the hydrodynamic short-range of the local weather forecasting · the hydrodynamic short-range of...

The hydrodynamic short-range of the local weather forecasting

L.V. Berkovich, Yu.V. TkachevaHydrometeorological Research Center of Russia

9-13 Bol’shoy Predtechenskiy Lane, 123242, Moscow-242, RussiaE-mail: [email protected]

A method of the hydrodynamic short-range forecast of weather [ 1,2 ] has been improved. Asystem coupling a basic large-scale model with a three-dimensional atmospheric boundarylayer (ABL) model has been applied in the real time [ 3,4 ]. The Weather ForecastingSystem (WFS) operated at the Hydrometeorological Research Centre of Russia . The basicprinciple of the system lies in reconstruction of both synoptic-scale and mesoscale weatherpatterns from the output product of a large-scale prediction model by means of the ABLmodel algorithms and locally-adapted parameterisations: of the radiative heating effects onthe near-surface air temperature variation, of a moist convection for prediction of convectiveprecipitation as well as of some other physics . The system provides 48 h forecast of weatherpatterns for cities in Russia and neighbouring regions. It includes the forecast of temperature,moisture, wind, cloudiness and precipitation as well as the profiles of temperature, moisture,wind and turbulence characteristics in the lower 2-km atmospheric layer. The system of localweather forecasting, which is currently operated at the Hydrometeorological Research Centerof Russia , has been developed further towards improved simulation of sharp weatherchanges. The goal of the system is 48 h forecasting weather elements for locations in a limitedarea and also for the urban and suburban areas of a longed city (like Moscow). Theexperience gained at its application indicates that the skill of the resulting forecasts is superiorto that of the synoptic forecasts as 83 to 49 in 2002. However, proper simulation of sharpweather changes connected with significant variations in the meteorological fields calls forfine resolution in the background prediction; at the same time, the background forecast areashould be sufficiently large to simulate correctly large-scale weather processes developingduring 48 hours over this area.

The skill of WFS is presented here by verification statistics . The corresponding dataon temperature, precipitation, and wind predictions for Moscow in 2002 are given in Tablewhich reproduces the mean absolute forecast error of daily minimum and daily maximumtemperatures (

�T min ,

�T max ; ° C), 12-h precipitation amount (

�P r , mm), and

wind speed (�V , m⋅s-1).

We demonstrate the improvement of WFS forecasts over persistence ( forecast error- FE and persistence error -PE).)., because there are no currently

available another objective short-range weather forecasting techniques for Moscow, which wecould compare our forecasts with.

dT min dT max dP r dV

Projection (h)

0 --- 36 00-12 12-24 24-36 12 24 36 48

FE 1.7 1.9 1.0 1.1 1.4 2.0 1.8 2.1 2.0

PE 2.9 2.9 2.0 1.9 1.9 2.5 2.2 2.9 2.6

Operational forecasts calculated from operational database for Russian citiesdemonstrate about the same skill as shown in Table.

Further improvement to the local weather forecasting, along with development ofmore advanced models, calls for more complete initial observation data. In the nested fine-grid domain considered early [5], the number of currently available surface stationobservations is about 1-2 % of the grid-point number. Data from observations of this type areinsufficient for mesoscale objective analysis, which calls for utilization of high-resolutiondata provided by all other available observation systems. Accordingly, data onthermophysical and dynamical characteristics of the underlying surface and on their seasonalvariability with spatial resolution comparable to the model resolution are necessary. Alsofurther developments being partly in progress now are:

•improvement of the air humidity prediction through implementation of a simplifiedland surface hydrology model;•parameterized treatment of the impact of atmospheric fronts upon the evolution ofmeteorological variables in the boundary layer and free atmosphere; and•development and implementation of parameterised treatment of the anthropogenicheating impacts to improve the technique of detailed weather forecasting for a largecity .

REFERENCES

1. Berkovich L.V., Belousov S.L., Kalugina G.Yu, Tkacheva Yu.V. Developments in theshort-term dynamical weather forecasting system at the Hydrometeorological ResearchCenter of Russia. – Res.Act. in Atmos. And Oceanic Modeling, 1999, No.28, 5.7-5.8.

2. Berkovich L.V., Belousov S.L , Tkacheva Yu.V. Operational hydrodynamicalforecasting of meteorological variables and weather patterns for locations in Russia andcontiguous counties. – Russian Meteorology and Hydrology, 1998, No.4, 11-22.

3. Berkovich L.V , Tarnopolskiy A.G., Shnaidman V.A. A hydrodynamic model of theatmospheric and oceanic boundary layers. – Ibid., 1997, No.7, 30-40.

4. Belousov S.L., Berkovich L.V., V.A.Shnaydman. Short-range forecasting ofmeteorological variables and weather patterns. – Res.Act. in Atmos. And OceanicModeling, 2001, No.31, 5.1- 5.2.

5. Berkovich L.V., Belousov S.L., Kalugina G.Yu, Tkacheva Yu.V. The hydrodynamicweather forecasting for Moscow and Moscow Region.– Res.Act. in Atmos. And OceanicModeling, 2001, No.31, 5.3-5.4.

Arome, a new operational mesoscale NWP system for Meteo-France

Francois Bouttier ([email protected])Centre National de Recherches Meteorologiques

Toulouse, France

In an effort to improve early warnings of severe convection events, Meteo-Franceis developing a convection-resolving data assimilation and limited area forecast model(LAM) called Arome (Applications of Research to Operations at Mesoscale). This willbe used at 2 to 3km horizontal resolution over domains the size of mainland France(approx. 2000km wide) for short range predictions in replacement of the current Al-adin local adaptation model (9.5km resolution today). The deadline for operationalimplementation is 2008. The Arome model will be coupled with Arpege, a variable-resolution global model.

Arome will reuse available software as much as possible. The software basis willbe the IFS/Arpege/Aladin system developed jointly by Meteo-France, ECMWF andAladin partner countries. The dynamical core will be the so-called Aladin-NH spectralsemi-implicit, semi-Lagrangian, non-hydrostatic mass-coordinate compressible modelin a recent, very stable version that allows timesteps in excess of 1 minute at 2km res-olution when minimal physics are used. Most of the physical parameterizations willbe imported from the Meso-NH mesoscale research, and they will be different fromthe ones in the French global Arpege model. They will comprise: ICE3 cloud micro-physics with 6 cloud and precipitation prognostic species, a version of the ECMWFFM radiation scheme, 1-D and 3-D turbulent mixing schemes with prognostic TKE,Kain-Fritsch-Bechtold subgrid-scale convection, and a PRISM/ALMA-compliant in-terface with externalized versions of the ISBA land surface scheme allowing e.g. CO2flux management, variational soil moisture initialization and runoff routing for hydro-logical models. Advection of arbitrary passive scalar fields is provided for chemicalsand aerosols, and later the Meso-NH chemical equation compiler will be included intothe software. This should provide a state-of-the-art model for resolutions between800m and 10km, with good scope for future interaction with the mesoscale researchcommunity on physics development and cloud-resolving model studies.

The main novelty will be in the data assimilation shceme. Arome being for short-range forecasting (3 to 36 hours) will run analyses using the Aladin 3D-Var-FGATsystem, which is closely related with the ECMWF operational 4D-Var and 3D-Var-FGAT used in ERA-40. The obs processing will encompass all types already usedin the global ECMWF and Arpege systems, with an emphasis on IR polar-orbitingsounders borne on the EOS, NPP, NPOESS and Metop platforms. Currently, ATOVSlevel 1C AMSU-A radiances are assimilated, with good progress being made on HIRS,

AMSU-B and AIRS. Arome will also assimilate rain radar reflectivities, conventionaldata with high density near the ground, and geostationary MSG/SEVIRI as radiancesfor clear-air humidity analysis, and as cloud analysis using pixel classification.

Current algorithmic developments include: downscaling of background error co-variances, with flow-dependent vertical humidity structure functions and enhancedresolution at low level; blending of the latest large-scale 4D-Var analyses into themesoscale 3D-Var assimilation; an accelerated version of the analysis for use in now-casting; automated tracking of precipitation fields on radar imagery for data qualitycontrol and verification of precipitation forecasts; studies of initialization using dig-ital filters at high resolution; optimization of the coupling between the global andmesoscale NWP systems.

Preliminary testing using data assimilation and breeding experiments suggest goodpotential for improving over the global-model forecasts. This is well known in casesof slow-moving convective systems. In most weather situations, the whole interior ofa 2000km-wide LAM forecast will be affected by the large-scale boundary conditionswithin 48 hours. On the other hand, there is evidence that perturbations to the LAMinitial state do affect the LAM forecast over the same range. It means that althoughone cannot ignore large-scale forecast errors when developing a LAM NWP system,one can still obtain good added value in the regional forecast products by investinginto a sophisticated LAM model and data assimilation.

�� ! #"%$&�'��)(*�+��,��-�.�0/21��3��45��6(�7&$8��9 �:�07��'(�;<1��0�

=?>A@CBCBEDGFIH�B >KJMLON P�BC@ QR>%S�@ >:JUTV>ALXWY'B[Z:\^]_>)>I`:a�b

LRc >Rdfe�ghTV>:i_jAP�k cml i c ZITnB >IZ:\%D�oA@CB[`.@p\)q6iEP�JrFIH8s3ZtJru)vxw�@ y�F_Jz`6i{|ZKBCBC@*\)qI}~F.JU`6i?e��IF6\:i�e��)��6D�D��

a P�F:F.�t>:JzZ)d�@ Q)> l >)u�>�ZAJzTAvxP�>:\)d�Jr>f}~F_J�P0Z)dOTAv:s8>:\)dk?�I`6JrF_B FtqA��i{�>IZ)d�v)>:J��IF.J�>IT)Z)u�d�@p\)q�w�JrF6oA�_i�D�o:Jr>IZ:oxFI} c >)dX>)F_JrF.B FAqt� l >)u�>IZtJzTAvxP�>:\)d�Jr>:i

c >AB��tF6o:J�\)>:i6Y�o)u�d�JzZABC@[Z_�c ZAJzTtv��6i��A�A�t�

�m�O��%��V O�X¡�¢~� £I��¤¥��¦M�X�3�O£��6¡O��¦M¢+¦~¡¨§O��©��§��¦E¤��X��V�X ��ª £I«��X�¬�£)�K®6��¯�°+±�²�³´�µ�¶�·�¸~¹»º §¥� ¦3¼�¦~�X¦f£I«O¢~� ¡f½I£�� O��¦ ´C¾�¸ ¢z£��©'«��p£V O�f¬�£)�*�X¡O��¦M¢~��O¿^¦XÀO��ª[ª ¹mº §O�^©��X¢Á§O£)��©'«�ª £�ÂX�X�Ã�p�¨©-£V O�X¦0¢Á§��O��X�<¢z£�¦X«Vª � ¢�¢Á§��8��R�¨�_�V��Ä� ��X�^��V��Vª ÂO¦Á� ¦E��O¢z£f¡X£I��¢~� ¿I¼O£n¼O¦�Ä��*�O��¦ ¶ ¤V¼O¢�¼�¦~�X¦��'¦X©�£R£V¢Á§��X¦~¦0¡X£n�O¦M¢Á�Ä��¢0¢z£8��¦X¼��*�'¢Á§¥��¢0¢Á§O��R�[��¿I��£�¦M�X�3 O��ª £)¡O� ¢rÂÅI��ª �< �� X¦�¦X©�£)£�¢Á§�ª Â'��¡��p£V¦~¦�¢Á§O��V�¥�Vª Â�¦�� ¦ ¹�º §O�'¦~¡¨§O��©��8§��¦'¤¥��8¦X§�£��0�8¢z£3��©�«��*£� ��O£��6¡O��¦M¢z¦�£�¬�«��p��¡O��«¥� ¢~��¢~� £n�Æ��¦M¦~£)¡O�[��¢z�X�x�� ¢Á§8¡XÂ�¡�ª £n�¥� ¡8¡��p¡�¼Vª ��¢~� £I�<��*£I¼V�O�^ª £��Ç«��*�X¦M¦X¼��*��*�O��¦ ¶ �V�O�^§��¦-�Vª ¦M£8��¨©-£n�O¦~¢Á��¢z�X�<¦�¼�«�� £)�E«��¬�£)�r©��V��¡X��n¼��O¿<¡��¦~�X¦�£�¬��©�¤��X�V��X�¡X£I�� O��¡X¢~� £I� ¹

±�«�¢~� ¡6½I£��©��X¢Á§�£R�V¦_�V��Ä� ��¬~�*£I©h�?�R��*�X¡X¢6�V«�«�ª � ¡��¢~� £I�E£¨¬t¢Á§O�0±�«�¢~� ¡-³nª £��xÈK£n�O¦M¢Á�Ä��¢É ±�³)È_ÊA�XËI¼��¢~� £I� ¹6º §O�+¢z��r©�¦Ì��+¢Á§¥� ¦.�XËI¼��¢~� £I�E�¨�p�+�� ¥�Vª[¼��¢z�X��£� ��I�+¦MËI¼��*�E¤Vª £)¡�À ¶ �0§�� ¡�§� ¦�¡X£n�O¦�� V��*�X��¢z£^©�£� ��'�� ¢Á§Æ�f¦Á��O¿Iª �� ª £)¡O� ¢rÂ ¹3º §�� ¦� ��ª £)¡O� ¢rÂ<� ¦�¬�£I¼V�O��¼O¦Á��O¿Í�Ãª �O��¦M¢¦~Ën¼¥�¨�p�X¦�©-��¢Á§O£)� ´[Î�¸~¹EÏ ª £)¡�À� ��ª £)¡O� ¢~� �X¦-�¨�*�-¦�©-£)£V¢Á§O�X� ¶ �V�O�3�V��O¦M¢~�V�O¢~�V�O��£I¼O¦��Ä��f�Ä��¢z�¬�£)�*�X¡O��¦~¢0� ¦�«��*£R�n¼O¡X�� ¹

º §��E«��¬�£)�r©��V��¡X��£¨¬I¢Á§O��O�X��O£��6¡O��¦M¢_¦~¡¨§O��©��0§¥��¦�¤��X��¡X£I©'«¥�¨�p��E�� ¢Á§0¢Á§¥��¢:£¨¬t¢Á§O�«��p�� ¥� £n¼O¦�®6��¯�°+±�²�³�� X¡X¢~� £n�E¦~¡¨§O��©�� É �O£V¢Ì��O¡¨ª[¼O�R��¿�¢Á§��£n¤UÐÑ�X¡X¢rÒz£)�Ä� ��¢z�X�'¡X£I��¡X��«O¢Á¼��Vªª � ¬��0¡XÂ�¡�ª �'©-£)�V��ª Ê ¹.º §O��«��C¬r£)��©��V�O¡X��£�¬:¢Á§O��¢��6£�¦M¡�§O��©��X¦+�+��¦+¡�£I©'«¥�¨�p�X�'£� ��:��«�� £)�£�¬ ¾ ©-£n�O¢Á§O¦ ¹_º §O��X�m¦~¡�§��©-��¦�§O£��6�X��¿��Ä�Vªt��©'«��*£� ��©-�¨�O¢��¬�£)�*�X¡O��¦M¢�¦�ÀO��ª[ª ¶ �V�O��Vª ¦~£��©�£I��¦~¢Á��¢z�X�E¦�¼�«�� £)�:«¥�¨�C¬�£)�r©��V�O¡X��£I�0�0�V¼�©'¤¥�¨�I£�¬:À��Â0¢z�X¦~¢:¡O��¦M�X¦ ¶ ��¡�ª[¼O�)��O¿�¢r�6£«��p��¡O��«¥� ¢~��¢~� £n��X O�¨�O¢z¦E��¦~¦M£)¡O�[��¢z�X�'�� ¢Á§-¦M�X ��p��½I£)£)�R��O¿ ¹ ±��E¢Á§��¤��¦�� ¦+£¨¬_¢Á§¥� ¦��X ¥�Vª[¼��¢~� £I�¢Á§O��£n«O¢~� ¡�½�£��»¦~¡�§��©-�f§¥��¦'¤��X��f¡�§O£V¦~��<�*��«Vª ��¡X��¢Á§O�3«��*�X �� £I¼O¦�®6��¯�°+±�²�³3�� X¡X¢~� £n�¦~¡�§��©�� ¹+º §��¤��O�ÓÅ�¢0£¨¬Ì¢Á§��O�X�Ô�� X¡X¢~� £I�'¦~¡�§��©��-£V O�¨�_¢Á§O��«��*�X �� £I¼�¦0¦~¡�§��©-�'� ¦��n¼O�©��Vª Â�¢z£3¢Á§O��R� Õ�¡�¼Vª ¢�Â3��<¼��V©�¤�� ¿I¼O£n¼O¦�ª Â8� ��¢~� ¬�Â¥��¿Í¡�£I�O¢~� ¿n¼O£n¼O¦��8��*�O��¦ ¶ �0§�� ¡�§Ö~×AØ[Ù~ÚMÛ�ÜrÝXÞ�ß[Ú:àXá�á�ÜrÙ~ârâMãRä.Ù~ß[Ø[Ø*å æ:ÝÁç:Ø[Ù~Üzè.é0Ù~ÛrÝÁê�ÚMÙXå ÚzÝXéë ×AØ[Ù~ÚMÛ�ÜrÝXÞ�ß[Ú:àXá�á�ÜrÙ~ârâMã�ìAØ[ß[íXÙ�å îtß[ÙMÜ�ÚMÙ�è.é0Ù~ÛrÝMê�ÚMÙXå ÚMÝ�éï ×AØ[Ù~ÚMÛ�ÜrÝXÞ�ß[Ú:àXá�á�ÜrÙ~ârâMãRðGå ñ�Ù~Ù~á�è.òVÝXé�å óXÝÁí�å àXô

õ

� ¦�«¥�¨�p¢~� ¡¨¼�ª �¨��ª Â�¢Á�*£I¼V¤�ª �X¦M£I©��£I��¦X©��Vª[ª��£n©��¦ ¶ ¦X¼O¡¨§��¦�¢Á§O�� ¹)º §V¼O¦E� ¢�� ¦0¡X£n�O¡�ª[¼��X�¢Á§¥��¢�¤Vª £R¡¨ÀÁÒ�¤��¦~�X�^©��X¢Á§�£R�V¦E��*��ª ��ÀÁ��ª Â'¢z£<¤��¦X¼V«��Ä� £)�+¢z£�£n¤UÐÑ�X¡X¢rÒÁ¤¥��¦~��^©-��¢Á§O£)��¦��¢Á§O�©��zÐÑ£)�Ä� ¢rÂ�£�¬6¡O��¦~�X¦ ¹

��

´�µÁ¸ ¯ ¹��:¹.Ï £��0ª ��+�V�O�xÈ ¹��:¹�� ¨�p¡X� ¶�� °6�X ��ª £I«V©-��¢0£�¬��'¦X§�£)�p¢��V�O¿V�� ³8�Vª ¿V£)�Ä� ¢Á§V©¤��¦~�X�x£I�8£n«O¢~� ¡½�£��¢z��¡�§�� ËI¼O��¦�� ¶��! �"$#&%'��(�)�*,+�'%'-.0/1%'243�5��76984:; <(�8>=@?0A>BÉz·0C0CV· Ê

´C·�¸ ¯ ¹��:¹6Ï £��0ª �� ¶ È ¹@�:¹@� � �¨�p¡X�8�V��3� ¹ED �X�X� ¶�� °6�X ��ª £I«V©-�¨�O¢E£¨¬��¦�§O£)�*¢��V�O¿V�� ³�Vª ¿V£)�Ä� ¢Á§V©�¤¥��¦M�X�3£I��£n«O¢~� ¡�½I£��Ô¢z�X¡�§V�¥� ËI¼��X¦�� ¶GFIH4JK�/L � M�'NO M8�P8>H4:M.0243�8Q�RES>N0:M80318IT0SÉz·0C0CV¾ Ê

´C¾�¸Ï+¹ � ¹U�A¹UV £)��V�O� Ï+¹ ® ¹0D ¡�§V¼��O¡¨À ¶_É�µXW0Ynµ Ê ¶�� °6��¢z��©��¥��¿�£I«�¢~� ¡O�Vª�½I£��@� ¶>Z�:� M/1[\%I/1243] .\ M�I3131/ T^�'.%'�X¶@_^`9a$_cb9d>e'f^g9h

iLjlknmporq<sUt�evuws^x�yMs^z'{@sU|&sU}�x�~�o��E�r�X�>|>a��'_U�^�9b��!a@�^��9t@��z'��9�@�!y��<��U��4�9��e�@�@�!sl�!yM�^}��ar��@��0��9�7�7�9��U��@�9��0��U��4�� 9¡w¢��£��¥¤��¦4�7�9�@a§f9��'_9�!a�d9�ce`9`

¨

�� "!#��$ &%'�(�� )�*,+.-,��/0 �/��213� 4� 516��87��,9;:�<��=��>��/��)

?A@CB�DEGF�HJI6KML DN @POQ@ F"H ? I�F DI3R @CS�T�U V IW�DI S"@�T X I B I�F X HJI V(VZY[U I�F Y I V HJIPL @�\ I T]T IÎ X HJIPL _ ` X=B E VaSCbJcI T ICdMe�F U f I T�V=U X8DIgHGh6iQh DI�j�I Ykc@ml EGF X=TnDI @ L

o�pkqr8satvuAwAx5y"saz uMr\{b I S h T]S E V IQE�| X=bJU V�} E T]~�U V�X E V�X h"H�� X=b I S�T I�H U Y=X�@ j U L �U X �kE�|;L U BÛ X I�H�� @�T I @PB E5HJI�L V�� R"` l�V��{U F @ F U HJI @ L U � I�H Y EGF��X I=� X d }Qb I T I f I T�U ��Yn@�X�U EGF3H @�X�@6@ F"H B E5HJI�L5I T]T E T�V�@�T I.F[E X@ F U V�V h[IC��E T]B�@ L ��d X=bJU V�~"U F"H�E�| V�X h"H�� U V.Yn@�T]T�U I�H�EGh X

ja� B I @ F V E�| @��v\�}�U F�KJ� S I T�U B I�F X�� d }QbJU Y"b�Y EGF V=U V�X�V�U FY E BkS"@�T�U F[� T hCF V E�| @ � U f I�F B E5HJI�L U F U X�U @ L U � I�H }�U X=b�V L U � b�X L �H U � I T I�F XgU F U X�U @ L Y EGF"H U X�U EGF V E f I T�X=b I VZ@CB I�HJE B�@JU F � `(F��X=b I V I Xg@ L �#��J�J� � � \{bJU V^@CSCS�T E @"Y"b d b E } I f I T d$HJE8I V F[E X@ L L E }�X E U F f I V�X�U � @�X I X=b IPI � I Y=X E�| X=b IPEGF[Ia� };@ ��F[I V�X�U F[�S�T E Y I�HGh T IÊGF X=b I S�T I�H U Y=X�@ j U L U X �� \{b I ��U �� T E X=b I T KJ�[�S I T�U B I�F X��k�� K �$};@[V HJI f I�L E S I�H U F(E T HJI T8X E @ F @ L � V I X=b II � I Y=X E�|5EGF[Ia� };@ ��F[I V�X�U F[��EGF U F X I T F @ L f"@�T�U @ j U L U X � @ F"H S�T Ia�H U Y=X�@ j U L U X �QE�|�R"` l�V(U F @�S I T |�I Y=X � B E5HJI�L Y EGF X I=� X �;� X�Y EGF��V=U V�X�V E�| @�bJU � b � T I V EGL h X�U EGF V=U B hCL @�X�U EGF U F @ L @�T �JI�HJE B�@JU FX=b"@�X(U V(Y EGF V=U HJI T I�H @[V(@3T Ia|�I T I�F Y I T hCF5d @ F"H Y E BkS"@�T I�H X EE X=b I TAT hCF V;S I T |�E T]B I�H }�U X=b�X=b I VZ@CB I B E5HJI�L"jCh X$}�U X=bJU F @VaB�@ L L I T HJE B�@JU F @ F"H6F[I V�X I�H U F X=b I T Ia|�I T I�F Y I V=U B hCL @�X�U EGF5�\{b I �� K�d5hCFCL U ~ I X E X=b I \�}�U F�KJ� S I T�U B I�F X d @"YnY EGhCF X�V |�E TX=b IgI � I Y=X E�| X=b IgEGF[Ia� };@ �PF[I V�X�U F[�PEGF S�T I�H U Y=X�@ j U L U X ��Apk�M 5¡�¢8t�z £6¢5r8s=¤5¥G¦�tv¤G£6¢J§{u8t�¨\{b I^© @ F @ H U @ F ? In� U EGF @ LC©�L U B�@�X I l E5HJI�L � © ? © lP� HJIa�VZY"T�U j�I�H U Fª© @ � @,@ F"H>R @CS�T�U V I � ��J�J� �PU V h V I�H«|�E TQ@,V Ia�T�U I V E�| V=U B hCL @�X�U EGF V�}�U X=b^¬ �"� ~�B'b E T�U � EGF X�@ L�� T�U H VaS"@"Y[U F[�Md

