Geographic Resource Solutions

  • Increase font size
  • Default font size
  • Decrease font size
Home Natural Resource Mapping Redwood National and State Parks Vegetation Mapping Project

Redwood National and State Parks Vegetation Mapping Project

Started in the spring of 2008 and completed in mid-2013, this NVCS vegetationRedwood Forestornia.    classification and mapping project performed for the National Park Service Inventory and Mapping Program encompassed over 136,000 acres of extremely steep, inaccessible, and rugged lands that ranged from flat beach sands to steep nearly barren ultra-maphic rock slopes, from meadows and prairies to oak woodlands, and from cutover young growth stands to cathedral-like virgin old-growth redwood-fir forests in Northernwestern Calif


Slopes were typically more than 50% and all access to field sample sites was on foot by cross-country hiking from the nearest road or trail system, sometimes more than 3 miles away. While some areas were rocky and nearly barren, the majority of the project area was covered by extremely dense and jungle-like vegetation, which in combination with the large woody debris and steep slopes made field conditions hazardous nearly every day.

GRS was the prime contractor on this project and was responsible for field sampling and vegetation mapping efforts. GRS acquired two scenes of Landsat TM5 imagery and performed an illumination correction on these images. GRS then classified the imagery to develop a stratified field sampling plan and assure the adequate sampling of the diversity of different vegetation types found in the project area. The resulting stratification identified approximately 90 separable spectral signatures in the project area that are believed to represent different vegetation and land-cover types or complexes of types. Field data were collected identifying all known species and collecting unknowns when identification was uncertain. Cover estimates by species were developed using line-point transect samples, or ocular estimates, and trace species were identified and noted as they were encountered. Transect sample data included individual tree diameter (dbh) and canopy position. In addition, Fire Monitoring (FireMon) sampling protocols were implemented to sample fine-woody debris size classes, coarse-woody debris decay classes, and soil-litter profiles. Each field site was documented with GPS point(s) and photography. All data were collected using TransIn running on ASUS PDAs that verified field data as it was collected. All field data were collected to be compatible for subsequent processing using Vegetation Classification tools, such as ordination and TWINSPAN analysis to identify the National Vegetation Classification System (NVCS) types that would be mapped during the image processing portion of this project. Development of the NVCS Classification was the responsibility of Dr. Ayzik Solomeshch from University of California at Davis. The RNSP Vegetation (NVCS) and Land-cover Key was initially developed by Dr. Solomeshch and completed by Ken Stumpf of GRS. Ken Stumpf performed Vegetation Classification quality control efforts to assure that field data were correctly classified using the NVCS Associations that had been developed. GRS’s Quality control feedback resulted in seven iterations of modifications to the RNSP Vegetation and Land-cover Key until it was finalized in March, 2013.

GRS classified the RNSP project area using our Discrete Classification Mapping Methodology.  Confusion was evaluated and training site fidelity was checked. GRS produced classification maps that contained 1,770 unique classes that represented the different NVCS Associations of the project area. GRS’s ecological rule-based pixel aggregation process aggregate was used to form polygons that meet minimum mapping unit size requirements of 0.5 hectares. This process aggregated pixels into polygons based on user defined rules and relationships that represent ecological similarities and dissimilarities of type characteristics. Polygon attributes include the NVCS Association and Alliance type names, average tree size, and cover density values by lifeform, as well as for major species.

An Accuracy Assessment being performed by an independent 3rd-party contractor is currently underway.

The following map shows some of the many detailed quantitative stand (polygon) attributes developed in the RNSP Discrete Classification map data set. While the map can be displayed using the color-coded NVCS Association names the data set represents a far greater level of attribution and detail which can be used to generate many, many different types of maps.


The following maps represent the cover distribution (magnitude and extent) of several major species found in the Project Area (SeqSem=Redwood, PseMen = Douglas-fir, AlnRub = Red Alder, and LitDen = Tanoak):


In summary, this project was mapped under the 12-Step Guidance of the NPS/USGS Vegetation Inventory Program. GRS participated in NVCS Vegetation Classification efforts and assisted in the development of the NVCS Key. All vegetation and land-cover type mapping by GRS was performed based upon the NVCS, with both Association and Alliance level information assigned to every classified pixel and each aggregated polygon. The project covered an area of rugged and inaccessible terrain encompassing more than 100,000 acres. GRS produced both a pixel/grid data set and a polygon data set. Ken Stumpf of GRS managed this project and performed the image classification and mapping efforts. Local ecological expertise was provided by Ken Stumpf of GRS.



GSA# GS-10F-0451NESRI Consultant