Monitoring coastal wetlands and near shore aquatic environments in response to the BP Horizon oil spill

Susan L. Ustin, University of California Davis, slustin@ucdavis.edu (Presenter)
Shruti Khanna, University of California Davis, shkhanna@ucdavis.edu
Alexander Koltunov, University of California, Davis, akoltunov@ucdavis.edu
Dar Alexander Roberts, UC Santa Barbara, dar@geog.ucsb.edu
Raymond Kokaly, US Geological Survey, raymond@usgs.gov

Globally, coastal wetlands and aquatic environments are biologically rich, highly productive ecosystems that provide critical habitat for both aquatic and terrestrial organisms, many of which are of significant economic and recreational importance, yet they are subject to frequent disturbance from weather, floods, and human activities, the long term impacts of which are poorly understood. There is a critical need to understand short and long term ecosystem and community responses to oil spills, recovery patterns and species sensitivities, as spills are becoming more common along coastal margins. Brackish and salt marshes are among the most vulnerable coastal ecosystems to oil spill impacts because oil tends to have a much long residence time compared to other environments. Currently, the United States Gulf of Mexico coastline supports ~50,000 wells (~30,000 are decommissioned or abandoned), that are frequently sources for oil spills. Using AVIRIS hyperspectral data flown over the Deepwater Horizon Gulf Oil Spill in July and September of 2010, we tested the impacts on the health and structure of the wetland. Everywhere oil came ashore it killed the macrophyte vegetation in the intertidal zone soon after contact. Oil in/on the vegetation can be identified because of the distinctive absorption bands across the VNIR-SWIR spectrum. Based on field surveys, the interior high marsh appeared to be mostly unaffected. We used AVIRIS spectral absorption methods on analysis of the September 2010 data to map and quantify pigment and water patterns interior to the dead zone in the intertidal. Multiple lines of evidence show that shorelines exposed to oil were significantly different than shorelines not exposed to oil. The impact of oil was significant at least 10-12m inland from the shoreline but not deeper because storms did not bring the tide farther inland. In the July dataset, near the time when the oil reached landfall, the effect of oil stress was less pronounced. A comparison of the green vegetation fraction between July and September showed no significant difference indicating that there was no significant loss of wetland area between July and September. This study illustrates the use of hyperspectral remote sensing in detecting ecosystem stress and monitoring recovery after a catastrophic event, such as an oil spill.

Presentation Type:  Plenary Talk

Session:  Poster Speed Talks:

Presentation Time:  Wed 9:55 AM  (5 minutes)