Modeled Biomass for Nearshore Flatfish Species, NOAA, 2011
Identification_Information
Data_Quality_Information
Spatial_Data_Organization_Information
Spatial_Reference_Information
Entity_and_Attribute_Information
Distribution_Information
Metadata_Reference_Information
Identification Information
Section Index
Citation:
Citation Information:
Originator: NOAA Biogeography Branch, Silver Spring, MD
Publication Date: 9/7/2011
Title: Modeled Biomass for Nearshore Flatfish Species, NOAA, 2011
Geospatial Data Presentation Form: raster digital data
Larger Work Citation:
Citation Information:
Originator: NOAA Biogeography Branch, Center for Coastal Monitoring and Assessment
Title: Exploratory analyses and predictive modeling of fish abundance and diversity across the NW Continental shelf
Other Citation Details: Unpublished data
Description:
Abstract:
Spatial predictive models for fish assemblage metrics were developed for waters offshore of Oregon. Analysis was completed using regional fishery-independent trawl datasets. This layer is the "Nearshore Flatfish species biomass probability prediction". Areas with greater probability are more likely to be hotspots. Nearshore Group included: Sand Sole, English Sole, Pacfic Sanddab, Speckled Sanddab, Petrale Sole, Starry Flounder, Butter Sole.
Purpose:
This modeling analysis and resulting layer products are created to assist ODFW and the state of Oregon in their efforts of marine spatial planning for wave energy development and siting, and in the context of updating the state's Territorial Sea Plan.
Supplemental Information:
Two different predictions were provided - a categorical prediction and a probability. This layer represents the probability prediction and is derived from the response variable output from the binary logistic regression. Areas with greater probability are more likely to be hotspots. The categorical prediction is derived from this probability. Probabilities > 50% are classified as a hotspot or "High" and probabilities <= 50% are classified as background or "Low".
Time Period of Content:
Time Period Information:
Multiple Dates/Times:
Single Date/Time:
Single Date/Time:
Single Date/Time:
Single Date/Time:
Single Date/Time:
Single Date/Time:
Single Date/Time:
Single Date/Time:
Single Date/Time:
Single Date/Time:
Single Date/Time:
Single Date/Time:
Single Date/Time:
Single Date/Time:
Single Date/Time:
Single Date/Time:
Single Date/Time:
Single Date/Time:
Single Date/Time:
Single Date/Time:
Single Date/Time:
Single Date/Time:
Single Date/Time:
Single Date/Time:
Single Date/Time:
Single Date/Time:
Single Date/Time:
Single Date/Time:
Single Date/Time:
Single Date/Time:
Single Date/Time:
Single Date/Time:
Currentness Reference: ground condition
Status:
Progress: Complete
Maintenance and Update Frequency: None planned
Spatial Domain:
Bounding Coordinates:
West Bounding Coordinate: -126.700294
East Bounding Coordinate: -122.563266
North Bounding Coordinate: 48.811589
South Bounding Coordinate: 39.987864
Keywords:
Theme:
Theme Keyword Thesaurus: N/A
Theme Keyword: fishery-independent
Theme Keyword: fish model
Access Constraints: Data may be viewed but not downloaded.
Use Constraints:
These data were developed for wave energy siting and other marine spatial planning efforts for the state of Oregon for use by ODFW. These data are not to be used by other entities or for purposes other than the original intent, without prior permission. DLCD is authorized to display these data in Oregon Marine Map for spatial planning purposes.
Point of Contact:
Contact Information:
Contact Person Primary:
Contact Person: Arlene Merems
Contact Organization: Oregon Dept. of Fish and Wildlife
Contact Address:
Address Type: mailing and physical address
Address: 2040 SE Marine Science Dr.
City: Newport
State or Province: OR
Postal Code: 97365
Contact Voice Telephone: 541-867-0300 x246
Data Set Credit: NOAA Biogeography Branch
Native Data Set Environment: Microsoft Windows Vista Version 6.1 (Build 7601) Service Pack 1; ESRI ArcCatalog 9.3.1.4000
Data Quality Information
Section Index
Logical Consistency Report:
An accuracy assessment using 20% of data removed prior to analysis was completed to assess the validity of model results in nearshore (<3nm) and offshore areas (>3nm). All models provide reasonably reliable results, except for diversity. The diversity model should NOT be used for further analyses. Overall accuracy ranged from 64% to 87%.
Lineage:
Source Information:
Source Citation Abbreviation: ODFW /OSU trawl data
Source Contribution:
All metrics were standardized by area swept (distance X trawl width) to allow distinct trawl datasets to be merged. Water hauls identified following Zimmerman et al. 2011 (collected before 1995 and with less than 1kg/10000m2) were removed from analysis. Water hauls are abnormal data. All data in the ODFW/OSU were analyzed.
