Home

1 second SRTM Derived Products User Guide

1. The water and void fill masks for the 1 second DSM were also applied to the DEM Further information is provided in the User Guide Geoscience Australia and CSIRO 2011 Positional accuracy The horizontal positional error is the same as for the raw SRTM 1 second data with 90 of tested locations within 7 2 m for Australia See Rodriguez et al 2006 for more information Attribute accuracy Accuracy was tested on the 1 second DEM using 1198 Permanent Survey Marks distributed across the Australian continent relative to the Australian Height Datum AHD71 Results of this comparison show the absolute accuracy of the data as tested relative to AHD71 to be 7 582 m at the 95 percentile with a RMS error of 3 868 in open flat terrain Ninety nine percent of points are within a height difference of less than 9 602 m The removal of striping artefacts improves the representation of the landform shape particularly in low relief areas but it is not clear whether this also produces an improvement in overall height accuracy Some striping remains in the data at a much reduced level mostly less than 0 3 m amplitude Additional artefacts including long wavelength 10km striping have not been corrected The removal of vegetation offsets provides a significant improvement in the representation of the landform shape particularly in low relief areas and areas of remnant vegetation Elevation accuracy varies in forested areas Comparisons with
2. A d ES E A a Figure 27 Typical noise in 1 second SRTM DEM Arthur River near Narrogin WA 117 38E 33 055 Elevation range 270 350 m One second SRTM Derived Digital Elevation Models User Guide v1 0 4 Page 38 of 106 Incomplete removal of vegetation offsets In some areas offsets due to trees have not been completely treated Some patches are completely untreated because they were not mapped as trees in the vegetation mapping while other areas are partially treated due to poor estimation of vegetation height offset One example is shown in the vegetation removal section above Actively managed forests such as pine plantations are particularly subject to this problem because the contrasts in height are obvious and mismatches in the date of the forest cover mapping create substantial errors Figure 28 shows a pine forest area in south west Victoria where some patches have been adequately treated but many obvious offsets remain This is one of the worst examples of this effect and most instances are much subtler than this In low relief areas residual vegetation offsets will significantly affect measures of local shape such as slope aspect flow direction and curvature Figure 28 Pine forests in Victoria not completely removed 140 959E 37 866S Elevation range 30 80 m Vegetation height over estimated In a small number of areas the offsets due to trees have been over estimated Figure 29 shows the lower end of the Glen
3. c d Figure 15 Drainage enforcement in a low relief floodplain Darling River at Bono near Menindee NSW 142 38E 32 57S Elevation range 50 80 m a Digital surface model DSM showing substantial noise and offsets due to trees along the river b Smoothed DEM S which includes some elevation offsets along the river c DEM H showing enforcement of river channels through the floodplain and preservation of numerous shallow depressions darker green areas in the floodplain d Landsat image Elevations near coasts and the water and ocean masks The original SRTM elevation data was prepared using edit rules which specified amongst other things that land areas adjacent to water bodies are at least 1 m above the water level note that the original SRTM data was in integer form so the 1 m increment was the smallest possible Any land elevations that were equal to or lower than the adjacent water elevations were raised to 1 m above the water elevation Land cells immediately adjacent to the coast were therefore at least 1 m in elevation although cells further inland can have lower elevations The modifications to the SRTM elevations to remove stripes remove offsets due to trees reduce noise and enforce drainage have induced changes to elevations that may result in lower elevations raising the possibility of logically inconsistent land elevations lower than the adjacent water elevations The DSM and DEM have had a modified version o
4. Australian Government Geoscience Australia 1 second SRTM Derived Products User Guide 1 arc second DSM DEM DEM S DEM H 3 arc second DSM DEM DEM S Version 1 0 4 October 2011 1 second DSM product Restricted for Government Use Only Supply Details Geoscience Australia Sales Centre GPO Box 378 Canberra ACT 2601 Email sales ga gov au Freecall Aus only 1800 800 173 Phone 61 2 6249 9966 Fax 61 2 6249 9960 Document Revisions Revision Revised Reason for update approved Version number by 23 12 2009 User Guide first published 01 02 2010 Typing error in definitions in Nomenclature section Revision Date 24 02 2010 Renaming of DEM and DSM mosaics to a non NW 1 0 2 numeric starting grid name to not violate ESRI naming limitations 30 08 2010 Inclusion of 1 second and 3 second DEM S Removal NW 1 0 3 of 1 second cover letter and licence to separate document 1 10 2011 Inclusion of 1 second DEM H Modification of 1 NW 1 0 4 second products for public release 1 second SRTM Derived Digital Elevation Models User Guide Version 1 0 4 PRODUCT DESCRIP TION occitano aia teste ete 5 CONTACIINFORMATION ssccnssssseraees tasstasieotasaelasconnen rm eataiant atone reir star Oise 5 TIS OLE netic o oca 5 SUI ata 5 CLIC TTS OCU CTU orner E a E ocitade 5 PA CKMOW CA CCTIVCINS di 5 LICENSING aeiio ad 6 Creative Commons I second amp 3 second Products excluding I second DSM
5. Grohman G Kroenung G and Strebeck J 2006 Filling SRTM voids The delta surface fill method Photogrammetric Engineering and Remote Sensing 72 3 213 216 Read A M Gallant J C and Dowling T I in prep Destriping and void filling methods used in SRTM 1 Second processing Rodriguez E Morris C S and Belz J E 2006 A global assessment of the SRTM performance Photogrammetric Engineering and Remote Sensing 72 3 249 260 Sibson R 1981 A brief description of natural neighbour interpolation In V Barnet editor Interpreting Multivariate Data pages 21 36 John Wiley amp Sons Chichester Slater J A Garvey G Johnston C Haase J Heady B Kroenung G and Little J 2006 The SRTM data finishing process and products Photogrammetric Engineering and Remote Sensing 72 3 237 247 For technical queries please contact Geoscience Australia elevation ga gov au or GA Sales on 02 6249 9966 Authors J ohn Gallant CSIRO Trevor Dowling CSIRO Arthur Read CSIRO Nerida Wilson Geoscience Australia Phil Tickle Geoscience Australia and Chris Inskeep Geoscience Australia One second SRTM Derived Digital Elevation Models User Guide v1 0 4 Page 92 of 106 Appendix F 3 second DEM Metadata Note This metadata describes the dataset in accordance with the ANZLIC Australia New Zealand Land Information Council Core Metadata Guidelines Version 2 Dataset citation ANZLIC unique identifier
6. Pyramid building may take a few moments Would pon like to create pyramids Help Mo Cancel Use my choice and do not show this dialog in the future Once the file opens you use the Navigation Tools to Zoom in Zoom Out Zoom to Full Extent Zoom to Previous etc al al ES e To load this toolbar right click in the black area at the top of ArcMap and scroll down until you get to Tools and tick to turn it on then position it in the window to suit To change the elevation colour ramp go to Layers Contents box and click on the coloured strip or right click on the layer Go to Properties or double click layer s name then under Symbology tab for further options One second SRTM Derived Digital Elevation Models User Guide v1 0 4 Page 104 of 106 Layers re Copy A Remove Layers dems3sw1_0 Joins and Relates b Value Zoom To Layer _ High 2221 08 lt Select Color Ramp HE Zoom To Raster Resolution Color Ramp Copy Raster Symbology Visible Scale Range b Data Save As Layer File Properties Once in the Properties Layer the min max values or other values can be altered as desired or you can apply a stretch apply a hillshade effect or change to a classified colour ramp defined colour for each range of elevation values selected Layer Properties General Source Extent Display symbology Fields Jons amp Relates Show Unique Values Draw raster stretching v
7. Vegetation masks at 1 8 x 1 8 degree resolution illustrating where vegetation was removed from the DEM and issues noted with the removal Tile indexes for the DEM S One second SRTM Derived Digital Elevation Models User Guide v1 0 4 Page 80 of 106 References Gallant J C 2011 An adaptive smoothing method for improving noisy DEMs http geomorphometry org 2011 Geoscience Australia 2008 GEODATA 9 Second DEM Version 3 Geoscience Australia and CSIRO 2011 1 Second SRTM Derived Digital Elevation Models User Guide Version 1 0 4 Geoscience Australia Rodriguez E Morris C S and Belz J E 2006 A global assessment of the SRTM performance Photogrammetric Engineering and Remote Sensing 72 3 249 260 Slater J A Garvey G Johnston C Haase J Heady B Kroenung G and Little J 2006 The SRTM data finishing process and products Photogrammetric Engineering and Remote Sensing 72 3 237 247 For technical queries please contact Geoscience Australia elevation ga gov au or GA Sales on 02 6249 9966 Authors J ohn Gallant CSIRO Trevor Dowling CSIRO Arthur Read CSIRO Nerida Wilson Geoscience Australia Phil Tickle Geoscience Australia and Chris Inskeep Geoscience Australia One second SRTM Derived Digital Elevation Models User Guide v1 0 4 Page 81 of 106 Appendix D 1 second DEM H Metadata Note This metadata describes the dataset in accordance with the ANZLIC Australia New Zealand L
8. Morris C S and Belz J E 2006 A global assessment of the SRTM performance Photogrammetric Engineering and Remote Sensing 72 3 249 260 Sibson R 1981 A brief description of natural neighbour interpolation In V Barnet editor Interpreting Multivariate Data pages 21 36 John Wiley amp Sons Chichester Slater J A Garvey G Johnston C Haase J Heady B Kroenung G and Little J 2006 The SRTM data finishing process and products Photogrammetric Engineering and Remote Sensing 72 3 237 247 For technical queries please contact Geoscience Australia elevation ga gov au or GA Sales on 02 6249 9966 Authors J ohn Gallant CSIRO Trevor Dowling CSIRO Arthur Read CSIRO Nerida Wilson Geoscience Australia Phil Tickle Geoscience Australia and Chris Inskeep Geoscience Australia One second SRTM Derived Digital Elevation Models User Guide v1 0 4 Page 76 of 106 Appendix C 1 second DEM S Metadata Note This metadata describes the dataset in accordance with the ANZLIC Australia New Zealand Land Information Council Core Metadata Guidelines Version 2 Dataset citation ANZLIC unique identifier ANZCWO703014016 Title 1 Second SRTM Derived Smoothed Digital Elevation Model DEM S version 1 0 Custodian Custodian Geoscience Australia Jurisdiction Australia Description Abstract The 1 second Shuttle Radar Topography Mission SRTM derived smoothed Digital Elevation Model DEM S Version
9. This DEM is based on the 1 second SRTM derived Digital Surface Model DSM that was itself derived from the 1 second Shuttle Radar Topography Mission data The DSM was produced by removing stripes filling voids and re flattening water bodies Further details are provided in the DSM metadata ANZCW0703013336 The vegetation removal used the DSM without voids filled so that vegetation height estimates would not be affected by interpolated heights and so that voids adjacent to vegetated areas could be filled using bare earth elevations DEM processing vegetation offset removal Vegetation offsets were identified using Landsat based mapping of woody vegetation The height offsets were estimated around the edges of vegetation patches then interpolated to a continuous surface of vegetation height offset that was subtracted from the DSM to produce a bare earth DEM Further details are provided in the DSM metadata ANZCW0703013336 DEM S adaptive smoothing The smoothing process was based on the amount of noise in the DEM The noise was estimated from the local variation in the difference between elevation and the mean of nearby elevations The adaptive smoothing process was designed to smooth flat areas to a greater degree than steep areas and to respond to the degree of noise so that very noisy flat areas are smoothed more than less noisy flat areas The process operated over multiple resolutions allowing smoothing over quite large distances in
10. WGS84 Vertical Datum EGM96 refer to Accuracy Assessment section for further information Additional Information The figures shown in this User Guide were created using the 1 second products with a hill shade applied and elevation values selected to highlight particular features in the dataset One second SRTM Derived Digital Elevation Models User Guide v1 0 4 Page 7 of 106 Introduction The User Guide provides an informative overview of the various products derived from the 1 second SRTM data including the Digital Surface Model DSM the Digital Elevation Model DEM the Smoothed Digital Elevation Model DEM S and the hydrologically enforced DEM H products It describes the characteristics of the data the differences between the different products examples of the data in good and poor areas known problems and comparisons between various elevation data and the SRTM derived products It does not describe the methods in any detail so users should refer to the product metadata and the references cited there for further information These products have been released in good faith that the user understands the limitations and inherent errors in the data The data should not be solely relied upon for decision making but rather as a supplementary dataset The errors associated with these elevation products will be minimised over time as more accurate national DEM products evolve Details of Known errors in the data are explained
11. several higher resolution datasets suggest that elevation accuracy varies depending on the height and structure of the existing vegetation quality of vegetation input masks and local relief Further details of these comparisons are provided in the User Guide Geoscience Australia and CSIRO 2011 Height accuracy Is likely to be poorer in areas where voids have been filled using the GEODATA 9 second DEM particularly in high relief areas Logical Consistency The DEM represents heights of the land surface Due to random noise the relative elevation between adjacent grid cells can be in error by several m The removal of vegetation involves estimation of vegetation height at the edges of vegetation patches and interpolation of those heights across areas of continuous vegetation cover Variations in vegetation height within large areas of vegetation are not captured by this method The vegetation removal process guarantees that no elevations have been increased as part of the process All void areas have been filled and there are no discontinuities due to tile boundaries The SRTM editing rules relating to water bodies have been respected in the processing lakes are flat rivers descend continuously in a downstream direction and sea surfaces are at O m elevation Flattened water bodies occupy the same areas as in the original SRTM 1 second data Grid cells adjacent to water bodies are at least 1 cm above the water surface Void areas within
12. software e Resampling of 1 second products to produce publicly available 3 second products DEM DSM and DEM S The main processing of the SRTM products has included e Removal of stripes e Void filling e Tree offset removal using automated methods e Adaptive smoothing of DEM v1 0 e Water masking to re flatten water bodies affected by processing e Drainage enforcement and flow direction checking Stripe removal Diagonal stripes exist across most of the SRTM DSM and are most visible in low relief landscapes The orientation of the stripes generally relates to the orbital path of the Space Shuttle The stripes are about 800 m apart and their amplitude is typically around 1 m but up to 4 m in places and can vary quite abruptly The stripes were treated using a 2 dimensional Fourier Transform method that detects features with a consistent orientation and spacing Stripes were detected One second SRTM Derived Digital Elevation Models User Guide v1 0 4 Page 10 of 106 and removed throughout Australia except where high relief masked their presence One of the ancillary data layers provided with the product shows the maximum magnitude of striping removed across the continent Figure 1 Maximum destripe m Max 4 O 250 500 1 000 Km Min O Figure 1 Magnitude and distribution of stripe cleaning Stripe removal was effective in most areas but in some locations where there were abrupt changes in the stripe amplitud
13. 146 99E 19 485 Elevation range 100 1100 m Black lines are clipped AusHydro stream lines and red lines are infill stream lines derived from DEM S One second SRTM Derived Digital Elevation Models User Guide v1 0 4 Page 45 of 106 Figure 36 Substantial reduction in height of mountain peak from DEM S to DEM H due to smoothing by ANUDEM Generally flat tiles with a few hills are the most affected The largest height reductions are nearly 200 m The smoothing has also raised part of the south eastern flank of the mountain by over 100 m Bluff Knoll and Coyanarup Peak Stirling Range WA 118 26E 34 38S Elevation range 100 1200 m One second SRTM Derived Digital Elevation Models User Guide v1 0 4 Page 46 of 106 Figure 37 Excessive incision downstream of open cut mines Latrobe Valley Vic 146 5E 38 25 Elevation range 100 400 m The incised drainage line at about 60 m elevation continues to the Gippsland Lakes 100 km downstream Incomplete Stream Enforcements A small number of tiles have been identified in which the AusHydro 1 250 000 steam lines have not been properly enforced into the DEM H The tiles identified are listed below with a rating of the degree of error based on an assessment by CSIRO and GA Larger errors seem to occur in the braided stream networks and often expand across neighbouring tiles These errors and any additional errors identified by users and emailed to elevation ga gov au will be looked at in future
14. 2009 has been released and this was considered as an alternate source of filling voids in the steep areas but rejected due to inconsistent quality The void filling method matches elevations around the edge of the void which avoids abrupt elevation changes at the void edges Some void fills are affected by erratic elevation values around the edge of the void particularly in salt lake areas in central Australia Figure 5 One second SRTM Derived Digital Elevation Models User Guide v1 0 4 Page 14 of 106 pu a Y A na Figure 4 Void filling in canyon area a fairly good result although canyon bottom has not quite been captured properly Colo River Wollemi National Park NSW 150 6E 33 35 Elevation range 0 800 m One second SRTM Derived Digital Elevation Models User Guide v1 0 4 Page 15 of 106 data isy the void are variable due to the noi Ing lac ions rep Poolowanna Lake S the elevati ds filled in a lake bed around the edge of the vo Figure 5 Vo Desert SA 137 6E 26 75 impson 20 70 m ion range Elevat Page 16 of 106 One second SRTM Derived Digital Elevation Models User Guide v1 0 4 Figure 6 Voids that are probably due to low reflectivity in dry sandy soils Coomallo Hill WA 115 4E 30 25 Elevation range 100 400 m Vegetation offset removal The radar used for the SRTM DSM does not penetrate vegetation so areas with a high tree density are visible in the DSM as raised
15. DATA 0 iia 104 MIDES ARCOS id 104 IO PUMCY DOWCS MIDI Or ASADA AAA AAA AAA AA AA Setuttodoudeels 106 Product Description Contact Information Custodian Geoscience Australia Cnr Jerrabomberra Avenue Hindmarsh Drive Symonston ACT 2609 Australia GPO Box 378 Canberra ACT 2601 Australia elevation ga gov au Supply Details Geoscience Australia Sales Centre GPO Box 378 Canberra ACT 2601 Phone 61 2 6249 9966 Freecall Aus only 1800 800 173 Fax 61 2 6249 9960 Email sales ga gov au Cite this document as Gallant J C Dowling T I Read A M Wilson N Tickle P Inskeep C 2011 1 second SRTM Derived Digital Elevation Models User Guide Geoscience Australia www ga gov au topographic mapping digital elevation data html Acknowledgements Authors include J ohn Gallant Trevor Dowling Arthur Read from CSIRO and from Geoscience Australia Nerida Wilson Phil Tickle Chris Inskeep Acknowledgement and input from Bureau of Meteorology BoM CSIRO and Australian National University ANU Fenner School of Environment and Society One second SRTM Derived Digital Elevation Models User Guide v1 0 4 Page 5 of 106 Licensing Creative Commons 1 second amp 3 second Products excluding 1 second DSM In June 2011 the Level 2 1 second or 30 m SRTM bare earth models DEM DEM S and DEM H were approved by Australia s Defence Imagery and Geospatial Organisation DIGO and the United States
16. Department of Defence for public release These products are now available under Creative Commons It does not include the Digital Surface Model DSM and associated vegetation layers which are still supplied under a government restricted licence The 3 second SRTM Digital Elevation Models DSM DEM and DEM S were released in August 2010 under Creative Commons Attribution 3 0 Australia licence Creative Commons means the data can be shared copied distributed and transmitted or adapted providing you acknowledge Geoscience Australia as the author or licensor Commonwealth of Australia Geoscience Australia 2011 This material is released under the Creative Commons Attribution 3 0 Australia Licence Further information on Creative Commons can be found on the website http creativecommons org licenses by 3 0 au Licence 1 second DSM Government Use Only The only product in the 1 second range that is not for public release is the Digital Surface Model DSM which is strictly for Government use only It is provided on request with specific licensing and release constraints agreed to by Australia s Defence Imagery and Geospatial Organisation DIGO and the United States Department of Defence It is therefore crucial that these conditions are adhered to both in terms of the source data and in the derivation of future products The release constraints are based on an assessment of the risk to national security of making the d
