Tropical Pacific Chlorophyll Algorithm (TPCA): Reprocessing v1.0

The Tropical Pacific Chlorophyll Algorithm (TPCA) regional reprocessing includes Level 3 Mapped Daily 9km Chlorophyll data for the tropical Pacific (10°N to 10°S and 150°E to 90°W). These products are provided for the satellite sensors SeaWiFS (1997 - 2010), MODIS-Aqua (2002 - Present) and MERIS (2002-2012). SeaWiFS and MODIS-Aqua had the TPCA algorithm applied to them, however MERIS performs well as is for the tropical Pacific, and thus a snapshot of the NASA chlor_a product has been provided here.
The tropical Pacific Ocean is a globally significant region of climate-driven biogeochemical variability. Satellite ocean color algorithms have been used for over 20 years and provide a substantial historical record of global ocean chlorophyll-a variability. Current chlorophyll algorithms perform better in the tropical Pacific than for the globe. Nevertheless, improvements can be made to create a robust historical record of chlorophyll variability, which is essential to accurately identify ocean-atmosphere carbon fluxes and long-term trends in ocean productivity. We use a large in situ chlorophyll database to tune empirical ocean color algorithms to remove bias in the equatorial Pacific. Traditional band ratio chlorophyll algorithms (OCx)   perform well, but exhibit errors at low chlorophyll concentrations. A new algorithm, the ocean Color Index (OCI; Hu et al., 2012), is effective at calculating low chlorophyll concentrations in the mesotrophic tropical Pacific Ocean. OCI effectiveness is limited to < 0.2mg m-3, with OCx outperforming it above this limit. Here, we present optimized ocean color algorithms with modified polynomials and updated OCx to OCI blending windows in chlorophyll space. Existing ocean color algorithms underestimate tropical Pacific chlorophyll by 5.8%, 14% and 2%  for SeaWiFS, MODIS-Aqua and MERIS. We develop regionally tuned sensor specific coefficients and blending windows to reduce systematic bias. We assess cross-sensor consistency to produce robust 21-year time series trends. These updated estimates increase chlorophyll concentrations in open water and decrease around island and warm-pool regions, with implications for our understanding of El Niño-Southern Oscillation driven carbon fluxes and net primary productivity.
The TPCA algorithm outlined in Pittman et al., 2019 (2019JC015498) has been applied to the relevant wavelengths for SeaWiFS and MODIS-Aqua. The Level 3 Mapped  Daily ; 9km Rrs files for each sensor were obtained from https://oceandata.sci.gsfc.nasa.gov/. The MERIS Snapshot was provided directly from the NASA chlor_a product. 
These files were then processed with the TPCA algorithm and are provided as yearly combined netcdf4 files.
5 wavelengths for SeaWiFS and 4 wavelengths for MODIS-Aqua were used in the creation of the TPCA reprocessing. 
SeaWiFS Rrs links:
https://oceandata.sci.gsfc.nasa.gov/SeaWiFS/Mapped/Daily/9km/Rrs_443/https://oceandata.sci.gsfc.nasa.gov/SeaWiFS/Mapped/Daily/9km/Rrs_490/
https://oceandata.sci.gsfc.nasa.gov/SeaWiFS/Mapped/Daily/9km/Rrs_510/https://oceandata.sci.gsfc.nasa.gov/SeaWiFS/Mapped/Daily/9km/Rrs_555/
https://oceandata.sci.gsfc.nasa.gov/SeaWiFS/Mapped/Daily/9km/Rrs_670/
MODIS-Aqua Rrs links: 
https://oceandata.sci.gsfc.nasa.gov/MODIS-Aqua/Mapped/Daily/9km/Rrs_443/
https://oceandata.sci.gsfc.nasa.gov/MODIS-Aqua/Mapped/Daily/9km/Rrs_488/
https://oceandata.sci.gsfc.nasa.gov/MODIS-Aqua/Mapped/Daily/9km/Rrs_547/
https://oceandata.sci.gsfc.nasa.gov/MODIS-Aqua/Mapped/Daily/9km/Rrs_667/
MERIS Snapshot was provided directly from the chlor_a files in:
https://oceandata.sci.gsfc.nasa.gov/MERIS/L3SMI/ 
Python 3.6 was used for the analysis and reprocessing for the TPCA. xarray 0.11.3 and netCDF4 were the primary libraries used for this reprocessing.

