Principal Investigator (PI): Hook Hua, NASA's Jet Propulsion Laboratory
Retrospective-analyses (reanalyses) and free-running climate models are critical tools in studying weather and climate variability and change. An important--but still incomplete--requirement is long-term continuous comparisons between models and observational data. From our recent peer-reviewed NASA-funded NEWS and Making Earth System Data Records for Use in Research Environments (MEaSUREs) projects, we have amassed a large collection of multi-sensor A-Train data and analysis capability spanning almost a decade. Level 2 comparisons of these data with models are an untapped science opportunity. For example, the Modern-Era Retrospective Analysis for Research and Applications (MERRA) project has generated a long-term record of the hydrological cycle, and fits well with model comparisons using our NEWS and MEaSUREs observations of the atmospheric water cycle. Modelers need data sets (such as NASA's multi-sensor A-Train data) to identify and correct biases in modeled physical processes occurring at many spatial or temporal scales. Collocation of satellite-measured water vapor, clouds, and temperature onto model grids is necessary for evaluation of model simulations of physical processes. In addition, scientific advances can be made by long-term global monitoring of climate-relevant processes quantified by the multi-year A-train record. The major difficulty with quantifying atmospheric phenomena over the lifetime of the A-train/NASA Earth Observing System (EOS) era is in efficient sub-sampling, quality control, and automation to obtain the necessary data (both from NASA data and modeling centers). Scientists need good dataset preparation and screening to properly compare Level 2 observations to sub-daily model grids. Tools and services are needed to streamline scientists' custom processing and analyses to assess and improve models. These are key uses of NASA data sets for improving the understanding of weather and climate variability and change.
We will reuse and integrate existing technologies and data sets (many developed from our prior Advancing Collaborative Connections for Earth System Science (ACCESS), MEaSUREs, NEWS, and American Recovery and Reinvestment Act [ARRA] grants) to reduce interoperability barriers and the unnecessary complexities in accessing, merging, and comparing multi-sensor satellite observations from the A-Train with climate models and reanalyses. Scientists performing model and observation comparisons for climate analyses face hurdles in data discovery, access, subsetting, subsampling, quality screening, regridding, collocation, analysis, and data sharing & collaboration. Our system, Collaborative Climate Model and Observational Data Services (CCMODS), will address these data access and interoperability issues that often exist in NASA Earth Science's distributed and heterogeneous data and information systems. CCMODS will also simplify the voluminous data transfer issue by automating observation and model data assembly, merging, and analysis on the server side. More importantly, these new capabilities address specific science gaps in the model evaluation process. Our science investigators will use this data system to move comparisons of satellite observations to models beyond monthly mean comparisons and simple statistics (mean differences and standard deviations). Our data system will be used to compare instantaneous Level-2 satellite observations of temperature, water vapor, etc. to three-hourly MERRA and European Center for Medium-Range Weather Forecasts (ECMWF) model grids. Our science team will also use the system to perform detailed model diagnosis (differences from Level-2 observations and trends) stratified by observed atmospheric processes (cloud scene and precipitation intensity), retaining the full Probability Density Functions (PDFs) of all relevant observed quantities.