5.   TASKS AND METHODS

5.1     Principal Task 1 (PT1): Dynamical downscaling

Principal Investigator: Thor Erik Nordeng, DNMI, Oslo

 

Major Objective:

To use a Regional Climate Model, in some cases coupled to a slab ocean model, to estimate the regional climate in Northern Europe and adjacent sea areas, given the best estimates of climate scenarioes from a coupled Atmospheric-Oceanic GCM

 

Tasks:

We use HIRHAM with one-way influence nesting based on the output from a highly rated AOGCM; In Phase I, data from the MPI's ECHAM4 are used, but other options will also be considered in Phase II A stepwise approach is chosen. First there is an implementation phase where the model system is set up. Second, the model is forced with large-scale observations (global analyses) at the lateral boundaries. We use data from the re-analysis archive at the European Centre for Medium Range Weather Forecasts (ECMWF) in Reading, UK. Then the first real climate experiment is performed by forcing the model with AOGCM-output for the present climate. Finally, we perform simulations where output from the same AOGCM for climate change scenarios will be used to force HIRHAM. In this way we will validate both the use of HIRHAM in our region, as well as the combined AOGCM-HIRHAM setup. In Phase I time-slices of 5 years with HIRHAM have been run. Our experience so far indicates that longer runs may be necessary. Several scenario-runs may also be necessary, reflecting to inter-decadal variabilities, as revealed by variations of the North Atlantic Oscillation (NAO) index. One experience from the work of Phase I is that the regional climate in our region is strongly dependent on the sea surface temperatures and wether the ocean is ice-covered or not. The same conclusion was drawn in the Sweclim project. They found that a ocean and sea ice model of the Baltic sea were necessary to achieve reasonable temperatures in northern Sweden and Finland (Ref. SMHI Reports No. 85, Feb. 1999). This is by no means surprising since sea suraface temperatures and ice cover are taken from the forcing model with its much coarser horisontal resolution than HIRHAM.

 

In Phase II we therefore plan to couple HIRHAM to a slab ocean model to achieve regional scale forcing from the ocean with the same resolution as in HIRHAM. Developing the coupled model will mainly take place in PT 7 (some resources in PT 1, however) while running the model will be incorporated into Tasks 1.3 and 1.4.

 

Nesting is normally performed by the boundary-relaxation method of Davies (1976) and Kållberg and Gibson (1977), where the variables from the forcing model and the nested model are blended within a relaxation zone along the boundary. RegClim has not found it necessary to use the spectral relaxation of K­ida et al. (1991), such as proposed in Iversen et al. (1997) for controling the larger scales (Jones et al., 1997).

 

Two tasks from PT5 are moved to PT1. Old tasks 5.8 (Unified convective scheme) and 5.9 (shallow convective closure) both end in Phase I. Siebsma and Cuijpers (1995) showed that current mass flux schemes for parameterisation of convective processes in General Circulation Models describe the average dynamics of the shallow convection reasonably well, but that entrainment and detrainment rates should be one order of magnitude larger. Nordeng (1994) suggested improvements to the Tiedtke (1989) scheme that actually treated entrainment and detrainment according to Siebesma and Cuijpers. In Phase I these ideas have been incorporated into HIRHAM and sensitivity runs have been made. Our results are, however, not conclusive and the ideas have not been used for the main runs so far. A report summarising the results of PT 5.8 and 5.9 (which end by 1. July 1999) is under preparation. The work will, however, continue in PT 1 by Thor Erik Nordeng as a contribution in kind from DNMI.

 

As a part of the validation of HIRHAM, DNMI will study the results from the dynamical downscaling with respect to the greenhouse effect over the North-Atlantic Ocean. In particular the contribution of water vapour and the climate feedback connected to this. The water vapour feedback, particularly the effect of vapour in the upper troposphere has been subject to uncertainty. (See pp. 200-201 in IPCC, 1996).

 

The greenhouse effect, defined as the difference of outgoing long wave radiation at the surface and at the top of atmosphere (TOA), will be described using results from HIRHAM-simulations (see e.g. Raval and Ramanathan, 1989). We will study its statistics in terms of dependence on surface temperature and possibly other model-parameters. In particular the spectral greenhouse parameter will be calculated for the ATOVS channels, with emphasis on water vapour channels. This will prepare the results for a direct comparison of the model statistics with statistics of actual satellite observations over the North-Atlantic Ocean. Earlier studies have shown deviations between ECHAM and TOVS-derived humidity data (see Chen et al., 1996). Differences between analysed ECMWF-fields and TOVS-radiances were described by Salathe and Chesters (1995).

Here, the HIRHAM-results for present-day climate runs will be used with boundary values from the ECHAM model (scenario-mode). Satellite observations of TOA radiation will be simulated from the model data to create statistics that are prepared for comparison with the statistics of real satellite data. Temperature and moisture profiles as well as other quantities connected to the radiation budget will be stored for generation of this statistics. A comparison with similar statistics from real satellite data will be performed using the water vapour channels of the HIRS and possibly AMSU instruments of the ATOVS instrument package to assess HIRHAM’s ability to simulate the observed water vapour climate feedback.

This validation will be done in addition to using statistics from more traditional (and regular) observations.

The following sub-tasks are put up for PT1:

 

Task 1.1:    Implementation and set-up of HIRHAM

The HIRHAM model is used for regional climate simulations at various centres (see discussion in chapter 1.3.1). DNMI is however at present using different computers than these centres and some resources are necessary in order to implement and make the model run efficiently at DNMI.. The original task is completed and finished in Phase I.

The task will, however, continue in Phase II when HIRHAM will be coupled to a slab ocean model. This requires somewhat more work with regard to making the system run efficiently.

The task is technical, and no separate report will be published.

 

Task 1.2:    HIRHAM forced with ECMWF analyses

In order to evaluate how the model behaves in reproducing the climate in our region and setting the standard for the coming tasks, the model is forced with large-scale observations (global analyses from ECMWF) at the lateral boundaries.

In Phase I we planned for a 5 year run. This is technically and scientifically completed, and preliminary reported in RegClim General Technical Report No. 1.

In order to compare the HIRHAM runs forced with an AOGCM for present climate with a statistically long enough period we will extend the runs to 15 years. This will be performed in Phase II.

 

Task 1.3a: HIRHAM forced with an AOGCM for present climate, high NAO-index

The AOGCM will have some characteristic climate signals, which may vary depending on large-scale flow patters (inter-decadal variablity). It is therefore necessary to perform more than one time-slice experiment and we will use the NAO-index to distinguish between various flow patterns. Results from this task will be compared to those from Task 1.2 to assess the difference between the climate as seen by the ARCM forced with an AOGCM as compared to the ARCM forced with “reality”.

In phase I we planned for a 5 year run. This is technically and scientifically completed, and preliminary reported in RegClim General Technical Report No. 2.

In Phase II the time-slice will be extended. In addition HIRHAM will be coupled to a slab ocean model.

 

Task 1.3b:  HIRHAM forced with an AOGCM for present climate, low NAO-index

Results from this task will be compared to those from Task 1.2 and 1.3 to assess the difference between the climate as seen by the ARCM forced with low and high index flows in the AOGCM.

In phase I we planned for a 5 year run. This is technically completed. Not yet analysed and reported.

In Phase II the time-slice will be extended. In addition HIRHAM will be coupled to a slab ocean model.

Task 1.4a:  HIRHAM forced with an AOGCM for future climate, high NAO-index

As for those experiments where the ARCM is forced with global model simulations of present climate, it is necessary to assess interdecadal variablity in a future climate by forcing the ARCM under various large scale flow patterns from the AOGCM. We will therefore use high and low NAO index here as well. Results from this task, as well as preceeding tasks, will be used to assess possible climate signals in our region.

In phase I we planned for a 5 year run. This is technically and scientifically completed, and preliminary reported in RegClim General Technical Report No. 2.

In Phase II the time-slice will be extended. In addition HIRHAM will be coupled to a slab ocean model

 

Task 1.4b:  HIRHAM forced with an AOGCM for future climate, low NAO-index

(See text related to 1.4a).

In phase I we planned for a 5 year run. This is technically completed. Not yet analysed and reported.

In Phase II the time-slice will be extended. In addition HIRHAM will be coupled to a slab ocean model.

 

Task 1.5: Further tests and experimentaions of slantwise convection.

This is done as a continuation of tasks 5.8 and 5.9 of PT 5. If successful, the revised convective scheme will be used in tasks 1.3 and 1.4.

 

Relations to other Principal Tasks:

PT 3:   Output of data from AOGCM and ARCM to be provided for empirical downscaling. Collaboration concerning evaluation of representativity of time slices used in dynamical downscaling experiments. Results of Empirical Downscaling will be held together with those of Dynamical Downscaling.

PT4, PT5 and PT6:    

Results from these tasks will be evaluated with the aim of improving process descriptions.

PT 7:   Output of data for AOGCM and ARCM for Sea-state modelling. The coupled HIRHAM/slab ocean model will mainly be developed in PT7.

PT 8:   High quality regional climate data are essential for all validation of ARCM and AGCMS.

PT9:     Interpretation of model results will be undertaken in close co-operation with PT9.


 

Milestones:

 

Table 1: Target schedule PT1

Task

Year

 

1997

1998

1999

2000

2001

2002

1.1

 

x

 

 

 

  x x        

1.2

  x x

 

 

 

x x x x    

1.3a

  x x

x

x

 

x

x x x    

1.3b

  x x x x

x

x

x

x

x

   
       1.4a   x x x x x x x x x    

1.4b

   x x

x

x

x

x

x

x x    
       1.5            x x x        

Finalization

                   

x

 

 

Institutions and Personnel:

Task 1.1-1.4:  

The work is to be carried out at DNMI by Dr. J. E. Haugen and Dr. D. Bjørge. Other scientists in the Section   for the atmosphere (DNMI) will be involved in several of the proposed tasks.

Task 1.5:

The work will be carried out by Prof. Thor Erik Nordeng

Budget:

The following numbers give contributions to the institution in units of Person-years. New in PHASE II  and PHASE II Extension are given in bold.

 

Contribution from the Research Council:

 

            2000    2001

Tasks 1.1 – 1.4

DNMI:                                     2.0       2.0

Finalisation

            DNMI                                                             

RegClim Phase II:                                 2.0       2.0      

Own funding:

PI of PT1 and Task 1.5

            DNMI                                      0.2       0.2      

 

Own funding:

Nordeng’s work-time (PI for PT1 and the contribution in Task 1.5) is a contribution in kind to the project from DNMI.

 

 

 

5.2         Principal Task 2 (PT2): Basin scale ocean modelling of the Nordic Seas

Principal Investigator: Helge Drange, NERSC, Bergen.

 

Major Objective:

            To establish and validate a model tool for oceanic circulation and thermodynamics with focus on the Nordic Seas, but suitable for coupling to a global atmospheric model.

 

Tasks:

The dynamics of the climate system of the Earth is to a large degree determined by the atmosphere and the oceans, and the interaction between these systems. Whereas the atmosphere is characterised by fluctuations on daily to weekly time scales and has a memory of a few months, anomalies in the ocean system may last for decades with potential long-term feedback on the climate system. It is therefore believed that simulations of the seasonal to decadal variabilities of the climate system - if possible at all - need proper description of the ocean component.  As a concequence, there is an urgent need to improve the understanding of the natural variability of the system, in order to assess large scale (or integrated) changes as the human generated pollution of the Earth's atmosphere continues, and to examine how the natural variability of the climate system may change due to human activities. Many of these problems rely on a proper description of the ocean component.

 

Up to now, rather coarse resolution versions of climate models have been used to study the Earth's climate system. This may give a reasonable description of large-scale features of the climate system, but regional climate simulations of variability on sub-basin scales are, in general, of poor quality. In particular, regions characterised by sharp fronts in the dynamic and/or the thermodynamic fields or by abrupt topography or bathymetry are especially affected. The North Atlantic region, characterised by sharp thermal fronts and dynamically unstable systems in both the atmosphere and the ocean, and with a complex topography, is an example of a typical problem region.

 

Deficiencies in coupled climate models, due to improper horizontal or vertical model resolution, improper parameterisation of sub-grid scale processes, or improper numerics and integration techniques, could be particularly severe in the North Atlantic as this region is frequently influenced by strong and persistent anomalies. In addition, the North Atlantic region is also unque in that there is an exceptional northward transport of heat in the ocean, and that the major part of the communication between the Arctic Ocean and the world oceans occur through this region.

