In social-based delay tolerant network (DTN) applications, hand-held mobile devices exchange information. The inherent social property of DTN has encouraged contemporary researchers in exploiting social metrics to devise forwarding techniques for efficient routing. This work observes evidence of seasonal behavior in contacts between node-pairs in real mobility traces, and exploits it to devise a novel seasonality aware similarity measure. We incorporate seasonality information into tie-strength, and then use it as link weight in a weighted similarity measure which we extend from Katz similarity index. We propose a Seasonality Aware Social-based (SAS) DTN forwarding technique based on the proposed similarity measure and ego-betweenness centrality. Finally we perform real trace driven simulations to show that SAS outperforms baseline social-based DTN forwarding methods significantly.