17. OF TSEC OND PRODUC IS lt eo 50 MAOCURAC ASSESSMENT tad Ad cag lesions 50 Permanent SUT VEY MODAS E AS SS 51 Tablelands Regional Council CONOUEDIAA lt A a das 53 COMPARISON WITH OTHER ELEVATION DATASETS cccccececscececececcscecececescscececccescscececesessecscecsascecececacess 57 A A c oct taht ig tana cain ark nae eee ee ase 37 TOW er Dar E L DA RA A A A AAA A 60 DERIVATION OF THE 3 SECOND PRODUCTS seseessoccessccceesssoccessscccesssooccesscoccessoccceessocccsssoceesssccesssso 63 PROCESSING OF THE 5 SECOND PRODUCTS riai iE E O oi 63 INCCURAGY ASSESSMENT ri sitio 63 CHOOSING THE CORRECTO DECOND ERODUC Torrado iaa 64 FUTURE DEVELOPMENTS ti acid 64 FEEDBACK rossi persons taens NE Oise 64 REFERENCES E A rs 64 APPENDIX A 1 SECOND DSM METADATA sccccssssssccsssscsscccsssssccscssssccssssssscccscsssscsscssssssscesssseees 66 APPENDIX B 1 SECOND DEM METADATA ccccssssssscssssssccssssssssccsssssscccsssssssccsssssssccesssssssscessesnss 71 APPENDIX C SECOND DEM S METADATA ide 77 APPENDIX D 1 SECOND DEM H METADATA sseesessssccessscoccesscoccessooccessscocesssocccesssoccesssccceesssoceesssoceeeo 82 APPENDIX E 3 SECOND DSM METADATA eeeessssocessscccessscoceessooccessoccceesscoceesscoceessococcessoocceesssccessssceeeeo 88 APPENDIX F 3 SECOND DEM METADATA dias 93 APPENDIX G 3 SECOND DEM S METADATA uuu ccsssssrccsssscsccccssssscccccsssscccsscssssscssssssscsssssssssscssssssees 99 APPENDIX H LOADING THE
18. and SRTM Water Body Data datasets Some fragments of mainland or pieces of islands may be missing Contact information Contact organisation Geoscience Australia GA Contact position Director Sales and Distribution ISB Mail address GPO Box 378 Mail address Locality Canberra State ACT Country Australia Postcode 2601 Telephone 61 2 6249 9966 Facsimile 61 2 6249 9960 Electronic mail address sales ga gov au Metadata information Metadata date 2010 08 30 Metadata Updated date 2011 09 01 Additional metadata Conversion to floating point format As a by product of the de striping process the integer data was converted to floating point format to allow for the continuously varying nature of the striping Areas where no de striping was required will contain unaltered integer values but represented in floating point format for consistency The 3 second data was produced as integer values as the decimal values are well below any effect on the accuracy One second SRTM Derived Digital Elevation Models User Guide v1 0 4 Page 91 of 106 Ancillary data layers distributed with the data A water mask at 1 second resolution showing the cells that are part of the flattened water bodies JPEG image of the 3 second DSM References Geoscience Australia 2008 GEODATA 9 Second DEM Version 3 Geoscience Australia and CSIRO 2011 1 Second SRTM Derived Digital Elevation Models User Guide Version 1 0 4 Geoscience Australia
19. and update frequency Updates and revisions are anticipated primarily to incorporate improvements to the bare earth DEM and DEM S over time A first revision is anticipated in 2011 and further revisions are likely Reference system Horizontal datum WGS84 Vertical datum EGM96 One second SRTM Derived Digital Elevation Models User Guide v1 0 4 Page 77 of 106 Access Stored data format DIGITAL ArcGIS grid Arclnfo grid Available format type DIGITAL ArcGIS grid Arcl nfo grid Access constraints From October 2011 this data is released under the Creative Commons Attribution 3 0 Australia Licence for use by government and the public http creativecommons orq licenses by 3 0 au Copyright O Commonwealth of Australia Geoscience Australia 2010 Data quality Lineage Source data 1 SRTM 1 second Version 2 data Slater et al 2006 supplied by Defence Imagery and Geospatial Organisation DIGO as 813 1 x 1 degree tiles Data was produced by NASA from radar data collected by the Shuttle Radar Topography Mission in February 2000 2 GEODATA 9 second DEM Version 3 Geoscience Australia 2008 used to fill voids 3 SRTM Water Body Data SWBD shapefile accompanying the SRTM data Slater et al 2006 This defines the coastline and larger inland waterbodies for the SRTM DEM and DSM 4 Vegetation masks and water masks applied to the DEM to remove vegetation 5 Adaptive smoothing applied to DEM to produce DEM S DSM processing
20. better In low relief landscapes the spatial error of the 1 250 000 stream lines is not a significant problem but in steeper areas the spatial offsets result in drainage lines being incised into hillslopes rather than valley floors To prevent this the mapped stream lines were used only where slope in DEM S was less than 10 degrees The excised segments were replaced with infilling stream lines derived from DEM S using a version of the AT search algorithm Ehlschlaeger 1989 known primarily as its implementation in GRASS as r watershed method that constructs flow lines through depressions without first filling the depressions to the outlet level The ANUDEM software cannot process the entire continent or entire drainage basins at the 1 second resolution in a single pass Drainage enforcement was therefore performed separately for each 1x1 degree tile using Y degree overlaps on each side The resulting 2x2 degree tiles were trimmed to a 100 cell overlap mosaicked with adjacent trimmed tiles then clipped to the 1x1 degree tile The mosaicking process does not guarantee the preservation of continuous descent along drainage lines so a final descent enforcement step was applied using the CheckStreamDescent program written for that purpose CheckStreamDescent processes all tiles as a single data set so that continuous descent of all stream lines to their termination points was ensured As a final step the ocean areas were set to no data Other
21. comprises a 5 m grid bare earth DEM derived from LiDAR data acquired in mid 2009 Figure 50 and the SRTM 1 second derived DEM Non ground points such as vegetation and man made structures were removed from the DEM so that it defines the bare earth ground surface The vertical accuracy of the LIDAR mass point data was verified at lt 15 cm 95 confidence LIDAR Comparison Lower Darling Project Elevation High 94 9553 Low 40 9575 0 15 30 Km Lo E E Figure 50 Lower Darling 5 m Grid 142 625E 32 477S ETA win aora ma 33 175 Lower Darling LIDAR statistics A difference surface Figure 51 was created by subtracting the LIDAR generated 5 m Grid from the DEM The difference surface shows some significant differences between the LIDAR and the SRTM DEM There are some offsets due to riparian vegetation that have not been removed from the STRM derived DEM There is also some striping identifiable that appears in the difference surface Differences in water surface heights in lakes are to be expected from data obtained at different times One second SRTM Derived Digital Elevation Models User Guide v1 0 4 Page 60 of 106 1sDEMv1 LowerDarling5mGRID Difference metres MN ss MN 35 4 21 2 Pe 21 1 12 A 11 9 6 9 EJ 5 8 4 4 MJ 43 28 E 2 7 1 5 1 4 0 2 0 11 23 i 24 33 E 34 42 ee 4 3 5 2 E 5 3 6 4 MN ss 7 7 MN 7 9 3 NN 1 5 MN 115 147 MI 14 5 32 F
22. due to smoothing is less than 1 5 m in 84 of tiles Mean elevation difference due to smoothing is less than 0 2 m across all tiles Figure 10 shows the changes made by the adaptive smoothing in an area of coastal NSW containing flat plateaus flat valley floors and steep escarpments The Smoothing has removed random variations in the flatter areas and left the steep areas essentially unchanged In the moderate relief areas there is some smoothing of topography with valleys raised and ridges lowered by a few m Figure 11 shows the profound impact of smoothing on a derived slope This area of Western Australia has subtle relief and relatively high noise levels the noise amplitude is 3 4 m in the noisier area and about 1 m in the less noisy area Before Smoothing the calculated slopes are heavily impacted by noise and there is little topographic structure apparent in the slope image After smoothing the topographic structure is clearly apparent One second SRTM Derived Digital Elevation Models User Guide v1 0 4 Page 21 of 106 PAE 7 f AS t f AO z we i He A y a a E O O Ly gt IAS po i ca Ror y E a i 5 yar z q 4 e ane gt a s by ont 7 Figure 10 Smoothed DEM S top The bottom image shows the difference between the DEM and smoothed DEM S with red showing areas that have been raised due to smoothing and blue those that have been lowered The differences in the low relief uplands and lowlands show rand
23. in this User Guide We urge users to provide feedback on any errors to Geoscience Australia at the following email address elevation ga gov au The 1 second resolution approximately 30 m products are available to government agencies and their collaborators and contractors Reduced resolution versions of the products at 3 second 90 m were released in 2010 Another product for public use is the GEODATA3 9 second 250 m DEM which has hydrological enforcement applied unlike the 3 second DEM and is available through Geoscience Australia Sales Centre Overview The 1 second DSM DEM DEM S and DEM H are national elevation data products derived from the Shuttle Radar Topography Mission SRTM data The SRTM data is not suitable for routine application due to various artefacts and noise The data has been treated with several processes to produce more usable products e A cleaned digital surface model DSM o regular grid representing ground surface topography as well as other features including vegetation and man made structures e A bare earth digital elevation model DEM o regular grid representing ground surface topography and where possible excluding other features such as vegetation and man made structures e A smoothed digital elevation model DEM S o A smoothed DEM based on the bare earth DEM that has been adaptively smoothed to reduce random noise typically associated with the SRTM data in low relief areas e A hydr
24. open water or dry sandy soils or topographic shadowing in high relief areas The Delta Surface Fill Method Grohman et al 2006 was adapted for this task using GEODATA 9 second DEM as the infill data source The 9 second data was refined to 1 second resolution using ANUDEM 5 2 without drainage enforcement Delta Surface Fill Method calculates height differences between SRTM and infill data to create a delta surface with voids where the SRTM has no values then interpolates across voids The void is then replaced by infill DEM adjusted by the interpolated delta surface resulting in an exact match of heights at the edges of each void Two changes to the Delta Surface Fill Method were made interpolation of the delta surface was achieved with natural One second SRTM Derived Digital Elevation Models User Guide v1 0 4 Page 89 of 106 neighbour interpolation Sibson 1981 implemented in ArcGIS 9 3 rather than inverse distance weighted interpolation and a mean plane inside larger voids was not used Water bodies Flat water bodies in the original 1 second data were modified as part of the de striping process and were re flattened afterwards SRTM Water Body Data was converted to a 1 second resolution grid then adjusted to match the extent of equal height pixels in original SRTM 1 second data Grid cells within that water mask were set to the original SRTM height Edit rules for land surrounding water bodies SRTM edit rules set all land adja
25. particular areas where removal of vegetation offsets to produce the DEM was ineffective will exhibit incorrect flow pathways Drainage enforcement modifies elevations and surface form significantly in some areas and applications that are concerned with landforms and elevations where drainage connectivity is not a critical factor should use DEM S in preference to DEM H The elevation error for DEM H ts difficult to characterise In general it will be similar to the raw SRTM 1 second data with 90 of tested heights within 9 8 m for Australia Rodriguez et al 2006 but significant changes to elevation have occurred due to the Smoothing and drainage enforcement processes As noted in the Quality Assessment section above errors as large as 200 m occur in some areas Further information on known errors is provided in the User Guide Geoscience Australia and CSIRO 2011 Logical Consistency The DEM H represents heights of the land surface modified to ensure that elevations decrease continuously in the downstream direction along drainage lines Smaller sinks have been cleared as part of the process but a large number of sinks remain where indicated by the elevation data Most of these are genuine topographic depressions but some are due to data errors There are no voids and there are no discontinuities due to tile boundaries Completeness The DEM H covers all of continental Australia and near coastal islands with land areas including al
26. shown in Figure 39 Figure 39 Medium error where part of the AusHydro stream network has not been enforced into the DEM H Adcock River north of Fitzroy Crossing WA 125 867E 17 325 Elevation range 196 934m Black lines are infill stream lines derived from DEM S and red lines are AusHydro stream lines Large errors are present in Cooper Creek NSW where a significant portion of a braided stream network has not been enforced into the DEM H Figure 40 l a gt H wae i il e a NINN Go te a eae qeg a eX Se Figure 40 Large error where a large fraction of a major braided network from AusHydro stream network has not been enforced into the DEM H Cooper Creek north west Bourke NSW 141 934 E 26 675S Elevation range 74 332m Black lines are infill stream lines derived from DEM S and red lines are AusHydro stream lines One second SRTM Derived Digital Elevation Models User Guide v1 0 4 Page 49 of 106 Evaluation of 1 Second Products Accuracy Assessment The SRTM data is part of a global dataset released by US Defence Department in the WGS84 projection with a height datum in the Earth Geopotential Model 1996 EGM96 The difference between EGM96 and the Australian Height Datum 1971 AHD71 is between 0 8 m and 1 2 m which is generally less than the uncertainty in the SRTM heights Given the minor difference between EGM96 and AHD71 relative to the vertical accuracy of the data and to maintain consistency with th
27. smaller offset 1 cm rather than 1 m could be used because the cleaned digital surface model is in floating point format rather than integer format of the original SRTM Some small islands within water bodies are represented as voids within the SRTM due to edit rules These voids are filled as part of void filling process and their elevations set to a minimum of 1 cm above surrounding water surface across the entire void fill DSM ancillary data layers Four additional data layers provide information about the alterations to the raw SRTM data to produce this DSM e A de stripe mask indicating which 1 4 x degree tiles have been affected by destriping and which have not been de striped e A striping magnitude layer showing the amplitude of the striping at 0 01 degree 1km resolution Restricted Licence for Government Use Only e A water mask at 1 second resolution showing the cells that are part of the flattened water bodies e Avoid mask showing cells that were no data in the raw SRTM and have been filled using the void filling algorithm Positional accuracy The horizontal positional error is the same as for the raw SRTM 1 second data with 90 of tested locations within 7 2 m for Australia See Rodriguez et al 2006 for more information Attribute accuracy Elevation accuracy is essentially the same as for the raw SRTM 1 second data with 90 of tested heights within 9 8 m for Australia Errors in height are still mostly due to r
28. surfaces are at 0 m elevation Flattened water bodies occupy the same areas as in the original SRTM 1 second data Grid cells adjacent to water bodies are at least 1 cm above the water surface Void areas within water bodies small islands not represented in the original SRTM data are at least 1 cm above the water surface over their entire area Completeness The DSM covers all of continental Australia and near coastal islands with land areas including all islands defined by the available SRTM 1 second elevation and SRTM Water Body Data datasets The following tiles containing fragments of mainland or pieces of islands were not supplied at 1 second resolution and are therefore missing from the DSM E112 S26 E124 S15 E142 S10 E113 S29 E125 S14 E143 S10 E118 S20 E132 11 E146 S17 E120 S35 E133 11 E150 S22 E121 S35 E134 S35 E152 S24 E123 S16 E141 S10 Note that the coordinates are of the south western corner of the tile Contact information Contact organisation Geoscience Australia GA Contact position Director Sales and Distribution ISB Mail address GPO Box 378 Mail address Locality Canberra State ACT Country Australia Postcode 2601 Telephone 61 2 6249 9966 Facsimile 61 2 6249 9960 Electronic mail address sales ga gov au Metadata information Metadata Created date 2009 12 23 Metadata Updated date 2010 08 30 Metadata Updated date 2011 09 01 Additional metadata Conversion to floating point fo
29. terrain and as such this national accuracy assessment should not be used in densely vegetated or high relief areas Results of this comparison show the absolute accuracy of the data as tested to be 7 582 m at the 95 percentile with a RMS error of 3 868 in open flat terrain Ninety nine percent of points are within a height difference of less than 9 602 m There are eight points with a height difference of more than 10 m most of which occur in high elevations on densely vegetated slopes where vegetation removal is likely to be the cause of the difference One second SRTM Derived Digital Elevation Models User Guide v1 0 4 Page 52 of 106 Tablelands Regional Council Contour Data The Tablelands Regional Council QLD provided contour data at a 2 m interval based on the AHD71 vertical datum covering an area south of Lake Tinaroo in the Atherton Tablelands Figure 43 Figure 43 Location of Atherton Tablelands 2 m Contour Data The contours were used to create a surface which was then compared with the 1 second DEM producing the difference surface shown in Figure 44 The statistical differences between the surfaces were as follows Mean 2085 mim 20909 ma DEM Atherton Difference Surface Statistics One second SRTM Derived Digital Elevation Models User Guide v1 0 4 Page 53 of 106 1sDEMv1 0 Atherton2m ag Figure 44 Atherton difference grid 145 604E 17 303S A total of 7881 points were randomly generated acro
30. the project in order to work effectively on the SRTM data as summarised in Hutchinson et al 2009 ANUDEM uses a discretised spline interpolation method that smooths the surface while enforcing continuous descent along supplied drainage lines and removing spurious sinks where that is consistent with the accuracy of the source elevation data DEM S The 1 250 000 scale stream line data used to produce the GEODATA 9 second DEM Version 3 was chosen as the source of drainage line data as this data set was the only available source of mostly cleaned and correctly oriented drainage lines covering the entire continent This 1 250 000 scale data with a spatial accuracy of about 200 m was significantly coarser than the 1 second DEM S which is accurate to 50 m or better In low relief landscapes the spatial error of the 1 250 000 stream lines is not a Significant problem but in steeper areas the spatial offsets result in drainage lines being incised into hillslopes rather than valley floors Apart from distorting the terrain surface this results in errors in the stream network since the valley floor already present in the DEM S remains and the enforced drainage line forms a parallel flow path To prevent this the mapped stream lines were used only where Slope in DEM S was less than 10 degrees The excised segments were replaced with infilling stream lines derived from DEM S using a version of the A search algorithm Ehlschlaeger 1989 known primari
31. to floating point format The smoothing process alters all data values in the 1 second DEM by varying amounts and the result is a floating point data set capturing in some places very small but meaningful differences in elevation between adjacent cells Ancillary data layers A water mask at 1 second resolution showing the cells that are part of the flattened water bodies JPEG image of the 3 second DEM S References Gallant J C in prep An adaptive smoothing method for improving noisy DEMs http geomorphometry org 2011 Geoscience Australia 2008 GEODATA 9 Second DEM Version 3 Geoscience Australia and CSIRO 2011 1 Second SRTM derived Digital Elevation Models User Guide Version 1 0 4 Geoscience Australia Rodriguez E Morris C S and Belz J E 2006 A global assessment of the SRTM performance Photogrammetric Engineering and Remote Sensing 72 3 249 260 One second SRTM Derived Digital Elevation Models User Guide v1 0 4 Page 102 of 106 Slater J A Garvey G Johnston C Haase J Heady B Kroenung G and Little J 2006 The SRTM data finishing process and products Photogrammetric Engineering and Remote Sensing 72 3 237 247 For technical queries please contact Geoscience Australia elevation ga gov au or GA Sales on 02 6249 9966 Authors J ohn Gallant CSIRO Trevor Dowling CSIRO Arthur Read CSIRO Nerida Wilson Geoscience Australia and Phil Tickle Geoscience Australia One sec
