In five subtropical gyres, 1998–2025 trends of low-chlorophyll area change sign between seven satellite products, and the 1998–2006 expansion did not continue at its early rate

Question: is the reported expansion of the least productive ocean water (surface chlorophyll at or below 0.04, 0.07 or 0.10 mg m-3) in the five subtropical gyres the same in independent satellite records? Method: we calculated monthly gyre areas below each threshold (5-45 degrees N/S, 1998-2025) from seven products: ESA OC-CCI v6.0, v5.0, v4.2 and v3.1, NASA SeaWiFS R2018, NASA MODIS-Aqua R2018 and NOAA VIIRS SNPP. We compared products only on the cells and months that both products observed. We used block-bootstrap intervals. Results: the ratio of OC-CCI v6.0 area to single-sensor area changed with time in all three sensor overlap windows. At CHL <= 0.07 for all gyres, the drift was +0.51 %/yr against SeaWiFS (1998-2010), +0.52 %/yr against MODIS-Aqua (2003-2021) and -1.47 %/yr against VIIRS (2013-2025). All three intervals exclude zero. At CHL <= 0.04, the drift was +1.30 and -5.67 %/yr. In two windows, the trend sign was opposite between products. The OC-CCI version also changed the result: the 1998-2006 all-gyre trend at CHL <= 0.07 was +0.89 %/yr in v6.0 and +1.01 %/yr in v5.0, but +0.07 to +0.11 %/yr in v3.1, v4.2 and SeaWiFS R2018. No product reproduced the reported 4.3 %/yr North Atlantic rate for 1998-2006. In OC-CCI v6.0, the 1998-2006 trend line predicted a 2016-2025 area 18 % above the 1998-2006 mean; the observed area was 6 % above it, and the flat mean predicted 2007-2025 better. With an annual-mean metric, OC-CCI v6.0 gave about +15 %/yr for CHL <= 0.04 area in the North Atlantic and North Pacific in 1998-2022, near published values. But in SeaWiFS, MODIS-Aqua and OC-CCI v3.1 and v4.2, this North Atlantic area was zero in most years, and the rates disagreed between products in matched windows. Conclusion: these satellite records do not identify the sign of the low-chlorophyll area trend since 1998 at the precision of the published rates. Limitations: we did not determine which product is correct. In-situ station data were not used. The gyre boxes and the 1/3 degree point sampling differ from earlier studies, but they are identical for all products.

ocean colour · chlorophyll-a · subtropical gyres · oligotrophic expansion · ocean deserts · ESA OC-CCI · SeaWiFS · MODIS-Aqua · VIIRS · sensor drift · trend robustness · out-of-sample test · phytoplankton · climate change detection