Letter to Richard Spinrad, Administrator of the National Oceanic and Atmospheric Administration - Wicker Opposes Noaa's Red Snapper Rule

Letter

Dear Dr. Spinrad,

This proposed rule is deeply flawed, based on bad science, and does a disservice to the public. When Amendment 50 gave Gulf States the ability to manage their own recreational red snapper quota, it was embraced throughout the Gulf. However, NOAA's implementation of the Amendment has only further highlighted the significant flaws with federal fishery management. I have expressed these frustrations for the past year to Secretary Raimondo, you, and numerous other NOAA officials. I have sent letters, asked questions at hearings and meetings, and published an op-ed. Despite these efforts, and despite the agency's public commitment to work with Congress on this issue, I have been informed that the only way my feedback will be considered is via a public comment in a rule making. Had NOAA been open to conversations over the past 18 months, there could have been a more thoughtful discussion. The agency was not interested in such a discussion. Instead, it proposed an arbitrary and capricious rule that uses fundamentally flawed data to cut Mississippi's recreational red snapper quota by 60 percent.

The Marine Recreational Information Program (MRIP)

NOAA's Marine Recreational Information Program (MRIP) is a well-intentioned but flawed and biased program. This program collects data in a suite of recreational fishing surveys and produces catch and effort estimates for recreational fisheries. It is of note that these survey methods were piloted in high-population states such as North Carolina and Florida, with the assumption that performance would be identical across other states. Despite this assumption, both state managers and stakeholders have raised concerns about this program, and many of these concerns have been documented by various National Academy of Sciences reports on the topic (2006, 2017, and 2021). The lack of confidence in MRIP is also demonstrated by the number of states that have chosen to leave the program and developed alternative surveys to improve catch estimates.

Of particular relevance, the 2021 report notes that "MRIP is designed as a general survey, and its precision is greatest for annual estimates at larger geographic scales; the precision of MRIP estimates is much lower of shorter periods of time and smaller geographic areas." The estimates from Mississippi, which has only 44 miles of coastline and accounts for less than three percent of the coastline of the United States Gulf Coast, routinely has high percent standard errors along with exceptionally volatile point estimates in the annual MRIP count. This clearly makes MRIP an inappropriate tool to apply to the state. Doing so is irresponsible and ignores the blatant biases produced by the overarching federal approach.

Many States Do Not Use Federal Data

It is also noteworthy that states with more resources have been allowed to transition away from federal surveys without calibration. In the mid-1990s, Washington State stopped participating in federal data collection for recreational fisheries. In Oregon, when discrepancies between a state recreational survey and MRIP's precursor survey (Marine Recreational Finfish Statistical Survey or MRFSS) were discovered, MRFSS was simply discontinued. In subsequent years, MRIP data was collected in Oregon for supplementing the state data, but the Oregon data was considered the base, as it should have been (https://www.dfw.state.or.us/MRP/salmon/docs/ORBS_Design_2021.pdf). Since 2004, West Coast States have all collected their own recreational fishery data. NOAA never demanded those states calibrate their data to federal data, despite their shared management of fish stocks. (https://www.federalregister.gov/documents/2004/03/09/04-4744/magnuson-stevens-act-provisions-fisheries-off-west-coast-states-and-in-the-western-pacific-pacific#p-61). This notion of a "common currency" seems to target specifically the Gulf of Mexico red snapper fishery.

Within the Gulf, Texas never participated in the MRIP program and is allowed simply to continue using its state data to comply with Amendment 50. In 2016, Louisiana stopped participating in MRIP.

Finally, it is worth noting that the two states most disadvantaged by this rule, Mississippi and Alabama, have the lowest per capita income of the Gulf States and are clearly the most disproportionately affected by the standardized MRIP methodology.

Concerns with the MRIP FES-CHTS Calibration

As NOAA has worked to improve MRIP, it transitioned from the Coastal Household Telephone Survey (CHTS) to a mail-based Fishing Effort Survey (FES) requiring a calibration between FES and CHTS for MRIP. The CHTS survey was terminated in 2016. Of relevance to Gulf of Mexico red snapper, the red snapper quota was established in MRIP CHTS pounds, but since the CHTS survey was discontinued, NOAA needed to compare the MRIP FES data to state data.

NOAA spent considerable time and effort developing a method to calibrate the MRIP FES data to the MRIP CHTS data. There was a workshop in 2017 that resulted in a proposed calibration model based on the Fay-Herriot small estimate model. Despite a large effort by NOAA to develop a scientifically defensible calibration, NOAA opted to use a simple linear calibration of the overlap in FES and CHTS data by state data. Different survey types are affected by different assumptions, sampling biases, and non-sampling errors, which is why a linear comparison between historical and new survey types is often avoided. The reason that this simple linear calibration was arbitrarily chosen when the agency had spent such considerable resources developing a scientifically robust calibration between FES and CHTS is not clear. What is clear is that this is another example of mismanagement of the fisheries by NOAA.

A state-by-state calibration, as used in this proposed rule, could amplify MRIP's known shortcomings with respect to small sample sizes. Moreover, if NOAA's assentation that continuous datasets are essential to fisheries management is correct, then NOAA should have produced a Gulf-wide calibration between FES and CHTS, rather than needing to calibrate data state by state on an ad hoc basis. If this calibration has not occurred, it is unclear how NOAA has been calculating the recreational quotas over the last several years. Finally, when my staff requested the CHTS and FES Mississippi data used to create the calibration, they were told that CHTS was terminated in 2016, but the calibration workshop documents described using CHTS data through 2019. This inconsistent nature of calibrating FES to CHTS data is not the best available science, and it is an inexcusable injustice to the recreational fishing industry of Mississippi.