 

For the reasons stated above, and the likely possibility that the North Atlantic Ocean is particularly sensitive to changes in the thermal and fresh water forcing, indicate that it is of paramount importance to further improve, test, and validate ocean general circulation models for the region. Therefore, a major challenge in climate modelling is to understand the mechanisms through which oceanic anomalies form, propagate, and decay, and to understand how their lifecycles are related to fluctuations in the atmosphere. The existence of correlations between atmospheric fluctuations such as the North Atlantic Oscillation and changes in the ocean circulation is now well established (Dickson et al., 1996; Curry et al., 1998). However, it is very difficult from such correlations alone to infer the crucial information about causality that is essential to understanding mechanisms.

 

Testing and validation of the ORCMs will be based on the knowledge achieved during OGCM intercomparison projects covering the North Atlantic Basin (for instance Chassignet et al, 1996; Roberts et al., 1996; New and Herrmann, 1997; and Herrmann, 1997), and the modelling studies of Oberhuber (1993), Aukrust and Oberhuber (1995), and Drange and Simonsen (1996a,b) in which both the Atlantic Ocean, the Nordic Seas, and the Arctic Basin are included.

 

An important part is to determine the model area. Both observations (Bjerknes, 1964; Sutton and Allen, 1997) and modelling studies with AOGCMs (Delworth, 1996; Griffies and Bryan, 1997) indicate that the surface and sub-surface waters of the Atlantic subtropical and subpolar gyres are slowly oscillating on decadal time scales. This oscillation, which can be identified as propagation of sea-surface temperatue anomalies, is probably one of the major sources for the well-known sea-level pressure variability over the North Atlantic/Nordic Seas and the Northern Europe. To minimize the computer requirements the possibility of using the stretched co-ordinates options is utilized. This allows the use of a fine mesh in areas where it is needed.

 

The work is divided into two main activities – to establish a coupled ice-ocean model for the region (Task 2.1), and to assess how the model simulates the mean ocean state and the observed variability over the last 50 years, including identifying and investigating mechanisms for the observed variability (Task 2.2-2.5).

 

Task 2.1 as described below is identical to Task 2.1 given in the project description of RegClim, Phase 1. The new Task 2.2 has been redefined and is split into more specific tasks, and consists of the activities described in the old Tasks 2.2 and 4.5 (new and old refer to phase 1 and 2 of RegClim).

 

Task 2.1:    Establish an ice-ocean regional circulation model

The ocean models to be used involves parameterizations of many complicated and poorly known physical processes. The relative importance of the included physical processes may change when applied to polar areas as proposed here. The choices for the parameterizations and their mathematical formulations in ocean models are known to have a significant impact on the simulation results (e.g., Manabe and Stouffer, 1988; Melsom, 1996).

 

In the present application both POM and MICOM will be run with much too coarse resolution to resolve meso-scale ocean features. These dynamic scales must therefore be parameterized. A recent approach has been suggested by Gent and McWilliams (1990), and Visbeck et al. (1997), and may be applied here. It consists of a combination of eddy transient advection and a horizontal diffusion parameterization. In contrast to MICOM, where the horizontal diffusion is automatically along isopycnals (except in the upper mixed layer), the original version of POM has large diapycnal mixing in areas where the isopycnal slopes with respect to the sigma coordinates. Rotation of the diffusivity tensor (Redi, 1982) is suggested as a solution.

 

Similarly, convective motion is not resolved and must be parameterized. Recent work by Marshall and coworkers (e.g., Marshall et al., 1997) and the parameterization by Cox and Bryan (1984) may form the basis for this activity.

 

Finally, the advection scheme in POM may not be adequate for the transport and conservation of water mass properties for longer climate runs. Alternative schemes are available (Berntsen, J. et al., 1996; Drange & Bleck, 1997).

Completed May 1999. Based on the obtained results and of the state of the models, it has been decided to contine the prognostic long-term simulations with MICOM. Sensitivity studies with POM will be made in the new PT7.

Reports: Budgell (1998), Drange (1998), Drange et al. (1999), Lunde and Hackett (1998), and Ådlandsvik (1998).

 

Old Task 2.2:  Model validation and comparison

Before any applications are to take place the performance of the models must be assessed. This will be done by comparing model generated hydrographic sections and transport rates through vertical sections derived from a 15 year run with ECMWF reanalyzed meteorological forcing fields (or for an even longer period with data from NCEP, USA) with observed quantities of the same. Observed transports through vertical sections will be based on current measurements and literature gathered in PT8 (Task 8.2).

Continues in PHASE II splitted into the tasks given below (2.2 is redefined).

Reports: Drange (1998), Drange et al. (1999), and Otterå (1999)

 

New Task 2.2:             Quantification of the mean ocean and sea ice states and the natural variability of the states during the last 50 years

The Nansen Center version of MICOM will here be forced with 12 hourly NCAR/NCEP atmospheric 10m wind, 2m air temperature, 2m humidity, total cloudiness, precipitation and sea level pressure analysis fields for the period 1958-98. The mean ice-ocean state will be determined, and deviations from the mean state will be identified and quantified.

 

New Task 2.3: Model validation

The mean ice-ocean state and identified anomalies obtained in T2.2 will be compared to available observations of the hydrography and circulation in the region, and to available literature estimates of integrated quantities like mass, heat, salt and ice transports through the passages and over the ridges confining the Nordic Seas.

 

New Task 2.4: Identification and quantification of the driving mechanisms for the simulated natural variability

The formation, propagation and decay of the anomalies identified in T2.2 will be examined. Special focus will put on the mechanisms for the formation of the anomalies.

 

New Task 2.5:             Sensitivity of the model results to model formulations

The role of increased horizontal and possibly vertical resolution will be assessed by performing at least 10 years integrations with different grid configurations of the model, including simulations with a horizontal grid resolution of about 20 km. In addition, the effect of including more sophisticated isopycnic (biharmonic versus harmonic mixing, Visbeck et al. (1997) parameterisation of isopycnic mixing, etc.) and diapycnic (improved parameterisation of the benthic boundary layer, velocity shear dependent mixing, etc.) mixing schemes will be addressed.

 

Relations to other Principal Tasks:

PT 1:   The POM model will be used for sensitivity studies. In Phase II this activity is moved to the redefined PT 7.

PT 4:   The MICOM model will be coupled to a global atmosphere model used in PT 4

PT 7:   The ice model and ocean-surface parameterisations developed under PT 7 for the POM model will be compared to similar processes in MICOM.

PT 8:   Data for validation of coupled ice-ocean results are to be provided by
PT 8.

PT9:     Interpretation of model results will be undertaken in close co-operation with PT9.

 

Milestones:

 

Table 2: Target schedule PT2

Task

Year

 

1997

1998

1999

2000

2001

2002

2.1

 

x

x

x

x

             

2.2

     

x

x

x

           

2.3

     

x

x

x

x

         

2.4

         

x

x

x

x

x

   

2.5

     

x

x

x

x

x

       

Finalization

                   

x

 

 

Institutions and Personnel:

Task 2.1-2.5:   

The work will be performed at NERSC and HI. At NERSC the work will be headed by Dr. Helge Drange (PI PT2), and will involve M. Bentsen and the modeling team affiliated with the G. C. Rieber Climate Institute within NERSC.

The work at IMR is headed by Dr. Bjørn Ådlandsvik with assistance of Dr. Paul Budgell.

Budget:

The following numbers give contributions to the different institutions in units of Person-years. New funds applied for in RegClim PHASE II are given in bold types.

 

Contribution from the Research Council:

                                               2000                2001   

Task 2.1

            Finished

Task 2.2-2.5:

DNMI: (finished)                                               

IMR:                             0.4     

           NERSC:                       0.9                  0.7         

RegClim Phase II:                     1.3                  0.7         

 

Contribution by NERSC and synergy with other projects at NERSC

The above stated tasks can not be completed based on the present requested funding only. Therefore, additional local funding and new projects are needed to support the proposed activites. At present, a conservative (low) estimate of external supported RegClim activities at NERSC is 3 person-years per year. External support on the present level for the following years is needed to some extent, but the details will depend on which projets that receive support in the 5th Framework Programme of EU.

 

5.3     Principal Task 3 (PT3): Empirical downscaling

 

Principal Investigator: Eirik Førland, DNMI, Oslo

 

Major Objective:

            To establish empirically based climate change scenarios on regional and local scales, with resolution in time and space appropriate for further impact modelling.

 

Tasks:

Empirical downscaling links large-scale meteorological fields (predictors) to local-scale climate elements (predictands). When such relations are established, scenarios for the predictands may be given based upon scenarios for the predictors. Scenarios will be developed for the predictands, with the time- and spatial resolutions given in the following table:

 

 

PREDICTAND

TIME  RESOLUTION

SPATIAL RESOLUTION

Temperature (mean, min., max.)

daily and monthly

Regions (~10 000 - 100 000 km 2 )

Precipitation

daily and monthly

Regions (~10 000 - 100 000 km 2 )

Wind and wave height

daily

key stations

Cloud cover

daily and monthly

key stations

Number of days with snow cover

seasonal

key stations

Degree days, growth days, frost days

seasonal

key stations

Number of days with precipitation exceeding threshold values.

Number of days with wind speed exceeding threshold values

monthly or seasonal

key stations

 

Scenarios for changes in mean temperature and precipitation will be made for regions where trends and variability of these elements have been similar during the period of historical records. In Phase I of RegClim, Norway has  been divided into 13 precipitation  regions and 6 temperature regions based upon a large number of long-term, monthly series ( Hanssen-Bauer & Førland, 1998; Hanssen-Bauer & Nordli, 1998). These regions will be fundemental in describing regional scenarios for temperature and precipitation. Scenarios for the other predictands will be developed only for some key stations. In Phase I the empirical downscaling has been concentrated on establishing scenarios for monthly values (Førland et al., 1998). In Phase II this analysis will continue, but the main emphasis will be on daily values (cf. Table above)

 

Task 3.1:    Establishing datasets for empirical downscaling

A comprehensive compilation of data sets have been implemented within RegClim in Phase I, according to the scheme presented in the project proposal (Iversen et al., 1997). The data sets contain three types of data:

a)      Historical observations for description of local/regional climate (predictands)

b)      Data for description of predictor fields (atmospheric circulation, surface temperature, sea surface temperature, geopotential heights, etc.)

c)      Data for evaluation of AOGCM control runs, and simulation of climate scenarios

COMPLETED Reported in: Benestad (1998a)

 

 

Task 3.2:    Development of empirical relationships between observed large-scale atmospheric fields (predictors) and local climate variables (predictands).

Several methods for empirical downscaling (cf. Introduction) have been evaluated in Phase I (Benestad, 1998b,c,d, Hanssen-Bauer & Førland, 1998a, Hanssen-Bauer,1999). It turned out that different methods will be preferential for different time-scales and climate elements.  Two main schemes of linking large-scale circulation to local climate have been considered:

 

a)    Two-step methods, where the first step is to classify the atmospheric circulation for each day or month, and the second step is to analyse the local climate for each class

Many downscaling techniques use some form of circulation pattern classification to create weather types. Techniques for objective classification based upon Principal Component Analysis (PCA) of the predictor fields over a specific area are outlined by e.g. Jones et al., 1993, Zorita et al., 1995. For the analysis of daily values in Phase II the analogue method (Zorita et al., 1995) seem to be appropriate. In Phase II models based on a combination of predictor fields will also be evaluated, e.g.  through a combined PCA of SLP and upper-air pressure and temperature fields (e.g. as described by Wilson et al., 1992). Objective weather classification schemes not based upon PCA will be applied only in case the PCA based methods seem inadequate. E.g. for downscaling of daily precipitation it is essential to distinguish between days with and without precipitation. In case the PCA based methods fail to do this, the CART technique (cf. section 1.3.1) is an alternative (Hughes et al., 1994, Zorita et al., 1995).

 

A crucial point concerning empirical downscaling is how to attach the local predictands to the weather classes. A problem when using classification methods based solely on SLP, is that only circulation induced changes in the local climate elements will be included. Since increased temperature affect the water budget of the atmosphere, the precipitation scenarios should also include the possibility of changes which are not connected to the circulation. A solution that will be investigated is, for each large-scale weather class, to connect the local climate elements to e.g. upper air temperatures using regression analysis.

 

b)    Methods based on correlation between time series of large-scale predictors and local climate elements.

