|
UGAMP Research
|
|
Developments with the Unified Model
Work has been progressing with the Unified Model over the past 6 months. There are now more than 10 individual projects utilising the Unified Model in UGAMP. These range from an investigation of the dynamics of large scale tropical circulations in an Aqua Planet version of the UM, to attempts to simulate hurricane formation in the UM limited area model. Details of some of these projects can be found elsewhere on this page (see also Atmosphere-Ocean Modelling and Palaeoclimate modelling).
The latest version of the UM physical parametrization package has been supplied to CGAM. This will be disseminated to the UGAMP community early in the Spring of 1997. The major changes from the 3rd Climate version are the new Edwards-Slingo Radiation scheme, the inclusion of cumulus momentum transport and the MOSES surface exchange scheme.
A 10 year integration of the Unified Model has been performed for the AMIP period (1979-88), using monthly mean Sea Surface Temperatures as a lower boundary condition. The model utilised 31 vertical levels compared with 19 in the standard climate configuration. The extra 12 levels were located below 500mb, providing greater vertical resolution in the lower troposphere. It is hoped this will allow for a better representation of crucial processes such as cloud formation and vertical turbulent fluxes. The results are being compared with three equivalent integrations performed at the Hadley Centre. These are directly comparable with the 31 level run except they differ in either the horizontal or vertical resolution. This work is progressing in collaboration with the Hadley Centre. Only a preliminary analysis has been performed to date. A more detailed investigation will be presented subsequently.
It was hoped that greater vertical resolution in the lower troposphere would produce a better representation of subtropical stratocumulus clouds. These play a crucial role in the radiation budget of the tropics. Fig. 7 shows a vertical cross section through the model atmosphere at 17.50S. The cross- section extends from 200E to 1200W, encompassing both the Namibian and Peruvian stratocumulus cloud decks. The figure shows the seasonal mean total cloud field for July/August/September for the period 1982-88. The standard 19 level model is compared directly with the 31 level model. The low level stratocumulus cloud decks clearly stand out, as maxima in cloud fraction at 100E and 750W, off the Namibian and Peruvian coasts respectively. There is a clear increase in the stratocumulus amounts in the 31 level model off the Namibian coast. A smoother transition to more scattered, deeper shallow cumulus clouds can also be seen in the 31 level model. This is an improvement when compared to observations. Improvements off the Peruvian coast are less clear. A detailed analysis of the cloud fields will be performed utilising ERBE satellite data.
Fig. 8 shows the seasonal mean total precipitation for Jan./Feb./Mar. 1982-88 in the 19 and 31 level models, compared to the Xie-Arkin climatology. A persistent problem with the UM precipitation field is the lack of rainfall over the maritime continent in the West Pacific. This problem appears to have been exasperated in the 31 level integration. Extremely high rain rates can be seen to the west of the maritime continent at 1650E and a distinct hole in the precipitation field over the maritime continent. This is not evident in the Xie- Arkin climatology. It is hoped that further analysis will explain some of the deficiencies in the model precipitation field.
SST Anomaly Experiments with the UM
The relationship between El Nino and the Asian Summer Monsoon (ASM) has been known for many years and has been studied extensively using both observations and numerical models. In essence, drought years over India are often, but not exclusively, related to warm SST anomalies in the equatorial central and East Pacific (El Nino), and wet years with anomalously cold SSTs (La Nina). As well as the sensitivity of the monsoon strength to the phase of El Nino, recent observational studies (Joseph et al. 1994), supported by modelling results obtained at CGAM (Ju and Slingo 1995; Soman and Slingo 1997), have found that the monsoon onset date is delayed significantly in El Nino years.
To investigate the influence of El Nino on the ASM further, a series of integrations with the UM have been performed in which an idealized El Nino SST anomaly distribution is imposed in the model. Original analysis of the results was carried out by Dr. Roy, a Commonwealth Fellow from Bangladesh, who visited CGAM for 10 months. Subsequently, the project has been extended in collaboration with Dr. Gill Martin at the Hadley Centre who is working on the SHIVA project.