Source Information:
Source Citation Abbreviation: NWFSC shelf and slope surveys
Source Contribution:
All metrics were standardized by area swept (distance X trawl width) to allow distinct trawl datasets to be merged. Water hauls identified following Zimmerman et al. 2011 (collected before 1995 and with less than 1kg/10000m2) were removed from analysis. Water hauls are abnormal data. Only data with haul type =3 and performance >=0 from NWFSC were analyzed.
Source Information:
Source Citation Abbreviation: AFSC shelf and slope surveys
Source Contribution:
All metrics were standardized by area swept (distance X trawl width) to allow distinct trawl datasets to be merged. Water hauls identified following Zimmerman et al. 2011 (collected before 1995 and with less than 1kg/10000m2) were removed from analysis. Water hauls are abnormal data. All data in the ODFW/OSU and AFSC were analyzed.
Process Step:
Process Description:
Spatial analysis was undertaken using a contiguous grid developed by joint efforts between the Biogeography Branch, ODFW and TNC. The grid was derived from BOEMRE lease blocks (http://www.boemre.gov/offshore/mapping/) and covered shelf and slope waters. Grid cells measured 1600m by 1600m, such that 9 grid cells fit within 1 lease block. The resolution of grid cells determined the resolution of output.
Fish metrics were predicted to all grid cells in the spatial analysis grid using fish-environment relationships modeled using boosted regression trees for continuous variables and binary logistic regression trees for categorical data (High/Low classes). High values (hotspots) were defined as the top 10% of values (=>90th percentile). For all assemblage level fish metrics, the binary classification outperformed continuous predictions. Due to poor performance and high uncertainty/error for continuous metrics, only hotspot maps were produced and assessed with an independent data set (random 20% of nearshore & offshore trawl samples held back n=1,534). See example in Fig 1 below.
The independent map accuracy assessment provided an error matrix with the percentage of correctly classified grid cells for each class. The continuous variables were assessed by the strength of correlation between predicted and actual values.
Upon request we mapped both continuous abundance and biomass and hotspots for the nearshore fish group.
Overall, the best performing models based on independent map accuracy assessment were logistic regression trees grown independently by the algorithm without cross-validation or any other form of testing. These models were computed using the maximum allowable number of trees (usually 1000 trees) and in most of the models all environmental predictors contributed to the final predictions, although the first three contributed most.
Models used 8 distinct spatial layers to derive 42 environmental predictors. Spatial layers and measurements used to develop predictors are defined below.
Basic Spatial Layer Measurement Criteria
• Distance from shore (Average distance to grid cell centroid Shore identified as 0 m on Coastal Relief Model)
• Distance from shelf edge (Average distance to grid cell centroid Shelf edge was 200m isobath)
• Benthic Habitat Richness* (Number of distinct habitat classes in grid cell)
• Habitat class derived from Benthic Habitat Majority * (Single unique habitat class with greatest area in grid cell)
• Chlorophyll hotspots (Presence of highest concentration in grid cell - Provided by TNC.)
Chlorophyll-a concentrations into two categories, high and low, using >=2 SD above the mean for the high, 1 to 2 SD above the mean for low concentrations
• Upwelling hotspots (Presence of longest upwelling persistence in grid cell - Provided by TNC.)
Upwelling split into two categories, high and low persistence, using >=1.5 SD below the mean for the high persistence and 0.5 to 1.5 SD below the mean for low persistence
• Depth** (Mean, minimum, maximum, range at 1600, 4800, 8000 spatial scales Coastal Relief Model (CRM) used)
• Bathymetric complexity** (Mean, minimum, maximum, range at 1600, 4800, 8000 spatial scales Derived from CRM using ESRI spatial analyst extension and 3 X 3 CRM cell neighborhood)
• Bathymetric slope** (Mean, minimum, maximum, range at 1600, 4800, 8000 spatial scales Derived from CRM using ESRI spatial analyst extension and 3 X 3 CRM cell neighborhood)
• Bathymetric Slope of Slope** (Mean, minimum, maximum, range at 1600, 4800, 8000 spatial scales Derived from CRM using ESRI spatial analyst extension and 3 X 3 CRM cell neighborhood)
• Bathymetric Aspect ** (Mean, minimum, maximum, range at 1600, 4800, 8000 spatial scales Derived from CRM using ESRI spatial analyst extension and 3 X 3 CRM cell neighborhood)
* Benthic habitat maps developed using best information available. Merged v3.1 and v3.5 maps provided by ODFW. Merge defaulted with 3.1, unless 3.5 was available, which occurred within State waters. These datasets were later determined to have errors. After careful deliberation and analysis of predictor contributions we determined that the contribution of these derived layers would not impact predicted spatial patterns.
