Scalar and Multivariate Approaches for Optimal Network Design in Antarctica

Thesis (Master's)--University of Washington, 2014 Observations are crucial for weather and climate, not only for daily forecasts and logistical purposes, for but maintaining representative records and for tuning atmospheric models. Here scalar theory for optimal network design is expanded in a...

Full description

Bibliographic Details
Main Author: Hryniw, Natalia
Other Authors: Hakim, Gregory J
Format: Thesis
Language:English
Published: 2014
Subjects:
Online Access:http://hdl.handle.net/1773/26157
Description
Summary:Thesis (Master's)--University of Washington, 2014 Observations are crucial for weather and climate, not only for daily forecasts and logistical purposes, for but maintaining representative records and for tuning atmospheric models. Here scalar theory for optimal network design is expanded in a multivariate framework, to allow for optimal station siting for full field optimization. Ensemble sensitivity theory is expanded to produce the covariance trace approach, which optimizes for the trace of the covariance matrix. Relative entropy is also used for multivariate optimization as an information theory approach for finding optimal locations. Antarctic surface temperature data is used as a testbed for these methods. Both methods produce different results which are tied to the fundamental physical parameters of the Antarctic temperature field.