Objectives for the cloud parameter validation project
There are three main purposes motivating this project:
1. To provide assistance to the SWECLIM/SMHI Rossby centre for regional climate simulations in the use of satellite-derived cloud parameters for validation of cloud forcing effects.
2. To utilise the updated NOAA AVHRR/3 instrument (recently equipped with a new spectral channel at 1.6 micron) to investigate the geographical distribution of ice and water clouds; an important cloud feature affecting the climate.
3. To strengthen the Swedish contribution and role in international co-operation projects aimed at satellite-based climate studies (i.e., the EUMETSAT Satellite Application Facility for Climate monitoring).
Background
The availability of a 10-year cloud climatology derived from satellite data and the provision of new satellite data and cloud parameter retreival schemes at SMHI enables new opportunities for validation and further development of cloud parameterisation schemes in climate simulations models. A co-operation with the SWECLIM/SMHI Rossby centre for regional climate simulations has therefore been established for this purpose within the framework of this SNSB project. The project will also serve as a link to the EUMETSAT Climate Monitoring SAF project (Hechler, 1997) where a more complete set of satellite-derived cloud parameters will be made available in the future.
Cloudiness has an important impact on climate by causing a strong regulation of radiation conditions at the earth’s surface as well as in the atmosphere. However, the impact of clouds is very complex and sensitive depending on the existing cloud types and the cloud amount (fractional cloud cover). By relating the effects of clouds on radiation conditions to the expected radiative warming due to increased CO2 concentrations in the atmosphere, one finds that changes in cloud appearance and cloud occurrence may have a potential to both increase (positive feedback) or, alternatively, compensate (negative feedback) CO2 warming substantially. In short, a positive feedback would result if an increase of the amount of semi-transparent ice clouds (transparent in visible but thermally non-transparent – thus imposing a greenhouse effect due to clouds) takes place while a negative feedback would be seen if the amount of dense and strongly reflecting water clouds increases.
The present knowledge is that the net radiative effect of clouds on a global basis results in a negative feedback, i.e., clouds are cooling down the earth-atmosphere system (Ramanathan et al., 1989_2). However, since the net effect is a result of a difference between two very large terms, small changes in cloudiness may have rather large consequences on the net effect on the radiation climate. In addition, since cloud conditions show a large geographical variability, one could expect to find large variations and deviations from the global net cooling effect in various regions.
A consequence of the discussion above is that the description of cloudiness in atmospheric general circulation models used for climate simulations represents a large potential source of uncertainty. Many climate models have a quite simple parameterisation of clouds and it can be questioned if this parameterisation is able to describe the rather complicated forcing of radiation conditions due to clouds. Thus, there is a need to validate cloud and radiation parameterisations by use of existing cloud observations. This need is especially important for regional studies where the net effect of clouds may be significantly different from the global mean effect.
Satellite measurements offer homogeneous coverage in both space and time and this application proposes the use of cloud information from NOAA AVHRR data for climate model validation purposes.
Use of satellite data sets for model validation/evaluation at the Rossby Centre
The Swedish Regional Climate Modelling Programme (SWECLIM) has as a goal to produce high resolution climate scenarios for the Nordic region. To do this a coupled Ocean-Atmosphere Regional Climate Model (RCM) is being developed at the Rossby Centre, located at SMHI (Rummukainen et al., 1999 and Räisänen et al., 1999). Multi-annual cycle integrations of this system must be performed, in order to produce usable statistics pertaining to climate and climate change over the Nordic region.
In long coupled model integrations, the representation of clouds and the interaction of clouds with the radiation field become vitally important. Uncertainty in the simulation of cloud-radiation interaction remains the largest source of error in attempts to model the present climate with coupled Atmosphere-Ocean General Circulation Models (AOGCMS) (IPCC 1995, Arking 1991)).
In simulating cloud-radiation interactions it is vital to get the following factors correct:
1. The total (horizontal) cloud fraction and its timing in the diurnal and seasonal cycle.
For correct radiation interactions it is absolutely fundamental that the simulated cloud fields have a diurnal and seasonal cycle as close to that observed as is possible. Information is needed on the observed mean diurnal & seasonal cycle of cloud amounts at ~20-80km resolution. This is directly available through the proposed satellite data set (described in next section) and will constitute a vital validation tool for the Rossby Centre Climate Model (RCM).
2. A correct vertical structure of the cloud fields.
As well as simulating the correct overall cloud field, radiative interactions require an accurate representation of the vertical structure of cloud fields. The overall radiative impact of clouds on the Atmosphere-Surface system is strongly a function of their location in the vertical (Ramanathan et al., 1989_1). Low level clouds in the mean tend to cool the surface through reflection of incoming solar radiation. High level clouds tend to warm the atmosphere-surface through their effect on upwelling terrestrial radiation. Data sets that can give information on the vertical location of observed clouds will aid in the validation of model clouds and in improving cloud representation in the RCM.