** Min, mean and max metrics used to characterize bathymetry and derived bathymetric variables (layers 5-8) were calculated for 3 spatial scales. Spatial scales were equivalent to a 1600m X 1600 m area, 4800 m X 4800 m area and a 8000m X 8000m area. These areas correspond to the lowest resolution possible given the spatial framework we were working in and from there.
Process Contact:
Contact Information:
Contact Person Primary:
Contact Person: Chris Caldow
Contact Organization: NOAA Biogeography Branch, Center for Coastal Monitoring and Assessment
Contact Position: Marine Biologist, Acting Branch Chief
Contact Address:
Address Type: mailing address
Address:
Biogeography Branch, NOAA/NOS/NCCOS/CCMA, N/SCI1,
1305 East-West Hwy, 9th floor,
City: Silver Spring
State or Province: MD
Postal Code: 20910
Contact Voice Telephone: 301-713-3028, ext. 164.
Contact Electronic Mail Address: chris.caldow@noaa.gov
Cloud Cover: 0
Spatial Data Organization Information
Section Index
Direct Spatial Reference Method: Vector
Point and Vector Object Information:
SDTS Terms Description:
SDTS Point and Vector Object Type: G-polygon
Point and Vector Object Count: 46220
Spatial Reference Information
Section Index
Horizontal Coordinate System Definition:
Planar:
Map Projection:
Map Projection Name: Lambert Conformal Conic
Lambert Conformal Conic:
Standard Parallel: 43
Standard Parallel: 45.5
Longitude of Central Meridian: -120.5
Latitude of Projection Origin: 41.75
False Easting: 1312335.958005
False Northing: 0
Planar Coordinate Information:
Planar Coordinate Encoding Method: Coordinate Pair
Coordinate Representation:
Abscissa Resolution: 0
Ordinate Resolution: 0
Planar Distance Units: international feet
Geodetic Model:
Horizontal Datum Name: North American Datum of 1983
Ellipsoid Name: Geodetic Reference System 80
Semi-major Axis: 6378137
Denominator of Flattening Ratio: 298.257
Entity and Attribute Information
Section Index
Detailed Description:
Entity Type:
Entity Type Label: Modeled_Biomass_for_Nearshore_Flatfish_Species_NOAA_2011
Attribute:
Attribute Label: FID
Attribute Definition: Internal feature number.
Attribute Definition Source: ESRI
Attribute Domain Values:
Unrepresentable Domain: Sequential unique whole numbers that are automatically generated.
Attribute:
Attribute:
Attribute Label: Latitude
Attribute:
Attribute:
Attribute Label: NEARBIOMAS
Attribute Definition:
Nearshore species biomass probability prediction. Derived from the response variable output from the binary logistic regression. Areas with greater probability are more likely to be hotspots. Nearshore Group included: Sand Sole, English Sole, Pacfic Sanddab, Speckled Sanddab, Petrale Sole, Starry Flounder, Butter Sole
Attribute:
Attribute Label: OBJECTID
Attribute:
Attribute Label: Shape
Attribute Definition: Feature geometry.
Attribute Definition Source: ESRI
Attribute Domain Values:
Unrepresentable Domain: Coordinates defining the features.
Distribution Information
Section Index
Standard Order Process:
Digital Form:
Digital Transfer Information:
Custom Order Process:
These data were developed for wave energy siting and other marine spatial planning efforts for the state of Oregon for use by ODFW. These data are not to be used by other entities or for purposes other than the original intent, without prior permission. Users may view this data in Oregon Marine Map for spatial planning purposes, but the data is not being provided for download.
Metadata Reference Information
Section Index
Metadata Date: 12/9/2011
Metadata Review Date:
Metadata Future Review Date:
Metadata Contact:
Contact Information:
Contact Organization Primary:
Contact Organization: Oregon Dept. of Fish and Wildlife
Contact Person: Arlene Merems
Contact Address:
Address Type: mailing and physical address
Address: 2040 SE Marine Science Dr.
City: Newport
State or Province: OR
Postal Code: 97365
Contact Voice Telephone: 541-867-0300 x246
Metadata Standard Name: FGDC Content Standards for Digital Geospatial Metadata
Metadata Standard Version: FGDC-STD-001-1998
Metadata Time Convention: local time
SMMS Metadata report generated 1/3/2012