热带太平洋叶绿素算法（Tropical Pacific Chlorophyll Algorithm, TPCA）区域重处理数据集涵盖热带太平洋（北纬10°至南纬10°、东经150°至西经90°）的三级映射每日9公里分辨率叶绿素数据（Level 3 Mapped Daily 9km Chlorophyll data）。该数据集覆盖SeaWiFS（1997-2010年）、MODIS-Aqua（2002年至今）和MERIS（2002-2012年）三款卫星传感器。其中，SeaWiFS与MODIS-Aqua均已应用TPCA算法进行重处理；而MERIS本身在热带太平洋海域表现优异，因此本数据集直接提供NASA chlor_a产品的快照版本。

热带太平洋是全球受气候驱动的生物地球化学变异性显著区域。卫星海洋色度算法已应用超过20年，为全球海洋叶绿素a浓度变异性提供了丰富的历史观测记录。当前的叶绿素算法在热带太平洋海域的表现优于全球平均水平，但仍可进一步优化，以构建可靠的叶绿素浓度变异性历史序列——这对精准识别海-气碳通量以及海洋生产力长期趋势至关重要。本研究利用大型原位叶绿素数据库对经验海洋色度算法进行校准，以消除赤道太平洋海域的算法偏差。传统波段比叶绿素算法（OCx）整体表现良好，但在低叶绿素浓度区间存在误差。新型算法——海洋颜色指数（Ocean Color Index, OCI; Hu et al., 2012）则可有效计算中营养热带太平洋海域的低浓度叶绿素，其有效适用范围为叶绿素浓度<0.2mg·m⁻³，当浓度高于该阈值时OCx算法表现更优。本研究提出了经过优化的海洋色度算法，包含修正后的多项式项以及叶绿素空间内OCx与OCI的混合窗口更新方案。现有海洋色度算法对热带太平洋叶绿素浓度的估算存在系统性低估：SeaWiFS、MODIS-Aqua和MERIS的低估幅度分别为5.8%、14%和2%。为此，我们开发了针对各传感器的区域校准系数与混合窗口，以降低系统性偏差。同时，我们通过评估跨传感器一致性，构建了可靠的21年时间序列趋势。更新后的估算结果显示，开阔海域的叶绿素浓度有所提升，而岛屿周边及暖池区域的浓度则有所降低，这将有助于我们深入理解厄尔尼诺-南方涛动驱动的碳通量与海洋净初级生产力变化。

本研究针对SeaWiFS与MODIS-Aqua的相关波段，应用了Pittman等人2019年（2019JC015498）提出的TPCA算法。各传感器的三级映射每日9公里分辨率遥感反射率文件（Level 3 Mapped Daily 9km Rrs files）均从https://oceandata.sci.gsfc.nasa.gov/获取。MERIS快照版本直接取自NASA chlor_a产品。

所有原始文件均通过TPCA算法完成重处理，并以年度合并的netCDF4格式文件提供。

TPCA重处理过程中，SeaWiFS传感器使用了5个波段，MODIS-Aqua传感器使用了4个波段。

SeaWiFS Rrs 链接：
https://oceandata.sci.gsfc.nasa.gov/SeaWiFS/Mapped/Daily/9km/Rrs_443/
https://oceandata.sci.gsfc.nasa.gov/SeaWiFS/Mapped/Daily/9km/Rrs_490/
https://oceandata.sci.gsfc.nasa.gov/SeaWiFS/Mapped/Daily/9km/Rrs_510/
https://oceandata.sci.gsfc.nasa.gov/SeaWiFS/Mapped/Daily/9km/Rrs_555/
https://oceandata.sci.gsfc.nasa.gov/SeaWiFS/Mapped/Daily/9km/Rrs_670/

MODIS-Aqua Rrs 链接：
https://oceandata.sci.gsfc.nasa.gov/MODIS-Aqua/Mapped/Daily/9km/Rrs_443/
https://oceandata.sci.gsfc.nasa.gov/MODIS-Aqua/Mapped/Daily/9km/Rrs_488/
https://oceandata.sci.gsfc.nasa.gov/MODIS-Aqua/Mapped/Daily/9km/Rrs_547/
https://oceandata.sci.gsfc.nasa.gov/MODIS-Aqua/Mapped/Daily/9km/Rrs_667/

MERIS快照数据直接取自以下路径的chlor_a文件：
https://oceandata.sci.gsfc.nasa.gov/MERIS/L3SMI/

本研究的TPCA分析与重处理均基于Python 3.6完成，主要使用的第三方库为xarray 0.11.3与netCDF4。