Comparing remotely sensed observations of clouds and aerosols in the Southern Ocean with climate model simulations
Southern Ocean (SO) shortwave (SW) radiation biases are a common problem in contemporary general circulation models (GCMs), with most models exhibiting a tendency to absorb too much incoming SW radiation. These biases have been attributed to deficiencies in the representation of clouds during the au...
Main Author: | |
---|---|
Format: | Thesis |
Language: | English |
Published: |
Zenodo
2020
|
Subjects: | |
Online Access: | https://dx.doi.org/10.5281/zenodo.4281575 https://zenodo.org/record/4281575 |
id |
ftdatacite:10.5281/zenodo.4281575 |
---|---|
record_format |
openpolar |
institution |
Open Polar |
collection |
DataCite Metadata Store (German National Library of Science and Technology) |
op_collection_id |
ftdatacite |
language |
English |
description |
Southern Ocean (SO) shortwave (SW) radiation biases are a common problem in contemporary general circulation models (GCMs), with most models exhibiting a tendency to absorb too much incoming SW radiation. These biases have been attributed to deficiencies in the representation of clouds during the austral summer months, either due to cloud cover or cloud albedo being too low. They affect simulation of New Zealand (NZ) and global climate in GCMs due to excessive heating of the sea surface and the effect on large-scale circulation. Therefore, improvement of GCMs is necessary for accurate prediction of future NZ and global climate. We performed ship-based lidar, radar, radiosonde and weather observations on two SO voyages and processed data from multiple past SO voyages. We used the observations and satellite measurements for evaluation of the Hadley Centre Global Environmental Model version 3 (HadGEM3) and contrasting with the Modern-Era Retrospective analysis for Research and Applications version 2 (MERRA-2) to better understand the source of the problem. Due to the nature of lidar observations (the laser signal is quickly attenuated by clouds) they cannot be used for 1:1 comparison with a model without using a lidar simulator, which performs atmospheric radiative transfer calculations of the laser signal. We modify an existing satellite lidar simulator present in the Cloud Feedback Model Intercomparison Project (CFMIP) Observational Simulator Package (COSP) for use with the ground-based lidars used in our observations by modifying the geometry of the radiative transfer calculations, Mie and Rayleigh scattering of the laser signal. We document and make the modified lidar simulator available to the scientific community as part of a newly-developed lidar processing tool called the Automatic Lidar and Ceilometer Framework (ALCF), which enables unbiased comparison between lidar observations and models by performing calibration of lidar backscatter, noise removal and consistent cloud detection. We apply the lidar simulator on HadGEM3 model fields. Significant SW radiation errors in the SO of up to 21 Wm⁻² are shown to be present in the model. Using the lidar observations, we find that the model underestimates overall cloud cover by about 9% and strongly underestimates boundary layer low-level stratocumulus (Sc) cloud below 1 km and fog. By using radiosonde observations, we find that the observed cloud was strongly linked to the boundary layer stability and sea surface temperature, while this relationship is weaker in the model. We identify that these errors are not due to misrepresentation of large-scale circulation, which is prescribed in our model based on global satellite observations by nudging. We conclude that the problem is likely in the subgrid-scale parametrisation schemes of the boundary layer, convection and large-scale could. In order to address the deficiencies identified we perform experimental simulations of HadGEM3 with modifications of the parametrisation schemes. We find that a three-layer cloud profiles were common in the Ross Sea region, consisting of cumulus (Cu) below Sc, and corresponding to local thermodynamic levels: lifting condensation level, dry and moist neutral buoyancy levels of parcels lifted from the surface. We find that not enough moisture is transported to the top of the boundary layer to form Sc clouds. By increasing surface moisture flux and convective mass flux in the model we improve the Sc cloud simulation, but we show that a lack of vertical moisture transport across the lifting condensation level from the surface layer to the zone of convective mass flux is a likely limiting factor. We show that the modifications had a positive impact on the Southern Ocean and global radiation balance of up to 5 Wm⁻² in zonal average over this limited time period. We suggest that further research should focus on the weak vertical coupling between the boundary layer turbulence and boundary layer convection parametrisation in the model, and that the lidar simulator framework is used as a cloud evaluation tool in further studies due to its benefits over more statistical approaches, which tend to hide compensating biases. |
format |
Thesis |
author |
Kuma, Peter |
spellingShingle |
Kuma, Peter Comparing remotely sensed observations of clouds and aerosols in the Southern Ocean with climate model simulations |
author_facet |
Kuma, Peter |
author_sort |
Kuma, Peter |
title |
Comparing remotely sensed observations of clouds and aerosols in the Southern Ocean with climate model simulations |
title_short |
Comparing remotely sensed observations of clouds and aerosols in the Southern Ocean with climate model simulations |
title_full |
Comparing remotely sensed observations of clouds and aerosols in the Southern Ocean with climate model simulations |
title_fullStr |
Comparing remotely sensed observations of clouds and aerosols in the Southern Ocean with climate model simulations |
title_full_unstemmed |
Comparing remotely sensed observations of clouds and aerosols in the Southern Ocean with climate model simulations |
