Effects of the changing heating profile associated with ... · EFFECTS OF THE CHANGING HEATING PROFILE ASSOCIATED WITH MELTING LAYERS 1 1. INTRODUCTION Convective heating and moisture

Bureau Research Report - 007

Effects of the changing heating profile associated with

melting layers

Hongyan Zhu and Rachel Stratton December 2015

EFFECTS OF THE CHANGING HEATING PROFILE ASSOCIATED WITH MELTING LAYERS

i

Effects of the changing heating profile associated

with melting layers

Hongyan Zhu 1 and Rachel Stratton 2

1 Bureau of Meteorology 2 UK Met Office

Bureau Research Report No. 007

December 2015

National Library of Australia Cataloguing-in-Publication entry

Author: Hongyan Zhu and Rachel Stratton Title: Effects of the changing heating profile associated with melting layers ISBN: 978-0-642-70670-6 Series: Bureau Research Report - BRR007


ii

Enquiries should be addressed to: Hongyan Zhu: Bureau of Meteorology GPO Box 1289, Melbourne Victoria 3001, Australia Contact Email: [email protected]

Copyright and Disclaimer

© 2015 Bureau of Meteorology. To the extent permitted by law, all rights are reserved and no part of

this publication covered by copyright may be reproduced or copied in any form or by any means

except with the written permission of the Bureau of Meteorology.

The Bureau of Meteorology advise that the information contained in this publication comprises

general statements based on scientific research. The reader is advised and needs to be aware that such

information may be incomplete or unable to be used in any specific situation. No reliance or actions

must therefore be made on that information without seeking prior expert professional, scientific and

technical advice. To the extent permitted by law and the Bureau of Meteorology (including each of its

employees and consultants) excludes all liability to any person for any consequences, including but

not limited to all losses, damages, costs, expenses and any other compensation, arising directly or

indirectly from using this publication (in part or in whole) and any information or material contained

in it.


iii

Contents

1. Introduction .......................................................................................................... 1

2. Model Description ................................................................................................ 1

3. Results and discussion ....................................................................................... 2

4. Conclusion and Discussion: .............................................................................. 8

5. Reference: ............................................................................................................ 8

Acknowledgement: ....................................................................................................... 9


iv

List of Figures

Fig. 1 Temperature anomaly relative to the heavy rainfall events. (60E-180E, 20S-20N). (a) Control experiment; and (b) Expt. 1. ............................................................................... 2

Fig. 2 Detrainment rate(/s) relative to the heavy rainfall events. (60E-180E, 20S-20N). (a) Control experiment; and (b) Expt. 1. ............................................................................... 3

Fig. 3 Difference between Expt.1 and control experiment for moisture anomaly relative to the heavy rainfall events. (60E-180E, 20S-20N). ............................................................... 4

Fig. 4 (a) Precipitation (mmd−1) as a function of surface temperature. (60E-180E, 20S-20N). (b) The ratio of surface temperature in the Control experiment. .................................... 4

Fig. 5 The model bias of precipitation rate of the control experiment comparing to the GPCP observation. For four years of 2006-2009. (a) The 4 years averaged model bias; (b) the averaged model bias for DJF seasons; and (c) the averaged model bias for JJA seasons. ......................................................................................................................... 5

Fig. 6 The upper panels are the model bias for DJF and JJA seasons and the lower panels are the difference between the Expt. 1 and the control experiment for DJF season ( left panels) and JJA (right panels). ....................................................................................... 6

Fig. 7 Lag correlation of equatorial intraseasonal OLR (a,c) and 850 hPa zonal wind (b,d) onto a reference zonal wind time series at 90E in ( The upper panels are for the control experiment (a and b) and Expt. 1 (c and d) ................................................................... 7


1

1. INTRODUCTION

Convective heating and moisture profile shows that the ACCESS model (Puri, 2010) has a large

cooling and drying spike at the freezing level (Zhu and Hendon, 2015) due to all the snow being

melted at the layer where the environmental temperature reaches the freezing level. For the high

vertical resolution model, it is likely that the freezing level is only represented by one model

level. Observation suggests that there is a mix of frozen and liquid precipitation over a layer of

several kilometres deep from the freezing level downwards (Illingworth and Thompson, 2011).

