Model Evaluation and Intercomparison of Marine Warm Low Cloud Fractions With Neural Network Ensembles. (23rd November 2021)

• Record Type: Journal Article
• Authors: Chen, Yao‐Sheng; Yamaguchi, Takanobu; Bogenschutz, Peter A.; Feingold, Graham
• Is Part Of: Journal of advances in modeling earth systems. Volume 13:Number 11(2021)
• Page Start: n/a
• Page End: n/a
• Publication Date: 2021-11-23
• Subjects: cloud controlling factors -- E3SM -- high resolution modeling -- machine learning -- shallow clouds; Geological modeling -- Periodicals; Climatology -- Periodicals; Geochemical modeling -- Periodicals; 551.5011
• DOI: 10.1029/2021MS002625
• Languages: English
• ISSNs: 1942-2466

Abstract

Abstract: Low cloud fractions (LCFs) and meteorological factors (MFs) over an oceanic region containing multiple cloud regimes are examined for three data sets: one Energy Exascale Earth System Model (E3SM) simulation with the default 72‐layer vertical grid (E3SM72), another one with 8‐times vertical resolution via the Framework for Improvement by Vertical Enhancement (E3SM × 8), and one with MFs from ERA5 reanalysis and LCFs from the CERES SSF product (ERA5‐SSF). Neural networks (NNs) are trained to capture the relationship between MFs and LCF and to select the best‐performing MF subsets for predicting LCF. NN ensembles are used to (a) confirm the performance of selected MF subsets, (b) to serve as proxy models for each data set to predict LCFs for MFs from all data sets, and (c) to classify MFs into those in shared and uniquely occupied MF subspaces. Overall, E3SM72 and E3SM × 8 have large fractions of MFs in shared MF subspace, but less so near the Californian and Peruvian stratocumulus decks. E3SM × 8 and ERA5 have small fractions of MFs in shared MF subspace but greater than E3SM72 and ERA5, especially in the Southeast Pacific. The differences in LCFs between three pairs of data sets are decomposed into those associated with the differences in the LCF‐MF relationship and those involving different MFs. Given the same MFs, LCFs produced by E3SM × 8 are greater than those produced by E3SM72 but are still different from those in ERA5‐SSF. In general, the shift in MFs dominates the difference in the LCFs. Plain Language Summary: Marine warm low clouds are critical for both present day and future climate because they reflect a lot of solar energy back to space. To make more reliable projections of our changing climate, scientists need to improve these clouds in climate models. One question that scientists ask is, why do the climate models predict so much less marine warm low cloud cover than the satellites see? Is it because the models misrepresent the meteorology (like temperature and humidity) or because they are not able to produce enough clouds even if they predict the meteorology well. In this work, we use neural networks, a machine learning technique, to answer these questions. We find that our recent efforts to improve a climate model help the model produce more marine warm low clouds given the same meteorology; these efforts also lead to changes in the meteorology predicted by the model. Further model improvements are needed to bring the model predictions closer to the observations. Key Points: Neural network ensembles are able to detect how different meteorological factors are between E3SM, E3SM‐FIVE, and observations E3SM‐FIVE high vertical resolution simulation improves marine warm low clouds fractions given the same meteorological factors Meteorological factors between the three data sets are large enough to dominate the differences in marine warm low cloud fractions