Stream order selection for model generalization of the topographic map of Indonesia

Abstract

The importance of stream networks is related to other features in the topographic map e.g. as weighted-parameter for contour derivation. The generalization of these features needs complex parameters specifically geometrical and conceptual aspects. Geometric parameters consist of stream length and vertices, while the conceptual part handles stream networks connectivity as logical consequences. Stream networks selection is a type of important step on map features analysis and in map databases. This paper proposes a new approach for stream networks generalization of Topographic Map of Indonesia (as known as RBI) for 1:5,000 to 1:25,000 of scale by using geometrical and conceptual parameters. Three stages used in this research were: data pre-processing (include resolving the topological errors), generating stream order (1:5,000 of scale as an input), comparing stream order algorithms (Strahler, Scheidegger, Shreve, and Drwal), and performing feature similarity-based analysis (comparison of stream ordering results and 1:25,000 of scale). The research resulted four different stream orders and eight different similarity-analysis values since each algorithm was tested in two scenarios (in 1st scenario, order > 1 were selected while in 2nd scenario, order > 2 were selected). Eventually, after comparing those results, the Scheidegger method obtained the highest similarity value in on both 1st and 2nd scenarios. Further, generalization by using stream order selection delivered the representation of river in constructing map elements of RBI.