3. The cloud albedo and cloud emissivity as a function of cloud phase (ice or water).
In the context of (2), a correct representation of the cloud albedo and emissivity is crucial for the correct radiative impact of a given cloud (Platt 1989, Roeckner 1987). Both of these quantities are dependant of the incloud constituents. In particular ice and water clouds have radically different cloud albedos. The transition between water and ice clouds is treated as a simple function of local model air temperature in the RCM. It is not known how accurate this transition is. Small errors in the relative amounts of water and ice clouds can have a large impact on the surface energy balance. The proposed satellite data set would allow a discrimination between the presence of water or ice clouds. This would be a first step towards validating the gross albedo of modelled clouds.
4. The diurnal cycle of convection.
With reference to (1), a known systematic error in a many models is an incorrect phasing of convection with the diurnal cycle. Convection tends to begin too early in the diurnal cycle, particularly in the summer (Ringer 1998). The clouds associated with this convection then tend to shade the surface and the effect of incoming solar radiation on surface temperature and evaporation is diminished. This leads to the shutting down of convection too early in the day. For numerical weather prediction models this often results in an incorrect timing of precipitation events. In climate modelling an incorrect phasing of clouds, associated with convection, to the diurnal cycle can potentially have a large effect on the surface temperature. This is particularly true in long climate integrations with a freely evolving ocean surface. Information on the diurnal cycle of convective clouds, and the fractional amount of cloud associated with that convection, would enable a diagnosis of the convection/cloud diurnal cycle in models. This will aid in the development of the RCM convection scheme.
All of the above factors will be amenable to evaluation with the suggested satellite data sets in this project. They are issues that must be addressed to improve the representation of cloud-radiation in climate models and therefore allow more accurate simulations of climate and climate change.
Existing and new satellite datasets suitable for climate model validation
Globally estimated cloud parameters based on satellite information have been available for model validation purposes by large international programmes (like the International Satellite Cloud Climatology Project (ISCCP) for more than a decade (Rossow and Garder, 1993 and Yu et al., 1996). However, these data sets have a rather coarse horizontal resolution (approximately 200-300 km) and they are based on a quite limited fraction of all available meteorological satellite data. It has also been noticed that the quality of the information is decreasing at high latitudes and near the poles (Mokhov and Schlesinger, 1993). For these reasons, it becomes increasingly evident that new satellite data sets which are more appropriate for regional model validation purposes have to be defined.
The SMHI SCANDIA model (SMHI Cloud Analysis model using digital NOAA AVHRR data - Karlsson, 1996) has been successfully used to create cloud climatologies from NOAA AVHRR data with a high horizontal resolution for an area covering a major part of northern Europe (Karlsson, 1997). Much of this work has been carried out in previous projects funded by SNSB (59/95, 113/96 and 152/98). A complete and consistent 10-year cloud analysis series containing data from the years 1991 to 2000 will be the expected final outcome of these projects. This long time series of SCANDIA cloud information is extracted with a frozen cloud retrieval scheme which implies a homogeneous error structure in the retrieved information without serious discontinuities and trends. This fact and that data is available for an entire decade make the data set interesting for the climate model validation tasks.
Most of the listed validation factors in the previous section concerning the climate
model of the Rossby centre could be studied by utilising the SCANDIA data set.
However, some adaptation or reprocessing of the data due to the specific model validation
needs seems necessary. It concerns mainly the restructuring of the data into a description
of the diurnal variation of cloudiness (especially the diurnal cycle of convective clouds)
and the extraction of information on cloud altitudes.
In addition, this proposal suggests the utilisation of results from a refined version of the SCANDIA cloud retrieval scheme taking into account the new spectral band at 1.6 micron (a channel denoted 3A in the AVHRR/3 instrument). This channel measures in a spectral region where the separability between water clouds and snow cover is greatly enhanced compared to the previous channel 3 (at 3.7 micron) used in the former AVHRR/2 instrument. An interesting consequence in the context of this research application is that it also permits an improved separation between water and ice clouds (Pilewskie and Twomey, 1987).
Data from the AVHRR/3 instrument on the NOAA-15 satellite have been made available to users during a test period in spring 1999 (from 9th March to 20th April). These data have been received and stored at SMHI. For the next series of NOAA-satellites (beginning with NOAA-16 to be launched later this year or in the beginning of year 2000), channel 3A will be operated on a continuous basis. The test data set and the expected continuous data stream from later NOAA-satellites can be used in a special validation study focused on the distribution of ice and water clouds. This study can furthermore be complemented and enhanced by utilising the SCANDIA 10-year data set (also including some information on the distribution of water and ice clouds as reported by Karlsson, 2000).
4 Links to international projects
SMHI is engaged in an EUMETSAT Satellite Application Facility (SAF) for climate monitoring (CM-SAF - Hechler, 1997). This SAF is hosted by the German Weather Service (DWD) and includes participation of scientists from Germany, the Netherlands, Finland and Sweden (SMHI). The SMHI contribution concerns assistance in the extraction of cloud parameters from satellite data. This work will be carried out in co-operation with German and Dutch scientists.