��,L I f I�L V�U F X=b I f I T X�U Yn@ L d @ F"H @� � b F[I V�X�U F[��| T I�®Gh[I�F Y ��` ��T�V�XPU F X In� T@�X�U EGF U V�B�@ HJI�|�E T�@ªB EGF X=b«U F @ HJE B�@JU FE�|k��J¯[�5��J¯m� T�U H S E U F X�VPU F X=b I b E T�U � EGF X�@ L �]Y I�F X I T I�H U F��I } KMF[�GL @ F"H � d�F[I V�X I�H }�U X=b ��©�KA° @ F @ L � V I V E�|g±"I�j T h��@�T �k��J� � \{bJU V�bJU � b � T I V EGL h X�U EGF V=U B hCL @�X�U EGF T Ia|�I T I�F Y I T hCFj�I Y E B I V�X=b I �]X=T h X=b"�kX E }QbJU Y"b E X=b I T�T hCF V�}�U L L�j�I Y E B �S"@�T I�HM� \{b ImEGh X=S h X�� I�L H VmS�T E5HGh Y I�H²ja� X=bJU V�T Ia|�I T I�F Y IT hCF @�T I X=b I�F�h V I�H X E�H T�U f I V=U B hCL @�X�U EGF VgS I T |�E T]B I�H.E f I [email protected]�@ L L I T HJE B�@JU F � ��³J³[�5��³J³ U F X=b I b E T�U � EGF X�@ L d ~ InI SJU F[�X=b I f I T X�U Yn@ L @ F"H b E T�U � EGF X�@ L T I V EGL h X�U EGF�hCF X EGh Y"b I�H � L E��Yn@�X I�H U F X=b I Y I�F X=T I6E�| X=b I�L @�T �JI T HJE B�@JU F5� \{bJU V^V I X h SS I T]BÛ X�V3X=b I Y E BkS"@�T�U V EGF�E�| X=b I�EGh X=S h X E�|�j�E X=b,V=U B h��L @�X�U EGF VPU F X=b I VZ@CB I T In� U EGF @ F"H X=b I T Ia|�E T I @[V�V I V�V I VgX=b I@ j U L U X �kE�| X=b IÊGF[Ia� };@ �PF[I V�X�U F[� X E T I S�T E5HGh Y I X=b I T I V hCL X�VE�| X=b I�L @�T �JI T HJE B�@JU F5�6´ U F Y Imj�E X=b.V=U B hCL @�X�U EGF V h V I X=b IVZ@CB I�|�E T]B hCL @�X�U EGF � H��CF @CBÛ YaV d SCb � V=U YaV d T I V EGL h X�U EGF5d(FCh��B I T�U YaV d(I X�Y�� d�H U � I T I�F Y I V�U F T I V hCL X�V�Yn@ Fµj�I @�XvX=T�U jCh X I�HhCF @CB j U �Gh[EGh V L � X E X=b I3F[I V�X�U F[� X I Y"b F U ®Gh[IC�� F#E T HJI T(X E B�@C~ I f"@ L h[I V3V�X�@�X�U V�X�U Yn@ L L � V�X�@ jCL ICd T I V hCL X�V} I T IPEGj X�@JU F[I�H.|�E T;¶�¬�T hCF V E�| X=b I VaB�@ L L �vHJE B�@JU F B E5HJI�LU F X In� T@�X I�H6HGh T�U F[� ¬ H @ � V d8I @"Y"b EGF[I V�X�@�T X�U F[�PEGF V h YnY I V �V=U f I�H @ � V E�|$±"I�j T h @�T �g��J� �A` B E T I�HJI X�@JU L I�HQHJI VZY"T�U S�X�U EGF

E�| X=bJU V I=� S I T�U B I�F X$Yn@ F�j�I�|�EGhCF"H U F{HJIQKML U @ I X$@ L � �]¶ ³J³ ¶C� �

·[¸J¹Zº�º�»½¼�¾�¹=¿aÀaÁ ¿�Â>Ã=ÄnÅ Æa¹Zº3Ã�ÀaÀaº�»½¼�¼�Ç0ÈCÉaÊkËÌ�Í�ÎaÏ«Ð�Ñ ËÒ ÉnÓgÔMÕ Ì�Öa×=ÏØGÙ�Ú�Û�Ó�Ü Ö Õ É ÍZÌ=Ý Ó«Þ=Þaß�à Ï�Ý�áµâ�ãZÏ Õ�ä�Õ ÌnÌ�å½Ï�âná�æ Ó�çZè á½ãêé[Ñ ÌnÌ ÕÓÛ Ì�Í�á Õ�ËÏ É Ñ Ó8ëCÙ�ÓCì�í�Ú�çaî8è�ÓJï8É Í É Î É æ�Ðnð Ê�É=ñ Ñ ò Õ Ï�Ñ ñ É=ó Ý�ô É æ ÖZõ ÉaÊ æ ô É

öAp�÷�z ø�¢8t½¢5r5y"¢Jùmú�¢Js�§(¢J¢5r«snû8¢�ü{§gz r«¤Gr5wªsnû8¢�ý�z þ5ÿý�tvu5snû8¢8t��M 5¡�¢8t�z £6¢5r8s�ù�� ±"E T5X=b I \�}�U FkKJ� S I T�U B I�F X�X=b I U F X I T F @ L f"@�T�U @ j U L U X � B�@ �

j�I�HJI � F[I�H @[V6X=b I ?�l ´>H U � I T I�F Y I�j�I X]} InI�F X]} E T hCF VV�X�@�T X I�H }�U X=bm�v@ L B E V�X��gU HJI�F X�U Yn@ L U F U X�U @ L Y EGF"H U X�U EGF V EGF X=b IVaB�@ L L8HJE B�@JU F5d @ F"H U X{Yn@ F6j�IÎ=� S�T I V�V I�H @[V

��! #"%$ &('%) *,+.-}Qb I T I �� @ F"H �� T I S�T I V I�F XQb E T�U � EGF X�@ L � I�L H V | T E B�X]} EH U � I T I�F XgT hCF V E�| X=b I VaB�@ L LGHJE B�@JU FP´ X=b"@�X�@�T I^|vhCF Y=X�U EGFE�| X�U B I0/ �� j�I U F[� Y EGF V=U HJI T I�H @[V � T EGhCF"H X=T h X=b d @ F"H ��@[V3X=b I S I T X h T j�I�H T hCF5� \{b I X I T]B "%$ &(' d X=b I#H U � I T I�F Y Ij�I X]} InI�F X=b I V I X]} E � I�L H V d T I S�T I V I�F X�V21 "%$ &('�*435�760- X=b IH U � I T I�F Y I U F U F U X�U @ L Y EGF"H U X�U EGF V d @ F"H "%$ &('�*4398:60- @ FkI V�X�U �B�@�X�U EGF3E�| X=b I U F X I T F @ L f"@�T�U @ j U L U X � @�X{V hCj V I�®Gh[I�F X�X�U B I V �\{b I X=bCT InI X I T]BQV�U F X=b IPI=� S�T I V�V=U EGF @�T I3|vhCF Y=X�U EGF6E�| X=b IU F X In� T@�X�U EGF X�U B I @ F"H @�T I Yn@[V I�HJI S I�F"HJI�F X � \{b I T E8E X �B I @ F�� V ®Gh @�T I.H U � I T I�F Y I6|�E T;X=bJU Vk\�}�U F«KJ� S I T�U B I�F XkYn@ Fj�IÎ=� S�T I V�V I�H @[V

;=<?>A@B �DCFEG*H��IJ��-,@K�DCFEL"%@$ &(' �NMO@$ &(' )P*RQ�-}Qb I T I X=b I V h BkB�@�X�U EGF U V#S I T |�E T]B I�H�E f I TQX=b I�HJE B�@JU FS �UT b I�F V I f I T@ L Yn@[V I V5V @�T I @�f"@JU L @ jCL I X=b I6I�F V I B jCL I@�f I T@ �JI T E8E X � B I @ F�� V ®Gh @�T IPH U � I T I�F Y I Yn@ F,j�I T I }QT�U XvX I�F@[V ;=<?> @B �WCYXZM @$ &(' � M @$ &('�[ *H\0-\{b I X I T]B EGF X=b I T�U � b�X.Vab EGhCL H�j�I U F X I T]S�T I X I�H @[V.X=b II�F V I B jCL I @�f I T@ �JI�E�| X=b I VaS"@�X�U @ L f"@�T�U @ F Y I @[V�V E Y[U @�X I�H}�U X=b{X=b I U F X I T F @ L f"@�T�U @ j U L U X � @�X�@ � U f I�F U F X In� T@�X�U EGF X�U B IC�] ��_^`� acbR^`�ed(�4��f� ��T b I�F X=b I Y E BkS"@�T�U V EGF;E�| X=b I VaB�@ L L �vHJE B�@JU F T hCF U V�Yn@�T �T�U I�H�EGh XQ@ � @JU F V�X(X=b I.L @�T �JIa�vHJE B�@JU F T Ia|�I T I�F Y I T hCF5d @ HCH U �X�U EGF @ L V EGh TY I V E�|�I T]T E TAV h Y"b h V�X=b I^F[I V�X�U F[� B I X=b E5HMd X=b I

F[I V�X�U F[�.| T I�®Gh[I�F Y ��d @ F"H X=b I T I V EGL h X�U EGF U F X=b I.j�EGhCF"H @�T �Y EGF"H U X�U EGF V.Vab EGhCL H j�I Y EGF V=U HJI T I�HM� ´ U F Y I X=b I U F X I T F @ Lf"@�T�U @ j U L U X � Yn@ FCF[E X j�I V I S"@�T@�X I�H#| T E B E X=b I T�V EGh TY I V E�|I T]T E T d @ L LAE�| X=b I B @�T I U F Y L h"HJI�H U F @.V=U F[�GL I X I T]B � \{bJU VYn@ F6j�IÎ=� S�T I V�V I�H @[V

��g� #"�h?) *4ij-}Qb I T I X=b I V hCj VZY"T�U S�Xlk V�X�@ F"H V |�E TkVaB�@ L L �vHJE B�@JU F T hCF5dm |�E T L @�T �JIa�vHJE B�@JU F T Ia|�I T I�F Y I T hCF5d @ F"Hln |�E T EGF[Ia� };@ �

�� J+� ��?��(d� �A�R�Hd(� d�� F�� (��A� ] ��d(��4��d(�� 2��(� � ��e�� %�!��d(��4��#"%$�&(bR��'`��(d(��R�(��*)!+9�� ,�_��?�4��f^9^9�`�F[I V�X�U F[�M�m±"EGL L E }�U F[� X=b I VZ@CB I S�T E Y I�HGh T I @[V3U F X=b I S�T Ia�fJU EGh V{V I Y=X�U EGF X=b I ?�l ´�E�| X=b I ��U �� T E X=b I T KJ� S I T�U B I�F XYn@ F6j�I }QT�U XvX I�F @[V

;=<?> @-�- � M @h ) **.�-}Qb I T I X=b I X I T]B EGF X=b I T�U � b�X � b"@ F"H V=U HJI T I S�T I V I�F X�V�X=b IVaS"@�X�U @ L f"@�T�U @ F Y I @[V�V E Y[U @�X I�H }�U X=bkU F X I T F @ L f"@�T�U @ j U L U X � @ F"HE X=b I T�V EGh TY I V E�|(I T]T E T �.´ U F Y I X=b I ��U �� T E X=b I T KJ� S I T�U �B I�F X�@"YnY EGhCF X�V |�E T�B E T I V EGh TY I V E�|AI T]T E T�X=b"@ F X=b I \�}�U FKJ� S I T�U B I�F X d X=bJU VQU F[I�®Gh @ L U X � B h V�X�b EGL H

;=<?> @B0/ ;=<?> @-�- [ *(10-� F X=b IP|�EGL L E }�U F[� V I Y=X�U EGF5d�j�E X=b�?�l ´ f"@ L h[I V�}�U L L�j�I6I V �X�U B�@�X I�H @[Vk@ |vhCF Y=X�U EGF.E�| VaS"@�X�U @ L VZYn@ L I }�U X=b H @�X�@ | T E BX=b I V=U B hCL @�X�U EGF V HJI VZY"T�U j�I�H U F V I Y=X�U EGF ¶ �2 p43�¢Jù�xG¥ s�ù\{b I6H U � I T I�F Y Imj�I X]} InI�F X=b I6�J�J³"� b ° @3f E T X�U Y[U X � � I�L H V

E�| X=b I T Ia|�I T I�F Y I T hCF @ F"H X=b I S I T X h T j�I�H V=U B hCL @�X�U EGF Vb"@[V j�InI�F�HJI Y E BkS E V I�H VaS I Y=X=T@ L L ��d @ F"H X=b I�F0F[E T]B�@ L �U � I�H�ja� X=b I @�f I T@ �JI f E T X�U Y[U X � VaS I Y=X=T h B E�| X=b I T Ia|�I T I�F Y IT hCF5� \{bJU V F[E T]B�@ L U � I�H VaS I Y=X=T h B B�@ �mj�I U F X I T]S�T I X I�H @[VX=b IgF[E T]B�@ L U � I�H T E8E X � B I @ F V ®Gh @�T I�H U � I T I�F Y I �Z?�l ´ � j�Ia�X]} InI�F X=b I � I�L H V3@[V3@ |vhCF Y=X�U EGF�E�| };@�f I�FCh B j�I T �65�I T IX=b I };@�f I�FCh B j�I T�V{@�T IPF[EGF"H U B I�F V=U EGF @ L U V I�H }�U X=b.T I VaS I Y=XX E @ Fg�J³"�½� T�U H S E U F X�}�U HJI�HJE B�@JU F�h V I�H X E3HJE X=b I�±"EGh T�U I T\�T@ F V |�E T]B �05�I�F Y I87 � Qc6 d$|�E T I=� @CBkS L ICd Y E T]T I VaS EGF"H VX E @k};@�f I�L I�F[� X=b E�|��J³ ~�B �\{b I X I BkS E T@ L5I f EGL h X�U EGFPE�| V I�L I Y=X I�H };@�f I�FCh B j�I T�VÛ VH U VaS L @ �nI�H U F�± U �M��g|�E T$X=b I ��U �� T E X=b I T KJ� S I T�U B I�F X �� XYn@ F.j�I V InI�F X=b"@�X(X=b IÎ T]T E T � T E };X=bPU V^bJU � b L �3HJI S I�F"HJI�F XEGF };@�f I�FCh B j�I T /8j�I U F[� f I T �kL U BÛ X I�HQ|�E T5VaB�@ L L };@�f I�FCh B �j�I T�Vg@ F"H @ L B E V�X^T I @"Y"bJU F[� X=b I Y"T�U X�U Yn@ L f"@ L h[I @[V�V E Y[U @�X I�H

�� Q9 �O�(��,�4: � a0�8; ] �A�<��d(�U�4��?�� b��F�}�U X=b hCF Y E T]T I�L @�X I�H V=U �GF @ L V��>= Q � |�E T L @�T �JI };@�f I�FCh B j�I T�V �\{bJU V j�I b"@�fJU E T dn|�EGhCF"H U F @ L L f"@�T�U @ jCL I V$V�X h"H U I�HMd Vab E }�VAX=b"@�XEGF[Ia� };@ ��F[I V�X�U F[� B E V�X L � Y EGF X=T EGL VPX=b I#I T]T E T � T E };X=b�@�XL @�T �JI VZYn@ L I V d�HJI VaSJU X I X=b I S�T I V I�F Y I6E�| @ L L };@�f I�FCh B j�I T�VU F X=b IPj�EGhCF"H @�T � Y EGF"H U X�U EGF V �\{b I X I BkS E T@ L�I f EGL h X�U EGFPE�| V I�L I Y=X I�H };@�f I�FCh B I T�V |�E TX=b I \�}�U F>KJ� S I T�U B I�F X�U V H U VaS L @ �nI�H U F²± U �M� ¶ ��` V#U F± U �M��Cd U X6Yn@ F�j�I V InI�F X=b"@�X � T E };X=b«U V�bJU � b L ��HJI S I�F��HJI�F X EGF };@�f I�FCh B j�I T�V dAj�I U F[� VaB�@ L LJ|�E T$X=b IPL @�T �JI VZYn@ L I V@ F"H.L @�T �JI�|�E TAX=b I VaB�@ L L VZYn@ L I V �?5�E } I f I T d U X;Yn@ FPj�I V InI�FX=b"@�X{?�l ´ f"@ L h[I V�VaB�@ L L I TGX=b"@ F X=b E V I(E�| X=b I ��U �� T E X=b I TKJ� S I T�U B I�F X�@�T I @�XvX�@JU F[I�H @�X;@ L L };@�f I�L I�F[� X=b[V d V h[�J�JI V�X�U F[�X=b"@�X{X=b I U F X I T F @ L f"@�T�U @ j U L U X �3E�| X=b I B E5HJI�LGHJE8I V F[E XQ@"Y �Y EGhCF X |�E T{X=b I X E X�@ L f"@�T�U @ j U L U X ��EGj V I T f I�H U F�± U �M�>�Cd @[VH U VZY h V�V I�H U F V I Y=X�U EGF �@ pBA�uMr5yJ¥ x8ùnz uMr8ù©�E BkS"@�T�U V EGF#E�| X=b I ?�l ´ f"@ L h[I V EGj X�@JU F[I�H�| T E B2X=b I��U �� T E X=b I T KJ� S I T�U B I�F Xg@ � @JU F V�XQX=b E V I3| T E B&X=b I \�}�U F

KJ� S I T�U B I�F X�U F"H U Yn@�X I V�X=b"@�X�X=b I�|�E T]B I T,@"YnY EGhCF X�V |�E T@ L @�T �JI T²U F X I T F @ L f"@�T�U @ j U L U X �� \{bJU V�T I V hCL X V h[�J�JI V�X�VX=b"@�X²B I @[V h T I>E�| S�T I�H U Y=X�@ j U L U X � EGj X�@JU F[I�H U F @�\�}�U F��KJ� S I T�U B I�F X S�T E X E Y EGL Y EGhCL H E f I T I T I V�X�U B�@�X I U X�V T I @ Lf"@ L h[IC��±"E T{X=bJU VmT I @[V EGF5d U X�B�@ ��j�I Y EGF f I�F U I�F XP@�X�X=b IX�U B I�E�| X I V�X�U F[� V I�F V=U X�U fJU X�U I V�X E Y I T X�@JU F S"@�T@CB I X I T�VkU F @T In� U EGF @ L B E5HJI�L X E�I f"@ L h @�X I }Qb I X=b I T�@�\�}�U F.E T�@��U ��T E X=b I T KJ� S I T�U B I�F X�@�T I B E T I @CSCS�T E S�T�U @�X I(|�E T�X=b I U F X I T �I V�X E�| @.S"@�T X�U Y hCL @�T�T I V I @�TY"b �3�¢�¦½¢8t½¢5r5y"¢Jù�C`(F X=b I V I Xg@ L � d��J�J��/<D ��(�(� �%�E� ��6F=�%de��G),��(�E� d<H�EGL �¶ ��dA��J�"� ¶ ³ ¶ �© @ � @ d W � @ F"H ? �ªR @CS�T�U V ICd��J�J��/ �?d(�� I�%��

� ��(� d J�KML$d �¬ �a� ¯ ¶ �HJI,KML U @ d ? � d ? ��R @CS�T�U V ICd @ F"H � � W I�F U V d ¶ ³J³ ¶�1 �?d(��

I�%��N� ��(� dOJ9PMQ�d ¶ ³J³J¯,R ¶ ³ ¶J �

�� ! "�� #��$�%��'&(�� )*�� ,+��-.�/��!0�213�4��&5�6&7�98:��13��&5�6�;��<3��=��><?��13��'&*��@-.��&A��B��'�� $��C

D/EGFH<3EJIK8��LNMPO=<3QR�=LNSTMU8V��WKLNXYL>�[ZA\P]^W7MYF_\PFGà��QRF4bcQdFfehgGMYEi�jlk�monRp*jlq"rsjutwv*xzy'p|{}rsx~mojK�$�Trsjutwr�nw�K��NiY��K��G�:��v5�R�Kmoj(��t�p5j��z�~��

� ��P�P�G�[�¡ �¢�£N��G¤¥�¦}§©¨�ª©¨|ª¬«}¦7P®�¯4°�±"²w«�²wª©´³�µ(¶|²N·7´´¸c¹}´º6¼»½¨|¦7¾¹l¿�À~³h¦(»~¼Á

