FORESTS OF LARGE CANOPY FORMING KELPS FOUND ON SHALLOW SUBTIDAL REEFS IN TEMPERATE WATERS AROUND THE WORLD PROVIDE THE FOUNDATION FOR SOME OF THE MOST PRODUCTIVE AND VALUABLE COASTAL ECOSYSTEMS IN THE WORLD. LARGE KELPS ARE ECOSYSTEM ENGINEERS PROVIDING BOTH FOOD AND HABITAT TO A DIVERSE ARRAY OF BIOLOGICALLY AND COMMERCIALLY IMPORTANT SPECIES OF ALGAE INVERTEBRATES FISH AND MARINE MAMMALS WHILE EXPORTING LARGE QUANTITIES OF ORGANIC MATTER TO ADJACENT LITTORAL AND CONTINENTAL SHELF ECOSYSTEMS. HOWEVER KELP IS PARTICULARLY SENSITIVE TO CHANGES IN ENVIRONMENTAL CONDITIONS AND CLIMATE CHANGE HAS ALREADY IMPACTED THE GLOBAL ABUNDANCE AND DISTRIBUTION OF KELP FORESTS. CHARACTERIZING THESE CHANGES ON REGIONAL TO GLOBAL SCALES IS CRITICAL IF WE ARE TO ANTICIPATE THE IMPACTS OF FUTURE ENVIRONMENTAL CHANGES ON THESE IMPORTANT ECOSYSTEMS. CURRENTLY HOWEVER THERE IS A MARKED LACK OF DATA ON VARIABILITY AND TRENDS IN GLOBAL KELP ABUNDANCE. A RECENT METANALYSIS OF GLOBAL KELP TIME SERIES DATA HIGHLIGHTED POOR SPATIAL AND TEMPORAL DATA COVERAGE FOR MANY PARTS OF THE WORLD. MUCH OF OUR UNDERSTANDING OF LONG-TERM TRENDS IN KELP ABUNDANCE COMES FROM A FEW WELL-STUDIED REGIONS E.G. SOUTHERN CALIFORNIA. WE PROPOSE TO DEVELOP A GLOBAL DATASET OF CHANGES IN KELP CANOPY ABUNDANCE FROM THE EARLY 1980S THROUGH THE PRESENT AND USE THIS DATA TO: CHARACTERIZE TRENDS IN KELP ABUNDANCE IDENTIFY THE DRIVERS OF THOSE TRENDS AND IDENTIFY THE REGION S MOST VULNERABLE TO THE EFFECTS OF CLIMATE CHANGE. WE WILL DEVELOP AND VALIDATE AN AUTOMATED ALGORITHM FOR IDENTIFYING KELP CANOPY FROM LANDSAT TM/ETM+/OLI. WE WILL VALIDATE THIS ALGORITHM WITH AN EXTENSIVE TRAINING DATA SET OF HUNDREDS OF THOUSANDS OF CITIZEN SCIENCE CLASSIFICATIONS THROUGHOUT THE WORLD THUS MINIMIZING REGIONAL BIAS OF PAST EFFORTS. WE WILL THEN APPLY THIS ALGORITHM TO THE GLOBAL LANDSAT ARCHIVE. WE WILL EXPLORE TWO DIFFERENT AUTOMATED IMAGE ANALYSIS APPROACHES TO CREATE THIS GLOBAL DATASET. THE FIRST BUILDS ON OUR PREVIOUS METHODS COMBINING RANDOM FOREST CLASSIFIERS AND MULTIPLE ENDMEMBER SPECTRAL UNMIXING ANALYSIS. THE SECOND APPROACH WILL DEVELOP NEW DEEP LEARNING CLASSIFICATION ALGORITHMS USING A SUBSET OF OUR CITIZEN SCIENCE KELP VALIDATION DATA. WE WILL DOCUMENT TRENDS IN KELP ABUNDANCE AROUND THE WORLD AND EXAMINE THE ROLE OF ENVIRONMENTAL FACTORS SUCH AS WATER TEMPERATURE (USING SATELLITE SEA SURFACE TEMPERATURE DATA) NUTRIENT LEVELS (DERIVED FROM SEA SURFACE TEMPERATURE-NITRATE RELATIONSHIPS IN CERTAIN REGIONS) PHOTOSYNTHETICALLY AVAILABLE RADIATION (USING SATELLITE OCEAN COLOR DATA) AND WAVE DISTURBANCE (USING REGIONAL AND GLOBAL WAVE MODELS). WE WILL MODEL THE IMPACTS OF CLIMATE CHANGE ON THE GLOBAL DISTRIBUTION OF KELP AND IDENTIFY REGIONS PARTICULARLY VULNERABLE TO FUTURE CHANGE. THE ACTIVITIES DESCRIBED ABOVE WILL PROVIDE INSIGHT INTO THE RELATIONSHIPS BETWEEN KELP PRESENCE AND ABUNDANCE AND VARIOUS ENVIRONMENTAL AND CLIMATE VARIABLES. WE WILL USE CLIMATE MODELS TO PROJECT HOW CLIMATE CHANGE WILL IMPACT THE DISTRIBUTION AND ABUNDANCE OF KELP ON REGIONAL TO GLOBAL SCALES. THIS PROJECT WILL DEVELOP NEW METHODS TO MONITOR CHANGE IN TEMPERATE REEF ECOSYSTEMS ON GLOBAL SCALES WHICH WILL IN TURN HELP US BETTER UNDERSTAND THE DRIVERS OF THESE CHANGES. OUR RESULTS WILL PROVIDE A FOUNDATION THAT CAN BE BUILT UPON BY FUTURE MISSIONS SUCH AS SBG AND PACE.