GPS-based analysis of shoreline change, 1995-2005, Mad Island Marsh Preserve, Matagorda County, Texas.
The Texas Nature Conservancy's Mad Island Marsh Preserve is a 29-[km.sup.2] tract containing the east arm of Mad Island Lake, its associated freshwater and brackish marshes, and surrounding upland prairie and shrub land habitats. The marshes on the preserve provide an important habitat for many aquatic organisms and an important wintering ground for millions of migratory birds that use the Central Flyway each year. Since 1993, the preserve has ranked among the top-five areas in the nation in number of species counted during the annual Audubon Society's Christmas Bird Count (L. Halsted, Texas Nature Conservancy, Pers. Comm., 2003). The shoreline of the preserve borders the Gulf Intracoastal Waterway (GIWW), a man-made canal constructed in 1941 (Figure 1a).
Barge traffic in the GIWW has caused considerable wave erosion of preserve shorelines over the last 65 years. Williams (1993a) calculated erosion rates along the shores of the preserve based on sequential aerial photographs covering the period 1943-1991. In 1993 Williams conducted a second study of shoreline erosion based on repeated measurements of the separation between the shoreline and a series of survey stakes set in the ground every 30.48 m along the shoreline. The surveys were conducted in June of 1992 and in May of 1993, and were used to derive 1-year erosion rates (Williams 1993b). The results of those studies showed long-term shoreline erosion approached 2 m/year on some parts of the preserve. The studies also suggested that erosion may threaten freshwater marsh habitats bordering Mad Island Lake, since erosive shortening of Mad Island Bayou--a tidal inlet connecting the lake to Matagorda Bay--is likely to increase salt water intrusion into the lake. In response to these findings, the Texas Nature Conservancy in 1996 constructed a protective concrete barrier along the shoreline bordering Mad Island Bayou in the hopes of preventing further erosion (Fig. 1b: section C).
The focus of this study is to determine recent shoreline change along a ca. 1.2 km stretch of shoreline bordering the mouth of Mad Island Bayou (Figure 1b). The objectives of the study are to: (1) measure shoreline change between 1995 and 2005, using a combination of high-resolution aerial photographs, GPS and GIS, (2) compare shoreline change derived for this study with the results of Williams (1993a; 1993b), (3) evaluate the effectiveness of the concrete mat emplaced in 1996 in preventing erosion (shoreline C in Figure 1b), and (4) investigate the variation in shoreline change between the different shoreline types found along this part of the preserve; these include an area of Spartina marsh growing out into the GIWW, an area of marsh bordered by a ca. 0.5 m cliff, the concrete erosion control mat, an area of dense woodland bordered by a ca. 1.5 m cliff and an area fronted by a gently sloping sandy beach (Figure 1b: sections A to E).
MATERIALS AND METHODS
1995 shoreline map. -- A map of the shoreline in 1995 was derived from a 1-meter-resolution Digital Ortho Quarter Quad (DOQQ) downloaded from the Texas Natural Resources Information System web site (TNRIS 2005). For the purposes of the study, the shoreline was defined as the edge of the erosional scarp that borders the majority of the GIWW in the study area. Where the erosional scarp was not present, the vegetation line was used. The shoreline was defined in this manner because it is readily identifiable on air photographs and is consistent with Williams' (1993a) methods. To improve accuracy and provide assessment of digitizing error, the shoreline was digitized three times at a scale of 1:500 (large visible pixels), 1:1,000 (medium visible pixels) and 1:1,500 (small visible pixels). The three resulting digitized shorelines were averaged within the GIS to create a single 1995 shoreline. The principles of the Digital Shoreline Analysis System (DSAS) extension, written for ArcView[TM] (Theiler et al. 2003), were used to create a single shoreline. DSAS is based on creating regularly spaced perpendicular transects to be used as measurement locations across multiple shorelines (Theiler et al. 2003). Following this methodology, five-meter-interval transects were created across the three shorelines and the distance across the shorelines was measured at each transect. The average distance across the three shorelines was 0.97 m. Based on this result, the digitizing error was assumed to be [+ or -] 1 m. The GIS was used to create a single 1995 shoreline passing through the average position of the three digitized shorelines along each transect (Figure 2).
