|
Poster Abstracts: Gulf Coast Science Conference
8: Northern Gulf of Mexico (NGOM) Project
Late 20th Century Land Use/Land Cover Changes in the Northern Gulf Coast
Brady R. Couvillion1 and J.A. Barras2
1 IAP World Services, National Wetlands Research Center, Lafayette, Louisiana
2 USGS National Wetlands Research Center, Coastal Restoration Field Station, Baton Rouge, Louisiana
Land cover is an invaluable data set when evaluating the impacts of urbanization, wetland loss, climate change, sea-level rise, and other environmental stressors on the resilience of ecosystems and communities. Land cover change detection requires examination of multiple land cover datasets over a period of time using consistent methodology. Although several national programs have developed land cover data sets, they lack the period of record and/or specificity of cover classes necessary for such in-depth examinations. This is particularly true of land cover classes for coastal wetlands and forests. Recent events including hurricanes Katrina and Rita have highlighted the importance of accurate and detailed land cover data in these environments for use in decision support systems. These events have also emphasized the importance of accurate baseline assessments to initiate projections of future landscape changes. In the current study, we have created a 2006 land cover data set for the northern Gulf of Mexico region with increased specificity in coastal wetland classes. Nationally recognized classification methodologies, such as those used in NLCD and C-CAP, were refined slightly to provide increased land cover class discrimination, yet retain consistency with the original classification scheme. This data set will provide managers with an increased ability to more accurately inventory resources. Future efforts of this task will use this same methodology to create land cover datasets for multiple time periods from 1983 through 2008. These datasets will greatly increase managers understanding of trends in land cover, as well as the ability to draw inferences and project patterns.
Contact Information: Brady Couvillion, IAP World Services, 700 Cajundome Blvd, Lafayette, LA 70506. phone : 225 578 7486; email: bcouvillion@usgs.gov
Recent Reductions of Subsidence Rates in the Mississippi River Delta Plain
Julie C. Bernier and Robert A. Morton
U.S. Geological Survey, Florida Integrated Science Center, St. Petersburg, Florida
The Mississippi River delta plain has long been characterized as an area with high rates of relative sea-level rise. This concept was tested by integrating National Ocean Service tide-gauge records with National Geodetic Survey benchmark releveling data and GPS elevations at Continuously Operating Reference Stations, providing a basis for understanding historical subsidence trends and most recent rates for southeastern coastal Louisiana. Tide-gauge records indicate that rates of relative sea-level rise at Grand Isle accelerated from about -2.2 mm/yr between 1947 and 1964 to about -11.5 mm/yr between 1964 and 1991 and then decelerated to about -3.4 mm/yr between 1991 and 2006. These trends and rates are independently verified by repeat leveling surveys that yielded average subsidence rates of -9.6 and -11 mm/yr from 1965/66 to 1993 at benchmarks between Raceland and Grand Isle and between Houma and Cocodrie, respectively, and GPS-derived elevation changes at Boothville, Houma, and Cocodrie that yielded average subsidence rates of -3.5 to -6.3 mm/yr from 2002/03 to 2007. The most recent slow rates of subsidence are similar to those averaged over geological time scales (e.g., radiocarbon-dated peats) that are attributed to natural sediment compaction and crustal loading.
The historical pattern of slow, then rapid, then slow subsidence may be caused by natural deep-basin processes (e.g., faulting, salt migration) but is more likely related to regional hydrocarbon production that followed the same general temporal trends. If accelerated subsidence was induced by reservoir compaction and fault reactivation associated with fluid withdrawal that also accelerated in the 1960s and 1970s, then the most recent reductions in subsidence rates likely reflect a balancing of subsurface stresses and a return to near-equilibrium conditions. Understanding historical and current trends in subsidence rates and their causes is critical for designing and successfully implementing coastal-restoration activities and for modeling and predicting expected impacts of relative sea-level rise on the delta plain.
