1: \begin{abstract}

2: 		The article presents the theory of stationary regularly varying random fields. In this context, we prove a new compound Poisson approximation theorem under appropriate dependence conditions, and demonstrate a couple of effective methods for checking its conditions.

3: 		The main result is based on an extension of the  Grigelionis' Poisson convergence theorem and a novel concept of anchoring of the tail process of regularly varying sequences and fields.

4: %		The theory is based on an extension of some standard point processes  limiting theory  to  general Polish spaces. 

5: 		We  apply these theoretical results  to shed entirely new light on the classical problem of evaluating local alignments of biological sequences.	

6: 		

7: 		{\textbf{Keywords}: compound Poisson approximation, \and random fields, \and regular variation, \and tail process, \and point process, \and local sequence alignment, \and Gumbel distribution}		

8: 		

9: 		{\textbf{MSC classes}: Primary 60G70; Secondary 60G60, 92D20}	

10: 	\end{abstract}

11: