Fishery Dependent Data Collection Office of Science and Technology MREP Southeast St. Petersburg, Florida April 22, 2014 Gordon Colvin
Summary Overview of Fishery-Dependent Sampling Methods: four basic types of data; census vs. sample Commercial fishery sampling methods Recreational fishery sampling methods MRIP U.S. Department of Commerce National Oceanic and Atmospheric Administration NOAA Fisheries Page 2
What are the four basic kinds of fishery-dependent data that we collect? Trip or Effort data Catch data Bycatch Biological data U.S. Department of Commerce National Oceanic and Atmospheric Administration NOAA Fisheries Page 3
Collect it and Forget it? No. Data Management is a large and frequently overlooked part of the job: Systems to receive and get data into database; Database design and management; QA/QC; Integration of data from different sources; Making data available to customers and stakeholders. U.S. Department of Commerce National Oceanic and Atmospheric Administration NOAA Fisheries Page 4
Census Surveys: Count Everything U.S. Department of Commerce National Oceanic and Atmospheric Administration NOAA Fisheries Page 5
Census Surveys: Requirements Collect complete data from all participants resulting in a total count of the measured parameters General requirements: Establish and maintain a complete list of the primary sampling unit (PSU)components; Include all members of the PSU in the fishery; Validate; Enforceable and enforced. U.S. Department of Commerce National Oceanic and Atmospheric Administration NOAA Fisheries Page 6
Is a Census Always Complete (and Accurate)? Sen. Chuck Schumer demands investigation into 2010 U.S. Census after fishy New York numbers BY TINA MOORE / DAILY NEWS STAFF WRITER SUNDAY, MARCH 27, 2011, 4:00 AM Schumer: NY Census senseless By CYNTHIA R. FAGEN Posted: 12:20 AM, March 27, 2011 New York officials have reached a consensus on the Census: It's a mess. No! So, we may need to employ statistical methods to correct for missing and misreported data. Validation is central to this. U.S. Department of Commerce National Oceanic and Atmospheric Administration NOAA Fisheries Page 7
Sample Surveys Sampling public opinion is like sampling soup: One spoonful can reflect the taste of the whole pot if the soup is well-stirred. George Gallup. U.S. Department of Commerce National Oceanic and Atmospheric Administration NOAA Fisheries Page 8
Sample Surveys: Requirements Collect data from a sample of the components of a fishery and expand to an estimate of the total. The estimate has two parts: Point estimate and margin of error. In general, the sample must be representative of the entire population (probability-based or PPS). Untested assumptions introduce potential for bias. Assumptions should be tested, bias measured if possible. U.S. Department of Commerce National Oceanic and Atmospheric Administration NOAA Fisheries Page 9
Anatomy of an Estimate Point Estimates Precision Removing potential bias affects both the point estimate and its associated measure of precision, expressed either as the Percent Standard Error (PSE) or graphically as the Confidence Interval.
The Effect of BIAS Precise, but inaccurate More accurate; Not precise Precise and accurate
To Improve Precision.. Increase Sample Size: Summer Flounder Landings 2007-2011 Atlantic Coast Annual Total Harvest (A+B1) PSE 3,108,974 6.0 2,362,303 9.1 1,829,579 6.8 1,510,707 7.5 1,847,156 7.7 Total Harvest (A+B1) Virginia Annual PSE 397,041 12.0 260,221 18.1 289,075 18.5 260,050 15.2 317,674 19.0 Total Harvest (A+B1) Virginia May-June PSE 146,492 17.7 128,936 29.7 117,149 33.1 85,991 22.3 149,367 34. U.S. Department of Commerce National Oceanic and Atmospheric Administration NOAA Fisheries Page 12
Commercial Fisheries Data Collection On-Board Observers VTR (logbook) VMS Dealer Reports; Dockside Sampling IVR On-Board Electronic Monitoring U.S. Department of Commerce National Oceanic and Atmospheric Administration NOAA Fisheries Page 13
Commercial Fisheries Data Collection Bycatch and Biological data Mainly by sample; bycatch reported in VTR Sea sampling (observers) and port sampling Dockside observers? U.S. Department of Commerce National Oceanic and Atmospheric Administration NOAA Fisheries Page 14
Commercial Fisheries Data Collection Sources of error and uncertainty: Measurement error? Reporting Errors: failure to report; over-orunder reporting; recall errors U.S. Department of Commerce National Oceanic and Atmospheric Administration NOAA Fisheries Page 15
Recreational Fisheries Data Collection Dockside Interviews Angler Diaries Headboat Logbooks and Sampling Biological Sampling Telephone and Mail Surveys U.S. Department of Commerce National Oceanic and Atmospheric Administration NOAA Fisheries Page 16
Recreational Fisheries Data Collection Catch data (including releases): mainly by sample Result is an estimate of average catch per trip: C/T U.S. Department of Commerce National Oceanic and Atmospheric Administration NOAA Fisheries Page 17
Recreational Fisheries Data Collection Trip data: mainly by sample: currently, by telephone But, that will change: Result is total number of trips: T U.S. Department of Commerce National Oceanic and Atmospheric Administration NOAA Fisheries Page 18
Recreational Catch Estimation Basically: C/T (T) = C, or: U.S. Department of Commerce National Oceanic and Atmospheric Administration NOAA Fisheries Page 19
