3  Data Products and Format

3.1 Overview

FastMAPOL (Fast Multi-Angle Polarimetric Ocean and Land algorithm) produces Level-2 (L2) geophysical retrievals of aerosol and surface properties from the multi-angle polarimetric measurements of the PACE polarimeters, including HARP2 and SPEXone.

The Level-2 product provides pixel-level estimates of aerosol optical properties, aerosol microphysical parameters, and ocean color quantities derived through atmospheric correction. In addition to the primary geophysical retrievals, the product includes diagnostic quantities describing retrieval performance and quality indicators for downstream processing.

The Level-2 product is distributed in NetCDF format and organized into hierarchical data groups to separate geolocation, geophysical variables, diagnostics, and sensor parameters.

The FastMAPOL Level-2 product is designed to support:

  • aerosol science applications
  • ocean color studies
  • land surface studies
  • atmospheric correction
  • algorithm validation
  • Level-3 gridded product generation

3.2 File Organization

The FastMAPOL Level-2 product is organized into several top-level groups. Each group contains variables associated with a specific component of the retrieval system.

Group Description
geolocation_data Pixel geolocation information
geophysical_data Primary aerosol, ocean, and land retrieval products
diagnostic_data Retrieval diagnostics and performance metrics
sensor_band_parameters Instrument spectral and angular parameters
processing_control Processing metadata and algorithm configuration

3.2.1 Dimensions

The FastMAPOL Level-2 product uses a set of core dimensions describing the spatial, spectral, and angular sampling of the observations and retrieval products.

Dimension Description
number_of_lines Number of along-track scan lines
pixels_per_line Number of cross-track pixels
wavelength Spectral wavelengths used in the retrieval products
number_of_views Number of viewing angles
intensity_bands_per_view Number of intensity bands per viewing angle
polarization_bands_per_view Number of polarization bands per viewing angle

3.2.2 Instrument Differences

HARP2

  • ~90 viewing angles
  • 4 spectral bands : 440, 550, 670, 870nm
  • wide swath (~2000km)

SPEXone

  • 5 viewing angles
  • hyperspectral polarimetric measurements 50
  • narrow swath (~100km)

A total of 34 specific wavelengths from the hyperspectral measurements for aerosol and surface retrieval.

3.3 Geolocation Data

Group: /geolocation_data

Variable Dimensions Description
latitude (number_of_lines, pixels_per_line) Pixel latitude
longitude (number_of_lines, pixels_per_line) Pixel longitude

These variables define the geographic location of each retrieval pixel.

3.4 Geophysical Retrieval Products

Group: /geophysical_data

The geophysical group contains aerosol optical properties, aerosol microphysical parameters, and ocean color products retrieved by FastMAPOL. Two data suites, MAPOL_OCEAN and MAPOL_LAND, are available for aerosol over ocean and land, respectively.

3.4.1 Aerosol Optical Properties

Product suites: MAPOL_OCEAN and MAPOL_LAND

Variable Dimensions Description
aot (number_of_lines, pixels_per_line, wavelength) Aerosol optical depth
aot_fine (number_of_lines, pixels_per_line, wavelength) Fine-mode aerosol optical depth
aot_coarse (number_of_lines, pixels_per_line, wavelength) Coarse-mode aerosol optical depth
fmf (number_of_lines, pixels_per_line) Fine-mode aerosol optical depth fraction
angstrom_440_870 (number_of_lines, pixels_per_line) Ångström exponent (440–870 nm)
angstrom_440_670 (number_of_lines, pixels_per_line) Ångström exponent (440–670 nm)
ssa (number_of_lines, pixels_per_line, wavelength) Aerosol single-scattering albedo
ssa_fine (number_of_lines, pixels_per_line, wavelength) Fine-mode single-scattering albedo
ssa_coarse (number_of_lines, pixels_per_line, wavelength) Coarse-mode single-scattering albedo
aerosol_lidar_ratio (number_of_lines, pixels_per_line, wavelength) Aerosol lidar ratio
aerosol_depol_ratio (number_of_lines, pixels_per_line, wavelength) Aerosol depolarization ratio

These parameters (aot,angstrom,fmf) describe aerosol optical loading and spectral behavior. Note that Ångström exponent (440-870 nm) is only available for HARP2 data since SPEXone do not have 870nm band. Single scattering albedo (ssa) describe aerosol absorption property.

Two new variables are added to V4 data including aerosol_lidar_ratio and aerosol_depol_ratio, for the comparison with lidar products.

