A FUNDAMENTAL RECENT FINDING IN CLIMATE DYNAMICS IS THAT RADIATIVE FEEDBACKS DEPEND ON CHANGING SURFACE TEMPERATURE PATTERNS. THIS PATTERN EFFECT IS CURRENTLY WIDELY EXPLORED IN THEORY AND CLIMATE MODELS BUT AN INVESTIGATION OF ITS OBSERVABILITY IS STILL MISSING. OBSERVABILITY HOWEVER IS CRUCIAL FOR TESTING THE REALISM OF CLIMATE MODEL SIMULATIONS AND THEORY AND THE FAR-REACHING IMPLICATIONS OF THE PATTERN EFFECT ON CLIMATE PREDICTABILITY. RADIATIVE FEEDBACKS ARE LOOSELY DEFINED AS A CHANGE IN THE TOP OF THE ATMOSPHERE (TOA) RADIATIVE IMBALANCE IN RESPONSE TO A SURFACE TEMPERATURE CHANGE. BOTH THESE FIELDS ARE STRONGLY INFLUENCED BY INTERNAL VARIABILITY AND THE TIMESCALES OF THE RELEVANCE OF THE PATTERN EFFECT ARE STILL UNCLEAR. WE AIM AT LEVERAGING THE 2000-2020 RECORD OF OBSERVED SPATIAL TOA RADIATIVE IMBALANCE AND SEA SURFACE TEMPERATURE (SST) TO 1. QUANTIFYING REALISM OF CLIMATE MODELS' TOA RADIATION AND SST PATTERN EVOLUTION 2. ASSESSING INFLUENCE OF MODELS' LOCAL TOA RADIATION AND SST BIASES ON THEIR ABILITY TO REPRODUCE OBSERVED PATTERN CHANGES 3. QUANTIFY OBSERVED FEEDBACK PATTERN CHANGES AND USE THEM TO CONSTRAIN MODELED LONG-TERM RADIATIVE FEEDBACK PATTERNS DATA AND METHODS WE WILL USE THE NASA DATA PRODUCTS FOR TOA RADIATION (CERES EBAF) SST (MODIS) AND TROPOSPHERIC STABILITY FIELDS (MERRA2) FOR THE PERIOD 2000-2020 IN CONJUNCTION WITH: (A) SEVEN SINGLE MODEL LARGE INITIAL CONDITIONS ENSEMBLES SAMPLING CLIMATE VARIABILITY AT AN UNPRECEDENTED PRECISION (B) CMIP6 MODELS WITH INITIAL CONDITION ENSEMBLES OF TEN OR MORE (C) A RECENT COLLECTION OF MILLENNIAL-LONG COUPLED CLIMATE MODEL SIMULATIONS LONGRUNMIP WITH AND WITHOUT EXTERNAL FORCING. SIMULATION OF LARGE (A) AND SMALL (B) INITIAL CONDITION ENSEMBLES WILL BE USED TO TEST WHETHER CLIMATE MODELS CAN SPONTANEOUSLY REPRODUCE THE OBSERVED TOA RADIATION AND RADIATIVE FEEDBACK PATTERN AND THEIR EVOLUTION IN RELEVANT REGIONS. SIMULATIONS OF SEVERAL THOUSAND YEARS (C) ALLOW US TO TEST THE LIKELIHOOD OBTAINING THE OBSERVED PATTERNS GIVEN MODEL INTERNAL VARIABILITY. OUR COMPARISON METHODS INCLUDE PATTERN CORRELATIONS TREND AND SPATIAL SMOOTHING ANALYSES RANK FREQUENCY HISTOGRAMS AND PRINCIPAL COMPONENT ANALYSES. SIGNIFICANCE TO SOLICITATION WE WILL SUPPORT THE GOALS OF THE EARTH SCIENCE RESEARCH PROGRAM THROUGH ADVANCING THE UNDERSTANDING OF CHANGES IN THE EARTH S RADIATION BALANCE AND IMPROVE THE ABILITY TO PREDICT CLIMATE CHANGES BY BETTER UNDERSTANDING THE ROLES AND INTERACTIONS OF THE OCEANS ATMOSPHERE LAND AND ICE IN THE CLIMATE SYSTEM . THE PROPOSED RESEARCH COMPLEMENTS THE EARTH SCIENCE DIVISION S FOCUS ON CLIMATE VARIABILITY AND CHANGE. OUR RESEARCH WILL CHARACTERIZE LIMITS OF VALIDITY OF MODELS SPECIFICALLY THE INTERACTION OF MODEL COMPONENTS OCEAN AND ATMOSPHERE. DIRECTLY RELEVANT FOR NASA WILL BE THE RESEARCH OUTCOME OF A) HOW LONG A CERES-LIKE RECORD WOULD NEED TO BE TO MEANINGFULLY LINK SHORT- AND LONG-TERM RADIATIVE FEEDBACKS AND DETECT LOCAL FEEDBACK CHANGES AND B) HOW UNCERTAINTY IN THE OBSERVED TOA RADIATION AND SST RECORDS HAMPERS THE ABILITY TO CONSTRAIN CLIMATE MODELS PATTERN EVOLUTION. PI MARIA RUGENSTEIN IS AN ASSISTANT PROFESSOR AT COLORADO STATE UNIVERSITY AND PIONEERED THE INTER COMPARISON OF MILLENNIAL-LENGTH GENERAL CIRCULATION MODELS (LONGRUNMIP). SHE IS AN EXPERT ON CLIMATE SENSITIVITY OCEAN HEAT UPTAKE AND TIME VARIATIONS IN RADIATIVE FEEDBACKS. THE RESEARCH WILL BE CONDUCTED BY THE PI ONE GRADUATE STUDENT AND TWO SUMMER INTERNS AT CSU. JPL-COLLABORATOR MARK RICHARDSON DEVELOPED AN OCO-2-CALIPSO CLOUD RETRIEVAL PRODUCT AND WILL SERVE AS NASA-PRODUCT LIAISON.