Marsh Die-back remote sensing update –
Infrastructure
The UGA Marine Extension service has spent nearly $13,000.00 on image processing infrastructure in light of the marsh die-off problem. In particular, we have bought a mass storage device with a total allowable capacity of 480/320 gigabytes to hold the imagery. In addition, we have purchased a dedicated Dell Workstation with an additional 55 gigabytes of space and a DVD burner, running both ArcView and Erdas Imagine applications. The latter is the program that we will use for the image processing. Finally, we bought a new large format (42-inch), high quality plotter to print hard copy maps and images of the affected areas that can be taken in to the field for verification.
Mass storage device $3679.20
Dedicate IP machine $1548.60
Large format printer $7567.00
Software licenses $200.00
Delineating dead marsh: Priceless.
Total infrastructure expenditure $12994.80
Georgia DNR
has purchased a state-wide set of the 1999 CIR DOQQs that we will use as the
reference (pre-dieback) imagery. Aimee
Gaddis and I will go to either the
Sapelo – Aimee Gaddis and I have been working with two graduate students from the Geography department to scan the hard copy photos and rectify them to work with other datasets. All of the photos have been scanned (as TIFF images) and then converted to the native format for Erdas Imagine. A total of 119 photos were scanned, each about 380,000 kb in size totaling 43 gigabytes. The rectification process is underway. Merryl Alber has a new graduate student, Sylvia Schaefer, who may step in and help with the processing.
Remainder of the coast – Jan McKinnon carefully selected and ordered a set of five color infrared and five true color images from a 2002 flight by the National Geodetic Survey. We will use these images to test our preliminary protocol for processing and then order the imagery for the remainder of the coast. We anticipate that there will be a trade-off made depending on which imagery we choose. Because all of the imagery was acquired in January of 2002, when the healthy marsh is senescent it may be best to use true color imagery in order to most accuracy map spatial extent. However, estimates of biomass or leaf-area indices are generally made using infrared film; however, again this point may be moot as the imagery was acquired in winter.
We will examine elevation models that are available for the coast to see if they can help refine mapped areas of dieback but most of the available models are too coarse for the problem at hand, or don’t cover areas far enough inland to cover the areas in question.
I have a short power point presentation that describes the protocol if anyone is interested they can contact me. In short: we will map the extent of die-off, which we define as both standing dead and denuded areas. We will use the field monitoring data and helicopter flights to assess the accuracy of the mapped extents and we will create a dataset of docks in the process. The accuracy of the dock dataset will be assessed in a joint project with Clarke Alexander at the Skidaway Oceanographic Institute.
-- (Karen Payne)