The results from the initial experiment with the El Nino imposed throughout the annual cycle showed the expected weakening of the ASM and a substantial delay in the monsoon onset. As noted in various modelling and observational studies, there are robust precursory signals which develop in association with El Nino, particularly an equatorwards shift of the subtropical jet. An important question raised by many studies has been the relative roles played by these precursory changes in the circulation (and possible associated Eurasian snow anomalies), compared with the direct response of the ASM to El Nino during the monsoon season itself. This is currently being investigated using further integrations of the UM designed to isolate these two particular aspects of the response of the ASM to El Nino. Preliminary results suggest that the direct influence of El Nino during the monsoon season itself is the dominant factor in producing a weak and late ASM. However, the influence of the precursory signals is not negligible and appears to be important in the early part of the monsoon season. Analysis of the results continues.
Investigation of the Physical Mechanisms involved in the Intraseasonal Variability of the Asian Monsoon using the UM.
One of the goals of SHIVA (newsletter 14) is to investigate the physical mechanisms involved in the intraseasonal variability (ISV) of the monsoon and its role in the potential predictability of the monsoon using a GCM. During the monsoon season, there are two preferred locations of the tropical convergence zone (TCZ), one over the continent and the other over the warm waters of the equatorial Indian Ocean. The northward propagation of the oceanic TCZ is poorly simulated by many GCMs and the physical processes that maintain the propagation is poorly understood. To gain some insight into the physical processes for the ISV, we use the Unified Model (UM) which has a realistic simulation of the mean monsoon precipitation, ISV and, particularly, the northward propagation of the TCZ.
We conduct a series of idealised numerical experiments in perpetual July mode to understand the basic characteristics of the ISV during the established monsoon. The goals of the study are: a) to investigate the role of land-atmosphere interaction (hydrology), b) to investigate the influence of Pacific SST anomalies (idealised El Nino), and c) to investigate the influence of equatorial Indian Ocean SST on the ISV of the monsoon. The model was run for 360 days to obtain statistically robust results.The model OLR and precipitation are analysed using empirical orthogonal function (EOF) analysis to extract the dominant modes of ISV.
The results with prescribed and interactive land surface hydrology are discussed. In the prescribed case mean soil moisture contents for July from a UM integration are used and kept fixed throughout the run. Fig. 9a shows the first EOF of model OLR with interactive hydrology and Fig. 9b that with prescribed hydrology. It is apparent that the presence of the two TCZs over the monsoon domain is the dominant mode of ISV in both the simulations; the similarity between their spatial patterns is very striking. This result indicates that the ISV is governed by internal dynamics and does not depend on land-atmosphere interaction. The time series (PC1), shown in Fig. 10 picks out the intraseasonal time scales of about 20-60 days. The interesting aspect is that in the interactive hydrology experiment, PC1 fluctuates at much longer time scales compared to PC1 in the prescribed hydrology case. Further investigation indicates that the northward propagation of the oceanic TCZ does not necessarily require the hydrology to be interactive. The remaining experiments are underway.
H. Annamalai
and Julia Slingo
(Reading)
Middle Atmosphere variability in the UKMO Unified Model
The Unified Model is the name given to a suite of numerical modelling software developed and used at the UKMO (Cullen 1993). The model is designed to be used for a wide range of different applications - notably Numerical Weather Prediction and climate simulation. The model- ling system also includes an ocean model, which can be run coupled with the atmosphere model, and a data assimilation scheme to assimilate meteorological observations. The Unified Model was introduced into operational service in 1991. Since then both its formulation and capabilities have been substantially enhanced.
The choice of horizontal and vertical resolution is arbitrary, although a number of standard resolutions tend to be used. Both the operational and climate models are currently run using a set of 19 levels, extending up to the middle stratosphere; the vertical resolution used for the top few levels is quite coarse (typically 5 km). By contrast, the 49-level configuration is designed to have a high vertical resolution of approximately 1.3 km throughout the stratosphere. The Unified Model uses a hybrid pressure/sigma vertical coordinate system (Simmons and Strufing 1983). The lowest four model levels are sigma levels, and the model levels gradually become constant pressure surfaces above about 50 hPa.