32. under Creative Commons licensing since October 2011 ANZLIC search words LAND Topography Models ECOLOGY Landscape WATER Hydrology Geographic extent name AUSTRALIA EXCLUDING EXTERNAL TERRITORIES AUS Australia Australia Geographic bounding box North bounding latitude 10 South bounding latitude 44 East bounding longitude 154 West bounding longitude 113 Data currency Beginning date 2000 2 11 Ending date 2000 2 22 Dataset status Progress Version 1 0 of the 1 second DEM H is complete as at 15 April 2011 Maintenance and update frequency Updates and revisions are anticipated to resolve some of the issues identified in the User Guide Geoscience Australia and CSIRO 2011 and Quality Assessment layers and to incorporate improvements in the source smoothed Digital Elevation Model DEM S and the DEM and DSM it is derived from Updates incorporating finer scale stream line One second SRTM Derived Digital Elevation Models User Guide v1 0 4 Page 82 of 106 data provided by State and Territory agencies are also anticipated progressively on a catchment by catchment basis Reference system Horizontal datum WGS84 Vertical datum EGM96 Access Stored data format DIGITAL ArcGIS grid Arclnfo grid Available format type DIGITAL ArcGIS grid Arclnfo grid Access constraints From October 2011 this data is released under the Creative Commons Attribution 3 0 Australia Licence for use by government and the
33. water bodies small islands not represented in the original SRTM data are at least 1 cm above the water surface over their entire area One second SRTM Derived Digital Elevation Models User Guide v1 0 4 Page 74 of 106 Completeness The DEM covers all of continental Australia and near coastal islands with land areas including all islands defined by the available SRTM 1 second elevation and SRTM Water Body Data datasets The following tiles containing fragments of mainland or pieces of islands were not supplied at 1 second resolution and are therefore missing from the DEM E112 S26 E124 S15 E142 S10 E113 S29 E125 S14 E143 S10 E118 S20 E132 S11 E146 S17 E120 S35 E133 S11 E150 S22 E121 S35 E134 S35 E152 S24 E123 S16 E141 S10 Note that the coordinates are of the south western corner of the tile Contact information Contact organisation Geoscience Australia GA Contact position Director Sales and Distribution ISB Mail address GPO Box 378 Mail address Locality Canberra State ACT Country Australia Postcode 2601 Telephone 61 2 6249 9966 Facsimile 61 2 6249 9960 Electronic mail address sales ga gov au Metadata information Metadata Created date 2009 12 23 Metadata Updated date 2010 08 30 Metadata Updated date 2011 09 01 Additional metadata Conversion to floating point format As a by product of the de striping process the integer data was converted to floating point format to allow for the contin
34. 0 has been applied to the tile names to differentiate between the elevation models and the version number in this case version 1 0 The 3 second DEM and DSM are available in integer format and the 3 second DEM S is in 32 bit Floating Point in a national mosaic If ordering the data through the Geoscience Australia Sales Centre the data will be Supplied as a national mosaic product in ESRI Grid format Geoscience Australia does not provide customised extents or file formats A subset of the data can be downloaded from the National Elevation Data Framework NEDF web portal at http nedf ga gov au in many formats in national coverage or tile format initially ESRI Grid Orders over a certain size will incur a data transfer cost Same as purchasing through Sales Further information on loading data into various software packages is explained in Appendix H Otherwise please consult your software vendor Data Extent Australia mainland and near shore islands North bounding latitude 10 South bounding latitude 44 East bounding longitude 154 West bounding longitude 113 The following tiles containing fragment or pieces of islands were not applied at 1 second resolution SRTM and therefore are missing from the 1 and 3 second products E112 S26 E120 S35 E124 S15 E133 S11 E142 S10 E150 S22 E113 S29 E121 35 E125 S14 E134 35 E143 S10 E152 S24 E118 S20 E123 S16 E132 S11 E141 S10 E146 S17 Reference System Horizontal Datum
35. 0 2 22 Dataset status Progress Version 1 0 of the 3 second bare earth DSM is complete as at 30 August 2010 Maintenance and update frequency Updates and revisions are anticipated primarily to incorporate improvements to the bare earth DEM over time Further revisions are likely once the 1 second products have been released One second SRTM Derived Digital Elevation Models User Guide v1 0 4 Page 88 of 106 Reference system Horizontal datum WGS84 Vertical datum EGM96 Access Stored data format DIGITAL ArcGIS grid Arclnfo grid Available format type DIGITAL ArcGIS grid Arclnfo grid Access constraints This data is released under the Creative Commons Attribution 3 0 Australia Licence for use by government and the public http creativecommons org licenses by 3 0 au Copyright Commonwealth of Australia Geoscience Australia 2010 Data quality Lineage Source data 1 SRTM 1 second Version 2 data Slater et al 2006 supplied by Defence Imagery and Geospatial Organisation DIGO as 813 1 x 1 degree tiles Data was produced by NASA from radar data collected by the Shuttle Radar Topography Mission in February 2000 2 GEODATA 9 second DEM Version 3 Geoscience Australia 2008 used to fill voids 3 SRTM Water Body Data SWBD shapefile accompanying the SRTM data Slater et al 2006 This defines the coastline and larger inland waterbodies for the SRTM DSM 4 1 second DSM resampled to 3 second DSM In order to und
36. 014216 DEM ANZCW0703014182 DEM S ANZCWO703014217 ANZLIC search words LAND Topography Models ECOLOGY Landscape Geographic extent name AUSTRALIA EXCLUDING EXTERNAL TERRITORIES AUS Australia Australia Geographic bounding box North bounding latitude 10 South bounding latitude 44 East bounding longitude 154 West bounding longitude 113 Data currency Beginning date 2000 2 11 Ending date 2000 2 22 Dataset status Progress Version 1 0 of the 3 second DEM S is complete as at 30 August 2010 Maintenance and update frequency Updates and revisions are anticipated primarily to incorporate improvements to the bare earth DEM over time Further revisions are likely once the 1 second products have been released One second SRTM Derived Digital Elevation Models User Guide v1 0 4 Page 99 of 106 Reference system Horizontal datum WGS84 Vertical datum EGM96 Access Stored data format DIGITAL ArcGIS grid Arclnfo grid Available format type DIGITAL ArcGIS grid Arclnfo grid Access constraints This data is released under the Creative Commons Attribution 3 0 Australia Licence for use by government and the public http creativecommons org licenses by 3 0 au Copyright Commonwealth of Australia Geoscience Australia 2010 Data quality Lineage Source data 1 SRTM 1 second Level 2 data Slater et al 2006 supplied by Defence Imagery and Geospatial Organisation DIGO as 813 1 x 1 degree tile
37. 1 0 is a 1 arc second 30 m gridded smoothed version of the DEM ANZCW0703013355 The DEM S represents ground surface topography excluding vegetation features and has been smoothed to reduce noise and improve the representation of surface shape The dataset was derived from the 1 second Digital Elevation Model Version 1 0 DSM ANZCW0703013336 by an adaptive smoothing process that applies more smoothing in flatter areas than hilly areas and more smoothing in noisier areas than in less noisy areas This DEM S supports calculation of local terrain shape attributes such as slope aspect and curvature that could not be reliably derived from the unsmoothed DEM because of noise A full description of the methods is in progress Gallant et al in prep The DEM S was used to create the hydrologically enforced product DEM H ANZCWO703014615 The three 1 second products DEM DEM S and DEM H were released under Creative Commons licensing from October 2011 ANZLIC search words LAND Topography Models ECOLOGY Landscape Geographic extent name AUSTRALIA EXCLUDING EXTERNAL TERRITORIES AUS Australia Australia Geographic bounding box North bounding latitude 10 South bounding latitude 44 East bounding longitude 154 West bounding longitude 113 Data currency Beginning date 2000 2 11 Ending date 2000 2 22 Dataset status Progress Version 1 0 of the 1 second smoothed DEM S is complete as at 30 August 2010 Maintenance
38. 106 pp 211 232 Hutchinson M F 2009 ANUDEM Version 5 2 Fenner School of Environment and Society Australian National University Available online at http fennerschool anu edu au publications software anudem php last accessed January 2011 Hutchinson M F Stein J A Stein J L and Xu T 2009 Locally adaptive gridding of noisy high resolution topographic data In Anderssen R S R D Braddock and L T H Newham eds 18th World IMACS Congress and MODSIMO9 International Congress on Modelling and Simulation Modelling and Simulation Society of Australia and New Zealand and International Association for Mathematics and Computers in Simulation July 2009 pp 2493 2499 ISBN 978 0 9758400 7 8 http www mssanz org au modsim09 F13 hutchinson pdf Kobrick M 2006 On the toes of giants how SRTM was born Photogrammetric Engineering and Remote Sensing Rodriguez E Morris C S Belz J E Chapin E C Martin J M Daffer W Hensley S 2005 An assessment of the SRTM Topographic Products J et Propulsion Laboratory D 31639 JPL Technical Memorandum http www2 jpl nasa gov srtm SRTM_D31639 pdf One second SRTM Derived Digital Elevation Models User Guide v1 0 4 Page 65 of 106 Appendix A 1 second DSM Metadata Note This metadata describes the dataset in accordance with the ANZLIC Australia New Zealand Land Information Council Core Metadata Guidelines Version 2 Dataset citation ANZLIC unique identifier
39. 2009 One second SRTM Derived Digital Elevation Models User Guide v1 0 4 Page 71 of 106 Maintenance and update frequency Updates and revisions are anticipated to resolve some of the issues identified in the User Guide Geoscience Australia and CSIRO 2011 and Quality Assessment layers and to incorporate improvements in the Digital Elevation Model Reference system Horizontal datum WGS84 Vertical datum EGM96 Access Stored data format DIGITAL ArcGIS grid Arclnfo grid Available format type DIGITAL ArcGIS grid Arclnfo grid Access constraints From October 2011 this data is released under the Creative Commons Attribution 3 0 Australia Licence for use by government and the public http creativecommons org licenses by 3 0 au Copyright Commonwealth of Australia Geoscience Australia 2010 Data quality Lineage Source data 1 SRTM 1 second Version 2 data Slater et al 2006 supplied by Defence Imagery and Geospatial Organisation DIGO as 813 1 x 1 degree tiles Data were produced by NASA from radar data collected by the Shuttle Radar Topography Mission in February 2000 2 GEODATA 9 second DEM Version 3 Geoscience Australia 2008 used to fill voids 3 SRTM Water Body Data SWBD shapefile accompanying the SRTM data Slater et al 2006 This defines the coastline and larger inland waterbodies for the DEM and DSM 4 Vegetation masks and water masks applied to the DEM to remove vegetation DSM processing Th
40. ANZCWO703013336 Title 1 second SRTM Derived Digital Surface Model DSM version 1 0 Custodian Custodian Geoscience Australia Jurisdiction Australia Description Abstract The 1 second Shuttle Radar Topography Mission SRTM derived Digital Surface Model DSM Version 1 0 is a 1 arc second 30 m gridded DSM that represents ground Surface topography as well as features above the ground such as vegetation and man made structures The dataset was derived from the SRTM data acquired in February 2000 supported by the GEODATA 9 second DEM in void areas and the SRTM Water Body Data Stripes and voids have been removed from the 1 second SRTM data to provide an enhanced and complete DSM for Australia and near shore islands A full description of the methods is in progress Read et al in prep This 1 second DSM forms the source for the 1 second DEM with vegetation offsets removed ANZCWO703013355 the smoothed DEM DEM S ANZCW0703014016 and hydrologically enforced DEM DEM H ANZCWO703014615 It is available under a government restricted licence only on request from elevation ga gov au An alternative DSM available under Creative Commons licensing is a resampled 3 second 90m version part of the 3 second product set ANZLIC search words LAND Topography Models ECOLOGY Landscape Geographic extent name AUSTRALIA EXCLUDING EXTERNAL TERRITORIES AUS Australia Australia Geographic bounding box North bounding latitude 10 So
41. ANZCWO703014182 Title 1 second SRTM Derived 3 second Digital Elevation Model DEM version 1 0 Custodian Custodian Geoscience Australia Jurisdiction Australia Description Abstract The 3 second 90 m Shuttle Radar Topography Mission SRTM Digital Elevation Model DEM version 1 0 was derived from resampling the 1 arc second 30 m gridded DEM ANZCW0703013355 The DEM represents ground surface topography and excludes vegetation features The dataset was derived from the 1 second Digital Surface Model DSM ANZCW0703013336 by automatically removing vegetation offsets identified using several vegetation maps and directly from the DSM The 1 second product provides substantial improvements in the quality and consistency of the data relative to the original SRTM data but is not free from artefacts Man made structures such as urban areas and power line towers have not been treated The removal of vegetation effects has produced satisfactory results over most of the continent and areas with defects are identified in the quality assessment layers distributed with the data and described in the User Guide Geoscience Australia and CSIRO 2011 A full description of the methods is in progress Read et al in prep Gallant et al in prep The 3 second DEM was produced for use by government and the public under Creative Commons attribution The 3 second DSM and smoothed DEM are also available as a product set DSM ANZCW0703014216 DEM S AN
42. Elevation Models User Guide v1 0 4 Page 96 of 106 Contact information Contact organisation Geoscience Australia GA Contact position Director Sales and Distribution Mail address GPO Box 378 Mail address Locality Canberra State ACT Country Australia Postcode 2601 Telephone 61 2 6249 9966 Facsimile 61 2 6249 9960 Electronic mail address sales ga gov au Metadata information Metadata date 2010 08 30 Metadata Updated date 2011 09 01 Additional metadata Conversion to floating point format As a by product of the de striping process the integer data was converted to floating point format to allow for the continuously varying nature of the striping Areas where no de striping was required will contain unaltered integer values but represented in floating point format for consistency The 3 second data was produced as integer values as the decimal values are well below any effect on the accuracy Ancillary data layers distributed with the data A water mask at 1 second resolution showing the cells that are part of the flattened water bodies JPEG Image of the 3 second DEM References Gallant J C Read A M Dowling T I and Austin J M in prep Vegetation Removal methods used in SRTM 1 Second processing Geoscience Australia 2008 GEODATA 9 Second DEM Version 3 Geoscience Australia and CSIRO 2011 1 Second SRTM Derived Digital Elevation Models User Guide Version 1 0 4 Geoscience Australia Gr
43. I pe i TE he EFA EIF shi we Gide wean AE TAN ra Figure 9 Poor vegetation removal much improved but many vegetation features remain Euroa VIC 145 5E 36 75 Elevation range 120 240 m One second SRTM Derived Digital Elevation Models User Guide v1 0 4 Page 20 of 106 Smoothing The 1 second DEM after vegetation removal still contains the random noise present in the original SRTM data The noise typically alters elevations by 2 3 m but in some cases by as much as 10 m In high relief areas where elevations change by many m from one grid cell to the next this noise is of little consequence In low relief areas the noise is usually larger than the actual height differences from one cell to the next and corrupts calculated surface properties like slope and flow direction that depend on local height differences The true topographic height variations become apparent over longer distances as the variations due to noise are averaged out Averaging over larger areas effectively eliminates noise but also smooths out real topography while averaging over small areas does not produce enough smoothing to eliminate noise in relatively flat areas The smoothing approach used to produce DEM S adapts the scale of smoothing in response to local relief and noise levels Broader scale averaging is used where the noise is large relative to the local relief while steep areas are left untouched or smoothed only slightly The effect of the Smooth
44. SRTM Some small islands within water bodies are represented as voids within the SRTM due to edit rules These voids are filled as part of the void filling process and their elevations set to a minimum of 1 cm above surrounding water surface across the entire void fill Overview of quality assessment The quality of vegetation offset removal was manually assessed on a 1 8 x1 8 degree grid Issues with the vegetation removal were identified and recorded in ancillary data layers The assessment was based on visible artefacts rather than comparison with reference data so relies on the detection of artefacts by edges The issues identified were e vegetation offsets are still visible not fully removed e vegetation offset over estimated e linear vegetation offset not fully removed e incomplete removal of built infrastructure and other minor issues DEM ancillary data layers The vegetation removal and assessment process produced two ancillary data layers e lt A shapefile of 1 8 x 1 8 degree tiles indicating which tiles have been affected by vegetation removal and any issue noted with the vegetation offset removal The water and void fill masks for the 1 second DSM were also applied to the DEM Further information is provided in the User Guide Geoscience Australia and CSIRO 2011 Resampling to 3 seconds The 1 second SRTM derived Digital Elevation Model DEM was resampled to 3 seconds of arc 90 m in ArcGIS software using an aggregation too
45. ZCW0703014217 ANZLIC search words LAND Topography Models ECOLOGY Landscape Geographic extent name AUSTRALIA EXCLUDING EXTERNAL TERRITORIES AUS Australia Australia Geographic bounding box North bounding latitude 10 South bounding latitude 44 East bounding longitude 154 West bounding longitude 113 Data currency Beginning date 2000 2 11 Ending date 2000 2 22 Dataset status Progress Version 1 0 of the 1 second bare earth DEM is complete as at 23 December 2009 Maintenance and update frequency Updates and revisions are anticipated primarily to incorporate improvements to the bare earth DEM over time Further revisions are likely once the 1 second products have been released One second SRTM Derived Digital Elevation Models User Guide v1 0 4 Page 93 of 106 Reference system Horizontal datum WGS84 Vertical datum EGM96 Access Stored data format DIGITAL ArcGIS grid Arclnfo grid Available format type DIGITAL ArcGIS grid Arclnfo grid Access constraints This data is released under the Creative Commons Attribution 3 0 Australia Licence for use by Government and the public http creativecommons org licenses by 3 0 au Copyright Commonwealth of Australia Geoscience Australia 2010 Data quality Lineage Source data 1 SRTM 1 second Version 2 data Slater et al 2006 supplied by Defence Imagery and Geospatial Organisation DIGO as 813 1 x 1 degree tiles Data was produced b
46. alues along a color ramp Import Classified F Value Label 2221 051055 High 2221 08 53375288 Low 65 9753 Color Ramp Display Background Value Use hillshade effect Display MoData as Stretch Type Standard Deviations Mone Standard Deviations Histogram Equalize Minimum Mazimum Histogram Specification If it asks for you to Compute Histogram say yes with the same for Calculate Statistics It is recommended that the Display NoData and Display Background Value as No Colour Now you are ready to query or view the data One second SRTM Derived Digital Elevation Models User Guide v1 0 4 Page 105 of 106 You can also view the data in ArcCatalog by clicking the Preview tab after navigating to the location of the data If you have not connected to a drive click the Connect to Folder button o to view a drive You will be asked to Build Pyramids if you have not already done this see above Into Pitney Bowes Mapl nfo Open MapInfo Professional Go to File gt Open Select ESRI grid and then navigate to the location of the SRTM data Select the hdr adf file Header File and press open This should load the ESRI data once it has created a TAB file in the folder Please consult your software company for technical support if required One second SRTM Derived Digital Elevation Models User Guide v1 0 4 Page 106 of 106
47. am line errors affect the hydrological quality of DEM H the other problems create incorrect elevations but the hydrological connectivity is correctly represented Permanent Survey Mark Data A total of 1198 Permanent Survey Marks PSM made available through State land survey agencies were used to assess the overall vertical accuracy of the data at the national level The PSM data uses AHD71 for the vertical datum and GDA94 for the horizontal datum Figure 41 shows the spatial distribution of points and the height differences relative to AHD71 for the 1 second DEM Figure 42 shows the histogram of differences SCIMS PSM Points AHD Difference metres 9 9 8 0 7 9 6 0 5 9 4 0 3 9 2 0 1 9 0 0 O 0 1 2 0 2 1 4 0 41 6 0 6 1 8 0 8 1 12 0 12 1 16 0 16 1 24 0 24 1 32 0 1 000 Km bool Figure 41 Height difference relative to the AHD71 vertical datum between the DEM and the PSM points One second SRTM Derived Digital Elevation Models User Guide v1 0 4 Page 51 of 106 Difference Histogram Frequency Figure 42 Height difference distribution between the DEM and the PSM points relative to the AHD71 vertical datum The following results were observed relative to the AHD71 vertical datum 1 287 1 668 9 882 31 285 3 868 DEM PSM points height statistics It is important to note that the PSM data are generally associated with open non vegetated and relatively flat
