1: \begin{abstract}

2: In this paper, we propose a space-time GMsFEM for transport equations.

3: Multiscale transport equations occur in many geoscientific applications,

4: which include subsurface transport, atmospheric pollution transport, and 

5: so on.

6: Most of existing multiscale approaches use spatial multiscale 

7: basis functions or upscaling, 

8: and there are very few works that design space-time multiscale functions

9: to solve the transport equation on a coarse grid.

10: For the time dependent problems, the use of space-time 

11: multiscale basis functions offers several advantages as

12: the spatial and temporal scales are intrinsically coupled.

13: By using the GMsFEM idea with a space-time framework, 

14: one obtains a better dimension reduction taking 

15: into account features of the solutions

16: in both space and time. 

17: In addition, the time-stepping can be performed using much 

18: coarser time step sizes compared to the case when 

19: spatial multiscale basis are used.

20: Our scheme is based on space-time snapshot spaces and

21: model reduction using space-time spectral problems derived from 

22: the analysis.

23: We give the analysis for the well-posedness and 

24: the spectral convergence of our method.

25: We also present some numerical examples to demonstrate the performance

26: of the method. In all examples, we observe a good accuracy with a few 

27: basis functions.

28: \end{abstract}

29: