|
Abstract:
|
Remote sensing instruments are a major source of the data we use to monitor and learn about climate today. In this talk we will shed light on some of the very interesting steps behind the measurements and climate data from satellites that we use for our statistical analysis and modeling. The focus point in this talk will be measurements of CO2 from NASA’s OCO-2 satellite which in fact involve Bayesian inferences. The OCO-2 instrument measures reflected sunlight in three spectral regions that make a single sounding. These soundings are then used to estimate the column averaged CO2 dry air mole fraction using a physical forward model for how a given CO2 concentration, and other atmospheric properties, affects a sounding. Due to the amount of data collected and complexity of the forward model, various simplifications are needed to obtain an estimate of the posterior mode and posterior variance using the so-called optimal estimation method. We will discuss some the challenges that come up, such as treating the posterior as a Gaussian and unaccounted for uncertainties due to e.g. uncertain parameter inputs and forward model discrepancy.
|