The Unified Model is grid-point and has a global domain with a horizontal resolution of 2.5o latitude and 3.75o longitude. This is the same resolution as that usually used for climate modelling, but three times coarser than the current operational resolution used for global forecasting.
Several integrations have now been run in the course of developing the 49-level configuration of the Unified Model. Two of these integrations have been carried out in a collaboration between the Centre for Global Atmospheric Modelling (CGAM) at the University of Reading and the UK Meteorological Office (UKMO). The first integration lasted for 15 months and is described in Met. Office Climate Research Technical Note No. 76 and UGAMP Technical report No. 42 (Swinbank et al. 1996). The second integration is five years long and is described in Swinbank et al. (1997).
The meteorological analyses used to validate the Unified Model are produced at the UK Meteorological Office using a stratospheric data assimilation system (Swinbank and O'Neill 1994). This system is based on the "Analysis Correction" data assimilation scheme which was originally developed for operational weather forecasting (Lorenc et al. 1991).
The UM produces a very good simulation of the stratospheric circulation and its seasonal evolution. As shown by Swinbank et al. (1996), the model reproduces synoptic features such as warmings and vortex mergers.
In common with most other middle atmosphere models, the UM does not simulate the quasi-biennial oscillation (QBO) which occurs in the tropical stratosphere. Instead, any tropical westerly winds in the initial data fade away over a period of several months, leading to an overall easterly bias in the tropical stratosphere. The UM does simulate a weak semi-annual oscillation (SAO), but the westerly phase tends to be much too weak.
(contd. at top of next column)
(contd. from previous column)
In general the UM shows some promising improvements over earlier stratosphere- troposphere models, although it still has some weaknesses, for example a cold bias, particularly in the middle to upper stratosphere (the cold bias is typically 10 K at 10 hPa). It is planned to address these weaknesses in future work. A notable positive feature of the UM is the synoptically realistic simulation of the winter circulation in both southern and northern winter (Figs. 11 and 12).
References
Cullen, M. J. P., 1993,: The Unified forecast / climate model. Meteorol. Mag., 122, 81-94.
Lorenc, A. C., R. S. Bell, and B. MacPherson, 1991: The Meteorological Office analysis correction data assimilation scheme. Quart. J. Roy. Met. Soc., 117, 59-89.
Simmons, A. J., and R. Strufing, 1983: Numerical forecasts of stratospheric warmings using a model with a hybrid vertical coordinate system. Quart. J. Roy. Met. Soc., 109, 81-111.
Swinbank, R., and A. OĠNeill, 1994: A stratosphere-troposphere data assimilation system. Mon. Wea. Rev., 122, 686-702.
Swinbank, R., W. Lahoz, C. Douglas, A. Heaps, R. Brugge, W, Norton, A. OĠNeill, and D. Podd, 1996: Middle Atmosphere Variability in the UKMO Unified Model. Met. Office Climate Research Technical Note no. 76 and UGAMP Technical report no. 42.
Swinbank, R., C. S. Douglas, W. A. Lahoz, D. Podd, A. Heaps, and W. A. Norton, 1997: The Stratosphere-Troposphere Configuration of the UK Meteorological Office Unified Model. First SPARC Assembly Proceedings, in press.