A simple way to alter the convection code is to allow a mix of liquid and frozen precipitation

from the freezing level to the freezing level plus 3 degrees. This change allows a mix of snow

and rain between the freezing level and 3K above this with the proportion of rain increasing

linearly from zero at the freezing level to one at the freezing level plus 3K.

2. MODEL DESCRIPTION

The UK Met Office Unified Model with GA2.0 model physics is used in this study (Zhu and

Hendon, 2015). The model horizontal resolution is 1.875°*1.25° and the integration time step

is 1200s.

The model uses a modified mass flux scheme based on Gregory and Rowntree (1990). The

convective diagnosis is based on an undiluted parcel ascent from the near surface. The

convective diagnosis is used to determine whether convection is possible from the boundary

layer and, if so, whether the convection is deep or shallow depending on the level of the cloud

top. The mid-level convection scheme operates on any instability found in a column above the

top of the deep/shallow convection or above the boundary layer in columns where the surface

layer is stable.

For deep convection, the cloud-base mass-flux is calculated based on the reduction to zero of

Convectively Available Potential Energy (CAPE) over a given timescale. The CAPE closure

has been modified in various ways to try to address model stability problems (grid point

storms). W based CAPE closure is the option used in most model configurations. In this

scheme, if the maximum large-scale vertical velocity, evaluated before convection, is larger

than the threshold vertical velocity, the CAPE timescale is reduced to remove the convective

instability faster. For deep convection, convective entrainment rates use prescribed profiles.

Many recent studies have found that increasing the convective entrainment rates will improve


2

MJO simulation in the model, therefore in this work the mixing entrainment rates in the

convection parameterization scheme are increased by 1.5. Mixing detrainment rates depend on

relative humidity and an adapted detrainment scheme is used to calculate the forced detrainment

rates. The representation of Convective momentum transports (CMT) is based on an eddy

viscosity model and a flux gradient approach has been introduced in the latest version of model.

The model uses the prognostic cloud fraction and prognostic condensate (PC2) scheme of

Wilson et al. 2008. For the boundary layer scheme, turbulent fluxes of heat, moisture and

horizontal momentum in the boundary layer are represented by a first-order K profile closure as

described by Lock et al. (2000). The model radiation scheme uses the modified version of

Edwards and Slingo (1996) scheme based on rigorous solution of the two-stream scattering

equations including partial cloud cover.

Model simulations use weekly observed SSTs and Sea Ice, and the simulations cover the

periods from 2006 to 2009.

3. RESULTS AND DISCUSSION

To show the temperature profile associated with the intense convection, in Fig.1, we plot the

lag and lead composites relative to the time of maximum rainfall anomaly. At each grid point,

we define a heavy rainfall event as an occurrence of rainfall anomaly that exceeds one standard

deviation, about 9.68 mm/day. Here we calculate the anomalies by removing the

climatologically seasonal cycle over the four years. In this fashion, we build up a one-

dimensional (height) lead-lag composite of the evolution of the intense rainfall events.

(a) (b)

Fig. 1 Temperature anomaly relative to the heavy rainfall events. (60E-180E, 20S-20N). (a) Control experiment; and (b) Expt. 1.


3

Figure 1 shows the temperature anomaly relative to the heavy rainfall events. At the time of

maximum precipitation, a warm center reaches 0.6 degree at the level of 10km due to

condensation heating in the upper troposphere. Meanwhile cooling anomalies occur in the

lower troposphere at day 0, which is a result of evaporation of convective downdraft. This

profile at day 0 is consistent with top-heavy diabatic heating as a result of stratiform

precipitation areas within deep convective systems. There is a cold spike at the level of 4.5km

with a minimum center of -0.6K. This cold spike is due to the reason that all the snow melts at

the freezing level leading to a strong cooling effect in a single level. By make the change

which allows a mix of snow and rain between the freezing level and 3K above this with the

proportion of rain increasing linearly from zero at the freezing level to one at the freezing level

plus 3K, the cooling center at the maximum rainfall rate has been reduced by 0.4K (Fig. 1b).

