Supplemental Information MRFSS Data Queries

SAMPLE COVERAGE

The Marine Recreational Fisheries Statistics Survey has undergone many changes in sampling scope since it began.

Changes in sampling coverage by wave, state, and mode affect summarized MRFSS estimates at the state level and higher (eg. subregion or regional estimates).

The NMFS Beaufort Laboratory conducts the Headboat Logbook Survey to provide headboat (partyboat) catch and effort for the Southeast Region (NC-TX). Data are available from:

Southeast Fisheries Science Center
Beaufort Laboratory
101 Pivers Island Road
Beaufort, NC 28516-9722

The Texas Parks and Wildlife Department has conducted their own survey of marine recreational fisheries since 1974. Estimates for Texas are available from:

Texas Parks and Wildlife Department,
8400 Smith School Road,
Austin, Texas 78744

The Pacific states conduct surveys of salmon fishing, ocean-boat fishing, and California passenger fishing vessels. Estimates for these fisheries are available from:

Pacific States Marine Fisheries Commission,
45 SE 82nd Drive, Suite 100,
Gladstone, OR 97027

The Alaska Department of Fish and Game conducts surveys of recreational fishing in that state.

PSE (Proportional Standard Error)

With one exception, no matter what type of information you requested, the PSE for that information was automatically included.

The exception is the PSE for mean lengths.

• Calculations of mean lengths involve weighting mean lengths by the estimated number of fish in a cell (state/wave/fishing mode/fishing area/species).
• These calculations are quite complicated and rely on various assumptions.
• Because of the assumptions, it is better that these calculations be made on a case-by-case basis.
• Use of mean lengths is not as useful as length distributions and is not used widely.
• Because of its limited use, we do not expect the absence of the PSE to cause any problems or lead to incorrect conclusions about a fishery.

The PSE, or proportional standard error, expresses the standard error of an estimate as a percentage of the estimate and is a measure of precision.

• Precision refers to the dispersion of sample measurements used to calculate an estimate and the resulting variability in the estimate.
• Large PSEs indicate high variability around estimates and therefore low precision.
• It is desirable to have small PSE's and more precise estimates.
• There is a direct relationship between precision and sample size.
• When we group year, state, wave, or mode estimates, sample size increases and so does precision.
• Catch estimates for commonly caught species that show up in the marine recreational fishery often are more precise than for rare event or pulse fisheries.

Data users should consider the width of confidence intervals surrounding estimates before drawing any far-reaching conclusions from point estimates.

• A confidence interval is calculated as the estimate minus 1.96 times the standard error (the lower limit) and the estimate plus 1.96 times the standard error (the upper limit).
• A 95% confidence interval means we are 95% sure that the true value lies between the lower limit and the upper limit.

CPUE (Catch per Unit Effort)

Catch per trip effort (cpue) estimates use effort based on all trips taken, not directed effort (trips targeting a specific species and trips where a specific species was caught, whether targeted or not). Direct effort queries for selected species will be made available in the future.

USE CAUTION WITH WEIGHT DATA

Weight estimates are minimums and may not reflect the actual total weight landed or harvested.

MRFSS weight estimates are calculated by multiplying the estimated number harvested in a cell (year/wave/state/mode/area/species) by the mean weight of the measured fish in that cell. Sometimes we have an estimate of harvest but no mean weight, either because

• the harvest is all reported by the anglers (B1), or
• because for some reason the interviewers couldn't weigh any fish (fish too big, already gutted and gilled, etc.).

If a cell is missing a mean weight, and if we have at least two fish measured in the state (all fishing areas and modes combined),

• We substitute the mean for the whole state for that wave.
• We need two measured fish to get a variance estimate.

After state substitution, if the mean weight is still missing,

• We use the mean from the whole subregion for that wave.
• The "two fish rule" still applies.

After subregional substitution, if the mean weight is STILL missing, we give up and leave a missing weight estimate. At that point,

• It is up to the user to determine whether to substitute, and
• What substitution is most appropriate to use (a mean from the preceding and following waves, the whole year, same wave over years, whole Atlantic & Gulf coast, some complicated regression model, whatever).
• We don't make those decisions because the information needs and sensitivity of the data vary among species.

The phenomenon of missing weights is more widespread with rarely caught species and with large fish (i.e. tunas).

The existence and/or extent of missing weights for your query can be examined by requesting data at the cell level: (by year/wave/state/by mode/by area/by species (time series)).

LENGTH DISTRIBUTIONS

Length distributions are available for selected species by inch group.

• Length distributions are available for the 5 most recent complete years on the Atlantic and Gulf coast and the 3 most recent years on the Pacific coast.
• Length distributions are based on cells (state/wave/fishing mode/fishing area/species) where at least five individual fish measurements were taken.
• This was done so that distributions are not skewed by unrepresentative fishing modes, areas, or waves.
• The raw length distributions are weighted by the estimated number of fish harvested (CATCH TYPE A + B1) in each cell so that the distributions are correctly summarized across time periods and geographic areas.

CATCH DISTRIBUTIONS

Catch (fish per angler) distributions are available for selected species.

• Catch distributions are available for the 5 most recent complete years on the Atlantic and Gulf coast and the 3 most recent years on the Pacific coast.
• Catch distributions are based on cells (state/wave/fishing mode/fishing area/species) where at least three interviews ocurred with anglers who harvested the selected species.
• The raw catch distributions are weighted by the estimated number of trips in each cell so that the distributions are correctly summarized across time periods and geographic areas.