Modeling Photosynthetically Available Radiation from Satellite and Field Data and Its Impact on Ocean Primary Production

Abstract

Ocean colour remote sensing is a useful tool and it provides quantitative information of seawater constituents. OCEANSAT-1 OCM and OCEANSAT-2 OCM are two Indian Ocean observation satellites which have been extensively used for various societal and scientific applications like Potential Fishing Zone identification, ocean primary productivity estimation, algal bloom detection and studying the coastal processes etc. Photosynthetically available radiation (PAR) with spatially invariant and uniform aerosol optical depth under clear sky condition has been used in current primary production models using OCM data. Thus, modeling PAR from OCM under variable aerosol loading and cloud coverage is a desirable input parameter to estimate ocean primary production from OCM more accurately.

PAR has been estimated from OCM using two different methods for spatially variable aerosols and cloudy condition. In method I, the atmosphere is treated as a single layer in clear sky condition or as double layer in cloudy conditions i.e., a layer above the cloud top and a layer from the cloud top downwards. The two main input parameters to estimate PAR i.e., aerosol optical depth (AOD) and cloud optical depth (COD) have been estimated from OCM data. Surface reflectance has been neglected in the first method. In method II, the effects of clouds and clear atmosphere can be decoupled with cloud system and surface albedo. In the second method, surface and cloud albedo have been estimated from OCM.

PAR estimated from OCM has been compared with PAR measured from surface downwelling flux measurement instrument during different seasons in the Arabian Sea. Other than in-situ data obtained from ship cruise, OCEANSAT-2 OCM estimated PAR have been validated with in-situ measured PAR obtained from surface irradiance data of a buoy located at the CAL-VAL site situated between Kavaratti and Agatti island in the Arabian Sea. The root mean square (r.m.s) difference between OCEANSAT-1 OCM PAR estimated using both the methods compared to in-situ measured PAR were within the accepted range (±10%) for PAR estimation from remote sensing. However, root mean square difference between PAR estimated from OCEANSAT-1 OCM and in-situ measured PAR was lower for method I compared to method II. OCEANSAT-1 & 2 OCM estimated PAR also has been compared with PAR estimated from other ocean colour sensors such as SeaWiFS (Sea-viewing Wide Field-of view Sensor) and MODIS (Moderate Resolution Imaging Spetroradiometer). OCEANSAT-1 OCM PAR estimated from method I had good correlation with correlation coefficient (r2) 0.95 with SeaWiFS PAR under clear sky condition. The correlation coefficient (r2) between OCEANSAT-1 OCM PAR estimated using method II with SeaWiFS PAR was 0.80. OCEANSAT-2 OCM estimated PAR using method I shows poor correlation (correlation coefficient 0.63) compared to OCEANSAT-1 OCM with in-situ measured PAR at CAL-VAL site situated between Kavaratti and Agatti island. The probable reason behind to get poor correlation between OCEANSAT-2 OCM PAR and in-situ measured PAR has been explained to the dynamic variability of clouds over Kavaratti region. However, OCEANSAT-2 OCM estimated PAR have good correlation (correlation coefficient r2 0.84) with MODIS Aqua estimated PAR at Kavaratti region.

In order to understand the variability of PAR under different aerosol types and different aerosol loading, a simulation study has been carried out using COART (Coupled Ocean Atmosphere Radiative Transfer) model. Euphotic primary production has been estimated using an analytical non-spectral model for various PAR values under different aerosol loading and cloud coverage conditions. Sensitivity analysis showed that for maritime, maritime polluted, and desert aerosols, PAR has attenuated to about 11–25%, whereas it has attenuated to 44% for urban aerosol type compared to clear sky. PAR has been reduced by ~57% for high aerosol loading and for overcast sky. The decrease in euphotic primary production under various aerosol loading and cloud coverage was observed to depend on the photoadaptation parameters. Euphotic primary production was reduced by 38% for maximum maritime aerosol loading and for overcast sky compared to clear sky.

Thus, PAR estimated under various aerosols loading and cloud coverage has shown improvement in the quantification of ocean euphotic primary production from OCM. Validation results between OCM PAR and in-situ measured PAR and comparative study between OCM PAR and SeaWiFS, MODIS estimated PAR shows that PAR from OCEANSAT-1 & 2 OCM could be produced reasonably accurately over the tropical Indian vii

Ocean region. This approach can be extended to future OCEANSAT-3 OCM data for operational estimation of PAR for regional marine ecosystem applications.