et al.
CA-380 UKF, or the “Conditional Autoregressive inversion of the UKF Parts” method of Bernstein et al., is a method for computing high-resolution estimates of subsurface temperature profiles from remotely-sensed temperature data. This method was designed to be an improvement over traditional autoregressive (AR) approaches, such as kriging, as it makes use of a variably-weighted, nonlinear fitting approach known as the Unscented Kalman Filter (UKF). This UKF approach offers improved resolution and better smoothing of temporal fluctuations when compared to traditional autoregressive approaches. In addition, the CA-380 UKF method of Bernstein et al. can use various datasets and parameters, such as soil surface temperatures, ground measurements, and land surface maps, and is well-suited for handling geographic and seasonal variations of temperatures over a particular area.