Overview of SMOS performance in terms of global soil moisture monitoring after six years in operation

Y. H. Kerr, A. Al-Yaari, N. Rodriguez-Fernandez, M. Parrens, B. Molero, D. Leroux, S. Bircher, A. Mahmoodi, A. Mialon, P. Richaume, S. Delwart, A. Al Bitar, T. Pellarin, R. Bindlish, T. J. Jackson, C. Rudiger, P. Waldteufel, S. Mecklenburg, J.-P. Wigneron

Research output: Contribution to journalArticleResearchpeer-review

146 Citations (Scopus)

Abstract

The Soil Moisture and Ocean Salinity satellite (SMOS) was launched in November 2009 and started delivering data in January 2010. The commissioning phase ended in May 2010. Subsequently, the satellite has been in operation for over six years while the retrieval algorithms from Level 1 (L1) to Level 2 (L2) underwent significant evolutions as knowledge improved. Moreover, other approaches for retrieval at L2 over land were investigated while Level 3 (L3) and Level 4 (L4) were initiated. In this paper, these improvements were assessed by inter-comparisons of the current L2 (V620) against the previous version (V551) and new products (using neural networks referred to as SMOS-NN) and L3 (referred to as SMOS-L3). In addition, a global evaluation of different SMOS soil moisture (SM) products (SMOS-L2, SMOS-L3, and SMOS-NN) was performed comparing products with those of model simulations and other satellites. Finally, all products were evaluated against in situ measurements of soil moisture (SM). To achieve such a goal a set of metrics to evaluate different satellite products are suggested.

The study demonstrated that the V620 shows a significant improvement (including those at L1 improving L2) with respect to the earlier version V551. Results also show that neural network based approaches can often yield excellent results over areas where other products are poor. Finally, global comparison indicates that SMOS behaves very well when compared to other sensors/approaches and gives consistent results over all surfaces from very dry (African Sahel, Arizona), to wet (tropical rain forests). RFI (Radio Frequency Interference) is still an issue even though detection has been greatly improved through the significant reduction of RFI sources in several areas of the world. When compared to other satellite products, the analysis shows that SMOS achieves its expected goals and is globally consistent over different eco climate regions from low to high latitudes and throughout the seasons.

Original languageEnglish
Pages (from-to)40-63
Number of pages24
JournalRemote Sensing of Environment
Volume180
DOIs
Publication statusPublished - Feb 2016

Keywords

  • Level 2
  • Level 3
  • Neural networks
  • Product inter-comparison
  • Retrieval accuracy metrics
  • SMOS
  • Soil moisture
  • Validation

Cite this