As eustatic sea level rise accelerates, coastal salt marsh ecosystems are increasingly vulnerable; vertical accretion rates must exceed or keep pace with rates of sea level rise to prevent transition to open water and inland migration of marsh vegetation. While some marsh systems have remained stable, the marshes of the Louisiana Gulf Coast have experienced high rates of conversion to open water. This study examines the pattern of historical salt marsh coverage at the mouth of the Pascagoula River in Jackson County, MS between 1996 and 2014. The goals of the study were 1) to determine whether marsh coverage has decreased over the study period and 2) to determine whether marsh vegetation migrated inland in response to local sea level rise. Color-infrared and textural image data from February 12, 1996, and October 5-16, 2014 were used to map coverage in each year at a 3-meter spatial resolution. Waterways were extracted using a near-infrared band threshold, and land cover was classified into three groups – marsh, woodland, and unvegetated – using a Maximum Likelihood Classifier. An in-situ 2016 vegetation survey provided reference data. The study area was divided into four zones of equal distance from north to south for change detection analysis.
Results/Conclusions
Change detection analysis revealed a total net marsh loss of 4.96% (280.27 ha) between 1996 and 2014. While every zone exhibited a net loss in marsh coverage, percent loss was greatest in Zone 2, where coverage decreased by 9.02%. This zone includes the central portion of the marsh just south of the Escatawpa and Pascagoula River confluence. The shoreline proximal to marsh vegetation in this area is subject to a higher degree of development than the other three zones. Loss in Zone 2 was dominated by conversion to open water, with 148.35 ha of marsh converted to open water between 1996 and 2014. Marsh loss was also high in Zone 4, the most inland zone, at 6.83%. However, this is due to an increase in woodland land cover rather than conversion to open water; 116.36 ha of marsh in this zone became woodland between 1996 and 2014. Zones 1 and 3 exhibited lower rates of loss at 0.16% and 2.07%. Accuracy was assessed using ENVI’s Confusion Matrix (99.88% for 2014 and 99.63% for 1996). Overall, the data suggest that the extent of marsh coverage will continue to decrease in this area rather than remaining stable or migrating inland.