Of course, it s a little more complicated Actual estimation formula: U.S. Department of Commerce National Oceanic and Atmospheric Administration NOAA Fisheries Page 20 1 1 H h n i b j k hijk hi hij n i h h hi h hij n y hi hi hi X X X X X π π
Recreational Data: Sources of Error and Uncertainty Potential bias uncertainty Actual bias errors; examples? Imprecision uncertainty Reporting errors: (Where were you fishing?) U.S. Department of Commerce National Oceanic and Atmospheric Administration NOAA Fisheries Page 21
What is MRIP? A complete overhaul of the way NOAA Fisheries conducts recreational fishing surveys. A meticulous process engaging data partners and stakeholders in an ongoing cycle of evaluation and improvement. A strategic approach to prioritizing and implementing changes to our surveys: Address Fundamental Design Issues Design, Review, and Certify New Methods Implement Improvements Scale Up to Enhance Precision, Timeliness and Coverage Ultimately, a series of regionally based recreational fishing surveys each using MRIP-certified methods. U.S. Department of Commerce National Oceanic and Atmospheric Administration NOAA Fisheries Page 22
Key MRIP Milestones Re-Estimation: 2012 Working with consultants, removed potential sources of bias from previous estimates dating back to 2004. Fundamental improvement and a response to one of the NRC s chief concerns. Recalculated all estimates 2004 to 2011; recalibrated earlier years National Saltwater Angler Registry Successfully launched January 1, 2010. Currently, registry information is collected at the state level for all recreational fishermen except those in HI, PR and USVI. New Catch Survey Design (APAIS) Rolled out in 2013, removes potential sources of design bias. Adjustments being made in 2014 to be responsive to state partner needs. Calibration Workshop this fall U.S. Department of Commerce National Oceanic and Atmospheric Administration NOAA Fisheries Page 23
Current MRIP Priorities Effort Surveys Intention is to abandon the current Coastal Household Telephone Survey: Why? Bias problems! Evaluating the use of mail surveys using a dual frame consisting of National Saltwater Angler Registry and USPS database. For-Hire Reporting Evaluating options for census-based approach, as well as universal electronic reporting. Current GOM & SA Council efforts. Supplemental Surveys Developing strategies to effectively manage pulse/rare event and other fisheries requiring finer data resolution, such as red snapper. West Coast: Reviewing, Improving, and Certifying Survey Methods Expanding both spatial and temporal survey coverage (WA and OR), video monitoring, assessing discards, electronic data collection Working with states to certify current sampling methods, such as the Puget Sound Sampling Program Transition to Implementation Role Regional Implementation MRIP will maintain central role in developing and certifying methods and establishing national standards Regions, through the FINs or equivalent, have responsibility of selecting survey methods and managing data collection Make key decisions on implementation of MRIP-certified methods, and set investment priorities for enhancing coverage, precision and timeliness. U.S. Department of Commerce National Oceanic and Atmospheric Administration NOAA Fisheries Page 24
Questions and Discussion U.S. Department of Commerce National Oceanic and Atmospheric Administration NOAA Fisheries Page 25
Supplemental Slides U.S. Department of Commerce National Oceanic and Atmospheric Administration NOAA Fisheries Page 26
Key Features of New APAIS Design Site assignments must be completed entirely. No substitution of alternate modes or alternate sites. Assignments will be in clusters of 1 to 3 locations, depending on the level of activity ( pressure rating ) of the sites. Sites in a cluster must be sampled in a specified order for a specified duration. Site assignments will cover four 6-hour time blocks, assuring coverage at all times of day. As at present, interviewers are instructed to try to sample all eligible trips, regardless of whether the anglers caught anything. In addition, complete counts of all eligible trips must be made, whether or not they were sampled. See MRIP website for more details. U.S. Department of Commerce National Oceanic and Atmospheric Administration NOAA Fisheries Page 27
Effort Survey Pilot Project MRIP has been conducting pilot studies of new effort survey methodologies since 2008. In general, we have determined that: Surveys that use only a single sample frame a coastal household telephone directory or state angler registry are subject to undercoverage bias. Telephone surveys generally have become subject to non-response error. Based on what we learned from pilot projects in 2008 2011, we have designed two major pilot projects, the results of which will enable a final decision on effort survey design: Dual frame (postal address and license registry), mixed-mode (mail and telephone) pilot conducted in 2011/2012 in FL, GA, SC, NC; A new Single-phase Address-based Sample pilot using a postal address frame with address matching from state angler registries in MA, NY, NC and FL in 2014. See MRIP website and recent Newscasts for more details. U.S. Department of Commerce National Oceanic and Atmospheric Administration NOAA Fisheries Page 28