Blog

Short title: streams_error

A Sextante implementation of these algorithms called FlowTools can be obtained here (contributed by Daniel Nüst). The project report can be downloaded as a PDF. You can access the Eclipse project “flowTools” in the following public Subversion repository:

http://svn.xp-dev.com/svn/FlowTools/

You can also download an initial version (might be outdated!) of the project folder as a zip file (12).

Purpose and use:

Extraction of stream networks from a DEM using error propagation technique.

Programming environment: R / S language
Status of work: Public Domain
Reference: On the uncertainty of stream networks derived from elevation data: the error propagation approach
Data set name: Baranja hill

Attachment:

streams_error_0.zip

stream_sims_0.zip

Wolfgang Schwanghart from the Geographisches Institut, Universität Basel has recently released a toolbox that allows analysis of relief and flow pathways in digital elevation models in Matlab. The toolbox can be used to visualize DEMs, extract simple derivatives, run (and modify) flow models, delineate drainage basins, produce hydrographs and implement similar DEM-based hyrdological analysis. For more info see the User Guide to TopoToolbox.

When you use TopoToolbox in your work, please refer to this publication:

Schwanghart, W., Kuhn, N. J. (2010): TopoToolbox: a set of Matlab functions for topographic analysis. Environmental Modelling & Software, in press.

Short title:geomorph

For a complete description of the processing steps, see the original publication.

Purpose and use:

Automated extraction of geomorphological features using digital elevation data: case study Drente; outputs: extracted classes and summary statistics; various plots and images.

Programming environment:R / S language
Status of work:Public Domain
Reference:{Semi-automated identification and extraction of geomorphological features using digital elevation data}
Data set name:Boschoord case study

Attachment:

maps_KML.zip

FYI: a short report from Geomorphometry 2009 by Bob MacMillan (ISRIC) published in the Pedometron newsletter #28:

“The main purpose of my participation was to keep informed about other efforts, similar to GlobalSoilMap.net, that have a interest in processing digital elevation data and other digital data sets globally or at least for extremely large areas. This conference actually contained a large number of presentations of direct relevance for the GlobalSoilMap.net project. Perhaps first and foremost were the descriptions of efforts being undertaken in Australia (Gallant and Read) and Europe (Köthe and Bock) to process SRTM DEM data at 30 m (Australia) and 90 m (Europe) grid resolution to reduce artefacts and produce a filtered and cleaned DEM that is more suitable for use to produce inputs for the GlobalSoilMap.net project. Both of these presentations highlighted the significant advantages that can be realised by applying a series of filtering and conditioning routines to the original raw SRTM DEM data. It is obvious that similar procedures would prove equally useful if applied to SRTM DEM data sets for other parts of the world under the jurisdiction of other GlobalSoilMap.net nodes. Gallant has offered to help with efforts in other Nodes if asked.

Also of great interest were several projects that demonstrated that it is indeed possible to process and produce digital output for global scale digital data sets, including global scale SRTM DEM data sets. Reuter and Nelson presented a description of WorldTerrain, a contribution of the Global Geomorphometric Atlas. Peter Guth described processing of global scale SRTM data to identify and classify organized linear landforms (dunes). Peter also provided examples of multiple scale analysis and illustrated what you get to “see” from DEMs of 1 m, 100 m and 2 km grid resolution. Guth intends to publish the many different grids of DEM derivatives he produced for his project and make these processed data available for free and widespread use by others. Marcello Gorini described a physiographic classification of the ocean flood using a multi-resolution geomorphometric approach.

Several authors presented methods that may prove of interest to the GlobalSoilMap.net project. Gallant and Hutchinson described a differential equation for computing specific catchment area that could be applied to produce an improved terrain covariate for use in the GlobalSoilMap.net project. Similarly, Peckham, gave a new algorithm for creating DEMs with smooth elevation profiles that could be used to condition rough SRTM or GDEM data sets to smooth out noise and produce more hydrologically plausible surfaces. This algorithm was of particular interest to the GlobalSoilMap.net project because it appeared to be able to introduce hydrologically and geomorphologically relevant detail into 90 m SRTM DEMs of relatively low spatial detail.