The idea is that this project, where cloud parameters from the SMHI satellite retrieval schemes are used in validation of the RCM models, would give an early experience of the direct use of the envisaged cloud parameters from CM-SAF in model validation activities. The links to the end-users are very important for all SAFs and this proposed research project may assist in this communication. Due to the special support from this research project, SMHI might also be given an increased influence on the final deliverables of CM-SAF. Of equal importance here is the fact that some additional cloud parameters of great value for model evaluation (e.g., cloud optical depth and cloud liquid water content) will become available from CM-SAF within a few years. The SMHI involvement in SAFCLIM combined with the activities in the proposed research project would then facilitate the use of these additional cloud parameters by guaranteeing an easy access of data in future model validation activities at the Rossby centre.
Finally, it must be mentioned that SMHI is involved in EU-funded research projects related to BALTEX (Baltic Sea Experiment) research activities. Of special interest here is the CLIWANET (Cloud Liquid Water Network) project where cloud liquid water retrievals will be carried out by use of a multi-satellite-sensor approach. Here, the SMHI task is to contribute with the basic NOAA AVHRR cloud processing algorithms in a co-operation with KNMI (Weather Service of the Netherlands) and IFM (Institute for Marine sciences, Kiel, Germany).
Current plans
A two-year project is envisaged with the following tentative project plan:
Tasks for year 2000
1. Planning and initiation of specific RCM model experiments for maximum use of the available satellite data set
2. Extraction and preparation of cloud parameters from the 10-year SCANDIA cloud climatology data set (i.e., diurnal and seasonal variation of fractional cloud cover - including the requested convective cloud data set - and cloud altitudes)
3. Extraction of an ice and water cloud climatology from the NOAA-15 AVHRR/3 test dataset and from the NOAA-16 AVHRR/3 operational data stream
4. Transfer of satellite-derived results to the RCM model domain
Tasks for year 2001
1. Finalising RCM simulations, possibly also complementing them with revised simulations depending on the evolution of the model development strategy at the Rossby Centre
2. Complementing the satellite validation data set with SAFCLIM or CLIWANET cloud information
3. Final evaluation of model validation results
4. Documentation and publication of results
References
Arking, A., 1991: The Radiative Effects of Clouds and their Impact on
Climate. Bull. Am. Met. Soc. 71, 795-813.
Hechler, P., 1997: Satellite Application Facility on Climate Monitoring,
Detailed Proposal, Deutscher Wetterdienst. (available from SMHI)
IPCC, 1995: Intergovernmental Panel on Climate Change (The Science of
Climate Change). Eds. Houghton J.T. et al., Cambridge University Press.
Karlsson, K.-G., 1996: Cloud classifications with the SCANDIA model,
SMHI Reports Meteorology and Climatology, No. 67, 36 pp.
Karlsson, K.-G., 1997: Cloud climate investigations in the Nordic region
using NOAA AVHRR data, Theor. Appl. Clim., 57, 181-195.
Karlsson, K.-G., 2000: Satellite sensing techniques and applications for
the purposes of BALTEX, Contr. Atm. Phys., In press.
Mokhov, I. I. and Schlesinger, M. E., 1993: Analysis of global cloudiness.
1. Comparison of Meteor,
Nimbus 7 and ISCCP satellite data, J. Geophys. Res., 98, 12 849-12 868.
Pilewskie, P. and Twomey, S., 1987: Cloud phase discrimination by
reflectance measurements near 1.6 and 2.2 micron, J. Atm. Sciences, 44,
3419-3420.
Platt, C.M.R., 1989: The Role of cloud microphysics in high-cloud feedback
effects on climate change, Nature, 341, 428-429.
Ramanathan, V. et al. 1989_1: Cloud-Radiative Forcing and Climate: Results
from the Earth Radiation Budget Experiment, Science, 243, 57-63.
Ramanathan, V., Barkstrom, B.R. and Harrison, E.F., 1989_2: Climate and
the earth’s radiation budget, Physics Today, 22-32.
Ringer, M.A., 1998: Tropical convective rainfall in the global UM:
Initial comparisons with estimates derived from Meteosat infrared imagery.
Forecasting Research, Techn.
Rep. No. 239. Meteorological Office, Bracknell, Berks. UK.
Roeckner E. et al., 1987: Cloud optical depth feedbacks and climate
modelling, Nature, 329, 138-140.
Rossow, W.B. and Garder, L.C., 1993: Cloud detection using satellite
measurements of infrared and visible radiances for ISCCP, J. Climate, 6,
2341-2368.
Rummukainen, M., Räisänen, J., Ullerstig, A., Bringfelt, B., Hansson,
U., and U. Willén, 1999: RCA - Rossby Centre regional Atmospheric climate
model: model description
and results form the first multi-year simulation, SMHI Reports Meteorology and
Climatology, No 83, 76pp.
Räisänen, J., Rummukainen, M., Ullerstig, A., Bringfelt, B., Hansson,
U., and U. Willén, 1999: The first Rossby Centre regional climate scenario -
Dynamical downscaling of
Co2-induced climate change in the HadCM2 GCM. SMHI Reports Meteorology and
Climatology, No 85, 76pp.
Yu, W., Doutriaux, M., Sèze, G., Le Treut, H. and Desbois, M., 1996: A
methodology study of the validation of clouds in GCMs using ISCCP
satellite observations, Climate
Dynamics, 12, 389-401.