Relationships between large-scale predictors and local climate elements may be established by multiple regression  (Schmidt & von Storch, 1993; Alexandersson et al., 1997) or by statistical techniques to detect coupled patterns in large-scale fields of climate data (Zorita & von Storch, 1997).  In Phase I, both multiple regression (Hanssen-Bauer & Førland, 1998a, Hanssen-Bauer, 1999) as well as coupled pattern techniques like Canonical Correlation Analysis (CCA) (Benestad, 1998b), Singular Vector Decomposition (SVD) (Benestad, 1998c) and Multivariate Regression (MVR)(Benestad, 1998d) have been evaluated and applied on monthly values of temperature and precipitation.

Almost finished for monthly values, in an initial phase for daily values. Will continue in Phase II.

Reported in:   Benestad (1998b,c,d), Førland et al. (1998), Hanssen-Bauer (1999), Hanssen-Bauer & Førland (1998a).

 

Task 3.3:    Testing the established relationships on independent observed data

One problem associated with empirical models and downscaling of future climate scenarios is that there is no guarantee that the past relationship between predictands and predictors holds for the future. In Phase I this was evaluated both for the multiple regression analysis of monthly values (Hanssen-Bauer, 1999) and in the CCA-models (Benestad, 1998b). Results from relationships based on SLP as well as surface temperatures and upper-air data have been compared (Benestad, 1999b), and trend analysis has been applied to the time series of residuals (Hanssen-Bauer, 1999). In Phase II similar tests for non-stationarity and trends in residuals will be performed on daily series.

Almost finished for monthly values. Not started for daily values. Will continue in Phase II.

Reported in:   Benestad (1998b), (1999b), Hanssen-Bauer (1999).

 

Task 3.4:    Evaluation of control simulations of atmospheric circulation, and use of the established relationships to check consistency between observed and simulated local climate

Use of empirical downscaling to give realistic local/regional climate scenarios based on AOGCM experiments depend on two assumptions:

a)    The control run simulations must represent the major characteristics of the observed SLP field;

b)    The spatial resolution of the AOGCM simulations of SLP must be sufficiently detailed to give adequate input to the empirical downscaling models.

 

Assumption a) has been accomplished for monthly values both by comparing the average and the standard deviation of the simulated and observed SLP fields, and by comparing the first 4-5 principal components of the control run SLP to the similar observed fields (Benestad et al., 1999). Assumption b) have been evaluated by comparing CTL simulations to observed values (site specific and for different scales) (Benestad, 1998a, Benestad et al., 1999)

 

In Phase I, Benestad et al (1999) found that for most climate elements, the ECHAM4/OPYC3 CTL integration gave a reasonable description of the average large-scale features. However, due to low frequency oscillations in the model integration, some statistical significant differences were found between different 50-year and 30-year periods. When using short ”time-slices” to establish scenarios it is crucial to be aware of this inter-decadal variability.

 

Differences between the EOFs of the model results and observations present a problem for downscaling of the model data to local climate variability.  In Phase II the problems caused by geographical model biases will be solved either by projecting the spatial model EOF patterns onto those of the observations (Sengupta & Boyles, 1998), or by merging observational and model datasets (e.g. by common EOFs ) (Barnett, 1999).

In progress for monthly values. Not started for daily values. Will continue in Phase II.

Reported in: Benestad et al. (1999).

 

Task 3.5:    Application of established relationships on large-scale fields projected by AOGCM to infer changes in local climate characteristics

The projected changes (up to year 2100) in local climate will be inferred by combining

a)      Empirical relationships between observed large scale predictors and local climate variables (Task 3.2), and

b)      Projected simulations of large-scale predictors

             Scenarios will be worked out for the various climate elements. By using the proposed downscaling-techniques (task 3.2), the scenarios may be given as both means, frequency distributions and extremes. The first tentative scenarios based on empirical downscaling of temperature and precipitation are presented by Benestad, 1999. In Phase II these monthly based scenarios will be further elaborated, and the main task will be to establish scenarios also based on daily values to accomplish the list of predictands presented in the first section of chapter 5.3

In progress for monthly values. Not started for daily values. Will continue in Phase II.

Reported in: Benestad (1999).

 

New Task 3.6 (dec00): Adaptation of downscaled scenarios for impact studies, with emphasis on hydrological consequences.

For impact studies, detailed downscaled scenarios with high spatial and temporal resolution are needed. This will be accomplished by developing fine-scale downscaling procedures for daily time-series of temperature (T2) and precipitation (P). Three different techniques will be applied and evaluated:   

 

a)      Model Output Statistics (MOS) applied directly on series of T2, P and SLP fields from the dynamical downscaling results in PT1. Basics for establishing the statistical relationships will be 15-year runs of ARCM (HIRHAM) forced with ECMWF ERA analyses (re. Task 1.2)

b)      Empirical downscaling of T2 and P from large-scale predictors (SLP, upper-air fields, etc.) from the HIRHAM-runs in PT1 (Tasks 1.3 & 1.4).

c)      Empirical downscaling of T2 and P from large-scale predictors from AOGCM-runs (re. Tasks 3.2-3.5).

 

The obtained empirically downscaled daily values of T2 and P will be used as input to a conceptual water-balance model, e.g. the Swedish HBV-model (Bergström, 1992) or the DNMI-model AMOR (Førland et al., 1996).  By applying regional parameters, gridded versions of these models will provide fine-scale values of evapotranspiration, snow reservoir (as water equivalent), soil-moisture reservoir and runoff. To keep internal consistency, the evapotranspiration calculated within the water-balance model will be used in the simulations. The difference between evapotranspiration directly from the HIRHAM-run and from the water-balance model will be evaluated. Water-balance simulations will be performed for selected gridcells in different regions/elevations representing different land use. The gridcells will primarily be allocated to a selection of DNMI weather stations. However, by spatial interpolation (GIS-based kriging) of the daily anomalies from the MOS-analysis, T2 and P will also be estimated for selected unmeasured grid-cells. For each of the selected gridcells, water-balance simulations will be performed for different land use characteristics. The estimates from single grid-cells can be integrated to cover larger areas or river basins.

 

The simulations will be run for 15-30 years ”time-slices”, both utilizing observations, as well as present day and climate scenario results deduced from the ECHAM4/OPYC3 GSDIO run. The results will be given both as ensembles of the water-balance throughout the year during the time slice-periods, as well as mean values, extremes and frequency distributions. In addition to projected scenario changes in the hydrological parameters (snow accumulation, snow reservoir, snowmelt, run-off, evaporation, soil-moisture reservoir), the fine scale downscaled results will be used to describe changes in degree-days, heating-season and growth season.

 

DNMI has offered to finance this incipient stage of hydrologic applications in RegClim, on the condition that the contribution is regarded as preliminary. Hydrologic modeling will be applied for later.

 

Relations to other Principal Tasks:

 

PT1:    Input of data from ARCM for empirical downscaling. Collaboration concerning evaluation of representativity of time slices used in dynamical downscaling experiments. Results of Empirical Downscaling will be held together with those of Dynamical Downscaling. The results from the HIRHAM runs will be adapted for impact studies.

 

PT9:    In task 9.1, Scripts developed within PT 3 will be incorporated in Task 9.1. In task 9.4, analyses of the flow regimes ECHAM4/OPYC3GSDIO integration will be accomplished in cooperation with PT3


 

Milestones:

Table 3: Target schedule PT 3

Task

Year

 

1997

1998

1999

2000

2001

2002

3.1

 

x

x

x

               

3.2

 

x

x

x

x

x

x

x

       

3.3

     

x

x

x

x

x

x

     

3.4

     

x

x

x

x

x

x

x

   

3.5

     

x

x

x

x

x

x

x

   

3.6

              x x x    

Finalization

                   

x

 

The first results are available for chapter 2 and 10 in the IPCC Third Assessment Report.

Institutions and Personnel:

 

Task 3.1-3.6:   The work is to be carried out at DNMI by Dr. R. Benestad, Dr. I. Hanssen-Bauer, senior scientists E. J. Førland (PI) and O.E. Tveito as well as other scientists at the DNMI Climatology Division. All own contributions from DNMI at present stage.

 

Budget:

The following numbers give contributions to the institution in units of Person-years. New funding in Phase II is given in bold types.

 

Contribution from the Research Council:

 

                                                            2000                2001   

Task 3.1-3.5 

            DNMI                                     1                       1         

RegClim Phase II:                                1                       1         

 

Own Funding

                                                          2000                2001

Task 3.6

            DNMI:                                  0.25*               0.5*    

*This contribution from DNMI is regarded as preliminary in the incipient stage of hydrology in RegClim.

 

5.4     Principal Task 4 (PT4) The role of the Nordic Seas: Atmosphere-Ocean feedback

Principal Investigator: Nils Gunnar Kvamstø, GfI-UiB, Bergen

 

Major Objective:

To improve our understanding of the relative role of mechanisms influencing the Norwegian Atlantic Current, and to estimate its impact on sea-surface temperature (SST) and sea-ice the Nordic Seas and the climate in our region.

Tasks:

Our aim is to give estimates for seasonal, interannual and decadal variation in regional climate under global warming. As far as computer power permits, also interdecadal variability will be addressed, but with only limited expectations.

To achieve this, a coupled atmospheric/ocean model with focus on the North Atlantic is implemented. In a fully coupled system, an integration interval of 10-15 years may be too short due to the fact that any drift in the mean (model) ocean state or misrepresentation of interdecadal coupled oscillations may disturb the signal of interest. We therefore suggest a setup with longer integration time in order to get a representative variability spectrum of the model and in this way better quantify any of the above mentioned “error” sources. For our purposes an integration interval of at least 50 years should be sufficient in order to obtain representative characteristics of the variability spectrum, up to decadal variations, of the coupled system (Sutton 1997, Hurrell 1995). Longer integrations will be made if computer resources are available. The coupled configuration may consist of MICOM with variable resolution that gives sufficient description of the dynamics and thermodynamics in the target area and ARPEGE with regular horizontal resolution of T42. Following studies like Huck et al (1997), Delworth et al. (1994) and Frakignoul et al. (1992), it seems that with a T42 (2.8 degrees) resolution, the atmospheric forcing (on the variably resolved ocean) should be represented in sufficient detail. At this resolution, the relevant atmospheric phenomena which forces the ocean on a long time scale, are resolved and the computer requirements will be met by the national resources. The main AOGCM integrations show substantial decadal variations, e.g. in storminess (L. Bengtsson, pers. comm.), and such variability will be studied from these runs. In our project, multidecadal time slices representing present and future climate will be considered.

 

In addition to the time slice experiments with coupling to the ocean, different kinds of sensitivity experiments are made. Such experiments are run with global models for periods of 10 years with and without a specified changed forcing, such as prescribed changes in SST and sea-ice cover in The Nordic Seas.

 

Task 4.1:         Control run with AGCMS

This is an AMIP-type (Gates, 1992) run using observed SST and sea-ice data from 1979 to 1988. The experiment cover the whole period with a stretched, global T79 resolution with focus in our area.

COMPLETED. Reported in Kvamstø (1998a). RegClim, General Technical Report No. 1.

 

Task 4.2:         Sensitivity experiments with AGCMS

Experiments similar to Task 4.1 will be made, but with perturbed SST and sea-ice conditions in The Nordic Seas and Arctic Seas. Imposed anomalies may be established on the basis of historical data, e.g. by identifying positive and negative extremes in a 100 years period and employ the difference between these extremes to establish sound anomalies. The results will be evaluated (analysed) against the control run described under Task 4.1.

Original experiment completed, but may be extended.

Reported in Kvamstø (1998b) RegClim, General Technical Report No. 1; and Kvamstø (1999) RegClim, General Technical Report No. 2.

 

Task 4.3:         Technical Coupling of ARPEGE/MICOM and initialisation

The coupling will be done with the OASIS software mentioned earlier. This is a technical task, which will be made with help from Dr. Terray at CERFACS. Test for smaller time periods will be made to ensure that the coupling works satisfactory. It is a complicated task to initialise a coupled model so that unrealistic climate drift is prevented. The main problem is to spin up the ocean model and long integrations are here usually needed to ensure the needed balance. In a RegClim setting, slice experiments with initial conditions from an AOGCM should be made. This means that the AORCM must be started from states predicted by an AOGCM. However, for the ocean part this will be a major task since the resolution in the regional version is considerably higher. The main steps of the initialisation procedure we will employ are as follows: i) A first initial state may be obtained from an ongoing experiment with MICOM, where the model will be run for at least 300 years with climatological atmospheric forcing (repeated annual cycle of monthly means). ii) Use the preliminary initial state and run MICOM with forcing fluxes from a selected year in the AOGCM experiment (from MPI) until a quasi equilibrium state is achieved (probably not more that 100 years of simulation). Again the atmospheric forcing will be monthly mean fluxes applied in a repeated annual cycle. At this stage the initial conditions of MICOM should correspond to the ocean state of the AOGCM at the beginning of the selected time-slice. One may note here that simulating 100 years with MICOM takes about 1 month (wall-clock time) on the SGI-Cray Origin 2000 computer.