Based on a presentation from NOAA's Southeast Regional Office (SERO) to the Science and Statistical Committee on August 11, 2020, the choice not to use the Fay-Herriot small estimate model for Mississippi was never properly explained or justified. For the calibration years used in this proposed rule, MRIP's percent standard errors (PSEs) for Mississippi red snapper catch were high (PSE of 51.5 in 2015, PSE of 32.2 in 2016, and PSE of 39 in 2017). Although I fully appreciate NOAA's interest in calibrating to a continuous time series, NOAA itself says data with PSEs over 50 are not published because they are unreliable and that even data with PSEs over 30 should be viewed with caution. Contrary to NOAA's assumption in using such data in this proposed rule, using bad data is often worse than including no data. An alternative approach could be to use Gulf-wide red snapper data to calculate a FES:CHTS ratio. This approach would take advantage of the larger Gulf-wide sample size. NOAA has a responsibility to minimize the impact of well-known issues for small sample sizes in the MRIP data sets.

MRIP Values for Mississippi

As noted above, NOAA's MRIP data for Mississippi is neither precise nor accurate. From 2015-2021, the percent standard error varied between 21.7 and 51.5 and point estimates of harvest fluctuated 400 percent across years with similar season structures. Moreover, the season of peak catch as inferred from the wave estimates are inconsistent from year to year. This means MRIP is failing to capture even the basic seasonal cycle of red snapper fishing or base recreational fishing in general. Additionally, the number of survey intercepts by MRIP's APAIS does not have any correlation with the estimates of catch. The MRIP data for Mississippi fails two basic gut checks based on data derived explicitly from the MRIP program itself.

Tails N' Scales is Best Available Science

The MRIP FES response rate is only 30 percent. Additionally, though NOAA can follow up on responses to the surveys, there is zero ability for NOAA to verify a survey response with on-the-water results. Mississippi's Tails N' Scales program has a 95 percent trip compliance rate. Between both water and on-land compliance checks and validation surveys, over 15 percent of the entire fleet is stopped during the Red Snapper season. The PSE value for Tails N' Scales for 2018 is 3.8; the PSE for MRIP's Mississippi red snapper data was 28.8. For 2019, the PSE for Tails N's Scales was 1.3; MRIP was 20.9. These differences are staggering and cannot reasonably be ignored. MRIP estimates show that Mississippi has 1,500 boats fishing for red snapper per day; the busiest day the Mississippi Department of Marine Resources saw this season was 268 boats. Again, with only a 44-mile coastline, the state knows what is happening in its waters.

It is incomprehensible not to recognize Tails N' Scales as a robust program that produces the only accurate estimates for harvest of Red Snapper in Mississippi. It is deeply frustrating that NOAA, fishermen, and environmentalists all agree on the need for better data and Mississippi has spent the resources needed to achieve that goal, only to have NOAA propose a 60 percent quota reduction based on a biased federal survey that shows no accuracy or consistency when producing estimates of harvest. Clearly NOAA is more interested in giving lip service to better data than properly managing this critical fishery for the public.

Calibration

As noted above, NOAA values its long data sets even where statistically useless and has required the states to calibrate the state data to the MRIP data, despite the many documented failings in MRIP described above. For the calibration of MRIP to Tails N' Scales, the Science and Statistical Committee (SSC) for the Gulf of Mexico Fishery Management Council (the Council) used two years of data (2018 and 2019) to create a calibration. The Committee did not consider 2020 in the calibration, even though the fishing year was done when they met. The high variance and imprecise nature of MRIP data means a seeming arbitrary choice, such as not including 2020 MRIP data, has a dramatic impact on the accuracy of the ratio. NOAA has made the informal suggestion that 2020 MRIP data should not be considered because pandemic-related disruptions resulted in some missing data. However, Mississippi continued to conduct surveys and ensure adequate MRIP coverage throughout the pandemic. NOAA used imputation, a statistically valid method, to compensate for the missing data, which creates a usable catch number. Moreover, its PSE of 33.3 is consistent with other years, suggesting the data should be included. To highlight the arbitrary nature and true volatility of the calibration, if only 2020 were used, Mississippi would have received a 48 percent quota increase. Notably, Louisiana's calibration was based on only a single year's data. Using 2018-2021 data would be statistically more robust, though again, the overall value of the MRIP data is questionable.

In their August 11-12, 2020 meeting, the SSC recognized the shortcoming of the calibration and realized that Mississippi was being treated unfairly. The meeting minutes state, "The SSC recognized that the difference in methodology by the state and federal surveys should be explored further, as to not penalize a state when the difference after calibration greatly reduce the state's quota." This is exactly what has happened to Mississippi. Further, with respect to whether Tails N' Scales data was actually being incorporated, "The SSC also agreed that scaling a state's data to MRIP-FES is not the same as calibrating those data, and that scaling to MRIP-FES is tantamount to using the MRIP-FES data." In short, this was not actually a "calibration," but instead was a dismissal of what the SSC described as "a near-census of that state's in-season catch and effort."

Nearly two years ago, the SSC realized the shortcoming of the calibration and suggested the need for more work. I have also repeatedly asked NOAA to focus on this issue. Congress appropriated $2 million for NOAA to work on this issue. The Gulf of Mexico Fishery Management Council allowed for a nearly two-year delay in the rule to take effect so that NOAA could work on this issue. This proposed rule makes no improvements to the calibration that two years ago was identified as flawed. It should be rejected by the Administrator and sent back to the Council.

Source