Contact Information: Julie C. Bernier, U.S. Geological Survey, Florida Integrated Science Center, 600 4th Street South, St. Petersburg, FL 33701; phone: 727-803-8747; email: jbernier@usgs.gov
Observations of the Spatial and Temporal Distribution of Hurricane-induced Land Loss in Coastal Louisiana Over the Latter Half of the 20th Century
John A. Barras, U.S. Geological Survey, National Wetlands Research Center, Baton Rouge, Louisiana
A combination of historical aerial photography and Landsat Thematic Mapper (TM) satellite imagery was used to identify hurricane-induced land loss in coastal Louisiana marshes from 1956 through 2005. Hurricane magnitude, track, and landfall information obtained from the National Oceanic and Atmospheric Administration (NOAA) were used to identify candidate storms. Landfall bracketing TM imagery and photography were then examined to identify probable storm-formed or storm-expanded water bodies. Most observed loss was related to the removal or partial removal of marsh vegetation by storm surge or to shoreline erosion caused by enhanced wave action. The satellite’s high temporal frequency was useful for identifying storm-induced land loss occurring from 1983 and 2005 although its moderate spatial resolution limited minimal detectable storm-formed water bodies to greater than 2 acres in size. Bracketing images were often obtained within months to weeks of a storm’s landfall. The TM imagery was successfully used to identify loss caused by Hurricanes Andrew (Aug. 26,1992), Lili (Oct. 3, 2002), Ivan (Sept. 16, 2004), Katrina (Aug. 29, 2005), and Rita (Sept. 4, 2005) and Tropical Strom Isadore (Sept. 26, 2002). The same techniques were applied to historical aerial photography to identify land loss caused by Hurricanes Audrey (June 27, 1957), Hilda (Oct. 3, 1964), and Betsy (Sept. 9, 1965). The photography lacked the temporal and spatial coverage of the TM imagery but was adequate for identifying historical hurricane-induced land loss.
Detectable hurricane-induced land loss increased with storm magnitude. Hurricane Audrey, a category 4 storm that made landfall near the Louisiana-Texas border, caused probable land loss 260 km east into western Terrebonne Parish. Category 2 or lesser storms caused detectable localized loss within 100 km east of landfall. Land loss magnitude and spatial distribution was greatest immediately east of storm landfall and then decreased eastward. Storm-induced land loss decreased immediately to the west of storm landfall implying most identifiable land loss was caused by storm surge rather than wind. Consecutive storm landfalls caused commingled land loss patterns of varying magnitude and spatial distributions consisting of new ponds and expanded ponds, some of which have remained in place since Hurricane Audrey’s landfall in 1957.
These observations suggest that hurricanes have and will continue to contribute significantly to coastal land loss and that future coastal restoration activities should account for both past and future periodic hurricane impacts.
Contact Information: John A. Barras, U.S. Geological Survey, National Wetlands Research Center, Baton Rouge, LA 70894; phone: 225 578 7486; email: barrasj@usgs.gov
The Development of Oyster Beds in Apalachicola Bay, FL during the late Holocene
D. Twichell1, L. Edmiston2, B. Andrews1, W. Stevenson3, J. Donoghue5, R. Poore4
1 U.S. Geological Survey, Woods Hole, Massachusetts
2 Apalachicola Bay National Estuarine Research Reserve, Apalachicola, Florida
3 NOAA Coastal Services Center, Charleston, South Carolina
4 U.S. Geological Survey, Florida Integrated Science Center, St. Petersburg, Florida
Apalachicola Bay harbors the largest oyster fishery in Florida, and the development and distribution of its oyster beds are products of late Holocene geologic evolution and subsequent estuarine conditions. The bay is shallow, having an average depth of approximately 2.5 m. Salinity is variable, but averages 6-20 ppt where oyster beds are found. Currents are primarily driven by tides, but are strongly affected by winds and river discharge. Net water movement is to the west, and velocities in the bay rarely exceed 0.5 m/sec, except near inlets.
Sidescan-sonar imagery, bathymetry, high-resolution seismic profiles, and sediment cores have been used to map the geologic evolution of a large part of Apalachicola Bay and St. George Sound. These data show that oyster beds occupy the crests of a series of shoals that are mostly 1-7 km in length, trend roughly north-south perpendicular to the long axis of the bay, and are asymmetrical with steeper sides facing to the west. Surface sediment samples show that the oyster bars consist of shelly sand, while much of the remainder of the bay floor is covered by mud delivered by the Apalachicola River. High-resolution seismic profiles and cores show two delta systems that advanced southward across the bay during the late Holocene when sea level was lower. Radiocarbon dates indicate that they initiated between 7,000 and 5,600 yr BP when sea level was 4-7 m lower than present. They were abandoned about 4,000 yr BP when deltaic deposition retreated landward of its present location in response to sea level rise. Oysters started colonizing sandy parts of these late Holocene deltas between 1,200 and 1,500 yr BP. Seismic profiles and cores indicate that oyster beds have been more extensive in the past, but some have been buried by the muddy prodelta deposits of the modern Apalachicola delta. Oyster bars that are still active have grown vertically and become asymmetrical, and internal bedding indicates they have migrated westward, presumably in response to the net westerly currents. The lithologic matrix of some oyster bars fines upward, suggesting that the sediment available for their development is becoming finer with time. Whether the increasingly limited availability of coarse-grained sediment will lead to eventual demise of the oyster bars is unknown.