William Lahoz
(Reading)
Richard Swinbank
(UKMO)
African Easterly Wave
In this research work we examine the African Easterly Wave (AEW) disturbances that emerged from the western coast of Northern Africa (mainly in 1995 which was one of the most active hurricane seasons of the century). We are particularly interested in their transformation into tropical cyclones in the Atlantic ocean. Radiosonde data, satellite imagery and the UKMO Global operational analyses have been used so far. After inspection of the global analyses, the UKMO Limited Area Model (LAM) will be run over Northern Africa and the tropical and subtropical Atlantic. The LAM is going to be used to make a high resolution examination of some intensifying and non-intensifying AEWs that occurred during the hurricane season of 1995. Also, the AEWs that were observed during this period, will be tracked, and their evolution determined. This is particularly important for the low-level warm core vortices that form north of the African Easterly Jet because their track has not been determined with much confidence so far; except in the work of Reed et al. (1988). Moreover, some experiments will be performed to assess the importance of Sea surface temperatures (SSTs) in the evolution of these hurricanes. One such experiment is to examine the influence of the warm SSTs that were observed in the tropical Atlantic on the evolution of the AEWs when they moved over the ocean and the subsequent transformation of some of them into tropical cyclones in this region. Saunders and Harris (1996) claim that the high SSTs were the dominant environmental influence behind the Atlantic's 1995 highly active season.
In preparation for this work, we have developed the ability to run the Limited Area version of the Unified Model with Initial and boundary conditions supplied directly from the operational analyses.
Reed, R.J., Klinker, E., and Hollingsworth, A., (1988). The Structure and Characteristics of African Easterly Wave Disturbances as Determined from the ECMWF Operational Analysis/Forecast System. Meteorol. Atmos. Phys., 38, 22-33.
Saunders, M.A., and Harris, A.R., (1996). Sea Surface Temperature and the Near-Record 1995 Atlantic Tropical Cyclone Activity. Submitted to Science.
Ioannis Pytharoulis,
Chris Thorncroft &
Colin Jones
(Reading)
Use of the UKMO unified model in aqua-planet mode
A number of 6-month integrations have been performed using the UKMO unified model in an aqua-planet configuration. This involves taking the standard climate setup of the UM and removing all the land points except for two latitude strips at each pole. The integrations are initialised using an arbitrary initial atmospheric state and reconfigured for zero orography. Forcing is provided by a fixed sea surface temperature (SST) distribution and perpetual March insolation.
Zonally symmetric SST distributions have been used in order to investigate how robust the large-scale axisymmetric circulations are in the absence of land. It has also proved useful to make comparisons with theory and simple axisymmetric models of the atmosphere to predict the nature of a circulation such as the Hadley Cell.
Preliminary results reveal a sub-tropical westerly jet maximum which is both stronger and further poleward than theory would suggest when considering angular momentum conservation and radiative equilibrium wind arguments.
Fig. 13 shows the zonally symmetric SST distributions used. Fig. 14 gives the zonally averaged zonal wind resulting from the control run SST profile which has a sine squared profile decreasing from 300K at the equator to 273K poleward of 60 degrees.
In addition to the use of zonally symmetric SST distributions integrations have been performed with anomalies superimposed onto the original distribution giving tropical 'warm' and 'cold pool' regions. Comparisons have been made between results using a 3K planetary wavenumber one anomaly and a more confined 3K anomaly covering just 60 degrees in longitude. Results from these two experiments differ substantially with the wave number one case forcing westerlies in excess of 45 metres per second downstream of the SST maximum whereas the confined anomaly only forces downstream westerlies of the order 6 metres per second. The mechanism for the production of such strong westerlies in the first case and the reason for the marked difference between the two cases is being investigated further. Fig. 15 shows the strong equatorial westerlies produced as a response to the 3K wavenumber one SST distribution and the equivalent response for the geographically constrained anomaly.
A future aim with the aqua-planet setup is to investigate the effects of land on this idealised ocean-atmosphere regime particularly in the tropics where surface interaction are central to the operation of the whole climate system. To facilitate this study an idealised continent will be reintroduced to the configuration and will hopefully yield an insight into the interaction locally and the communication of its effects tropic-wide.
Richard Neale,
Brian Hoskins &
Colin Jones
(Reading)
Predictability of the Intraseasonal oscillation in the UK Hadley Centre Climate Model
In collaboration with the Hadley Centre an ensemble of GCM integrations have been used to investigate the interannual variability of the tropical intraseasonal oscillation. The intraseasonal oscillation, (ISO), or 30-60 day oscillation, is an important component of tropical variability. The oscillation can be seen clearly as large scale convective centres which move eastward from the Indian Ocean to the central Pacific Ocean. It is also evident as anomalies in the upper troposphere which often propagate the full circumference of the globe, (Madden and Julian, 1994).