The cooling maximum now is in the lower troposphere instead of the middle troposphere.

Reducing cooling spike at the melting level has a direct impact on the detrainment rates. Figure

2 shows the detrainment rate relative to the intense rainfall events for the control experiment

and Expt.1. Reducing the cooling spike at the melting level has reduced the associated adaptive

detrainment by half near to the melting levels.

(a) (b)

Fig. 2 Detrainment rate(/s) relative to the heavy rainfall events. (60E-180E, 20S-20N). (a) Control experiment; and (b) Expt. 1.


4

With the change of the structure of massflux, in Expt.1, the convection is more efficient to

transfer moisture into troposphere. In Fig. 3, the difference of the moisture anomaly relative to

the intense rainfall events shows that there are more moisture being transferred into troposphere.

Fig. 3 Difference between Expt.1 and control experiment for moisture anomaly relative to the heavy rainfall events. (60E-180E, 20S-20N).

To further investigate the impacts of this change on the model simulation, firstly we studied the

relationship between precipitation and surface temperature.

(a) (b)

Fig. 4 (a) Precipitation (mmd−1) as a function of surface temperature. (60E-180E, 20S-20N). (b) The ratio of surface temperature in the Control experiment.

Figure 4 shows the relationship of daily precipitation and surface temperatures. In the control

experiment, the rainfall rates has a relatively constant value of 2mm/day between 295 K and

298K, then the rainfall rate starts to increase exponentially and reach a maximum value at

303K, followed by a rapidly decrease to 2mm/day at 304K. For Expt.1, the rainfall rate


5

increase is slower than the control experiment between 298K and 302K. The rainfall rate in

Expt. 1 is stronger than the control experiment when the surface temperature reaches 302K, and

the rainfall rate maximum is reached at higher surface temperature value, 303.25K, and also the

maximum value is about 2 mm/day stronger than the control experiment. The difference

between Expt. 1 and control experiment mainly occurs for the temperature range of 302K and

304K, with Expt. 1 having stronger rainfall rates. Figure 4b shows that the most rainfall events

are happening for the temperature range between 300K and 304K. So the difference shown in

Fig.2a has impacts on most rainfall events in the Tropics.

Fig. 5 The model bias of precipitation rate of the control experiment comparing to the GPCP observation. For four years of 2006-2009. (a) The 4 years averaged model bias; (b) the averaged model bias for DJF seasons; and (c) the averaged model bias for JJA seasons.

(a)

(b)

(c)


6

Figure 5 shows the rainfall bias of the control experiment comparing to the GPCP observation.

The four year avergared modle bias (Fig.5a) mainly has wet model bias in the west Indian

Ocean region and Western Pacific region in the southen hemisphere. In between the wet bias,

there is a dry bias in the region of the Maritime continent and the north part of Australia. For the

north hemisphere, there is dry bias in the Indian monsoon region and wet bais in the north west

Pacific. Figure 5b shows the 4 year model bias for the DJF season. Comparing to Fig. 5a, the

wet-dry-wet model bias in the Inidian Ocean, Maritime continent and South West Pacific region

is consistent with those for the 4 year averged model bias in those regions. Figure 5c is the 4

year model bias for JJA season, which shows that the model bias for the Indian monsoon region

and the North West Pacific in Fig.5a mainly comes from JJA season.

Fig. 6 The upper panels are the model bias for DJF and JJA seasons and the lower panels are the difference between the Expt. 1 and the control experiment for DJF season ( left panels) and JJA (right panels).

With the changes made with the melting level, the differences between Expt.1 and the control

experiment are shown in Fig.6, together with the corresponding model bias. For DJF season, the

difference between the two experiments has the opposite sign to the model bias for DJF season.

The wet tendency in the Maritime continent region and north Australia is beneficial for the MJO

simulation, which will be discussed in the Fig.7. There is not much improvement for the

summer model bias, and the wet bias in the north west Pacific region is getting worse in Expt. 1.


