From: hogrefek@geo.oregonstate.edu Date: February 5, 2009 4:01:43 PM PST To: Joyce Miller Cc: Scott Ferguson , John Rooney , Jonathan Weiss , Frances Lichowski , Tomoko Acoba , Dawn@science.oregonstate.edu, Deepsea Subject: Tinian Deliverable LiDAR Update In the zip file you will find: 1 MS Word file: Tin_ErrorAnalysis_020509 3 ESRI Grid files (in their own folder): TinLiDBallmos: final mosaic of 6 derived bathy products from three images. TinLiDBallerr: error grid, Error = LiDAR depth - derived depth TinMBLiDBmos: mosaic of the multibeam sonar grid, LiDAR grid and the DB mosaic. 2 Text (metadata) files: TinLiDBallMos_metadata TinMBLiDBmos_metadata 14 PDF files: Tin233_LiDB: Image Tin233, product DB - derived with original MLR variables. Tin233_LiDB1: Image Tin233, product DB1 - derived with reduced Y-int value. Tin233_LiDBDB1_mos: Image Tin233 mosaic - data from DB given priority. Tin29700_LiDB: Image Tin29700, product DB - derived with original MLR variables. Tin29700_LiDB4: Image Tin29700, product DB3 - derived with reduced slope values. Tin29700_LiDBDB4_mos: Image Tin29700 mosaic - data from DB given priority. Tin29701_LiDB: Image Tin29701, product DB - derived with original MLR variables. Tin29701_LiDB1: Image Tin29701, product DB1 - derived with reduced Y-intercept. Tin29701_LiDBDB1_mos: Image Tin297-01 mosaic - data from DB given priority. Tin_LiDBall_mos: Mosaic of the DB mosaics from all images (file: TinLiDBallmos) Tin_LiDBall_error: Error grid (file: TinDBallerr) Tin_MBLiDB_mos: Mosaic of multibeam,LiDAR and derived bathy (file: TinMBLiDBmos) Tin_MB_AddLicoverage: Basically a MB/LiDAR mosaic with the LiDAR bathy highlighted Tin_MBLi_AddDBcoverage: Basically a MB/LiDAR/DB mosaic with the derived bathy highlighted BIG(ish) news: I added a third section to the error analysis to accentuate a pretty cool additional outcome to working with the LiDAR. Nutshell: One wouldn't guess it from looking at the stats from the last Tinian deliverable, but it appears that the bathymetry derived using the multibeam sonar is almost as statistically valid as the bathymetry derived using the LiDAR as a baseline - even with the (shallow) depth limitations of the multibeam.  I used the LiDAR dataset to extract LiDAR depth and depth values derived from multibeam (last deliverable) everywhere the datasets overlapped (usual Part 2 error assessment) and the stats are only slightly inferior to those for the LiDAR derived bathy.  Implication - even with a baseline dataset of limtted spatial/depth coverage into the less than 15m range and no data less than 8m (like Tinian), this bathy derivation method can be depended upon to produce quite accurate data (though in most cases the more extensive coverage needed to test this accuracy will be lacking).  Might be worth another paper ...