3.4.2 Aerosol Microphysical Properties

Product suites: MAPOL_OCEAN and MAPOL_LAND

Variable Dimensions Description
reff_fine (number_of_lines, pixels_per_line) Fine-mode effective radius
reff_coarse (number_of_lines, pixels_per_line) Coarse-mode effective radius
veff_fine (number_of_lines, pixels_per_line) Fine-mode effective variance
veff_coarse (number_of_lines, pixels_per_line) Coarse-mode effective variance
mr_fine (number_of_lines, pixels_per_line, wavelength) Fine-mode real refractive index
mr_coarse (number_of_lines, pixels_per_line, wavelength) Coarse-mode real refractive index
mr (number_of_lines, pixels_per_line, wavelength) Mode-averaged real refractive index
mi_fine (number_of_lines, pixels_per_line, wavelength) Fine-mode imaginary refractive index
mi_coarse (number_of_lines, pixels_per_line, wavelength) Coarse-mode imaginary refractive index
mi (number_of_lines, pixels_per_line, wavelength) Mode-averaged imaginary refractive index
sph_fine (number_of_lines, pixels_per_line) Fine-mode spherical particle fraction
sph_coarse (number_of_lines, pixels_per_line) Coarse-mode spherical particle fraction
sph (number_of_lines, pixels_per_line) Volume-averaged spherical particle fraction
vd_fine (number_of_lines, pixels_per_line) Fine-mode aerosol volume density
vd_coarse (number_of_lines, pixels_per_line) Coarse-mode aerosol volume density
fvf (number_of_lines, pixels_per_line) Fine-mode volume fraction
alh (number_of_lines, pixels_per_line) Aerosol layer height

These variables constrain aerosol particle size distribution, composition, and shape.

3.4.3 Ocean Products

Product suites: MAPOL_OCEAN

Variable Dimensions Description
Rrs_angular (number_of_lines, pixels_per_line, number_of_views, intensity_bands_per_view) Angular remote sensing reflectance before bidirectional reflectance correction
Rrs_nadir (number_of_lines, pixels_per_line, number_of_views, intensity_bands_per_view) Nadir-adjusted remote sensing reflectance after bidirectional reflectance correction
Rrs_angular_mean (number_of_lines, pixels_per_line, wavelength) Mean of Rrs_angular over viewing angles
Rrs_angular_std (number_of_lines, pixels_per_line, wavelength) Standard deviation of Rrs_angular over viewing angles
Rrs_nadir_mean (number_of_lines, pixels_per_line, wavelength) Mean of Rrs_nadir over viewing angles
Rrs_nadir_std (number_of_lines, pixels_per_line, wavelength) Standard deviation of Rrs_nadir over viewing angles
wind_speed (number_of_lines, pixels_per_line) Retrieved ocean surface wind speed
chla (number_of_lines, pixels_per_line) Retrieved chlorophyll-a concentration

The dimensions of the angular reflectance products depend on the PACE polarimeter. Example dimensions are:

  • SPEXone: number_of_lines = 395, pixels_per_line = 29, number_of_views = 5, and intensity_bands_per_view = 34.
  • HARP2: number_of_lines = 395, pixels_per_line = 519, number_of_views = 90, and intensity_bands_per_view = 1.

The remote sensing reflectance represents ocean-leaving reflectance after atmospheric correction. To compare with AERONET OC or other dataset for validation purpose after BRDF correction, variable Rrs_nadir_mean is suggested to use.

Note that V3 data contains Rrs1 and Rrs2, which are renmaed as Rrs_angular and Rrs_nadir in current V4 data.

3.4.4 Land surface product

Product suites: MAPOL_LAND

Variable Dimensions Description
rhos_angular (number_of_lines, pixels_per_line, number_of_views, intensity_bands_per_view) Angular land surface reflectance before bidirectional reflectance correction
rhos_nadir (number_of_lines, pixels_per_line, number_of_views, intensity_bands_per_view) Nadir-adjusted land surface reflectance after bidirectional reflectance correction
rhos_angular_mean (number_of_lines, pixels_per_line, wavelength) Mean of rhos_angular over viewing angles
rhos_angular_std (number_of_lines, pixels_per_line, wavelength) Standard deviation of rhos_angular over viewing angles
rhos_nadir_mean (number_of_lines, pixels_per_line, wavelength) Mean of rhos_nadir over viewing angles
rhos_nadir_std (number_of_lines, pixels_per_line, wavelength) Standard deviation of rhos_nadir over viewing angles

The land surface reflectance represents reflectance after atmospheric correction. To compare with validation dataset after BRDF correction, variable rhos_nadir_mean is suggested to use.

