Title: A landscape classification approach for watersheds of the Pacific Northwest: is aquaticecosubregiona
1A landscape classification approach for
watersheds of the Pacific Northwest is
aquaticecosubregionalizationeven a word?
- Chris Jordan, Steve Rentmeester, Carol Volk, Mimi
DIorio, George Pess, Tim Beechie - NOAA-NWFSC, Seattle
2What are we doing, and why?
- Classify the aquatic-landscape of the Pacific
Northwest based on relevant broad-scale
characteristics - Major determinants of watershed processes
- Immutable geomorphic characteristics
- Human impact
- Data analysis support
- Environmental variance partitioning
- Evaluation tool for site selection
3A Made-up Example of What We Want the Output to
Look Like
4A couple of examples of something similar, but
not quite the same
- Hessburg et al. 2000. Ecological subregions of
the ICRB based on PVG, Temp-precip, solar
radiation, elevation. - Omernik et al. 1999, US EPA Level III IV
Ecoregions based on terrestrial vegetation
assemblages.
5(No Transcript)
6(No Transcript)
7What are we doing, and why?
- Classify the aquatic-landscape of the Pacific
Northwest based on relevant broad-scale
characteristics - Data analysis support
- Evaluation tool for site selection
- Assess representativeness of current monitoring
and restoration efforts. - Locate additional monitoring and restoration
projects.
8(No Transcript)
9How are we doing this?
- Taking commonly available spatial data w/
consistent coverage across study area. - Generating functional data layers from above.
- Attributing 6th field watersheds with a single
value for each input data layer. - Grouping watersheds into clusters of like, or
classes.
10Input Data
Climate
- Annual Precipitation
- Month of Max Precipitation
- Growing Degree Day
Topography
- Median Elevation
- Median Hill Slope
Geology
- Stream sediment production
- Water chemistry
Channel Network
- Density (by gradient)
- Complexity (valley width)
- Stream power
- Tributary junctions
- Watershed shape
11How are we doing this?
- Taking commonly available spatial data w/
consistent coverage across study area. - Generating functional data layers from above.
- Attributing 6th field watersheds with a single
value for each input data layer. - Grouping watersheds into clusters of like, or
classes.
12(No Transcript)
13(No Transcript)
14How are we doing this?
- Taking commonly available spatial data w/
consistent coverage across study area. - Generating functional data layers from above.
- Attributing 6th field watersheds with a single
value for each input data layer. - Grouping watersheds into clusters of like, or
classes.
15Hydrologic Unit Code
6th field HUCs Sub-watersheds (10,000-40,000
ac)
16Five data layers 6th field watersheds with a
single values for each input characteristic.
17Five data layers 6th field watersheds with a
single values for each input characteristic.
18How are we doing this?
- Taking commonly available spatial data w/
consistent coverage across study area. - Generating functional data layers from above.
- Attributing 6th field watersheds with a single
value for each input data layer. - Grouping watersheds into clusters of like, or
classes.
19Processing Step
Processing Tools
Spatial Analyst Zonal Statistics Reclassify
Raster
Compile categorical data for 6th order HUCS and
build as attributes into a GIS shapefile Convert
features from vectors to 200m raster
grids Stack separate raster integer grids into
one multi-band raster file Apply ISOCLUSTER and
Maximum Likelihood Classification algorithms to
separate classes based on pixel
spectra Evaluate spatial patterns using
Fragstats
Spatial Analyst Convert Features to Raster
Raster Calculator or Command Line Make Grid
Stack or Composite Bands Tool
Spatial Analyst Tools Command Line ISOCLUSTER
Fragstats Patch Class and Landscape Metrics
20(No Transcript)
21Where are we and next steps
- Need to resolve 200m pixel v. 6th HUC grain
- Need to clean up a few more data layers
- Erosion potential v. Slope x Area
- T, R, S
- Month of max ppt v. hydro regime
- Need to resolve classification tool
- ISODATA v. MCLUST
- Need to make maps and get feedback
- Need to move on to anthropogenic layers