Combining Datasets of Satellite-Retrieved Products. Part I: Methodology and Water Budget Closure

• DOI: 10.1175/jhm-d-13-0148.1
• Publication Date: 2014-04-26T05:46:57Z
• Authors: Filipe Aires
• Keywords: Water cycle

Abstract This study addresses in general terms the problem of optimal combination multiple observation datasets. Only satellite-retrieved geophysical parameter datasets are considered here (not raw satellite observations). focuses on terrestrial water cycle and presents methodologies to obtain a coherent dataset four key components: precipitation, evapotranspiration, runoff, storage. Various innovative “integration” introduced: simple weighting (SW), constrained linear (CL), interpolation (OI), neural networks (NN). The term will be used here, not “assimilation,” as no model included data fusion process. A postprocessing filtering (PF) step can impose budget closure after integration method. It is shown that this constraint actually improves estimation components. techniques tested using real over Mississippi Niger basins from situ measurements. Monte Carlo experiment with synthetic uncertainty perturbation measure ability SW, OI, NN, or without PF step, retrieve Once added, have equivalent accuracies. need for these types should increase future since now available climate community needs combine them into unique, optimal, parameters. companion paper test at basin global scales.