3.5 Diagnostic Data

Product suites: MAPOL_OCEAN and MAPOL_LAND

Group: /diagnostic_data

These variables describe retrieval convergence and performance.

Variable Dimensions Description
chi2 (number_of_lines, pixels_per_line) Retrieval cost function
chi2_first_guess (number_of_lines, pixels_per_line) Cost function at the initial state
nv_ref (number_of_lines, pixels_per_line) Number of reflectance measurements used in the retrieval
nv_dolp (number_of_lines, pixels_per_line) Number of DoLP measurements used in the retrieval
mask_ref (number_of_lines, pixels_per_line, number_of_views, intensity_bands_per_view) Reflectance measurement screening mask
mask_dolp (number_of_lines, pixels_per_line, number_of_views, polarization_bands_per_view) DoLP measurement screening mask
quality_flag (number_of_lines, pixels_per_line) Retrieval quality indicator
nfev (number_of_lines, pixels_per_line) Number of forward-model evaluations
njev (number_of_lines, pixels_per_line) Number of Jacobian evaluations
timing (number_of_lines, pixels_per_line) Retrieval runtime
ozone (number_of_lines, pixels_per_line) Total column ozone
surface_pressure (number_of_lines, pixels_per_line) Surface pressure

Further details on chi2, nv_ref, nv_dolp, mask_ref, mask_dolp, and quality_flag are provided below.

The remaining variables are primarily included for diagnostic and debugging purposes:

  • nfev, njev, and timing are used to track the inversion process and its computational performance.

  • ozone and surface_pressure are typically obtained from ancillary data but may be updated within the retrieval algorithm to provide values at higher spatial resolution.

3.5.1 Retrieval Quality Metrics Based on the Cost Function

FastMAPOL retrievals are performed using an optimal estimation framework that minimizes a cost function chi2. The variable chi2 represents the final value of the cost function after the inversion has converged. chi2_first_guess represents the \(\chi^2\) value obtained from the first-guess retrieval using the smaller neural network in the initial retrieval stage (see Appendix G).

Usually lower values of \(\chi^2\) indicate better agreement between the measurements and forward model simulations.

3.5.2 Retrieval Quality Metrics Based on Fitted Measurements

FastMAPOL employs adaptive data screening to remove measurements at specific viewing angles, spectral bands, or their combinations that cannot be adequately fitted by the forward model (see Appendix F). Such measurements are often associated with clouds or other anomalies.

The total numbers of reflectance and DoLP measurements retained in the retrieval are denoted by nv_ref and nv_dolp, respectively. These variables also provide useful indicators of the retrieval quality for each pixel. If a large fraction of the measurements is removed, less information is available to constrain the retrieved parameters. Therefore, larger values of nv_ref and nv_dolp generally indicate that more measurement information is available for the retrieval.

The measurement masks for individual viewing angles and spectral bands are also provided in the diagnostic data group as mask_ref and mask_dolp. These masks allow users to identify which measurements are retained or excluded at each viewing angle and wavelength. The total numbers of valid measurements indicated by mask_ref and mask_dolp are consistent with the corresponding values of nv_ref and nv_dolp, respectively.

3.5.3 Retrieval Quality Flag

A quality_flag is provided to facilitate data filtering and Level-3 processing. The flag is determined from a combination of chi2, nv_ref, and nv_dolp. In general, a value of 0 indicates a smaller chi2, corresponding to better agreement between the measurements and forward model simulations, together with larger nv_ref and nv_dolp, indicating that a larger fraction of the measurements is retained in the retrieval.

Larger quality_flag values generally indicate a larger chi2 and/or smaller numbers of retained measurements, and therefore reduced retrieval quality. This may result from reduced information content or potential contamination by effects such as clouds, land, or other adjacency effects.

quality_flag Description
0 Highest retrieval quality
1 Good retrieval quality
>1 Reduced retrieval quality

3.6 Sensor Band Parameters

Product suites: MAPOL_OCEAN and MAPOL_LAND

Group: /sensor_band_parameters

Variable Dimensions Description
wavelength (wavelength) Retrieval wavelengths
sensor_view_angle (number_of_views) Sensor viewing angles
intensity_wavelength (number_of_views, intensity_bands_per_view) Intensity field center wavelengths at each view