Romstad and Etzelmuller described a new approach for segmenting hillslopes into landform elements by applying a watershed algorithm to a surface defined by the total curvature at a point instead of the raw elevation value. The resulting watersheds were bounded by lines of maximum curvature, effectively structuring each hillslope into components partitioned by lines of maximum local curvature. This is harder to explain than to understand when illustrated but it is remarkably simple to implement and may provide a new way of automatically segmenting hillslopes in a simple and efficient fashion.

Metz and others presented an algorithm for fast and efficient processing of massive DEMs to extract drainage networks and flow paths. This is of considerable interest and relevance to the GlobalSoilMap.net project because of the project’s need to process SRTM data globally to compute hydrological flow networks and various indices that are computed based on flow networks (e.g. elevation above channel, distance from divide). This algorithm can process data sets of hundreds of millions of cells (11,424 rows by 13,691 cols) in a few minutes instead of a few days (or not at all for some algorithms that fail on data sets this large).

Overall, this was an excellent conference, dominated by leading edge research in the area of geomorphic processing of digital elevation data that is of direct relevance and interest to the GlobalSoilMap.net project. We have much to learn from these researchers and much to benefit from maintaining contacts and working relationships with them.”

The case study “Boschoord” (3024 ha) is a small area located in the province of Drenthe, in the northern part of the Netherlands. The Boschoord area is part of the Drenthe Plateau which is underlain by boulder clay deposited by the second last (Saalien) ice sheet. This rather complex genesis created a fragmented landscape in which hydrological differences are strongly linked to this polycyclic landscape development. What makes this dataset especially interesting is that it is an area of low relief, but with distinct geomorphological classes that have been mapped with relatively high accuracy (Koomen and Maas, 2004). The elevations range from 3 to 10 m above the sea level, with a standard deviation of 1.54 m; changes in topography are difficult to notice even in the field. This data set is available only for collaborators. To receive a password to use the data, please contact the authors.

Fig: Location of the study area (a) and the two main DEM data sources used for analysis: DEM25TOPO – generated using ordinary kriging (b) and DEM25LIDAR (c).

Available layers:

- Elevation datasets – This includes the 5 m LiDAR DEM (surveyed in 2004), and a point dataset with 5010 measurements of heights (surveyed in 1960-69). Both datasets show elevations measured with a high precision (±10-20 cm).
- Geomorphological map (GKN50) – The map contains of 12 classes: Ground moraine (3L1), Low plains with ridges (3N3), Peat bog depressions (2R4), Cover sand undulated (3L5), Low plains/depressions without ridges (3N4), Low dunes + plains (3L8), Low plains/depressions without ridges (3N4), Cover sand undulated (3K14), Undulating ground moraines (3L1), Ground moraine (high) (3L2a), Low dunes + plains (3L9), Areas partially covered with cover sand (2M14), Low dunes (4K19), and Cover sand areas (2M13).
- Topographic data – Includes all roads and infrastructure, land use classes and similar features from the TOP10VECTOR basic topographic map of the Netherlands (1:5000 scale). This data is used only for orientation purposes.Grid definition:

ncols: 1110
nrows: 1081
xllcorner: 208297
yllcorner: 543057
cellsize: 5proj4:+proj=sterea +lat_0=52.15616055555555 +lon_0=5.38763888888889 +k=0.999908 +x_0=155000 +y_0=463000 +ellps=bessel +towgs84=565.237,50.0087,465.658,-0.406857,0.350733,-1.87035,4.0812 +units=m +no_defsLineage:

The 5020 field measurements of elevation (land survey) collected in the 1960s by the ‘Meetkundige Dienst Rijkswaterstaat’. This was used to generate the 25 m DEM25TOPO. The 5 m LiDAR-based DEM distributed by the Ministry of transportation and water management (measurements in cm). This dataset is also known as “Actueel Hoogtebestand Nederland” (AHN ) (van Heerd et al. 2008).

Data owner:Universiteit van Amsterdam
Reference:Semi-automated identification and extraction of geomorphological features using digital elevation data
Location: Boschoord Netherlands52° 53’ 55.7772” N,6° 13’ 41.2896” E
See map:Google Maps