Various mechanisms have been suggested which might affect the oscillation and the aim of this work was to investigate whether the mean state of the tropics was a factor involved in the predictability of the ISO. The major forcing on the mean state of the tropics is the sea surface temperature, (SST), distribution, and changes in the distribution, as in El Nino years, can cause changes in the whole structure of the tropical troposphere, (Philander, 1990). Therefore, using an ensemble of integrations the reproducibility of the ISO on the interannual timescales has been examined.
The model data used in this study was created using the Hadley Centre climate model, version HADAM2a, which is a version of the UK Meteorological Office, (UKMO), Unified Model. Four integrations had been carried out whilst applying the Hadley Centre Global Ice and Sea Surface Temperature, (GISST), data set as a boundary forcing for the period 1948-1993. For the purpose of this study a four member ensemble of data sets was available for the 45 year period.
In this work the variability of the ISO has been examined by calculating an ISO Index. The variance of the band-pass filtered, (20-100 day), 200mb equatorial zonal mean wind, (10N-10S), is calculated and then low pass filtered, (100 day running mean), as described by Slingo et al. 1996, (their figure 14). This essentially shows how active the ISO is and how the activity varies on the seasonal to interannual timescales. Fig. 16 shows the ISO index for the four integrations for the period 1949-1993. The interannual variability of the ISO is shown quite clearly, with periods of very little activity in the mid-1960s and periods of great activity in the early-1980s. The figure also shows that ISO activity is more likely to occur in the late winter or early spring seasons and a calculation of the average seasonal cycle, (not shown) displays the ISO index to be largest in the months January- March. This is in agreement with results shown in Madden and Julian (1994) where the oscillation in atmospheric data is shown to be largest during the December-February period.
The reproducibility of the ISO index in the four time-series was measured using the 'Analysis of Variance' method, (Rowell et al. 1995). The calculation essentially estimates what percentage of the variance is the same in the four time-series. In this case only 10% of the variance was found to the same in each of the integrations and, therefore, only 10% of the total variance can be attributed to the SST forcing. The other 90% is due to internal variability. A correlation of each of the time- series with the Southern Oscillation Index, (SOI), also gives a correlation of only -0.09. These results show that SSTs have no discernable impact on the variability of the ISO in this model and that there is no reproducibility of the oscillation in atmosphere-only integrations forced with observed SSTs. This result questions the predictability of the Intraseasonal Oscillation given the mean state of the tropical atmosphere, but does not answer the question of whether the oscillation is inherently randomly initiated or whether the mechanism which initiates the oscillation is a feature not allowed within the model such as air - sea interaction.
Madden, R.A., and P.R. Julian, 1994: Observations of the 40-60 Day Tropical Oscillation - A Review. Mon.Wea.Rev., 122, 814- 835.
Philander, S.G.H., 1990: El Nino, La Nina, and the Southern Oscillation. Academic Press, 242pp. Rowell, D.P., C.K. Folland, K. Maskell, and M.N. Ward, 1995: Variability of Summer rainfall over Tropical North Africa (1902-92): Observations and Modelling. Q.J.R.Meteorol.Soc., 121, 699- 704.
Slingo, J.M., K.R. Sperber, J.S. Boyle, J.-P. Ceron, M. Dix, B. Dugas, W. Ebisuzaki, J. Fyfe, D. Gregory, J.-F. Gueremy, J. Hack, A. Harzalla, P. Inness, A. Kitoh, W.K.-M. Lau, B. McAvaney, R. Madden, A. Matthews, T.N. Palmer, C.-K. Park, D. Randall, N. Renno, 1996: Intraseasonal oscillations in 15 Atmospheric General Circulation Models: Results from an AMIP diagnostics subproject. Climate Dynamics, 12, 325-357.
F.Nortley and
J.Slingo
(Reading)
D.Rowell
(Hadley Centre).