TY - JOUR
T1 - Two-Stage Artificial Intelligence Algorithm for Calculating Moisture-Tracking Atmospheric Motion Vectors
AU - Ouyed, Amir
AU - Zeng, Xubin
AU - Wu, Longtao
AU - Posselt, Derek
AU - Su, Hui
N1 - Funding Information: Acknowledgments. This work is funded by NASA Grant 80NSSC19K0442 in support of ACTIVATE, which is a NASA Earth Venture Suborbital-3 (EVS-3) investigation funded by NASA’s Earth Science Division and managed through the Earth System Science Pathfinder Program Office, and by the Arizona Space Institute. The research was carried out at the Jet Propulsion Laboratory, California Institute of Technology, under a contract with the National Aeronautics and Space Administration (80NM0018D0004). We thank the reviewers for their constructive comments and suggestions for our significant revision of the original manuscript. Funding Information: This work is funded by NASA Grant 80NSSC19K0442 in support of ACTIVATE, which is a NASA Earth Venture Suborbital-3 (EVS-3) investigation funded by NASA’s Earth Science Division and managed through the Earth System Science Pathfinder Program Office, and by the Arizona Space Institute. The research was carried out at the Jet Propulsion Laboratory, California Institute of Technology, under a contract with the National Aeronautics and Space Administration (80NM0018D0004). We thank the reviewers for their constructive comments and suggestions for our significant revision of the original manuscript. Publisher Copyright: © 2021 American Meteorological Society.
PY - 2021/12
Y1 - 2021/12
N2 - Much of the errors of atmospheric motion vectors (AMV) may be a consequence of algorithms not incor-porating dynamical information. A physics-informed, artificial intelligence algorithm was developed that corrects errors of moisture tracking AMV (from the movement of water vapor) using numerical weather prediction (NWP) fields. The University of Arizona (UA) algorithm uses a variational method as a first step (fsUA); the second step then filters the first-stage AMVs using a random forest model that learns the error correction from NWP fields. The UA algorithm is compared with a traditional image feature tracking algorithm (JPL) using a global nature run as the “ground truth.” Experiments use global all-sky humidity fields at 500 and 850 hPa for 1–3 January 2006 and 1–3 July 2006. UA outputs AMVs with root-mean-square vector differences (RMSVDs) of 2 m s-1 for the tropics and ∼2–3 ms-1 for midlatitudes and the poles, whereas JPL outputs much higher RMSVDs of ∼3 ms-1 for the tropics and ∼3–9 ms-1 for the midlatitudes and poles. Although the algorithm fsUA produces approximately the same global RMSVDs as the JPL algorithm, fsUA has a higher resolution since it outputs an AMV per pixel, whereas the JPL algorithm uses a target box that effectively smooths the vec-tors. Furthermore, UA’s RMSVDs are lower than the intrinsic error (calculated from the differences between two reanaly-sis datasets). Even for error-prone regions with low moisture gradients and where winds are oriented along moisture isolines, UA’s absolute speed difference with “truth” stays within ∼3ms-1.
AB - Much of the errors of atmospheric motion vectors (AMV) may be a consequence of algorithms not incor-porating dynamical information. A physics-informed, artificial intelligence algorithm was developed that corrects errors of moisture tracking AMV (from the movement of water vapor) using numerical weather prediction (NWP) fields. The University of Arizona (UA) algorithm uses a variational method as a first step (fsUA); the second step then filters the first-stage AMVs using a random forest model that learns the error correction from NWP fields. The UA algorithm is compared with a traditional image feature tracking algorithm (JPL) using a global nature run as the “ground truth.” Experiments use global all-sky humidity fields at 500 and 850 hPa for 1–3 January 2006 and 1–3 July 2006. UA outputs AMVs with root-mean-square vector differences (RMSVDs) of 2 m s-1 for the tropics and ∼2–3 ms-1 for midlatitudes and the poles, whereas JPL outputs much higher RMSVDs of ∼3 ms-1 for the tropics and ∼3–9 ms-1 for the midlatitudes and poles. Although the algorithm fsUA produces approximately the same global RMSVDs as the JPL algorithm, fsUA has a higher resolution since it outputs an AMV per pixel, whereas the JPL algorithm uses a target box that effectively smooths the vec-tors. Furthermore, UA’s RMSVDs are lower than the intrinsic error (calculated from the differences between two reanaly-sis datasets). Even for error-prone regions with low moisture gradients and where winds are oriented along moisture isolines, UA’s absolute speed difference with “truth” stays within ∼3ms-1.
KW - Artificial intelligence
KW - Cloud tracking/cloud motion winds
KW - Machine learning
UR - http://www.scopus.com/inward/record.url?scp=85135864875&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85135864875&partnerID=8YFLogxK
U2 - 10.1175/JAMC-D-21-0070.1
DO - 10.1175/JAMC-D-21-0070.1
M3 - Article
SN - 1558-8424
VL - 60
SP - 1671
EP - 1684
JO - Journal of Applied Meteorology and Climatology
JF - Journal of Applied Meteorology and Climatology
IS - 12
ER -