Landsat Analysis of the Texas Coastal Zone

Abstract
Mapping based on Landsat imagery was initiated along the Texas coast for four test sites selected for contrasting vegetation, environmental geology, and levels of economic development. Standard Landsat transparencies of part of a 1:1,000,000 band 7 image were optically enlarged to fit an existing 1:125,000 map base. Land-water, cropland, and other distinctive boundaries were extracted from the enlargement and compiled on an overlay on which additional unit boundaries from other bands and from the false-color composite of bands 4, 5, and 7 were added. Each resulting map unit was then classified according to a bilevel scheme modified from published systems to meet the priorities of this study. These priorities are, in order of greatest emphasis, the inventorying and monitoring of wetlands, land use, beaches and dunes, and bay systems. Primary classes include urban and built-up land, agricultural land, grassland, rangeland, woodland, water, wetland, and barren land. Twenty -three secondary categories include five wetland classes: topographically low marshes, topographically high marshes, tidal flat, submerged sea-grass beds and algal flats, and vegetated dredge spoil. Water was separated into four classes on the basis of turbidity to emphasize flow pattern and suspended-sediment distribution rather than water-body type. Change detection was accomplished by overlaying enlarged positive and negative transparencies. Areas of radiance change were caused by seasonal growth stage of wetland vegetation, burning of rangeland, and status of cropland. Multiple transparencies in bands 5 and 7 were used to map the addition of barren dredge spoil, a unit that was easily detected as a result of its extremely high reflectance. Band 5 images provided a synoptic view of the turbidity plume issuing from Aransas Pass and indicated that turbid water was derived from both the inlet and the surf zone adjacent to the jetties. The test sites mapped were the Freeport - San Luis Pass area, the upper San Antonio Bay - Espiritu Santo Bay vicinity, southern Padre Island - Laguna Madre, and the Harbor Island area. Substantial variation in areal extent and species content of wetlands was found. The limits of delineation of all categories depended on the size of the map unit relative to Landsat's limit of resolution, reflectivity contrast with surrounding units, and unique characteristics of each image. Mapping of the Harbor Island site from three successive scenes showed significant variations in the detectability of urban and transportation features. An evaluation of the Landsat mapping was made by a random-point comparison with aerial photographs. A mean accuracy of 87.4 percent was obtained for the 806 points checked; within each category, results varied from 64.9-percent accuracy for undifferentiated barren land to 93.1-percent accuracy for cropland. The results indicate that Landsat offers a valuable source of current information on coastal resources for use in conjunction with the existing data base on Texas coastal resources.
Authors
Robert J. Finley
Citation

Finley, R. J., 1979, Landsat Analysis of the Texas Coastal Zone: The University of Texas at Austin, Bureau of Economic Geology, Report of Investigations No. 93, 71 p.

Code
RI093
ISSN
2475-367X
Number
93
Number of figures
29
Number of pages
71
Number of plates
1
Publisher
The University of Texas at Austin, Bureau of Economic Geology
Series
Report of Investigation
Year
1979