Fig 1.
Study area of the eight Distributed Biological Observatory (DBO) sites across the Pacific Arctic region as they relate to the mean sea ice edge, determined with a 15% sea ice concentration threshold using SMMR, SSM/I, and SSMIS satellite data (decadal means as well as the last three years individually: 2018, 2019, and 2020).
All sea ice edge contours north (south) of Bering Strait represent September (March) conditions. Basemap datasets from Natural Earth [49] and ESRI [50].
Fig 2.
Scatterplot comparisons of the Oregon State University (OSU) primary productivity data compared to the primary productivity values in this study over the Pacific Arctic study region for (a) April, (b) May, (c) June, (d) July, (e) August, and (f) September of 2019.
Fig 3.
Maps of the pan-Arctic region for August 2014 showing (a) primary productivity values as derived in this study, (b) primary productivity values as derived by Kahru et al. (2016), and (c) the difference map of (a)–(b). Basemap datasets from Natural Earth [49] and ESRI [50].
Fig 4.
Time series of primary productivity derived in this study (GSFC PP) and those derived from Kahru et al. (2016) [45] (Kahru VGPM PP) across the Pacific Arctic study region for (a) monthly average values of primary productivity and (b) monthly anomalies of primary productivity.
Fig 5.
Mean values for each of the eight DBO sites (over the years 2003–2020) for (a) satellite-derived monthly sea surface temperatures, (b) world ocean atlas-derived monthly surface nitrate concentrations, (c) satellite-derived monthly sea ice concentrations, (d) satellite-derived timing of sea ice breakup and formation (with the mean number of annual open water days also designated), (e) satellite-derived monthly chlorophyll-a concentrations, and (f) satellite-derived monthly primary productivity. Color coding for DBO sites designated in (a) is the same for all composite plots.
Fig 6.
Theil-Sen median trends in satellite-derived monthly sea surface temperature (SST) where the hatched pattern indicates a statistically significant Mann-Kendall test for trend (p<0.1) over the 2003–2020 period.
Fig 7.
Theil-Sen median trends in satellite-derived monthly sea ice concentration where the hatched pattern indicates a statistically significant Mann-Kendall test for trend (p<0.1) over the 2003–2020 period.
Fig 8.
Theil-Sen median trends in sea ice events across the Pacific Arctic region: (a) annual sea ice persistence, (b) timing of sea ice breakup, and (c) timing of sea ice formation over the 2003–2020 period. The hatched pattern indicates a statistically significant (p<0.1 using the Mann-Kendall test for trend) over the 2003–2020 period. Basemap datasets from Natural Earth [49] and ESRI [50].
Fig 9.
Theil-Sen median trends in satellite-derived monthly chlorophyll-a concentrations where the hatched pattern indicates a statistically significant (p<0.1 using the Mann-Kendall test for trend) over the 2003–2020 period.
Fig 10.
Thiel-Sen median trends in satellite-derived monthly primary productivity where the hatched pattern indicates a statistically significant (p<0.1 using the Mann-Kendall test for trend) over the 2003–2020 period.
Fig 11.
(a) Mean annual (March-September) primary productivity and (b) Theil-Sen median trends in annual (March-September) primary productivity where the hatched pattern indicates statistically significant trends (p<0.1 using the Mann-Kendall test for trend) over the 2003–2020 period. Basemap datasets from Natural Earth [49] and ESRI [50].
Fig 12.
Thiel-Sen median trends in satellite-derived monthly photosynthetically available radiation where the hatched pattern indicates a statistically significant (p<0.1 using the Mann-Kendall test for trend) over the 2003–2020 period.
Fig 13.
Annual sea ice persistence, timing of sea ice breakup/formation, and annual primary productivity for all eight DBO sites.
Trend lines (Sen median slope) are only shown for those time series (over the years 2003–2020) that show statistically significant trends (p<0.1 using the Mann-Kendall test for trend). Error bar shading indicates ±1SD of the spatial variability of values within each DBO site.
Fig 14.
Summary plots of decadal trends (as indicated by the Theil-Sen median slope) in (a) sea surface temperature, (b) sea ice concentration, (c) chlorophyll-a, (d) primary productivity, (e) annual primary productivity, and (f) sea ice events (timing of sea ice breakup/formation and annual sea ice persistence). Statistical significance (based on a Mann-Kendall test for trend) is indicated by * (p < 0.1), ** (p < 0.05), or *** (p < 0.01).
Fig 15.
Relationships between length of open water season (March-September) and annual primary productivity (March-September) for each of the eight DBO sites.
Linear regressions with 95% confidence intervals (shaded regions) are shown, with the year of each datapoint color coded using the color scale shown in (a).
Fig 16.
Results from the linear regression of annual primary productivity (March-September) and number of open water days (March-September) over the years 2003–2020 across the DBO sites, including both (a) slope and (b) R2. Basemap datasets from Natural Earth [49] and ESRI [50].