title_sort |
comparing remotely sensed observations of clouds and aerosols in the southern ocean with climate model simulations |
publisher |
Zenodo |
publishDate |
2020 |
url |
https://dx.doi.org/10.5281/zenodo.4281575 https://zenodo.org/record/4281575 |
long_lat |
ENVELOPE(12.615,12.615,65.816,65.816) |
geographic |
Southern Ocean Austral Ross Sea New Zealand Merra |
geographic_facet |
Southern Ocean Austral Ross Sea New Zealand Merra |
genre |
Ross Sea Southern Ocean |
genre_facet |
Ross Sea Southern Ocean |
op_relation |
https://dx.doi.org/10.5281/zenodo.3865849 |
op_rights |
Open Access Creative Commons Attribution 4.0 International https://creativecommons.org/licenses/by/4.0/legalcode cc-by-4.0 info:eu-repo/semantics/openAccess |
op_rightsnorm |
CC-BY |
op_doi |
https://doi.org/10.5281/zenodo.4281575 https://doi.org/10.5281/zenodo.3865849 |
_version_ |
1766179789760102400 |
spelling |
ftdatacite:10.5281/zenodo.4281575 2023-05-15T18:07:35+02:00 Comparing remotely sensed observations of clouds and aerosols in the Southern Ocean with climate model simulations Kuma, Peter 2020 https://dx.doi.org/10.5281/zenodo.4281575 https://zenodo.org/record/4281575 en eng Zenodo https://dx.doi.org/10.5281/zenodo.3865849 Open Access Creative Commons Attribution 4.0 International https://creativecommons.org/licenses/by/4.0/legalcode cc-by-4.0 info:eu-repo/semantics/openAccess CC-BY Text Thesis article-journal ScholarlyArticle 2020 ftdatacite https://doi.org/10.5281/zenodo.4281575 https://doi.org/10.5281/zenodo.3865849 2021-11-05T12:55:41Z Southern Ocean (SO) shortwave (SW) radiation biases are a common problem in contemporary general circulation models (GCMs), with most models exhibiting a tendency to absorb too much incoming SW radiation. These biases have been attributed to deficiencies in the representation of clouds during the austral summer months, either due to cloud cover or cloud albedo being too low. They affect simulation of New Zealand (NZ) and global climate in GCMs due to excessive heating of the sea surface and the effect on large-scale circulation. Therefore, improvement of GCMs is necessary for accurate prediction of future NZ and global climate. We performed ship-based lidar, radar, radiosonde and weather observations on two SO voyages and processed data from multiple past SO voyages. We used the observations and satellite measurements for evaluation of the Hadley Centre Global Environmental Model version 3 (HadGEM3) and contrasting with the Modern-Era Retrospective analysis for Research and Applications version 2 (MERRA-2) to better understand the source of the problem. Due to the nature of lidar observations (the laser signal is quickly attenuated by clouds) they cannot be used for 1:1 comparison with a model without using a lidar simulator, which performs atmospheric radiative transfer calculations of the laser signal. We modify an existing satellite lidar simulator present in the Cloud Feedback Model Intercomparison Project (CFMIP) Observational Simulator Package (COSP) for use with the ground-based lidars used in our observations by modifying the geometry of the radiative transfer calculations, Mie and Rayleigh scattering of the laser signal. We document and make the modified lidar simulator available to the scientific community as part of a newly-developed lidar processing tool called the Automatic Lidar and Ceilometer Framework (ALCF), which enables unbiased comparison between lidar observations and models by performing calibration of lidar backscatter, noise removal and consistent cloud detection. We apply the lidar simulator on HadGEM3 model fields. Significant SW radiation errors in the SO of up to 21 Wm⁻² are shown to be present in the model. Using the lidar observations, we find that the model underestimates overall cloud cover by about 9% and strongly underestimates boundary layer low-level stratocumulus (Sc) cloud below 1 km and fog. By using radiosonde observations, we find that the observed cloud was strongly linked to the boundary layer stability and sea surface temperature, while this relationship is weaker in the model. We identify that these errors are not due to misrepresentation of large-scale circulation, which is prescribed in our model based on global satellite observations by nudging. We conclude that the problem is likely in the subgrid-scale parametrisation schemes of the boundary layer, convection and large-scale could. In order to address the deficiencies identified we perform experimental simulations of HadGEM3 with modifications of the parametrisation schemes. We find that a three-layer cloud profiles were common in the Ross Sea region, consisting of cumulus (Cu) below Sc, and corresponding to local thermodynamic levels: lifting condensation level, dry and moist neutral buoyancy levels of parcels lifted from the surface. We find that not enough moisture is transported to the top of the boundary layer to form Sc clouds. By increasing surface moisture flux and convective mass flux in the model we improve the Sc cloud simulation, but we show that a lack of vertical moisture transport across the lifting condensation level from the surface layer to the zone of convective mass flux is a likely limiting factor. We show that the modifications had a positive impact on the Southern Ocean and global radiation balance of up to 5 Wm⁻² in zonal average over this limited time period. We suggest that further research should focus on the weak vertical coupling between the boundary layer turbulence and boundary layer convection parametrisation in the model, and that the lidar simulator framework is used as a cloud evaluation tool in further studies due to its benefits over more statistical approaches, which tend to hide compensating biases. Thesis Ross Sea Southern Ocean DataCite Metadata Store (German National Library of Science and Technology) Southern Ocean Austral Ross Sea New Zealand Merra ENVELOPE(12.615,12.615,65.816,65.816) |