[FIGURE 1 OMITTED]
2005 shoreline map. -- In February of 2005, a sub-meter accuracy backpack GPS was used to map the shoreline. To improve accuracy and provide assessment of GPS mapping error, the shoreline was mapped three times. The GPS files were differentially corrected and imported into the GIS. The separation of the three resulting shorelines was measured along transects at 5-m intervals. The average distance across the three shorelines was 0.69 m. Based on this result, the GPS mapping error was also assumed to be [+ or -] 1 m. The three shorelines were averaged within the GIS to create a single 2005 shoreline.
Additional error assessment. -- Because mapping error plays a critical role in determining the magnitude of shoreline change, two additional error assessment techniques were employed: the accuracy of the 1995 DOQQ was further assessed by collecting GPS positions of a number of Ground Control Points (GCPs) within the study area. Eleven GCPs, readily located in the field and identifiable as a single pixel on the DOQQ, were selected (examples include the corner of a building and the center of a road intersection). The GPS positions were imported into the GIS and compared to the position of the GCPs on the DOQQ: all 11 GPS points were located within the correct pixel on the map, suggesting a level of accuracy of [+ or -] 1 m.
[FIGURE 2 OMITTED]
To further assess the accuracy of GPS-based distance measurements, a GPS point was recorded for seven survey stakes remaining from Williams' 1993b study (stake numbers 1, 3, 8, 9, 10, 11 and 13-all within section A of the study area (Fig. 1b); other stakes from Williams' 1993b study could not be found within the study area and are presumably missing). The perpendicular distance from each stake to the shoreline was measured with a 30m tape and the shoreline position was recorded with a GPS point. The tape was also used to measure the distances between stakes 1 and 3, 8 and 9, 9 and 10, 10 and 11 and 11 and 13. The GPS points were imported into the GIS and distances between points were measured for comparison to the same distances measured by tape (Table 1).
All the GPS-based measurements were within 0.4 m of the tape measurements (mean difference was 0.13 m). This result is consistent with the reported sub-meter accuracy of the GPS and is within the [+ or -] 1 m mapping error derived from shoreline mapping. Overall, the assessment of error suggests a point accuracy of [+ or -] 1 m for both the DOQQ and GPS-based shoreline maps. This provides a distance measurement accuracy of [+ or -] 2m (each end of a measured line is considered a point with [+ or -] 1 m accuracy).
Shoreline change: 1995-2005. -- The GIS was used to create shoreline-perpendicular transects at 10-m spacing along the shoreline of the entire study area. The separation of the 1995 and 2005 shorelines was measured along the resulting 110 transect lines. Although shoreline retreat (erosion) was apparent along many transects, shoreline accretion (growth of the shoreline out into the GIWW) was found at approximately one third of measurement locations.
Because one of the objectives of the study is to compare shoreline change between different shoreline types, the results are discussed in the context of the shoreline divisions shown in Figure 1b. As the measurement error for distance is [+ or -] 2 m, the error for calculated rates of shoreline change over the 10 year period of the study is [+ or -] 0.2 m/year.
Section A. -- This section of shoreline consists of Spartina marsh growing out into the GIWW with no visible scarp at the water's edge. Changes along this ca. 430-m-long section are a mixture of erosion and accretion (Figure 3). Along many transects, the shoreline position between 1995 and 2005 is essentially unchanged, the erosion or accretion amount falling within the 2-m measurement error. There are however pockets of erosion and accretion where shoreline changes exceed the measurement error. The maximum erosion is between 5.2 and 9.2 m (7.2 [+ or -] 2 m). The maximum accretion is between 2 and 6 m (4 [+ or -] 2 m). The average change for the entire section is erosion between 0.2 and 4.2 m (2.2 [+ or -] 2 m; equivalent to 0.02 to 0.42 m/year); however, the use of an average is misleading in this instance, because the small pockets of erosion and accretion, which are clearly present, cancel out when averaged.
These findings contrast with those of Williams (1993a), who found no evidence of accretion between 1943 and 1991, and calculated an average erosion rate for this section of shoreline of 1.18 m/year (based on two measurement locations, about 200 m apart). In a follow-up study, Williams (1993b) found one-year erosion rates averaging 0.69 m/year for this section of shoreline, based on 14 measurement locations spaced about 30 m apart. Williams (1993b) also reported that a small erosive scarp was present along most of this shoreline in 1993; this scarp was not observed in 2005.
