Graduation Year

2018

Document Type

Dissertation

Degree

Ph.D.

Degree Name

Doctor of Philosophy (Ph.D.)

Degree Granting Department

Geography, Environment and Planning

Major Professor

Ruiliang Pu, Ph.D.

Committee Member

Susan Bell, Ph.D.

Committee Member

Barnali Dixon, Ph.D.

Committee Member

Joni Firat, Ph.D.

Committee Member

Graham Tobin, Ph.D.

Keywords

Landscape Metrics, Remote Sensing, Spatial Scale

Abstract

Florida’s seagrasses are ecologically important marine environments which have suffered major degradation caused by increasing anthropogenic pressures. A 2011 seagrass die-off event caused by an algal bloom in the Florida Indian River Lagoon (IRL) was particularly severe with a majority of seagrass lost in areas such as the Banana River. An understanding of how this coastal marine environment changed is an important step toward better managing resources for conservation. Modern tools and methods provide new opportunities to study these changes at the landscape scale, a scale that informs on the larger more comprehensive state of a system. Classified satellite imagery and spatial landscape metrics were used to quantify changes in IRL Banana River seagrass landscape patterns following the die-off event.

Thirty-six landscape metrics in four categories (Area-Edge, Shape, Core Area and Aggregation) were used to discern the spatial complexities of habitat changes over space and time in the IRL study area. Seagrass loss from 2011 to 2013 based on image classifications was as high as 91% in the Banana River study areas. Landscape metrics indicate that following the seagrass die-off in the IRL, meadows became more fragmented, patches became more isolated, and the amount and spatial complexity of meadow edge was reduced. For the most part, these landscape structural changes in the IRL increased with more severe amounts of seagrass loss.

The metrics were evaluated and scored for their effectiveness in detecting seagrass landscape changes and their ability to provide consistent detection with variable resolution imagery. The top metrics in order of highest evaluation score were Total Edge, Splitting Index, Total Core Area, Effective Mesh Size, Landscape Shape Index, Edge Density, Perimeter-Area Ratio Distribution, Average Core Area, Disjunct Core Area Distribution Mean and Patch Shape Index. Area-edge and aggregation type metrics were identified as the best metrics for evaluating landscape changes under different degrees of seagrass loss in the IRL. Landscape metrics applied to classified images have the ability to provide quantitative and informative techniques for monitoring seagrass health.

Share

COinS