³h´¸5ª©¾¹G¶|¸(¹J§©Â(¸J²�À½¸}ÃÝ¥4¦�§©À½»o¿5Ä|Å�Å*Æ4¶�ªÈÇÉ�°ÈÊ�ËÍÌu§d¶�Ã¼ª©¨�¸hRª�¶�»oÎ½¿Æ¾Ï|Ï�Ð*Ñ�Ò/Â�¶|¸K¹�Ìl§d¶|Ã¼ª©¨|¸u¿ÈÄ�Å�Å�Æ¾ÓRÎ�Ô¬ªGÃ´¨|¸(²�À~²wª©²c¨�ÕÆ¾Å"³h´³c·7R§©²Ö À×ª©µØª Ö ¨:§©´Ù�À~¨�¸(¶|»Ú³h¨*¹�´»~²UËÍ°Nªd¶h¶�¸K¹[¯�®(Û"ÓP¶�ª4Ü6Ð�Ý*³�§©¼²�¨�»~Â}Áª©À~¨�¸Ø¨sº6R§4Ç4¨|§�ª©µ[¥4³h¼§©À~Ã¾¶:¹�¨�³�¶�À½¸lÎÈ°�¶�ÃRµ"³h¨*¹�´»AÃ´¨�¸5ª�§©À~·(Â�ª©´²Þ º�VËàß|ÓY³h´³J·7¼§©²4À~¸}À~ª©À½¶sª©¾¹ Ö À×ª©µ,¹}À×áA¼§©´¸5ª�¦7¼§�ª©Â}§©·7¾¹"À~¸(À×ª©À½¶|»Ã´¨|¸K¹}À×ª©À~¨�¸}²PËoÔ�É�²dÓuÃR§©¾¶�ª©¾¹U·5«/ª©µ(È·}§©¼¾¹}À~¸(ÙY³h¼ª©µ}¨*¹�ËÍÌu¨|ª©µG¶|¸(¹â ¶|»~¸(¶¾«5¿�Æ¾Ï|Ï�ã6ÓRÎØÔ¬ªU§©Â(¸}²Uª Ö À~Ã´ä¶Ø¹(¶´«�ËÍÅ�Ï|åä¶�¸K¹�Ä5Æ´åsÓ Ö À×ª©µ�¶Õæ¨�§©´Ã¾¶�²wª�»½¼¸(Ù|ª©µ�¨�Õ�ç|ã�µ}¨�Â�§©²´Î�Ì�µ�§©´Vª¬«}¦7¼²J¨�Õ®*¯�°È±9¦�§©¨*¹�ÁÂ(ÃRª©²´¿cÀ~¸(Ã¼»½Â(¹}À~¸(Ù$´¸(²�´³J·(»~�³h¾¶�¸7è�²�¦�§©¾¶�¹A¿c²�¦(¶|Ù|µ(¼ª�ª©ÀJÃRµ(¶�§�ª©²¶|¸(¹h¦}§©¨�·(¶|·}À½»~À~²wª©À~Ã4Õé¨|§©´Ã¾¶�²wª©²P¨|Õu³�¶|¸5«:²�¼»½¼Ã¼ª©¾¹ Þ ´»½¹}²´¿�¶s§©4¦�§©¨|Á¹}Â}Ã´¾¹ä¶|¸(¹ä¹�À~²�¦(»½¶´«5¾¹�À½¸ä§©¾¶|»7ª©À~³h4¶�ªPª©µ(YÕæ¨|»½»~¨ Ö À~¸(Ù Ö ´·V²�À×ª©�êµ5ª�ª©¦uê è�è ÖPÖPÖ ª¾Î ´³hÃ�Î ¸(Ã´¼¦uÎ ¸(¨�¶�¶�Î Ù|¨�º}ès³h³c·7è|®�¯4°�±Nè|®�¯4°�±�Î µ5ª©³h»Êz§©´»~À½³hÀ~¸(¶�§�«"§©´²�Â}»~ª©²4Õé§©¨�³ëª©µ(À~²/¦}§©¨�ª©¨|ª¬«}¦7c®�¯4°È±$²w«�²wª©´³¶�§©U¦�§©¨�³hÀ~²�À~¸(Ù"ËÍÌu§d¶�Ã¼ª©¨�¸�¶|¸(¹[Ò/Âl¿AÄ|Å|Å�Æ¾ÓRÎ�¥Y¸ØÀ~³h¦7¨|§�ªd¶�¸*ªY»~´²wÁ²�¨�¸ Ö ¶|²�¶|»~²�Ü»½´¶�§©¸(´¹cÕé§©¨�³@¨�Â�§�¦(¶�§�ª©À~Ã´À~¦K¶sª©À~¨�¸hÀ~¸:ª©µ(4®}¥�ÛØ°Èì¼í�¦7¼§©À~³h´¸5ªYËoî�¨|ÂäRªP¶|»oÎ~¿(Ä|Å|Å�ÆsÓ�¹}Â}§©À~¸}ÙUª©µ}T¾¶�§©»×«�²wªd¶|Ù|�¨�Õlª©µ}®�¯4°È±9²w«�²wª©´³ï¹}´º6¼»½¨|¦7´³h´¸5ª¾Ñ�ðÍñæò¾ñ�¦7¼§�ª©Â}§©·(¶�ª©À~¨�¸}²cÀ~¸�ó^ôPÉ�²µK¶¾º6,¶�·(À~Ù�À~³h¦K¶|ÃRªØ¨�¸�´¸}²�´³c·}»~�²�¦�§©¾¶�¹ Ö µ(¼¸�ª©µ}�³h¨*¹}´»Ã´¨|³h¦(Â}ªd¶sª©À~¨�¸�¹�¨�³�¶|À~¸CÀ~²�ª©¨5¨�²�³�¶|»~»�À~¸C²�À~å´'ËoÒ/Â�¶�¸K¹CÌl§d¶|Ã¼Áª©¨�¸l¿PÆ¾Ï|Ï�Ï6ÓRÎ,Û[¨|²wª:¨�Õ4¨|Â}§J¦}§©´º*À~¨�Â}²G§©¼¦7¨|§�ª©²UÕæ¨5Ã´Â}²�¾¹�³�¶|À~¸(»×«¨�¸�³JÂ(»×ª©À×Á¬ÔwÉ'¶�²�¦7´Ã¼ª©²´Î�Ì�µ(À~²�§©¼¦7¨|§�ª Ö À~»~»o¿(µ}¨ Ö ´º6R§¾¿*³�¶|À~¸(»×«h¶|¹�Á¹�§©¼²�²Y³JÂ(»×ª©À×Áà³h¨*¹}´»A¶�²�¦7´Ã¼ª©²´Î

õ öa�U£��÷¾��£.ø*�Pù[¤ ú ��÷´�Pû¼üuý.��£��J¤ÿþ�÷�� ¢,£N��G¤!£®�À~¸}ÃŃ·7¨|ª©µU¨�Â�§Ú¦}§©¼»½À~³hÀ~¸K¶s§�«�§©´²�Â}»×ª©²z¶|¸(¹�ª©µ(�®}¥/ÛØ°Nì�§©¼Á

¦7¨|§�ª�²�Â(Ù|Ù�´²wª©´¹hª©µ(¶�ª�¶|¸h¼¸(²�´³J·(»~P²w«�²wª©´³ Ö À~ª©µh³JÂ(»×ª©À�³h¨*¹}´»~²Ö ¶�²È¶|» Ö ¶´«�²N·7¼ª�ª©¼§zª©µK¶�¸c¨|¸( Ö À×ª©µh¶Y²�À½¸}Ù�»~P³h¨*¹}¼»Í¿�À~¸Jª©¼§©³h²Ú¨�Õ´¸}²�´³c·}»½J³h¾¶|¸�¶�¸K¹�¦}§©¨�·(¶|·}À½»~À~²wª©À~Ã:Õæ¨�§©´Ã¾¶|²wª©²´¿z¶V³cÂ}»×ª©À~Áà³h¨*¹�´»¶|¦}¦}§©¨�¶|ÃdµäÀ~²�¨�¸(T¨�Õuª©µ}4¹�À×§©´Ã¼ª©À~¨�¸}² Ö 4¶�§©Y¼í�¦(»~¨�§©À½¸}Ù}ÎÈÌ^ÜÃ´¨|¸}Áª©À~¸*Â(Vª©¨�¼í�¦(»~¨�§©[À~¸'ª©µ}À~²h¹}À×§©´Ã¼ª©À~¨|¸u¿P¶�¸(¨�ª©µ(¼§h²�Rª�¨|Õ Þ º6�Ëàß�Ó³h´³J·7¼§©²ÚÕé§©¨|³>°zªd¶³h¨*¹}´»o¿5·}Â}ª Ö À×ª©µ â ¶|À~¸}Á¬±}§©À×ª©Ãdµ:Ã´¨�¸*º6¼Ã¼ª©À~º6²�Ãdµ(´³hª©¨:§©´¦}»�¶�Ã´Jª©µ}Gô�¼ª�ª©²wÁ Û[À~»½»~¼§�Á � ¶�¸(À~Ã:²�Ãdµ(´³h¸}¨|§©³�¶�»½»×«Â(²�´¹c·5«Gª©µ}T°Nªd¶/³h¨*¹�´»o¿6µ(¶sº�P·7´´¸h¶|¹(¹}´¹Jª©¨/ª©µ(/Æ´Å�Áà³h´³c·7R§®�¯4°È±[²w«�²wª©´³#ª©¨�³�¶�Ý6�¶Tª©¨|ªd¶�»�¨�ÕÚÆsßT³h´³c·7R§©²�Ëo¸(¨�ª©�êÚ²wª�§©À~Ã¼ª©»×«²�¦7¾¶�Ý�À~¸}Ù}¿Kª©µ}À½²4²wª©´¦"À~²/¶:³JÂ(»×ª©À×Á¬¦}µ5«}²�À~Ã´²P§d¶sª©µ(¼§Yª©µK¶�¸�¶:³cÂ(»×ª©À×Á³h¨*¹}¼»u¶�¦(¦�§©¨6¶|Ãdµ7ÓRÎ±ÚÀ~Ù�Â�§©´²�ÆÈ¶�¸K¹UÄ�²�µ(¨ Ö ª©µ(�®�´¦�ª©´³c·7R§uÄ|Å�Å�ÄY¶¾º6R§d¶|Ù�È²�Ã´¨�§©´²¨|Õ^Ü6Ð�Ý�³>´¸(²�´³J·(»~�³h´¹}À½¶|¸hÕæ¨�§©´Ã¾¶�²wª©²´¿ Ö À×ª©µ[º|¶�§�«�À~¸(ÙG¸*Â(³c·7R§©²¨|Õ�³h¨*¹}´»~²JÀ½¸*º6¨|»~º6¾¹l¿�¶�² Ö ´»~»T¶|²:ª©µ}V¨|¦7¼§d¶�ª©À~¨|¸K¶|»UÆsÄsÝ*³_°zªd¶

Õæ¨�§©´Ã¾¶�²wª©²´Î�±}¨|§P¶�»~»7²�À~íc²�´»~´Ã¼ª©¾¹�º|¶�§©À½¶|·}»½¼²´¿*À½¸}Ã¼§©¾¶�²��À~¸:Õæ¨�§©´Ã¾¶|²wª¶|Ã¼Ã´Â}§d¶�Ã¼«�¨|Õlª©µ(Y´¸(²�¼³c·(»~Y³h¾¹}À½¶�¸ä¨�§P³h¾¶�¸:Õæ¨|§©¼Ã¾¶|²wª©²Y¶�§©�¶|²wÁª©¨�¸}À~²�µ(À~¸(Ù|»~«�²�À½Ù|¸(À Þ Ã´¶|¸5ª¾¿�¶|²cª©µ}ä¸*Â}³c·7¼§G¨|ÕY³h¨*¹�´»~²cÀ~¸(Ã¼§©´¶|²�Õé§©¨|³�¨�¸}/³h¨*¹}´»zËàßU³h´³c·7R§©²dÓNª©¨Jª Ö ¨J³h¨*¹}´»~²/ËwÆ´Åc³h´³c·7R§©²dÓ¶|¸(¹'ª©¨�ª©µ}§©´Ø³h¨*¹�´»~²"ËwÆsß�³h´³c·7¼§©²dÓRÎ�Ô¬ª�� ²�¶�»½²�¨�À½¸5ª©¼§©¼²wª©À½¸}Ùª©¨G²�´4ª©µ(¶�ª�ª©µ}�´¸}²�´³c·}»½Y³h¾¹�À�¶�¸hÕé§©¨�³@ª©µ(4Ü6Ð|Ý*³.ßsÁà³h´³c·7R§Æ¼Áà³h¨*¹�´»N²w«}²wª©¼³�¦7¼§�Õæ¨�§©³h²G¸}¨Ø·7¼ª�ª©¼§ª©µ(¶|¸�ª©µ(h²�À~¸(Ù|»½"Æ¾Ä�Ý*³°zªd¶�¿ Ö µ}À~»½Vª©µK¶�ªJÕé§©¨|³ Æ¾ÅsÁ¬³h¼³c·7¼§J¨|ÕÄsÁà³h¨*¹}´»�²w«}²wª©¼³_¦7¼§�ÁÕæ¨�§©³h²/Ù�¼¸(¼§d¶�»½»×«�·7¼ª�ª©R§/ª©µK¶�¸�¨�§U¶�ª�»~¾¶|²wªUÃ´¨|³h¦K¶s§d¶|·(»~Gª©¨�ª©µ}

�� ! " $#&%('�)+*�,.-�%/,1032�4/5�687:9<;�-�=>�? 'A@�-.B+C$-�9"[email protected]$D ? -.KL6MC$-�9AC$'NEO/6M;QPR0RS�T�U�UH4/,�9 VXWY@�7EI10Q6G7EB+Z2�Z.' ? -.K�[�0.F\%]T�U�^:_�`�0ba/2�4dc�C=eBf7 ? D3gh*�,�-1) %/,�i/-.@�7�7�) K�@ >�j

ÆsÄsÝ*³ °zªd¶�¶|¸K¹$ª©µ(¶�ªäÕé§©¨|³2Æ¾ßsÁà³h´³c·7R§V¨�ÕJã�Áà³h¨*¹}´»4²w«�²wª©´³¨�Â�ª©¦7¼§�Õæ¨�§©³h²:ª©µ(�ÆsÄ�Ý*³a°zªd¶,²�À~Ù|¸(À Þ Ã¾¶�¸5ª©»~«RkC®�À~³hÀ~»½¶�§h§©´²�Â}»~ª©²´¿·(Â�ª Ö À×ª©µ�»½¼²�²�¼§c²�À~Ù|¸(À Þ Ã¾¶�¸(Ã´|¿NÃ¾¶|¸�·7h²�¼´¸�Õæ¨�§c¦�§©´Ã´À~¦(À×ªd¶sª©À½¨|¸Õæ¨�§©´Ã¾¶�²wª©²4ª©¨5¨"ËÍ±ÚÀ~Ù}Î�ã�¶6ÓRÎ�±ÚÀ~Ù�Â}§©Uã�·"²�µ(¨ Ö ²Tª©µ(À~³h¦}§©¨sº6´³h¼¸*ª¨|Õ�¦}§©¨|·K¶�·(À~»~À½²wª©À~ÃcÕé¨|§©´Ã´¶|²wª©²³h¾¶�²�Â}§©¾¹"·5«"¯4¶|¸(Ý�¾¹"ÊN§©¨|·K¶�·(À~»~À½²wÁª©À~ÃJ®�Ý*À~»½»z®�Ã´¨�§©äËo¯4Ê�®(®(ÓP¶�²4ª©µ(¸*Â}³c·7¼§T¨�Õ�³h¨*¹�´»~²4À~¸}Ã¼§©¾¶|²�¼²´Î

l/m�n�oRp:qsrbtvu3qJw3xzy�{�{�y|p:}�~�qv�b�f��b�1}�q��~!q�}��R��qv}�q��3m��b��bp:q��b~!x:~��+m�x��r��"}��3�3q��|� �.�bq�p:�bnbq��bq�p��:��"�<x:��R��:��u3 s�¡�b��¢�E�.�£\lRqJx:�b��t¤�:�b��¥��¦y.�"}��3�3qJ�X��E�.��§xE�¡¨©��u3 s�Rtª�Jr�{��¥v��¬«��"}��RqJ�X��E�.��§x:�®¨©�.��u3 ¯¨°��£\lRqJx:��Rt8�Jr.��¥�u3�f�±l²~�³3~$x:qJ}X~��b��´�bwQqJpE��x:m��µr�y��1}�§xE��!��wR�¶��¥�·§¸¹��x:��§x:��b��|ºHl§u|��³~�m¶~¹�bp�qHo�~�q��Y·

» ¸([³hÀ½Ù|µ5ª[¶s§©Ù�Â}�ª©µ(¶�ªä²�Â(Ãdµ$²�À~Ù|¸(À Þ Ã¾¶�¸5ª[À~³h¦}§©¨sº6¼³h´¸5ª©²ª©µ}§©¨|Â(Ù|µ$¶�³cÂ(»×ª©À×Áà³h¨*¹}´»�¶|¦}¦}§©¨6¶�ÃRµ�Ã´¨�Â}»½¹'²wª©´³ïÕé§©¨�³_³h¨*¹}´»·(À½¶�²hÃ¾¶|¸}Ã´´»~»½¶�ª©À~¨�¸ Ö µ}´¸�«5¨�Â'Ã´¨�³J·(À~¸(�Õæ¨�§©´Ã¾¶�²wª©²�Õ�§©¨�³H³h¨�§©ª©µK¶�¸$¨|¸(Ø³h¨*¹�´»oÎ�Ì�µ(¼§©RÕæ¨|§©|¿�¶,²�À½³h¦}»~[ª©´Ãdµ(¸(À½¼5Â( Ö ¶�²V¶�¦}Á¦(»~À~¾¹,ª©¨ Þ §©²wªU§©´³h¨sº6:·(À½¶�²�´²cÀ~¸K¹�À~º�À½¹�ÂK¶|»~»×«�Õé§©¨�³ ¾¶�ÃRµ�³h¨*¹}´»¶|¸(¹Vª©µ(¼¸[ª©¨:§©´¦7¾¶sª4ª©µ}U¶|·7¨sº6Ã¾¶�»½Ã¼Â(»½¶�ª©À~¨�¸lÎTÔ¬ªYÀ~²YÕé¨�Â}¸K¹Vª©µ(¶�ªª©µ(È§©´²�Â(»×ª©²N¶s§©�º*À×§�ª©ÂK¶|»~»×«GÂ(¸}ÃRµ(¶|¸(Ù|¾¹hËo¸(¨�ªz²�µ(¨ Ö ¸7ÓRÎ » Õ(Ã´¨|Â}§©²��¿³h¨|§©�²wª©ÂK¹*«UÀ½²N¸}´¾¹}´¹Uª©¨4º6¼§©À×Õé«Uª©µ(À~²´ÎÈ®*À~¸(Ã´Èª©µ(�´¸}²�´³c·}»~�²�À~å´

¾�¿LÀdÁÂ¾�ÃQÄEÅ:Æ1Ç È3ÉQÅHÈ1ÊdË�ÌÍ�ÎÆ Æ�Ï�ÐYÆbÑ�ÅEÆbÒ ÒMÓ½Ô ÄEÐMË�Æ�ÌÒÕ�Ö1×+Ø\Õ�Ùb×EÚYÛ�ÜµØ\×EØµÝ3×:Þ$ß!à á�â�ã ä�äbå ä.ã æ�äbå â�ã ç1Ö1Ù�×:Þ§Ù�èéßê3×:Þ$ßèéë.×

Ä�Ó�Ï]ìMÌ1Ä�Ó¶ÑîíGÆbï¶ÒQÄ ðAñ ÕHØ\Õ�Ù�×:Ú½ßfÛ"æEò¦Ø\×EØ¦Ý3×�Þ$ß!à ó�Ü�ã æEôbå ó�ã õbå ô.ã Ö1×�ö�Þ<ê3×:Þ÷¶×Eø ðÛeß$ö�Ø\×&ö�ß±è½Ö�ùMèéú�å±æ�à ð$û Þ$×:×&Ø\Õ�Ù�×EÚéßfÛ"æ�Ü�Ø\×EØ¦Ý3×:Þ$ß!à ò.ã æ�ó�å Ü�ã ü$æ�ó�å Ü�ã Õ�ë.×:ÞdÙ�èéßê3×:Þ$ßèéë.×

Õ�Ö1×+Ø\Õ�Ùb×EÚYÛ�ÜµØ\×EØµÝ3×:Þ$ß!à ä.â�ã ä�äbå ä.ã äbå â�ã Ö1×�ö�Þ<ê3×:Þ÷¶×Eø ðÐQËJÆbÑ.Ó�ÐGÓ�ÅEÌ�Å�Ó�È3Í ðAñ ÕHØ\Õ�Ù�×:Ú½ßfÛ"æEò¦Ø\×EØ¦Ý3×�Þ$ß!à ó�ó�ã æEôbå ó�ã äbå ô.ã Ö1×�ö�Þ<ê3×:Þ÷¶×Eø ð

ð$û Þ$×:×&Ø\Õ�Ù�×EÚéßfÛ"æ�Ü�Ø\×EØ¦Ý3×:Þ$ß!à æ�Ü�ã æ�ó�å Ü�ã ó�å Ü�ã Ö1×�ö�Þ<ê3×:Þ÷¶×Eø ð

ý/þ�ÿ�� :þ�� :þ��§þ��þ�� "!��#��$%��þ&�"�"� ' ÿ(�)�*�&�"��Jþ��+�"��',��-�"��$îÿ��.�Eþ��-��\þ��/$%��þ��"0��"��'vÿ(�|ý(þ��þ�1��Eþ��'2' !3�4��"!�ÿ�0 ��!��&�5' 0��!��)6(� � �"� $�ÿ7��98�:�:�8 ;

ö�ÚéßÕ�èéÖ�ø:Þ$×�ö�ß×:ß�Û¶÷�Þ$Õ�Ø�Ü ð Õ æ�Ü�à ñ<û ×EÖ ð$û ×¦Öbç�Ø¦Ý3×:Þ¹Õ�÷/Ø\Õ�Ù�×EÚéßèéÖ1ø:Þ$×Eö�ß×Eß�Û¶÷�Þ$Õ�Ø¢æ ð ÕHä.à èéÖ ð$û èéß<ß ð çÙ�<>=�Õ�Ö1×+øEÕ�ç1Ú½ÙXöJÞ$ú�ç1× ð$û ö ðð$û èéß§Ý�è½ú¦è½Ø\êbÞ$Õ�ë.×EØ\×EÖ ð Þ$×Eßç1Ú ð ×�ÙL÷�Þ$Õ�Ø ð$û ×&èéÖ1ø:Þ$×Eö�ß×+èéÖ�×EÖ�ß×EØHüÝ1Úé×<ßè@?E×<Þ!ö ð$û ×�Þ ð$û ö�Ö\èéÖHØ\Õ�Ùb×EÚ1Ö�ç1ØµÝ3×:Þ�å�A¹Õ ñ ×Eë�×:ÞB=�ö&ê�Þ$×:ëbèéÕ�ç1ßß ð çÙ�<�ÛDC&çv× ð ö�Úeå@=hæBE�E�â�à&ß û Õ ñ ß ð$û ö ð èéØ\ê�Þ$Õ�ë.×EØ\×:Ö ð ß±èéÖv÷�Õ�Þ$×:üø�öJß ð öJøEøEç�Þ!öJø�< ð$û Þ$Õ�ç�ú û ×EÖ�ß×EØ¦Ý�Úé×\ö�ë.×:Þ!ö�ú�èéÖ1úvÙ�×Eø�Þ$×�ö�ß× ñ<û ×:Ö

FG!��&0��.8��H63þ�$%�\þ��/FG!��-;��#I��#� � ��*��J�#�&�"��¶þ>��!��&�K��L%1�#!��;

×EÖ�ß×EØ¦Ý�Ú½×Xßè@?E×vèéÖ1ø:Þ$×Eö�ß×Eßîè½Ö ö�Õ�Ö1×:ü�Ø\Õ�Ùb×EÚd×:Ö1ß×EØµÝ1Úé×XßM<bß ð ×EØÛeØ¦ç�Ú ð èéüONMP�Õ�Ö�Ú@<1à�åRQ|× ð$û ×�Þ$×:÷¶ÕJÞ$×XÝ3×EÚéè½×:ë.× ð$û ö ð\ð$û ×Eß×Xßèéú�Ö1èTS1üø�öJÖ ð èéØ\ê�Þ$Õ�ë.×EØ\×:Ö ð ß¹ÛeÖ1Õ ð × U ð$û ×+èéØ\ê�Þ$Õ�ë.×:Ø\×EÖ ð ÷�Þ$Õ�Ø�æEò ð Õ�æ�ÜØ\×EØµÝ3×:Þ$ß ñ ö�ß§×Eë�×EÖîØ\ÕJÞ$×dßèéú�Ö1èTSø�öJÖ ð§ð$û ö�Ö ð$û ö ð ÷�Þ$Õ�Ø¤Ü ð Õ�æEòØ\×EØµÝ3×:Þ$ßWV à�ö�Þ$×µöHÞ$×Eßç�Ú ð ÕJ÷(ø:Õ�Ø¦Ý�èéÖ1èéÖ1úHØµç1Ú ð è üONXP ñ è ð$û Ø¦ç1Ú ð è üØ\Õ�Ù�×:Ú*=-YOZ3[BZ]\_^�` aDYDbc["d�e�["\.fW`g[/e4h e4ai["\_e�j+klk(mXnBj�o+p�V

q öJÝ1Úé×|æ�èéßLövßç1Ø\ØLöJÞX<|Õ�÷ ð$û ×�ê3×:Þ$ø:×EÖ ð ö�ú�×:ß ñ<û ×:Þ$× ð Þ$ç�×ö ð Ø\Õ�ßê û ×:Þ$× ñ öJßµÝ3×�<.Õ�ÖÙ ð$û ×HÞ!öJÖ1ú�×LêbÞ$×�Ù�èéø ð ×�Ù Ý�<v×EÖ�ß×EØ¦Ý�Úé×

Ù�çbÞ$èéÖ1úsrb×:ê ð ×EØµÝ3×:Þ¹ó�ò�ò�ó�åtN ð øEö�ÖzÝ3×fß×E×EÖ ð$û ö ðdð$û ×µ×EÖ�ß×EØ¦Ý�Úé×ßê�Þ$×Eö�Ù�èéßJ^�d(u>["m"bcu Y�eOk&["m+e4Y3v [dèéÖXö¦Õ�Ö1×:ü�Ø\Õ�Ùb×EÚ�ßM<bß ð ×EØ Û�Ü�Ø\×EØHüÝ3×:Þ$ß!à"=QÝ1ç ð n�v ["m�biu Y3e*k&["m+e+Y�v [µèéÖ�ö ð$û Þ$×E×:ü�Ø\Õ�Ù�×EÚ<Û"æ�ÜîØ\×EØ¦Ý3×�Þ$ß!àßM<bß ð ×EØ�åxwJ[4j�m�b@k&["m3yW[4o�aze*k(mX[4j�u èéß�ß×E×:Ö èéÖ ð$û ×vêbÞ$×EøEèéê1è ð ö ð è½Õ�Ö÷¶ÕJÞ$×Eø�öJß ð ßEå