Contact Information: Dave Twichell, U.S. Geological Survey, Woods Hole CMG Field Station, 384 Woods Hole Rd, Woods Hole, MA 02543; phone: 508 457 2266; email: dtwichell@usgs.gov
High-Resolution Coastal Land-Cover Classification in the Northern Gulf of Mexico
Amar Nayegandhi1, Joyce Fry2, Christine Kranenburg1, and John Brock3
1 Jacobs Technology, Inc., Florida Integrated Science Center, St. Petersburg, Florida
2 SGT, contracted to U.S. Geological Survey, Earth Resources Observation and Science Center (EROS), Sioux Falls, South Dakota.
3 U.S. Geological Survey, Coastal and Marine Geology Program, National Center, Reston, Virginia
The northern Gulf of Mexico landscape contains about 30 percent of the Nation’s wetlands, but accounts for 90 percent of annual wetland losses. Several remote-sensing-based studies in the northern Gulf of Mexico region have been conducted to refine quantification of these losses. Most of these studies have been based on moderate resolution (30-m) Landsat imagery. Although Landsat imagery provides excellent results at a reasonable cost for large-scale mapping, the land-cover classification map from a 900-m2 Landsat pixel is not designed to capture local change from catastrophic storm events, such as Hurricanes Katrina and Rita, or small-scale anthropogenic influences. The objectives of this research are two fold: 1) to use high-resolution Quickbird (2.4-m multispectral) imagery to improve the specificity of the current Landsat-based post-Hurricane Katrina land-cover classification map and 2) to investigate the effects on classification accuracy by combining advanced image-segmentation techniques and high-resolution Lidar-derived elevation data with established National Land Cover Database (NLCD) classification methods. Three pilot regions were chosen in the coastal regions of Louisiana and Mississippi: an ~776-km2 region in the Delacroix river basin, an ~346-km2 region covering the lower Pearl River basin, and an ~75 km2 region encompassing the Barataria Preserve at Jean Lafitte National Park. The land-cover classification maps developed for the three study areas had a final target spatial resolution of 5 m and a suitable classification legend that represents the biophysical structure of wetland and coastal areas.
Contact Information: Amar Nayagandhi, Jacobs Technology, Inc., 600 4th Street S, St. Petersburg, FL 33701; phone: 727 803 8747; email: anayegandhi@usgs.gov
Land Area Change in Coastal Louisiana: 1956 to 2006
John A. Barras1, Julie C. Bernier2, Robert A. Morton3
1 U.S. Geological Survey, National Wetlands Research Center, Baton Rouge, Louisiana
2 U.S. Geological Survey, Florida Integrated Science Center, St. Petersburg, Florida
3 U.S. Geological Survey, Austin, Texas
The USGS analyzed changes in land and water area in coastal Louisiana from 1956 to 2006 using a series of 14 sequential datasets. These datasets provide a spatially and temporally consistent source of land area information by physiographic province for coastal Louisiana. Land-area changes were interpreted using spatial analysis and linear regression. The spatial analysis quantified and spatially depicted coastwide land-area changes for five time periods: 1956 to 1978, 1978 to 1990, 1990 to 2001, 2001 to 2004 and 2004 to 2006. Linear regression analysis used data derived from a single source (classified Landsat TM imagery) to provide a robust estimate of recent land-area change from 1985 to 2004, prior to the 2005 landfalls of Hurricanes Katrina and Rita, and from 1985 to 2006.
Total net land loss for coastal Louisiana from 1956 to 2006 was 3,493.9 km2. Annual land-loss rates were highest from 1956 to 1978 at 101.7 km2 /yr and accounted for 64% of the total land loss over the entire 50-year time period. Land loss occurring from 1978 to 2004 accounted for 21% of the total land lost during the 50-year time period. The 512.8 km2 potential loss from the 2005 hurricanes accounts for 14.7% of the total loss from 1956 to 2006 and for 40.7% of the total loss from 1978 to 2006. The annual loss rate, based on linear regression analysis, declined between 1985 and 2004 to 30.7 ± 5.7 km2/yr.
The majority of land loss from 1985 to 2004 occurred in the Deltaic Plain and was slightly offset by small gains in the Marginal Deltaic Plain due to land gains in the Atchafalaya River and Wax Lake Deltas. The Chenier Plain was stable during this period. The rapid decrease in land loss rates after the 1970s confirms findings of past studies although the rates observed in this study are lower. The permanency of the 2005 hurricanes’ contribution to land loss is uncertain. Initial observations suggest that large hurricanes can contribute significantly to coastal land loss and that these significant episodic events can alter the long-term, time-averaged trends of landscape change. Prior trend assessments may lack the temporal resolution to identify the land loss contributions of past episodic events. Trend projections based on simple extrapolation of current rates may not adequately account for potential future episodic events.