7

Fig. 7 Lag correlation of equatorial intraseasonal OLR (a,c) and 850 hPa zonal wind (b,d) onto a reference zonal wind time series at 90E in ( The upper panels are for the control experiment (a and b) and Expt. 1 (c and d)

With the improvement in the Indian Ocean and the South West Pacific region, the model MJO

simulation has been improved. To demonstrate the ability of the model to simulate eastward

propagating intraseasonal variability, in Fig. 7, we calculated the lead-lag correlation

coefficients between 20-100 days bandpass filtered data using the central Indian Ocean regional

time series and the associated near-equatorial data at all longitudes to produce a time-longitude

plot of correlation values for both OLR and 850 hPa zonal wind. For the control experiment, the

convection is mainly located in the Indian Ocean region, and fails to propagate eastwards

beyond the 120E. Also the 850 hPa zonal wind doesn’t show an organized eastwards

propagation. In Expt. 1, with the improvement of the convection simulation in the Maritime

continent region Model exhibits coherent eastward propagation across the Indian Ocean and the

West Pacific Ocean.

(a) Control (b) Control

(c) Expt. 1 (d) Expt. 1


8

With the modification of the snow melting in the convection code, the convection is stronger for

the higher surface temperature region in the Tropics. This change leads to the improved

precipitation forecast for the Indian Ocean region, Maritime continent and north Australia,

especially for the DJF seasons. As a result of this improved climatology, the model MJO

simulation is also improved.

4. CONCLUSION AND DISCUSSION:

Convective heating tendencies shows that the UM has a large cooling spike at the freezing level

due to all the snow being melted in the layer where the environmental temperature reaches the

freezing level. By allowing a mixture of snow and rain between the freezing level and 3K above

this, the convection becomes stronger for the higher surface temperature regions, especially in

the region of the Maritime Continent and north Australia. As a result, model bias for DJF season

in the tropical Indian Ocean and West Pacific region has been largely improved. The increased

convection in these regions helps the eastward propagation of organized convection and

improves the model climatology in the Tropics.

5. REFERENCE:

Illingworth, A. and Thomson, R. 2011: Radar bright band correction using the linear depolarisation ratio. Weather Radar and Hydrology, proceedings of a symposium held in Exter, UK, April 2011. (IAHS Publ. 3XX, 2011) Edwards, J.M. and Slingo, A. 1996: Studies with a flexible new radiation code. I: Choosing a configration for a large-scale model, Quart. J. Roy. Meteorol. Soc., 122, 689-719. Gregory, D. and Rowntree, P.R., 1990: A mass-flux convection scheme with representation of cloud ensemble characteristics and stability dependence closure. Mon. Weather Rev., 118:1483-1506. Puri, K. and co-authors (2010): Preliminary results from Numerical Weather Prediction implementation of ACCESS CAWCR Research Letters Issue 5,December 2010 Lock, A.P., Brown, A.R., Bush, M.R., Martin, G.M. and Smith, R.N.B. 2000: A new boundary layer mixing scheme. Part I: Scheme description and single column model tests. Mon. Weather Rev., 128, 3187-3199. Stratton. R., Willett, M., Derbyshire, S. and Wong, R. 2012: Convection Scheme,Unified Model Document paper, 27. Zhu, H. and Hendon, H. 2015: Role of large scale moisture advection for simulation of the MJO with increased entrainment. Quart. J. Roy Met. Soc. (Accepted manuscript online: 24 DEC 2014 04:30AM EST | DOI: 10.1002/qj.2510).


9

ACKNOWLEDGEMENT:

This research is supported by the Australian Government Department of the Environment, the

Bureau of Meteorology and· CSIRO through the Australian Climate Change Science

Programme

Documents

Effects of the changing heating profile associated with ... · EFFECTS OF THE CHANGING HEATING PROFILE ASSOCIATED WITH MELTING LAYERS 1 1. INTRODUCTION Convective heating and moisture