Predictive Model of summer Short-beaked Common Dolphin Densities in the California Current Ecosystem, NOAA SWFSC, 2009

Citation:

Citation Information:

Originator: NOAA Southwest Fisheries Science Center
Publication Date: May, 2009
Title: Predictive Model of summer Short-beaked Common Dolphin Densities in the California Current Ecosystem, NOAA SWFSC, 2009
Geospatial Data Presentation Form: vector digital data
Other Citation Details:
Final Technical Report: Predictive Modeling of Cetacean Densities in the Eastern Pacific Ocean (SI-1391).
Prepared for the U.S. Department of Defense, Strategic Environmental Research and Development Program By the U.S. Department of Commerce, NOAA Fisheries, Southwest Fisheries Science Center.

Online Linkage: http://seamap.env.duke.edu

Description:

Abstract:
Best estimates of summer Short-beaked Common Dolphin density (Animals per Square Kilomoter) over the past 15 years in the California Current Ecosystem. Modeled from 16 ship-based cetacean and ecosystem assessment surveys. All data were collected by NOAAÂ’s Southwest Fisheries Science Center (SWFSC) from 1986-2006 using accepted, peer-reviewed survey methods. Data include over 17,000 sightings of cetacean groups on transects covering over 400,000 km.
Purpose:
The Navy and other users of the marine environment conduct many activities that can potentially harm marine mammals. Consequently, these entities are required to complete Environmental Assessments and Environmental Impact Statements to determine the likely impact of their activities. Specifically, those documents require an estimate of the number of animals that might be harmed or disturbed. A key element of this estimation is knowledge of cetacean (whale, dolphin, and porpoise) densities in specific areas where those activities will occur.
Supplemental Information:
Cetacean densities are typically estimated by line-transect surveys. Within United States Exclusive Economic Zone (US EEZ) waters and in the Eastern Tropical Pacific (ETP), most cetacean surveys have been conducted by the US National Marine Fisheries Service as part of their stock assessment research and typically result in estimates of cetacean densities in very large geographic strata (e.g., the entire US West Coast). Although estimates are sometimes available for smaller strata (e.g., the waters off southern California), these areas are still much larger than the operational areas where impacts may occur (e.g., the NavyÂ’s Southern California Offshore Range (SCORE) off San Clemente Island). Stratification methods cannot provide accurate density estimates for small areas because sample size (i.e., the number of cetacean sightings) becomes limiting as areas become smaller. Recently, habitat modeling has been developed as a method to estimate cetacean densities. These models allow predictions of cetacean densities on a finer spatial scale than traditional line-transect analyses because cetacean densities are estimated as a continuous function of habitat variables (i.e., sea surface temperature, seafloor depth, distance from shore, prey density, etc.). Cetacean densities can then be predicted wherever these habitat variables can be measured or estimated, within the area that was modeled.
Time Period of Content:

Time Period Information:

Single Date/Time:

Calendar Date:
Time of Day: Unknown

Currentness Reference: publication date

Status:

Progress: Complete
Maintenance and Update Frequency: None planned

Spatial Domain:

Bounding Coordinates:

West Bounding Coordinate: -132.529867
East Bounding Coordinate: -116.140028
North Bounding Coordinate: 48.582736
South Bounding Coordinate: 29.666627

Keywords:

Theme:

Theme Keyword Thesaurus: N/A
Theme Keyword: Short-beaked Common Dolphin
Theme Keyword: Delphinus delphis
Theme Keyword: Cetacean density

Place:

Place Keyword Thesaurus: N/A
Place Keyword: California Current Ecosystem
Place Keyword: West Coast

Temporal:

Temporal Keyword Thesaurus: N/A
Temporal Keyword: Summer

Access Constraints: Approved for public release; distribution unlimited.
Point of Contact:

Contact Information:

Contact Person Primary:

Contact Person: Jay Barlow
Contact Organization: Southwest Fisheries Science Center, NOAA National Marine Fisheries Service

Contact Address:

Address Type: mailing address
Address: 3333 N. Torrey Pines Court
City: La Jolla
State or Province: CA
Postal Code: 92037-1022
Country: U.S.A

Contact Voice Telephone: (858) 546-7178
Contact Facsimile Telephone: (858) 546-7003
Contact Electronic Mail Address: Jay.Barlow@noaa.gov

Data Set Credit: Barlow, Jay, Megan C. Ferguson, Elizabeth A. Becker, Jessica V. Redfern, Karin A. Forney, Ignacio L. Vilchis, Paul C. Fiedler, Tim Gerrodette, Lisa T. Ballance. 2009.
Native Data Set Environment: Microsoft Windows Vista Version 6.1 (Build 7601) Service Pack 1; ESRI ArcCatalog 9.3.1.4000

Lineage:

Source Information:

