 |
| Titel |
4DVar Data Assimilation for Dust Emission Parameter Estimation over East Asia Area |
| VerfasserIn |
Jianbing Jin, Hai Xiang Lin, Arnold Heemink, Arjo Segers |
| Konferenz |
EGU General Assembly 2017
|
| Medientyp |
Artikel
|
| Sprache |
en
|
| Digitales Dokument |
PDF |
| Erschienen |
In: GRA - Volume 19 (2017) |
| Datensatznummer |
250139567
|
| Publikation (Nr.) |
 EGU/EGU2017-2831.pdf |
|
|
|
|
|
| Zusammenfassung |
| The Severe Dust Storms (SDS) have long played a wide and negative impact on the
atmospheric environment and climate system. To reduce the social and economic
influences caused by the SDS, various model-based dust forecasting and early warning
systems have been developed. However, the simulated dust concentrations by those
existing models sometimes show a discrepancy of more than two orders of magnitudes
from the observations. The most important reason for such large differences is the
difficulty in accurately identifying the dust emission source region and emission
rate.
In our study, the LOTOS-EUROS/Dust is used to simulate the SDS over East Asia areas. A
geographic dependent friction velocity threshold (FVT), instead of a spatially constant
one, is introduced in the dust emission equation. A trajectory-based 4DVar data
assimilation scheme is designed to estimate the spatially different FVTs. By using the
trajectories (the ensemble model realization perturbations with the Monte Carlo sampled
FVTs), an accurate approximation of the expected FVTs with high efficiency can be
obtained.
Twin experiments have been implemented, where the 2D Aerosol Optical Depth (AOD)
observations transformed from the expected model realization are assimilated, both the
estimated FVTs and the forecast dust concentrations are evaluated. Besides, the improved
FVTs (for trajectories) sampling scheme and model-based FVTs reduction are also
implemented, which can further improve the forecast accuracy without increasing the number
of trajectories. |
|
|
|
|
|
|