48. and Information Council Core Metadata Guidelines Version 2 Dataset citation ANZLIC unique identifier ANZCWO703014615 Title 1 Second SRTM Derived Hydrological Digital Elevation Model DEM H Version 1 0 Custodian Custodian Geoscience Australia Jurisdiction Australia Description Abstract The 1 second SRTM derived DEM H Version 1 0 is a 1 arc second 30 m gridded digital elevation model DEM that has been hydrologically conditioned and drainage enforced The DEM H captures flow paths based on SRTM elevations and mapped stream lines and supports delineation of catchments and related hydrological attributes The dataset was derived from the 1 second smoothed Digital Elevation Model DEM S ANZCWO703014016 by enforcing hydrological connectivity with the ANUDEM software using selected AusHydro V1 6 February 2010 1 250 000 scale watercourse lines ANZCWO503900101 and lines derived from DEM S to define the watercourses The drainage enforcement has produced a consistent representation of hydrological connectivity with some elevation artefacts resulting from the drainage enforcement A full description of the methods is in preparation Dowling et al in prep This product is the last of the Version 1 0 series derived from the 1 second SRTM DSM DEM DEM S and DEM H and provides a DEM suitable for use in hydrological analysis such as catchment definition and flow routing The 1 second products DEM DEM S and DEM H have been released
49. andom variation noise that is spatially uncorrelated beyond distances of about 100 m but there are some broader scale errors The noise component is typically about 2 m but in some areas is much larger See Rodriguez et al 2006 for more information The removal of striping artefacts improves the representation of the landform shape particularly in low relief areas but it is not clear whether this also produces an improvement in overall height accuracy Some striping remains in the data at a much reduced level mostly less than 0 3 m amplitude Additional artefacts including long wavelength 10km striping have not been corrected Height accuracy Is likely to be poorer in areas where voids have been filled using the GEODATA 9 second DEM particularly in high relief areas Logical Consistency The DSM represents heights of the land surface or buildings or vegetation above the land surface Due to random noise the relative elevation between adjacent grid cells can be in error by several m The removal of striping has improved the representation of local landform shape particularly in low relief areas All void areas have been filled and there are no discontinuities due to tile boundaries One second SRTM Derived Digital Elevation Models User Guide v1 0 4 Page 68 of 106 The SRTM editing rules relating to water bodies have been respected in the processing lakes are flat rivers descend continuously in a downstream direction and sea
50. areas of very low relief The smoothing was performed on overlapping tiles with sufficient overlap that cells used in the final product were not impacted by edge effects In essence the smoothing process operated by comparing the variance of elevations in a 3x3 group of cells with the mean noise variance in the group If the elevation variance was larger than the mean noise it was considered to be due to real topographic variation and the elevations were left unchanged while if it was smaller it was One second SRTM Derived Digital Elevation Models User Guide v1 0 4 Page 78 of 106 considered to be due to noise and the elevations were replaced by the mean elevation in the group This was applied at successively coarser resolutions producing smoothing over large areas where the topographic variation was small compared to the noise levels The algorithm used statistical tests to make the decisions and combined the multiple estimates of elevation at different resolutions using variance weighting Water bodies Water bodies defined from the SRTM Water Body Data as part of the DSM processing were set to the same elevations as in the DSM after the smoothing The water bodies were also removed from the DEM set to null before the smoothing operation to prevent them unduly affecting the land elevations One cell of water adjacent to land is retained to prevent shoreline elevations from being raised to match the higher elevations further from the
51. as it becomes available http www ga gov au topographic mapping digital elevation data html Feedback This is an evolving product which requires government support and feedback to improve the accuracy of data and to refine processing techniques Please direct feedback to elevation ga gov au References Ehlschlaeger C R 1989 Using the AT search algorithm to develop hydrologic models from digital elevation data Proceedings of the International Geographic Information System IGIS Symposium Baltimore MD 275 281 http chuck ehlschlaeger info older 1GIS paper html ERSDAC June 2009 ASTER Global DEM Validation Summary Report Earth Remote Sensing Data Analysis Centre Japan http www gdem aster ersdac or jp index jsp Farr T G Rosen P A Caro E Crippen R et al 2007 The shuttle radar topography mission Reviews of Geophysics 45 RG2004 Hutchinson M F 1988 Calculation of hydrologically sound digital elevation models Proceedings of the Third International Symposium on Spatial Data Handling August 17 19 1988 Sydney Australia 117 133 International One second SRTM Derived Digital Elevation Models User Guide v1 0 4 Page 64 of 106 Geographical Union Commission on Geographical Data Sensing and Processing Ohio State University Columbus Ohio Hutchinson M F 1989 A new procedure for gridding elevation and stream line data with automatic removal of spurious pits Journal of Hydrology
52. ata available the uniqueness of the information the requirement to protect source capability and an assessment of the net benefit societal and otherwise of disseminating the data compared to restricting access Subject to developments in technology and capabilities these release constraints will likely be revisited in the future A licence agreement is required to obtain the SRTM derived 1 second DSM product The data Is available to government agencies and their collaborators and contractors who sign a copy of the licence and return it to Geoscience Australia The 1 second DSM is not available to universities or students unless they are working on a government project The licence agreement will cover all versions of DSM product derived from the SRTM data once signed Under the agreement with DIGO Geoscience Australia is required to keep a record of all government agencies that have received the DSM data The data is subject to Commonwealth of Australia Copyright The 1 second DSM data is provided upon request to eligible parties by contacting elevation ga gov au One second SRTM Derived Digital Elevation Models User Guide v1 0 4 Page 6 of 106 Data Schema Data Format Data is stored as continuous 32 bit Floating Point ESRI Grids tiles and mosaic and ESRI shapefiles for some reference data One second 30 m Grid tiles are named per the latitude and longitude of the south west corner A suffix of dem1_0 dems1_0 or demh1_
53. ce The GEODATA 9 second DEM data were refined to 1 second resolution using ANUDEM 5 2 without drainage enforcement Delta Surface Fill Method calculates height differences between SRTM and infill data to create a delta surface with voids where the SRTM has no values then interpolates across voids The void is then replaced by infill DEM adjusted by the interpolated delta surface resulting in an exact match of heights at the edges of each void Two changes to the Delta Surface Fill Method were made interpolation of the delta surface was achieved with natural neighbour interpolation Sibson 1981 implemented in ArcGIS 9 3 rather than inverse distance weighted interpolation and a mean plane inside larger voids was not used Water bodies Water bodies defined from the SRTM Water Body Data as part of the DSM processing were set to the same elevations as in the DSM One second SRTM Derived Digital Elevation Models User Guide v1 0 4 Page 94 of 106 Edit rules for land surrounding water bodies SRTM edit rules set all land adjacent to water at least 1 m above water level to ensure containment of water Slater et al 2006 Following vegetation removal void filling and water flattening the heights of all grid cells adjacent to water were set to at least 1 cm above the water surface The smaller offset 1cm rather than 1 m could be used because the cleaned digital surface model is in floating point format rather than integer format of the original
54. cent to water at least 1 m above water level to ensure containment of water Slater et al 2006 Following de striping void filling and water flattening the heights of all grid cells adjacent to water was set to at least 1 cm above the water surface The smaller offset 1 cm rather than 1 m could be used because the cleaned digital surface model is in floating point format rather than integer format of the original SRTM Some small islands within water bodies are represented as voids within the SRTM due to edit rules These voids are filled as part of void filling process and their elevations set to a minimum of 1 cm above the surrounding water surface across the entire void fill DSM ancillary data layers Four additional data layers were used to make alterations to the raw SRTM data to produce the 1 second DSM e A de stripe mask indicating which x Y4 degree tiles have been affected by de striping and which have not been de striped e A striping magnitude layer showing the amplitude of the striping at 0 01 degree 1km resolution e A water mask at 1 second resolution showing the cells that are part of the flattened water bodies e Avoid mask showing cells that were no data in the raw SRTM and have been filled using the void filling algorithm Re sampling to 3 seconds The 1 second SRTM derived Digital Surface Model DSM mosaic was resampled to 3 seconds of arc 90 m in ArcGIS software using an aggregation tool This tool determines a n
55. cludes all water features in the SWBD that are connected to the ocean and have zero elevation Also note that there are some areas particularly on the southern margin of the Gulf of Carpentaria where there are water bodies with zero elevation in the original SRTM data and in DSM near the coast but separated from the coast by a narrow strip of land often covered by mangroves These are typically areas subject to tidal inundation and should not be considered part of the ocean and are not included in the ocean mask Figure 16 illustrates many of these effects SWBD Ocean PSA SWBD River Figure 16 An estuary flowing into the Gulf of Carpentaria showing DSM on the left and DEM H on the right The effect of SRTM edit rules on original elevations adjacent to the coast is apparent in the DSM The SRTM Water Body Data SWBD distinguishes between ocean and river in the absence of elevation changes but the ocean mask includes all water cells adjacent to the ocean with O m elevation yellow and these cells are set to NODATA white in DEM H Also note the elevations below zero blue in DEM H adjacent to the ocean the edit rules are not applied to DEM H One second SRTM Derived Digital Elevation Models User Guide v1 0 4 Page 30 of 106 Dataset Examples This section shows some examples of the 1 second bare earth DEM from a range of landscapes around Australia highlighting the capabilities of this DEM Figures 17 to 23 The exam
56. degree than steep areas and to respond to the degree of noise so that very noisy flat areas are smoothed more than less noisy flat areas The process operated over multiple resolutions allowing smoothing over quite large distances in areas of very low relief The smoothing was performed on overlapping tiles with sufficient overlap that cells used in the final product were not impacted by edge effects The smoothing process was based on the amount of noise in the 1 second DEM The noise was estimated from the local variation in the difference between elevation and the mean of nearby elevations In essence the smoothing process operated by comparing the variance of elevations in a 3x3 group of cells with the mean noise variance in the group If the elevation variance was larger than the mean noise it was considered to be due to real topographic One second SRTM Derived Digital Elevation Models User Guide v1 0 4 Page 100 of 106 variation and the elevations were left unchanged while if it was smaller it was considered to be due to noise and the elevations were replaced by the mean elevation in the group This was applied at successively coarser resolutions producing smoothing over large areas where the topographic variation was small compared to the noise levels The algorithm used statistical tests to make the decisions and combined the multiple estimates of elevation at different resolutions using variance weighting Water bodies Water bodi
57. e Radar Topography Mission SRTM derived Digital Surface Model DSM Version 1 0 was derived from resampling the 1 arc second 30 m gridded DSM ANZCW0703013336 that represents ground surface topography as well as features above the ground such as vegetation and man made structures The 1 second DSM was derived from the SRTM data acquired in February 2000 supported by the GEODATA 9 second DEM in void areas and the SRTM Water Body Data Stripes and voids have been removed from the 1 second SRTM data to provide an enhanced and complete DSM for Australia and near shore islands A full description of the methods is In progress Read et al in prep The 3 second DEM was produced for use by government and the public under Creative Commons attribution Further information can be found in the User Guide The 1 second DSM forms the source for the 1 second DEM with vegetation offsets removed ANZCWO703013355 and the smoothed version ANZCW0703014016 All 1 second products resampled to 3 seconds are available DSM ANZCWO703014216 DEM ANZCW0703014182 DEM S ANZCW0703014217 ANZLIC search words LAND Topography Models ECOLOGY Landscape Geographic extent name AUSTRALIA EXCLUDING EXTERNAL TERRITORIES AUS Australia Australia Geographic bounding box North bounding latitude 10 South bounding latitude 44 East bounding longitude 154 West bounding longitude 113 Data currency Beginning date 2000 2 11 Ending date 200
58. e global dataset it was therefore decided not to make any corrections between EGM96 and AHD71 in Version 1 of the derived datasets In future versions the minor correction to the AHD71 vertical height datum will be undertaken following more thorough analysis of higher quality datasets For most purposes the SRTM data can therefore be considered to be AHD heights In order to quantify the absolute vertical accuracy of the datasets relative to AHD71 analyses have been undertaken at national State and local levels A number of existing elevation products were used to compare the vertical accuracy of the surface including permanent survey mark data PSM sample contour data for the Atherton Tableland area Queensland a LIDAR derived DEM for Lower Darling and the Victorian DEM State wide product VicMap Elevation DTM 20 m Each of these analyses is described below The accuracy assessment was completed on the base product the 1 second DEM Analysis showed little difference in the vertical accuracy of the DSM DEM and DEM S Relative elevation accuracy between adjacent cells is improved in DEM S and DEM H due to the reduction in noise levels this has not been quantified but is evident in the comparison of slopes calculated before and after smoothing as shown in this User Guide refer to the Smoothing Section The elevation error for DEM H is difficult to characterise In general it will be similar to the raw SRTM 1 second data with 90 of test
59. e the stripes are still apparent Figures 2 amp 3 One second SRTM Derived Digital Elevation Models User Guide v1 0 4 Page 11 of 106 e Y gt A Figure 2 Example of good stripe removal Southesk Tablelands in Great Sandy Desert WA 126 3E 20 2S Elevation range 220 320 m One second SRTM Derived Digital Elevation Models User Guide v1 0 4 Page 12 of 106 A ci Shi AR E Figure 3 Example of poor stripe removal there are also some voids among the stripes Bogan River near Brewarrina NSW 146 7E 30 25 Elevation range 105 160 m One second SRTM Derived Digital Elevation Models User Guide v1 0 4 Page 13 of 106 Void filling The SRTM DSM contains voids areas without data where the surface did not produce a good radar signal There are several reasons for these voids e Steep areas like canyons where the radar could not see the ground because of its 45 look angle Figure 4 e Water bodies that did not reflect a radar signal back to the Shuttle Figure 5 e Dry sandy areas that did not reflect a radar signal back to the Shuttle Figure 6 Voids are filled by replacing the missing data with elevations from another source in this case the GEODATA 9 second DEM While this DEM is much lower in resolution than the SRTM data it provides a much better representation of the landscape in the steep areas than just filling in the missing areas by interpolation Since completing the void filling the ASTER G DEM ERSDAC
60. eam lines where none were mapped e g in tile e129s25 e Errors in the 1 250 000 stream lines e g 148 29 E 35 35 S One second SRTM Derived Digital Elevation Models User Guide v1 0 4 Page 84 of 106 Note that only the last two issues extraneous infill lines and stream line errors affect the hydrological quality of DEM H the other problems create incorrect elevations but the hydrological connectivity is correctly represented DEM H ancillary data layers Flow direction grids along watercourses have been included in 1 degree tiles Positional accuracy The positional accuracy of watercourses in flatter areas is the same as for the 1 250 000 stream line data about 200 m For other features the horizontal positional error is generally the same as for the raw SRTM 1 second data with 90 of tested locations within 7 2 m for Australia See Rodriguez et al 2006 for more information Attribute accuracy The primary purpose of a hydrological DEM is to support hydrological analysis related to connectivity of flow paths and hydrological properties of catchments and stream lines The combination of ANUDEM and the CheckStreamDescent analysis ensures that DEM H V1 0 correctly represents flow pathways as defined by the AusHydro 1 250 000 mapping modified by the infill streams defined from DEM S Flow paths where there are no mapped streams reflect the surface topography as represented in DEM S so will be affected by errors in that DEM In
61. ed heights within 9 8 m for Australia but significant changes to elevation have occurred due to the smoothing and drainage enforcement processes Differences in height between DEM S and DEM H were examined to identify areas where defects were created by the drainage enforcement process Some large elevation differences up to 290 m were due to valid drainage enforcements in canyons Other significant differences are related to various problems including e Excessive height reductions on steep slopes due to multiple parallel infill stream lines e g 152 295 E 30 943 S e Excessive smoothing lowering of hilltops and raising of lower slopes in some areas e g the eastern peaks of the Stirling Range WA around 118 28 E 34 36 S with hilltops lowered by around 200 m e Drainage enforcements to the level of open cut mines traversed by mapped stream lines resulting in deep incisions extending long distances downstream of the mines with the worst instance being from the coal mines in Latrobe Valley Victoria to the outlet of the Gippsland Lakes to an elevation of about 60 m for about 180 km e A few extraneous infill stream lines in inland areas creating long stream lines where none were mapped e g in tile e129s25 e Errors in the 1 250 000 stream lines e g 148 29 E 35 35 S One second SRTM Derived Digital Elevation Models User Guide v1 0 4 Page 50 of 106 Note that only the last two issues extraneous infill lines and stre
62. elg River in south western Victoria before and after removal of vegetation offset The areas around the river are covered in mostly low vegetation probably with little impact on the DSM Unfortunately the edge of the vegetated area corresponds to the edge of the river gorge so the difference in height between vegetated and non vegetated areas includes the depth of the gorge The adjustment for this apparent vegetation offset almost eliminates the gorge itself One second SRTM Derived Digital Elevation Models User Guide v1 0 4 Page 39 of 106 Where the over estimation is associated with regularly shaped patches of trees the effect can be seen as shallow depressions bounded by relatively straight lines corresponding to the edges of the mapped area of trees Vegetated dunes can also be subject to this problem and it results in either attenuation or removal of the dune features in the DEM Small tree covered hills in cleared landscapes may also be affected by this problem Figure 29 The outlet of the Glenelg River Victoria around 141 00E 38 005 adjacent to the area of Figure 28 The gorge has almost disappeared because the cliffs have been mistakenly identified as vegetation offsets Elevation range O 150 m One second SRTM Derived Digital Elevation Models User Guide v1 0 4 Page 40 of 106 Incomplete removal of urban and built infrastructure Buildings and structures above the ground are seen by the SRTM radar if they are sufficientl
