Enhancing Retrievals of Air–Sea Heat Fluxes from AMSR2 Microwave Observations Based on Deep Learning

<p><span>The primary generated products are </span><span>2m</span><span> air temperature (Ta)</span><span> and</span><span> specific humidity (Qa), as well as the resulting fluxes of sensible heat (SHF)</span><span> and </span><span>latent heat (LHF). The data records for Ta</span><span> </span><span>and Qa are generated from </span><span>sea surface temperature (T</span><span>s</span><span>), column water vapour</span><span> (WV),</span><span> </span><span>wind speed (WS)</span><span>, column cloud liquid water</span><span> (CLW) </span><span>and rain rate</span><span> (RR)</span><span> data from JAXA&rsquo;s Advanced Microwave Scanning Radiometer 2 (AMSR2) onboard the Global Change Observation Mission 1st-Water (GCOM-W1)</span><span> </span><span>spacecraft</span><span>.</span><span> The </span><span>model</span><span> for determining Ta, Qa, and U10 is the Matrices-Points Fusion Network (MPFNet). The resulting SHF</span><span> and</span><span> LHF fluxes are calculated from these fields of Ta, Qa, </span><span>WS</span><span>, and </span><span>Ts</span><span> using the Coupled Ocean-Atmosphere Response Experiment (COARE) 3.</span><span>6</span><span> flux algorithm</span><span>. </span><span>The </span><span>daily </span><span>MPFNet dataset is stored in </span><span>&ldquo;nc&rdquo;</span><span> format on 0.25&deg;</span><span> </span><span><span>&nbsp;</span>0.25&deg; gridded maps and span from July 2012 to December 2023.</span></p>