FG!��&0��|{-�}6Rþ�$%�´þ��~FG!��;��8ÿ�0 �~�*�&��þ&�,�dýJ6��1þ��#��/!g��:-;��"�-�"��+��&��'��|þ��-'��ÿ��|�J��6�6��>�� -&��t��Jþ��"��&��!��G0-�"!��-��8��$��]�Eþsþ&�z��#�*��#��*�&�"��1�Jþ&�4��;]�_�t�� #!��vþ��þ�� +!3�H!3�0��+��')!��)�"�-�� "!��þ>�"!��;

� �¡]¢¤£5¥�¦�§©¨«ªz¬~® �¥�ªz¬~¨

¯ ö�ß×EÙ]Õ�Ö ð$û ×�ê3Õ�ßè ð èéë.×�Þ$×:ßç1Ú ð ßW=±°�P±²´³µr�¶�²hù¡ßM<bß ð ×EØñ èéÚéÚ�øEÕ�Ö ð èéÖ�ç1× ð Õ<Ý3×hÙ�×Eë.×:Ú½Õ�ê3×�Ù&öJÚ½Õ�Ö1ú ð$û ×GÚéèéÖ1×hÕ�÷bØ¦ç�Ú ð èéü�×EÖ�ß×EØ¦Ý�Úé×ßM<bß ð ×EØ\ßEå�P/çbÞÞ$×EÖ ð ÚT<>= ðAñ ÕvÖ1× ñ ßç1Ý�ß× ð ßHÕJ÷+Ü�Ø\×EØµÝ3×:Þ$ßµ×Eö�ø ûñ è ð$û ¶]·/P]ö�ÖÙ¤¸�¶�³Gr\Þ$×Eú�èéÕ�Öö�Ú�Ø\Õ�Ù�×EÚéß Þ$×Eßê3×:ø ð èéë.×EÚT<�ö�Þ$×&Ý3×:üèéÖ1ú ð ×Eß ð ×�Ù�ö�Ö1Ù ñ èéÚéÚ�Ý3×dèéØ\ê1Úé×EØ\×EÖ ð ×�ÙHè ÷MêbÞ$Õ�ë.×EÖ ð Õ¦ö�Ù1ÙLë�öJÚéç1×�å¹ ë.×:Þü Ùbè½ßê3×�Þ$ßè½Õ�Ö�èéÖ�ßê�Þ$×�ö�Ùîèéß±öµê3Õ�ßßèéÝ�Ú½×+êbÞ$Õ�Ý�Ú½×:Ø²èéÖ ð$û è½ß<öJê�üê�Þ$Õ�ö�ø û�ð$û ö ð ØLöW<îÖ�×E×�Ù ð ÕHÝ3×&øEÕJÞÞ$×Eø ð ×�ÙRå

¸dÖ�Õ ð$û ×:Þfê�Þ$Õ�Ý1Úé×EØ Þ$×EÚ½ö ð ×�Ù ð Õ ð$û èéß�ö�ê1êbÞ$Õ.öJø û èéß ð$û ö ð�ð$û ××EÖ�ß×EØ¦Ý�Ú½×zØ\×EØµÝ3×:Þ$ß�øEÚéç1ß ð ×:Þ�Ý�< Ø\Õ�Ù�×EÚ¹è½Ö ßÕ�Ø\× ßè ð çö ð èéÕ�Ö1ßEåq ÕvèéÖ1ø:Þ$×�öJß× ð$û ×�Ùb×Eú�Þ$×E×îÕ�÷±÷�Þ$×E×EÙ�Õ�Ø%=hú�è½ë�×EÖ ð$û ×�è½Ø\ê3ÕJÞ ð ö�Ö�øE×Õ�÷/ç1Ö�øE×:Þ ð ö�èéÖ ð <�è½Ö Ø\Õ�Ù�×EÚGê û <�ßèéøEß¦ÛOr ð ×EÖ�ß"Þ$çÙX× ð öJÚ�å@=Gó�ò�ò�ò.à"=MöÖ1× ñ r�¶�²hù ßM<bß ð ×EØ¢èéß¦ç�ÖÙ�×�Þ¦øEÕ�Ö1ß ð Þ$ç�ø ð èéÕ�Ö�åKN ð�ñ èéÚéÚ§ö�Ù�ÙbÞ$×Eßßê û <bßèéøEßdÙ�èéë.×:Þ$ßè ð <îèéÖX×�ö�ø û Ø\Õ�Ù�×:ÚQÝ><�ç�ßèéÖ1úLÙ�èTºR×:Þ$×:Ö ð ö�Ú ð ×:Þ$Ö1öJüð èéë.×¦ê û <�ßèéøEßdêöJø"»�ö�ú�×Eß&÷¶ÕJÞ�ö�ú�è½ë�×EÖ Ø\Õ�Ùb×EÚeå�ùMèéÖö�ÚéÚT<>= ð$û ×¦ø:ç�ÞüÞ$×EÖ ð á.ô�»bØ�ßM<bß ð ×EØ%=�Þ$ç1Ö1Ö�èéÖ1ú ð ñ è½ø:×dê3×:Þ/Ù1öW<>= ñ è½ÚéÚÝ3×§ç�ê1úJÞ!ö�Ù�×EÙð Õ\ä�ó�»�Ø%=�Þ$ç1Ö1Ö�èéÖ1úµ÷�Õ�çbÞdø"<�ø:Ú½×:ß±ê3×�Þ¹Ù�öW<>=�èéÖ ð$û ×&Ö1×EöJÞ<÷¶ç ð ç�Þ$×�å

¼¾½J¿�½�À]½�Á~Â/½JÃ

Ä�Å-Æ�Ç&È@Æ±É ÈËÊ�È�Ì�Å�Í@Í@ÎWÏ�Æ´Ð�Ï(Ñ5ÒGÈ�É�Ð Ï�Ñ&Î�Ó4ÔWÆHÕBÖ Öl×>Ø¤É�Ù�Ú�ÓXÛXÜ*Ó+Ð Ï�Ý ÎÎWÏ�ÔXÎWÞ9ß�Í@ÎHà3Ú Ó4Î�áBÐ ÔMÛ4â@Ï�Ý.Ú�àËã>Å(Ð�Ï>Û4â@Û+Ð#Û4â@älÎ�å�Ó4ÎWáWâ@å�âTÛ+Ð#Û4âgÚ Ï�È(æKç è éê�ëXì>é-íHë"î égÆ�ï�ð(ñ-Æ�ò#ó>ò ×#Ücò�ólô Ö�È

Ä�Å-Æ>Ç&È@Æ�Ð�Ï(ÑzÌKÈ�É Èlõ-Ó+Ð á�Û4Ú Ï�Æ&ò ö ö�Õ Ø�÷cÞ.å�Í@ÎWÞ.Î�Ï�Û+Ð#Û4â@Ú�Ï�Ú à�Ð�Ó4ÎBÐ�Í@ÜÛ4âgÞ.ÎÔXÙ�Ú�ÓXÛXÜ*Ó+Ð Ï�Ý Î�ÎWÏ&ÔXÎWÞ_ß&ÍgÎHà3Ú�Ó4ÎWáBÐ�ÔMÛ4âgÏ&Ý.ÔMø�ÔMÛ4ÎWÞùÐ#Û]ú/û±ü�ýHØÐ ÏzÅ�å�Ñ&Ð�Û4Î�È�ýËÓ4ÎWå�Ó4âgÏ>Û4ÔWÆ�þ ÿ��ç è��Bë��4ë"è�+ë�ç èzæ,ë�"ç��4ì �gë��Xç��"ë��ë�4Æ>Ò7ÛBÈ#Ê�Ð Å�Ñ&Î�Ó+Ñ&Ð Í@Î�Æ Ò�Í@Ú Ó4âgÑ�Ð�Æ��Þ.Î�ÓBÈ#Ì�Î�Û4Î�Ú ÓBÈ É�Ú>á ÈgÆ��ô�ô#Ü�lô��

Ä�Å-Æ,Ç�ÈgÆ,Ð Ï�Ñ ÌKÈ�É�Èsõ-Ó+Ð á�Û4Ú Ï�Æ ÕBÖ�Ö Ö�Ø ÷iÞ.å(Ð á"ÛRÚ à,ÍgÐ�Û4Î�Ó+Ð�Íß Ú�Å�Ï�Ñ�Ð#ÓXø á�Ú�Ï(Ñ�âTÛ4âgÚ Ï�Ô)Ú�Ï Ó4ÎWÝ�â@Ú�Ï�Ð ÍTÜOÞ.Ú�Ñ&ÎWÍ9ÎWÏ�ÔXÎ�Þ_ß�Í@Î~å�Ó4Î�ÜÑ&â@á�Û4â@Ú�Ï-È´íë��ëXì��|ì��ÿ��3î��3ÿ��Oë��3è ì ÿ��ç�� ë��)ì è"!Kç��+ë4ì�è#��.ç�!>ë$�%� ��è'&)(DÎBÑ�âTÛ4ÎBÑ2ß>ø+*�È-,�âTÛ4á+Ù�â@Î/."Æ0,�Î�å Ú ÓXÛ9ò�1�ÆËû��tÉ#2�Ç>É�û3~Ú Ó�4�â@Ï�Ý65�Ó4Ú�Å&å|útÅ&Þ.Î�Ó4â@áBÐ Íü87�å Î�Ó4â@Þ.ÎWÏ>Û+Ð�Û4â@Ú�Ï9(%3:5tútü;."Æ3�Ì=<>2 õ±ÄtÜOú�Ú&È@Æ"?A@�ð-Æ"��È ×#ÜB��È 1�È

*tÚ Å�Æ�Ä_È@Æ�üHÈ CJÐ Í@Ï(ÐWø>Æ&Ð Ï�Ñ)C�ÈAC�È�ÄtÓ4Ú>ÎWÝ ÎWÞ.ÎWâ@Î�ÓBÆ(ò ö�ö�Õ�Ø�<tßADcÎWá�ÜÛ4âgä�Î�älÎ�Ó4âFE(áBÐ#Û4â@Ú�Ï Ú à/Û4Ù&ÎKÉ ��Ì%üHGJI Ö�1~ÎWÏ�ÔXÎWÞ9ß�Í@Î%à3Ú�Ó4ÎWáBÐ�ÔMÛ4ÔWÈæKç è�étê%ëXì>é-íë"î égÆ-ï�ð?�Æ�×��#ÜiÖ�Õ�È

É�Û4Î�Ï�ÔMÓ4Å(Ñ�Æ Ä_È-Ç&È@Æ�Ç&È Ü�3�È0K±Ð�Ú&Æ�Ð Ï�Ñ2õtÈ�õtÈ"3~Ð�Ó4Ï�Î"ÓBÆ�ò ö�ö ö�ØMLtÔMÜâgÏ&Ý�â@Ï�âTÛ4âgÐ Í-áWÚ�Ï�Ñ&âTÛ4â@Ú�Ï�Ð�Ï(ÑsÞ.Ú�Ñ&ÎWÍ�å�Ù>ø�ÔXâ@áWÔ]å Î�ÓXÛ4Å�Ó4ß(Ð�Û4â@Ú Ï�ÔHâ@ÏÔXÙ�Ú ÓXÛXÜ*Ó+Ð�Ï�Ý�Î�Î�Ï�ÔXÎWÞ9ß�Í@Î�ÔXâ@Þ_Å�ÍgÐ#Û4âgÚ Ï�ÔzÚ àJÞ.Î�ÔXÚ�ÔXáBÐ�ÍgÎ2áWÚ Ï�ä�ÎWá�ÜÛ4âgä�Î�ÔMø�ÔMÛ4ÎWÞ.ÔWÈ�æ2ç è�étê%ë4ìlé�íHë"î égÆ-ï�ðN-Æ(ò ö�×�×BÜMò>ÕBö�×�È

õ�Ó+Ð�á�Û4Ú�Ï-ÆùÌKÈRÉ�È@ÆµÇ&È Ä�Å-ÆgÆ ò�ö�ö�Õ�ØO��å&å�Í@â@áBÐ�Û4â@Ú�Ï�Ú à�Û4Ù&Îú/û±ü�ý;2�üGÌ�û ÔXÙ&Ú ÓXÛXÜ*Ó+Ð Ï&Ý�Î ÎWÏ�ÔXÎ�Þ_ß�Í@Î¾à3Ú�Ó4ÎWáBÐ�ÔMÛ¾ÔMø�ÔMÛ4ÎWÞ(OÉA,�ü´ÒH. Û4Ú å&Ó4ÎWÑ&â@á�Û4â@Ï�Ý Î$7�ÛXÓ4ÎWÞ.Î å�Ó4ÎWáWâ@å�âTÛ+Ð#Û4âgÚ Ï¾ÎWälÎWÏ>Û4ÔWÈý�Ó4ÎWå�Ó4â@Ï�Û4ÔWÆQP R��ç��%S'� ç èT��4ëU�$� �#�3ÿOì ÿ��*ç è6VXW ÿ��4ë��zë�$Y)�Q�Xë$�!��ÿ��*ç è Z0[��ì��ÿ��ZËì è!/íHë�\��ç�è#�ë�4ÆA��Í@ß�Å�ã�Î"Ó+ã�Å&Î�Æ�útÎ$]|Ì�Î$7�Üâgá�Ú&Æ"�Þ.Î�ÓBÈ&Ì�Î�Û4ÎWÚ�ÓBÈ�É�Ú>á�È@Æ"� ó�ÜB�lô�È

õ�Ó+Ð�á�Û4Ú�Ï-Æ�ÌKÈ�É ÈgÆlÇ&ÈlÄtÅ�Æ�^ËÈ�õ7Ú Û4Ù-Æ�Ð Ï�ÑJ*�ÈlÇ�Å�Ð Ï&Ý&Æ(ÕWÖ�Ö 1�Ø´É�Ù&Ú ÓXÛXÜÓ+Ð Ï�Ý Î ÎWÏ&ÔXÎWÞ_ß&Í@Î�à�Ú Ó4ÎWáBÐ�ÔMÛ4â@Ï�Ý_(OÉ',tü�ÒH. Ð#Û ú/û±ü�ý`2 ü´Ì�û�Èý�Ó4ÎWå�Ó4â@Ï�Û4ÔWÆba/clÿ��+��ç è��Bé�ç è�deS'�.ë��4ì �±ê%ëXì ÿ��ë��f��4ëU!��ÿ��*ç è�Æý�Ù&Ú>ÎWÏ�âF7�Æ�Þ.Î�ÓBÈ�Ì�Î�Û4ÎWÚ�ÓBÈ�É�Ú>á ÈgÆ(ò�� Ö�Ücò�× ò�È

õ7Ú�Û4Ù�ÆH^�È@ÆËÐ�Ï(ÑgCJÐ�ÍgÏ�ÐWø>Æ�ü]È@Æ±ÕBÖ Ö ��Ø_ü�Ï�ÔXÎWÞ9ß�Í@Î.Ò�Ú�Ó4ÎWáBÐ ÔMÛ4â@Ï&Ý)Ð�ÛÛ4Ù�Îtú�Ì�û�Ø�õ±Ù�Î�Ý ÎWÏ�Î"Ó+Ð�Û4â@Ú�Ï¤Ú à�å Î"ÓXÛ4Å&Ó4ß(Ð#Û4â@Ú�Ï�ÔWÈ#hQS ��gé"if�zë��Wéæ,ë"ÿiëXç��Xç �géjP ç��#é@Æ-kA@�Æ�ò��Õ#×BÜMòl� � ö�È

�� !�� "��#��$�#��&%�� '�(�#�)�*�+%,�-�'��. /�#��0��&12�3�� 4�� /��0

576989896;:'<>=�?�@A:'8;89B#C�D#E'F�?G:'H#IKJLCM69ION'F2:PE&:'HOQ'N�RS.T�U�V�?WN'XG@YT�N�8ZN'[�69\�:'8�:'H#I]=^VM<PN�C_U�ÒT�F269\ba�\�69T3H#\cT'DOJ+N d!:ea"VM:fV_TbgYH#6ZR'T�F2C26ZV�h'D�=Y<PT3C�DOJL=

T3<i:'698kjldm[�:'898;B#ConA69:'CMVM:fV_T'?pT3I#B

=rq�stQ�<uR'T�F2CM6ZN�H7N'X�VMÒTwvAxzyW{|yWVM:}<PN"I#T38Y`�:'C$~�T�T3H7F2B�H(N R'T�F�:�FMN�BO[�`#89h�q�s's's��mq�s's'sQ"<�ION�<):'69H�\cT3H&V_T�F2T3I�N R'T�F�VMÒT$g�?�a�?�B#U#U�T�F��6;IOd�T3CMVbX/N'F��fs�d!:fFo<�C_T3:'CMN�H�<PT3CMN�CM\�:'8ZTe\cN�H&R'T3\��VM6ZR'T$C_h�C_V_T3<�\�:'C_T3C�?�a�6;<�B#8;:fVM6ZN�H#CAd!T�FMT)U�T�FMX/N'F2<PT3I�B#C269HO[K~�N'VM`�VMÒT��!T�V_VMCL�+��6;898ZT�F_�+E&:'H��_6;\]��^�4E�:'H#I��b:'69H��OF269VMCM\2`��G��\cN�HR'T3\cVM6ZR'TKCM\2`#T3<PT3C�?]JLH*:'I#I�6ZVM6ZN�H�D�~�N'VM`�CM\2`#T3<PT3C�d�T�FMT�B#C_T3I�69H�F2B�H#C69H#6ZVM69:'8;6Z��T3I�d�6ZVM`�CMT�R'T�F2:'8WI#6Z��T�FMT3HV�V_T3\o`#H#69�"BOT3C�V_N�~�T�V_V_T�F�FMT�U�FMT3C_T3HVb<PT3C_N�CM\�:'89TiX/T3:fVMBOFMT3C�?K�mÒT3CMTV_T3\2`#H#6;�BOT3C�69H#\�8;B#IOT3I}��6�� :'H�:'I��MB#C_VM<PT3HV�V_N�IOT�U�69\cV�\cN�89I�U�N"N�89C�D!��6;6��A69H#\�8;B#CM6ZN�H4N'Xm<PT3C_N�H#T�V�d!N'FMQCMBOFMX�:'\cTKI#:fVM:�\cN�BOU�8ZT3I�V_Nw:�IOT�T�U�T�F�8;:3h'T�FPVMÒFMN�B#[�`�B�C_T$N'X.VMÒTe<PN"I#T38k¡¢C�N�d�H�R'T�FMVM69\�:'8zI�6£��B�CM6ZN�HN�R'T�Fi:�¤4ÒN�BOF�:'CMCM69<)6;89:fVM6ZN�HwU�T�F26ZN�I�D¥:'H#I¦��69696��AT38969<)69H�:fVM6ZN�H�N'X�8;:3h'T�F2C�d�ÒT�F2T$FMT38;:fVM6ZR'T�`"B#<)6;I#6ZV�hd!:'C.8ZT3CMC^VM`#:'H�§fs&¨©6ZX-F2:'I�:fFmT3\2ÒNid!:'C�U#FMT3C_T3HV�?G�ON'F.:'8;8�N'X�VMÒT FoB#H#C�D�VMÒT F2:'6;HOX/:'8;8�X/N'FMT3\�:'C_VMC�d�T�F2TTcª�:'<)6;HOT3I�V_N�I#T�V_T�F2<)69HOTA6ZX-:'\�\�B#F2:'\chid^:'C^FMT389:fV_T3IeV_N�VMÒT <):f[�H#69VMB#IOT�N'X,89:fFM['T�F_��C2\�:'8ZTAX�N'F2\�6;HO[�:'H�IVMÒT�F2<PN�IOh�H#:'<)69\�C�?

�-NKU�T�FMX/N'F2<«VMÒTP:'H#:'8Zh�CM69C3D�¬�R'T�U�:fF2:'<PT�V_T�F2Ci�®�s'sK<�~wR'T�FMVM69\�:'8l<)N'VM6ZN�H�D-�fs's]<�~�I#6ZR'T�FM['T3H�\cT'D¯ s's�<�~�R'N'FMVM69\�6ZV+hi:'IOR'T3\cVM6ZN�H�D#§ ¯ s�<�~$V_T3<PU�T�F2:fVMB#FMT.:'IOR'T3\cVM69N�H�DO:'H#IiCMB#FMX/:'\cTYX/FMN�H&V_N'['T3H#T3CM69Co�Gd�T�F2T:3R'T�F2:f['T3IeN R'T�FzVMÒT.:fFMT3:�d�6ZVM`�69Hi°��fs's�Q�<±N'X�VMÒTY\cT3H&V_F2N�69IiN'X�VM`#TY\cN�HR'T3\cVM6ZR'T C_h�C_V_T3<²I�BOF269HO[ VM`#T¬#F2C_Vz³b`#N�BOF2C�N'X�VMÒTACM69<�B#89:fVM6ZN�Hid�ÒT3H$N'FM[�:'H#6Z��T3IeU#F2T3\�6ZU�6ZVM:fVM6ZN�Hid!:'C!N'~�CMT�FMR'T3I�?G�mÒTAC_V_FMN�H#['T3C_V�°� ¯ ¨«N'X�:'898�T�R'T3HVMC�:'H#I�d!T3:fQ'T3C_Vb� ¯ ¨«d!T�FMTP69IOT3HVM6Z¬#T3I469H�:'8;8 ¯ \�:fV_T�['N'F269T3C�?�=±F2:'HOQ�69HO[�N'X.q�d!:'C:'CMCM6Z[�HOT3I�V_N�VMÒT)d!T3:fQ'T3C_V�X�N'Fo\�69HO[�C�:'H#I�¤eV_N]VMÒTiCMV_FMN�HO['T3C_V�D¥d�6ZVM`��KX�N'F�VMÒTi<)69I�I#8ZT ¯ s&¨�?��ÒTF2:'HOQ�69HO[�C,69H�:'898 ¯ \�:fV_T�['N'F269T3Cld�T�F2T!VM`#T3H�C2B#<)<PT3I�D':'H#I�VMÒT!V_N'Ui� ¯ ¨Ń'X�VM`#T!CM\cN'F2T3C¥d�T�FMTzIOT�¬�HOT3I�V_N~�T�C_V_FMN�HO[�8Zh�®X/N'F2\cT3IKT�R'T3HVMC�D�:'H�I]VMÒTb89N�d!T3C_VY� ¯ ¨©N'X¥\�:'CMT3C�d!T3:fQ"89h&�®X/N'F2\cT3I�?^�ON'FYVMÒT3C_T�µLTcª�V_FMT3<PT3¶C_T�VMClN'X�\�:'CMT3C�� ¯ C_V_FMN�HO[�89h&�®X/N'F2\cT3IP:'H#I ¯ d!T3:fQ"8Zh�®X/N'F2\cT3I��oD:�R'T�F2:f['TmT3�"B#6ZVM:f~�89T!VM`#FMT3:fVm��yW�.�,:'H#I�~�6;:'CCM\cN'FMT3C^d!T�FMT IOT�V_T�F2<)69H#T3IiX/N'Fm³ ��ÒN�BOF28ZhPFo:'69HOX�:'898k?��ÒTYCMV_FMN�HO[�8Zh�®X�N'Fo\cT3I�\�:'CMT3C^d�T�FMTAFMT�U�FMT3C_T3HV_T3I$~"hq3�K³ ��ÒN�BOFAU�T�F26ZN�I#CmN'XW:'\cVM69R'TPF2:'69HOX�:'898kD�d�`#698ZTbVMÒT�d�T3:fQ�8Zh�®X�N'F2\cT3Iw\�:'CMT3C d�T�F2T�FMT�U�FMT3C_T3HV_T3I�~h*q�s�?=.I��MB#C_V_T3Iwy��·CM\cN'FMT3C�d�T�FMT$\cN�<PU�BOV_T3I�B�CM69HO[�VMÒTPV_T3\2`�H#69�"BOTPN'X!¸Y:'<)698;8^�Lq3¹'¹'¹��V_N]T3�"B#:'896Z��T�~�6;:'C~�T�V+d�T�T3H�VMÒTbC_T�VMC�N'X,\�:'CMT3C�?