63. elta Surface Fill Method Grohman et al 2006 was adapted for this task using GEODATA 9 second DEM as the infill data source The 9 second data was refined to 1 second resolution using ANUDEM 5 2 without drainage enforcement Delta Surface Fill Method calculates height differences between SRTM and infill data to create a delta surface with voids where the SRTM has no values then interpolates across voids The void is then replaced by infill DEM adjusted by the interpolated delta surface resulting in an exact match of heights at the edges of each void Two changes to the Delta Surface Fill Method were made interpolation of the delta surface was achieved with natural neighbour interpolation Sibson 1981 implemented in ArcGIS 9 3 rather than inverse distance weighted interpolation and a mean plane inside larger voids was not used Water bodies Water bodies defined from the SRTM Water Body Data as part of the DSM processing were set to the same elevations as in the DSM Edit rules for land surrounding water bodies SRTM edit rules set all land adjacent to water at least 1 m above water level to ensure containment of water Slater et al 2006 Following vegetation removal void filling and water flattening the heights of all grid cells adjacent to water were set to at least 1 centimetre above the water surface The smaller offset 1 cm rather than 1 m could be used because the cleaned digital surface model is in floating point format rath
64. emoval in the 1 second DEM and has an over estimated elevation by 4 16 m 2 Some areas mapped as woody vegetation particularly the forests in north west Victoria have been lowered too much by the vegetation offset removal the height of vegetation has been over estimated 3 These rectangular patterns of difference have not been explained but do not appear to be an SRTM artefact It is thought that this is a photogrammetric One second SRTM Derived Digital Elevation Models User Guide v1 0 4 Page 57 of 106 error that has been passed onto the VicMap DTM as this effect was also seen in the GEODATA 9 second DEM 4 Residual striping from the SRTM data is visible in the north western part of the State These are being carried through into the difference surface from the SRTM DEM and are typically of a magnitude of around 1 2 m 5 The Orange Red colour 16 24 m occurs in areas where the vegetation has not been sufficiently removed or treated This can occur for a range of reasons Riparian and remnant vegetation which was not adequately mapped and therefore hard to remove from the SRTM derived DEM Continuously forested hilly areas where the vegetation offset has been systematically under estimated e Areas of narrow gorges or cliffs where due to the angle of the SRTM no readings were recorded for the valley floor and this results in a more generalized valley Lake and water levels for the SRTM are set to highest water mark th
65. er than integer format of the original SRTM Some small islands within water bodies are represented as voids within the SRTM due to edit rules These voids are filled as part of void filling process and their elevations set to a minimum of 1 cm above surrounding water surface across the entire void fill Overview of quality assessment The quality of vegetation offset removal was manually assessed on a 1 8 x1 8 degree grid Issues with the vegetation removal were identified and recorded in ancillary data layers The assessment was based on visible artefacts rather than comparison with reference data and relies on the detection of artefacts by edges The issues identified were vegetation offsets are still visible not fully removed vegetation offset over estimated linear vegetation offset not fully removed incomplete removal of built infrastructure and other minor issues DEM ancillary data layers The vegetation removal and assessment process produced two ancillary data layers e lt A shapefile of 1 8 x 1 8 degree tiles indicating which tiles have been affected by vegetation removal and any issue noted with the vegetation offset removal One second SRTM Derived Digital Elevation Models User Guide v1 0 4 Page 73 of 106 e A difference surface showing the vegetation offset that has been removed this shows the effect of vegetation on heights as observed by the SRTM radar instrument and is related to vegetation height density and structure
66. eral vegetation maps and directly from the DSM This product provides substantial improvements in the quality and consistency of the data relative to the original SRTM data but is not free from artefacts Man made structures such as urban areas and power line towers have not been treated The removal of vegetation effects has produced satisfactory results over most of the continent and areas with defects are identified in the quality assessment layers distributed with the data and described in the User Guide Geoscience Australia and CSIRO 2011 A full description of the methods is in progress Read et al in prep Gallant et al in prep Smoothed DEM DEM S ANZCW0703014016 was released in August 2010 as a derivative product of the DEM and the DSM ANZCWO703013336 and the drainage enforced version DEM H ANZCWO703014615 was released in October 2011 The three products DEM DEM S and DEM H have been released under Creative Commons licensing since October 2011 ANZLIC search words LAND Topography Models ECOLOGY Landscape Geographic extent name AUSTRALIA EXCLUDING EXTERNAL TERRITORIES AUS Australia Australia Geographic bounding box North bounding latitude 10 South bounding latitude 44 East bounding longitude 154 West bounding longitude 113 Data currency Beginning date 2000 2 11 Ending date 2000 2 22 Dataset status Progress Version 1 0 of the 1 second bare earth DEM is complete as at 23 December
67. erence of less than 29 97 m One second SRTM Derived Digital Elevation Models User Guide v1 0 4 Page 63 of 106 The following results were observed relative to the AHD71 vertical datum 7 029 3 second DEM S PSM points height statistics Choosing the Correct 3 Second Product Given the accuracy of the 3 second product it is advised that you consider the desired application of the data and which of the 3 second products to use This product has been released in good faith that the user understands the limitations and inherent errors in the data The data should not be solely relied upon for decision making The 3 second products are not suitable for finer scale applications requiring accuracy greater than the specified vertical accuracy of approximately 21m There is currently no hydrologically enforced 3 second product If you require a hydrologically enforced product use either the 1 second DEM H or the 9 second GEODATA3 DEM which are available through the GA Sales Centre although these have a poorer accuracy A 3 second DEM H may be produced in the future if there is sufficient demand Future Developments The products described in this User Guide are version 1 0 meaning that they are the relatively new versions of each product Work is continuing to treat some of the known issues and further releases of the products are planned in 2012 Product information will be added to the Geoscience Australia Digital Elevation Data webpage
68. erstand the 3 second DSM the processing of the parent dataset the 1 second DSM is described below 1 second DSM processing The 1 second SRTM derived Digital Surface Model DSM was derived from the 1 second Shuttle Radar Topography Mission data by removing stripes filling voids and re flattening water bodies Further details are provided in the 1 second DSM metadata ANZCWO703013336 and the User Guide Geoscience Australia and CSIRO 2010 De striping SRTM data contains striping artefacts oriented approximately NE SW and NW SE that vary in amplitude from about 0 2 m to nearly 4 m The wavelength of the striping is approximately 800 m Stripes were detected in the elevation data using a 2 dimensional Fast Fourier Transform Peaks in the spectra were visually identified and manually delineated using a tool designed specifically for this purpose Striping occurred everywhere except where relief was high enough to obscure striping Spectral analysis was performed on sub tiles to account for spatial variation in the intensity and direction of striping Fourier transform was applied to overlapping sub tiles covering 1536 x 1536 cells 0 43 x 0 43 degrees Central 1024 x 1024 cells were retained each comprising one sixteenth of a 1 x 1 degree tile 900 x 900 cells with a 62 cell overlap on each edge to provide smooth transitions between sub tiles Void filling Voids areas without data occur in the data due to low radar reflectance typically
69. es defined from the SRTM Water Body Data as part of the DSM processing were set to the same elevations as in the DSM after the smoothing The water bodies were also removed from the DEM set to null before the smoothing operation to prevent them affecting the land elevations unduly One cell of water adjacent to land is retained to prevent shoreline elevations from being raised to match the higher elevations further from the shore Re sampling to 3 seconds The 1 second SRTM derived smoothed Digital Elevation Model DEM S was re sampled to 3 seconds of arc 90 m in ArcGIS software using an aggregation tool This tool determines a new cell value based on multiplying the cell resolution by a factor of the Input in this case three and determines the mean value of input cells with the new extent of the cell i e the mean value of the 3x3 input cells The 3 second SRTM was left in floating point format which does make this dataset slower to open run Further information on the processing is provided in the User Guide Geoscience Australia and CSIRO 2011 Positional accuracy The horizontal positional error is estimated to be three times that of the 1 second products The 1 second products are the same as the raw SRTM 1 second data with 90 of tested locations within 7 2 m for Australia See Rodriguez et al 2006 for more information on SRTM accuracy Attribute accuracy Accuracy of the 3 second DEM S was tested using the same 1198 Per
70. etails are provided in the DEM metadata ANZCW0703013355 DEM S adaptive smoothing The DEM was smoothed by averaging elevations over distances ranging from 90 m to several kms depending on the level of noise and the local relief The smoothing removes most of the local noise and allows measurement of slopes down to less than 0 1 Further details are provided in the DEM S metadata ANZCW0703014016 One second SRTM Derived Digital Elevation Models User Guide v1 0 4 Page 83 of 106 Drainage enforcement The 1 second Hydrological Digital Elevation Model DEM H was derived from the 1 second Smoothed Digital Elevation Model DEM S using the ANUDEM software Hutchinson 1988 1989 2009 version 5 2 5 dated 1 December 2010 This version of ANUDEM was modified to work effectively on the SRTM data as briefly described in Hutchinson et al 2009 ANUDEM uses a spline interpolation method that smooths the surface enforces continuous descent along supplied drainage lines and removes sinks consistent with the accuracy of the source elevation data DEM S The 1 250 000 scale stream line data used to produce the GEODATA 9 second DEM Version 3 was chosen as the source of drainage line data as it was the only available source of cleaned and correctly oriented drainage lines and it covered the entire continent This 1 250 000 scale data with a spatial accuracy of about 200 m was Significantly coarser than the 1 second DEM S that is accurate to 50 m or
71. ew cell value based on multiplying the cell resolution by a factor of the input in this case three and determines the mean value of input cells with the new extent of the cell i e the mean value of the 3x3 input cells The 3 second DSM mosaic was converted to integer format to make the file size more manageable It does not affect the accuracy of the data at this resolution Further information on the processing is provided in the User Guide Geoscience Australia and CSIRO 2010 Positional accuracy The horizontal positional error is estimated to be three times that of the 1 second products The 1 second products are the same as the raw SRTM 1 second data with 90 of tested locations within 7 2 m for Australia See Rodriguez et al 2006 for more information on SRTM accuracy Attribute accuracy Elevation accuracy is essentially three times the raw SRTM 1 second data accuracy with 90 of tested heights within 9 8 m for Australia which makes the 3 second DSM accuracy about 29 m Errors in height are still mostly due to random variation noise that is spatially uncorrelated beyond distances of about 100 m 1 second DSM but there are some broader scale errors The noise component Is typically about 2 m in One second SRTM Derived Digital Elevation Models User Guide v1 0 4 Page 90 of 106 the 1 second DSM but in some areas is much larger See Rodriguez et al 2006 for more information The removal of striping artefacts impr
72. f the edit rules applied to ensure that land areas adjacent to coast are at least 0 01 m Note again that the edit rules only apply to cells immediately adjacent to the coast so cells further inland can have negative elevations In some cases this is due to over estimation of vegetation heights in other cases it is due to the original SRTM data The edit rules were not applied after the adaptive smoothing to produce DEM S so there are some areas where land elevations adjacent to water bodies including the ocean are lower than the adjacent water elevations In DEM H ocean areas have been set to NODATA water bodies have not been re flattened and the edit rules have not been applied to avoid corrupting the hydrological enforcement As a result there are some areas immediately adjacent to the coast with elevations below zero The ocean and water masks should be used to determine whether any cell belongs to the land or a water body rather than relying on elevation values above or below zero The ocean mask is 1 for ocean cells and NODATA for all other cells The water mask is 1 for all water bodies including the ocean and NODATA for land cells These masks have been derived from the SRTM Water Body Data SWBD which One second SRTM Derived Digital Elevation Models User Guide v1 0 4 Page 29 of 106 identifies water bodies as ocean lake or river with relatively arbitrary boundaries between river and ocean in estuaries The ocean mask in
73. ffsets as seen by the SRTM instrument The estimation of the vegetation offsets can also be under or over estimated if vegetation and topographic patterns coincide such as trees on hilltops or dune ridges or in inset floodplains or swamps The height offsets at vegetation edges are interpolated within vegetation patches to estimate the effects within the patches The best results tend to be in small patches such as remnant tree patches In continuously forested areas with few edges for estimating the offsets the heights are likely to be less reliable and there is no information at all on variations of the height offset within continuous forests The removal of vegetation has been quite effective overall but there are many areas that contain either untreated or incompletely treated vegetation effects such as the area shown in Figure 9 One second SRTM Derived Digital Elevation Models User Guide v1 0 4 Page 18 of 106 ak A p ra y O i 3 r alt r A F F i L 4 y F Ea F ut s ar E 7 pine j l pe P these g PA Figure 8 An example of effective removal of vegetation offset Culcairn NSW 147 0E 35 75 Elevation range 150 600 m One second SRTM Derived Digital Elevation Models User Guide v1 0 4 Page 19 of 106 Ar e b r a cart a F 3 i 7 7 gt 1 q p wy Ge Ae ov i a y d a Pi 7 ca e AN fae GE iO E pate r s i EA i Pon ee desd LACA E par AA AE Al af dl pza te A t b
74. has been discovered affecting the Grose Valley in the Blue Mountains of NSW Figure 26 This valley is surrounded by cliffs which resulted in voids around most of the valley floor The edges of the valley floor have been erroneously assigned heights consistent with the surrounding plateau ignoring the cliffs so they are about 200 m too high No other errors approaching this magnitude have been detected PE We ne r Oe s Fi f y J Y 4 y a lt a r A Figure 26 The black polygon encloses the affected area Grose Valley NSW 150 345E 33 602S Elevation range 100 1200 m One second SRTM Derived Digital Elevation Models User Guide v1 0 4 Page 37 of 106 Noise The SRTM DSM is affected by intrinsic noise due to the nature of the radar acquisition and processing Figure 27 also visible in many other images The noise has no directional character and has a short range correlation over a distance of about 100 m appearing as humps and hollows in flat areas This noise typically has an amplitude of 2 3 m but can be much larger up to about 10 m In areas of low relief this noise significantly impacts measures of local shape such as slope aspect flow direction and curvature It creates a multitude of small sinks and peaks although there are often real sinks in those landscapes too In steep areas it is essentially inconsequential The smoothed version of the dataset DEM S has most of this noise removed
75. http chuck ehlschlaeger info older GIS paper htm Farr T G Rosen P A Caro E Crippen R and others 2007 The shuttle radar topography mission Reviews of Geophysics 45 RG2004 Geoscience Australia 2008 GEODATA 9 Second DEM Version 3 Geoscience Australia and CSIRO 2011 1 Second SRTM Derived Digital Elevation Models User Guide Version 1 0 4 Geoscience Australia Hutchinson M F 1988 Calculation of hydrologically sound digital elevation models Proceedings of the Third International Symposium on Spatial Data Handling August 17 19 1988 Sydney Australia 117 133 International Geographical Union Commission on Geographical Data Sensing and Processing Ohio State University Columbus Ohio Hutchinson M F 1989 A new procedure for gridding elevation and stream line data with automatic removal of spurious pits Journal of Hydrology 106 211 232 Hutchinson M F 2009 ANUDEM Version 5 2 Fenner School of Environment and Society Australian National University Available online at http fennerschool anu edu au publications software anudem php last accessed January 2011 One second SRTM Derived Digital Elevation Models User Guide v1 0 4 Page 86 of 106 Hutchinson M F Stein J A Stein J L and Xu T 2009 Locally adaptive gridding of noisy high resolution topographic data In Anderssen R S R D Braddock and L T H Newham eds 18th World IMACS Congress and MODSIMO9 International Congress on Model
76. igure 51 Lower Darling Difference Grid DEM LIDAR 142 625E 32 477S Some 10 000 points were randomly selected across the project area to extract values for further statistical analysis Figure 52 The following results were obtained Statistics of Lower Darling LIDAR comparison One second SRTM Derived Digital Elevation Models User Guide v1 0 4 Page 61 of 106 Difference Histogram gt Do Cc oa 3 ml w i LL Figure 52 Histogram of Difference in Elevation between Lower Darling LIDAR and SRTM DEM A few significant differences were observed between the Lower Darling LiDAR generated DEM and the SRTM derived DEM The largest differences relate to riparian vegetation that has not been removed from the SRTM DEM Also the Striping apparent in the SRTM is visible in the difference surface see Striping Section One second SRTM Derived Digital Elevation Models User Guide v1 0 4 Page 62 of 106 Derivation of the 3 Second Products Processing of the 3 Second Products The processing described in Processing of the SRTM Data was completed on the 1 second data as the parent datasets As the 3 second data is a derived product there are inherent improvements in the 3 second products also This does not include the DEM H which has not been produced at a 3 second resolution and should be not resampled to a coarser resolution as the drainage would be affected Refer to the above section for further information Below
77. in a height difference of less than 9 602 m The removal of striping artefacts from the 1 second DEM improves the representation of the landform shape particularly in low relief areas but it is not clear whether this also produces an improvement in overall height accuracy Some striping remains in the data at a much reduced level mostly less than 0 3 m amplitude Additional artefacts including long wavelength 10km striping have not been corrected The removal of vegetation offsets in the 1 second DEM provides a significant Improvement in the representation of the landform shape particularly in low relief areas and areas of remnant vegetation Elevation accuracy varies in forested areas Comparisons with several higher resolution datasets suggest that elevation accuracy varies depending on the height and structure of the existing vegetation quality of vegetation input masks and local relief Further details of these comparisons are provided in the User Guide Geoscience Australia and CSIRO 2011 Height accuracy Is likely to be poorer in areas where voids have been filled using the GEODATA 9 second DEM particularly in high relief areas Logical Consistency The DEM represents elevation Due to random noise the relative elevation between adjacent grid cells can be in error by several m The removal of vegetation involves estimation of vegetation height at the edges of vegetation patches and interpolation of those heights across area
