Pilot Sentinel Survey to State Work Groups and Recommendations for Pilot Scale Monitoring

After generating a long list of candidate sentinels for monitoring in Long Island Sound and its coastal ecoregions (see Sentinel Indices Matrix, Appendix H), the list of sentinels were prioritized for potential inclusion in a pilot program. The set of desired sentinel attributes that was agreed upon by the technical work groups was also used in the prioritization process. These attributes were as follows:

• They can be measured at multiple sites, so that comparison between sites can be made;
• The climate change signal for the indicator can be distinguished from natural variations or anthropogenic stressors with the appropriate sampling resolution; For biological indicators, they are:
  • representative of regional biological communities, processes, ecosystems and/or
• a species at the edge of its range (fringe) or in a habitat that is limited
• They have an existing or potential data record that would allow comparison of historic, current, and future conditions
• They can be measured and studied feasibly with respect to cost and available technology (or new technology can be developed in order to support their measurement).

Although all five attributes were considered during creation of the Sentinel Indices Table, the bistate work group decided to focus on the first two attributes to prioritize the list of sentinels for pilot scale implementation. The first attribute was selected as a requirement for the success of any proposed sentinel. The second attribute was chosen due to the temporal and budgetary restrictions of a pilot-scale study. The remaining attributes had been used previously by the technical work groups to narrow the list of sentinels in the matrix. The technical work groups advised that analysis of historical/existing monitoring data could yield information in the short-term on climate change signals already present in Long Island Sound. The pilot study is intended to obtain information for use by managers within a two to three year time frame, to be used to leverage funding for a larger, longer-term monitoring program, and to be as cost-effective as possible. For these reasons, a pilot-scale study that combines analysis of existing data with on-the-ground monitoring has been identified as the optimal approach.

An online survey was designed that focused on these two attributes for prioritization of the candidate sentinels. The survey link was distributed to the two state-level technical work groups and each member was asked to rate each sentinel based upon the main attributes, which were described in the survey as follows: 1) A sufficient data record exists to allow comparison of current conditions to relative historic conditions for the sentinel in question in order to identify long-term trends that may be occurring (or have occurred) related to climate change; and, 2) A climate change signal could in theory be distinguished from natural variations or anthropogenic stressors with the appropriate sampling resolution. The categories for rating were: Strongly Disagree, Disagree, Agree, Strongly Agree or Unsure. Four categories of rating were deliberately chosen in order to force participants to give an opinion and avoid the statistical middle in responses. Technical work group members were asked to only respond to sentinels for which they felt comfortable assigning a rating. If they lacked sufficient knowledge to assign a rating, they were asked to respond with “Unsure”. Work group members were given approximately one month to respond to the survey and reminders were periodically sent.

Twenty-three work group members from Connecticut and ten work group members from New York responded to the survey. For a complete list of survey results, please see Appendix K. Responses were assigned a numerical rating: Strongly Disagree (1), Disagree (2), Agree (3) and Strongly Agree (4). “Unsure” ratings received no number and did not affect the analysis. Data from the two states was analyzed separately to prevent the larger response from Connecticut from biasing the outcome. The average rating for each attribute was generated for each of the 37 candidate sentinels. An average rating of 2.5 or higher was considered general agreement by work group members of a sentinel’s potential value. Sentinels were included in the short list if their average rating for both attributes, by both state work groups, was greater than or equal to 2.5. Of the 37 original candidates, this left 17 priority sentinels. These 17 priority sentinels were as follows:

1. Areal extent and distribution of eelgrass
2. Areal extent, diversity, and composition of brackish marshes
3. Areal extent, diversity, and composition of freshwater tidal marshes
4. Areal extent, diversity, and composition of salt marshes
5. Changes in diadromous fish run timing
6. Changes in distribution and marine transgression of marshes
7. Distribution, abundance, and species composition of marsh birds, colonial nesting birds, shorebirds, waterfowl
8. Distribution, composition, and abundance of terrestrial invasive species
9. Extent and distribution of barrier beaches/islands
10. Extent and distribution of coastal forests, shrublands and grasslands
11. Extent and distribution of habitats associated with coastal embayments (e.g., fringe marsh, shorelines and tidal creeks)
12. Extent and distribution of sea cliffs/bluff and escarpments
13. Extent and distribution of unvegetated nearshore (submerged and intertidal) habitats (e.g., mudflats, sandflats, rocky intertidal)
14. Finfish biomass, species composition, and abundance
15. Lobster abundance (based on fishery-independent measurements)
16. Phytoplankton biomass, species composition, and timing of blooms
17. Species composition within coastal forests, shrublands, and grasslands

Before moving forward with development of a Request for Proposals based on these 17 sentinels, the bi-state work group worked to verify data availability in both states. Datasets identified during earlier technical work group meetings were compiled, technical work group members provided additional information, and state agencies were consulted for data availability. The datasets identified through this process are listed in Appendix L and will be listed in the data citation clearinghouse (see Section VIII). The list of candidate sentinels was then further narrowed down based on the availability of existing data in both states in multiple locations as sentinels with greater data availability were ideal for a pilot study, as indicated above. Again, this does not diminish the potential importance of sentinels with little or no current data availability to a larger program, but suggests that they are not appropriate for a small-scale, short-term pilot study. The details of this discussion to narrow the candidate sentinels is included with the data availability in Appendix L. The final list is as follows:

1. Distribution, abundance, and species composition of marsh birds, colonial nesting birds, shorebirds, waterfowl
2. Finfish biomass, species composition, and abundance
3. Lobster abundance (based on fishery-independent measurements)
4. Phytoplankton biomass, species composition, and timing of blooms
5. Species composition within coastal forests, shrublands, and grasslands
6. Areal extent, diversity, composition, and marine transgression of salt marshes

There are two important considerations for the pilot-scale awardee: 1) Ensure ready access to data collected at sites chosen for monitoring; 2) If there are multiple sites, data collection will have to be replicated at each site; therefore, methods must be standardized.