Started and will be continued in Phase II.

 

Task 4.4:         Control run with AORCMS on present climate

Internationally, atmospheric climate models are often tested on present climate in so-called AMIP runs (Gates, 1992). The test period is usually 10 years. Coupled models are now also being tested in similar so-called CMIP1 runs (http://www-pcmdi.llnl.gov/covey/cmip/cmiphome.html), where external forcing terms (CO2, solar luminosity, etc.) are kept constant. The ARPEGE model is being tested in an AMIP-type of run for 10 years (task 4.1). Similarly, the MICOM model is being tested in a longer run that also cover the same 10 years (task 2.2). We propose to make a kind of a CMIP1 run with AORCMS as described above. The length of the integration period should be up to 50 years.

For Phase II - Extension:

The results will be evaluated for both atmosphere and ocean and compared with the results of the stand-alone simulations in Task 4.1 and Task 2.2 and with available observations of the atmosphere and ocean. IMR will contribute to this evaluation with emphasis in the oceanic circulation in the Nordic Seas as simulated by the coupled model.

 

Task 4.5:         Future climate simulations with coupling.

We here propose a similar run as in 4.4, with the difference that the initial conditions will be taken from a simulated future state of the atmospheric/oceanic system by an AOGCM. The results will be analysed and compared to the results from the control run (Task 4.4). The results will also be compared to selections of data sets by IMR.

Task 4.5 was prolonged into 2002, due to a half-year delay of the initiation of RegClim Extension of June 1998.

Relations to other Principal Tasks

PT1:    Results from PT4 will be evaluated with the aim of improving process descriptions in the models applied in PT1. Possible improvements of the parameterization of air-sea interaction processes will be implemented in the HIRHAM model.

PT2:    PT 2 will provide the ocean model to be used in PT 4.

PT7:    The ice model in MICOM will be compared to the ice-model developed in PT 7 for use in dynamical downscaling.

PT8:    Data for model validation are to provided by PT 8.

PT9:      Interpretation of model results will be undertaken in close co-operation with PT9.

Milestones

Table 4: Target Schedule PT 4

 

Year

Task

1997

1998

1999

2000

2001

2002

4.1

 

x

x

x

               

4.2

   

x

x

               

4.3

       

x

x

x

x

x

     

4.4

         

x

x

x

x

x

   

4.5

             

x

x

x

x

 

Finalization

                   

x

 

 

Institutions and personnel

The PI, Dr. N.G. Kvamstø, has a new, permanent position at GfI-UiB from 1. June 1999. He will continue as PI for PT4, with about 0.25 person-years per year as a voluntary contribution to RegClim from GfI-UiB. Paul Skeie will be hired on full time on PT4 for the funds granted from RC. All in all, the extra resources needed to educate the new scientist on the job is estimated to be balanced by voluntary contribution by Kvamstø.

Dr. Tore Furevik is presently hired by GfI/UiB until 30.06.1999, thus assuring his participation in RegClim from that date onwards on the extension of PT4 endorsed by the Programme Committee in 1998. His working time will be evenly shared between GfI-UiB and NERSC.

 

Task 4.1-4.2: 

is to be carried out at GfI-UiB by Dr. N. G. Kvamstø and Paul Skeie.

Task 4.3-4.5: 

Dr. T. Furevik will work full time on 4.3-4.5 together with Dr. H. Drange, M.Bentsen, Dr. N.G. Kvamstø and Paul Skeie. IMR will contribute, headed by Dr. Bjørn Ådlandsvik who will work together with Dr. Paul Budgell and Kjell Arne Mork.

Budget for PT4

The following numbers give contributions to the different institutions in units of Person-years, including earlier funds granted.

 

Contributions from the Research Council:

 

                                                2000                2001                2002   

Task 4.1-4.5

            Gfi-UiB                        1.45*               1.4*                 0.15**

NERSC                        0.7                   0.9                   0.15**

            IMR                             0.2                   0.1                              

RegClim Phase II:                     2.35*               2.4*                 0.3**  

*0.05 (2000) and 0.1(2001) person-years are transferred to PT9 (Task 9.1).

**Not formal part of present contract

 

 

5.5     Principal Task 5 (PT5): Indirect aerosol effects.

Principal Investigator: Jón Egill Kristjánsson, IGf-UiO, Oslo

 

Major Objective:

To estimate indirect aerosol effects on the climate of our region, by linking aerosols to cloud microphysics, precipitation and cloud-radiation interaction; and to refine parameterizations of certain cloud-processes .

 

Tasks:

In order to describe the indirect effect of aerosols, a close combination of several model components is needed. In each model grid point the aerosol burden and their composition with respect to water activity must be known. This requires a comprehensive aerosol life cycle model, that includes both chemistry, aerosol physics and transport modules.

 

Within the RC-project "Impacts of anthropogenic aerosols on climate (IAAC)", which ended in 1998, a life cycle scheme for sulfate and black carbon aerosols was developed and tested in a hemispheric-scale off-line transport model (Seland and Iversen, 1999). The scheme is now implemented in the NCAR CCM3 global climate model, and work is under way to refine and improve the scheme, e.g., by including other aerosol types, such as sea-salt and organic carbon.

 

Early work on the indirect effect (e.g., Boucher and Lohmann, 1995) used crude parameterizations to obtain cloud droplet number concentration merely as a prescribed function of sulfate concentrations. More recently, modellers have not only started taking into account other aerosol types, but have paid more attention to source and sink processes involving a size distributed aerosol (e.g., Lohmann et al., 1999). We follow this latter philosophy in PT5 by combining information on aerosol concentrations with assumptions on size distributions of natural and anthropogenic aerosols. Source and sink processes, including humidity swelling are accounted for, and the droplet number concentration is obtained based on Köhler theory (Kirkevåg et al., 1998). We aim at establishing and implementing parametric relations and tabulations which are closer to first principles for aerosol-cloud-radiation effects than those presently in use in climate models, and to run time-slice climate simulations with aerosols and tropospheric ozone as active constituents

By using the new prognostic cloud water scheme, developed by Rasch and Kristjansson (1998) for NCAR CCM3, the droplet number concentration information can be incorporated into cloud microphysics, hence allowing a simulation of "the Albrecht effect" (Albrecht, 1989). In order to treat the other component of the indirect effect, i.e., "the Twomey effect" (Twomey, 1977), droplet sizes computed from the cloud condensation scheme then need to be fed into the radiation scheme.

 

Task 5.1:         Development of a simplified aerosol modelling scheme suited for estimating size segregated particle composition

Such a scheme for sulphate and BC has been developed and tested in the hemispheric CTM, and implemented in NCAR-CCM3. Sensitivity studies for estimating the importance of uncertain parameters have been made. Preliminary model calculations will be made during the development of this task, in order to provide input data which are crucial for other tasks under PT5 and PT6. We continue to build on the competence developed at Igf-UiO in order to include background primary particles and OC in the calculations, and thus develop improved relations between aerosols and cloud properties. One scientist is dedicated to operate the RegClim version of the NCAR-CCM model, but several scientists will work with it in parallell.

Dispersion modelling of the primary background aerosol will also be made as an integrated part of re aerosol modelling. For sea-salt we have available the emission inventory and the sea-salt climatology calculated by Gong et al. (1997 a and b), obtained through co-operative exchange with the Canadian NARCM-project (Northern Aerosol Regional Climate Modelling lead by Len Barrie and Jean-Pierre Blanchet). This is also the basis for emissions used for IPCC TAR scenarioes. Through IPCC emissions of crustal dust aerosols, sulphur compunds, BC and OC are now available.

In progress. Will continue in Phase II.

Publications: 1) Seland, Ø., and Iversen, T.,(1999a); 2) Seland, Ø., and Iversen, T., (1998); 3) Seland, Ø., and Iversen, T., (1999b)

 

Task 5.2:    Parameterization of CCN concentrations for given aerosol loading: Tabulations and implementation in NCAR-CCM3

Assuming that the number-concentration and size-segregated composition are given, we need to develop methods to determine CCN-amounts. First estimates are being made by using BC and sulphate fields and assuming fixed super-saturations for different cloud-types. In order to open for the possibility that the super-saturation is influenced by the aerosol properties, a separate thermodynamic box-model with given aerosols will be developed. Cooling rates will be given based on expected updraft-velocities and turbulent fluctuations in determining variables for different cloud types given in open literature. Tabulations will be implemented in CCM3 in similar ways as tables for optical parameters, or, if suitable, by developing parametric relations.

In progress. Will continue in Phase II.    

Publications: 1) Kirkevåg, A., Iversen, T., and Kristjansson, J. E., (1998); 2) Kirkevåg, A., Iversen, T., and Kristjansson, J. E., (1999)

 

Task 5.3:    Develop methods to incorporate CCN information into the cloud parameterization scheme of NCAR-CCM3

When we know the CCN-concentrations for given possible cloud types (e.g. convective or stratiform), this knowledge must influence the properties of the clouds which are realized in a GCM grid-cell. It is necessary to estimate at least the first two moments of the cloud-droplet distribution, in order to calculate effective optical parameters, and the efficiency of coalescence in water clouds and accretion and riming in mixed-phase clouds. The latter is important in order to evaluate changes in certain precipitation efficiencies (e.g. the Albrecht effect).

In progress. Will continue in Phase II.

Publications: Kristjansson, J. E., Kirkevåg, A., Seland, Ø., and Iversen, T., (1999)

 

Task 5.4:    Investigate need for introducing new prognostic variables (precipitation, snow, ice) or new diagnostic quantities (more detailed treatment of cloud droplet spectra, etc.)

As we seek to account for the effect of anthropogenic aerosols, questions will inevitably arise as to whether the cloud parameterization scheme itself needs to be reformulated. For instance, if release of precipitation were to be considered a crucial process, then it might be necessary to treat precipitating water prognostically, instead of the simpler diagnostic formulation used at present.

In progress. Will continue in Phase II. Publications to come.

 

Task 5.5:    Develop SCM for cloud microphysics investigations

The development work described under subtasks 5.1-5.3 above will involve running a lot of test simulations. It is essential for this test phase to have efficient tools, such as a 1-dimensional (vertical) model, to accompany the more cumbersome 3-dimensional GCM. This single column model (SCM) will probably be a 1-dimensional version of the cloud microphysics module in the GCM.

Finished. Technical task, not to be published.

 

Task 5.6:    Sensitivity experiments in SCM and GCM.

With the SCM and GCM model tools in place, code development can be done efficiently, e.g. through sensitivity experiments. Such experiments will be important during the development phase as well as for production of results in later phases of the project

In progress, will continue in Phase II. Technical task, not to be published.

 

Task 5.7:    Derivation of cloud optical properties due to aerosols

This is the so-called Twomey-effect. In Task 6.6 (PT 6) tabulations of aerosol-optical parameters are made. To start with it is necessary to link the radiation-effective droplet size with the average size of cloud-droplets, which are presently used parameters in CCM3. Depending on the findings in 5.3, tables similar to those for optical properties of aerosols are planned. However, effects on long-wave terrestrial radiation should also be investigated and possibly parameterized. Tests of the parameterizations will be made by comparison with exact calculations in the radiative transfer model of NCAR-CCM3. Alternative experiments with the DISORT model (Stamnes, 1988) are also planned.

In progress. Will be finished in Phase I.

Publications: 1) Kirkevåg, A., Iversen, T., and Dahlback, A.,(1999).

 

Old Task 5.8:  Unified convective scheme

Nordeng’s (1994) approach is coded in the ECHAM4-physics which is part of the HIRHAM model, but only for deep convection. In this work we will make a unified scheme where these ideas are implemented also for shallow convection. The scheme will be incorporated in HIRHAM and NCAR CCM3.

