The utility of simulated ocean chlorophyll observations: a case study with the Chlorophyll Observation Simulator Package (version 1) in CESMv2.2

• Main Authors: Clow, Genevieve, Lovenduski, Nicole, Levy, Michael, Lindsay, Keith, Kay, Jennifer
• Format: Text
• Language: English
• Published: 2023
• Subjects: Sea ice
• Online Access: https://doi.org/10.5194/gmd-2023-143; https://gmd.copernicus.org/preprints/gmd-2023-143/

For several decades, a suite of satellite sensors has enabled us to study the global spatiotemporal distribution of phytoplankton through remote sensing of chlorophyll. However, the satellite record has extensive missing data, partially due to cloud cover; regions characterized by the highest phytoplankton abundance are also some of the cloudiest. To quantify potential sampling biases due to missing data, we developed a satellite simulator for ocean chlorophyll in the Community Earth System Model (CESM) that mimics what a satellite would detect if it were present in the model-generated world. Our Chlorophyll Observation Simulator Package (ChlOSP) generates synthetic chlorophyll observations at model runtime. ChlOSP accounts for missing data – due to low light, sea ice, and cloud cover – and it can implement swath sampling. Results from a 50-year pre-industrial control simulation of CESM-ChlOSP suggest that missing data impacts the apparent mean state and variability of chlorophyll. The simulated observations exhibit a nearly -20 % difference in global mean chlorophyll compared with the standard model output, which is the same order of magnitude as the projected change in chlorophyll by the end of the century. Additionally, missing data impacts the apparent seasonal cycle of chlorophyll in subpolar regions. We highlight four potential future applications of ChlOSP: (1) refined model tuning, (2) evaluating chlorophyll-based NPP algorithms, (3) revised time to emergence of anthropogenic chlorophyll trends, and (4) a testbed for the assessment of gap-filling approaches for missing satellite chlorophyll data.