Data Management Plan

To understand various ecosystem dynamics that could explain observed changes and differences in monk seal biomass at French Frigate Shoals (FFS) and Laysan (LAY), Ecopath with Ecosim (EwE) version 6.4.4 modeling software was used to conduct two types of simulations, sensitivity (to understand the system) and hindcast (to evaluate drivers of monk seal population dynamics). EwE models for LAY and FFS were developed and model output was compared between the 2 models. The model output data included in the data sets represent the most important conclusions as reported in the publication from this study. These are (1) the mean of the last 5 years of each perturbation for sensitivity assessment; and (2) hindcast simulations. The first data set was used to understand the main drivers of the trophic structure and energy flow in the ecosystems. Four perturbation scenarios were simulated for a time span of 40 years (2010 through 2049): (a) baseline (compare static structural ecosystem composition between FFS and LAY), (b) altered predator abundance (30% decrease in biomass of main predators in the ecosystem), (c) altered monk seal prey abundance (30% decrease in biomass of main monk seal prey groups, with and without bottomfish to test how much bottomfish accounts for changes, and (d) altered primary productivity (change productivity annually by 10% for both phytoplankton and benthic algae through an annual forcing factor of 0.9 and 1.1 on both groups). The hindcast simulations were performed to evaluate the importance of historical stressors to monk seal population dynamics by selecting the best-fit model for the monk seal biomass time series (1998-2015). Stressors included were (a) fishing with time series based on commercial fishery data, (b) environment with time series of the monthly PDO index as a multiplier for primary productivity, (c) additional mortality simulated with a constant removal of monk seals of varied intensities, and (4) combinations of these 3 stressors (2a and 2c being the most important with regard to the observed trend in monk seal population). Time series observation data of monk seal biomass and benthic bottomfish biomass, as well as catch time series of bank sharks (only for FFS), benthic and demersal bottomfish, bank jacks, and macroheterotrophs (lobsters) were loaded into Ecosim for model fitting. Refer to Weijerman et al (2017) for complete details.

Laysan and French Frigate Shoals of the Northwestern Hawaiian Islands.

Lineage Statement:
Input data was collected for the Ecopath with Ecosim model (EwE) from Parrish et al 2011; PIFSC Ecosystem Sciences Division data, monk seal data, bottomfish stock asessment data, and other sources defined in Weijerman et al 2017 MEPS. The model output data is generated with the EwE model. See Weijerman et al 2017 MEPS for details.

Process Steps:

• Fish species were aggregated into functional groups. All biomass data was converted to metric tons per square kilometer and used as input data for the EwE model. Rate parameters were calculated based on a weighted average by species composition in the functional groups. Landings data of bottomfish were aggregated by year for the spatial areas encompassing the two ecosystems of interest (Laysan and French Frigate Shoals). Model output data is reported without any additional process steps. (Citation: Weijerman M, Robinson S, Parrish F, Polovina J, Littnan C (2017) Comparative application of trophic ecosystem models to evaluate drivers of endangered Hawaiian monk seal populations. Mar Ecol Prog Ser 582:215-229. )

Model dynamics were validated with a hindcast simulation (1998-2015). Time series data of monk seal biomass and bottomfish biomass were loaded into Ecosim for validation and to fine-tune the model by adjusting vulnerability values and so minimize the residuals between predicted and observed time series of monk seal and bottomfish biomass using a least-square fitting criterion. We manually adjusted vulnerability values of interest based on the results from the perturbation scenarios and restricted them to ecological meaningful values (i.e., preferably between 1 and 3 with 10 as maximum value).

Data can be accessed online via the NOAA National Centers for Environmental Information (NCEI) Ocean Archive.

Model data is stored on an external hard drive and backed up to a server at PIFSC weekly by the data steward. Data on the PIFSC server are managed by NOAA IRC and NOAA Fisheries ITS.