 

Old Task 5.9:  Shallow convective closure

The knowledge of how to close shallow convection (how to relate convective activity to large scale, resolved flow) is poor. In HIRHAM shallow convection is closed by assuming a stationary boundary layer which clearly is a poor approximation when there are large scale temperature and moisture changes as in e.g. cold air outbreaks. In this task we will investigate and test if there are alternative closures. This task is closely related to Task 5.7.

Old Tasks 5.8 and 5.9 are finished and a report is being prepared.

The results were however not so conclusive that the revised convective scheme will be put into the HIRHAM model for scenario runs yet. The work will however continue under PT 1 (task 1.5) as a contribution in kind from DNMI.

 

New Task 5.8: (common with PT6:  task 6.7):           Interactive modeling of the total effects of aerosols in AGCM

This Task is common and the same for PT5 and PT6.

The emphasis in these AGCM experiments is on the dynamic effects of the regional radiative forcing patterns that result from the direct radiative forcing of particles, as well as from the modification of the cloud fields caused by aerosols. For the latter, at first simple parameterizations can be used to study the sensitivity of the model to changes in the effective radius of water clouds. A special study of the indirect aerosol effects will also be considered. Finally, we will combine the modules developed in PT 5 and PT 6 for computing radiative forcing due to both direct and indirect effects of aerosols. The forcing will now be calculated in full interaction with the model dynamics of the AGCM, so that the response of the forcing will be calculated in terms changes in the pressure, wind and temperature fields etc. The results of these experiments will be investigated in close co-operation with Task 9.6 (PT 9).

 

Not started. Will continue in Phase II. Publications to come.

 

Relations to other Principal Tasks

PT6:   PT5 and PT6 are closely linked, and end up with a common Task 5.10/6.7. Moreover, experiments under Task 5.6 will be co-ordinated with Task 6.6. Task 5.7 will be completed in co-operation with PT6.

PT8:    Data for validation of the SCM are to be provided by PT 8.

PT6 and PT9:

             Interpretation of model results will be undertaken in close co-operation with PT6 and PT9.

 

Milestones

Table 5: Target schedule PT 5

Task

Year

 

1997

1998

1999

2000

2001

2002

5.1

(x)

(x)

(x)

x

x x x x x x    

5.2

 

(x)

(x)

(x)

x

x

x

 

       

5.3

     

(x)

x

x

x

x

 

     

5.4

     

(x)

x

x

x

x

 

 

   

5.5

     

(x)

x

 

 

 

 

 

   

5.6

      x x x x x x x    

5.7

  x x x x x            

5.8/6.6

          x x x x x x  

Finalization

                   

x

 

(x) = activities under an earlier project (referred to as IAAC in Iversen et al. 1997) which was prolongated in RegClim.

Institutions and Personnel

Task 5.1:         To be carried out at IGf-UiO by Ø. Seland, with own contribution from Prof. T. Iversen and Dr. J.E. Kristjánsson.

Task 5.2, 5.3, 5.5:       

                        To be carried out at IGf-UiO chiefly by A. Kirkevåg, with own contributions from Prof. T.Iversen and Dr. J.E. Kristánsson.

Task 5.4:          To be carried out with own contributions from Igf-UiO by Dr. J.E. Kristánsson.

Task 5.6:          To be carried out at IGf-UiO by A. Kirkevåg and Ø. Seland, with own contributions from Prof. T. Iversen and Dr. J.E. Kristjánsson.

Task 5.7:          To be carried out partly at IGf-UiO by Alf Kirkevåg.

Task 5.8 / Task 6.6:

            The work is to be carried out at IGf-UiO by Ø. Seland and Dr. A. Kirkevåg, with own contributions from Prof. T. Iversen, Dr. J. E. Kristjánsson and at NILU by Prof. F. Stordal .

 

Budget for PT5

The following numbers give contributions to the different institutions in units of Person-years. New funding in RegClim PHASE II is given in bold types.

 

Contributions from the Research Council:

 

                                                            2000                2001                2002   

Task 5.1 –5.7

            IGf-UiO                                   1.4                   1.05    

Task 5.8/6.6  (PT5-part)

            IGf-UiO                                   0.4                   0.95                 0.25*  

RegClim Phase II                                 1.8                   2.0                   0.25*                          

*Not formal part of present contract.

 

 

Own funding

The Department of Geophysics, University of Oslo, will also contribute through the work of Jón Egill Kristjánsson and Trond Iversen. This amounts to 0.3 Person-years per year.

 

5.6     Principal Task 6 (PT6): Direct climate effects of regional contaminants

Principal Investigator: Frode Stordal, NILU, Kjeller

 

Major Objective:

            To calculate the regional distribution of direct radiative forcing in our region due to tropospheric ozone, sulphate, organic aerosols and black carbon, and to estimate the effects on the climate of our region.

 

Tasks:

Work will be mainly based on existing model tools, which have already been developed by or used by the participating groups. Such models include Chemistry-Transport Models (CTMs), Radiative Transfer Models (RTMs), an Atmospheric General (global) Circulation Model (AGCM) and an Atmospheric Regional Climate Model (ARCM). The studies based on CTMs will be process oriented. Such studies are needed as there are still significant uncertainties in many of the processes leading to changes in the concentrations of ozone and aerosols and in the associated radiative forcing. The ARCM studeis are included in order to study coupled chemistry-radiation-dynamics effects-

 

Three different CTMs will be used. A hemispheric model has already been in use for some time to study the cycling of sulphur and BC components from SO2 and BC emissions (CTM-A, Iversen et al., 1997). The second CTM (CTM-B, Berntsen and Isaksen, 1997) has been used to study formation of ozone as well as sulphate particles. An improved version of this model, with improved transport (e.g. convection), will be used in this subproject (CTM-C, Sundet, 1997. Radiative transfer modelling in the longwave region will mainly be based on a broad band model (BBM, Myhre and Stordal, 1997). Some comparative experiments will be made also with a line-by-line (LBL) model (Edwards, 1992). In the shortwave region, models based on the discrete ordinates method (DISORT) will be used (Stamnes et al., 1988). The AGCM model that will be used in this subproject to study effects of regional climate forcing, is the NCAR CCM3.

 

In the following, the status of the work under each of the sub-tasks is described, and an outline of the work to be carried out in phase-II is presented.

 

Old Task 6.1:  Establish emissions of ozone precursors

Current emissions of NOx, NMHC and CO will be estimated. Since there are free tropospheric interactions through transport over large regions in the Northern Hemisphere, particularly at mid-latitudes, data from areas outside Europe will also be used. Emission data will be taken from the Global Emission Inventory Activity (GEIA, IGAC), the European Evaluation and Monitoring Program (EMEP), and the European Environmental Agency (EEA). Future scenarios will also be established, taking into account current and planned international conventions regulating the emissions of NOx and NMHC and also expected development in industrial technologies and transportation. Collaboration will be sought with organizations actively working on future emission scenarios, in particular IIASA. The work in this task is completed. In phase II emissions will be taken directly from GEIA, IPCC scenarios and other sources.

The Task is therefore not included anymore, and the remaining tasks are renumbered

 

New Task 6.1: Modelling regional distribution of ozone and the resulting radiative forcing

Transport of ozone and its precursors can take place from the boundary layer to the upper troposphere in regions with efficient convective transport. This can affect the chemical distribution and the ozone concentrations in the upper troposphere (Berntsen et al., 1996), with influence on the regional radiative forcing. CTM-B will be used to study ozone generation and estimate contributions from different chemical sources, e.g. in situ production due to processes occurring in the free troposphere due to transport from polluted regions, NOx production from lightning, as well as ozone transport from different regions; transport from the surface boundary layer and tropospheric transport which can occur over large distances (Berntsen et al., 1996). Ozone distributions calculated in the CTM(s) will be used in calculations of regional radiative forcing with the BBM model in the longwave and the DISORT model in the shortwave region.

Calculations of changes in ozone since preindustrial times have been performed along with calculations of the associated radiative forcing. Two studies have been completed, one comparing the present atmosphere with the preindustrial one, and one where the time evolution has been studied, in terms of time slices.

In Phase II only CTM-C will be used for studies of radiative impacts of ozone changes. An improvement is that calculations will be made in a high resolution; some experiments will be performed in a T42 resolution, a few cases in T63 will be considered. Increased resolution is considered necessary to describe regional effects, especially due to non-linearity in the ozone chemistry. We will focus on past changes in ozone as well as possible future ozone changes, taking also into account two effects of changes in the stratospheric ozone concentrations; increased penetration of UV radiation to the troposphere and subsequent impact on the photochemistry, and reduced fluxes of ozone from the stratosphere to the troposphere. 

 

New Task 6.2: Modelling regional distribution of aerosols and the resulting direct radiative forcing

Likewise, CTM-A as well as CTM-B or CTM-C (possibly also CTM-D) will be used to estimate distributions of sulphate concentrations which are formed in the atmosphere as a result of man-made and natural emissions of SO2. These calculations will be followed by calculations of the direct radiative forcing due to aerosols, based on the radiation code of NCAR CCM3 as well as the DISORT model.

Several calculations of the direct radiative effect of aerosols have been performed, based partly on experiments with CTM-A and CTM-B, partly on an assumed particle concentration in a transect through Europe. Sulphate and black carbon have been considered.

In Phase II, we will continue process studies using the same approaches, now with a main focus on the CTM-C. In addition to calculating new concentration fields of sulphate based on improved resolution in CTM-C, we will include organic aerosols. The model is particularly suited for studies of organic compounds as it contains a full ozone chemistry including several organic gaseous compounds.

 

New Task 6.3: Development and implementation of simplified chemistry schemes

The chemical module used in the 3-D CTMs used for ozone studies has an extended chemical scheme, which includes approximately 50 chemical compounds involved in the tropospheric ozone chemistry. A simpler scheme will be developed. This work includes reducing the number of compounds that are transported. Preliminary studies indicate that it is possible to reduce the number of transported compounds to around 15 within acceptable uncertainty limits. At present a version of the simplified ozone scheme has been developed. The simplified scheme will be implemented in the AGCM.

In Phase II, it is applied for another 0.5 person-years in 2000 and 0.3 person-years in 2001 to cover up this very important part of the project, which so far has been covered by own funding and synergy with other projects (Igf-UiO). The simplified scheme will be implemented in the AGCM under Task 6.5 in phase-II. Based on the experience in that development, we will consider making modifications to the simplified scheme.

 

New Task 6.4: Validation and possible improvement of radiation schemes in AGCM

The models used to calculate radiative effects of ozone and particles need further validation and possibly improvements. Shine et al. (1995) have performed a comparison of radiation codes used to estimate the climatic effects of ozone. An idealised set of assumptions will be made and the AGCM radiation schemes and the RTMs of this subproject will be compared to their results. In the comparison the optical parameters, albedo and zenith angle will be varied. If the results of the testing of radiation schemes show unsatisfactory accuracies in the AGCM radiation schemes, efforts will be made to improve the schemes.

Comparisons have been made between a broad band scheme for IR radiation. For solar radiation various methods have been intercompared.

In Phase II, this work on testing and validation of schemes for calculation of optical properties of aerosols and for radiative transfer will be continued also in phase-II.  There are still significant uncertainties, in particular regarding optical properties of aerosols that need further investigation. E.g. we will examine the role of various assumptions on mixing of different types of particles.

 

New Task 6.5: Interactive modelling of particles (direct effect) in AGCM

Coupled chemistry-radiation-dynamics calculations in the AGCM will be performed. Separate runs will be made to study the effects of ozone and the direct aerosol effect. Several types of aerosols are considered: primary background, sulphate, and organic aerosols. Based on detailed radiation calculations a parameterisation scheme will be developed for the radiative properties of the two types of aerosols (tabulations). This scheme has to be economical enough to be run at a reasonably low computational cost in the AGCM. Under this task a large effort has been put in preparations for inclusion of aerosols in the AGCM. At present aerosol tabulations are finished and tested off-line. The integration of ozone in the AGCM will be initiated.

In Phase II, The work in this task will be an important part in addition to the process- oriented work in the above tasks. Recent GCM calculations have shown that ozone and aerosols may have had a substantial impact on the climate change up till now compared to the impact of the well-mixed greenhouse gases. The plan is still to perform separate experiments for ozone and aerosols, as it is important to understand the coupling in both cases. There are some important differences between the two cases, both in the regional patterns and (may be more importantly) in the altitudinal distributions of the implied radiative impact. Interpretational work will be performed in co-operation with Task 9.6 (PT 9).