78. ing is therefore most apparent in the flattest and noisiest areas The adaptive smoothing process was designed to smooth flat areas to a greater degree than steep areas and to respond to the degree of noise so that very noisy flat areas are smoothed more than less noisy flat areas The process operated over multiple resolutions allowing smoothing over quite large distances in areas of very low relief The smoothing was performed on overlapping tiles with sufficient overlap that cells used in the final product were not impacted by edge effects In essence the smoothing process operated by comparing the variance of elevations in a 3x3 group of cells with the mean noise variance in the group If the elevation variance was larger than the mean noise it was considered to be due to real topographic variation and the elevations were left unchanged however if it was Smaller it was considered to be due to noise and the elevations were replaced by the mean elevation in the group This was applied at successively coarser resolutions producing smoothing over large areas where the topographic variation was small compared to the noise levels The algorithm used statistical tests to make the decisions and combined the multiple estimates of elevation at different resolutions using variance weighting Differences due to smoothing can be as large as 110 m although the maximum change is less than 50 m in 87 of tiles The standard deviation of elevation change
79. is was the same for VicMap DTM unless it used other readings lower than the high water mark 6 This area is an artefact in the VicMap DTM and is the result of a contour used to create that VicMap DTM which should not have existed or was mislabelled The extremes shown in pink and purple in the figure below in the difference surface are minor and insignificant The negative range of differences are randomly scattered in minute areas and the positive range of differences are the same These show that some lakes have been over estimated by the SRTM DEM in particular Lake Dartmouth north east of Mount Beauty Victoria 147 545E 36 574S Difference Surface m MI 140 7 54 MN 63 9 16 Figure 48 DEM VicMap DTM Difference Surface 141 306E 38 049S One second SRTM Derived Digital Elevation Models User Guide v1 0 4 Page 58 of 106 Significant differences were observed between the VicMap DTM and the SRTM derived DEM see table below A total of 4193 points were randomly created across the surface for a more in depth analysis The following statistics were obtained from the sample points DEM VicMap DTM Sample Points Statistics This produced the following difference histogram Figure 49 Difference Histogram Frequency Figure 49 DEM VicMap DTM Sample Points Histogram One second SRTM Derived Digital Elevation Models User Guide v1 0 4 Page 59 of 106 Lower Darling LIDAR The Lower Darling LIDAR comparison
80. is DEM is based on the 1 second SRTM derived Digital Surface Model DSM that was itself derived from the 1 second Shuttle Radar Topography Mission data The DSM was produced by removing stripes filling voids and re flattening water bodies Further details are provided in the DSM metadata ANZCW0703013336 The vegetation removal used the DSM without voids filled so that vegetation height estimates would not be affected by interpolated heights and so that voids adjacent to vegetated areas could be filled using bare earth elevations Vegetation offset removal The processing of vegetation offsets to produce the DEM relies on Landsat based mapping of woody vegetation to define where the offsets are likely to occur The mapped extents of woody vegetation were adjusted using an edge matching process to better represent the extents of areas affected by vegetation offsets in the SRTM DSM Vegetation was processed across approximately 40 of Australia as shown in the vegetation mask ancillary dataset and in the User Guide Geoscience Australia and CSIRO 2011 Vegetation offset processing involves detecting vegetation patches measuring the height offset around the edges interpolating the height offset across the vegetated areas and subtracting the offset from the DSM The heights of the offsets are estimated by measuring height differences across the boundaries of the vegetation patches The method provides good estimates of the offsets in flat landscape
81. is the coverage of the 3 second DEM S Figure 53 1 3 second SRTM m High 2221 08 O 250 500 1 000 Km Loa laa Low 65 9753 Figure 53 3 second National DEM coverage Resampling the data to 3 seconds from the 1 second product was completed in ESRI ArcGIS software with an aggregation tool using mean cell values This tool determines a new cell value based on multiplying the cell resolution by a factor of the input in this case three and determines the mean value of input cells with the new extent of the cell i e the mean value of the 3x3 input cells The 3 second DSM and DEM products have been converted to integer format to reduce the file size this is considered to have little impact on accuracy for these products The DEM S has been retained in floating point format to preserve the subtle variations in height that the adaptive smoothing method produces Accuracy Assessment The same Permanent Survey Mark PSM data comparison was conducted on the 3 second DEM S using the same 1198 points It was expected that the 3 second would be approximately three times that of the 1 second product given the resolution of the data and additional smoothing that was applied to the 3 second DEM and its parent 1 second product Results showed the absolute accuracy of the data as tested to be 14 54 m at the 95 percentile with a RMS error of 7 029 in open flat terrain Ninety nine percent of points are within a height diff
82. l Core Metadata Guidelines Version 2 Dataset citation ANZLIC unique identifier ANZCWO703014217 Title 1 Second SRTM Derived 3 second Smoothed Digital Elevation Model DEM S version 1 0 Custodian Custodian Geoscience Australia Jurisdiction Australia Description Abstract The 3 second 90 m Smoothed Digital Elevation Model DEM S Version 1 0 was derived from resampling the 1 second SRTM derived DEM S gridded smoothed digital elevation model ANZCWO703014016 The DEM represents ground surface topography excluding vegetation features and has been smoothed to reduce noise and improve the representation of surface shape The DEM S was derived from the 1 second Digital Surface Model DSM ANZCW0703013336 and the Digital Elevation Model Version 1 0 DEM ANZCW0703013355 by an adaptive smoothing process that applies more smoothing in flatter areas than hilly areas and more smoothing in noisier areas than in less noisy areas This DEM S supports calculation of local terrain shape attributes such as slope aspect and curvatures that could not be reliably derived from the unsmoothed 1 second DEM because of noise A full description of the methods is in progress Gallant et al in prep and in the 1 second User Guide The 3 second DEM was produced for use by government and the public under the Creative Commons attribution The 1 second DSM and DEM that form the basis of the product are also available as 3 second products DSM ANZCW0703
83. l This tool determines a new cell value based on multiplying the cell resolution by a factor of the input in this case three and determines the mean value of input cells with the new extent of the cell i e the mean value of the 3x3 input cells The 3 second SRTM was converted to integer format for the national mosaic to make the file size more manageable It does not affect the accuracy of the data at this resolution Further information on the processing Is provided in the User Guide Geoscience Australia and CSIRO 2011 Positional accuracy The horizontal positional error is estimated to be three times that of the 1 second products The 1 second products are the same as the raw SRTM 1 second data with 90 of tested locations within 7 2 m for Australia See Rodriguez et al 2006 for more information on SRTM accuracy Attribute accuracy The accuracy of the 3 second DEM is determined to be three times that of the accuracy of the 1 second DEM This is approximately 22 m Accuracy was tested on the 1 second DEM using 1198 Permanent Survey Marks distributed across the Australian continent relative to the Australian Height Datum AHD71 Results of this comparison show the absolute accuracy of the data as tested One second SRTM Derived Digital Elevation Models User Guide v1 0 4 Page 95 of 106 relative to AHD71 to be 7 582 m at the 95 percentile with a RMS error of 3 868 in open flat terrain Ninety nine percent of points are with
84. l islands defined by the available SRTM 1 second elevation and SRTM Water Body Data datasets The following tiles containing fragments of mainland or pieces of islands were not Supplied at 1 second resolution and are therefore missing from the DEM H One second SRTM Derived Digital Elevation Models User Guide v1 0 4 Page 85 of 106 E112 S26 E113 S29 E118 S20 E120 S35 E121 S35 E123 S16 E124 515 E125 S14 E132 11 E133 11 E134 S35 E141 S10 E142 510 E143 S10 E146 S17 E150 S22 E152 S24 Note that the coordinates are of the south western corner of the tile Contact information Contact organisation Geoscience Australia GA Contact position Director Sales and Distribution ISB Mail address GPO Box 378 Mail address Locality Canberra State ACT Country Australia Postcode 2601 Telephone 61 2 6249 9966 Facsimile 61 2 6249 9960 Electronic mail address sales ga gov au Metadata information Metadata Created date 2011 04 15 Metadata Updated date 2011 09 01 Additional metadata Metadata reference XHTML NA Metadata reference XML NA References Dowling T I Read A M Hutchinson M F and Gallant J C in prep Drainage enforcement of the 1 second SRTM DEM for Australia Ehlschlaeger C R 1989 Using the AT search algorithm to develop hydrologic models from digital elevation data Proceedings of the International Geographic Information System IGIS Symposium Baltimore MD 275 281
85. ling and Simulation Modelling and Simulation Society of Australia and New Zealand and International Association for Mathematics and Computers in Simulation July 2009 pp 2493 2499 ISBN 978 0 9758400 7 8 http www mssanz org au modsim09 F13 hutchinson pdf Rodriguez E Morris C S and Belz J E 2006 A global assessment of the SRTM performance Photogrammetric Engineering and Remote Sensing 72 3 249 260 Slater J A Garvey G Johnston C Haase J Heady B Kroenung G and Little J 2006 The SRTM data finishing process and products Photogrammetric Engineering and Remote Sensing 72 3 237 247 For technical queries please contact Geoscience Australia elevation ga gov au or GA Sales on 02 6249 9966 Authors J ohn Gallant CSIRO Trevor Dowling CSIRO Arthur Read CSIRO Michael Hutchinson ANU Nerida Wilson Geoscience Australia and Phil Tickle Geoscience Australia One second SRTM Derived Digital Elevation Models User Guide v1 0 4 Page 87 of 106 Appendix E 3 second DSM Metadata Note This metadata describes the dataset in accordance with the ANZLIC Australia New Zealand Land Information Council Core Metadata Guidelines Version 2 Dataset citation ANZLIC unique identifier ANZCWO703014216 Title 1 second SRTM Derived 3 second Digital Surface Model DSM version 1 0 Custodian Custodian Geoscience Australia Jurisdiction Australia Description Abstract The 3 second 90 m Shuttl
86. ly as its implementation in GRASS as r watershed that constructs flow lines through depressions without first filling the depressions to the outlet level Figure 12 a e illustrates this process The ANUDEM software cannot process the entire continent or entire drainage basins at the 1 second resolution in a single pass Drainage enforcement was therefore performed separately for each 1x1 degree tile using Y degree overlaps on each side The resulting 2x2 degree tiles were trimmed to a 100 cell overlap mosaicked with adjacent trimmed tiles then clipped to the 1x1 degree tile The mosaicking with 100 cell overlaps was done to help ensure there are no elevation discontinuities at tile boundaries However there are differences in smoothing on different tiles due to the automatic adjustment of smoothing parameters in ANUDEM The mosaicking process does not guarantee the preservation of continuous descent along drainage lines so a final descent enforcement step was applied using the CheckStreamDescent program written for that purpose CheckStreamDescent processes all tiles as a single data set so continuous descent of all stream lines to their termination points was ensured As a final step the ocean areas were set to nodata Other water bodies have not been altered after drainage enforcement and most water bodies include a drainage line through them reflecting the connectors in the AusHydro data Note that this is in contrast to the fi
87. manent Survey Marks PSM as the 1 second DEM accuracy assessment Results of the comparison Showed the absolute accuracy of the data as tested relative to AHD71 to be 14 54 m at the 95 percentile with an RMS error of 7 029 m in open flat terrain Ninety nine percent of points are within a height difference of less than 29 97 m Relative elevation accuracy between adjacent cells is improved in the DEM S due to the reduction in noise levels this has not been quantified but is evident in the comparison of slopes calculated before and after smoothing as shown in the User Guide Geoscience Australia and CSIRO 2011 The smoothing process estimated typical improvements of the order of 2 3 m in the 1 second DEM S Height accuracy Is likely to be poorer in areas where voids have been filled using the GEODATA 9 second DEM particularly in high relief areas Logical Consistency The 1 second DEM S represents ground elevation with greatly improved relative elevations between adjacent grid cells in low relief areas due to the smoothing process Slopes as small as 0 02 2 m in 10 km can be resolved in the DEM S The removal of vegetation involves estimation of vegetation height at the edges of vegetation patches and interpolation of those heights across areas of continuous vegetation cover Variations in vegetation height within large areas of vegetation are One second SRTM Derived Digital Elevation Models User Guide v1 0 4 Page 101 of 106 not ca
88. ments of mainland or pieces of islands were not supplied at 1 second resolution and are therefore missing from the DEM E112 S26 E113 S29 E118 S20 E120 S35 E121 S35 E123 S16 E124 S15 E125 S14 E132 S11 E133 S11 E134 S35 E141 S10 E142 S10 E143 S10 E146 S17 E150 S22 E152 S24 Note that the coordinates are of the south western corner of the tile Contact information Contact organisation Geoscience Australia GA Contact position Director Sales and Distribution ISB Mail address GPO Box 378 Mail address Locality Canberra State ACT Country Australia Postcode 2601 Telephone 61 2 6249 9966 Facsimile 61 2 6249 9960 Electronic mail address sales ga gov au Metadata information Metadata created date 2010 08 30 Metadata Updated date 2011 09 01 Additional metadata Conversion to floating point format The smoothing process alters all data values in the DEM by varying amounts and the result is a floating point data set capturing in some places very small but meaningful differences in elevation between adjacent cells Ancillary data layers distributed with the data Five additional data layers provide information about the alterations made to the raw SRTM data to produce this DEM A water mask at 1 second resolution showing the cells that are part of the flattened water bodies A void mask showing cells that were no data in the raw SRTM and have been filled using the void filling algorithm
89. n late 2009 Version 1 0 of the DEM S was released in July 2010 and version 1 0 of the hydrologically enforced DEM H was released in October 2011 These products provide substantial improvements in the quality and consistency of the data relative to the original SRTM data but are not free from artefacts Improved products will be released over time The 3 second products were derived from the 1 second data and version 1 0 was released in August 2010 Future releases of these products will occur when the 1 second products have been improved At this stage there is no 3 second DEM H product which requires re interpolation with drainage enforcement at that resolution Nomenclature There is no universal agreement about the use of the terms digital surface model DSM digital elevation model DEM and digital terrain model DTM The usage adopted for the SRTM derived 1 second products is that a DSM represents a regular grid of ground surface topography and height as well as other features including vegetation and man made structures while a DEM represents a regular grid of ground surface topography and where possible excludes other features such as vegetation and man made structures In some areas the term DTM is used for the land surface model with the DEM having a more generic meaning as a DIM or DSM but in Australia the term DEM is generally accepted to mean a land surface model such as the GEODATA 9 second DEM and we have chosen to contin
90. ndscapes Figure 36 e Excessive incision downstream of open cut mines Figure 37 Apart from the first two causes the large and undesirable changes in elevation do not affect the hydrological connectivity of the landscape Figure 33 An extraneous infill stream line red created due to a defect in the line creation algorithm The line results in the creation of a drainage feature in DEM H that does not exist in the landscape except as a series of gradually descending depressions note the absence of mapped AusHydro lines blue in the flat areas The effect of this issue is small since the drainage line is incised to a depth of only 1 2 m and follows the natural drainage direction of the landscape but suggests the existence of a channel where none actually exists Near Newhaven Station NT 131 25E 22 75S Elevation range 500 800 m One second SRTM Derived Digital Elevation Models User Guide v1 0 4 Page 43 of 106 Figure 34 Error in AusHydro 1 250 000 stream line DEM S top and DEM H bottom with unmodified AusHydro stream lines The incorrect stream line has been cut by the 10 degree slope threshold near the top of the hill preventing it from gouging all the way through the hill Near Tumut NSW 148 29E 35 345 Elevation range 250 400 m One second SRTM Derived Digital Elevation Models User Guide v1 0 4 Page 44 of 106 Figure 35 Excessive incision due to closely spaced infill streams Mt Elliot near Townsville Qld
91. nishing of the DSM DEM and DEM S which all contain flattened water bodies and used the SRTM edit rules to ensure that land adjacent to water bodies is at a higher elevation than the water Examples of drainage enforcement in different environments are shown in Figures 12 15 One second SRTM Derived Digital Elevation Models User Guide v1 0 4 Page 24 of 106 a 0 0 5 1 km One second SRTM Derived Digital Elevation Models User Guide v1 0 4 Page 25 of 106 Figure 12 Illustration of the improvement from combining clipped AusHydro stream lines and infill lines for drainage enforcement in the Kiewa River below Dartmouth Dam Victoria 147 42E 36 525 Elevation range 250 350 m a DEM S with no drainage enforcement the valley is clearly visible but consists of a series of depressions with lower green areas separated by higher yellow elevations b Drainage enforcement using unmodified AusHydro stream lines blue note the duplication of drainage structure caused by spatial offsets of the mapped stream lines c Drainage enforcement using AusHydro lines clipped to areas where slope is less than 10 degrees black note the incomplete drainage enforcement along the river d DEM H drainage enforced using both clipped AusHydro lines black and infill lines red derived from DEM S drainage enforcement is continuous along the river line The clipped segments of AusHydro produce erroneous drainage enforcement in some areas but they d
92. o not create significant hydrological network problems because they are isolated e The several versions of drainage lines superimposed on Landsat imagery showing the good agreement between the infill lines red derived from DEM S and the true river channel One second SRTM Derived Digital Elevation Models User Guide v1 0 4 Page 26 of 106 Figure 13 Drainage enforcement in a canyon area Capertee River in Wollemi National Park NSW 150 43E 33 16S Elevation range 150 650 m DEM S top and DEM H bottom with the stream lines used for the enforcement comprised of segments of AusHydro stream lines and infill lines derived from DEM S One second SRTM Derived Digital Elevation Models User Guide v1 0 4 Page 27 of 106 0 1 2km Figure 14 Drainage enforcement in high and low relief areas Avon River at Mt Mambup on the north eastern edge of Perth WA 116 05E 31 77S Elevation range O 250 m DEM S top and DEM H bottom with AusHydro stream lines blue clipped AusHydro lines black and infill lines red derived from DEM S Note that in the low relief areas the AusHydro lines are used unmodified while in higher relief areas they are replaced by the infill lines derived from DEM S While the AusHydro lines in low relief areas are still commonly misplaced by up to 200 m the spatial error does not cause significant hydrological anomalies One second SRTM Derived Digital Elevation Models User Guide v1 0 4 Page 28 of 106
