Introduction

Under the broad theme of habitat loss and fragmentation, Dr. Patrick Biber and his group in the Department of Coastal Sciences at the Gulf Coast Research Laboratory have been conducting a historical change study of seagrasses in the Mississippi and Chandeleur Sounds. Seagrasses are an important submerged aquatic resource that provides key ecosystem services yet are subjected to numerous stressors in the coastal environment.

Seagrass mapping data from a multitude of previously completed projects in the region, including field surveys, herbarium records, remote sensing, and GIS, were gathered and combined to provide information on seagrass change from 1940 to 2011. The study area generally lost seagrasses over the 71-year period in terms of both species richness and areal coverage, ostensibly due to reduction of protective island barriers and declines in water quality. The research results also indicated that comparison of seagrass area among various studies that used different mapping methods can result in overestimation of area change and potentially misleading conclusions about habitat loss.

Subsequent data analysis proceeded by utilizing theories and techniques from landscape ecology to quantify patterns and trends of seagrass habitat loss and especially landscape fragmentation in the Sounds. Fragmentation of seagrass landscapes has been mentioned in only a few studies to date mainly due to data limitations, however, the time series of fine-resolution maps (example above is image of Horn Island in 2003 with seagrass patches highlighted in red) made it possible not only to interpret fragmentation pattern at one snapshot but also to analyze changes in fragmentation processes over the last seven decades. Different patch-based and landscape-based metrics from Fragstats showed that seagrasses in the Mississippi and Chandeleur Sounds underwent habitat loss and fragmentation.

A synthesis of these findings were used to construct a model linking seagrass landscape spatial pattern with temporal dynamics using Bayesian inference. A hierarchical Bayesian model using Markov chain Monte Carlo method was used to quantify uncertainties, assimilate data covariances, and incorporate expert knowledge, which would have been impossible with traditional statistics. Overall, this study demonstrated the need to approach the multiple-source data from different angles to obtain a comprehensive view of seagrass landscape change in the Mississippi and Chandeleur Sounds, which in turn will be helpful for natural resource management and conservation of this changing habitat.