New Task 6.6: (common with PT5:  task 5.8):           Interactive modeling of the total effects of aerosols in AGCM

This Task is common and for both PT5 and PT6.

We here combine the modules developed in PT 5 and PT 6 for computing radiative forcing due to both direct and indirect effects of aerosols. The forcing will now be calculated in full interaction with the model dynamics of the AGCM, so that the response of the forcing will be calculated in terms changes in the pressure, wind and temperature fields etc. The results of these experiments will be investigated in close co-operation with Task 9.6 (PT 9).

 

Milestones

Update Table 6: Target schedule PT 6

Task

Year

 

1997

1998

1999

2000

2001

2002

6.1

 

x

x

x

    x x x      

6.2

 

x

x

x

x

x

x

x

x      

6.3

  x x

x

 

x

 x

 x

x

     

6.4

    x

x

 x

 x

 x

x

 

   

6.5

    x

x

x

x

x

x

 x

 x

   

6.6/5.8

          x x x x x    

Finalization

                   

x

 

Institutions and Personnel

Task 6.1:         The work is to be carried out at IGf-UiO by a Ph.D. student or a Post.doc under the supervision of Dr. I. S. A. Isaksen.  At NILU Dr. F. Stordal will be responsible for the work, which will be carried out also by other NILU scientists.

Task 6.2:         The work is to be carried out at IGf-UiO by Ø. Seland and A. Kirkevåg (and with own contributions from Dr. I. S. A. Isaksen, Dr. T. Iversen and Dr. J. E. Kristjánsson). At NILU Dr. F. Stordal will be responsible for the work, which will be carried out also by other NILU scientists.

Task 6.3:         The work is to be carried out at IGf-UiO by Dr. Terje Berntsen and Dr. Jostein Sundet under the supervision of Dr. I. S. A. Isaksen.  At NILU Dr. F. Stordal will be responsible for the work, which will be carried out also by other NILU scientists.

Task 6.4:          At NILU Dr. F. Stordal will be responsible for the work, which will be carried out also by other NILU scientists (e.g. Dr. Gunnar Myhre).

Task 6.5:         The work is to be carried out at IGf-UiO by Ø. Seland and A. Kirkevåg (particles) (and with own contributions from Drs. T. Iversen and J. E. Kristjánsson), and Dr. Terje Berntsen and Dr. Jostein Sundet (ozone) (under the supervision of I. S. A. Isaksen). At NILU Dr. F. Stordal will be responsible for the work, which will be carried out also by other NILU scientists.

Task 5.8 / Task 6.6:
The work is to be carried out at IGf-UiO by Ø. Seland and Dr. A. Kirkevåg, with own contributions from Prof. T. Iversen, Dr. J. E. Kristjánsson and at NILU by Prof. F. Stordal.

 

Relations to other Principal Tasks

 

PT 5:    PT5. and PT6. are closely linked. Task 6.7 is the same as Task 5.10. Task 6.6 will be co-ordinated with Task 5.6.

PT 8:    Data for validation of CTMs are to be provided by PT 8.

 

Budget for PT 6

 

The following numbers give contributions to the different institutions in units of Person-years.

 

                                                                            2000                2001   

Task 6.1-6.2,  6.3 and 6.5(ozone)

IGf-UiO                                               0.2      

IGf-UiO                                               0.3*                  0.5*    

IGf-UiO                                               0.15                 0.15                

NILU                                                   0.1      

Task 6.5(particles)

IGf-UiO                                               0.2

Task 6.4

            NILU                                                   0.9                   0.9      

RegCim Phase II                                              1.85                 1.55    

*transferred from Task 6.6/5.8(PT6-part) as a consequence of reduced budget.

 

Own Contribution

Task 6.6/5.8 (PT6-part) The resources allocated to this are now completely under PT5. Any contribution in the interpretation phase will be taken through synergy with Task 6.5.

 

5.7     Principal Task 7 (PT7): Air -ice-ocean interface processes and sea state modeling

Principal Investigator: Lars Petter Røed, DNMI, Oslo

 

Major Objective:

            To develop suitable dynamic boundary conditions for dynamical downscaling, to estimate the effect of atmospheric anamolies on the ocean circulation and hydrography, and to model the Nordic Seas climate with respect to storm surges and waves.

 

Tasks:

Sea ice

One of the major problems in climate modelling is the inclusion of ice in the coupled climate system. This is due to the sensitivity of the climate system to the extent of the ice cover (dramatic changes in albedo and fluxes), and is compounded by the sensitivity to melting and freezing conditions. The Nordic Seas is an ice infested sea and the ice extent and the location of the ice edge is of vital importance for the regional as well as the global climate. The aim here is to provide a verified (but not necessarily validated) state-of-the-art sea ice model siutable for coupling to the primary and secondary ocean model of PT 2, as well as a slab ocean model.

 

Coupled ice-ocean models

The results of validation experiments utilizing the coupled ice-ocean model system, consisting of DNMI's POM version and the improved Hakkinen-Mellor-Kantha ice model, has shown that the ice extent and ice thickness distribution does not validate properly. In particular it reveals that the ice edge is too diffuse, and that the ice thickness tends to build up in areas where it is not supposed to, e.g., close to islands and other obstructions (Sætra et al., 1998, Røed et al., 1999). To remedy this several improvements to the ice model have been suggested, among them the implementation of a positive definite advection scheme to remedy the tendency of a too diffuse ice edge, and a different rheology (elastic-viscous-plastic or EVP rheology) to destiffen the ice motion. The advection scheme is already in place (Røed et al., 1999), while the EVP rehology wil be implemented in the near future.

 

Dynamic boundary conditions

In the dynamical downscaling experimants in PT 1, detailed atmospheric structures are interpreted. In order to possibly improve the simulated atmospheric state, chosen time-slices of 5 years in a coupled simulation with the ARCM, HIRHAM are run and analysed. We aim at coupling HIRHAM with a simple ocean model (e.g. a 1 1/2-layer MICOM version or slab ocean model) to allow HIRHAM to develop its own SST and ice/snow cover climate over the ocean surface part of its domain. This is consistent with what is done over land surfaces. In Sweclim it has been shown imperative that the downscaling model has the power to influence the ice-cover in particular, since the interpolated fields from the AOGCM are too coarse.

 

Sensitivity of the Nordic Seas circulation and hydrography to forcing anamolies

The Norwegian Atlantic Current (NAC) flowing along the continental slope west of Norway is the main reason for the relatively mild climate experienced in continental Norway (Aas, 1994). Hence any change in this current, either in its volume flux or heat content, may significantly influence the continetal Norwegian climate. Here we aim to enhance our understanding of the driving mechanisms for the NAC. Is it purely wind-driven, or purly driven by thermohaline forces such as deep water formation? What is the influence of a change in the freshwater discgharges? In these studies use will be made of the coupled ice-ocean model system consisting of DNMI's POM version and the improved Häkkinen-Mellor-Kantha ice model.

 

Air-sea interaction processes

Considered is the ocean surface flux parametrizations for momentum, heat and moisture in the HIRHAM model used in RegClim. The investigation will be based on the hypothesis that the wave-age dependent momentum and heat flux may have a strong impact on the development, life cycle and intensity of small scale atmospheric instabilities (polar lows etc.). For these phenomena, the instability is to a large extent governed by the heat and moisture flux from the ocean, and to a less extent by the baroclinicity. Earlier investigations have shown that parameters such as wind at 10m height and sea-state may be significantly influenced by taking this effect into account (Lionelli et. al. 1998, Janssen 1994). These parameters are key climatic parameters in RegClim.

 

We will focus on air-sea interaction processes that are important when modelling regional climate at high latitudes with high resolution. This region is characterized by large thermal contrasts between ice-covered and open sea. These contrasts have significant impacts on the atmospheric flow pattern. Here, we will study how phenomena that are governed by the heat flux from the ocean, may be affected by a wave-age dependent sea-surface roughness. Such phenomena are related to shallow convection (studied under PT5 and in Phase II in PT1), including polar lows, which are strong small-scale vortices. The evolution of polar lows is controlled by the air-sea heat- and moisture fluxes (Økland and Schyberg, 1987).

We will use the ocean surface flux formulations which are part of the atmosphere models used in RegClim, and those will be evaluated in such situations as discussed above, in particular the evolution of polar lows and arctic air outbreaks. We will address the effect of neglecting the wave-age dependence of momentum, heat and moisture on climate simulations. If there turns out to be major deficiencies, we will address the necessity for coupling with an ocean wave model. The flux parametrizations are also important for computing climate parameters such as surface temperature, 10 m wind, and the ocean state. These are all key climate parameters in the RegClim.

To be able to study the wave-age dependent surface roughness, coupling between the atmospheric circulation and ocean wave model is necessary. Earlier investigations have been carried out using coupled atmosphere/wave models for studying both idealized cases (Lionelli et.al 1998) and for more realistic situations (Doyle 1994, Janssen 1994). In these studies, the horizontal grid resolution have been relatively coarse (>50 km), and they have mostly focused on the atmospheric dynamics in the mid-latitudes. Although there is some controversy on the importance of the effect, they all seems to agree that: (i) the coupling has a significant effect on the development of the ocean wave field, wind at 10 m height and the momentum transfer; (ii) it has only a minor effect on the development and life cycle of the mid-latitude baroclinacally unstable low pressure systems; (iii) the effect is stronger for smaller scale atmospheric structures that appear when the horizontal grid resolution increases. Since we intend to investigate this effect on small scale atmospheric phenomena, a horizontal resolution of < 10km may be necessary.

The ultimate goal is to develop improved parameterizations of air-sea fluxes which can be used in other models in RegClim such as HIRHAM and ARPEGE.

Observations and validation

Since very few observations of surface fluxes are available, the surface flux parameterization must be evaluated indirectly. An important tool will be the ability to use satellite observation in a coupled objective analysis system for ocean surface wind and sea-state. In this system we will assimilate satellite altimeter observations of wind speed and wave height from ERS and ENVISAT satellite radar altimeters. DNMI has a large database of ERS altimeter and scatterometer observations back to 1992 (Breivik and Reistad, 1994) and we aim at processing observations from ENVISAT from 2000.

The analysis system is a further development of the currently operational altimeter wave data assimilation system at DNMI, and by using a coupled system that lets observations simultaneously affect both wind field and wave field the observations will be utilized in a more optimal way.

The optimal surface analyses produced this way can be used as an initial state of wind field and sea-state for model runs of interesting cases with different formulations for the surface stresses. The prognoses using the various surface flux parameterizations can then be evaluated against satellite observations in the area.

Task 7.1:    Development of a sea-ice model

The present ice model coupled to POM is a reduced version of the full thermodynamic-dynamic, sea-ice model of Häkkinen and Mellor (1990, 1992). In this task it will be extended to the full model and its performance verified against simple idealized cases (Røed, 1995).

This task is completed. The Hakkinen-Mellor-Kantha ice model has been implemented and its performance verified (Sætra et al., 1998). The ice model has been improved upon by implementing a new positive definite advection scheme The model code has been made available to NERSC.

 

New Task 7.2: Coupled ice-ocean model (former 7.2 and parts of 2.2)

In this task the full ice model will be coupled to POM and MICOM and its performance verified against simple test cases. These test cases will focus on the sensitivity of the location and extent of ice cover to various forcing scenarios. If possible the Häkkinen and Mellor ice model will be compared with the present ice model coupled to MICOM.

This task is partly completed. The sea-ice model developed in Task 7.1 has been succesfully coupled to DNMI's POM version, and its performance has been verified against benchmark simulations kindly made available to the project by Dr. Sirpa Hakkinen. However, as explained above (Røed et al. 1999) the present ice rheology makes the ice motion too stiff. This task will therefore be prolonged, but no cost extension is asked for. Although not originally applied for, it is proposed to bring IMR into this task.

 

New Task 7.3: Technical Coupling of HIRHAM and WAM. (former 4.7)

The coupling between the atmosphere model HIRHAM and the wave model WAM will be carried out using the OASIS software. The model will be set up to run in parallel mode on the Cray T3E computer at NTNU. In this task the model domain will be an idealized zonal channel with cyclic boundary conditions.

Various formulations for the air-sea fluxes will be implemented. In particular, we will study the existing models for calculation of the wave induced stress from the ocean wave spectrum.