93. ohman G Kroenung G and Strebeck J 2006 Filling SRTM voids The delta surface fill method Photogrammetric Engineering and Remote Sensing 72 3 213 216 Read A M Gallant J C and Dowling T I in prep Destriping and void filling methods used in SRTM 1 Second processing Rodr guez E Morris C S and Belz J E 2006 A global assessment of the SRTM performance Photogrammetric Engineering and Remote Sensing 72 3 249 260 Sibson R 1981 A brief description of natural neighbour interpolation In V Barnet editor Interpreting Multivariate Data pages 21 36 John Wiley Sons Chichester Slater J A Garvey G Johnston C Haase J Heady B Kroenung G and Little J 2006 The SRTM data finishing process and products Photogrammetric Engineering and Remote Sensing 72 3 237 247 One second SRTM Derived Digital Elevation Models User Guide v1 0 4 Page 97 of 106 For technical queries please contact Geoscience Australia elevation ga gov au or GA Sales on 02 6249 9966 Authors J ohn Gallant CSIRO Trevor Dowling CSIRO Arthur Read CSIRO Nerida Wilson Geoscience Australia Phil Tickle Geoscience Australia and Chris Inskeep Geoscience Australia One second SRTM Derived Digital Elevation Models User Guide v1 0 4 Page 98 of 106 Appendix G 3 second DEM S Metadata Note This metadata describes the dataset in accordance with the ANZLIC Australia New Zealand Land Information Counci
94. ologically enforced digital elevation model DEM H o A hydrologically enforced DEM is based on DEM S that has had drainage lines imposed and been further smoothed using the ANUDEM interpolation software One second SRTM Derived Digital Elevation Models User Guide v1 0 4 Page 8 of 106 The last product a hydrologically enforced DEM is most similar to the DEMs commonly in use around Australia such as the GEODATA 9 Second DEM and the 25 m resolution DEMs produced by State and Territory agencies from digitised topographic maps For any analysis where surface shape is important one of the smoothed DEMs DEM S or DEM H should be used DEM S is preferred for shape and vertical accuracy and DEM H for hydrological connectivity The DSM s suitable if you want to see the vegetation as well as the land surface height There are few cases where DEM ts the best data source unless access to a less processed product is necessary The 1 second DEM in its various incarnations has quite different characteristics to DEMs derived by interpolation from topographic data Those DEMs are typically quite smooth and are based on fairly accurate but sparse source data usually contours and spot heights supplemented by drainage lines The SRTM data Is derived from radar measurements that are dense there is essentially a measurement at almost every grid cell but noisy Version 1 0 of the DSM was released in early 2009 and version 1 0 of the DEM was released i
95. om patterns of noise that have been removed In the moderate relief areas some topographic structure has been lost due to smoothing while in the steepest areas there has been very little change Jamberoo NSW 150 7E 34 65 Elevation range O 750 m One second SRTM Derived Digital Elevation Models User Guide v1 0 4 Page 22 of 106 Figure 11 Slope before top and after smoothing bottom the slope colour scale is the same in both cases The broad diagonal stripe of higher slopes is due to particularly high noise levels in that area almost completely obscuring the topographic structure in the slope map After smoothing the flat valley floor in the middle of the area is apparent Lake Bryde area WA 118 8E 33 45 Elevation range 280 390 m One second SRTM Derived Digital Elevation Models User Guide v1 0 4 Page 23 of 106 Drainage Enforcement The SRTM DEM does not represent channels except where they are quite large This is partly due to the SRTM radar s inherent resolution of around 50 m Farr et al 2007 and partly due to the prevalence of trees on drainage lines in much of Australia that obscure and effectively raise the channel Drainage enforcement using independently mapped stream lines is therefore required to produce a DEM that properly represents flow paths through the landscape The ANUDEM software Hutchinson 1988 1989 2009 version 5 2 5 dated 1 December 2010 was the result of significant revisions to the code during
96. ond resolution using ANUDEM 5 2 without drainage enforcement The Delta Surface Fill Method calculates height differences between SRTM and infill data to create a delta surface with voids where the SRTM has no values then interpolates across voids The void is then replaced by infill DEM adjusted by the interpolated delta surface resulting In an exact match of heights at the edges of each void Two changes to the Delta Surface Fill Method were made interpolation of the delta surface was achieved with natural neighbour interpolation Sibson 1981 implemented in ArcGIS 9 3 rather than Inverse distance weighted interpolation and a mean plane inside larger voids was not used r bodies ke bodies in the original 1 second data were modified as part of the de striping process and were re flattened afterwards SRTM Water Body data were converted to a 1 second resolution grid then adjusted to match the extent of equal height pixels in One second SRTM Derived Digital Elevation Models User Guide v1 0 4 Page 67 of 106 original SRTM 1 second data Grid cells within that water mask were set to the original SRTM height Edit rules for land surrounding water bodies SRTM edit rules set all land adjacent to water at least 1 m above water level to ensure containment of water Slater et al 2006 Following de striping void filling and water flattening the heights of all grid cells adjacent to water was set to at least 1 cm above the water surface The
97. ond SRTM Derived Digital Elevation Models User Guide v1 0 4 Page 103 of 106 Appendix H Loading the data To assist you in loading and viewing the data into your preferred software some basic instructions are described as a guide if you are unfamiliar with raster data This is a guide only and there may be other ways to import the data Geoscience Australia is able to provide further information on the product and data format but is not able to provide specific software advice For this please consult your software company for technical support As data processing capacity is improved other software packages will be detailed Into ESRI ArcGIS To open ArcMap either using the button in ArcCatalog a or through Windows Start Program menu Either open a new mxd or a blank mxd when prompted To add the SRTM either press the Add Data icon or go to File gt Add Data Navigate to the directory where you have stored the SRTM data The raster grid should appear with an icon like this 24 next to the file name When asked if you would like to Create Pyramids it is advised that you click yes This will take some time now but will save time when viewing the files later Please note that Building Pyramids almost doubles the file size but at a later stage you can delete the pyramids if needed Create pyramids for dems3sv1_0 49200 x 40800 This raster data source does not have pyramids Pyramids allow for rapid display at Waring resolutions
98. ons between adjacent grid cells in low relief areas due to the smoothing process Slopes as small as 0 02 2 m in 10 km can be resolved in this DEM S The removal of vegetation involves estimation of vegetation height at the edges of vegetation patches and interpolation of those heights across areas of continuous vegetation cover Variations in vegetation height within large areas of vegetation are not captured by this method The vegetation removal process guarantees that no elevations have been increased as part of the process All void areas have been filled and there are no discontinuities due to tile boundaries The SRTM editing rules relating to water bodies have been respected in the processing lakes are flat rivers descend continuously in a downstream direction and sea surfaces are at O m elevation Flattened water bodies occupy the same areas as in the original SRTM 1 second data Grid cells adjacent to water bodies are at least 1 cm above the One second SRTM Derived Digital Elevation Models User Guide v1 0 4 Page 79 of 106 water surface Void areas within water bodies small islands not represented in the original SRTM data are at least 1 cm above the water surface over their entire area Completeness The DEM covers all of continental Australia and near coastal islands with land areas including all islands defined by the available SRTM 1 second elevation and SRTM Water Body Data datasets The following tiles containing frag
99. ooonnnnnnicnnnnnninnnnnarnnccnnnnos 6 Licence 1 second DSM Government Use Only ecccccccccssssscccccccccccesssecceeeeeaaeesseccceeesaaeeseeeeseseesauaaeseeeeeeeenes DATA SCHEMA nic ion ls a pits 7 DQG TE OVI seria oc edaaaanc neues 7 Daa LAT A e E T E Nacaunuhen Oe uiatcanben aes 7 WSCA CAME LO ISC a bs T 7 AD DIGIONALAINEOR MATION ada 7 INTRODUCTION lt a oa 8 VERVIEW sentada plena iras aia ts 8 NOMENCLATURE ia 9 SRTM B ACK GROIN eds 9 PROCESSING OF THE SRTM DATA urna soir o deea adiro sieis 10 SUPE TCI OV Al iio 10 void TUU unberden a diia iaa 14 Wes cta tono Sel FONOVISA A is 17 MOI AS A Oe 21 Drona OC LCAN ed Ase ET eR 24 Elevations near coasts and the water ANd ocean MASKS oocccccccnncnnccnnnnnnnnnonnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnss 29 DATASETENAMBLES edo 31 KNOWN ISSUES usada 35 Res dual Spears 35 Brodd S COLE SIPOS AA A A AAA AE da 35 O A EE EIET EENT Pee ere RATE ee E REET ear EAIA SETS CME Rarer eae Tete Cee PET E art a Renee Seen eer re 36 ETA IS CLS eer AE sala tot E A AA oes Jf NOS A A AAA AA A A OSO 38 Incomplete removal Of vegetal on Offsets id 59 VECEIGIION hetent Over estimated lt a A A AAA A see tatess 39 Incomplete removal of urban and built Infrastructure ooonnnnnccnnnnnnnnnnnononnnnnnnnnnnncnnnnnnnnnnnnnnnnnnnnnnnnnncnnnnnnnns 4 DFANAa Se REI GICE ASS US e EE EEE E E AEO T A E 43 incoinplete Stream FN OV CCIMENIS visssce sess ce cask ca cuestenssebunea xe tas tap a ian AONE AARNE A 47 EVALUATION
100. oves the representation of the landform shape particularly in low relief areas but it is not clear whether this also produces an improvement in overall height accuracy Some striping remains in the data at a much reduced level mostly less than 0 3 m amplitude in the 1 second DSM Additional artefacts including long wavelength 10km striping have not been corrected in the 1 second DSM Height accuracy is likely to be poorer in areas where voids have been filled using the GEODATA 9 second DEM particularly in high relief areas Logical Consistency The DSM represents elevation Due to random noise the relative elevation between adjacent grid cells can be in error by several m All void areas have been filled and there are no discontinuities due to tile boundaries The SRTM editing rules relating to water bodies have been respected in the processing lakes are flat rivers descend continuously in a downstream direction and sea surfaces are at O m elevation Flattened water bodies occupy the same areas as in the original SRTM 1 second data Grid cells adjacent to water bodies are at least 1 cm above the water surface Void areas within water bodies small islands not represented in the original SRTM data are at least 1 cm above the water surface over their entire area Completeness The DSM covers all of continental Australia and near coastal islands with land areas including all islands defined by the available SRTM 1 second elevation
101. patches Lower and less dense vegetation including crops do not appear to cause any significant offset The treatment of vegetation offsets to produce the DEM relies heavily on Landsat based mapping of woody vegetation to define where the offsets are likely to occur The mapped extents of woody vegetation were adjusted using an edge matching process to better represent the extents of areas affected by vegetation offsets in the SRTM DSM Vegetation treatment was undertaken across about 40 of One second SRTM Derived Digital Elevation Models User Guide v1 0 4 Page 17 of 106 Australia The extent of treatment as shown in Figure 7 below is provided as an ancillary dataset Vegetation Removal 0 250 500 1 000 Km Yes Si came a Figure 7 Distribution of vegetation removal in the SRTM The tree offsets are treated by detecting affected areas measuring the height offset around the edges interpolating the height offset across the tree vegetated areas and subtracting the offset from the DSM Figure 8 The heights of the offsets are estimated by measuring height differences across the boundaries of the vegetation patches The method provides good estimates of the offsets in flat landscapes with well mapped vegetation boundaries The effect of sloping terrain is accounted for in the estimation of the offsets but the results are less reliable in hilly terrain where the mapped vegetation extents do not match the extents of vegetation o
102. ples focus on low relief landforms where the SRTM based 1 second DEM is significantly superior to DEMs based on interpolated contour data and in the arid zone previously only covered by the GEODATA 9 second DEM pe fe a 4 el i AF ji J 4 Figure 17 The linear features in the bottom centre are open cut coal mines Moura QLD 150 0E 24 5S Elevation range 220 320 m One second SRTM Derived Digital Elevation Models User Guide v1 0 4 Page 31 of 106 0 30 60 km A AS ES ee Figure 18 Escarpment at the southern edge of the Nullarbor Plain Madura WA 126 8E 31 95 Elevation range O 170 m A ya 4 1 y a ee i Te a a Figure 19 Wilbrunga Range Tanami Desert NT 129 5E 21 5S Elevation range 350 480 m One second SRTM Derived Digital Elevation Models User Guide v1 0 4 Page 32 of 106 i a R B ARA la i e e r 4 Ta f J r hao l ae TA gt A y i e A A i y A 3 a o IE E p is ea Elo rae Figure 20 Victoria River WA 131 2E 16 6S Elevation range 70 200 m i y pe J a e i r j 5 rF E mF pa 1 ES r P ds AA Y iE y pe i 5 fae ti AL pa RRA AIRE MAP AI Liia ut dr Fo r pr E lt mi f 15 fe 4 Hr Y F e gt ws dal Wh 1 r ar sm d i AE E i i i x 4 en wah E Mita REEN jz E art e t Figure 21 Ouyen VIC 142 2E 35 1S Elevation range 30 120 m me rE diye ge ee 7 A AA ee e a F 3 Ej _ A erat oe y fe ps at z er ti en Par i g
103. preting Multivariate Data pages 21 36 John Wiley Sons Chichester Slater J A Garvey G Johnston C Haase J Heady B Kroenung G and Little J 2006 The SRTM data finishing process and products Photogrammetric Engineering and Remote Sensing 72 3 237 247 For technical queries please contact Geoscience Australia elevation ga gov au or GA Sales on 02 6249 9966 Authors J ohn Gallant CSIRO Trevor Dowling CSIRO Arthur Read CSIRO Nerida Wilson Geoscience Australia Phil Tickle Geoscience Australia and Chris Inskeep Geoscience Australia One second SRTM Derived Digital Elevation Models User Guide v1 0 4 Page 70 of 106 Appendix B 1 second DEM Metadata Note This metadata describes the dataset in accordance with the ANZLIC Australia New Zealand Land Information Council Core Metadata Guidelines Version 2 Dataset citation ANZLIC unique identifier ANZCWO703013355 Title 1 second SRTM Derived Digital Elevation Model DEM version 1 0 Custodian Custodian Geoscience Australia Jurisdiction Australia Description Abstract The 1 second Shuttle Radar Topography Mission SRTM derived Digital Elevation Model DEM Version 1 0 is a 1 arc second 30 m gridded DEM The DEM represents ground surface topography and excludes vegetation features The dataset was derived from the 1 second Digital Surface Model DSM ANZCW0703013336 by automatically removing vegetation offsets identified using sev
104. ptured by this method The vegetation removal process guarantees that no elevations have been increased as part of the process All void areas have been filled and there are no discontinuities due to original tile boundaries The SRTM editing rules relating to water bodies have been respected in the processing lakes are flat rivers descend continuously in a downstream direction and sea surfaces are at O m elevation Flattened water bodies occupy the same areas as in the original SRTM 1 second data Grid cells adjacent to water bodies are at least 1cm above the water surface Void areas within water bodies small islands not represented in the original SRTM data are at least 1cm above the water surface over their entire area Completeness The DEM S covers all of continental Australia and near coastal islands with land areas including all islands defined by the available SRTM 1 second elevation and SRTM Water Body Data datasets Some fragments of mainland or pieces of islands may be missing Contact information Contact organisation Geoscience Australia GA Contact position Director Sales and Distribution Mail address GPO Box 378 Mail address Locality Canberra State ACT Country Australia Postcode 2601 Telephone 61 2 6249 9966 Facsimile 61 2 6249 9960 Electronic mail address sales ga gov au Metadata information Metadata date 2010 08 30 Metadata Updated date 2011 09 01 Additional metadata Conversion
105. public http creativecommons org licenses by 3 0 au Copyright Commonwealth of Australia Geoscience Australia 2010 Data quality Lineage Source data 1 SRTM 1 second Version 2 data Farr et al 2007 supplied by Defence Imagery and Geospatial Organisation DIGO as 813 1 x 1 degree tiles Data was produced by NASA from radar data collected by the Shuttle Radar Topography Mission in February 2000 2 GEODATA 9 second DEM Version 3 Geoscience Australia 2008 used to fill voids 3 SRTM Water Body Data SWBD shapefile accompanying the SRTM data Slater et al 2006 This defines the coastline and larger inland waterbodies for the DEM and DSM 4 Vegetation masks and water masks applied to the DEM to remove vegetation 5 AusHydro V1 6 1 250 000 watercourse lines DSM processing This DEM is based on the 1 second SRTM derived Digital Surface Model DSM that was itself derived from the 1 second Shuttle Radar Topography Mission data The DSM was produced by removing stripes filling voids and re flattening water bodies Further details are provided in the DSM metadata ANZCW0703013336 DEM processing vegetation offset removal Vegetation offsets were identified using Landsat based mapping of woody vegetation The height offsets were estimated around the edges of vegetation patches then interpolated to a continuous surface of vegetation height offset that was subtracted from the DSM to produce a bare earth DEM Further d
106. revisions and removed where possible Error rating 123 19 WA FitzroyRiver sf hare les fis WA Adcock River north Fitzroy Crsing Med NT McArthur ower 36 7 NT manne arg 141 17 QLD Staaten River from Gulf Carpentaria Medium _ E E E ee ee eee ee QLD Cooper es east SA QLD border CE on Buroo River Large 144 28 QLD Bulloo River sf hare TAS River Derwent north west Hobart 149 36 NSW Lake George 150 36 NSW various rivers from the coast inland includes Moruya Small River Clyde River at Batemans Bay Conjola Creek One second SRTM Derived Digital Elevation Models User Guide v1 0 4 Page 47 of 106 Longitude Latitude Location An example of a small error is shown below Figure 38 in the Derwent River near Hobart Tasmania Lines that have been missed in the drainage enforcement process are visible as red lines AusHydro stream lines with no corresponding black lines drainage enforcements Figure 38 Small errors where parts of the AusHydro stream network have not been enforced into the DEM H Derwent River north west Hobart Tas 147 075E 42 765S Elevation range 4 964 m Black shows where stream enforcement has occurred and red lines are AusHydro stream lines One second SRTM Derived Digital Elevation Models User Guide v1 0 4 Page 48 of 106 Medium errors are those that have a portion of the AusHydro braided stream network omitted from the network enforced into the DEM H as
107. rmat As a by product of the de striping process the integer data was converted to floating point format to allow for the continuously varying nature of the striping Areas where no de striping was required will contain unaltered integer values but are represented in floating point format for consistency Data layers distributed with the data Five additional data layers provide information about this DSM The four DSM ancillary layers and the DSM tile index have been provided with the data Vegetation offset Removal Grids are for government use only and will be supplied with the DSM if approved by elevation ga gov au One second SRTM Derived Digital Elevation Models User Guide v1 0 4 Page 69 of 106 References Geoscience Australia 2008 GEODATA 9 Second DEM Version 3 Geoscience Australia and CSIRO 2011 1 Second SRTM Derived Digital Elevation Models User Guide Version 1 0 4 Geoscience Australia Grohman G Kroenung G and Strebeck J 2006 Filling SRTM voids The delta surface fill method Photogrammetric Engineering and Remote Sensing 72 3 213 216 Read A M Gallant J C and Dowling T I in prep Destriping and void filling the 1 second SRTM DEM for Australia Rodr guez E Morris C S and Belz J E 2006 A global assessment of the SRTM performance Photogrammetric Engineering and Remote Sensing 72 3 249 260 Sibson R 1981 A brief description of natural neighbour interpolation In V Barnet editor Inter
108. s Data was produced by NASA from radar data collected by the Shuttle Radar Topography Mission in February 2000 2 GEODATA 9 second DEM Version 3 Geoscience Australia 2008 used to fill voids 3 SRTM Water Body Data SWBD shapefile accompanying the SRTM data Slater et al 2006 This defines the coastline and larger inland waterbodies for the SRTM DEM and DSM 4 Vegetation masks and water masks applied to the DEM to remove vegetation 5 Adaptive smoothing applied to DEM to produce 1 second DEM S 6 1 second DEM S resampled to 3 second DEM S In order to understand the 3 second DEM the processing of the parent dataset the 1 second DEM S is described below 1 second DSM processing The 1 second SRTM derived Digital Surface Model DSM was derived from the 1 second Shuttle Radar Topography Mission data by removing stripes filling voids and re flattening water bodies Further details are provided in the DSM metadata ANZCWO703013336 1 second DEM processing vegetation offset removal Vegetation offsets were identified using Landsat based mapping of woody vegetation The height offsets were estimated around the edges of vegetation patches then interpolated to a continuous surface of vegetation height offset that was subtracted from the DSM to produce a bare earth DEM Further details are provided in the DSM metadata ANZCW0703013355 Adaptive smoothing The adaptive smoothing process was designed to smooth flat areas to a greater
