A Combined Active and Passive SWE Retrieval Algorithm Using SnowEx 2017 and Finland NoSREx datasets
Abstract
Active based retrieval algorithm of snow water equivalent (SWE) has been developed based on X- and Ku- band (10 and 17 GHz) and single polarization (VV-polarization) radar observations. To validate proposed algorithm, we have applied three sets of airborne SnowSAR data (including 2011 and 2012 campaigns in Finland; 2013 campaign in Canada, achieving a root-mean-square error (RSME) below 30 mm of SWE and a correlation coefficient above 0.64. In SnowEx 2017 winter campaign, both ground-based scatterometer and radiometer were deployed in Grand Mesa Colorado. While the scatterometer observation is from UWScat which is a dual frequency (10 and 17 GHz) radar system from the University of Waterloo, Canada, the passive radiometers operating at 19 GHz and 37GHz were provided by University of Michigan to acquire the Brightness Temperature, Tb data. Both are ground-based remote sensing dataset. Previous ground-based active and passive microwave datasets are available from Finland (ESA NoSREx; 2009-2013). By correlating these active and passive measurements with in situ field measurements supported by DMRT-bicontinuous based forward modeling, we are proposing a retrieval algorithm combined active and passive observations.
One of the challenging aspects for the active only SWE retrieval algorithm is to obtain a proper priori estimation, scattering albedo or effective grain size. We combine the high sensitivity of passive microwave with the high resolution of the active microwave using synthetic aperture radar (SAR) techniques to overcome. In the combined algorithm, we utilize the sensitivity of the passive measurements to estimate an accurate initial guess of the scattering albedo. With this initial guess, we classify the observed snow into two types: snow with large and small scattering albedo. For each class, we give an uniform priori of scattering albedo for the active retrieval algorithm in low spatial resolution. Then, the active algorithm retrieves both albedo and SWE in a much finer spatial resolution. The combined algorithm is validated against available ground based Finland NoSREx and SnowEx 207 dataset and proved to be effective. The performance of retrieval algorithm shows achieved root-mean-square error (RSME) between retrieval and measurement SWE is less than 30 mm.- Publication:
-
AGU Fall Meeting Abstracts
- Pub Date:
- December 2018
- Bibcode:
- 2018AGUFM.C12A..06Z
- Keywords:
-
- 0736 Snow;
- CRYOSPHEREDE: 0758 Remote sensing;
- CRYOSPHEREDE: 0794 Instruments and techniques;
- CRYOSPHEREDE: 0798 Modeling;
- CRYOSPHERE