[FIGURE 3 OMITTED]
The reason(s) for these changes are unknown; this section of shoreline appears to be stabilizing: on average, erosion has slowed considerably and no longer affects the entire section of shoreline. Shoreline accretion suggests that sediment trapping is occurring in places, presumably aided by the Spartina marsh. A possible explanation for diminishing wave erosion is that wave energy is being dissipated over a progressively widening wave-cut platform--a possibility suggested by Williams (1993a).
Section B. -- This section of shoreline consists of an area of marsh bordered by a prominent ca. 0.5-m-high cliff. With the exception of a small pocket of accretion at the western end, this shoreline has undergone considerable erosion between 1995 and 2005 (Figure 3). Maximum erosion is between 9.6 and 13.6 m (11.6 [+ or -] 2 m). The average change for the entire section is erosion between 4.9 and 8.9 m (6.9 [+ or -] 2 m; equivalent to 0.49 to 0.89 m/year). These findings are more in line with those of Williams (1993a; 1993b). Williams (1993a) calculated an average erosion rate of 1.04 m/year based on two measurement locations about 200 m apart. Williams (1993b) calculated an average one-year erosion rate of 1.26 m/year for this section of shoreline, based on five measurement locations spaced about 30 m apart.
Section C. -- This section of shoreline consists of a thin peninsula of land bordered by the concrete erosion control mat emplaced in 1996. Shoreline position along this section of shoreline is essentially unchanged between 1995 and 2005 (Figure 3). The separation between the 1995 and 2005 shorelines along all 13 transects located in this section is less than the 2 m measurement error. This is not a surprising result because the concrete mat forms the shoreline along this section and it has been in place nine out of the ten years of the study period. Williams (1993a) calculated an erosion rate of 0.9 m/year for this section of shoreline, based on one measurement location. Williams (1993b) calculated an average one-year erosion rate of 0.82 m/year, based on four measurement locations spaced about 30 m apart. Although it appears that this section of shoreline has stabilized and further erosion has been prevented, this result is misleading because field observations indicate that in several places soil underlying the concrete mat has been washed out, leaving the structure hollow and prone to collapse (Figure 4).
Section D. -- This section of shoreline consists of dense woodland bordered by a ca. 1.5 m cliff. Shoreline changes along this section are a mixture of erosion and accretion (Figure 3). Shoreline change is negligible along most transects, the measured erosion or accretion amount falling within the 2-m measurement error. The maximum erosion is between 2.5 and 6.5 m (4.5 [+ or -] 2 m). The maximum accretion is between 4.4 and 8.4 m (6.4 [+ or -] 2 m).
Williams (1993a) calculated an erosion rate for this section of shoreline of 0.61 m/year (based on one measurement location). Williams (1993b) found one-year erosion rates averaging 0.45 m/year, based on eight measurement locations spaced 15-30 m apart. These results suggest that erosion has slowed considerably on this section of shoreline and it has become relatively stable. The reason(s) for this change is unknown. A dense root mat was observed in the cliff face bordering this shoreline in 2005--it may be that the root mat is helping to strengthen and stabilize the cliff and retard the effects of wave erosion. The effects of a widening wave-cut platform may also contribute to a reduction of erosion at this site.
Section E. -- This 130-m-long section of shoreline consists of a gently sloping sandy beach. Approximately the western two-thirds of this shoreline underwent erosion between 1995 and 2005 (Figure 3). The maximum erosion was measured near the center of this section, where the shoreline retreated between 5.4 and 9.4 m (7.4 [+ or -] 2 m). Measurements along three transects within the eastern one-third of this shoreline indicate a small amount of accretion, but the amount of accretion is either below or close to the 2-m measurement error. The average change for the entire section is erosion between 1.3 and 5.3 m (3.3 [+ or -] 2 m; equivalent to 0.13 to 0.53 m/year). These findings suggest erosion has slowed along this shoreline in the last decade. Williams (1993a) calculated an average erosion rate of 1.06 m/year based on one measurement location. Williams (1993b) calculated an average one-year erosion rate of 0.88 m/year for this section of shoreline, based on seven measurement locations spaced 15-30 m apart. The reason(s) for the apparent slowing of erosion is unknown--it may also be the result of a widening wave-cut platform, has as been suggested for other shoreline sections where erosion appeared to have declined over the last decade.
