Improving the Mean and Uncertainty of Ultraviolet Multi-Filter Rotating Shadowband Radiometer In-Situ Calibration Factors: Utilizing Gaussian Process Regression with a New Method to Estimate Dynamic Input Uncertainty

Abstract

Abstract. To recover the actual responsivity for Ultraviolet Multi-Filter Rotating Shadowband Radiometer (UV-MFRSR), the complex (e.g. unstable, noisy, and with gaps) time series of its in-situ calibration factors (Vo) need to be smoothed. Many smoothing techniques require accurate input uncertainty of the time series. A new method is proposed to estimate the dynamic input uncertainty by examining overall variation and subgroup means within a moving time window. Using this calculated dynamic input uncertainty within Gaussian Process regression (GP) provides the mean and uncertainty functions of the time series. This proposed GP solution was first applied on a synthetic signal and showed significant smaller RMSEs than a Gaussian Process regression performed with constant values of input uncertainty and the mean function. GP was then applied to three UV-MFRSR Vo time series at three ground sites; The method appropriately accounted for variation in slopes, noises, and gaps at all sites. The validation results demonstrated that the agreement between aerosol optical depths (AODs) calculated using Vo determined by the GP mean function and AERONET AODs were consistently better than those calculated using Vo from standard techniques (e.g. moving average). The improved accuracy of in-situ UVMRP Vo values suggests the GP solution is a robust technique for accurate analysis of complex time series and may be applicable to other fields.