JºH*~�N'VM`*CM69<�B#89:fVM69N�H#C B#CM69H#[KVMÒTi�z��E4:'H#I*��»\cN�HR'T3\cVM6ZR'TKCM\2ÒT3<)T3C�D,CMV_FMN�HO[�8Zh�®X�N'Fo\cT3I*\�:'C_T3CT3:fF2HOT3I}`#6Z[�`#T�F�y��¼CM\cN'FMT3C��,:f~�8ZT�q��X�N'F)<)N�C_V�F2:'69HOX�:'898�VMÒFMT3CMÒN�8;I#C�?�J+VPCM`#N�B#89I�~�TKHON'V_T3ItVM`�:fV:'8ZVMÒN�BO[�`tVMÒTe~�6;:'CL�®T3�B�:'896Z�3:fVM6ZN�H�:'I��MB#C_VM<PT3HVPI#69I�\2`#:'HO['T]<)N�C_VPyW�¼CM\cN'FMT3C�D�VMÒTK\2`#:'H#['T3CPd�T�F2THON'V.89:fF2['TbT3HON�BO[�`�V_Ne:��T3\cVYV_FMT3H#I#C.CMÒN d�H�69H�VMÒTbVM:f~�8ZT'?m�mÒTbI�6£��T�F2T3H#\cTb69H�yW��C2\cN'FMT3CY~�T�V+d�T�T3HC_V_FMN�HO[�8Zh$:'H#Iid!T3:fQ�8ZhiX/N'F2\cT3Ie\�:'C_T3C^d^:'C�F2N�BO[�`#8Zh)CM6;<)6989:fFGd.ÒT3H�~�N'VM`�CM\oÒT3<PT3C^d!T�FMTYB�C_T3I�D"Tcª�\cT�U#V:fV�VMÒTeÒT3:�R"6ZT3CMV�VMÒF2T3CMÒN�89Iwd.ÒT�FMT$�z�4E�F2B#H#C�`#:'I�H#N'VM69\cT3:f~�8Zhw`#6Z[�`#T�F�yW�>CM\cN'FMT3C�X/N'F�CMV_FMN�HO[�8Zh�X�N'F2\cT3I(T�R'T3H&VMC3D!~�BOV)��½F2B#H#CP`�:'I¾CM69<)6;89:fFPyW�©CM\cN'FMT3CiFMT�[�:fF2I#89T3CMCPN'XAVMÒTKX/N'F2\�69HO[O?}JLH}['T3HOT�F2:'8®Dd�T3:fQ�8Zh)X/N'F2\cT3Ie\�:'C_T3C^d�6ZVM`$~�N'VM`$C2\2ÒT3<PT3Cz`�:'I$`#69[�ÒT�F�~�69:'C_T3C�VM`#:'HeC_V_FMN�HO[�8Zh�®X/N'F2\cT3I�T�R'T3H&VMC��/VM:f~�89THON'V.CM`#N�d�H��o?z�ON'FA�z��E$F2B�H#C�DOVMÒT�~�69:'CzN'XlVMÒT�C_V_FMN�HO[�89h&�®X/N'F2\cT3I]T�R'T3HVMCYd!:'CY\�8ZN�C_T�V_N�qf?¢s�D�C_NiVM`�:fVVMÒT�~�69:'C.T�FMFMN'F2C��/N�R'T�FMU�FMT3I#69\cVM6ZN�H4N'X�F2:'69HOX�:'898�:fFMT3:&�.d!T�FMT�d�N'FoC_T�d�6ZVM`]d!T3:fQ�8Zh&�®X/N'F2\cT3I�T�R'T3HVMC�?��ON'F��(FoB#H#C�D�C_V_FMN�HO[�89h&�®X/N'F2\cT3I4T�R'T3H&VMC�['T3HOT�F2:'898Zh�`#:'I4~�69:'C_T3CA8ZT3CMCYVM`#:'Htqf?¢s�D-:'H�I�VM`#T�~�69:'CYT�FMFMN'FAd!:'C8ZT3CMCzX/N'FmVMÒT�d!T3:fQ�8Zh&�®X/N'F2\cT3I�T�R'T3H&VMC�?

a�69<)6;89:fF�\cN�<PU�:fF269C_N�H#C�d!T�FMTÛ�T�F2X�N'F2<)T3IbX/N'FlBOU#U�T�F_��8ZT�R'T38IOh"H�:'<)69\GX/N'F2\�69H#[��/FMT�U�FMT3C_T3HV_T3I�~h ¯ s's<b~4R'N'FMVM69\�6ZV+h�:'IOR'T3\cVM6ZN�H��:'H�I�8ZN d�T�FM��8ZT�R'T38-X/N'F2\�69H#[4�/FMT�U#F2T3C_T3H&V_T3I4~"h]CMBOFMX�:'\cTbX/FMN�HV_N'['T3HOT3CM69C��o?Yy��CM\cN'FMT3C�I�69IeHON'VmI#6£��T�Fz<�B#\o`K~�T�V�d!T�T3H�\�:'C_T3C�d�6ZVMè89:fFM['T�:'H�IKCM<):'898�:'<PN�B�H&VMCzN'X ¯ s's)<�~]R'N'FMVM69\�69V�h:'IOR'T3\cVM6ZN�H�?4=|89:fFM['T�F�I#6£��T�FMT3H#\cT$d!:'C�U#FMT3C_T3HVbd.ÒT3H�\cN�<)U�:fF269HO[�C_V_FMN�HO[4X�F2N�H&V_N'['T3HOT3C269C�\�:'CMT3C�V_Nd�T3:fQeN�HOT3C�D�T3C_U�T3\�69:'898Zh)X/N'F��z�4E)F2B�H#C ��HON'V�CMÒN d�H��o?G�m`#T3C_T FMT3CMB#89VMCz<):3hKCMBO['['T3CMVmVM`#:fVzU#FMN�\cT3CMC_T3C

¿�À"Á�Â9Ã�Ä'Å9ÅZÀ�ÆMÇ3ÈMÉ�ÅZ¿MÂ9ÊOË Â9ÊPÈMÌiÄ'Å9ÅZÇ�Æ_Í�ÈMÃ�Ä'Å9Ç'Î&ÌPÏ'ÆMÇ�Â;Ê&¿_Ç3Ê#ÈMÇmÉOÁÐ^ÄfÆ2Ñ)Ì)Ï'¿MÂZÏ�Ê�Î&ÈMÉ�Ã2ÒiÄ'ÈWÓ�Æ2Ï�Ê&¿_Ï'Ë'Ç3ÊOÇ3È2Â9È�Î&Á�Å9Ä3ÀÄPË'ÆMÇ3Äf¿_Ç�ÆYÆMÏ�ÅZÇ�Â9ÊKÂ9Ê#Ô�ÉOÇ3Ê#Ã�Â9Ê#Ëb¿MÒOÇbÄ'Ã�Ã�É#Æ2Ä'ÃcÀKÏ'Ó�Ð^ÄfÆ2Ì|ÈMÇ3Ä'È_Ï�Ê�ÃcÏ�ÊÕ'Ç3Ãc¿MÂZÕ'Ç�Æ2Ä'Â9ÊOÓ�Ä'Å9Å�Ó�Ï'Æ2Ç3Ã�Ä'È_¿MÈm¿MÒ#Ä'ÊÖ É#Ä'È2Â£Í®Ë'Ç�Ï�È_¿_ÆMÏ'Á�Ò�Â9ÃAÓ�Ï'ÆoÃ�Â9ÊOË)ÌPÇ3ÃoÒ#Ä'Ê#Â9ÈMÌiÈmÈMÉ#ÃoÒ]Ä'È�×fØ'ØiÌ�Ù]Õ'Ï'ÆM¿MÂ;Ã�ÂZ¿�ÀKÄ'ÑOÕ'Ç3Ãc¿MÂZÏ�Ê�Ú

Û Æ2Ç3Ã�ÂZÁ�ÂZ¿MÄf¿MÂZÏ�ÊKÜ�ÒOÆMÇ3ÈMÒOÏ�Å;Ñ�Ý�ÌiÌ$Þ

ß Ï�ÑOÇ3Å;à á#Ï'Æ2Ã�Â9ÊOË Úãâ'×�ä â"Úã×�ä å"Úãæ'× ç3â"Ú�è â'×"Úpäézß�ê Í�È_¿_ÆMÏ�ÊOË Úãâ'ë'æ Úãâ�èfå Úãâ�ä&× Ú9ç3ìfØ Ú9ç'ç'çézß�ê Í®Ð�Ç3Äfí ÚãâfØ�å Ú9ç3å�ä Ú9ç3æ'ì Ú9ç�Ø'Ø Ú¢ØOç3×î á�Í�È_¿_ÆMÏ�ÊOË ÚãæfØ�â Úãâ'ì'× Úãâ'ìfØ Ú9ç3ì'× Ú¢Ø�åfØî á�Í®Ð�Ç3Äfí Úãâ�ç3å Ú9ç�èfæ Ú9çcä&ì Ú9ç'ç'ç Ú¢Ø�å�è

Ü�ÄfÙ�ÅZÇ)çfï.ðYÑ�ñMÉ#È_¿_Ç3Ñ4òWÜ½ÈMÃcÏ'ÆMÇ3È)Ý/¿_ÏeÇ Ö É�Ä'Å9ÂZó�Ç�Ù�Â9Ä'ÈoÞ�Ä3Õ'Ç�Æ2ÄfË'Ç3Ñ�Ó�Ï'Æ È_¿_ÆMÏ�Ê#Ë�ÅZÀ]Ä'Ê#Ñ4Ð�Ç3Äfí�ÅZÀ]Ó/Ï'Æ2ÃcÇ3Ñ�Ã�Ä'È_Ç3ÈÓ/Ï'Æ�ÈMÂ9Ì�É#Å9Äf¿MÂ9Ï�Ê#È^É#ÈMÂ9ÊOË�¿MÒOÇ ézß4ê Ä'Ê�Ñ î átÈMÃ2Ò#Ç3ÌPÇ3È�Ú

áOÏ'ÆMÇ3Ã�Ä'È_¿lÄ'Ã�Ã�ÉOÆoÄ'ÃcÀAÄ'È,ÄzÓ�É#Ê#Ãc¿MÂZÏ�ÊAÏ'ÓOÈ_Ç�Õ'Ç�ÆoÄ'Å�¿MÒOÇ�Æ2Ì)Ï"ÑOÀ�Ê#Ä'Ì)Â;Ã¥Á�ÄfÆ2Ä'ÌPÇ�¿_Ç�Æ2È-Ð!Ä'È�Ä'Å9È_ÏmÇ�ÕfÄ'Å9É#Äf¿_Ç3Ñ�Úô Ê ézß4ê Æ2É#Ê�È�Î"ò�Ü¦ÈMÃcÏ'ÆMÇ3ÈmÑ�Â9ÑeÊOÏ'¿zÕ�ÄfÆ2À$Ì�É#Ã2ÒeÙ�Ç�¿�Ð!Ç�Ç3Ê�Ã�Ä'ÈMÇ3ÈzÐ�ÂZ¿MÒeÅ9ÄfÆMË'Ç�Ä'Ê�ÑKÈMÌ)Ä'Å9ÅlÝ/ÆMÇ3Å9Äf¿MÂ9Õ'Ç ¿_ÏÂ9Ê#ÂZ¿MÂ9Ä'Å;ÂZó3Äf¿MÂZÏ�Êw¿MÂ9Ì)Ç$Ä3Õ'Ç�ÆoÄfË'Ç Þ�ÃcÏ�Ê&Õ'Ç3Ãc¿MÂZÕ'Ç]Ä�Õ�Ä'Â;Å9ÄfÙ�ÅZÇ)Á�Ï'¿_Ç3Ê¿MÂ9Ä'Å�Ç3ÊOÇ�ÆMË'ÀõÝ®ö^ð Û ò^ÞoÚ�÷.Ï�Ð!Ç�Õ'Ç�Æ3ÎWÐ.ÂZ¿MÒî á!Î,ò�Ü²ÈMÃcÏ'Æ2Ç3È�Ð!Ç�ÆMÇeÌ�É�Ã2Ò�Ò#ÂZË�ÒOÇ�Æ�Ð�ÒOÇ3Ê�özð Û ò´Ð^Ä'È�Å9ÄfÆMË'Ç�Æ3Ú�áOÏ'ÆPÙ�Ï'¿MÒ*ÈMÃoÒOÇ3ÌPÇ3È�ÎGÃ�Ä'È_Ç3È�Ð.ÂZ¿MÒÈMÌ)Ä'Å9ÅlÃcÏ�ÊÕ'Ç3Ãc¿MÂZÕ'ÇeÂ9Ê#Ò#ÂZÙ�ÂZ¿MÂZÏ�Ê�Ý®ö ôºø ÞYÇ3ÄfÆ2ÊOÇ3Ñ*Ò#ÂZË�ÒOÇ�Æ�ò�Ü·ÈMÃcÏ'ÆMÇ3Èb¿MÒ#Ä'Êw¿MÒOÏ�È_ÇiÐ�ÂZ¿MÒwÅ9ÄfÆMË'Çeö ôºø Ó�Ï'ÆÅ9ÂZË�Ò¿_Ç�ÆbÆ2Ä'Â9Ê#Ó/Ä'Å9Å¥¿MÒ#ÆMÇ3ÈMÒOÏ�Å9Ñ�È�Î�Ù�ÉO¿�Ñ�Â£ù�Ç�Æ2Ç3Ñ�Å9ÂZ¿_¿MÅZÇPÓ/Ï'Æ�Ò#Ç3Ä3Õ�ÂZÇ�ÆbÆ2Ä'Â;ÊOÓ/Ä'Å;ÅkÚiòWÜ²ÈMÃcÏ'ÆMÇ3È�Ñ#Â9ÑwÊOÏ'¿�Ñ#Â£ù�Ç�ÆÌ�É#ÃoÒ)Ù�Ç�¿�Ð!Ç�Ç3ÊiÃ�Ä'ÈMÇ3ÈGÐ�ÂZ¿MÒPÒ�ÂZË�ÒPÄ'Ê#ÑPÅZÏ Ð�ç�Ø'Ø'Ø�Í+×fØ'Ø�ÌbÙ)Æ2Ç3Å9Äf¿MÂZÕ'ÇmÒ"É#Ì)Â9Ñ#Â9¿�ÀbÏ'ÆWÁ#ÆMÇ3Ã�ÂZÁ�Â9¿MÄfÙ�ÅZÇ!Ð!Äf¿_Ç�Æ3Ú

ÜmÒOÇzÂ9ÌPÁ�Ä'Ãc¿MÈ�Ï'Ó�¿MÒOÇzÂ9Ê#Â9¿MÂ9Ä'Å9ÂZó3Äf¿MÂZÏ�Ê�Ä'Ñ�ñMÉ#È_¿MÌPÇ3Ê¿MÈ¥Ä'ÈGÄYÓ/É#Ê�Ãc¿MÂZÏ�Ê�Ï'Ó�Å9ÄfÆMË'ÇcÍ�È2Ã�Ä'ÅZÇzÓ�Ï'ÆoÃ�Â9ÊOËYÐ!Ä'ÈlÄ'Å9ÈMÏÇcú�Ä'Ì)Â;ÊOÇ3Ñ�ÚGÜmÒOÇAÃcÏ�Å9Ñ�Á�ÏÏ�Å�Ä'Ñ�ñMÉ#È_¿MÌPÇ3Ê¿^Ò#Ä'ÑeÅ9ÂZ¿_¿MÅZÇAÂ9ÌPÁ�Ä'Ãc¿!ÆMÇ�Ë�ÄfÆ2Ñ�ÅZÇ3ÈMÈ!Ï'Ó�¿MÒOÇAÓ/Ï'Æ2Ã�Â9ÊOË�Ì)ÄfË�Ê#ÂZ¿MÉ�ÑOÇ'ÚÜmÒOÇ�Â9Ê#Ã�Å;É#ÈMÂZÏ�ÊõÏ'Ó�Ì)Ç3È_Ï�ÈMÃ�Ä'ÅZÇ�Ï'Ù�È_Ç�ÆMÕfÄf¿MÂZÏ�Ê#È]Â9Ì)Á#ÆMÏ�Õ'Ç3Ñ�É#Ê#Ä'Ñ�ñMÉ#È_¿_Ç3Ñ7òWÜûÈMÃcÏ'ÆMÇ3È�ÎAÙ�ÉO¿KÁ�Æ2Â9Ì)ÄfÆ2Â;ÅZÀÐ�ÒOÇ3ÊwÓ/Ï'Æ2Ã�Â9ÊOËeÐ^Ä'ÈbÈ_¿_Æ2Ï�ÊOËOÚ)ÜmÒ#Â9ÈYÆMÇ3È2É#ÅZ¿AÐ!Ä'È�Ñ#Â9ÈMÄfÁ�Á�Ï�Â;Ê&¿MÂ9Ê#Ë$ÈMÂ9Ê�ÃcÇ)ÂZ¿AÐ!Ä'È�ÒOÏ'Á�Ç3Ñ�Ð�Ç3Äfí�ÅZÀ4Ó�Ï'Æ2ÃcÇ3ÑÇ�Õ'Ç3Ê&¿MÈlÐ!Ï�É#Å9Ñ�Ù�Ç3ÊOÇ�ü#¿¥ÌPÏ'ÆMÇ^Ù"ÀbÒ#Ä�Õ"Â9Ê#Ë.Ù�Ç�¿_¿_Ç�ÆlÆMÇ�Á#Æ2Ç3È_Ç3Ê&¿MÄf¿MÂ9Ï�Ê�Ï'Ó�ÌPÇ3È_Ï�È2Ã�Ä'ÅZÇzÑOÇ�¿MÄ'Â9Å9È�Ú¥ÜmÒOÇ!ÆMÇ3Å9Äf¿MÂ9Õ'ÇÒÉ#ÌiÂ9Ñ#ÂZ¿+À&Í®Æ2Ä'Ñ�ÄfÆiÇ3Ã2Ò#ÏtÄ'Ñfñ_É#ÈM¿MÌPÇ3Ê&¿�Â9ÌPÁ#Æ2Ï�Õ'Ç3Ñ�É#Ê#Ä'Ñ�ñMÉ#È_¿_Ç3ÑõòWÜ«ÈMÃcÏ'ÆMÇ3È�Ó�Ï'ÆeÙ�Ï'¿MÒ�È_¿_ÆMÏ�ÊOË�ÅZÀÍ�Ä'Ê�ÑÐ�Ç3Äfí�ÅZÀÍ®Ó�Ï'Æ2ÃcÇ3Ñ�Ç�Õ'Ç3Ê&¿MÈ�Ð�ÒOÇ3Êe¿MÒOÇ ézß4ê ÈMÃoÒOÇ3ÌPÇ�Ð^Ä'ÈmÉ#ÈMÇ3Ñ�Î�Ù�ÉO¿^Ï�Ê#ÅZÀ$Ó�Ï'Æ�ÈM¿_ÆMÏ�ÊOËPÇ�Õ'Ç3Ê¿MÈmÐ�ÒOÇ3Êe¿MÒ#Çî átÈMÃ2ÒOÇ3Ì)Ç Ð!Ä'ÈmÉ�È_Ç3Ñ�Ú¥ý�Â9¿MÒ é^ß4ê Î#Â;ÌPÁ#ÆMÏ Õ'Ç3ÌPÇ3Ê&¿mÂ;Ê$¿MÒ#ÇYÐ!Ç3Äfí�ÅZÀ&Í®Ó/Ï'Æ2ÃcÇ3Ñ]Ã�Ä'ÈMÇ3ÈmÐ^Ä'ÈmÃcÏ�ÌPÁ�ÄfÆ2ÄfÙ�ÅZÇ¿_ÏK¿MÒ#Äf¿ Ï'Ó�¿MÒ#Ç�ÈM¿_ÆMÏ�ÊOËKÏ�ÊOÇ3ÈAÓ/Ï'Æ�Å9ÂZË�Ò¿A¿MÒOÆMÇ3ÈMÒOÏ�Å;Ñ#È�Î�ÈMÉOË'Ë'Ç3È_¿MÂ9Ê#ËK¿MÒ#Äf¿ Ä'ÅZ¿MÒOÏ�ÉOË�ÒwÌPÇ3ÈMÏ�ÈMÃ�Ä'ÅZÇ)ÑOÇ�¿MÄ'Â9Å;ÈÄfÆMÇ Â9ÌPÁ�Ï'ÆM¿MÄ'Ê¿!Â9ÊeÌPÏ�ÑOÇ3Å�Â9Ê#ÂZ¿MÂ;Ä'Å9ÂZó3Äf¿MÂZÏ�Ê�Î&¿MÒOÇ ÌPÏ'ÆMÇAÁ�Ï�ÈMÂZ¿MÂZÕ'ÇAÂ9ÌPÁ�Ä'Ãc¿!Ï�Ã�Ã�ÉOÆ2È^Ð�ÒOÇ3Ê$¿MÒOÇ ÑOÇ�¿MÄ'Â9Å;È!ÄfÆ2ÇÁ#ÆMÏ�Õ�Â9ÑOÇ3Ñ�¿MÒOÆ2Ï�ÉOË�ÒPÄ ÑOÇ�Ç�Á)Äf¿MÌPÏ�È_Á�ÒOÇ�Æ2Â9Ã^Å9Ä�À'Ç�Æ3Ú é Ï�É�Ê#Ñ#ÄfÆMÀ�Å9Ä3À'Ç�ÆWÂ;ÊOÓ�Ï'ÆoÌ)Äf¿MÂZÏ�Ê�Ä'ÅZÏ�ÊOÇzÌ)Ä�À�ÊOÏ'¿GÒ#Ç3ÅZÁ¿MÒOÇ�Æ2Ä'Â9Ê#Ó/Ä'Å9Å�Ó/Ï'ÆMÇ3Ã�Ä'È_¿�Ì�É#ÃoÒ�Ú

ô Ê�ÈMÉ#Ì)Ì)ÄfÆ2À'ÎOÐ�Ç�Ò#Ä�Õ'ÇbÓ/Ï�É#Ê#ÑeÓ/ÆMÏ�Ì©ÄPÆ2Äf¿MÒ#Ç�Æ�Å9ÄfÆMË'Ç�ÈMÄ'Ì)Á�ÅZÇ Ï'Ó¥Ã�Ä'È_Ç3Èm¿MÒ#Äf¿.ÌPÏ"Ñ#Ç3Å�Ó�Ï'ÆMÇ3Ã�Ä'ÈM¿MÈmÏ'ÓÐ!ÄfÆ2ÌþÈ_Ç3Ä'È_Ï�Ê�Æ2Ä'Â9ÊOÓ�Ä'Å9Å�ÄfÆMÇbÌPÏ'Æ2ÇbÄ'Ã�Ã�ÉOÆ2Äf¿_Ç�Ð�ÒOÇ3Ê]Å9ÄfÆ2Ë'ÇcÍ�ÈMÃ�Ä'ÅZÇ�Ó/Ï'Æ2Ã�Â9ÊOË)Â;ÈmÈ_¿_ÆMÏ�ÊOËOÚ�ÜmÒOÇ�Ä'Ã�Ã�ÉOÆ2Ä'ÃcÀ]Â;ÈÌPÏ'ÆMÇbÈ_Ç3Ê�ÈMÂZ¿MÂZÕ'Ç ¿_ÏiÅ9Ï�ÐzÍ�ÅZÇ�Õ'Ç3Å-Ó�Ï'ÆoÃ�Â9ÊOË)ÌPÇ3ÃoÒ#Ä'Ê#Â9ÈMÌiÈbÝ�ÈMÉ#ÃoÒ�Ä'È�È2ÉOÆMÓ�Ä'ÃcÇ Ó�ÆMÏ�Ê¿_Ï'Ë'Ç3ÊOÇ3ÈMÂ;ÈoÞ!¿MÒ#Ä'Ê�Ó�Ï'Æ2Ã�Â;ÊOËÄf¿�ÉOÁ#Á�Ç�Æ ÅZÇ�Õ'Ç3Å9È$Ý/Ç'ÚpËOÚZÎl×fØ'Ø�ÌbÙ*Õ'Ï'ÆM¿MÂ9Ã�ÂZ¿+À�Ä'Ñ#Õ'Ç3Ãc¿MÂZÏ�Ê�ÞoÚ ß Ï'ÆMÇ�Ä'Ã�Ã�ÉOÆ2Äf¿_ÇiÓ/Ï'ÆMÇ3Ã�Ä'È_¿MÈ�ÄfÆMÇiÄ'Å9ÈMÏ�Å9ÂZí'Ç3ÅZÀÐ�ÒOÇ3Ê�ö^ð Û òõÂ9ÈzÅ9ÄfÆMË'Ç'Î�Ê#Ï'¿mÈMÌ)Ä'Å9ÅkÎOÂZÓ�¿MÒOÇ î átÈMÃoÒOÇ3ÌPÇbÂ;ÈzÉ#È_Ç3Ñ�Î�Ù�ÉO¿mÅ;ÂZ¿_¿MÅZÇ Ñ#Â£ù�Ç�ÆMÇ3Ê#ÃcÇAÏ�Ã�Ã�ÉOÆ2ÈzÐ�Ò#Ç3Ê¿MÒOÇ ézß4ê ÈMÃ2ÒOÇ3Ì)Ç�Â9È�É#È_Ç3Ñ�Úlÿ�Ì)Ä'Å9Å�ö ôºø ÕfÄ'Å9ÉOÇ3È-Ó/Ä�Õ'Ï'Æ,Ù�Ç�¿_¿_Ç�Æ¥Ä'Ã�Ã�ÉOÆ2Ä'ÃcÀ�¿MÒ�Ä'Ê�Å9ÄfÆ2Ë'ÇWÕfÄ'Å9ÉOÇ3È�Ó/Ï'Æ¥Å9ÂZË�Ò¿_Ç�Æ¿MÒOÆMÇ3ÈMÒOÏ�Å;Ñ#È�Î�Ù�ÉO¿.ö ôLø ÑOÏÇ3È�ÊOÏ'¿mÈ_¿_ÆMÏ�ÊOË�Å9À�Â9Ê#Ô�ÉOÇ3Ê#ÃcÇ òWÜ�ÈMÃcÏ'ÆMÇ3ÈmÓ/Ï'Æ�ÒOÇ3Ä�Õ"Â9Ç�ÆmÆ2Ä'Â9ÊOÓ�Ä'Å9Å�Ä'ÌPÏ�É#Ê¿MÈ�Ú��

ÜmÒ#Â9È�Ð!Ï'ÆMí�Ð^Ä'È�È2ÉOÁ#Á�Ï'ÆM¿_Ç3Ñ�Ù"À ø ÿ�á»Ë'ÆoÄ'Ê&¿PðlÜ ß Í+ë'ëfØ�ì'ë'æ'â�Â9Ê�¿MÒOÇ��Aÿ�ý�� Û Á#ÆMÏ'Ë'Æ2Ä'ÌKÎGÄ'Ê�ÑË'Æ2Ä'Ê&¿.ðlÜ ß Í�Ø�â'â'åfØ�×'ë"Ú� �� !" #��$ �

÷.Ä'Ì)Â;Å9ÅkÎ'Ü Ú ß ÚZÎ�ç3ë'ë'ë"ï¥÷.À"Á�Ï'¿MÒ#Ç3ÈMÂ9È,¿_Ç3È_¿MÈlÓ/Ï'ÆlÇ�Õ�Ä'Å;É#Äf¿MÂ9ÊOË Ê"É#ÌPÇ�Æ2Â9Ã�Ä'ÅÁ#ÆMÇ3Ã�Â9Á�ÂZ¿MÄf¿MÂZÏ�Ê�Ó/Ï'ÆMÇ3Ã�Ä'È_¿MÈ�Ú&%('*)$+,�-/. '*0*)�13254�67�Î�89�Î«ç3×'× Í2ç3å�èÚ

NWP research in Austria T. Haiden Central Institute for Meteorology and Geodynamics [email protected] 1. Operational forecast system Operational limited area weather forecasts in Austria are made using version AL15 of the ARPEGE/ALADIN modelling system. ALADIN forecasts are made on two Central European domains, with horizontal resolutions of 12.1 km and 9.6 km, respectively. The number of levels in the vertical is 37 in both cases. The model is spectral, run in hydrostatic mode, with a semi-implicit, semi-Lagrangian advection scheme. Initial and boundary conditions are taken from the global model ARPEGE. A modified Bougeault scheme is used for deep convection, a first-oder closure for turbulent vertical transports, and the ISBA (Interaction Soil-Biosphere-Atmosphere) scheme is used to represent surface processes. Coupling frequency is 3 hours. Integrations up to +48 hours are performed twice a day. 2. Research a. Numerical prediction of inversion fog and low stratus The prediction of low-level cloudiness is a challenge for current NWP models. The underprediction of low stratus (high inversion fog) is a major forecasting problem in areas of eastern central Europe. Here the negative bias in cloud cover is the primary source of error in 2m temperature forecasts during wintertime (Greilberger and Haiden, 2003). Low stratus events are typically of large scale in the horizontal, and quasi-stationary over several days. Radiation and vertical mixing are the dominant cloud forcing mechanisms. According to 1-d and 3-d modelling studies the underprediction of inversion cloudiness is mainly due to a too smooth temperature profile across the inversion. In the case of the ALADIN model the smoothness is not caused by the limited vertical resolution but a side-effect of the specific formulation of the Richardson number dependency of the first-order turbulence scheme. This formulation sets a lower limit to vertical diffusion in the case of very stable stratification in order to avoid a reinforcing feedback loop of very cold air close to the surface. Ongoing work focuses on the development stage of the inversion, and on the comparison of first order and prognostic TKE turbulence closures in the prediction of inversion development. Also, alternative formulations for the cloudiness parameterizations have been tested (Seidl and Kann, 2002). The work is part of COST Action 722 ‘Short-Range Forecasting Methods of Fog, Visibility and Low Clouds’. b. Deep convection triggering On small grid scales (10 km and below) humidity convergence fields computed by a model represent not only synoptic-scale convergence but also meso-scale effects. Furthermore, one cannot expect the assumption of equilibrium between humidity convergence and convection intensity to hold on this scale. Similarly, on the small scale the triggering of convective precipitation should probably not be linked any more to the presence of humidity convergence. Convection should be allowed to become ‘active’ in the sense that it creates its

own humidity convergence. The problem of the diurnal cycle of convective precipitation, i.e. that the precipitating stage is reached too early in NWP models, is well known. Previous studies with ALADIN have shown that using a prognostic deep convection scheme tends to improve the mesoscale precipitation structures but does not solve the timing problem as long as the trigger function is kept unchanged. Simply using CAPE triggering does not solve the problem either. It appears that an improved trigger function needs to address more explicitly the cloud growth from Cu to Cu-cong into Cb. Starting in summer 2003, radar will be used to investigate the problem in more detail, and to test systematically different trigger and closure functions for deep convection. c. Prediction of cold air pools and katabatic flows The fact that NWP models usually employ a terrain-following coordinate system at low levels poses a problem in the forecasting of cold air pools in complex terrain. Problems also occur as aresult of the use of envelope orography, such that generally cold air pools contained within alpine basins are not simulated well. Moreover, katabatic flows, which significantly contribute to the build-up of cold air pools in mountain basins, are often too shallow to be realistically simulated in operational NWP models (Haiden and Whiteman, 2002; Zhong et al., 2002). A research initiative has been started to investigate the role of the coordinate system (terrain-following vs. z-coordinate), hydrostatic vs. nonhydrostatic simulation (Haiden, 2003b), model resolution, and turbulence parameterization on the prediction of the stable PBL in complex terrain. d. Prediction of heavy rainfall As a response to the August 2002 floods in Central Europe, which severely affected large parts of Austria (Haiden, 2003a), research has started on the potential of combining radar data and model results in the nowcasting of rainfall amounts for hydrological applications. References Greilberger, S., and T. Haiden, 2003: T2m nowcasting. ALADIN Newsletter, 23 (in press). Haiden, T., 2003a: On the performance of ALADIN during the August 2002 floods. ALADIN

Newsletter, 23 (in press). Haiden, T., 2003b: A note on the pressure field in the slope wind layer. J. Atmos. Sci., 60 (in press). Haiden, T., and C. D. Whiteman, 2002: The bulk momentum budget in katabatic flow: observations

and hydraulic model results. Preprints, 10th Conference on Mountain Meteorology, Amer. Meteor. Soc., Park City, Utah, 26-29.

Seidl, H., and A. Kann, 2002: New approaches to stratus diagnosis in ALADIN. ALADIN Newsletter,

22, 106-108. Zhong, S., C. D. Whiteman, and T. Haiden, 2002: How well can mesoscale models capture katabatic

flows observed in a large valley? Preprints, 10th Conference on Mountain Meteorology, Amer. Meteor. Soc., Park City, Utah, 69-72.

THE AUSTRALIAN AIR QUALITY FORECASTING SYSTEM

G. D. Hess1, M. E. Cope2,3, S. Lee2, P. Manins2, K. Puri1, K. Tory1, M. Burgers4, and M. Young5

1Bureau of Meteorology Research Centre, Melbourne, Victoria, Australia, 2CSIRO AtmosphericResearch, Aspendale, Victoria, Australia, 3CSIRO Energy Technology, North Ryde, New SouthWales, Australia, 4Environment Protection Authority of Victoria, Melbourne, Victoria, Australia,5New South Wales Environment Protection Authority, Lidcombe NSW, Australia. E-mail:[email protected]

The Australian Air Quality Forecasting System (AAQFS) is a joint project between theBureau of Meteorology (BoM), CSIRO Atmospheric Research (CAR), CSIRO EnergyTechnology (CET), the Environment Protection Authority of Victoria (EPA Victoria) and theNew South Wales Environment Protection Authority (NSW EPA) to develop a high-resolution airquality forecasting system. The initial development of AAQFS was funded by the Air Pollution inMajor Cities Program (sponsored by Environment Australia).

The project has a number of specific goals: to provide the ability to generate 24/36-hour airquality forecasts twice per day (available 9 am and 3 pm); provide forecasts for a range of airpollutants including oxides of nitrogen (NOx), ozone (O3), sulfur dioxide (SO2), carbon monoxide(CO), benzene (C6H6), formaldehyde (CH2O) and particulate matter (PM10 and PM2.5); provideforecasts at a resolution sufficient to consider suburban variations in air quality; and to providethe ability to generate simultaneous forecasts for a ‘business-as-usual’ emissions scenario and a‘green emissions’ forecast. The latter scenario may correspond to minimal motor vehicle-usage,for example, and which could be used to indicate the reduction in population exposure that couldresult from a concerted public response to a forecast of poor air quality for the next day.

The AAQFS consists of five major components: a numerical weather prediction (NWP)system, an emissions inventory module (EIM), a chemical transport module (CTM) for air qualitymodelling, an evaluation module, and a data archiving and display module.

The BoM’s operational Limited Area Prediction System (LAPS) system has been adaptedfor the AAQFS NWP component. Comprehensive numerics and physics packages are includedand recent work has paid special attention to the resolution and treatment of surface processes.The model has 29 vertical levels and a horizontal resolution of 0.05° (covering the State ofVictoria and most of New South Wales). This model is nested in LAPS at 0.375o resolution,which in turn is nested in the BoM global model, GASP.

EPA Victoria and CSIRO, with support from NSW EPA, have developed the emissionsinventory. The inventory component includes estimates of size-fractionated and speciated particleemissions, 0.01° gridded area sources over the densely populated regions and meteorologically-dependent emissions that are generated based on LAPS predictions.

The CTM has been custom-built for the project using state-of-the-art methodologies. Anotable inclusion to the CTM is the Generic Reaction Set photochemical mechanism, a highlycondensed (7 species and 7 reactions) photochemical transformation mechanism featuringminimal computational overhead. Parallel tests of a more comprehensive photochemistry, CarbonBond IV, are now being conducted. Particle transformation is modelled by a sectionally-basedparticle scheme. The transport fields are updated every 60 minutes. The CTM has 17 verticallevels, and simulations use a 0.05° outer grid, with nested 0.01° inner grids for major urban areas. While the focus of the AAQFS to date has been the forecasting of urban air quality, otherapplications are being considered. Australia is currently suffering from a prolonged drought,which has given rise to frequent dust storms. The feasibility of forecasting these dust storms isbeing investigated (Fig. 1). Another current activity is verifying long-range transport of theMelbourne urban plume (Fig.2).

Figure 1. Simulation of a major dust storm at 1200 Australian Eastern Standard Time, 23 October2002, which engulfed eastern cities from northern Victoria to Queensland in a deep cloud of dust.

The concentrations indicate 1-hour average PM10 values.

King Island

Tasmania

Geelong

Melbourne

Figure 2. Simulation of the long-range transport of the carbon monoxide urban plume from theMelbourne-Geelong region to northern Tasmania at 0600 Australian Eastern Standard Time,19 September 2001. The shaded region indicates concentrations ranging from 80–140+ ppb.

THE NCEP NONHYDROSTATIC MESOSCALE FORECASTING MODEL

Z. I. Janjic, T. L. Black, E. Rogers, H. Chuang and G. DiMego

Mesoscale Modeling Branch, Environmental Modeling Center NOAA/NWS/National Centers for Environmental Prediction,

5200 Auth Rd., Camp Springs, MD 20746; [email protected] Considerable experience with nonhydrostatic models has been accumulated at the scales of convective clouds and storms. However, numerical weather prediction (NWP) deals with motions over a much wider range of temporal and spatial scales. Difficulties that may not be significant or may go unnoticed at the smaller scales, may become important in NWP applications. For example, an erratic gain or loss of mass would be hard to tolerate in operational NWP applications. Another problem may arise regarding the control of spurious motions generated at upper levels by nonhydrostatic dynamics and numerics. Forcing the variables in the top layers toward a constant in time basic state in response to this problem appears to be inadequate for NWP. On the other hand, specifying time dependent computational top boundary conditions further limits the ability of the regional nonhydrostatic model to produce more accurate forecasts than the parent hydrostatic model. Based on these considerations, a new approach has been applied in developing the NCEP Nonhydrostatic Meso Model (NMM) within the WRF effort (Janjic et al., 2001, Janjic, 2002). Namely, instead of extending the cloud models to synoptic scales, the hydrostatic approximation is relaxed in a hydrostatic model formulation. In this way the validity of the model dynamics is extended to nonhydrostatic motions, the number of prognostic equations remains the same as in the hydrostatic model, and at the same time the favorable features of the hydrostatic formulation are preserved. This approach does not involve any additional approximation. “Isotropic” horizontal finite differencing employed in the model conserves a variety of basic and derived dynamical and quadratic quantities. Among these, the conservation of energy and enstrophy improves the accuracy of the nonlinear dynamics of the model. In the vertical, the hybrid pressure-sigma coordinate has been chosen as the primary option. The forward-backward scheme is used for horizontally propagating fast waves, and an implicit scheme is used for vertically propagating sound waves. The inexpensive Adams-Bashforth scheme is applied for non-split horizontal advection of the basic dynamical variables and for the Coriolis force. In real data runs the nonhydrostatic dynamics does not require extra computational boundary conditions at the top. The computational cost of the nonhydrostatic extension is about 20% of the cost of the hydrostatic dynamics, both in terms of computer time and memory. The relatively low cost of the nonhydrostatic dynamics justifies its application even at medium resolutions. In high resolution NWP applications, the efficiency of the described computational algorithm significantly exceeds those of several established state-of-the-art nonhydrostatic models. It is argued that the high computational efficiency has been achieved primarily due to the design of the time-stepping procedure. The high computational efficiency of the model demonstrates that meaningful nonhydrostatic forecasting/simulations are rapidly becoming feasible at smaller centers also, using workstations and PC’s. The dynamical core of the NMM has been adopted as one of the three major alternative dynamical cores within the WRF model. The conversion of the NMM code to the WRF standards and its incorporation into the WRF modeling infrastructure are nearing completion. The NMM has been run operationally at NCEP (Black et al., 2002). In high resolution NWP applications, the model has been highly competitive with mature hydrostatic NWP models and with other nonhydrostatic models. The 42 hour forecasts of 10 m wind in Southern California, valid February 6, 2003, 00 UTC, obtained using the operational hydrostatic 12 km Eta model (left panel) and the 8 km NMM (right panel) are shown in Fig. 1. The heavy arrows represent the observed winds. As can be seen from the plots,

the NMM forecast agrees much better with the observations than does the Eta forecast. Further examples of the NMM forecasts can be found in Black et al. (2002).

Fig. 1. The 42 hour forecasts of 10 m wind in Southern California valid February 6, 2003, 00 UTC obtained using the operational hydrostatic 12 km Eta model (left panel) and the 8 km NMM (right panel). The heavy arrows represent the observed winds.

REFERENCES

Black, T., E. Rogers, Z. Janjic, H. Chuang, and G. DiMego, 2002: Forecast guidance from NCEP's high resolution nonhydrostatic mesoscale model. Preprints 15th Conference on Numerical Weather Prediction, San Antonio, TX, Amer. Meteor. Soc., J23-J24.

Janjic Z. I., J. P. Gerrity, Jr. and S. Nickovic, 2001: An Alternative Approach to Nonhydrostatic Modeling. Monthly Weather Review, Vol. 129, 1164-1178.

Janjic, Z. I., 2002: A Nonhydrostatic Model Based on a New Approach. Meteorology and Atmospheric Physics, DOI 10.1007/s00703-001-0587-6.

Convective Asymmetries Associated with Tropical Cyclone Landfall Part I: F-Plane Simulations

Xudong Liang*1 2and Johnny C L Chan 2

1 Shanghai Typhoon Institute, China Meteorological Administration 2Laboratory for Atmospheric Research, Department of Physics and Materials Science

City University of Hong Kong *Email: [email protected]

1. Introduction The problem of tropical cyclone (TC) landfall has received more attention in recent years especially after it has been listed as one of the foci of the US Weather Research Program. The earlier modeling efforts by Ooyama (1969), Rosenthal (1971), Tuleya and Kurihara (1978) and Tuleya et al. (1984) have given some characters of the landfall TC. This study therefore represents an attempt to investigate the physical processes that, under idealized conditions, lead to changes in the rainfall distribution in a TC prior to, during and after landfall using the mesoscale model MM5. The following experiments on f plane and without mean-flow are performed: turning off the sensible heat flux over land (Expt. 2), turning off the moisture flux over land (Expt. 3), and setting the roughness length to be 0.25 m over land (Expt. 4). The coastline is moved to represent the movement TCs In all of the experiments. 2. Results The results of the experiments suggested that cutting off the moisture flux over land should have a profound effect on the TC characteristics while changing the sensible heat flux has very little impact. In Expt.3, before and during TC landfall, the maximum rainfall primarily occurs in upper and upper left quadrants (Fig. 1), which are very similar to that in Expt.4. Because the change in the roughness length of the underlying surface in Expt.4 causes a change in the wind speed in the PBL, which will then reduce the moisture flux. Therefore, both in Expt.3 and 4, the moisture flux on the land surface is smaller than that on the sea surface.

These results also suggest that the moisture flux is very important to determine the rainfall pattern before and during TC landfall, and the convergence/divergence in the onshore/offshore areas may not play the key role in determining the distribution of precipitation.

Fig. 1 Azimuthally-summed 1-h rainfall (cm) from 1 to 36 h in Expt. 3. Shaded is the increased rainfall compared with the control run. The 1-h rainfall is obtained by radially summing the hourly rainfall within each 15-km, 1 degree azimuth box (centered on the TC center ) out to 300 km at each degree from east. The x-axis is the degree relative to east (anti-clockwise, 1o ), left ordinate is the integrating time and right ordinate is the distance of the TC center relative to the coastline in km.

3. Effect of moisture flux During TC landfall, moisture flux over land is less than that on the sea, but it is not always favorable to reduce the rainfall. In fact, it usually increases the rainfall in upper and upper left quadrants related to the TC and reduces it behind. The reason is that the dry air over land and moist air over the sea are advected into the inner area of the TC and raised by rising motion. As a result, the instability should be decreased/increased by the inflow of dry/moist air in lower levels with moist/dry air in upper levels. The asymmetric distributions of ∂θse/∂z in Expt. 3 (Fig. 2) indicate that the lower levels in inner areas to the east and north of TC center are more unstable than those to the west and south while to the east and north are more stable in the higher levels before landfall. The increased/decreased instability will lead to increase/decrease of upward motion. Therefore, the anomalous upward motion begins to the east and north of TC at lower levels and increases to the west and south at high levels before and during landfall. The rainfall anomaly caused by upward motion does not directly occur in the east and north, because the rainfall occurs only after the air is raised up to a certain height. For a TC, the average horizontal wind speed is very large; when the air is rising, it is cyclonically advected to the downstream areas so that enhanced rainfall exists from northwest to southwest of the TC due to the anomalous upward motion. References Tuleya R. E. and Y. Kurihara, 1978: A

numerical simulation of tropical cyclones. J. Atmos. Sci., 35, 242-257.

Tuleya R. E., M. A. Bender and Y. Kurihara, 1984: A simulation study of the landfall of tropical cyclones using a movable nested-mesh model. Mon. Wea. Rev, 112, 124-136.

Ooyama, K., 1969: Numerical simulation of the life cycle of tropical cyclones. J. Atmos. Sci., 26, 3-40.

Rosenthal, S., 1971: The response of a tropical cyclone model to variations in boundary layer parameters, initial conditions, lateral boundary conditions and domain size. Mon. Wea. Rev, 99, 767-777.

Fig. 2. Anomalies of ∂θse/∂z (unit: K km-1) at σ levels 3 (~ 920 hPa), 5 (~ 850 hPa) and 7 (~ 750 hPa) (from left column to right) in Expt. 3 at 9, 18, 27 and 36 h (from top row to bottom). The abscissa and ordinate are respectively the zonal and meridional distances (km) relative to the TC center, which is represented by the typhoon symbol. The straight long dashed line indicates the coastline.

DEVELOPMENT OF A TWO-WAY MULTIPLY-NESTED MOVABLE MESH TYPHOON MODEL USING THE CLOUD RESOLVING NONHYDROSTATIC MODEL

W. Mashiko and C. Muroi Meteorological Research Institute, JMA, 1-1 Nagamine, Tsukuba, Ibaraki 305-0052, Japan

(corresponding author : [email protected]) 1. Introduction

In order to get better prediction of typhoon track, intensity, heavy rainfall and strong wind distribution, and inner structure, we are on the way to develop a two-way multiply-nested movable mesh typhoon model based on Meteorological Research Institute/Numerical Prediction Division unified nonhydrostatic model of the Japan Meteorological Agency (MRI/NPD-NHM; Saito et al., 2001).

MRI/NPD-NHM has been used to simulate various cases of mesoscale phenomena. However, numerical experiments of typhoon in high resolution have rarely been conducted due to the limited computer resources.

For the precise prediction of typhoon, enormous computer resources are needed because environment fields around a typhoon largely affect the typhoon evolution and therefore a large model domain is needed. Moreover, since convective processes play an important role in the typhoon dynamics, fine mesh (horizontal grid size; less than 1-2km) that represents convection is desirable especially near the typhoon center.

The major advantage of the two-way nested model is that it can effectively reduce computational time and memory requirements because both the large domain and high resolution can be accomplished together by arranging fine mesh over the typhoon area and coarse mesh over the environment. 2. Model description

The main procedures of the two-way interactive mesh refinement scheme are (1)construction of an interface between coarse mesh and fine mesh, (2)interpolation that is used to provide initial and boundary values of fine mesh, (3)feedback from fine mesh to coarse mesh, and (4)boundaries conditions for fine mesh.