109. s one of which was inside the Shuttle and the other was on a 60 m boom The Shuttle was oriented to point the antennas at 45 to the ground to optimise the effect of topography on the interference patterns but this also has the effect of obscuring any steep areas facing away from the Shuttle This is mostly overcome by collecting overlapping swathes from different orbits although some canyons and steep areas have no data Other areas that did not produce a good radar return signal also have no data Comparison of SRTM with reference data Rodriguez et al 2006 showed that 90 of tested heights were within 6 m of the reference heights In much of the clear flat areas of Australia the height errors are less than 3 m although there are some areas where the errors are much larger Trees and buildings produce offsets in the elevation much larger than these height errors since the radar frequency used by SRTM is reflected by them Processing of the SRTM data The processing of the SRTM DSM has produced a series of products e The cleaned digital surface model DSM is the 1 second SRTM with stripes removed and voids filled e The bare earth DEM is based on the cleaned DSM and has had tree offsets removed using automated methods e The DEM S is based on the bare earth DEM and has been adaptively Smoothed e The hydrologically enforced DEM H is based on the bare earth DEM S and has had drainage lines imposed and been smoothed using the ANUDEM
110. s of continuous vegetation cover Variations in vegetation height within large areas of vegetation are not captured by this method The vegetation removal process guarantees that no elevations have been increased as part of the process All void areas have been filled and there are no discontinuities due to tile boundaries The SRTM editing rules relating to water bodies have been respected in the processing lakes are flat rivers decline continuously in a downstream direction and sea surfaces are at O m elevation Flattened water bodies occupy the same areas as in the original SRTM 1 second data Grid cells adjacent to water bodies are at least 1cm above the water surface Void areas within water bodies small islands not represented in the original SRTM data are at least 1cm above the water surface over their entire area Completeness The DEM covers all of continental Australia and near coastal islands with land areas including all islands defined by the available SRTM 1 second elevation and SRTM Water Body Data datasets The following tiles containing fragments of mainland or pieces of islands were not Supplied at 1 second resolution and are therefore missing from the DEM E112 S26 E124 515 E142 10 E113 S29 E125 S14 E143 S10 E118 S20 E132 11 E146 S17 E120 S35 E133 S11 E150 S22 E121 S35 E134 S35 E152 S24 E123 S16 E141 S10 Note that the coordinates are of the south western corner of the tile One second SRTM Derived Digital
111. s with well mapped vegetation boundaries The effect of sloping terrain is accounted for in the estimation of the offsets but the results are less reliable in hilly terrain Estimates of the offsets can also be very poor where the mapped vegetation extents do not match the extents of One second SRTM Derived Digital Elevation Models User Guide v1 0 4 Page 72 of 106 vegetation offsets as seen by the SRTM instrument The estimation of the vegetation offsets can also be under or over estimated if vegetation and topographic patterns coincide such as trees on hilltops or dune ridges or in inset floodplains or swamps The height offsets at vegetation edges are interpolated within vegetation patches to estimate the effects within the patches The best results tend to be in small patches such as remnant tree patches In continuously forested areas with few edges for estimating the offsets the heights are likely to be less reliable and there is no information at all on variations of the height offset within continuous forests The removal of vegetation has been quite effective overall but there are many areas that contain either untreated or incompletely treated vegetation effects The methods will be fully described in Read et al in prep and Gallant et al in prep Void filling Voids areas without data occur in the data due to low radar reflectance typically open water or dry sandy soils or topographic shadowing in high relief areas The D
112. shore Further information is provided in the User Guide Geoscience Australia and CSIRO 2011 Positional accuracy The horizontal positional error is the same as for the raw SRTM 1 second data with 90 of tested locations within 7 2 m for Australia See Rodriguez et al 2006 for more information Attribute accuracy Accuracy of the 1 second DEM before smoothing to form DEM S was tested using 1198 Permanent Survey Marks PSM distributed across the Australian continent relative to the Australian Height Datum AHD71 Results of this comparison show the absolute accuracy of the data as tested relative to AHD71 to be 7 582 m at the 95 percentile with a RMS error of 3 868 in open flat terrain Ninety nine percent of points are within a height difference of less than 9 602 m The smoothing process estimated typical improvements in the order of 2 3 m This would make the DEM S accuracy to be of approximately 5 m Relative elevation accuracy between adjacent cells is improved in DEM S due to the reduction in noise levels this has not been quantified but is evident in the comparison of slopes calculated before and after smoothing as shown in the User Guide Geoscience Australia and CSIRO 2011 Height accuracy Is likely to be poorer in areas where voids have been filled using the GEODATA 9 second DEM particularly in high relief areas Logical Consistency The DEM S represents ground elevation with greatly improved relative elevati
113. ss the surface for statistical analysis Figure 45 shows the histogram of the differences The following statistics were then obtained Mean 3 001 St Dev 9 863 Atherton Sample Points Statistics The 1 second DEM and Atherton data provide similar heights normally in the range of 3 m to 6 m except in densely vegetated areas or lakes where the DEM has under estimated the heights by 8 40 m This results in the mean difference being much higher than the average height difference over non vegetated areas which is closer to 0 2 m for the majority of the sample area One second SRTM Derived Digital Elevation Models User Guide v1 0 4 Page 54 of 106 Difference Histogram ho a a a Frequency RP Ln DO a Us a ta i y Me AA at 22 4 20 0 19 9 16 0 15 9 12 0 11 9 8 0 7 9 4 0 3 9 2 0 1 9 0 0 o 0 1 2 0 21 40 4 1 6 0 6 1 8 0 8 1 12 0 12 1 16 0 16 1 20 0 20 1 24 0 24 1 30 0 30 1 36 0 36 1 42 0 Figure 46 Atherton Grid difference points 145 604E 17 303S One second SRTM Derived Digital Elevation Models User Guide v1 0 4 Page 55 of 106 The areas of large difference between the 1 second DEM and the Atherton surface Figure 46 are primarily forested areas and water surfaces Lake Eacham and Lake Barrine both lakes in volcanic craters The differences in the water surfaces are most likely due to incorrect water surface heights assigned to the SRTM data although
114. t pa Br z a Pa One second SRTM Derived Digital Elevation Models User Guide v1 0 4 Page 33 of 106 SA e 5 j 4 a pe so o nT z ii e An E j A B cd r Ki r j a a ys j g 2 2 a ad a j Figure 22 Longford TAS 147 2E 41 6S Elevation range O 700 m k NR AE pr E ia ee he Figure 23 Braidwood NSW 149 8E 35 3S Elevation range 400 900 m One second SRTM Derived Digital Elevation Models User Guide v1 0 4 Page 34 of 106 Known ssues All DEMs are imperfect representations of the earth s land surface with their particular foibles and the products derived from the 1 second SRTM are no exception There are a number of known issues with the products described below These problems are being addressed in various ways and subsequent releases of the products will improve or resolve these issues Residual stripes Some areas of the DSM and derived DEMs contain stripes that could not be removed using CSIRO custom made de striping tool An example is shown in the Stripe Removal section Figures 2 amp 3 Residual stripes are relatively rare These Stripes will significantly affect measures of surface shape such as slope aspect flow direction and curvature Broad scale stripes In a few areas notably the Hay Plain in southern NSW Figure 24 there are gentle undulations similar to the stripes but with much longer wavelength about 10 km rather than the 800 m of the widespread stripes and amplitude of
115. ue with that term For further information users should refer to the Intergovernmental Committee on Surveying and Mapping Guidelines for Digital Elevation Data http www icsm gov au icsm elevation index html SRTM Background During eleven days in February 2000 Space Shuttle Endeavour collected global elevation data on the SRTM Acquired by the National Geospatial ntelligence Agency NGA and National Aeronautics and Space Administration NASA the data is publicly available globally at three arc second 90 metre resolution and one arc second 30 metre resolution over the United States Australia s Defence Imagery and Geospatial Organisation DIGO were provided access to the 1 second SRTM data over Australia Following the establishment of the National Elevation Data Framework NEDF and passing of the Water Act in 2007 Geoscience Australia GA the Bureau of Meteorology BoM the One second SRTM Derived Digital Elevation Models User Guide v1 0 4 Page 9 of 106 Commonwealth Scientific and Industrial Research Organisation CSIRO and the Australian National University ANU formalised a collaboration to derive a series of 1 second SRTM DEMs that would underpin the new NEDF and the BoMs Australian Hydrological Geospatial Fabric or Geofabric The data was acquired by interferometric synthetic aperture radar meaning that the information is contained in the interference patterns between the radar signals collected by two antenna
116. uously varying nature of the striping Areas where no de striping was required will contain unaltered integer values but are represented in floating point format for consistency Data layers distributed with the data Four additional data layers provide information about the alterations made to the raw SRTM data to produce this DEM A water mask at 1 second resolution showing the cells that are part of the flattened water bodies A void mask showing cells that were no data in the raw SRTM and have been filled using the void filling algorithm Vegetation masks at 1 8 x 1 8 degree resolution illustrating where vegetation was removed from the DEM and issues noted with the removal Tile indexes for the DEM One second SRTM Derived Digital Elevation Models User Guide v1 0 4 Page 75 of 106 References Gallant J C Read A M Dowling T I and Austin J M in prep Removing vegetation offsets from the 1 second SRTM DEM for Australia Geoscience Australia 2008 GEODATA 9 Second DEM Version 3 Geoscience Australia and CSIRO 2011 1 Second SRTM Derived Digital Elevation Models User Guide Version 1 0 Geoscience Australia Grohman G Kroenung G and Strebeck J 2006 Filling SRTM voids The delta surface fill method Photogrammetric Engineering and Remote Sensing 72 3 213 216 Read A M Gallant J C and Dowling T I in prep Destriping and void filling methods used in the 1 second SRTM DEM for Australia Rodriguez E
117. up to 4 m These have not yet been treated Due to the very low gradient of the terrain in the Hay Plain as low as 1 m per 10 kms these stripes will impact surface shape and flow patterns at the 10 km scale Pee a peak oe Veli ui AS A e Figure 24 Broad striping in Hay Plain NSW 144 8E 34 55 Elevation range 70 110 m One second SRTM Derived Digital Elevation Models User Guide v1 0 4 Page 35 of 106 Steps There are several places where there are steps in elevation along straight lines extending many kms These lines are oriented along orbital paths in the same way as the fine scale stripes The most obvious example is north of Balranald in south western NSW where a step of up to 7 m extends along a 30 km line Figure 25 Areas with steps also tend to have a higher noise level which obscures the details of the step but the steps appear to be gradual rather than abrupt with the elevation change occurring over a distance of about 1 km Another clear example extends from south of St George QLD 148 47E 28 265 to Mungindi near the NSW QLD border 149 18E 29 105 These steps will affect measures of local shape such as slope aspect flow direction and curvature They may also disrupt drainage patterns Figure 25 Step north of Balranald NSW 144 0E 33 8S Elevation range 60 90 m One second SRTM Derived Digital Elevation Models User Guide v1 0 4 Page 36 of 106 Large offsets One example of large offsets in elevation
118. ustralia 2008 used to fill voids 3 SRTM Water Body Data SWBD shapefile accompanying the SRTM data Slater et al 2006 This defines the coastline and larger inland waterbodies for the SRTM DSM De striping SRTM data contains striping artefacts oriented approximately NE SW and NW SE that vary in amplitude from about 0 2 m to nearly 4 m The wavelength of the striping is approximately 800 m Stripes were detected in the elevation data using a 2 dimensional Fast Fourier Transform Peaks in the spectra were visually identified and manually delineated using a tool designed specifically for this purpose Striping occurred everywhere except where relief was high enough to obscure striping Spectral analysis was performed on sub tiles to account for spatial variation in the intensity and direction of striping Fourier transform was applied to overlapping sub tiles covering 1536 x 1536 cells 0 43 x 0 43 degrees Central 1024 x 1024 cells were retained each comprising one sixteenth of a 1 x 1 degree tile 900 x 900 cells with a 62 cell overlap on each edge to provide smooth transitions between sub tiles Void filling Voids areas without data occur in the data due to low radar reflectance typically open water or dry sandy soils or topographic shadowing in high relief areas Delta Surface Fill Method Grohman et al 2006 was adapted for this task using GEODATA 9 second DEM as infill data source The GEODATA 9 second data were refined to 1 sec
119. uth bounding latitude 44 East bounding longitude 154 West bounding longitude 113 Data currency Beginning date 2000 2 11 Ending date 2000 2 22 Dataset status Progress Version 1 0 of the 1 second bare earth DSM is complete as at 23 December 2009 Maintenance and update frequency Updates and revisions are anticipated to resolve some of the issues identified in the User Guide Geoscience Australia and CSIRO 2011 and Quality Assessment layers and to incorporate improvements in the Digital Surface Model One second SRTM Derived Digital Elevation Models User Guide v1 0 4 Page 66 of 106 Reference system Horizontal datum WGS84 Vertical datum EGM96 Access Stored data format DIGITAL ArcGIS grid Arclnfo grid Available format type DIGITAL ArcGIS grid Arclnfo grid Access constraints The 1 second DSM data are subject to Commonwealth of Australia Copyright A licence agreement is required and a licence fee is also applicable for packaged data included in the purchase price This data is strictly for government use only and may be provided upon request to elevation ga gov au Data quality Lineage Source data 1 SRTM 1 second Version 2 data Slater et al 2006 supplied by Defence Imagery and Geospatial Organisation DIGO as 813 1 x 1 degree tiles Data were produced by NASA from radar data collected by the Shuttle Radar Topography Mission in February 2000 2 GEODATA 9 second DEM Version 3 Geoscience A
120. variations in height over time may also be involved The height differences in forested areas indicate that the vegetation offset has been significantly under estimated in those areas One second SRTM Derived Digital Elevation Models User Guide v1 0 4 Page 56 of 106 Comparison with Other Elevation Datasets VicMap Elevation The VicMap 20 m DTM coverage extends across the whole of Victoria and 10 km into bordering States Figure 47 The dataset consists of a wide variety of input source data varying in currency from 1974 to 2006 The DTM is hydrologically enforced to represent the mapped surface drainage system The spatial accuracy for VicMap Elevation DTM 20 m and DTM 10 m is inherited from the spatial accuracies of its many source datasets The most consistently used and therefore the base for positional accuracy is the VicMap Elevation 10 20 m Contours amp Relief Therefore the positional accuracy for VicMap Elevation DTM 20 m and DTM 10 m is 12 5 m horizontally and 5 m AHD71 vertically or better barring Known significant errors in those data VIC DEM 20m Elevation Figure 47 Extent of VicMap DTM 20 m 141 306E 38 049S The VicMap DTM was subtracted from the SRTM derived DEM to produce a difference surface Figure 48 The differences are due to a number of factors numbers correspond to approximate location on the figure below 1 Riparian vegetation along the Murray River is a known area of inadequate vegetation r
121. water bodies have not been altered after drainage enforcement and most water bodies include a drainage line through them reflecting the connectors in the AusHydro data Note that this is in contrast to the finishing of the DSM DEM and DEM S which all contain flattened water bodies and used the SRTM edit rules to ensure that land adjacent to water bodies is at a higher elevation than the water Overview of quality assessment Differences in height between DEM S and DEM H were examined to identify areas where defects were created by the drainage enforcement process Some large elevation differences up to 290 m were due to valid drainage enforcements in canyons Other Significant differences are related to various problems including e Excessive height reductions on steep slopes due to multiple parallel infill stream lines e g 152 295 E 30 943 S e Excessive smoothing lowering of hilltops and raising of lower slopes in some areas e g the eastern peaks of the Stirling Range WA around 118 28 E 34 36 S with hilltops lowered by around 200 m e Drainage enforcements to the level of open cut mines traversed by mapped stream lines resulting in deep incisions extending long distances downstream of the mines with the worst instance being from the coal mines in Latrobe Valley Victoria to the outlet of the Gippsland Lakes to an elevation of about 60 m for about 180 km e A few extraneous infill stream lines in inland areas creating long str
122. y NASA from radar data collected by the Shuttle Radar Topography Mission in February 2000 2 GEODATA 9 second DEM Version 3 Geoscience Australia 2008 used to fill voids 3 SRTM Water Body Data SWBD shapefile accompanying the SRTM data Slater et al 2006 This defines the coastline and larger inland waterbodies for the DEM and DSM 4 Vegetation masks and water masks applied to the DEM to remove vegetation 5 1 second DEM re sampled to 3 second DEM 1 second DSM processing The 1 second SRTM derived Digital Surface Model DSM was derived from the 1 second Shuttle Radar Topography Mission data by removing stripes filling voids and re flattening water bodies Further details are provided in the DSM metadata ANZCW0703013336 1 second DEM processing vegetation offset removal Vegetation offsets were identified using Landsat based mapping of woody vegetation The height offsets were estimated around the edges of vegetation patches then interpolated to a continuous surface of vegetation height offset that was subtracted from the DSM to produce a bare earth DEM Further details are provided in the 1 second DSM metadata ANZCW0703013355 Void filling Voids areas without data occur in the data due to low radar reflectance typically open water or dry sandy soils or topographic shadowing in high relief areas Delta Surface Fill Method Grohman et al 2006 was adapted for this task using GEODATA 9 second DEM as infill data sour
123. y large or dense in the same way as vegetation No attempt has been made in this version to remove these features although the adaptive smoothing used to produce DEM S has removed these features in some cases The most visible examples are major city centres and power line towers less dense urban areas are mostly free of such offsets Figures 30 to 32 Figure 30 Power line transmission towers near Mortlake western Victoria appearing as bumps up to 20 m high 142 92E 38 055 Elevation range 120 200 m Figure 31 Central Sydney showing significant visible offsets of up to 30 m in the CBD area and isolated features elsewhere 141 21E 33 875 Elevation range O 100 m One second SRTM Derived Digital Elevation Models User Guide v1 0 4 Page 41 of 106 Figure 32 Albury Wodonga area showing very few artefacts due to urban structures 146 9E 36 15 Elevation range 100 400 m One second SRTM Derived Digital Elevation Models User Guide v1 0 4 Page 42 of 106 Drainage Related Issues A number of known defects have been introduced through the drainage enforcement process e Extraneous lines created by the stream line infill process Figure 33 e Errors in the AusHydro 1 250 000 stream lines Figure 34 e Excessive incision on steep slopes due to closely spaced infill streams Figure 35 e Substantial reductions and in some cases increases in surface elevations due to smoothing by ANUDEM in steep areas in otherwise low relief la

1 second SRTM Derived Products User Guide

Contents

Download Pdf Manuals

Related Search

Related Contents