Will be initiated during the fall of 1999, To be continued in PHASE II

New Task 7.4: Air-Sea Interaction Case studies (former 4.8)

Weather charts and literature will be examined and a suitable situation will be selected for a case study. Here, we will focus our attention on situations with strong thermal differences across the marginal ice zone, and the subsequent formation of atmospheric instabilities (polar lows etc.). The case studies will be carried out using the coupled atmosphere/wave model and the coupled system for data assimilation. The prognoses using the various surface flux parameterisations can then be evaluated against satellite observations in the area.

The models for air-sea fluxes will be evaluated, and if possible we will suggest improvements of these models.

To be developed as a part of PHASE II

 

New Task 7.5: Sensitivity of the Norwegian Atlantic Current to anomolous forcing (former parts of 4.5)

DNMI's POM version coupled to the improved Hakkinen-Mellor-Kantha ice model will be run for various forcing scenarios for time slices of 5 - 10 years and results compared to a present day time slice run with the ECMWF reanalyzed forcing fields (ERA 15) combined with their operational analysis.

Initiated (see Røed et al., 1999). To be continued as part of PHASE II

 

New Task 7.6: Dynamic boundary conditions

This task is fully justified in section 5.1 under Principal Task 1: Dynamic downscaling. The aim is to develop the necessary slab ocean tool for coupling to HIRHAM, and to couple it to HIRHAM. Essentially the slab ocean model (or mixed layer model) will be a simplified 1 ½-layer version of MICOM and/or OSMOM (Røed, 1995) coupled to the improved ice model developed in Task 7.1 and 7.2. The coupling tool will be the OASIS program. The work will be performed in close collaboration with scientist working on PT 1..

To be developed as a part of PHASE II

 

New Task 7.7: Storm surge and wave statistics (former 2.3)

This task is an end product since the results will not influence the work to be carried out under other tasks. The standard wave and storm surge models at DNMI (Engedahl, 1995; Sætra and Reistad, 1997) will be run to produce the necessary quantitative statistics on wave height and sea surface elevation. The driving meteorological fields will be taken from results of PT 1.

To be developed as a part of PHASE II

 

Relations to other Principal Tasks:

PT 1:   Task 7.4 depends on results from PT 1. Task 7.6. is a prerquisite for PT 1. Also the coupling in Task 7.6 will be part of PT 1. Task 7.4 will be co-ordinated with the investigations of shallow convection in New Task 1.5.

PT 2:   The ice model and ocean-surface parameterisations developed under PT 7, i.a. for the POM model, will be compared to similar processes in MICOM run under PT 2.

PT 8:    Data for model validation are to be provided by PT 8.

PT9:     Interpretation of model results will be undertaken in close co-operation with PT9.

 

Milestones:

Task 7.4 was delayed 0.5 years in 1998

 

Table 7: Target schedule PT 7

Task

Year

 

1997

1998

1999

2000

2001

2002

7.1

 

x

                   

7.2

   

x

x

x

x

x

         

7.3

         

x

x

x

x

x

   

7.4

               

x

x

x

 

7.5

         

x

x

x

x

x

   

7.6

         

x

x

x

       

7.7

               

x

x

   

Finalization

                   

x

 

 

Institutions and Personnel

Task 7.1-7.2, and 7.6-7.7:
The work is to be carried out at DNMI by Prof. L. P. Røed (PI), Dr. A. Melsom, Dr. Ø Sætra, and Dr. H. Engedahl,  Dr. P. Budgell at IMR will be consulted. Other scientific personnel at DNMI will be used if necessary.

Task 7.3-7.4:   The work will be carried out at DNMI in Oslo by Dr. Ø. Sætra, Dr. J.E. Haugen, H. Schyberg and L.A. Breivik.

Task 7.5          The work is to be carried out at DNMI, and will be headed by Dr. H. Engedahl. He will work in close collaboration with Dr., Ø. Sætra. Also the remaining modeling team at DNMI will be involved in this task.

 

Budget:

The following numbers give contributions to the different institutions in units of Person-years.

 

Contribution from the Research Council:

 

                                                2000                2001                2002

Task 7.1

Finished           

Task 7.2

DNMI:                         0.1         

Task 7.3-7.7

DNMI                          1.5                   2.3                   0.2*

IMR                             0.2                   0.1

RegClim – Phase II:                  1.8                   2.4                   0.2*

*Not formal part of present contract.

 

 

5.8     Principal Task 8 (PT8): Data for model evaluation

 

Principal Investigator: Knut Arne Iden, DNMI, Oslo.

 

Major Objective:

To establish links to data-sets for model-validation and to maintain an information page on availability of different data for model-validation in the project.

 

Tasks:

The proposals for the other PT’s contain a paragraph where data relevant for model evaluation are listed. Further systematisation will be based upon information from PIs of the relevant principal tasks.

It is intended on behalf of RegClim to encourage the National Climate Service Centre (NoSerC) to take on some of the responsibility to provide observation data to modellers. One important aspect in this connection is to establish agreements with other institutions in Norway and abroad (IMR, NILU, NERSC, Norwegian Polar Institute, and climate centres abroad) that have the wide range of data for comparison with climate models. On behalf of RegClim, the PI of PT8 will be contact person in these data matters with NoSerC.

Task 8.1:    Organize meteorological climate data for Northern Europe for model validation

In addition to data in Task 3.1, the following data for model evaluation are available at DNMI:

Upper Air Observations:

  • 6 hourly radiosonde data from station «M» from 1961, 12 hourly 1949-1960.

  • Controlled radiosonde data from Europe (1980-1990).

  • ATOVS indirect measurements of humidity and temperature through radiants. These measurements are available from spring 1998 and give better spatial coverage and more accurate information about the humidity profile.

Meteorological fields:

  • Gridded 6 hourly sea level pressure, wind speed (u,v), significant wave height (Hm0) (Hindcast arhive, DNMI). Period 1955-1998, Area : 50°- 90°N, 40°W-40°E, Grid 75 km

  • Weekly sea ice concentration maps (1966-1993) with SST included from 1972. Area covered by the mapping varies.

  • Time series of wind speed, wind direction, total sea, wind sea and swell for 100 grid points in the Hindcast archive (Period 1955-1998).

Historical observations from ships and maritime stations with fixed positions:

  • 3 hourly values of standard meteorological parameters including significant wave height Ship «M» (1949-), «Ami» (1976-1984), «Famita» (1959-1978 (winter seasons)), Ekofisk (1980- ), Frigg (1978-), Statfjord A (1978-1989), Gullfaks C (1989-), Sleipner A (1994-), Heidrun (1995-), Draugen (1995-)

The data mentioned are available at DNMI, but further systematization may be needed for  use in the different PT’s. The activities within this activity will be co-ordinated with the new NFR project NorSerC.

 

Task 8.2:    Organise oceanographic data aimed at model validation

Hydrographic data for the Nordic Seas are available from the CTD database of IMR and World Ocean Atlas 1994. Transport estimates through various vertical sections are found in the literature or estimated directly from current measurements. Suitable ice data sets are available (Johannessen et al., 1996; Johannessen et al., 1995; Walsh and Johnson, 1978).

 

The transport estimates and gridded ice data sets are useful for validation as they are. For hydrography, a set of standard vertical sections will be determined. CTD-data from these sections will be put into a standard format for easy comparison with model results. Where available, hydrographic time series will be made accessible.

This part of the activity was finished at IMR in 1998.

 

Extended in Phase II (Dec00):

For comparison with model results, the hydrographic data from IMR’s standard Norwegian Sea sections (Svinøy-NW, Gimsøy-NW, Bjørnøya-W) will bee quality controlled and gridded. The results will be saved in a standard netCDF format. Statistical products, such as climatological means and variance estimates will also be computed. This can be used to make weighted discrepancy measures between model results and observations.  Similar gridding in the horizontal will be provided for selected synoptic cruises. The gridded data products will be made available within RegClim, and links will be provided to the www.info.sheet , and/or NoSerC (re. Task 8.5).

Task 8.3:    Organise air quality data aimed at model validation

There is only a limited number of observations of free tropospheric ozone. Measurement data for ozone and related species will be collected and used in evaluation of CTMs used in this subproject. Data will be included for instance from aircraft measurements (e.g. the MOZAIC project) and from ozone sondes which are performed on a routine basis at a few stations over Europe and during intensive campaigns over limited time periods (e.g. SESAME). Such observations have shown that there are large temporal and spatial variations in ozone and related gases, and that there are large regional areas which are affected by the emission of pollutants from industrial activity (WMO, 1995).

Pointers to these data will be established within 1999 and the task finished. The resouces to NILU (0.1 person-years) are transfered to PT 6 for Phase II

 

Extended in Phase II (Dec00):

Modeling of aerosols is one of the elements in PT5 and PT6. On one hand the distribution of sulphate aerosols is modelled in a CTM as well as a GCM, on the other hand radiative transfer models calculate the impact on the radiative forcing, and finally the climate impact will be estimated by a GCM. Several observations have been made of relevant aerosol parameters, both in long-term observation networks and in measurement campaigns. It is the aim of this task to put together a dataset of aerosol data for validation of the models used in RegClim. The main emphasis will be put on data covering the region of primary interest in RegClim.

The parameters that are of interest are total mass, size distribution and chemical composition for validation of calculated concentration distributions. To validate and yield improved input to radiative transfer models optical properties are needed, such as optical depth, extinction coefficient, single scattering albedo and asymmetry factor, or related properties.

With the limited funding within  RegClim, we will focus mainly on satellite data. Optical depths will be derived from AVHRR (e.g. Nakajima and Higurashi, 1998) and POLDER (e.g. Deuzé et al., 1999). In addition we will use data from AERONET (Holben et al., 1998), which is a groundbased aerosol monitoring system that offers a standardisation for a regional to global scale monitoring and characterisation network. 

The responsible scientist at NILU will be Dr. Mihalis Lazaridis. The work in this task will benefit from NILU’s participation in the project Synthesis of Integrated Global Aerosol Datasets (SINGADS) financed by the EC.

Technical assistance will be sought from NoSerC.

Task 8.4:    Compile sufficient metadata for the data types appropriate for the users

Metadata (i.e. information concerning the data) for the various datasets will be supplied from Tasks 8.1-8.3.

Metadata are presented on the www-info-sheet and are updated as use of the different datasets reveals weakness and/or strength. The activity will be co-ordinated with the new NFR project NoSerC.

 

Task 8.5:    Establish and maintain www-info-sheet

The info-sheet will contain information on:

·      what data are available;

·      where these data are stored;

·      who to contact;

·      data format.

The WWW-info-sheet is established and are regularly updated (http://projects.dnmi.no/~regclim). The activity will be co-ordinated with the new NFR project NoSerC.

 

New Task 8.6: Ice concentration maps and other relevant sea ice information

a) NERSC.

The sea ice satellite sea data are obtained from the Scanning Multimeter Microwave Radiometer (SMMR) carried onboard the Nimbus 7 satellite, and from the Special Sensor Microwave/Imager (SSMI) onboard the Defence Meteorological Satellite Program (DMSP) satellites F8, F11 and F13. The Nimbus 7 satellite operated from the October 26, 1978, to the August 20, 1987. The DMSP satellites together provide data from July 9, 1987, to present.

The NORSEX algorithm (Svendsen et al., 1983) will be used to calculate sea ice concentrations from SMMR/SSMI individual channels. This algorithm uses sensor channels 18 GHz V and 37 GHz V for SMMR, 19GHz V and 37GHz V for SSMI and monthly average atmospheric surface temperature to calculate sea ice concentrations. The algorithm uses values for emissitivities and atmospheric opacities taken from measurements in the Arctic (NORSEX Group, 1983). Weekly means of the sea ice extent, concentrations and age (i.e., first year versus multi year ice) will be provided on a standard grid.

b) DNMI.

 Regional weekly ice maps have been prepared by NPI/DNMI since 1966. The ice maps are digitised at NPI  for the period 1966-1993. The dataset is available at DNMI (presently on CD-rom). Some transformation (gridding) has to be done prior to applications in model validation.

Through different sources a series of ice maps is established (Vinje, NPI) covering the years 1552-1965. This series is also available at DNMI, and will probably be organised in co-operation with NoSerC.

New Task 8.7:   Establish a common library with programs/scripts.

The format of the data in the archives of the different participating institutes, as well as data retrieved from other archives, may have different formats. The same is valid for output from model simulations as well. The purpose of his task is to make the transformation between different formats easy through a library of programs/scripts with these functionalities. For reference, see Benestad (1999 c,d). This task is duplicated by Task 4.2 and 4.4 in the new NFR project NoSerC. It is expected that the task will be financed there.

