Here is a summary of some things I learned at the meeting in Sioux Falls.
From Tuesday, January 25 to Thursday, January 27, I attended the NED-H Collaborators' meeting held at the U.S. Geological Survey/EROS Data Center located in Sioux Falls, SD. NED-H stands for National Elevation Dataset -- Hydrologic derivatives. I learned a lot of interesting information at this meeting related to NED-H and other great data sets and related activities.
Over 30 people attended this meeting. Several institutions were represented including the USGS, EPA, BLM (Bureau of Land Management), ESRI (software vendors - Arc/Info, ArcView etc.), University of Wyoming, and NOAA (including myself and 3 people from the National Severe Storms Lab-NSSL).
There were informative discussions regarding NED-H, and also NED, NHD, and NLCD (more info below).
NED -- National Elevation Dataset
NED-H -- National Elevation Dataset Hydrologic Derivatives
NHD -- National Hydrography Dataset
NLCD -- National Land Cover Data Set
An excellent website explaining NHD, NED, and NLCD is gisdata.usgs.gov. The development and maintenance of these data sets is a huge, ongoing, interagency effort. Technologies for serving geospatial data on the Internet to a variety of audiences were also a major part of the discussion at this meeting. In addition, a good part of the discussion on Thursday was devoted to a new 3-year Cooperative Research and Development Agreement between the USGS and ESRI that includes emphasis on (1) data maintenance, access, and distribution, (2) developing better tools for preparing data for gridded models (the application focus here is on the groundwater model MODFLOW), and (3) developing methods and tools for using spatial information with decision support systems.
The acronym NED by itself stands for National Elevation Data Set. NED is a 1 arc-second digital elevation model being produced at the USGS EROS data center. NED is a seamless grid with national coverage. The data are stored in 7.5 minute quadrangle blocks (of which there are about 53,800 in the conterminous U.S.). The quality of the data varies with locations around the country due to different types of source data and processing techniques used to derive the data. Quality control and updating the data is an ongoing process at the USGS EROS data center. 1 arc-second DEMs are sometimes also referred to as 30-m DEMs because 1 arc-second corresponds to a distance of near 30-m at the Equator.
NED-H stands for National Elevation Dataset - Hydrologic Derivatives. Hydrologic derivatives include products derived from NED that are useful for hydrologic applications. These derivatives include but are not be limited to:
Although NED is distributed in geographic coordinates, to be pragmatic, NED-H grids will be created in an Albers Equal-Area projection.
The NED-H project is a massive undertaking initiated at the USGS and is just getting started. The USGS estimates that the NED-H project will take on the order of 50 person-years worth of effort, although it is difficult to make an accurate estimate of this at such an early stage. This large amount of work is the reason for collaboration. The final size of NED-H is estimated to be 1.2 terabytes.
The main reason for the high amount of effort required for this project is that although programs for creating the hydrologic derivatives are automated, checks will be required to ensure consistency with reality, and corrections will be made to the DEM to make as accurate a drainage network as possible.
NED-H project coordinators have outlined the project in terms of 3 basic steps.
It is very fortunate that the folks at NSSL have been pro-active in working with the USGS to obtain the best possible data for use in deriving subbasins for FFMP. The agreement reached between the EROS Data Center and NSSL is that EROS will provide the NED data to NSSL, NSSL will do the "blind pass" processing (and in the process derive subbasins needed for FFMP) and return the derivatives to EROS. NSSL has purchased a raid disk and has 6 full-time and 4 part-time Sun Workstations online to facilitate the processing. Although there will certainly be problem areas after the blind pass processing at NSSL, the resulting basin boundaries will be better than anything else that is currently available. I am working with NSSL to include in their processing the calculation of some basic subbasin parameters that will be useful for threshold runoff calculations and perhaps other modeling efforts being undertaken here at HRL.
The National Hydrography Dataset (NHD) is a national data set of lines and polygons that represent surface water features. NHD is a collaborative effort between the USGS and EPA to build a seamless geospatial data set of surface water features from the RF3 files and USGS 1:100,000 scale digital line graph files. The stream detail of NHD varies significantly throughout the country due to variations in source data. NHD provides much more information than RF1 files currently used in IHABBS and included in the threshold runoff database. To give an example, the NHD file for the Tenkiller, OK, HUC contains more than 100 times the number of features in the RF1 file for the same area (2926 line features for NHD compared with 21 for the RF1).
National Land Cover Dataset: 30-m land cover grid of the United States derived from Landsat Thematic Mapper satellite images taken in the early-to-mid 1990's. One of 21 land cover classifications is assigned to each 30-m pixel.
The NHD, NED, and NLCD are national databases with unprecedented spatial resolution. These datasets are ripe for use in hydrologic applications. Some uses of these data may be obvious and straightforward while others may require more model development and testing. For example, NHD files may be useful simply for display purposes or for evaluating existing subbasin boundaries, while the land cover information in NLCD would need to be translated into useful hydrologic information like impervious percent.
HRL should start taking advantage of these data sets in developing new hydrologic applications. The work being done at NSSL to produce a basin boundary data set for FFMP is a good start. The data that comes out of this project will also provide useful inputs for threshold runoff calculations and for testing distributed modeling approaches.