Source Contribution:
We use data from 16 ship-based cetacean and ecosystem assessment surveys to develop habitat models to predict density for 15 cetacean species in the ETP and for 12 cetacean species in the California Current Ecosystem (CCE). All data were collected by NOAAÂ’s Southwest Fisheries Science Center (SWFSC) from 1986-2006 using accepted, peer-reviewed survey methods. Data include over 17,000 sightings of cetacean groups on transects covering over 400,000 km.
The expected number of groups seen per transect segment and the expected size of groups were modeled separately as functions of habitat variables. Model predictions were then used in standard line-transect formulae to estimate density for each transect segment for each survey year. Predicted densities for each year were smoothed with geospatial methods to obtain a continuous grid of density estimates for the surveyed area. These annual grids were then averaged to obtain a composite grid that represents our best estimates of cetacean density over the past 20 years in the ETP and the past 15 years in the CCE. Many methodological choices were required for every aspect of this modeling. In completing this project, we explored as many of these choices as possible and used the choices that resulted in the best predictive models. To evaluate predictive power, we used cross-validation (leaving out one survey year and predicting densities for that year with models built using only the other years). Data from the two most recent surveys (2005 in the CCE and 2006 in the ETP) were used for this model validation step.

Process Step:

Process Description:
We explored three modeling approaches to predict cetacean densities from habitat variables: Generalized Linear Models (GLMs) with polynomials, Generalized Additive Models (GAMs) with nonparametric smoothing functions, and Regression Trees. Within the category of GAMs, we tested and compared several software implementations. In summary, we found that Regression Trees could not deal effectively with the large number of transect segments containing zero sightings. GLMs and GAMs both performed well and differences between the models built using these methods were typically small. Different GAM implementations also gave similar, but not identical results. We chose the GAM framework to build our best-and-final models. In some cases, only the linear terms were selected, making them equivalent to GLMs.
Process Step:

Process Description:
We explored the effects of two aspects of sampling scale (resolution and extent) on our cetacean density models. To explore the effect of resolution, we sampled transect segments on scales ranging from 2 to 120 km. We found that differences in segment lengths within this range had virtually no effect on our models in the ETP, but that scale affected the models for some species in the CCE where habitats are more geographically variable. For our best-and-final models, we accommodated this regional scale difference by using a longer segment length in the ETP (10 km) than in the CCE (5 km). To explore the effect of extent, we constructed models using data from the ETP and CCE separately and for the two ecosystems combined. We found that the best predictive models were based on data from only one ecosystem; therefore, all our best-and-final models are specific to either the CCE or the ETP.
Process Step:

Process Description:
We explored five methods of interpolating oceanographic measurements to obtain continuous grids of our in situ oceanographic habitat variables. Cross-validation of the interpolations gave similar results for all methods. Ordinary kriging was chosen as our preferred method because it is widely used and because, qualitatively, it did not produce unrealistic “bull’s eyes” in the continuous grids.
Process Step:

Process Description:
We explored the use of CCE oceanographic habitat data from two available sources: in situ measurements collected during cetacean surveys and remotely sensed measurements from satellites. Only sea surface temperature (SST) and measures of its variance were available from remotely sensed sources, whereas the in situ measurements also included sea surface salinity, surface chlorophyll and vertical properties of the water-column. We conducted a comparison of the predictive ability of models built using in situ, remotely sensed, or combined data and found that the combined models typically resulted in the best density predictions for a novel year of data. In our best-and-final CCE models we therefore used the combination of in situ and remotely sensed data that gave the best predictive power.
Process Step:

Process Description:
In some years, in situ data also included net tows and acoustic backscatter. We explored whether indices of “mid-trophic” species abundance derived from these sources improved the predictive power of our models. The plankton and small nekton (mid-trophic level species) sampled by these methods are likely to include cetacean prey and were therefore expected to be closely correlated with cetacean abundance. We tested the predictive power of models built with 1) only physical oceanographic and chlorophyll data, 2) only net-tow indices, 3) only acoustic backscatter indices, or 4) the optimal combination of all three in situ data sources. We found that models for some species were improved by using mid-trophic measures of their habitat, but the improvement was marginal in most cases. Although the results look promising, our best-and-final models do not include indices of mid-trophic species abundance because acoustic backscatter was measured on too few surveys.
Process Step:

Process Description:
We explored the effect of seasonality on our models using aerial survey data collected in February and March of 1991 and 1992. Due to logistic constraints, our ship survey data are limited to summer and fall seasons, corresponding to the “warm-season” for cetaceans in the CCE. Although some data in winter and spring (the “cold-season”) are available from aerial surveys in California, these data are too sparse to develop habitat models. We therefore tested whether models built from data collected during multiple warm seasons could be use to predict density patterns in the cold season. We used the 1991-92 aerial surveys to test these predictions.
Although the warm-season models were able to predict cold-season density patterns for some species, they could not do so reliably, because some of the cold-season habitat variables were outside the range of values used to build the models. Furthermore, the two available years of cold-season data did not include a full range of inter-annual variation in winter oceanographic conditions. An additional complication is that some cetaceans found in the CCE during the warm season are migratory and nearly absent in the cold season. For these reasons, our best-and-final models based on warm-season data in the CCE should not be used to predict cetacean densities for the cold season.

