Marine Data Products

Citizen Monitoring
Profiles
Time Series

ORCA Buoy
Profiles
Time Series

MMP Profiler + ADCP

Historical Comparisons

Comparative Cross-Sections

Freshwater & Terrestrial Data Products

Freshwater Sampling

Land Use

Geology

Marine Life Studies

Diver Observations

Bloom & Fish Kill Observations

Data Access

Land Use

Land Use/Land Cover Analysis

Terrestrial land use patterns and modifications can significantly affect the chemistry and quality of streamwater and runoff draining into Hood Canal. HCDOP is currently assessing nutrient loading from streams with different land use characteristics in order to test the hypothesis that anthropogenic loads contribute to dissolved oxygen depletion. Potential impacts of future land use change are being explored through HCDOP’s terrestrial modeling efforts. In order to provide the most accurate land use/land cover data possible, HCDOP is working on the development of a new land use classification map focused specifically on the Hood Canal watershed. The new map was created through the supervised classification of a Landsat Enhanced Thematic Mapper Plus (ETM+) image from July 30, 2000. Following this supervised classification, urban areas were masked based on their location the PRISM 2002 landcover map and subalpine forests were masked based on proximity to snow-covered areas. The resultant map has 14 land use/land cover categories:

1) Bare Ground/Clearcut
2) Grass/Shrubs/Crops/Early Regrowth
3) Marsh/Wetland/Shoreline/Shallow Water
4) Deciduous/Mixed Forest
5) Water
6) Mature Coniferous Forest
7) Young Coniferous Forest
8) Snow/Ice
9) Cloud
10) Open Forest/Regrowth
11) Sub-Alpine Forest
12) Low Density Urban
13) High Density Urban
14) Cloud Shadow

To obtain a preliminary estimate of the new land use/land cover map’s accuracy, 100 randomly located test points were referenced against aerial photographs from 2003 or parts of the original Landsat ETM+ image that were not used as training sites during supervised classification. Quantitative comparison between these reference data and the new map indicated that it has an overall classification accuracy of >90%. User accuracy levels (computed by dividing the number of points classified correctly in each category by the total number of points classified in that category) were above 90% for all categories except marsh/wetland/shoreline/shallow water (66.67%, 6 test points), open forest/regrowth (83.33%, 6 test points) and high density urban areas (50%, 2 test points). Due to the fact that test point locations were chosen randomly, no test points were located in the cloud or cloud shadow categories, which cover a very small portion of the map’s total area. Qualitative assessment indicates that the cloud shadow category has fairly high classification accuracy, while the cloud category has lower accuracy.