 

Relations to other Principal Tasks:

PT1-PT7 and PT9:

Survey of available data for model evaluation. Task 8.7 and 9.1 will be co-ordinated.

 

Milestones:

Table 8: Target schedule PT 8

Task

Year

 

1997

1998

1999

2000

2001

2002

8.1

 

x

x

x

x

x

x

x

x

x

   

8.2

 

x

x

x

      x x x    

8.3

 

x

x

x

x

x

  x x x    

8.4

 

x

x

x

x

x

x

x

x

x

   

8.5

   

x

x

x

x

x

x

x

x

   
8.6               x x x    
8.7               x x x    

Finalization

                   

x

 

Institutions and Personnel:

Task 8.1, 8.4, 8.5, 8.6b, 8.7:

                        The work is to be carried out at DNMI by K. A. Iden and other DNMI scientists.

Task 8.2:          The work is to be carried out by Bjørn Ådlandsvik, Kjell Arne Mork and other   IMR scientists.

Task 8.3:          The work is to be carried out at NILU by Mihalis Lazaridis

Task 8.6a:        The work is to be carried out at NERSC by Helge Drange and other scientists.

 

Budget:

The following numbers give contributions in units of Person-years.

 

Contribution from the Research Council:

                                                            2000              2001   

Task 8.2

IMR:                                        0.15                 0.1

Task 8.3

NILU:                                      0.15                 0.1

Task 8.6a

            NERSC                                   0.15                 0.1      

RegClim Phase II                                 0.45                 0.3      

 

Own funding:

                                                            2000                2001   

PI for PT8, Task 8.1,8.4,8.5*

DNMI:                                      0.2                  0.2      

Technical assistance to all tasks and

additional resources for 8.6b) and 8.7:

            DNMI/NoSerC**                                            0.5      

*Already included in RegClim Phase II of Dec. 1999
**Proposal from RegClim, agreement with NoSerC will be sought
.

Own funding:

At present there is no funds from the Research Council for PT8 in Phase II. DNMI will contribute with own funding through the work of PI Knut A. Iden. This amounts to 0.3 Person-years per year for the work as PI for the years 2000-2001.

 

5.9       Principal Task 9 (PT9): Advanced analysis and interpretation of climate model results and observations

Principal Investigator: Inger Hanssen-Bauer, DNMI, Oslo

 

This is a completely new Principal Task.

 

Major Objective:

To detect forced and internal variability on different time-scales in the climate system with focus on our region, and to improve the physical interpretation of the non-linear dynamical interactions that cause the variability.

Tasks:

Since the climate system is a highly non-linear chaotic system, internal variability is large, and special attention must be given to the interpretation of observations and model results. In order to understand and interpret the changes and variabilities found, sophisticated statistical tools will be used. In PT 9, such tools will be used to interpret results from model simulations of various types that have been produced in other Principal Tasks. Also, studies of the sensitivities of prevailing atmospheric flow regimes for changes in forcing will be initiated.

The advanced interpretations in PT9 will take place in co-operation with scientists who produce model results  in other PTs, and it is anticipated that these scientists will be involved .  Thus the total effort in RegClim spent on such interpretations should become considerably larger than the budget for PT9 alone indicates. 

Task: 9.1    Preparation (and consultancy in use) of methods for advanced statistical analysis.

There are various program packages that can be used for advanced data analysis, but few are tailored for large geophysical/geographical data. Many of these packages (i.e. Matlab, IDL, Ferret) require scripts to do specific jobs, and some may even use FORTRAN or C. There is now a large number of scripts written within the RegClim project (e.g. Benestad, 1998b,c,d).  In task 9.1 tools for advanced statistical analyses,  (e.g. EOF-Analysis, CCA, SVD, MVR, Wavelet Analysis, and Spectral Analysis) will be made available over internet for all RegClim scientists.  The routines will be available as MatLab-routines (m-files), FORTRAN-routines, IDL-scripts or Ferret-scripts.  The scripts and codes will be well documented, so that they can be used straightforwardly. A group of RegClim scientists with experience in the use of statistical methods will be available for giving advice in choice of methods etc.

Task 9.2:    A tool for estimating forcing sensitivities and optimal excitation of flow regimes.

The T21L3 atmospheric, quasi-geostrophic (QG) model originally developed by Marshall and Molteni (1993) now includes a facility to calculate leading forcing singular vectors. The whole package can be run in “climate mode” (without input of actual analysed atmospheric data) on modern workstations, and can be set up to estimate regimes’ sensitivity for small changes in external forcing, and forcing-structures that optimally trigger unstable growth over finite time-intervals  (forcing singular vectors). Such results can be of considerable interpretational value, and be a first step towards generalizing the “fingerprint” method to non-linearity (Palmer, 1999). Any parts of PT9 (and RegClim) that involve atmospheric flow regimes may benefit from such simplified diagnostics at present state of science, even though results will be far from conclusive.

 

Task: 9.3:   Analyses of model results from dynamical downscaling vs. observations in the context of flow-regimes

In PT1 two 20-year time-slice experiments have been run and the data tested in terms of standard statistical parameters. In phase II data from these runs will be evaluated further in terms of flow regimes possibly estimated based on leading common EOFs, and compared to those of the global forcing models (see Task 9.4 below). Establishing to what degree regionalisation improves the coarse global model results in the region will be addressed. It is believed that results from Task 9.2 will be of guidance in the latter analysis.

If time and resources permit, similar evaluation will be performed for runs with the coupled atmosphere/slab ocean model as soon as results from these runs are available.

Task: 9.4:   Analyses of flow regimes in the ECHAM4/OPYC3 model results vs. observations on a regional scale

In PT3 the ECHAM4/OPYC3 control and GSDIO integrations have been validated in the North Atlantic region on a monthly basis using various statistical techniques (e.g. Benestad et al. 1999). Common EOF-analysis indicated an increase in the average loading of the first EOF (a NAO-like structure) in the period with increasing radiative forcing (Hanssen-Bauer 2000). - In phase II, the ECHAM4/OPYC3 GSDIO integration will be evaluated by using common EOFs based upon daily observed and modelled regional SLP fields. Flow regimes will be defined based upon the first few common EOFs, either by using one of the classification methods suggested by Huth (1996) or by application of PDFs. Frequencies of different regimes and transition between them will be compared for model results and observations, and for model time-slices including different radiative forcing. 

Task: 9.5:   Storm track diagnostics in relation to oceanic forcing  

Simulations carried out in Task 4.4 and 4.5 of PT 4 will be used as a basis for an examination of weather systems and their interaction with the underlying sea (ice) surface. Existing techniques (Hodges (1994, 1995, 1996) will be used for a detailed cataloguing of simulated storms seen in surface pressure and to identify weather systems in a range of parameters. Numerous diagnostics of weather systems behaviour will be produced and stratified using the NAO/AO and other relevant indices of the large-scale flow. Indices such as AO and NAO have shown to be important in the description of the SST and sea ice cover variability as well as for the atmospheric variability. This stratification should thus be useful in order to identify the role of the synoptic weather systems in the coupled system. Particular attention will be paid to weather systems that either directly influence the Nordic region or are associated with Arctic cold air outbreaks.

Optimal forcing patterns and forcing sensitivity diagnosed by the simplified T21L3 QG (Task 9.2) may be used for guidance in the interpretation.

Task: 9.6:   Analyses of impact of inhomogeneous radiative forcing

Multi-year simulations of direct and indirect effect of aerosols, as well as the radiative impact of changes in ozone distribution, will be carried out in the near future. Based on results so far (Berntsen et al. 1997; Berntsen et al. 2000; Kirkevåg et al., 1999; Kristjansson, 1999; Kirkevåg et al., 2000; Myhre et al., 2000) it is clear that the radiative forcing have significant spatial inhomogeneities. What is the impact of these inhomogeneous forcings on the atmospheric flow patterns in the North Atlantic and its vicinity?  In order to answer this question the multi-year results need to be analysed using appropriate statistical tools (e.g. regression correlation patterns, spectral analysis, cluster analysis, distribution functions).  One particularly interesting question is to what extent flow patterns such as the Arctic Oscillation and/or the North Atlantic Oscillation are affected.  Regime estimations in a low-dimensional common EOF-space will probably be an important tool for this purpose. Also estimates of forcing singular vectors and sensitivity patterns developed in Task 9.2 may be of valuable qualitative guidance at this stage. (However, the missing vertical resolution limits the applicability for the AO.)

The non-linear interactions between different processes (dynamical feedbacks) often render it difficult to explain the course of events leading to the final differences in results between different simulations. This can be partly alleviated by the use of sophisticated statistical tools such as Multivariate Regression (e.g., Benestad 1999e), establishing links between different variables that may differ from one simulation to another.

Task: 9.7:   Upper ocean heat content diagnosis

A high-resolution isopycnal model (MICOM) covering the North Atlantic and Arctic Ocean has been integrated for 40 years using the NCEP/NCAR reanalysis data as forcing fields. The output from this model will be analysed in terms of the upper ocean heat and salinity fluxes. The results will be compared to available data from hydrographical sections spanning from the inflowing area in the Faroe Shetland Channel to the Barents Sea and Fram Strait. The heat and fresh water budget for the Arctic Ocean will be emphasised.

The anomalies in the model, will be analysed in terms of complex empirical orthogonal functions (CEOFs), which have the advantage of detecting propagating features as well as standing oscillations which is the case for the traditional empirical orthogonal function analysis (Horel 1984). The method has been used on satellite observed sea surface temperatures from the Nordic Seas (Furevik 2000b), and it was shown that the upper ocean heat anomalies could be tracked from space. The CEOF results from the model will be compared with these observations.

Relations to other Principal Tasks:

PT 1:   Task 9.3 performs statistical analyses of results obtained within  PT 1.

PT 2:   Task 9.7 performs statistical analyses of results obtained within PT 2

PT 3:   In task 9.1, scripts developed mainly within PT 3 will be made available for other PTs.  In task 9.4, analyses of the flow regimes of ECHAM4/OPYC3GSDIO integration will be accomplished in cooperation with PT3

PT 4:   Task 9.5 performs statistical analyses of results obtained within tasks 4.4 and 4.5

PT 5 and PT 6:

            Task 9.6 performs a statistical analysis of results obtained within these PTs, mainly tasks 5.8 and 6.7

PT 7:   Task 9.3 performs statistical analyses of results obtained within PT 7.

PT 8:   Task 9.1 and task 8.7 will be co-ordinated.

Milestones (all new activities in the last extension of RegClim in dec. 2000):

 

 

Table 9: Target schedule PT9

Task

Year

 

1997

1998

1999

2000

2001

2002

9.1

              x x x    

9.2

                x x    

9.3

              x x x    

9.4

                x x    

9.5

              x x x    
9.6               x x x    
9.7               x x x    

Finalization

                    x  

 

Institutions and Personnel:

Task 9.1:          Rasmus Benestad, DNMI,  Paul Skeie and Tore Furevik, GfI-UiB

Task 9.2:          Inger-Lise Frogner, Igf-UiO and Trond Iversen, (Igf-UiO, own contribution).

Task 9.3:          Viel Ødegaard (DNMI) and possibly Trond Iversen (Igf-UiO,own contribution)

Task 9.4:          Rasmus Benestad /Inger Hanssen-Bauer/Eirik Førland/Viel Ødegaard (DNMI)

Task 9.5:          Dr. Ina T. Kindem, GfI-UiB

Task 9.6:          Inger-Lise Frogner, Igf-UiO and scientists involved in PT5 and 6

Task 9.7:          Dr. Kjetil Lygre, NERSC

 

Budget in person-years: (this extension in bold)

Contribution from Research Council:

                                                            2000                2001   

Task 9.1

DNMI                                      0.05                 0.1      

GfI-UiB                                    0.05*               0.1*    

Task 9.2

IGf-UiO                                                           0.27    

Tasks 9.3 and 9.4

DNMI                                      0.3                   0.08    

Task 9.5

GfI-UiB                                    0.1                   0.5      

Task 9.6

IGf-UiO                                                            0.4      

Task 9.7

            NERSC                                    0.3                   0.2      

RegClim Phase II:                                 0.8                   1.65    

Own funding:

                                                            2000                2001   

PI for PT 9 and partly Task 9.4

DNMI:                                      0.1                  0.2      

Task 9.2 (and possibly 9.3)

            Igf-UiO:                                    0.4                  0.1      

 

*Transferred from PT4