Estimation of volcanic ash emissions through assimilating satellite data and ground-based observations

Sha Lu, Hai Xiang Lin, Arnold Heemink, Arjo Segers, Guangliang Fu

Research output: Contribution to journalArticleScientificpeer-review

9 Citations (Scopus)
23 Downloads (Pure)


In this paper, we reconstruct the vertical profile of volcanic ash emissions by assimilating satellite data and ground-based observations using a modified trajectory-based 4D-Var (Trj4DVar) approach. In our previous work, we found that the lack of vertical resolution in satellite ash column data can result in a poor estimation of the injection layer where the ash is emitted into the atmosphere. The injection layer is crucial for the forecast of volcanic ash clouds. To improve estimation, Trj4DVar was implemented, and it has shown increased performance in twin experiments using synthetic observations. However, there are some cases with real satellite data where Trj4DVar has difficulty in obtaining an accurate estimation of the injection layer. To remedy this, we propose a modification of Trj4DVar, test it with synthetic twin experiments, and evaluate real data performance. The results show that the modified Trj4DVar is able to accurately estimate the injection height (location of the maximal emission rate) by incorporating the plume height (top of the ash plume) and mass eruption rate data obtained from ground-based observations near the source into the assimilation system. This will produce more accurate emission estimations and more reliable forecasts of volcanic ash clouds. Also provided are two strategies on the preprocessing and proper use of satellite data.
Original languageEnglish
Pages (from-to)10.971–10.994
Number of pages24
JournalJournal Of Geophysical Research-Atmospheres
Issue number18
Publication statusPublished - 27 Sep 2016


  • Data assimilation
  • volcanic ash
  • satellite data
  • emission estimation


Dive into the research topics of 'Estimation of volcanic ash emissions through assimilating satellite data and ground-based observations'. Together they form a unique fingerprint.

Cite this