Contact Information: John A. Barras, U.S. Geological Survey, National Wetlands Research Center, Baton Rouge, LA 70894; Phone: 225 578 7486; email: barrasj@usgs.gov
Holocene Evolution of Apalachicola Bay, Florida
Lisa E. Osterman1, David C. Twichell2, Richard Z. Poore1
1 U.S. Geological Survey, Florida Integrated Science Center, St. Petersburg, Florida
2 U.S. Geological Survey, Woods Hole, Massachusetts
In order to understand the Holocene evolution of Apalachicola Bay in Florida better, a program of geophysical mapping was followed by sedimentological and faunal analyses of vibracores. Ten vibracores, collected in two to four meters of water, contain chronologies provided by 34 AMS 14C dates on shells and wood. As sea level rose after the last glacial maximum, fluvial deposits filled in the Apalachicola River paleo-channel, which extended southward across the central part of the bay. Sediment cores to either side of the paleo-channel contain abundant wood fragments documenting forested areas on the shelf at 8,000 14C years. Marsh sediments with agglutinated foraminifers and diatoms overlie the wood-bearing sediments and indicate that the initial marine transgression occurred at about 7,000 14C years. After the initial flooding of the area presently covered by the bay, the Apalachicola paleo-delta readvanced onto the inner shelf west of the filled paleo-channel after 7,000 14C years. Later, a shift in the path of the river system allowed the establishment of an eastern deltaic lobe at 5,600 14C years.
The timing of barrier-island formation is based on establishment of estuarine conditions in Apalachicola Bay. Estuarine benthic foraminiferal assemblages occurred as early as 6,500 14C years in the western bay and after 5,200 14C years in the eastern bay. The faunal assemblage supports the interpretation that the barrier islands developed from sand reworked from the paleo-delta lobes and expanded first in the western bay and later in the eastern bay. In addition, at the base of two cores, open-marine benthic foraminifers record an older marine highstand that indicates the absence or more landward occurrence of an earlier barrier-islands system.
Contact Information: Lisa E. Osterman, U.S. Geological Survey, Florida Integrated Science Center, 600 Fourth St. South, St. Petersburg, Florida 33701; phone: 727-803-8747 x 3084; fax: 727-803-2032; email: osterman@usgs.gov
Unsupervised classification of vegetation communities using airborne LiDAR data at Naval Live Oaks Reservation, Florida
Monica Palaseanu-Lovejoy1, Amar Nayegandhi1, John Brock2 and David Nagle1
1 Jacobs Technology, Inc., Florida Integrated Science Center, St. Petersburg, Florida
2 U.S. Geological Survey, Coastal and Marine Geology Program, National Center, Reston, Virginia
This study evaluates the capability of Experimental Advanced Airborne Research Lidar (EAARL) to delineate vegetation communities in Naval Live Oaks Reservation (NLO), Florida using unsupervised k-means clustering. Five-meter-resolution grids of bare earth (BE), canopy height (CH), canopy-reflection ratio (CRR), and height of the median energy (HOME) were derived from spatially dense EAARL data acquired in 2005 and 2007.
The EAARL is a temporal waveform green-laser (532 nm), small-footprint airborne lidar instrument. To describe the vertical structure of vegetation, several individual small-footprints are combined to make a composite “large-footprint” waveform. The size of the composite footprint is a variable determined in the post-flight processing software. BE is derived directly from the small footprints, while CH, CRR and HOME are derived from the composite footprints. CRR represents a relative measure of the canopy closure, while HOME is sensitive to changes in vegetation structure and the degree of canopy openness.
A principal component analysis (PCA) of the four lidar metrics was performed and the k-means clustering was conducted on the PCA components weighted by their respective proportion of explained variances. A majority filter was applied on the vegetation classification map in order to increase class cohesiveness. Both filtered and unfiltered classifications were compared with the color infrared imagery obtained during same EAARL flight mission to assess the reliability of the vegetation classification categories.
Contact Information: Monica Palaseanu-Lovejoy, Jacobs Technology, Inc., Florida Integrated Science Center, 600 4th Street S, St. Petersburg, FL 33701; phone: 727 803 8747; email: mpal@usgs.gov
John Brock, U.S. Geological Survey, 12201 Sunrise Valley Dr., Reston, Virginia 20192, jbrock@usgs.gov
|
URL of this page is: http://ngom.usgs.gov/gomsc/meetings/gcsm2008/abstract_gom_poster8.html