We employed monotone advection algorithms for the interpolation procedure(2) and one point feedback with a smoother–desmoother used in Penn State/NCAR Mesoscale Model (Grell et al., 1995) for the feedback procedure(3). As for boundaries conditions(4), we adopted the radiative-nesting boundary conditions considering the difference between coarse and fine mesh’s values (Chen 1991). Some Fortran90 modules concerning two-way nesting were added to MRI/NPD-NHM in such a way that minimizes the change of its original source code. 3. Basic verification against 3-D mountain waves

To check the dynamical core of the model, we performed three-dimensional numerical simulations of a mountain flow with the model. The two-way nested model that covers the near-mountain area with fine mesh and elsewhere with coarse mesh produced almost the same results as those from the model with fine mesh all over the model domain and no erroneous wave reflection occurred near the boundaries (Fig.1). And computational time and memory could be reduced, compared with the model covered with all area fine mesh.

a. b. Fig.1 Vertical cross section of vertical wind through mountain top at 2-hour a. all area fine mesh (magnified corresponding to 2-way fine mesh area) b. fine mesh (2way)

915

920

925

930

950

955

960

965

970

0 3 6 9 12 15 18 21 24 27 30 33 36 39

MODEL HOUR

（hpa

）

ST TRACK

RSM

NHM（2way）

935

940

BE 945

4. Numerical experiment of Typhoon Saomai Numerical experiment of Typhoon Saomai (2000) was conducted with the two-way

double-nested movable mesh model. The model is nested with the Regional Spectral Model (RSM : horizontal resolution about 20km), an operational hydrostatic model of JMA. The initial condition is also supplied from RSM. The coarse mesh contains 200x200 grid points horizontally (grid size 18km), which covers a domain of (3600km)2 shown in Fig.3. The squares in Fig.3 show the domain of the fine mesh (horizontal grid size 6km : 151x151), which moves with the typhoon. Further descriptions of the model and nesting procedure are given in Table1 and Fig.2, respectively. Cold-rain explicit cloud microphysics were employed and convective adjustment parameterization is used jointly only in the coarse mesh. Fig.4 compares the central pressure of the model storm with the best track one determined by the RSMC-Tokyo. Fig.5 shows the sea-level pressure field and hourly accumulated rainfall of fine mesh at 34 hour forecast. As a whole, the model reproduced not only typhoon intensity but also rain distribution and the scale of eye well.

118E 120E 122E124E 126E 128E 130E132E134E 136E138E 140E142E144E 146E148E

104E

106E

108E

110E

112E

114E

116E

148E

150E

20N

22N

24N

26N

28N

30N

32N

34N

36N

38N

40N

42N

44N

46N

48N

50N

0 500 1000 1500

Fig.2 Relationship of nesting between numerical models

△ した

RSM 51h

NHM18km 39h

9/11 12

00UTC 00

NHM6km 39h2way

1way

は時

13

00

△ shows the timing of fine mesh’s movement

Table1. The model design

Fig.3 Orography of coarse mesh and domain of fine mesh Grid size (km)

Domain Coarse mesh Fine meshDimension (x,y) 200x200 151x151

Area coverage (km2) 3600x3600 906x906

Vertical levels 38 3818 6

Time step (s) 30 15

Fig.4 Time series of the minimum central pressure of

Typhoon Saomai

Fig5. Sea-level pressure and hourly rainfall at 34-hour forecast of fine mesh (22UTC 12Sep) Contour interval is 2hpa. The interval of hatch is 0.1,1,10,20mm.

5. Future plan We are planning to conduct further high resolution experiments near typhoon center with

triple-nested model in order to resolve the details of inner structure. Moreover, we’ll examine the effect of two-way nesting.

References C.Chen : A Nested Grid, Nonhydrostatic, Elastic Model Using a Terrain-following Coordinate

Transformation The Radiative-nesting Boundary Conditions. Mon.Wea.Rev119(1991) 2852-2869 G.A.Grell, J Dudhia, D.R.Stauffer ：A Description of the Fifth-Generation Penn State/NCAR

Mesoscale Model(MM5). NCAR TECHNICAL NOTE (1995) 122p Saito, K., T. Kato, H. Eito and C. Muroi, 2001 : Documentation of the Meteorological Research

Institute/Numerical Prediction Division unified nonhydrostatic model. Tech. Rep. MRI, 42

�� !��"�$#�� %�&!��'! ��

( )+*�,�-.-/,10325476869,;:=<>)�? 4.@A,3BDCE:�?�)+F�,3B/@7C9036GCE:=H�)IFKJL2NMO,32NCE:�P�)+Q�47B/JAR.RSC ,32�TVU5)W(ICYX.,369T3C

Z\[�]�Z�^`_`ab[dc�[WeSfhgEihjNk3l+ame7n/eSiho�ihlEihfhgEpAk3l+_qeSosrqgEpSe3cut�ihlEihfhjNkvcvwOn.k3l9xpSiho�o�eSy�z�ihj`{`gEj`f|k3}vn/~`ihoS��pS��k3o�y�gEfhlEg8�\y��|oS� k3o�zNkv� eS�|oS� g9n

��$�I�E�1��D�.�1��1��G�h�.�1��D��1��.�1�.�� D�A��$�.��.�1�D�L��+��;��s��1��G�D�$�;��$��1�q�G�D��$��1�.��;�G�\�q��D�1�D� �3��D��;��.�1�G��1��¡�1��1�1�s�A�q��s¢L��D��A��D�¡��h�/�D��D�W�.�1�£��.�s�G�;¤��G�V��d�G��$�\��D��/��A�q��¦¥�G�Y�£�1� ��L��D��>�1��b�G�h�.�1��D�m�.�1�D�L��.� §��G��$�;¤$�;�m��$�¨¤��1� ��G�;�3¢L��D��;��¤$��G�;�m��d�G��D�3¢3�A��D�V��h�$�D��©«ª §&¬|��1� ��7�1�D��$�&��$�1��®¢L�s�1�$�1�&��D��/��G�;�d�1��G�h�.�1��G��L��D�� .�1�D�L��.¯��G�=��;��.��D�=��G°�¥�.¤��G��E�1��D�.�1��1�.�.¤$�s�A��D�G�¨��$�D�G� ��q�1�/��¢L��W�.�1�;��¤$��G�;��E�1� �s�1�$�1�D� �G�;�$�1�.��1��±L²d�$�;¤$�s�.�I��1�$�h��W�D��D��q�G�=��D��/��G�;��h��D�d� ©«³�´=µ ��s¯L�d�1�L�� D�A��$�.� �/�D�L��D�.¯3�1�� ³=¶ ¬V·�¸¡�1��;�$��$�D��¯L��D�1��;�G�1�D�¹��1��>�1��b��1��D��q��1��W��$�>��1�q�A�q��\�A��;�1�s�N��º��.�;�A�s��$�|��$�>¤��1� ��d�G��D�3¢�A��D��h�$�D�� G��d�D��D��q�G��h��D�d�.¯�§�º ´ §�¥ µ ¬|º»�$�.�1�D�G�1�`�D� ª�³�´=µ ¯h� ª ��d�G��D�3¢3�A��D� ³ ��D��q�G�´ ��D��/��G�;� µ �h��D��¼&¤$�I�1�N�|½��¤$�`�.�O¢L�D��D��q�G�D¾&�1�I¿DÀ1Á��D�G�D��D�L��1�`��$��.��/�;��D�.¤$��G�1� �1�`�.�s�A��G�;��1�³=¶ ¬V·�¸ ³�´=µ �s¤��.¯`Â��1��.��¤��.�s�&�A��d��$�D�L��GÂ��D��1��N¯N�E�1��D�1�s� �1��$�D��¯u��.��¤��.�&��.��D��D�L��A��G�1��D��`�.� © º�¬¡�� s�$�;��D�N� ³ �1�s�|�1�I�¤��s�|º�¬|� ��7�h��$�D��`�1��V��G��1�N�1��V�`�;¤��A��¨�/�;��G��G�;��E�1�� ª §&¬Ã��L��.�1�s�A��G�;�N¯`�1�D�$�.�s�A��$�¨��|�� =�D��>��d�1�� ¢3��GÄ.�1¯`��G�;�3¢L��D��;��¤$��G�;��D��D��q�G�1��$� ª�³�´=µ��.��$�h�$�;�G�1�1�¹�1��G�7� ��m�/�;��q��$��$��`�D�$�.Â��\�1��$�V��1�q�A�q��d�A��;�1�s� © �£��.��1��¦Å`�.��D�L��.�1�;��¤$��G�;�m��/�D��A�s�G�;��7�h��$�D�¡��|��$��E�1��D�.�1��.�/&� �G�� $��G�;�3¢L��D��;��¤$��G�;� �$�.��1��+�1��$� ª §&¬��L��.�1�s�A��G�;��.¯1� �G��&��d�G��D��/�;��q¤$��A��G�;��1�h��L�1�D��D�L� © ¬|�A�s��G�;��L�.�+�1�8�G¯LÆAÇ1Ç$¿A¯�È=É�ÊAË/ÌÎÍ�Ï$Í$Ð�ÑAÒ;Í�ÓmÔ/ÕÌYÔÑAË.Í�ÖvÑD×7Í�¯�ØhÙvÚv¯vÆAÇ;Û1ÀD¢hÆh¿1¿DÀ3Üu¬V�;�G��D��`�.� �1�8�G¯qÆAÇ1Ç$¿A¯�È=É�ÊAË/ÌÎÍ�Ï$ÍuÐ�ÑAÒ;Í�ÓmÔ.ÌYÔÑAË.Í5ÖvÑD×7Í�¯�ØhÙvÚv¯vÆAÇ;Ý1ÛD¢hÆAÇ;ÛS±$Ü¬V�;�L��1��v�.��1�8�G¯$ÆAÇ1Ç$¿A¯`Þ�ÑAß�à áEßvÍ`â=Ë�ÑD×7Í=ã�ÔÑsäqåLÒSæ.Í�¯vçN¯$Á1²;è.¢hÁ1Û1Û;s�=éO��$�&ê3¤$�1��D�>��.�E�.��D��/�D�=�G� ��1� �`�.�D��$�7� �|��A�.¯`�7�1�.�W�>�3¤��`�.� �1�I��D��.�1��D�&�1��|�E�1�&��.�1�.�s�1�5�E�1��D�.�1��&�s�1�$�1�D� © ±L²D¢v¿DÆAÇ��$�;¤$�s��s¯ ª�³�´=µ�`�.��E�1�s�d�=�`�.��.��1� ³�´=µ �/�;��/�.�s��$��$�Wê3¤��1�L��G��A��G�1�W�E�1��D�.�1��1��L��D��&��D�.�G�q�G��A��G�;�N¯$�1�� D��1�d��$��1�.�1�1�s�A�q��.�1��G�h�.�1��A��G�;�b�1�&��$��.�;�G�;��d��;��d��G�1�D�G�b�SÅ`�D�/��D��3�b��$��.�1�D�L��.�'¸$�;��G�7� ��$��$�>�D��/�;¤$�s�A�;��$��D�¤��G��1� �1�b�D�A�s�G�£�/ëh�`�.�s��D�L��1��q��1��1¯+��$�>�1�D�$�.�s�A��G�;�b�1� �1�ì½O�/ëh�`�.�s��D�L��1� ¢�1�`�.�s�A��G�;��1��¾��d�G��D�3¢3�A��D�\�D��D��q�G��D��/��G�;�V��h��D��¯q��$�dí�î�ïuðdî�ñvòhó`ô`ï õAö9÷ø®ù.ú/û�¯N��1� ��D�/�D�L��G��A��D��;�W��$� ³=¶ ¬V·�¸|�/�;��q¤$��.��h��D�ü¤��$�.�u��$� �1¤��q��/�D�u�1� ¶ ¼ µ ¬¡¼�� ¶ ¼ µ ¬¡¼ ©Î¶ ¼&��1��¤��E�1� µ �d�1�� ¢3��.�1�G��¬¡¼&�$�D��$�$¯`��W� �/�;��S¢3��h�$�D�8� �1��3&��W��/�;��1��¤��ý��L�1�;�G�h��$�|þ��.�s�d�1�L�1¯5é®��1�G�1¯µ � �G��Ä.�.�s��1��N¯1þ��.�D�/��1�� ´ �;��1��;��$� �$�.�1�D�G�1�`�D��D�L�5�1�$��$�=��d�G��D�3¢3�A��D��$�;�3¢3�L�h�$��;��A��h�$�D�ª �1�A�1��¬V�h�$�D�� ©«ª ¬¡s� ¶ ¼ µ ¬¡¼�¢ ª�³�´=µ �1��d�5��$�.��.�E�1��A��$��$�.�1�D�G�1�q��D�L�I�1��¢L�1�`�.�s�A��G�;��1�3��D��1�v�¨½��$�1��D��¤��¢L�s�1�$�1�D¾ © ±L²D¢v¿DÆAÇW�$�;¤$�s��N��1�q�A�q��E�1��D�.�1��$��h��D�ÿ¤��$�� ª §&¬ �7�1�.��¨�/�;��q�A�s�A��G�1�D�G�|��A��1�\�$�;�d�1�� © ��.�\Â��$��¿7s¯u�/�7�1�.�s��$��1��5�/�;¤��L��s�G�D��L�1�;�G�1�D�¡�� ¶ ¼ µ ¬¡¼��v��1�$�h�

276

280

280

284

284

288

288

288

292

292

292

296

296300

304

35°N35°N

40°N 40°N

45°N45°N

50°N 50°N

55°N55°N

10°W

10°W 5°W

5°W 0°

0° 5°E

5°E 10°E

10°E 15°E

15°E 20°E

20°E 25°E

25°E 30°E

30°E

Monday 27 January 2003 12UTC OFFNB Forecast t+108 VT: Saturday 1 February 2003 00UTC 700hPa geopotential height

¸��G�;¤$��d¿�¼��`�.�s�A��G�;��1� ¶ ¼ µ ¬¡¼�¢ ª�³�´=µ �$�;�d�1��N�

��W��$��/ëh�`�.�s�G�D��/��;�1��$�D�d��¤$�s��$��$��D�A�s�G��/ëh�`�.�s��D�L��1�h�q��1��1¯L�G�+�=�1��$�D�.��$�D�\��.�®¢3¤$�d��$� �¤��G��1��E�;��G�7� �

�

COSMO−LEPS products ready

0h

0h

0h +24h

+12h+12h

+96h

+144h+120h

+108h +132h

+120h

Start of LM integrationsClustering Times

1 5

3 M

E M

B E

R S

N − 2 N − 1 N N + 1 N + 2 N + 3 N + 4

youngest EPS ready

OLDEST EPS

MIDDLE EPS

YOUNGEST EPS

Start of LM integrations End of LM integrations

12 12 12 1200 00 00 00 000012 12 12O

¸��G�;¤$��WÆ�� .��1��1� ¶ ¼ µ ¬¡¼�¢ ª�³�´=µ �¤��G��1�

� ��$��.�\�¤��.�/�D��G�1�>¿DÆD¢3�$�;¤$�O¢3��A�1�1�D� ³�´=µ �s¤�� © ��A��D��A�\¿DÆ�� ¶ �1�I��D��Þ��;¯`�A�&Ç1Ç>�1��m¿DÆ�� ¶ �1��D�VÞ��7 �A��1��;¤$�`�D�>��1�1�.��$�.��d�1��1�D�$�.�s�A��¿DÀ1ÁD¢3��D��`�.� �¤$�`�.�O¢L�D��D��q�G�1Ü© ��.��Â��$��Æ;sÜ

� ��G�.�s�A�s�s��.�1�h�.��¤��.�I�1��1�G�h��5�`�.��E�1�s��D��;�\��$��¿DÀ1Á��D��`�.�s�+��1�+��1��;¤$��1��h�D�G�D��D�L��L��À��.��¤��.�s� © �1�$��¦Å`�.��D�L�`�`�1�q¤��A��G�;��sÜ7��$��.��¤��.�s��$�=�A�A�s��A�q�G�D�N�A��5¯ ��¯ � �1�� © ��`�D�.�GÂq��3¤��d��G�Y��A��$��.�\��D��¤$��\�G�.�1�D�� © ÀAÇ1Çh¯�èSÇ1Çh¯�²1ÀAÇ�� ´ �L��1�� A��Y� ��E�1��D�.�1��D� © �«��Û1Ý�1��'�«��\¿DÆAÇb�E�1�¨��$��½O�1�;¤��$�1�D��¾ ³�´=µ sÜ ��$�¡�.��¤��.��$�;�d�1��'�/�7�1�.�s�¨��$�¡��.�;�G�;��ÁAÇ��¢hÝAÇ��¯ÆAÇA·�¢L±;Ç ³ Ü

� � �G��W�D�1�s��.��¤��.�D¯A�;�$� ��.��D��D�L��A��G�1�=��D��`�.� © º�¬¡u��5��D�G�D�/��D�\�1�.�/�1�s��$��&��$� �E�;��G�7� ��$��/�s�G��.�s��

��$��º�¬ ��N��A�u�D�G�D��D�L�5�.�G�;��D��u�� $��D��`�.�s�N�1�$�G��N�7� ��.��¤��.�s�u�1��W��;��1�L�

�E��;��$��D��`�.�s� �1��$�&�1��$�.��.��¤��.�s�.Ü$��1��/�D�=�A��.�1��.¤��A��D��¤��$��$��1��&�A�A�s��A�q�G�D��1��$��1��.��s��¤��D��¨��$��.��¤��.� �1��1�G�h��.Ü$�$�D��/�1¯vÀ\º�¬|� �A��W��D�G�D�/��D�NÜ

� �D�1�s�|º�¬ ��7�h��$�D�=��G��1�q�1��¨�`�;¤��A��d�/�;��G��G�;��E�1��$��L��.�1�s�A��G�;��=� �G�� ª ¬£¯h� ��s�¨��s¤��¨À��D��E�1��¿DÆAÇ\�$�;¤$�s�.¯h�1�G�=�D�h��A��$��A��¿DÆ�� ¶ �1��D�¨Þ��&�1��D��$��A�&¿DÆ�� ¶ �1��D�|Þ��$Ü

� ��$� ª ¬"��1� ��$�1�s�GÄ.�;�L��1�q��D��;��¤$��G�;��ü¿.Ç��h��¯$Á1ÆW�1�.��.�1�v�G�.�1�D�� 1��d��$� ��¢3��.��¤��D��E�1� ��$�W��L��.�1�s�A��G�;�� ÝAÇd��D�AÜ

� ��1�q�A�q��G�Y��d�A�q��q�1��D�£�;� ª ¬ �s¤��A��1�D�$�.�s�A��D�m�3� �1��G�;��$�¨��|�D�1�s� ª ¬ ��L��.�1�s�A��G�;��£� �D�G�;�L��1�`�1��G�;��1�=��m��$�|�`�1�q¤��A��G�;�¹�1�&��$�|�.��¤��.��E��;� � ��s� ��$�|º�¬ © ��7�h��$��G��1�u�1��|�`�;¤��A��¨�/�;��G��G�;��=�=�1��D�G�D�/��D�NÜu�$�.��.�s�d��&��h��¤��/�� © ��A�&��.¯v��$�\À ª ¬��/�D��A�s�G�;� ��¨��.�s�d��1��¤$��«�1�/��1��V¤$��`�.�O¢3�G�.�1�D�vÂ��D��A��1��h��¤��/�D�NÜ

� ª ¬ �1�s�G�h¢LÂq�G�D��A��W��D�d��A��D�¨��d��$� ¶ ¼ µ ¬¡¼ì�/�;�d��¤��G�Y�>�E�1��.�A�1��¤��A��G�;�NÜ¶ ¼ µ ¬¡¼�¢ ª�³�´=µ ��h��¤��/�� A��\¤��¤��1��G��D�A�1��A�q�G��3��Û�� ¶ �1��D� Þ�¯q� �D��5��V��¨�`��.�A�1�¦¥

¤��A��D�¨�3��1�`�.�s�A��G�;��1�v�E�1��D�.�1��.�s�.� ¶ ¼ µ ¬¡¼�¢ ª�³�´=µ ��D�d��A��G�;��A��D�V��¤$�s��$�� 7�1�D��`�.��ÆAÇ1Ç;Æ�1��N¯3�A� ��$� ��1�N��s�G��$� © ¬|�A�s�s��ÆAÇ1Ç;Á;s¯$��$��h��D� ��h�`�D��$�W��D��D��\�1��D�� G��¤��.�«¤��$�D��\��.�®¢L�1�q�+��3�;�d�.¯;�q�A��.¤��A�s�G�W��.�s�d��1�q��$��1��1��/��;�G�1�D�\��E�1��D�.�1��.�s�I��\�.�1��D�I�1�q�/ëh��D��.�1�D�L��.� §��1�! ��D�/��G�1��1�q�A�q��1�.�s�GÂq�.�A��G�;��1�5��$��h��D� ��1��`�D��$��.�A��s�G�D�¨�;�N�

Analysis of the evolution and motion of tropical cyclones on the basis of calculationsusing the ETA model and satellite data

(Section 5)A.E.Pokhil, A.D.Naumov

Moscow, Hydrometeorological Research Center of Russia 9-11 Bolshoy Predtechenski Lane, 123242 Moskow,Russia. E-mail: [email protected]

and [email protected]

It hasbeeninvestigatedwhetherit is possibleto takeinto accountmultiple interactionoftropical cyclonesin the forecastof their motion in the north-westof the Pacific basedoncalculationsof pressurefields and the wind using the ETA model and the comparisonof thecalculation results with the GMS satellite data and actual TC trajectories. The performedanalysisof thebehaviorof the interactingTCs hasrevealed,in particular,strangemotionof thevorticesandtheir suddendisappearance.In the typhoonseasonof this yeartwo TC pairs,whichinteractedduringdefinite time intervals,wereof particularinterestfor investigation.ThesewereChataanandNakri (9 - 10 July); HalongandNakri (12 - 14 July). In the first caseChataanwasthe leadingone.The distanceat which their interactionbecamedetectablewas1000km. It iswell seenat thefields calculatedusingtheETA model(Fig. 1a).SimultaneouslyChataanbreaksin two vortices,its southpart keepsinteractingwith Nakri: on 10 July the trajectoryof Nakrisharplychangesits direction,andtheTC movesright theeasttowardsthenewly formedvortex.On 11 - 12 July theymergeinto a singlevortex,which canbeeasilyseenfrom thesatellitedata.On 13 July at midday the distancebetweenthe centersof the approachedby that momentTCHalongandNakri constitutes800 - 900 km. Intensiveinteractionis observed,with the strongertyphoonHalongentrappingNakri andthe lastoneceasingto exist asa separateformation(Fig.1b). Furthermotion of the typhoontakesplacewith a 90° turn. Having turnedthe TC movesalong the polar front in the north-east direction.

Thecalculationof thewind andpressurefields usingtheETA modelreflectsratherwellthe location of strong formations (STS and Ty), whereasweak vortices are not distinctlydetectablein the analysisof fields. Besidesthat, the pressurein the centerof vortices is notconsistentwith actualpressure.As far asthe characterof the vortex interactionis concerned,itdependsto a considerableextenton theparametersof thevorticesthemselves.It hasbeenshownin our previous works devoted to the experimentswith vortex pairs using barotropic andbaroclinicmodels/1, 2, 3/. So,to obtaingoodcalculationresults,it is necessaryto synthesizethealgorithmof therestorationof the initial vortex,which modelstheTCsin thefields of objectiveanalysis (“vortex initialization”).

For understandingthe evolution of vortices and their interaction the comparisonofcalculations with the satellite data is quite useful.

References 1. A.E.Pokhil, A.V.Nikolaeva. The behavior of a pair of interacting cyclonic vortices in thebarotropic model of the atmosphere. Meteorology and Hydrology., 2000, No. 11. 2. Pokhil A.E., Sitnikov I.G.Diagnostics of motion and development of two series of tropicalcyclones for the 2000 typhoon season. -CAS/JSC Working Group on numericalexperimentation. Research activities in atmospheric and oceanic modelling. Report N 31, april2001. WMO/TD-N 1064.

3. A.E.Pokhil, A.V.Nikolaeva. Numerical experiments with a pair of interacting vortices in abaroclinic model. Meteorology and Hydrology, 2002, No. 3.

4.

Sea level pressure, wind 1000 hPaa)

09.07.2002b)

14.07.2002Fig. 1a,b.

Study on limited-area short-range ensembleapproaches targeted for heavy rain in Europe

Kai Sattler and Henrik FeddersenDanish Meteorological Institute, Lyngbyvej 100, DK-2100 Copenhagen Ø

Email: [email protected]

A common way to address the uncertainty in numerical weather prediction (NWP) is to designan ensemble, attempting to forecast the likelihood for the expected future event on basis of thebest possible knowledge about the atmospheric state at the time of the forecast production.The interest of doing this with a high-resolution limited-area NWP model (LAM) has grownduring the last years, and there are many different possibilities of designing a LAM ensemble.This study investigated two different approaches of creating a small LAM ensemble for rainfallprediction. They are both based on the Danish Meteorological Institute’s (DMI) version of theHIgh Resolution Limited-Area Model DMI-HIRLAM, which has been configured as a doubly1-way nested system, were the outer model receives initial as well as lateral boundary data fromthe ECMWF ensemble prediction system (EPS) in a 6 hour frequency. This outer model coversEurope and the North Atlantic as well as parts of North America. The inner DMI-HIRLAMmodel is the actual model of interest and covers Europe with a 0.1 ◦ grid. Integrations for thisstudy were performed up to 72 hours ahead.