[FIGURE 4 OMITTED]
Combining a sub-meter-accuracy Global Positioning System and a 1-m-resolution DOQQ to map shoreline change provided a level of point accuracy of [+ or -] 1 m and distance accuracy of [+ or -] 2 m. The results of the study suggest that shoreline change varies between different shoreline types and that rates of shoreline change over the last decade vary from those found by earlier studies. Overall, erosion appears to have declined within the study area, with erosion slowing in some places and shorelines becoming relatively stable in other places. Shoreline sections A and D, which displayed consistent erosion between 1943 and 1993, underwent relatively little change between 1995 and 2005 and appear to have stabilized. The reason(s) for this change is uncertain--it may be that a progressively widening wave-cut platform is dissipating wave energy along section A; whereas a dense woodland root mat may be retarding wave erosion along the shoreline bordering section D. Other shoreline sections where erosion has apparently slowed may also be showing the effects of a widening wave-cut platform. The shoreline bordered by the concrete erosion control mat showed negligible change in position between 1995 and 2005. Although the concrete mat has apparently stabilized this shoreline, this result is misleading because soil beneath the mat has been washed away in several places and tidal waters from the GIWW flow freely into Mad Island Bayou through the lattice structure of the mat. This will presumably increase saltwater intrusion into Mad Island Lake--an outcome the mat was designed to prevent.
Halsted, L. 2003. Texas Nature Conservancy Internal Document: Conservation Plan for Mad Island Marsh--Oyster Lake, from conservationonline.org/2003/07/h/CAP_Mad_Island.
Texas Natural Resource Information System (TNRIS). 1995. DOQQ Map: Palacios NW.
Theiler, R., D. Martin & A. Ergul. 2003. Digital Shoreline Analysis System (DSAS) version 2.0: An ArcView extension for calculating shoreline change. United States Geological Survey.
Williams, H. 1993a. Shoreline Erosion at Mad Island Marsh Preserve, Matagorda County, Texas. Texas J. Science, 45(4):300-309.
Williams, H. 1993b. Shoreline Erosion Monitoring Network, Mad Island Marsh Preserve, Matagorda County, Texas: First Year Results (1992-1993). The Nature Conservancy of Texas, Research Report, 12 pp.
HW at: firstname.lastname@example.org
Webster Mangham and Harry Williams
Trinity River Authority of Texas, 5300 South Collins Arlington, Texas 76004-0060 and Geography Department, University of North Texas Enton, Texas 76203
Table 1. Comparison of measurements made by tape and GPS. Measurement Tape (m) GPS (m) Difference (m) Stake 1 to shoreline 20.0 20.1 0.1 Stake 3 to shoreline 14.9 15.1 0.2 Stake 8 to shoreline 22.1 22.1 0.0 Stake 9 to shoreline 21.6 21.7 0.1 Stake 10 to shoreline 19.9 19.8 0.1 Stake 11 to shoreline 18.2 18.3 0.1 Stake 13 to shoreline 5.4 5.3 0.1 Stake 1 to stake 3 62.0 62.2 0.2 Stake 8 to stake 9 30.4 30.5 0.1 Stake 9 to stake 10 30.7 30.7 0.0 Stake 10 to stake 11 30.4 30.2 0.2 Stake 11 to stake 13 60.9 60.5 0.4 Mean 0.13
|Printer friendly Cite/link Email Feedback|
|Author:||Mangham, Webster; Williams, Harry|
|Publication:||The Texas Journal of Science|
|Date:||Feb 1, 2007|
|Previous Article:||Effect of human disturbance on the abundance and spatial distribution of the Atlantic Ghost crab (Ocypode quadrata) (Fabricius 1798) on a Texas beach.|
|Next Article:||Demise of an introduced population of the sailfin molly, Poecilia latipinna, in the upper Guadalupe River of central Texas.|
|Shoreline erosion at Mad Island Marsh Preserve, Matagorda County, Texas.|
|Recent records of the river otter (Lutra canadensis) along the Texas Gulf Coast.|