, 2008). Land cover/use layers from 1895, 1975, 1989, 2000, and 2010 were used at the largest scale of analysis, which encompassed Pool 6 and its floodplains, covering an area of 72.2 km2. The

1895 dataset from the Mississippi River Commission (USACE, 1895) was digitized by the USGS. The remaining land cover/use data sets were digitized by the USGS, with polygon interpretations based on photo overlays, EROS satellite imagery, aerial imagery, and color infrared imagery (http://www.umesc.usgs.gov/rivers/upper_mississippi/reach1/pool_6/p6_gis_data.html). For the 1931 Brown Survey (USACE, 1931), land/water features were digitized for this project. Details of the coding of these layers are in Freyer (2013). In this analysis, land contiguous with uplands Selleck Volasertib or levees was not distinguished from mid-channel islands. Within the P6 and using the same data sources, a Pool 6 Managed Channel (P6MC) area focuses analyses on the active channels, covering an area of 29.9 km2. Areas outside of the levees, railroads, and managed areas adjacent to the active channel such as docks and ports were excluded. The second scale of analysis encompassed 3.65 km2 of the lower portion of Pool 6 (LP6) from Lock and Dam

6 upstream to river mile 716.5. In addition to the above datasets, historical aerial photos (Table 1) were scanned R428 clinical trial and imported to ArcGIS. Methods for georeferencing and registering imagery were adopted from previous studies (Zanoni et al., 2008). Each of the images

was georeferenced from previously orthorectified Digital Orthophoto Quadrangles and color infrared mosaic images. Ten to twenty ground control points (GCPs) were identified on each aerial photo. GCPs were bridges, structures associated Lock and Dam 6, and road intersections adjacent to the river. The maximum acceptable RMS error value for this study was <1, giving ground measurements an average error of ±1 m. Final rectification employed a cubic convolution image resample (Zanoni et al., 2008), and emergent areas were digitized. RMS values represent only some of the errors that should be considered in GIS analysis of aerial photography; Hughes et al. (2006) and Day et al. (2013) provide more comprehensive treatments of this topic. Selecting only photos during that corresponded with normal pool elevations for post-dam datasets (643–645.35 ft) (Table 1) also minimized error. The LP6 area was divided into 10 sectors with boundaries chosen to minimize division of 1895 and 2010 contiguous land areas into multiple sectors (Fig. 5). Sectors 1 and 10 consist of land attached to the river banks, while sectors 2–9 consist of mid-channel islands. Four sets of bathymetric data surround an island group known unofficially as the Mobile Islands (Fremling et al., 1973). These datasets include 1897 soundings completed as part of the Mississippi River Commission Survey (USACE, 1895), the 1931 Brown Survey (USACE, 1931), a 1972 survey (Fremling et al.