Subtask: Late 20^th Century Landcover Change in the Northern Gulf Coast

Subtask Leader: Brady Couvillion - USGS National Wetlands Research Center, Baton Rouge, Louisiana

Aim

Spatial analysis of Landsat data from 1970s to today will reveal the landscape-level impacts of hurricanes on coastal vegetation, from both short-term damages to long-term marsh loss and recovery. [Enlargement]

Land cover provides baseline information when measuring the impacts of urbanization, wetland loss, climate change, sea-level rise, or other environmental stressors on the resilience of communities. Using satellite-derived land cover data, we can examine multi-decade evolution of landscape change using consistent methodology and contiguous data. In addition, a closer look at wetland trends in high temporal detail will further elucidate the effect of environmental factors on wetland loss in the northern Gulf of Mexico.

The applications of land cover data for NGOM scientific endeavors are numerous, from landscape spatial modeling in Subtask 4.3 to land-loss time series analysis in Subtask 3.5. These data will provide a significant baseline data set for decadal landscape evolution for the entire NGOM study area and marsh loss studies for coastal Louisiana. A second, coupled study focusing on land cover changes and wetland impacts due to historic hurricanes will complement these analyses and will further support NGOM research in Task 4, particularly subtasks dealing with catastrophic disturbances and landscape modeling.

The goal of this subtask is to identify and document land-cover changes within the NGOM project area by acquiring and classifying historical Landsat satellite imagery using a standardized project classification scheme on a decadal basis from the 1970s to the present. Annual to biannual classified Landsat TM land-water data sets will be used to refine trend timing and variance over the NGOM landscape. In addition, Landsat data and other remotely-sensed data sets will be examined to understand the impact of hurricane-induced landscape changes as well as changes from other episodic events. The use of wet marsh as a potential stress or loss indicator will be investigated for selected areas.

Specific objectives of this subtask include:

• Building on existing data sets to develop land cover data spanning 1970s to present using Landsat TM and MSS and conduct land cover change analyses.
• Expand the spatial and temporal frequency, as well as the thematic resolution of wetland data and examine the variance in wetland type and area under various conditions.
• Examine wetland loss and recovery following hurricanes and tropical storms.
• Examine high resolution data as a means of training moderate resolution classifiers to increase the content and utility of moderate resolution datasets.

Approach

Nationally recognized classification methodologies, such as those utilized in NLCD and C-CAP, were altered slightly to provide increased value, yet retain consistency with these programs. Much of the increased value results from more specific land cover classes with regard to wetlands. 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 datasets will provide managers with an increased ability to inventory resources, understand past trends, and subsequently, draw inferences and project patterns.

Late 20th Century Land Use/Land Cover Changes in the Northern Gulf Coast. [Enlargement]

Accomplishments

• Land cover data for the NGOM study region has been created for the entire study area for 2008, 2006, and 2004. These classifications have obtained accuracies similar to those of CCAP and NLCD despite an increase in the number of categories classified.
• Land cover data has been created for the Mississippi and Alabama portions of the study area for 2008, 2006, 2004, 2000, 1996, 1992, 1988, and 1984. Change analyses of these data are currently taking place.
• Land cover data is being utilized to feed modeling applications, as well as foster improvements to other data sets such as topography.

Publications

Manuscripts

Abstracts and Presentations