The vast majority of theses in this collection are open access and freely available. There are a small number of theses that have access restricted to the WWU campus. For off-campus access to a thesis labeled "Campus Only Access," please log in here with your WWU universal ID, or talk to your librarian about requesting the restricted thesis through interlibrary loan.

Date Permissions Signed


Date of Award


Document Type

Masters Thesis

Degree Name

Master of Science (MS)


Environmental Studies

First Advisor

Miles, Scott B.

Second Advisor

Medler, Michael J.

Third Advisor

Mitchell, Robert J. (Geologist)


Flow directions and catchment algorithms have historically utilized raster-based data models. A significant body of literature focuses on raster-based interpolation errors, and the subsequent surface reconditioning to compensate for those errors, that together degrade the accuracy of the derived flow directions and catchments. This research seeks to improve upon the raster-based approach by developing and evaluating a vector-based approach to generating flow directions and delineating catchments that preserves the accuracy of the input point data through the use of irregular tessellated surface models. Specifically, the Python computer programming language was used in conjunction with a geographic information system (GIS) to develop ITSMHydro, a custom toolset that creates a Delaunay triangulated irregular network (TIN) from LiDAR bare-earth sample point data, and subsequently generates flow directions, delineates basins, and processes spurious sink catchments. Surface model accuracy, and area, shape, and overlap of the resulting catchments were compared with catchments delineated using industry-standard raster-based digital terrain models. The vector-based approach implemented through ITSMHydro was limited to file sizes less than approximately 120,000 LiDAR strikes that processed in approximately 30 hours, whereas the industry-standard raster-based approach transformed 111,000,000 LiDAR strikes across the study area into a 3-feet pixel surface model and generated catchment boundaries in approximately 36 hours. A root mean square analysis of surface models indicates that surface model quality is more heavily degraded when LiDAR sample points are interpolated to raster grids as opposed to surface models relying on Delaunay TIN interpolation, suggesting that the vector-based approach maintains the quality and precision of the LiDAR input data. For the four test areas in which the two approaches were compared, ITSMHydro generated catchments that were generally smaller (percent difference in areas ranged from -83.97% to 9.39%) and with more complex boundaries (i.e. lower isoperimetric quotient in 3 out of 4 test areas) than the associated raster-based catchments. Coefficient of areal correspondence (CAC), a measure of overlap between catchments generated by the two methods where a value of 1 indicates perfect overlap, ranged from 0.28 to 0.80 in the four test areas. Given the lower relative accuracy of raster-based surface models evident in the study area, these differences suggest use of the raster-based approach may compromise accuracy in area, shape, and location of the resulting catchments. A vector-based approach that preserves the accuracy of the input data is preferred, especially in areas of low topographic relief. The file size constraints limit application of the approach developed herein, however, at least until technological advances and/or code revisions improve computer processing speed and file size capacity.




Western Washington University

OCLC Number


Digital Format


Geographic Coverage

Lummi Reservation (Wash.)


Academic theses




Copying of this thesis in whole or in part is allowable only for scholarly purposes. It is understood, however, that any copying or publication of this thesis for commercial purposes, or for financial gain, shall not be allowed without the author's written permission.

Included in

Geography Commons