Process Step:

Process Description:
Our best-and-final models for the CCE and the ETP have been incorporated into a web-based GIS software system developed by Duke UniversityÂ’s SERDP Team in close collaboration with our SWFSC SERDP Team. The web site (<http://serdp.env.duke.edu/>) is currently hosted at Duke University but needs to be transitioned to a permanent home. The software, called the Spatial Decision Support System (SDSS), allows the user to view our model outputs as color-coded maps of cetacean density as well as maps that depict the precision of the models (expressed as point-wise standard errors and log-normal 90% confidence intervals). The user can pan and zoom to their area of interest. To obtain quantitative information about cetacean densities, including the coefficients of variation, the user can define a specific operational area either by 1) choosing one from a pull-down menu, 2) uploading a shape file defining that area, or 3) interactively choosing perimeter points. Density estimates for a user-selected area are produced along with estimates of their uncertainty.
Process Step:

Process Description:
Although our models include most of the species found in the CCE and the ETP, sample sizes were too small to model density for rarely seen species. Additionally, we could not develop models for the cold season in the CCE or for areas around the Hawaiian Islands due to data limitations. To provide the best available density estimates for these data-limited cases, we have included stratified estimates of density from traditional line-transect analyses in the SDSS where available: cold-season estimates from aerial surveys off California, estimates from ship surveys in the US EEZ around Hawaii, and estimates for rarely seen species found in the CCE and the ETP.
Cloud Cover: 0

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: 1892

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

Detailed Description:

Entity Type:

Entity Type Label: Short-beaked_Common_Dolphin_Modeled_Density_NOAA_SWFSC_2009
Entity Type Definition: Cetacean Densities (Animals per Square Kilomoter) in the California California Ecosystem (CCE)

Attribute:

Attribute Label: cid
Attribute Definition: cell ID used by Duke SDSS

Attribute:

Attribute Label: Dde_i_u_h
Attribute Definition: Short-beaked common dolphin (Delphinus delphis) in summer using in-situ model high 90% CI

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 Label: gid
Attribute Definition: grid cell ID from original SURFER file

Attribute:

Attribute Label: Shape
Attribute Definition: Feature geometry.
Attribute Definition Source: ESRI
Attribute Domain Values:

Unrepresentable Domain: Coordinates defining the features.

Overview Description:

Entity and Attribute Overview:
Since fieldnames of shapefiles are limited to 10 characters, the following name scheme was used:
species_model_season_output

species:
Bba = Guild: Berardius (Berardius)
Bmu = blue whale (Balaenoptera musculus)
Bph = fin whale (Balaenoptera physalus)
Dde = short-beaked common dolphin (Delphinus delphis)
Ggr = Risso's dolphin (Grampus griseus)
Lbo = northern right whale dolphin (Lissodelphis borealis)
Lob = Pacific white-sided dolphin (Lagenorhynchus obliquidens)
Mno = humpback whale (Megaptera novaeangliae)
Pda = Dall's porpoise (Phocoenoides dalli)
Pma = sperm whale (Physeter macrocephalus)
Sco = striped dolphin (Stenella coeruleoalba)
Zsm = Guild: small beaked whale (Ziphius and Mesoplodon)

model:
i = in-situ
r = remote-sensed

season:
u = summer

output type:
h = high 90% CI density
l = low 90% CI density
d = average density
e = standard error of density

Distribution Liability:
This report was prepared under contract to the Department of Defense Strategic Environmental Research and Development Program (SERDP). The publication of this report does not indicate endorsement by the Department of Defense, nor should the contents be construed as reflecting the official policy or position of the Department of Defense. Reference herein to any specific commercial product, process, or service by trade name, trademark, manufacturer, or otherwise, does not necessarily constitute or imply its endorsement, recommendation, or favoring by the Department of Defense.
Standard Order Process:

Digital Form:

Digital Transfer Information:

Transfer Size: 0.258

Metadata Date: 11/21/2011
Metadata Review Date:
Metadata Future Review Date:
Metadata Contact:

Contact Information:

Contact Organization Primary:

Contact Organization: Southwest Fisheries Science Center, NOAA National Marine Fisheries Service
Contact Person: Jay Barlow

Contact Address:

Address Type: mailing address
Address: 3333 North Torrey Pines Court
City: La Jolla
State or Province: CA
Postal Code: 92037-1022
Country: U.S.A.

Contact Voice Telephone: (858) 546-7178
Contact Facsimile Telephone: (858) 546-7003
Contact Electronic Mail Address: Jay.Barlow@noaa.gov

Metadata Standard Name: FGDC Content Standards for Digital Geospatial Metadata
Metadata Standard Version: FGDC-STD-001-1998
Metadata Time Convention: local time