The first LAM ensemble is designed to represent an uncertainty in the initial condition (atmo-spheric state) as well as at the lateral LAM boundaries. It is created on basis of the ECMWF-EPS with 50+1 members, from which a small set of significant members is selected (selectionensemble (SE)). The significance of a member is estimated on basis of accumulated precipitationover the area of interest, which includes the river basin that may potentially experience floodingdue to heavy rain. The control forecast is allways included in the selection. The selected mem-bers from the global EPS then make up the initial and boundary conditions for the members ofthis first LAM ensemble design. The size of the selection ensemble has been defined to be 5+1.

The second LAM ensemble tries to address model related uncertainty in the rainfall forecastsand consists of the variation of the physical representation of the condensation and convectionprocesses within the LAM (parameterization ensemble (PE)). A similar approach has e. g. beeninvestigated by Stensrud et al. (2000), even though with different schemes. Five parameteriza-tion schemes were available for this study:

• Stratiform condensation scheme (STC; Kallberg, 1989)• Anthes-Kuo scheme (AKC; Kuo et al., 1974; Anthes, 1977)• Sundquist scheme (SQS; Sundqvist, 1993)• Tiedtke scheme (TDS; Tiedtke, 1993)• STRACO (Sass et al., 1997)

This results in an ensemble size of 4+1, were the last member (STRACO) is identical with thecontrol of the SE design.

The two ensemble designs SE and PE were tested on three historical heavy rain events indifferent European areas and of different synoptic and dynamic characteristics. The cases arethe heavy rain over the Piemonte region in Northern Italy from November 1994, the strongprecipitation during the end of January 1995 over the basins of the rivers Rhine and Meuse,

and the heavy rain over the Odra basin in Poland/Germany/Chzech Republic in July 1997. Avalidation using ranked histogramms shows that both designs exhibit too little ensemble spread(not shown), the PE designs, however, shows a larger spread than the SE. Reliability diagramsfor the 5 mm threshold of 24 hour accumulated rainfall between 42–66 hours forecast rangeare shown in Fig. 1. The two ensemble designs perform comparably well, with the PE designtending to exhibit a slightly better reliability. Both designs generally tend to overestimate higherobservation frequencies at this threshold.

Obs

erve

d fr

eque

ncy

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1Forecast probability

2-10 Jul97; t+66h;W+C Europe; 24h precip > 5 mm

PESE

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1

Obs

erve

d fr

eque

ncy

Forecast probability

23Jan-2Feb95; t+66h;W+C Europe; 24h precip > 5 mm

PESE

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1

Obs

erve

d fr

eque

ncy

Forecast probability

1-6 Nov94; t+66h;W+C Europe; 24h precip > 5 mm

PESE

Fig. 1: Reliability diagrams for the 5 mm threshold of accumulated precipitation between 42–66 hoursfor the selection ensemble (green) and the parameterization ensemble (red), for the three cases of heavyrain: Piemonte case November 1994 (left), Rhine/Meuse case January 1995 (middle) and Odra case July1997 (right).

The performance of the ensemble designs reveals to vary strongly from case to case and even ona daily basis, and a combination of both ensembles seems advantageous in many cases, but notin all. A weakness of the SE design is that the selection of the members is based on the globalEPS rainfall fields only, because small scale effects as well as non-linear developments withinthe LAM integration are not known a priori. This may in some cases have the effect that theensemble spread of the SE in the LAM rainfall is smaller compared to an ensemble integrationof all 50+1 members with the LAM, even though this is the case in the global rainfall fields.The PE design is quite promising, as long as all schemes perform comparably well in order notto degrade the ensemble performance.

References

Anthes, R.A., 1977: A Cumulus Parameterization Scheme Utilizing a One-Dimensional CloudModel. Mon. Wea. Rev., 105, 270–286.Kuo, H.L., 1974: Further Studies of the Parameterization of the Influence of Cumulus Convectionon Large-Scale Flow. J. Atmos. Sci., 31, 1232–1240.Kallberg, P., 1989: HIRLAM Forecast Model Level 1 Documentation Manual. 77pp. Availablefrom the Swedish Meteorological and Hydrological Institute (SMHI) Norrkoping, Sweden.Sass, B.H., 1997: Reduction of numerical noise connected to the parameterization of cloud andcondensation processes in the HIRLAM model. HIRLAM Newsletter, 29, 37–45. Available fromthe Swedish Meteorological and Hydrological Institute (SMHI) Norrkoping, Sweden.Stensrud, D.J., Bao, J.-W., Warner, T.T., 2000: Using initial condition and model physicsperturbations in short-range ensembles. Mon. Wea. Rev., 128, 2077–2107.Sundqvist, H., 1993: Inclusion of Ice Phase of Hydrometeors in Cloud Parameterization forMesoscale and Large Scale Models. Beitr. Phys. Atmos., 66, 137–147.Tiedtke, M., 1993: Representation of Clouds in Large-Scale Models. Mon. Wea. Rev., 121,3040–3061.

NUMERICAL STUDY OF THE SHAPES AND MAINTENANCE MECHANISMS OF MESO-β SCALE LINE-SHAPED PRECIPITATION SYSTEM IN THE MIDDLE-LATITUDES

Hiromu Seko(1), Hajime Nakamura(2)

(1) Meteorological Research Institute, Japan Meteorological Agency, Tsukuba, Japan

(2) Numerical Prediction Division, Japan Meteorological Agency, Tokyo, Japan 1. Introduction

Precipitation systems that cause severe disasters often have the line-shaped precipitation regions. These precipitation system has a scale of 100km (meso-β scale), and the convective cells generate one after another in the systems. The repeating generation of the cells leads to the long life of the systems.

Line-shaped precipitation systems reported so far are divided into mainly two types; squall line (SL) type and back building (BB) type. The structure of the SL type is similar to the squall lines observed in the Midwest of U.S. In the BB system, the convective cells generate in the tip of the system repeatedly, and move backward. Besides these types, the systems with the carrot-shaped cloud region often inflict heavy rainfalls.

The structure and long-lasting mechanism of the SL type are relatively well-known. However, other types are not fully investigated yet. Therefore, the precipitation systems of each type were analyzed by special observation data and the outputs of the numerical simulation. The results of the analyses are summarized in the section 2.

Because the results of the analyses indicate that the environment of the systems (e.g. vertical profiles of horizontal wind and humidity) depends on their types, the numerical experiments were performed by changing the environment conditions to investigate the factors that determine the type of the precipitation systems. 2. Case studies In this section, the observed features of the observed line-shaped precipitation systems are summarized. 2.1 Squall line (SL) type A line-shaped precipitation system developed over the Kanto area on 16 August 1995 (Seko et al,1998). The intense

convections developed along the front side of the system, and the wide weak precipitation region existed on its rear side. The airflow structure consisted of the two airflows: One is the low-level inflow of warm and humid air from the south. It was lifted up by the outflow of the system along its front-side edge. And another flow is the middle-level rear inflow that invaded into the system. When this system passed, the temperature descended more than 10°C and the gust of 24m/s was observed. This indicates that the rear inflow generated the cold pool by the evaporation of the water substances. 2.2 Back Building (BB) type The cloud cluster that consisted of the several meso-βscale precipitation systems stayed over the southern Kyushu on 7 July

1996 (Seko et al. 2000). The meso-β scale system was further composed of several short convective bands. The short band in the meso-β scale system indicates that a scale exists between the meso-β scale and meso-γ scale. Hereinafter, this scale of the short bands is referred to as the meso-βs scale. In these meso-βs bands, the new convective cells generate at the tip of the band repeatedly, and moved backward. Because the generation point of the cells was only near the tip of the band, the length of the bands was shorter than that of the meso-β scale system. The direction of the middle-level inflow was almost same to that of the low-level inflow, and the bands extended to the leeward side of the middle-level wind. The low-level inflows entered the meso-β system from the south and ascended at the tip of the meso-βs scale bands. The middle-level airflows passed between the meso-βs scale bands. 2.3 Back and Side Building (BSB) type The precipitation system with the carrot shaped cloud region was observed over the Kanto area when the typhoon 9426 was

approaching to Japan (Seko et al.1999). The precipitation system extended to the same direction of the middle-level wind and the width of the cloud became wider at the leeward side, producing the carrot-shaped cloud region. The relation of the low-level inflow and middle-level wind was different from those of the SL type and the BB type: the direction of the middle-wind was perpendicular to that of the low-level inflow. The convective cells were generated on the upwind side of the middle-level wind (back building) and moved to the leeward side, and merged into the long intense convective band. After merging with the band, the convective cell was enhanced by the low-level inflow from its lateral side (side building). From these features, this carrot-shaped band had the new mechanism that should be called 'back and side building type'. 3. Numerical experiment under the simplified environmental condition The comparison of the observed precipitation systems suggests the environmental factors that determine their

type. The low-level convergence was common in aforementioned three cases. On the other hand, the middle-level wind direction to the low-level inflow and the middle-level humidity depended on the types. Thus, the numerical simulations were performed by changing these factors. In this study, non-hydrostatic model of Meteorological Research Institute was used. The horizontal grid interval of 2km was adapted to reproduce the convective cells. When the direction of the middle-level wind was opposite to that of the low-level inflow, the intense convective

band developed along the convergence line. The rear inflow at the middle-level descended and the cold pool was produced near the surface (Fig. 1a). When the directions of the middle-level wind was same as that of the low-level inflow, the short convective bands

generated side by side along the convergence line in the meso-β scale system (Fig. 1b). The short convective bands extended to the direction of the middle-level wind. In these bands, the new convective cells were generated at the upwind tip of the bands, and moved to leeward. When the direction of the middle-level wind was perpendicular to the low-level inflow, the intense band developed (Fig. 1c).

*Corresponding author address: Hiromu Seko Meteorological Research Institute, 1-1, Nagamine, Tsukuba, Ibaraki, 305-0052, Japan, e-mail:[email protected]

Although the shape of the band was similar to the band in which the direction of the middle-level wind was opposite to that of the low-level inflow, the middle-level airflow did not descend and the intense cold pool was not produced. The convective cells generated on the upwind side of the middle-level flow, and moved to leeward. These simulated features in three

experiments were the same as the observed ones. This result indicates that the direction of the middle-level wind is one of factors that determine the type of the precipitation system. When the relative humidity between the heights of 1.5km and 4.5km was reduced, the relation between the type of the precipitation systems and the directions of the middle-level wind and low-level inflow was not changed, although the weak precipitation generated more widely. This result indicates that the middle-level humidity

is not the primary factor that determines the type of the system.

4. SUMMARY Besides the precipitation systems of the SL type and the BB type that have been investigated so far, it was found

that the precipitation system of the BSB type that often caused the heavy rainfall exists. When three types of the precipitation bands were compared, there are the differences in their environments and structures. As for the environment of the systems, when the direction of the middle-level wind is opposite, same and perpendicular to that of the low-level inflow, the precipitation system of the SL type, the BB type and the BSB type developed.

The shapes and structures of these systems were explained from the viewpoint of the middle-level wind (Fig.2). In the SL, the middle-level rear wind descends in the precipitation system and enhances the outflow of cold air. The outflow intensifies the convergence with the low-level inflow, and the narrow intense convective band developed along the convergence line. In the BB type, the middle-level wind moves the convective cells to leeward, producing the meso-βs scale convective bands. Since the directions of the low-level and middle-level winds are same, the structure, with which the meso-βs scale bands develop side by side, is favorable for the low-level inflow to ascend without disturbing the middle-level wind. In BSB type, the new cell generates on the upwind side of the band, and moves to leeward with being intensified by the low-level inflow from the lateral side. This evolution of the cells

produces the carrot-shaped precipitation region. The numerical experiments were also

performed with changing the middle-level humidity. The types of the systems were not changed when the middle-level humidity was reduced. However, the decrease of precipitation amount depends on the types. Precipitation amount was decreased slightly in the SL type, while it was decreased significantly in the BB type. This difference can be explained by the middle-revel flow. In the BB type, the middle-level flows entered the meso-β scale precipitation system and evaporated water substances while they crossed the precipitation regions. In the BSB type, the middle-level flow made a detour to avoid the system and the middle-level airflow did not evaporate water substances of the system. In the SL type, the cool pool that was intensified by the middle-level rear inflow strengthened the low-level convergence, enhancing the convections, although the rear inflow evaporated water substances.

Fig.2 Schematic illustration of the shapes and structures of the precipitation systems. Right column indicates the decrease rate of precipitation amount

Fig.1 Precipitation system simulated with the simplified environment condition. (a) Horizontal distribution of equivalent potential temperature when the middle-level wind direction is opposite to the low-level inflow. (b)Precipitation distribution when the directions of the middle-level wind and low-level inflow are same. (c) Same as (b) except for the middle-level direction perpendicular to the low-level inflow.

(a) (b) (c)

Model of real classification of meteorological parameters for regional prognosisS. Senotrusova, V. Svinuhov*

Far Eastern State University

30-61 Aleutskaya St., 690000 Vladivostok, RussiaE-mail:[email protected]

Offered model of short-term regional prognosis. The main idea of the method is topick out air mass of area investigation, using common meteorological variable and synopticalparameters. Thus, initial point of offered method is the classification of air mass. Certainlimits are put on meteorological variable quantity. Firstly, they are to be all to distinguish willair mass and, at the same time, they are to change a little within air mass itself. Secondly, theyare to be regularly observed by variable quantities in the net of the stations.

Presently initial information is usually given as a set of quantitative (numerical)indications. When transforming of initial information is the method of main components oftenused, which is will known in the meteorology as the method of decomposition into naturalorthogonal components [1]. Decomposition of fields according to space is usually given. Incontrast to that method in this paper decomposition was given according to time [1,2].Fulfilling this work the data of aerological station “Sad-gorod” (Garden-City) and table TM-1meteostation in Vladivostok (Primorie Territory) within 1975-1995 were used. Analysingthese findings many factors were under consideration. Such as: direction and speed wind,temperature and air humidity, pressure, temperature of dew-point on the mixing surface level,and speed of wind, temperature air humidity, pressure above 500-1500 meters sea-level.According to aerological station height of air layer shift was counted average daily. As aparameter showing vertical atmosphere steadiness difference altitude temperature and thetemperature 500 ms above surface of earth. Thus, matrix of meteorological parameters andsynoptical variable quantities measured by 20*174 was used. Decomposition of variablequantities into natural orthogonal functions selected eleven components, describing 86 percent of variable dispersion for cold season and 82 per sent for warm season. Coefficient ofdecomposition matrix results (measured 11*174) may be classified in groups by differentmethods. And finally, statistic method of objective classification was used in this work as wellfor revealing homogeneous local groups in certain point totality of multi-measured sign space.The task was being solved with the help of criteria searching, finding of it gave the possibilityto evaluate the homogeneity of totality under consideration. Analogy criterion by Pirson wasused in this work for classify of succession of variable quantities. Classify decompositioncoefficients according to this method let us receive ten groups for cold half year and 12groups for warm six months. Every of these groups can be characterized by the particularitiesof general circulation and complex of meteorological parameters. Mean quantity ofmeteorological variable quantities (such as: direction and wind speed, pressure, temperature,air humidity on the level of stretching surface, temperature, wind speed, pressure and airhumidity on the level 500 meters and 1,5 km above earth surface and others) were counted foreach group. Moreover, every group was depicted by typical synoptical situation let’s take anexample of a few typical synoptical situation describing. Let’s describe a few typicalsynoptical situations as an example. Synoptical situation in the South region of PrimorieTerritory for the first group is determined by south cyclone. Trajectory of centre cyclonedisplacement passes from Yellow Sea region along Japanese Islands. South regions ofPrimorsky Territory are under influence of the North originally and then North-West

periphery of cyclone. As a result, we could observe relative low temperatures (-11C, -15C) inthe investigated region. 10-15 m/sec speed North-East and North-Wind mainly. Sea airmasses of temperate latitudes dominate over the South of Primoriye. The other group ischaracterized by intensive cyclonic activity over Far-Eastern Seas. This is a crest of Asiananticyclone over Primorye. Thus, each group had the following characteristics: typical airmasses, meteorological parameters and certain synoptical situation. This approach may beuseful for pollution prognosis in the close land air mass and population morbidity in theindustrial city prognosis. Analysis give the possibility to pick out classters (and consequently,groups of synoptical situations and meteorological parameters) which greatly determine thehigh level atmosphere pollution and therefore the morbidity of population. Our investigationshows that in winter 3 synoptical situations from 10 being under study has high potential ofatmosphere pollution. Low height of mixing atmosphere layer in the morning 70 meters, lowwind speed at the earth surface (3,5m/s),little displacement of wind speed in the close landlayer 3,5 m/s and 7,4 m/s) high background of temperature at the earth surface (-3,8 C) are thetypical characteristics situation for these groups. Taking the second step, prognosis of acertain synoptical group gives the possibility to expand the results of out investigation forprognostic some meteorological components, potential of atmosphere pollution andpopulation morbidity, using equation of step by step regression or discriminant analyses. Tosolve the second and the third tasks for building prognosis equation it is necessary to use theinformation on the atmosphere pollution and population morbidity in the cities.

This approach is very effective for getting good results, which areimpossible to receive under traditional methods.They are:To identify specific synoptical situations and air mass, describe typical potentialof air pollution for these synoptical situations;To pick out the groups of dominating air mass, where the potential of pollutionespecially high;To identify the groups of parameters, characterizing the conditions of low andhigh potential of atmosphere pollution;To determine groups parameters, describing conditions of appearing close earthand slightly lifted inversion.

RFERENCES

1. Danltrey S. Principal components analisis. Concept and technigues inModern Geography. 1976,8/1.-51p.

2. Warld J. Hierarchical grouping to optimize an objective function// J.Amer.Statistic.Ass.,1963,58,pp.236-244.

Ensemble Forecasting of Tropical Cyclone Motion Using A Baroclinic Model

Xiaqiong Zhou1 2and Johnny C L Chan 2

1 Shanghai Typhoon Institute, China Meteorological Administration 2Laboratory for Atmospheric Research, Department of Physics and Materials Science

City University of Hong Kong (email: [email protected])

1. Introduction While some studies have been conducted in the ensemble forecasting (EF) of tropical cyclone (TC) motion (Aberson et al. 1995, 1998; Cheung and Chan 1999a, b, hereafter CCa and CCb respectively; Cheung 2001), much more efforts are needed to establish EF as a viable alternative to the traditional single solution from a numerical-weather-prediction model for TC motion forecast. In this paper, the lagged-averaged forecast (LAF) and regional breeding of growing modes (BGM) perturbation methods are applied to the baroclinic model of the University of New South Wales (Leslie et al. 1985) to estimate the effectiveness of these EF techniques in TC motion forecast. Five cases from three typhoons that occurred during the Tropical Cyclone Motion Experiment (TCM-90) are chosen for study.

2. Model and data The model is a hydrostatic primitive equation model with a horizontal grid spacing of 30 km and 24 internal vertical sigma levels. The modified Kuo and Kain-Fritsch cumulus parameterization schemes are used. The initial fields are taken from the TCM-90 dataset. The bogussing scheme of the Typhoon Model (TYM) of the Japan Meteorological Agency is used in some of experiments after filtering the analysis vortex. Forecast time is up to 72 h. The LAF perturbations are generated from the differences between the short-range forecasts. No vortex-environment flow separation is applied (Cheung 2001). The BGM perturbation scheme is similar to that of CCa, CCb. When employing the BGM technique, TC vortex–environment separation is applied (BGMV and BGME respectively). Six pairs of ensemble members are generated in all three methods: LAF, BGMV and BGME.

3. Forecast results a. Typhoon Yancy Typhoon Yancy of 1990 at the initial time of 00UTC 17 August (to be labeled as Y1700) is chosen as an example. In all three sets of experiments no bogus scheme is included because the analysis vortex seems to be well represented by the TYM-90 analysis. Most of the LAF members predicted the landfall of Yancy over Taiwan except one that passed to its north as the control (Fig. 1). Two of the landfall members showed a jump across the Central Mountain Range. The ensemble mean track is much closer to the best track than that in the control run. In the BGME and BGMV experiments, the mean forecast tracks also show an improvement over the control though landfall is not predicted (not shown).

Fig. 1. Ensemble tracks (thin lines) of Yancy (Y1700) using the LAF technique. Best track – closed circle, control - open circle, ensemble mean – triangle. Positions are plotted every 6 h. b. Ed and Flo 1) SENSITIVITY EXPERIMENTS The forecast track of Ed turns towards the south instead of a steady westward movement, resulting in a 1000-km position error by 72 h (Fig. 2). Possible factors contributing to the failure of the forecasts of Ed such as cumulus parameterization,

bogussing scheme and lateral boundary, the influences of Flo and subtropical ridge are examined from 00UTC 14 September 1990 and 00UTC 15 September 1990 (to be labeled as E1400 and E1500 for the Ed, and F1400 and F1500 for the Flo respectively).

Fig. 2. As Fig. 1 except for Ed and Flo (E1400, F1400). The tracks of Ed are surprisingly insensitive to the possible factors (not shown). The forecast errors of the environment, which includes the subtropical ridge and the outer flow of Flo, largely contribute to forecast failure of the movement of Ed. The weakening subtropical ridge over the South China is replaced by a continuously deepening trough. The northwesterly flow upstream of the westerly trough pushes Ed to the southeast. At the same time, the outer flows of Ed and Flo merge due to the lack of the interference of the subtropical ridge. When Flo moves to the northeast of Ed, the outer flow of Flo makes Ed accelerate and move to the southeast. 2) ENSEMBLE FORECASTS OF FLO AND ED The three EF techniques are performed on the cases of Ed and Flo. For Ed, no significant improvement is obtained because the forecast of the western part of the subtropical ridge is not improved when the analysis is perturbed by the LAF errors (Fig. 2). For F1400, the control run performs well, and the position errors remain below 100 km from 48 to 72 h. The mean track is close to the best track except that it has a larger movement speed.

Bogussing scheme is used in the BGMV and BGME experiments. In the EF experiments of Ed, although some members have more westward movement when the vortex or the environment is perturbed, the evolution of the environment in EF forecasts is not improved apparently. The ensemble mean is similar to that of the control. The forecast errors come not only from the errors of the initial state also from the model. According to the results of the sensitivity experiments and the EF experiments, the main problem in this study is the large forecast errors of Ed. The model cannot predict the evolution of the subtropical ridge correctly in any of experiments. It appears that the forecasts moved along with a certain attractor of the model.

Ed

Flo

References: Aberson, S. D., S. J. Lord, M. DeMaria,

and M. S. Tracton, 1995: Short-range ensemble forecasting of hurricane tracks. 21st Conf. Hurr. Trop. Meteor., Miami, Amer. Meteor. Soc., 494-497.

Aberson, S. D., M. A. Bender, and R. E. Tuleya, 1998: Ensemble forecasting of tropical cyclone tracks. 12th Conf. Num. Wea. Pred., Phoenix, Amer. Meteor. Soc., 290-292.

Cheung, K. K. W., 2001: Ensemble forecasting of tropical cyclone motion: comparison between regional bred modes and random perturbations. Meteor. Atmos. Phys., 78, 23-34.

Cheung, K. K. W. and J. C. L Chan, 1999a: Ensemble forecasting of tropical cyclone motion using a barotropic model. Part I: Perturbations for the environment. Mon. Wea. Rev. 127, 1229-1243.

Cheung, K. K. W. and J. C. L Chan, 1999b: Ensemble forecasting of tropical cyclone motion using a barotropic model. Part II: Perturbations of vortex. Mon. Wea. Rev., 127, 2617-2640.

Leslie, L. M., G. A. Mills, L. Logan, D. J. Gauntlett, G. A. Kelly, M. J. Manton, J. McGregor, and J. Sardie, 1985: A high resolution primitive equations NWP model for operations and research. Aust. Met. Mag., 33, 11-35.

the hydrodynamic short-range of the local weather forecasting · the hydrodynamic short-range of...

Documents