEMENT, AND ENERGY.