Partial validation of a lossy compression approach to computing radiative transfer in cloud system‐resolving models. (9th December 2020)

• Record Type: Journal Article
• Authors: Barker, Howard W.; Qu, Zhipeng; Dhanraj, Varun; Cole, Jason N. S.
• Is Part Of: Quarterly journal of the Royal Meteorological Society. Volume 147:Number 734(2021)
• Page Start: 363
• Page End: 381
• Publication Date: 2020-12-09
• Subjects: 3. physical phenomenon: clouds -- 3. physical phenomenon: radiation -- 4. geophysical sphere: atmosphere; Meteorology -- Periodicals; 551.5
• DOI: 10.1002/qj.3922
• Languages: English
• ISSNs: 0035-9009

Abstract

Abstract: Cloud system‐resolving models (CSRMs) routinely calculate radiative flux profiles via the Independent Column Approximation (ICA). The ICA applies 1D radiative transfer models (RTMs) to all N columns in a CSRM's domain. For this study, the Partitioned Gauss–Legendre Quadrature (PGLQ) method replaced the ICA in a CSRM. The PGLQ applies RTMs to n G = N columns, identified with GLQ rules, and distributes their flux profiles to the other N − n G columns. The PGLQ approach is likened to a lossy compression algorithm that trades information for efficiency. While verification and validation of an audio compression algorithm rest, respectively, on file size reduction and sound quality according to listeners, for the PGLQ they rest on increasing f RT = N / n G while maintaining, according to experimenters, the integrity of CSRM simulations. A CSRM was run for 80 days in radiative‐convective equilibrium (1, 024 × 1, 024 columns and horizontal grid‐spacing of 0.25 km) for sea‐surface temperature SST = 295 and 300 K; the last 40 days were analysed. Simulations using the ICA represent the control ; experiments used the PGLQ calling the RTMs f RT = 200, 5, 000 and 50, 000 fewer times than the control . For f RT = 50, 000, several key variables spanning time/domain‐averaged cloud and radiation properties, a measure of cloud (convection) aggregation, and horizontal fluctuations of cloud and radiation fields, differ significantly from the control . In contrast, corresponding differences between the control and PGLQ with f RT ≤ 5, 000, for both a given SST and differences between SST s, are often minor, in all respects, and appear to be drawn from a single population. While these results partially validate the PGLQ method for f RT ≤ 5, 000, they also indicate that overly large reductions in radiative information are detrimental. Abstract : Schematic diagram that portrays the ICA and Barker and Li's (2019) PGLQ algorithms as file transfer processes. The CSRM provides N profiles of information x N . For the ICA, RT models, denoted as T, act on them and produce N flux profiles F N that are passed back to the CSRM. This is analogous to transferring an original (large) data file. For PGLQ, x N go through an encoder E that sorts and identifies n G = N / f RT = N columns, via GLQ rules, that get acted on by T to produce n G flux profiles F n G . This is analogous to transferring a compressed (small) data file. Last, the n G profiles go through a decoder D that distributes them to all N columns and back to the CSRM.